Volume 71, Number 6, November-December 2019
Siriraj Medical Journal
SMJ
ISSN 2629-995XE-ISSN 2228-8082
ORIGINAL ARTICLE
432
Molecular Subtypes of Ductal Carcinoma In Situ (DCIS)
Identified by Expression of ER, PR, HER2, and Ki-67
Warapan Numprasit, et al.
438 Effect of Age on Recurrence in Hormone Receptor-Positive
Breast Cancer, and Factors Significantly Associated with
Cancer Recurrence
Warapan Numprasit, et al.
446 Outcomes of the Recombinant Human Epidermal Growth
Factor Addition to Chloramphenicol Ointment for Facial
Burn Wound Healing
Pornprom Muangman, et al.
450 Decision-making Factors in Non-operative Management of
Zygomatic Fractures
Somphon Roeksomtawin, et al.
457 Retrospective Analysis of Outcomes in Elderly Patients
with Adenocarcinoma of Stomach and Esophagogastric
Junction Following Three Different Treatments
Thammawat Parakonthun, et al.
466 The Malignant Potential in Adult Choledochal Cysts; an
Awareness Issue
Pholasith Sangserestid, et al
472 Survival and Factors Predictive of Survival in Patients with
Thymic Carcinoma
Suparauk Geanphun, et al.
480 Impact of Accuracy of Preoperative Transthoracic
Echocardigraphy on Complex Congenital Heart Surgery in
Pediatrics
Vutthipong Sanphasitvong, et al..
486 Presence of Residual Venous Thrombus at Warfarin
Withdrawal: Is It a Predictor for Recurrence After a First
Episode of Symptomatic Provoked Proximal Deep Venous
Thrombosis in the Thai Population?
Kanin Pruekprasert, et al.
491 Endovascular Thrombectomy versus Open Surgical
Thrombectomy for Thrombosed Arteriovenous
Hemodialysis Graft
Nattawut Puangpunngam, et al.
499 Outcomes of Autogenous Snuffbox Radiocephalic
Arteriovenous Fistula-First Strategy for Hemodialysis Access
Banjerd Praditsuktavorn, et al.
506 Formulaic Prediction of Z-Plasty Outcomes Relative to Z
Scar Lengthening, Z Flap Tension, and Area of Distortion,
and Determination of the Multiple Z-Plasty Configurations
that Optimally Increase Z Scar Length, Decrease Z Flap
Tension, and Decrease Area of Distortion
Nutthawut Akaranuchat, et al.
www.smj.si.mahidol.ac.th
E-mail: sijournal@mahidol.ac.th
International Association of Surgeons
Gastroenterologists & Oncologists
Thailand Chapter
Thai Association for Gastrointestinal
Endoscopy
Indexed by
SURGERY
SIRIRAJ MEDICAL JOURNAL
First Editor: Ouay Ketusinh Emeritus Editors: Somchai Bovornkitti, Adulya Viriyavejakul, Sommai Toongsuwan,
Nanta Maranetra, Niphon Poungvarin, Prasit Watanapa, Vithya Vathanophas, Pipop Jirapinyo, Sanya Sukpanichnant,
Somboon Kunathikom
Executive Editor: Prasit Watanapa Editorial Director: Manee Rattanachaiyanont
Managing Editor: Gulapar Srisawasdi, Chenchit Chayachinda
Editor-in-Chief: awatchai Akaraviputh
Associate Editor: Varut Lohsiriwat, Prapat Wanitpongpan Online Editor: Puttinun Patpituck
SIRIRAJ MEDICAL JOURNAL is published bimonthly, 6 issues a year (Jan-Feb, Mar-Apr, May-Jun, Jul-Aug, Sep-Oct and Nov-Dec)
and distributed by the end of the last month of that issue.
SIRIRAJ MEDICAL JOURNAL is listed as a journal following the Uniform Requirements for Manuscripts Submitted to Biomedical Journals (URM)
by the International Committee of Medical Journal Editors (ICMJE) since 9 July 2010 [http://www.icmje.org/journals.html].
Philip Board (Australian National University, Australia)
Richard J. Deckelbaum (Columbia University, USA)
Yozo Miyake (Aichi Medical University, Japan)
Yik Ying Teo (National University of Singapore, Singapore)
Harland Winter (Massachusetts General Hospital, USA)
Philip A. Brunell (State University of New York At Bualo, USA)
Noritaka Isogai (Kinki University, Japan)
Yuji Murata (Aizenbashi Hospital, Japan)
Keiichi Akita (Tokyo Medical and Dental University Hospital, Japan)
Shuji Shimizu (Kyushu University Hospital, Japan)
David S. Sheps (University of Florida, USA)
Robin CN Williamson (Royal Postgraduate Medical School, UK)
Tai-Soon Yong (Yonsei University, Korea)
Anusak Yiengpruksawan (e Valley Robotic Institute, USA)
Stanlay James Rogers (University of California, San Francisco, USA)
Kyoichi Takaori (Kyoto University Hospital, Japan)
Tomohisa Uchida (Oita University, Japan)
Yoshiki Hirooka (Nagoya University Hospital, Japan)
Hidemi Goto (Nagoya University Graduate School of Medicine, Japan)
Kazuo Hara (Aichi Cancer Center Hospital, Japan)
Shomei Ryozawa (Saitama Medical University, Japan)
Christopher Khor (Singapore General Hospital, Singapore)
Yasushi Sano (Director of Gastrointestinal Center, Japan)
Mitsuhiro Kida (Kitasato University & Hospital, Japan)
Seigo Kitano (Oita University, Japan)
Ichizo Nishino (National Institute of Neuroscience NCNP, Japan)
Masakazu Yamamoto (Tokyo Womens Medical University, Japan)
Dong-Wan Seo (University of Ulsan College of Medicine, Korea)
George S. Baillie (University of Glasgow, UK)
G. Allen Finley (Delhousie University, Canada)
Sara Schwanke Khilji (Oregon Health & Science University, USA)
Matthew S. Dunne (Institute of Food, Nutrition, and Health, Switzerland) 
Marianne Hokland (University of Aarhus, Denmark)
Marcela Hermoso Ramello (University of Chile, Chile)
Ciro Isidoro (University of Novara, Italy)
Moses Rodriguez (Mayo Clinic, USA)
Robert W. Mann (University of Hawaii, USA)
Wikrom Karnsakul (Johns Hopkins Childrens Center, USA)
Frans Laurens Moll (University Medical Center Ultrecht, Netherlands)
James P. Dolan (Oregon Health & Science University, USA)
John Hunter (Oregon Health & Science University, USA)
Nima Rezaei (Tehran University of Medical Sciences, Iran)
Dennis J. Janisse (Subsidiary of DJO Global, USA)
Folker Meyer (Argonne National Laboratory, USA)
David Wayne Ussery (University of Arkansas for Medical Sciences, USA)
Intawat Nookaew (University of Arkansas for Medical Sciences, USA)
Victor Manuel Charoenrook de la Fuente 
(Centro de Oalmologia Barraquer, Spain)
Karl omas Moritz
(Swedish University of Agricultural Sciences, Sweden)
Nam H. CHO (University School of Medicine and Hospital, Korea)
www.smj.si.mahidol.ac.th
SMJ
Statistician: Saowalak Hunnangkul (Mahidol University, ailand)
Medical Illustrator: Chananya Hokierti (Nopparat Rajathanee Hospital, ailand)
Online Assistant: Surang Promsorn, Wilailuck Amornmontien, Hatairat Ruangsuwan Editorial Oce Secretary: Amornrat Sangkaew
International Editorial Board
Editorial Board
Watchara Kasinrerk (Chiang Mai University, ailand)
Rungroj Krittayaphong (Siriraj Hospital, Mahidol University, ailand)
Wiroon Laupattrakasem (Khon Kaen University, ailand)
Anuwat Pongkunakorn (Lampang Hospital, ailand)
Nopporn Sittisombut (Chiang Mai University, ailand)
Vasant Sumethkul (Ramathibodi Hospital, Mahidol University, ailand)
Yuen Tanniradorm (Chulalongkorn University, ailand)
Saranatra Waikakul (Siriraj Hospital, Mahidol University, ailand)
Pa-thai Yenchitsomanus (Siriraj Hospital, Mahidol University, ailand)
Surapol Issaragrisil (Siriraj Hospital, Mahidol University,ailand)
Jaturat Kanpittaya (Khon Kaen University, ailand)
Suneerat Kongsayreepong (Siriraj Hospital, Mahidol University, ailand)
Pornchai O-Charoenrat (Siriraj Hospital, Mahidol University, ailand)
Nopphol Pausawasdi (Siriraj Hospital, Mahidol University, ailand)
Supakorn Rojananin (Siriraj Hospital, Mahidol University, ailand)
Jarupim Soongswang (Siriraj Hospital, Mahidol University, ailand)
Suttipong Wacharasindhu (Chulalongkorn University, ailand)
Prapon Wilairat (Mahidol University, ailand)
Pornprom Muangman (Siriraj Hospital, Mahidol University, ailand)
Ampaiwan Chuansumrit
(Ramathibodi Hospital, Mahidol University, ailand)
Sayomporn Sirinavin
(Ramathibodi Hospital, Mahidol University, ailand)
Vitoon Chinswangwatanakul
(Siriraj Hospital, Mahidol University, ailand)
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
432
Original Article
SMJ
Warapan Numprasit, M.D.*, Irin Chowchankit, M.D.*, Suebwong Chuthapisith, M.D.*, Malee Warnnissorn,
M.D.**, Waraporn Imruetaicharoenchoke, M.D.*
*Department of Surgery, **Department of Pathology, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok 10700, ailand.
Molecular Subtypes of Ductal Carcinoma In Situ
(DCIS) Identied by Expression of ER, PR, HER2,
and Ki-67
ABSTRACT
Objective: e estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2
(HER2), and Ki-67 are used to classify invasive breast cancer into various subtypes. In the case of ductal carcinoma
in situ (DCIS), however, the information relating to subtypes using these markers is still limited.
Methods: e pathological specimens of 267 patients diagnosed with DCIS at Siriraj Hospital were analyzed. By
using the expressions of ER, PR, HER2, and Ki-67, breast cancer patients were classied into the ve molecular
subtypes: luminal A, luminal B/HER2 negative, luminal B/HER2 positive, HER2 overexpression, and triple-negative.
Based on the specic molecular subtypes, age, clinical presentation, tumor size, and tumor grade were analyzed
separately using univariate analysis.
Results: 135 (50.6%), 1 (0.4%), 58 (21.7%), 59 (22.1%), and 14 (5.2%) DCIS were luminal A, luminal B/HER2 negative,
luminal B/HER2 positive, HER2 overexpression, and triple-negative, respectively. Patients with luminal A DCIS
signicantly presented with mass, compared to the other subtypes that displayed with mass and microcalcication
in combination (p = 0.008). e luminal A subtype was also associated with tumors of a smaller size (less than 2
cm; p = 0.007) and a lower nuclear grade (p = 0.000) than the HER2 overexpression subtype.
Conclusion: Although luminal A DCIS was the most common DCIS subtype of ai women, HER2 overexpression
DCIS was signicantly correlated with poor clinicopathological features, including a large tumor size and a high
nuclear grade, which are recognized prognostic factors for tumor recurrence.
Keywords: Clinicopathological features; ductal carcinoma in situ; molecular subtypes; prevalence (Siriraj Med J
2019; 71: 432-437)
Corresponding author: Waraporn Imruetaicharoenchoke
E-mail: waraporn.imr@mahidol.ac.th
Received 31 May 2019 Revised 14 August 2019 Accepted 23 September 2019
ORCID ID: http://orcid.org/0000-0003-1566-5307
http://dx.doi.org/10.33192/Smj.2019.64
INTRODUCTION
DNA microarray proling of invasive breast cancer
is used to classify breast cancers into distinct subtypes that
are explicitly associated with oncological outcomes, and to
identify therapeutic options.
1
In addition, various studies
have conrmed that immunohistochemical staining of
paran sections are well-correlated with, and are used
as reliable surrogate markers for, molecular subtype
classications categorized by DNA microarrays.
1
e
staining of immunohistochemistry (IHC) applied to
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
433
Numprasit et al.
pathological specimens includes antibodies for estrogen
receptor (ER), progesterone receptor (PR), human epidermal
growth factor receptor 2 (HER2), and Ki-67.
2
Based on the protein expressions of these IHC
markers, invasive breast cancer subtypes are categorized
into luminal A, luminal B/HER2 negative, luminal B/
HER2 positive, HER2 overexpression, and triple-negative.
2
Multiple previous studies have shown that luminal A has
the best prognosis and the lowest rate of locoregional
recurrence, whereas the HER2 overexpression and triple-
negative subtypes are strongly associated with a high
risk of regional recurrence and distant metastases.
3-5
Ductal carcinoma in situ (DCIS), which accounts
for 20%–25% of new breast cancer diagnoses, is the
proliferation of malignant mammary ductal epithelial cells
conned within the basement membrane.
6
Identication
of the predictive and prognostic markers associated with
the biology of DCIS would provide better therapeutic
options and oncological outcomes.
However, the study of subtype classications in DCIS
has been limited to date. is study primarily aimed to
establish the incidence of DCIS sorting by individual
molecular subtypes. e secondary aim was to identify
the relationship between each DCIS subtype and the
associated clinicopathological features and outcomes
in ai women.
MATERIALS AND METHODS
Data collection
A retrospective review was performed of pathological
specimens obtained from patients diagnosed with pure
DCIS at the Department of Surgery, Faculty of Medicine
Siriraj Hospital, between December 2006 and September
2012. A total of 267 pathological specimens were collected.
Protein expressions of ER, PR, HER2, and Ki-67 were
detected by IHC. Pathological specimens containing
microinvasions, extensive intraductal components
and nodal metastasis were excluded. Information on
the patient and tumor characteristics (comprising age,
clinical presentation, tumor size, and tumor grading)
was collected. is study was approved by the Ethics
Committee of the Siriraj Institutional Review Board,
Faculty of Medicine Siriraj Hospital (Si 435/2010).
Molecular subtype classication
Based on expressions of ER, PR, HER2, and Ki-67,
the tumors were classied into the following subtypes:
luminal A: ER positive, PR positive, HER2 negative,
and Ki-67 < 14%;
luminal B/HER2 negative: ER positive, PR positive,
HER2 negative, and Ki-67 > 14%;
luminal B/HER2 positive: ER positive, any PR,
HER2 positive, and any Ki-67;
HER2 overexpression: ER negative, PR negative,
and HER2 positive; and
triple-negative: all ER, PR, and HER2 negative.
Immunohistochemistry (IHC)
IHC was performed on 3-µm formalin-xed, paran-
embedded, pathological specimens. e expression of
each marker was detected by IHC using rabbit anti-ER
monoclonal antibody (ready-to-use; clone SP1, Ventana
Laboratories, Tucson, AZ); rabbit anti-PR monoclonal
antibody (ready-to-use; clone 1E2, Ventana Laboratories,
Tucson, AZ); rabbit anti-HER2 monoclonal antibody
(ready-to-use; clone 4B5, Ventana Laboratories, Tucson,
AZ); and rabbit anti-Ki-67 monoclonal antibody (1:300
dilution; clone MIB-1, DAKO Laboratories, Carpinteria,
CA).
Interpretation of staining
According to the American Society of Clinical
Oncology and the College of American Pathologists, ER
and PR assays are considered positive if there is at least
1% positive tumor nuclei staining in the samples.
7
HER2
assays are considered positive if there is IHC staining
of 3+ (uniform, with intense membrane staining of
> 30% of the tumor cells), but it is considered negative
if there is IHC staining of 0 (no immunostaining) or
1+ (weak immunostaining, with < 30% of the tumor
cells stained). However, IHC staining of 2+ (complete
membrane staining, with either uniform or weak staining
in at least 10% of the cells) is considered equivocal.
8
In
this study, we did not perform in situ hybridization in all
equivocal HER2 (2+); instead, we considered equivocal
HER2 as the HER2 negative subtype. In accordance with
the invasive breast cancer subtype classication from St.
Gallen in 2011, Ki-67 was used to classify patients into 2
subtypes: low Ki-67 < 14% as luminal A, and high Ki-67
≥ 14% as luminal B/HER2 negative.
9
Statistical analysis
A chi-squared test was used for the binary variables
and an analysis of the variance of the continuous variables
in order to compare the clinicopathological characteristics
of the four subtypes. Univariate logistic regression
was used to determine any associations between each
molecular subtype and age, clinical presentation, tumor
size, and tumor grade. A p-value < 0.05 was considered
to be statistically signicant. All statistical analyses were
performed by SPSS Statistics for Windows, version 15.0
(SPSS Inc., Chicago, Ill., USA).
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
434
Original Article
SMJ
RESULTS
e population baseline characteristics are described
in Table 1. Of 267 patients with DCIS, the majority
(86.5%) were over 40 years of age. e DCIS patients
more commonly presented with a combination of mass
and microcalcication (36.7%), a tumor size of generally
less than 2 cm (60.7%), and a high nuclear-grade tumor
(52.4%). The results of the IHC staining are shown
in Table 2. e number of positive specimens for ER,
PR, and HER2 was 189 (70.8%), 181 (67.8 %), and 117
(43.8%), respectively (Table 2). Aer Ki-67 staining was
performed on the 136 patients with ER/PR positive and
HER2 negative, only 1 patient (0.7%) had high Ki-67
(≥ 14%). According to the ER, PR, HER2, and Ki-67
statuses, the DCIS was divided into 4 subtypes (Table 3),
with around half of the patients (50.6%) classied as
luminal A. Table 4 illustrates the correlations between each
tumor subtype and the clinicopathological characteristics
determined by a univariate analysis. In terms of age, there
were no dierences between each molecular subtype.
Moreover, the patients diagnosed with the luminal A
subtype signicantly exhibited mass, in contrast to the
other subtypes, which presented with mass combined
with microcalcication (p = 0.008). e tumor sizes of
the luminal A and B groups were signicantly smaller
than those of DCIS carrying HER2 overexpression (p =
0.007). In addition, approximately three quarters (72.6%)
of the luminal A DCIS had a signicantly low nuclear
grade, whereas the clear majority of the pathological
specimens with the other subtypes (91.5% of those with
the HER2 overexpression, 85.7% of the triple-negative,
100% of the luminal B/HER2 negative, and 63.8% of
the luminal B/HER2 positive) displayed a high nuclear
tumor grade (p < 0.000).
DISCUSSION
Ductal carcinoma in situ, a heterogenous disease,
has similar molecular proles to invasive breast cancer by
genetic proling.
10
IHC staining has replaced microarray
genetic testing to classify breast cancer into 4 subtypes,
based on the expressions of ER, PR, HER2, and Ki-67.
ese surrogate markers can be used to establish the
treatment and prognosis of breast cancer.
e present study demonstrated that luminal A DCIS
was the most common subtype (50.6%), followed by HER2
overexpression (22.1%), luminal B/HER2 positive (21.7%),
triple-negative (5.2%), and luminal B/HER2 negative
(0.4%). e DCIS incidences in our study were similar
to those found by two previous studies, with luminal A
DCIS being ranked rst.
11,12
However, the current study
found that the HER2 overexpression subgroup had a
TABLE 1. Population characteristics of the 256 patients
included in this study.
TABLE 2. Results of immunohistochemistry staining
of all 267 patients in this study.
Characteristics Number (%)
Age, n (%)
≤40years 36(13.5)
>40years 231(86.5)
Clinicalpresentation
Mass 76(28.5)
Microcalcication 93(34.8)
Combination 98(36.7)
Tumorsize
≤2cm 162(60.7)
>2cmto≤5cm 83(31.1)
>5cm 22(8.2)
Tumorgrading
Lowgrade 38(14.2)
Intermediategrade 88(33.0)
Highgrade 141(52.8)
IHC study Number (%)
ER
Positive 189(70.8)
Negative 78(29.2)
PR
Positive 181(67.8)
Negative 86(32.2)
HER2
Positive 117(43.8)
Negative 150(56.2)
Ki-67(total136)
Low(<14%) 135(99.3)
High 1(0.7)
Abbreviations: ER, estrogen receptor; HER2, human epidermal
growth factor 2; PR, progesterone receptor
TABLE 3. Tumors classied according to molecular
subtype.
Subtype Number (%)
LuminalA 135(50.6)
LuminalB/HER2negative 1(0.4)
LuminalB/HER2positive 58(21.7)
HER2overexpression 59(22.1)
Triple-negative 14(5.2)
Total 267
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
435
TABLE 4. Univariate analysis to determine the associations between the clinicopathological features and subtypes.
Age (n%)
≤40years 16(11.9) 0(0) 10(17.2) 9(15.3) 1(7.1) 0.669
>40years 119(88.1) 1(100) 48(82.8) 50(84.7) 13(92.9)
Clinicalpresentation(n%) 0.008
Mass 47(34.8) 0(0) 15(25.9) 11(18.6) 3(21.4)
Microcalcication 54(40.0) 0(0) 18(31.0) 18(30.5) 3(21.4)
Combination 34(25.2) 1(100) 25(43.1) 30(50.8) 8(57.1)
Tumorsize(n%) 0.007
≤2cm 93(68.9) 1(100) 36(62.1) 25(42.4) 7(50.0)
>2cmto≤5cm 34(25.2) 0(0) 18(31.0) 24(40.7) 7(50.0)
>5cm 8(5.9) 0(0) 4(6.9) 10(16.9) 0(0)
Tumorgrade(n%)
Lowgrade 35(25.9) 0(0) 2(3.4) 1(1.7) 0(0) <0.001
Intermediategrade 63(46.7) 0(0) 19(32.8) 4(6.8) 2(14.3)
Highgrade 37(27.4) 1(100) 37(63.8) 54(91.5) 12(85.7)
* p<0.05 = signicant
Clinicopathological Luminal A Luminal Luminal HER2 Triple-
Variables (n = 135) B/HER2 B/HER2 overexpression negative P-value*
(%) negative positive (n = 59) (n = 14)
(n = 1) (n = 58) (%) (%)
(%) (%)
slightly higher proportion than the corresponding gures
reported by those other publications (13% and 16%).
e dierences in the DCIS subtypes might relate to
the inequity of their prognoses; likewise, the dierences
in the molecular subtypes might predict the prognoses of
invasive breast cancers. Although the present study could
not nd any relationships between the DCIS subtypes
and prognoses due to this not being included in our
objectives, many studies, such as that by Williams et al.,
have demonstrated that luminal A DCIS had a lower
recurrence either locoregionally or in regard to distance
than the other subgroups (7.6% vs 15.5%–36.1%).
13
Moreover, HER2 overexpression in DCIS has been found
to be an independent prognostic factor for invasive
recurrence compared to the luminal A subtype (HR
17.8; 95% CI 2.14–148; p = 0.008).
13,14
e correlations between the clinicopathological
features and DCIS dier for the various subtypes.
15,16
In our
study, despite mass associated with microcalcication being
a major complaint of DCIS patients, the luminal A DCIS
subgroup signicantly presented with a solitary, smaller
mass. is phenomenon implies that the aggressiveness
of the disease is related to microcalcication. However,
this study could not identify the characteristics of
microcalcication, such as comedonecrosis, which is
used to determine DCIS prognoses. In addition, tumor
size was an important factor, with approximately 65%
of the luminal DCIS patients having a smaller-sized
tumor than the patients in the other subgroups (< 2 cm).
is nding suggests that the less invasive properties
of luminal DCIS, with its smaller tumor size, present a
lower risk of recurrence.
15
In addition, we found that a high nuclear grade was
mostly associated with the luminal B, triple-negative, and
in particular, HER2 overexpression DCIS subgroups. To
our knowledge, nuclear grade is an important factor,
with a higher grade correlating with a higher risk of
recurrence and thus being associated with poor molecular
Numprasit et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
436
Original Article
SMJ
subtypes.
17
Recent publications have reported that a high
nuclear grade was signicantly found in non-luminal A
subtypes, and it was considered to be an independent
predictor for overall recurrence.
15,16
In this study, we could prove a relationship between the
molecular subtypes and the clinicopathological backgrounds.
All of these ndings could predict the prognosis of DCIS;
for instance, luminal A DCIS tended to have better
clinicopathological factors. We acknowledge that patient
age, tumor size, comedonecrosis, nuclear grade, surgical
margin, and adjuvant treatment are mandatory factors
for recurrence.
18
e goal of treatment should therefore
be to decrease the opportunities for disease relapse.
However, the most suitable DCIS treatment approach
has been long debated, particular in the case of breast
conserving surgery (BCS).
19
Heretofore, there has been
strong evidence of the benet of using radiation to decrease
local recurrence, regardless of the DCIS subgroup.
20,21
On the other hand, recent reports have challenged this
standard treatment by omitting radiation from the treatment
given to some patients in light of the growing knowledge
of the distinct dierences in the risk of recurrence of each
DCIS clinicopathological subgroup.
22
Unfortunately,
healthcare professional teams give little consideration
to the DCIS molecular subtypes when making treatment
decisions, even though breast cancer treatment is tending
to become more personalized.
Furthermore, in the near future, decision-making
about the treatment approach for DCIS patients will
not only depend on the clinicopathological parameters
or molecular subtypes, but will also include multigene
assays, which are more specic to individuals. As an
example, Oncotype Dx can estimate the 10-year risk
of locoregional recurrence based on an analysis of 12
genes.
23
is can be useful for escalating or de-escalating
the treatment for DCIS patients.
This study has several limitations. Firstly, its
retrospective design presented inherent limitations,
such as missing data. In addition, the positivity of HER2
depended only on IHC; HER2 2+ on IHC was not further
conrmed with in situ hybridization, and we used the
early denition of Ki-67 score which have altered from
present recommendation, however, it was slightly change
in DCIS incidence but not eect to other study results.
Lastly, as there were no available data on recurrence,
the eects of the DCIS molecular subtypes on the risks
of locoregional recurrence could not be evaluated.
CONCLUSION
In our study, luminal A DCIS was the main DCIS
molecular subtype and was related to less aggressive
clinicopathological features. Moreover, the HER2
overexpression subtype was mostly found concomitant
with a larger tumor size and a higher nuclear grade. e
use of the ER, PR, HER2, and Ki-67 statuses on IHC
may suggest the need for further aggressive therapy for
some particular subtypes of DCIS.
ACKNOWLEDGMENTS
is study was nancially supported by the Siriraj
Hospital Research Fund of the Faculty of Medicine.
e authors thank to Mr. Suthipol Udompunthurak for
his excellent assistance with the statistical analyses. All
authors declare that there are no conicts of interest.
REFERENCES
1. Sandhu R, Parker JS, Jones WD, Livasy CA, Coleman WB.
Microarray-Based Gene Expression Proling for Molecular
Classication of Breast Cancer and Identication of New
Targets for erapy. Laboratory Medicine 2010;41:364-72.
2. Goldhirsch A, Wood WC, Coates AS, Gelber RD, ürlimann
B, Senn H-J, et al. Strategies for subtypes—dealing with the
diversity of breast cancer: highlights of the St Gallen International
Expert Consensus on the Primary erapy of Early Breast
Cancer 2011. Ann Oncol 2011;22:1736-47.
3. Braunstein LZ, Taghian AG, Niemierko A, Salama L, Capuco
A, Bellon JR, et al. Breast-cancer subtype, age, and lymph node
status as predictors of local recurrence following breast-conserving
therapy. Breast Cancer Res Treat 2017;161:173-9.
4. Hennigs A, Riedel F, Gondos A, Sinn P, Schirmacher P, Marme F,
et al. Prognosis of breast cancer molecular subtypes in routine
clinical care: A large prospective cohort study. BMC Cancer
2016;16:734.
5. Fallahpour S, Navaneelan T, De P, Borgo A. Breast cancer
survival by molecular subtype: a population-based analysis of
cancer registry data. CMAJ Open 2017;5:E734-E9.
6. Kerlikowske K. Epidemiology of ductal carcinoma in situ. J
Natl Cancer Inst Monogr 2010;2010:139-41.
7. Hammond ME, Hayes DF, Dowsett M, Allred DC, Hagerty KL,
Badve S, et al. American Society of Clinical Oncology/
College of American Pathologists guideline recommendations
for immunohistochemical testing of estrogen and progesterone
receptors in breast cancer (unabridged version). Arch Pathol
Lab Med 2010;134:e48-72.
8. Wol AC, Hammond ME, Schwartz JN, Hagerty KL, Allred
DC, Cote RJ, et al. American Society of Clinical Oncology/
College of American Pathologists guideline recommendations
for human epidermal growth factor receptor 2 testing in breast
cancer. Arch Pathol Lab Med 2007;131:18-43.
9. Goldhirsch A, Wood WC, Coates AS, Gelber RD, ürlimann
B, Senn HJ. Strategies for subtypes--dealing with the diversity
of breast cancer: highlights of the St. Gallen International
Expert Consensus on the Primary erapy of Early Breast
Cancer 2011. Ann Oncol 2011;22:1736-47.
10. Muggerud AA, Hallett M, Johnsen H, Kleivi K, Zhou W,
Tahmasebpoor S, et al. Molecular diversity in ductal carcinoma
in situ (DCIS) and early invasive breast cancer. Mol Oncol
2010;4:357-68.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
437
11. Zhou W, Jirström K, Amini RM, Fjällskog ML, Sollie T, Lindman
H, et al. Molecular subtypes in ductal carcinoma in situ of the
breast and their relation to prognosis: a population-based
cohort study. BMC Cancer 2013;13:512.
12. Tamimi RM, Baer HJ, Marotti J, Galan M, Galaburda L, Fu
Y, et al. Comparison of molecular phenotypes of ductal carcinoma
in situ and invasive breast cancer. Breast Cancer Res 2008;10:R67.
13. Williams KE, Barnes NL, Cramer A, Johnson R, Cheema K,
Morris J, et al. Molecular phenotypes of DCIS predict overall
and invasive recurrence. Ann Oncol 2015;26:1019-25.
14. Di Cesare P, Pavesi L, Villani L, Battaglia A, Da Prada GA,
Riccardi A, et al. e Relationships between HER2 Overexpression
and DCIS Characteristics. Breast J 2017;23:307-14.
15. Van Bockstal M, Lambein K, Denys H, Braems G, Nuyts A,
Van den Broecke R, et al. Histopathological characterization
of ductal carcinoma in situ (DCIS) of the breast according to
HER2 amplication status and molecular subtype. Virchows
Arch 2014;465:275-89.
16. Curigliano G, Disalvatore D, Esposito A, Pruneri G, Lazzeroni
M, Guerrieri-Gonzaga A, et al. Risk of subsequent in situ and
invasive breast cancer in human epidermal growth factor receptor
2-positive ductal carcinoma in situ. Ann Oncol 2015;26:682-7.
17. Shoker BS, Sloane JP. DCIS grading schemes and clinical
implications. Histopathology 1999;35:393-400.
18. Silverstein MJ. e University of Southern California/Van
Nuys prognostic index for ductal carcinoma in situ of the
breast. Am J Surg 2003;186:337-43.
19. Toss M, Miligy I, ompson AM, Khout H, Green AR, Ellis
IO, et al. Current trials to reduce surgical intervention in ductal
carcinoma in situ of the breast: Critical review. Breast (Edinburgh,
Scotland) 2017;35:151-6.
20. Donker M, Litiere S, Werutsky G, Julien JP, Fentiman IS, Agresti
R, et al. Breast-conserving treatment with or without radiotherapy
in ductal carcinoma In Situ: 15-year recurrence rates and
outcome aer a recurrence, from the EORTC 10853 randomized
phase III trial. J Clin Oncol 2013;31:4054-9.
21. Wapnir IL, Dignam JJ, Fisher B, Mamounas EP, Anderson SJ,
Julian TB, et al. Long-term outcomes of invasive ipsilateral
breast tumor recurrences aer lumpectomy in NSABP B-17
and B-24 randomized clinical trials for DCIS. J Natl Cancer
Inst 2011;103:478-88.
22. Solin LJ, Gray R, Hughes LL, Wood WC, Lowen MA, Badve
SS, et al. Surgical Excision Without Radiation for Ductal
Carcinoma in Situ of the Breast: 12-Year Results From the
ECOG-ACRIN E5194 Study. J Clin Oncol 2015;33:3938-44.
23. Manders JB, Kuerer HM, Smith BD, McCluskey C, Farrar WB,
Frazier TG, et al. Clinical Utility of the 12-Gene DCIS Score
Assay: Impact on Radiotherapy Recommendations for Patients
with Ductal Carcinoma In Situ. Ann Surg Oncol 2017;24:660-8.
Numprasit et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
438
Original Article
SMJ
Warapan Numprasit, M.D.*, Chaninporn Saengsri, M.D.*, Ng char Hong, MBChB, MS.**, Pongthep Pisarnturakit,
M.D.*
*Department of Surgery, Faculty of Medicine Siriraj Hospital Mahidol University, Bangkok 10700, ailand, **Sunway Medical Centre, Malaysia.
Effect of Age on Recurrence in Hormone Receptor-
Positive Breast Cancer, and Factors Signicantly
Associated with Cancer Recurrence
ABSTRACT
Objective: Age is an important factor for predicting survival, with worse prognosis among young women compared
to middle-aged women. Hormone receptor-positive breast cancer is the most commonly reported diagnoses among
younger women. is study aimed to investigate the eect of age on recurrence in hormone receptor-positive breast
cancer patients, and to identify its signicant related factors.
Methods: Operable hormone receptor-positive breast cancer patients who underwent surgery at the Division
of Head, Neck, and Breast Surgery, Department of Surgery, Faculty of Medicine Siriraj Hospital, Mahidol
University, Bangkok, ailand during 2008-2013 were retrospectively recruited. Age at diagnosis, follow-up
time, staging, tumor characteristics, treatment, and date of recurrence were collected, recorded, and analyzed.
Results: Of the 431 patients that were included, 145 patients were aged 40 years or younger, and 286 patients
were aged older than 40 years. e median follow-up time was 4.1 years. In multivariate analysis, the unadjusted
recurrence rate was higher in T3-4, node positive, high pathological grade, and lymphovascular invasion. Aer
adjusting for age, only N stage N1 remained statistically signicant (hazard ratio [HR]: 2.75, 95% condence interval
[CI]: 1.18-6.40; p=0.19). e recurrence rate was found to be non-signicantly higher in younger patients than in
older patients (11% vs. 5.6%, p=0.21).
Conclusion: e results of this study revealed no signicant dierence between age groups for recurrence-free
survival in women with hormone receptor-positive breast cancer; however, younger women did demonstrate a
higher rate of recurrence. N stage N1 is an independent predictor of cancer recurrence.
Keywords: Age; hormone receptor-positive breast cancer; recurrence (Siriraj Med J 2019; 71: 438-445)
Corresponding author: Pongthep Pisarnturakit
E-mail: pongthep.pis@mahidol.ac.th
Received 31 May 2019 Revised 19 November 2019 Accepted 20 November 2019
ORCID ID: http://orcid.org/0000-0002-2047-0849
http://dx.doi.org/10.33192/Smj.2019.65
INTRODUCTION
Breast cancer is the most common cancer among
women in ailand. Even though breast cancer more
oen occurs in women over age 50, it also aects younger
aged women. In 2018, there were 12,770 cases of breast
cancer diagnosed in women under the age of 40 in the
US.
1
Age is an important factor for predicting survival,
with worse prognosis among young women compared to
middle-aged women.
2
Some studies reported that breast
cancer among younger-aged women is usually found at a
later stage and that it has more aggressive subtypes, such
as triple negative and HER2 overexpression.
3,4
However
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
439
luminal subtype or hormone receptor positive breast
cancer has been and continues to be the most common
diagnosis in this younger age group.
Hormone receptor-positive breast cancer has better
prognosis relative to survival and recurrence when
compared to other types of breast cancer.
5,6
us, the
hypothesis that the observed poor survival among young
women depends solely on cancer subtype is questionable.
Previous studies reported that hormone receptor status
aects survival outcome in early-onset breast cancer,
but their results are controversial. One large cohort
study found that younger aged patients (≤40 years) with
luminal A subtype had poorer disease-free survival and
poorer distant metastasis-free survival than patients
in the middle-aged group (41-60 years).
7
In contrast,
Fallahpour, et al. found mortality rate increase with
increasing age regardless of molecular subtype.
6
e aim of this study was to investigate the eect of
age on recurrence in hormone receptor-positive breast
cancer patients, and to identify factors signicantly
associated with cancer recurrence.
MATERIALS AND METHODS
Study design and population
is retrospective study included operable hormone
receptor-positive breast cancer patients who underwent
surgery at the Division of Head, Neck, and Breast Surgery,
Department of Surgery, Faculty of Medicine Siriraj
Hospital, Mahidol University, Bangkok, ailand during
the 2008-2013 study period. All included patients had
histologically-proven hormone receptor-positive breast
cancer. Patients with metastatic disease, with history of
neoadjuvant treatment, and/or with less than 2 years
of follow-up were excluded. e protocol for this study
was approved by the Siriraj Institutional Review Board
(Si 658/2016), and the requirement to obtain written
informed consent was waived due to the retrospective
design of this study.
Data collection
Age at diagnosis, date of last follow-up, pathological
staging, tumor characteristics, surgical treatment, adjuvant
treatment, and date of recurrence diagnosis were collected,
recorded, and analyzed.
Statistical analysis
Bivariate analysis of age, stage, tumor characteristics,
surgical treatment, adjuvant treatment, and recurrence
was performed. Statistical analyses were performed using
SPSS for Windows version 22.0 (SPSS, Inc.; IBM Corp.,
Armonk, New York, USA). Categorical variables were
compared using chi-square test or Fisher’s exact test.
Hazard ratios were used to compare recurrence between
groups. Multivariate analysis by Cox proportional hazards
model was performed to identify independent predictors
of disease recurrence. Survival analysis was performed
using the Kaplan-Meier method, with comparisons
between groups performed using log-rank analysis.
A result was considered statistically signicant it its
pvalue was less than 0.05.
RESULTS
Demographics
Of the 431 patients that were included, 145 patients
were aged 40 years or younger, and 286 patients were aged
older than 40 years. e median follow-up time was 4.1
years. ere was no signicant dierence between age
groups for T stage, N stage, lymphovascular invasion, or
hormone receptor status. e recurrence rate was found
to be non-signicantly higher in younger patients than
in older patients. Pathological grade was signicantly
higher in the younger age group than in the older age
group (p<0.001) (Table 1).
Recurrence-free survival
Recurrence-free survival at 5 years in patients aged
≤40 years was 87.3% compared with 93.1% among patients
aged >40 years (p=0.21) (Fig 1). Recurrence-free survival
was signicantly higher in both lower T stage (p=0.07)
and lower N stage (p=0.003) (Fig 2 and 3).
Univariate analysis
Univariate analysis revealed T3-4, node positive,
high pathological grade, and lymphovascular invasion
to be statistically signicantly associated with breast
cancer recurrence. e recurrence rate was found to be
higher among younger patients (11.0%) than among
older (5.6%) patients; however, the dierence between
groups failed to achieve statistical signicance (hazard
ratio [HR]: 1.59, 95% condence interval [CI]: 0.76-3.30;
p=0.21) (Table 2).
Multivariate analysis
Aer adjusting for age, N stage N1 was identied
as the only independent predictor of cancer recurrence
(HR: 2.75, 95% CI: 1.18-6.40; p=0.19) (Table 3).
DISCUSSION
Age at diagnosis has long been established as an
important prognostic factor in terms of predicting survival
and recurrence. Many studies reported the survival of
young breast cancer (YBC) patients to be signicantly
Numprasit et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
440
Original Article
SMJ
TABLE 1. Staging, tumor characteristics, and treatments compared between age groups.
Variables Age ≤40 (n=145) Age >40 (n=286) p-value
Number Percentage Number Percentage
Follow-up time (yrs), 4.53±1.27 4.03±0.46 <0.001
mean±SD
T stage
T1 67 46.2% 152 53.1% 0.54
T2 72 49.7% 126 44.1%
T3 5 3.4% 7 2.4%
T4 1 7.0% 1 3.0%
N stage
N0 85 58.6% 192 67.1%
N1 34 23.4% 59 20.6% 0.05
N2 19 13.1% 17 5.9%
N3 7 4.8% 18 6.3%
Pathological grade
Low 21 14.5% 56 19.6%
Intermediate 80 55.2% 190 66.4% <0.001
High 44 30.3% 40 14.0%
Lymphovascular invasion
Negative 84 69.4% 174 74.7% 0.31
Positive 37 30.6% 59 25.3%
HER-2
Negative 91 77.1% 203 86.4% 0.03
Positive 27 22.9% 32 13.6%
%ER, mean±SD 69.7±19.7 71.3±17.4 0.41
%PR, mean±SD 54.0±30.8 53.8±28.9 0.95
Breast surgery
Wideexcision 62 42.8% 92 32.2% 0.03
Mastectomy 83 57.2% 194 67.8%
Axillary surgery
SLNB 78 53.8% 188 65.7% 0.02
ALND 67 46.2% 98 34.3%
Radiation
No 129 89.0% 270 94.4% 0.017
Yes 16 11.0% 16 5.6%
Chemotherapy
No 20 13.8% 121() 42.3% <0.001
Yes 125 86.2% 165() 57.7%
Recurrence
No 129 89.0% 270 94.4% 0.05
Yes 16 11.0% 16 5.6%
A p-value<0.05 indicates statistical signicance
Abbreviations: SD, standard deviation; T stage, tumor stage; N stage, lymph node stage; HER-2, human epidermal growth factor receptor
2; ER, estrogen receptor; PR, progesterone receptor; SLNB, sentinel lymph node biopsy; ALND, axillary lymph node dissection
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
441
Fig 1. Recurrence-free survival stratied by age.
Fig 2. Recurrence-free survival stratied by T-stage.
Fig 3. Recurrence-free survival stratied by N-stage.
Numprasit et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
442
Original Article
SMJ
TABLE 2. Univariate analysis for factors signicantly associated with cancer recurrence.
Factors No recurrence Recurrence Crude HR 95% CI p-value
Number Percentage Number Percentage
Age
≤40 129 89.0% 16 11.0% 1.59 0.76-3.30 0.22
>40 270 94.4% 16 5.6% 1
T stage
T1 208 95.0% 11 5.0% 1
T2 180 90.9% 18 9.1% 1.76 0.83-3.73 0.14
T3-4 11 78.6% 3 21.4% 3.79 1.05-13.68 0.04
N stage
N0 266 96.0% 11 4.0% 1
N1 81 87.1% 12 12.9% 3.20 1.41-7.27 0.005
N2 32 88.9% 4 11.1% 2.68 0.85-8.43 0.09
N3 20 80.0% 5 20.0% 4.98 1.72-14.41 0.003
Pathological grade
Low 74 96.1% 3 3.9% 1
Intermediate 254 94.1% 16 5.9% 1.46 0.42-5.03 0.54
High 71 84.5% 13 15.5% 4.00 1.14-14.06 0.03
Lymphovascular invasion
Negative 246 95.3% 12 4.7% 1
Positive 83 86.5% 13 13.5% 3.03 1.38-6.65 0.006
HER-2
Negative 272 92.5% 22 7.5% 1
Positive 53 89.8% 6 10.2% 1.26 0.51-3.12 0.61
Breast surgery
Wideexcision 146 94.8% 8 5.2% 1
Mastectomy 253 91.3% 24 8.7% 1.77 0.79-3.96 0.16
Axillary surgery
SLNB 256 96.2% 10 3.8% 1
ALND 143 86.7% 22 13.3% 3.42 1.62-7.24 0.001
Radiation
No 180 95.2% 9 4.8% 1
Yes 219 90.5% 23 9.5% 1.94 0.89-4.20 0.092
Chemotherapy
No 139 98.6% 2 1.4% 1
Yes 260 89.7% 30 10.3% 6.48 1.54-27.2 0.011
A p-value<0.05 indicates statistical signicance
Abbreviations: HR, hazard ratio; CI, condence interval; T stage, tumor stage; N stage, lymph node stage; HER-2, human epidermal growth
factor receptor 2; SLNB, sentinel lymph node biopsy; ALND, axillary lymph node dissection
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
443
lower than that of older breast cancer patients.
8-10
Even
though the denition of YBC varies, 40 years of age is
a generally accepted cuto value that is used in most
studies.
10,11
Brandt, et al. compared breast cancer-specic
mortality between women aged <40 and women aged 40-49,
and they found that younger women had a signicantly
worse 10-year-mortality rate (risk ratio [RR]: 1.40, 95%
CI: 1.04-1.88).
10
A recent meta-analysis reported 5- and 10-year
local recurrence risk, and both were higher in YBC
than in older-onset patients (5-year – RR: 2.64, 95%
CI: 1.94-3.60; and, 10-year – RR: 2.37, 95% CI: 1.57-
3.58).
12
In contrast, the present study, which focused on
disease recurrence, found no signicant dierence in
RFS between the younger-onset and older-onset groups
(11.0% vs. 5.6%, respectively; p=0.22). is nding may be
explained by the fact that a signicantly greater number
of young women received chemotherapy (CMT) than
the number older women who received CMT (86.2% vs.
57.7%, respectively; p<0.001). Receiving CMT could imply
more aggressive phenotypes and higher recurrence risk
among younger aged patients. On the other hand, CMT
may lower the risk of recurrence in this group. One study
found signicant improvement in the local recurrence
rate among YBC who underwent breast conserving
surgery and that were treated systemically (HR: 0.42,
95% CI: 0.28-0.60; p<0.0001).
13
e largest retrospective
study analyzed the risk of local and distant recurrence
compared between YBC and older-onset patients, and
they found no signicant dierence between groups for
either parameter (5-year LR: 3.9% vs. 4.5%; and, 5year
DFS: 75.3% vs. 77.7%).
14
e unfavorable prognosis of YBC patients may
be explained by the fact that they had more aggressive
biological subtypes and/or clinicopathological baselines.
Typically, YBC is associated with poor prognostic factors,
such as large tumor size, nodal involvement, high-
grade tumor, absence of hormone receptor, and HER2
TABLE 3. Multivariate analysis for factors that independently predict cancer recurrence.
No recurrence Recurrence Crude 95% CI p-value Adjusted 95% CI p-value
Number Percentage Number Percentage HR HR
Age
≤40 129 89.0% 16 11.0% 1.59 0.76-3.30 0.22 0.81 0.37-1.73 0.59
>40 270 94.4% 16 5.6% 1 1
T stage
T1 208 95.0% 11 5.0% 1 1
T2 180 90.9% 18 9.1% 1.76 0.83-3.73 0.14 1.23 0.55-2.74 0.61
T3-4 11 78.6% 3 21.4% 3.79 1.05-13.68 0.04 1.77 0.40-7.89 0.45
N stage
N0 266 96.0% 11 4.0% 1 1
N1 81 87.1% 12 12.9% 3.20 1.41-7.27 0.005 2.75 1.18-6.40 0.02
N2 32 88.9% 4 11.1% 2.68 0.85-8.43 0.09 2.16 0.65-7.16 0.21
N3 20 80.0% 5 20.0% 4.98 1.72-14.41 0.003 2.87 0.82-9.94 0.09
Pathological grade
Low 74 96.1% 3 3.9% 1 1
Intermediate 254 94.1% 16 5.9% 1.46 0.42-5.03 0.54 1.05 0.30-3.73 0.93
High 71 84.5% 13 15.5% 4.00 1.14-14.06 0.03 2.16 0.56-8.28 0.26
A p-value<0.05 indicates statistical signicance
Abbreviations: HR, hazard ratio; CI, condence interval; T stage, tumor stage; N stage, lymph node stage
Numprasit et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
444
Original Article
SMJ
overexpression.
8,15,16
In our experience, the hormone
receptor-positive subtype has better prognosis compared
to the others; especially luminal A, which has the best
survival. A large population-based study in Surveillance,
Epidemiology, and End Results (SEER) data assessed the
survival of operable breast cancer according to molecular
subtypes, and they found the greatest 5-year breast cancer-
specic survival in the hormone receptor positive and
HER2 negative group (HR+/HER2-) (95.5%) compared
to the triple-negative group (83.1%).
17
e present study
also found superior overall survival in the HR+/HER2-
subtype. Although hormone receptor positive breast
cancer has a more favorable prognosis, many clinicians
remain concerned that the age of diagnosis may adversely
aect survival. Several publications reported disparate
survival and recurrence results when age was used to
assess the outcomes. Liu, et al. reported poor 5-year
disease-free survival (DFS) and poor distant metastasis-
free survival (DMFS) in luminal A young women (age
≤40 years) compared to age 41-60 (5-year DFS: 80% vs.
91.6%; p<0.001; and, 5-year DMFS: 83.8% vs. 93.8%;
p<0.001). ese dierences between groups may be due
to more aggressive clinicopathological characteristics
in younger-onset patients. In our study, young women
had signicantly higher in tumor size (T3-4), more
nodal metastasis, more lymphovascular invasion, and
high nuclear grade; however, the dierence in survival
between groups was not signicant. Worse outcome
among younger women may also be explained by the
fact that younger women continue to menstruate aer
chemotherapy. SOFT and TEXT trial could clarify in this
case with their results showed that breast cancer patients
whom still premenopausal status take the benet of
adding ovarian suppression to tamoxifen or exemestane,
in order to increase disease free survival and freedom
from distant recurrence, in patients who previously
received chemotherapy.
18
Limitations
is study has several limitations. First, the retrospective
design of our study confers inherent weaknesses that
include missing or incomplete data. Second, our relatively
small sample size and short follow-up time (median: 4.1
years) may have limited the power of our survival analysis,
and our ability to identify all signicant associations
and dierences. ird, trastuzumab was not a standard
treatment for breast cancer patients who had HER2
overexpression during the study period, so the observed
high recurrence rate may have been due to inadequate
treatment. Fourth, our data was from a single center that
is ailand’s largest national tertiary referral center. As
such, we are oen referred cases that are judged to be
dicult to manage in other care settings. is may limit
the generalizability of our data to other care settings.
Fih and last, we did not have genetic prole data that
can be used to predict the benet of chemotherapy in
high-risk groups.
CONCLUSION
e results of this study revealed no signicant
dierence between age groups for recurrence-free survival
in women with hormone receptor-positive breast cancer;
however, younger women did demonstrate a higher rate
of recurrence. N stage N1 is an independent predictor
of cancer recurrence.
ACKNOWLEDGMENTS
e authors gratefully acknowledge Miss Julaporn
Pooliam of the Division of Clinical Epidemiology,
Department of Research, Faculty of Medicine Siriraj
Hospital, Mahidol University, Bangkok, ailand for
assistance with statistical analysis.
Conict of interest declaration: All authors declare no
personal or professional conicts of interest relating to
any aspect of this study.
Funding disclosure: is was an unfunded study.
REFERENCES
1. Society AC. Breast Cancer Facts & Figures 2017-2018. Atlanta:
American Cancer Society, Inc. 2017.
2. Adami HO, Malker B, Holmberg L, Persson I, Stone B. e
relation between survival and age at diagnosis in breast cancer.
N Engl J Med 1986;315:559-63.
3. Carvalho FM, Bacchi LM, Santos PP, Bacchi CE. Triple-negative
breast carcinomas are a heterogeneous entity that differs
between young and old patients. Clinics (Sao Paulo) 2010;65:
1033-6.
4. Collins LC, Marotti JD, Gelber S, Cole K, Ruddy K, Kereakoglow
S, et al. Pathologic features and molecular phenotype by patient
age in a large cohort of young women with breast cancer.
Breast Cancer Res Treat 2012;131:1061-6.
5. Pan XB, Chen RJ, Huang ST, Jiang YM, Zhu XD. Systematic
review and metaanalysis of the ecacy of breast conservation
therapy followed by radiotherapy in four breast cancer subtypes.
Oncotarget 2017;8:57414-20.
6. Fallahpour S, Navaneelan T, De P, Borgo A. Breast cancer
survival by molecular subtype: a population-based analysis of
cancer registry data. CMAJ Open 2017;5:E734-e9.
7. Liu Z, Sahli Z, Wang Y, Wol AC, Cope LM, Umbricht CB.
Young age at diagnosis is associated with worse prognosis in
the Luminal A breast cancer subtype: a retrospective institutional
cohort study. Breast Cancer Res Treat 2018;172:689-702.
8. Han W, Kang SY. Relationship between age at diagnosis and
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
445
outcome of premenopausal breast cancer: age less than 35 years
is a reasonable cut-o for dening young age-onset breast
cancer. Breast Cancer Res Treat 2010;119:193-200.
9. Johansson ALV, Trewin CB, Hjerkind KV, Ellingjord-Dale M,
Johannesen TB, Ursin G. Breast cancer-specic survival by
clinical subtype aer 7 years follow-up of young and elderly
women in a nationwide cohort. Int J Cancer 2019;144:1251-61.
10. Brandt J, Garne JP, Tengrup I, Manjer J. Age at diagnosis in
relation to survival following breast cancer: a cohort study.
World J Surg Oncol 2015;13:33.
11. Anders CK, Johnson R, Litton J, Phillips M, Bleyer A. Breast
cancer before age 40 years. Semin Oncol 2009;36:237-49.
12. He XM, Zou DH. e association of young age with local
recurrence in women with early-stage breast cancer aer
breast-conserving therapy: a meta-analysis. Sci Rep 2017;7:11058.
13. van Laar C, van der Sangen MJ, Poortmans PM, Nieuwenhuijzen
GA, Roukema JA, Roumen RM, et al. Local recurrence following
breast-conserving treatment in women aged 40 years or younger:
trends in risk and the impact on prognosis in a populationbased
cohort of 1143 patients. Eur J Cancer 2013;49:3093-101.
14. Radosa JC, Eaton A, Stempel M, Khander A, Liedtke C, Solomayer
EF, et al. Evaluation of Local and Distant Recurrence Patterns
in Patients with Triple-Negative Breast Cancer According to
Age. Ann Surg Oncol 2017;24:698-704.
15. Kataoka A, Iwamoto T, Tokunaga E, Tomotaki A, Kumamaru
H, Miyata H, et al. Young adult breast cancer patients have
a poor prognosis independent of prognostic clinicopathological
factors: a study from the Japanese Breast Cancer Registry.
Breast Cancer Res Treat 2016;160:163-72.
16. Kheirelseid EH, Boggs JM, Curran C, Glynn RW, Dooley C,
Sweeney KJ, et al. Younger age as a prognostic indicator in
breast cancer: a cohort study. BMC Cancer 2011;11:383.
17. Hwang K-T, Kim J, Jung J, Chang JH, Chai YJ, Oh SW, et al.
Impact of Breast Cancer Subtypes on Prognosis of Women
with Operable Invasive Breast Cancer: A Populationbased Study
Using SEER Database. Clinical Cancer Research 2019;25:
1970-9.
18. Francis PA, Pagani O, Fleming GF, Walley BA, Colleoni M,
Lang I, et al. Tailoring Adjuvant Endocrine Therapy for
Premenopausal Breast Cancer. N Engl J Med 2018;379:122-37.
Numprasit et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
446
Original Article
SMJ
Pornprom Muangman, M.D.*, Nattakarn Changchit, M.D.*, Kusuma Chinaroonchai, M.D.*, Nantaporn
Namviriyachote, M.D.*, Natthida Owattanapanich, M.D.*, Harikrishna K.R. Nair, M.D.**
*Department of Surgery, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok 10700, ailand, ** Wound Care Unit, Department of
Internal Medicine, Kuala Lumpur Hospital, Malaysia.
Outcomes of the Recombinant Human Epidermal
Growth Factor Addition to Chloramphenicol
Ointment for Facial Burn Wound Healing
ABSTRACT
Objective: Delayed healing of facial burns can result in scarring and psychological morbidity. Epidermal growth
factor may promote wound granulation and angiogenesis to enhance wound healing. We compared the eect of
Chloramphenicol ointment alone with Recombinant Human Epidermal Growth Factor (rhEGF) plus chloramphenicol
on facial burn wound healing.
Methods: A randomized controlled trial was conducted in patients admitted to the Burn Unit. Subjects aged 18 to
65 years with acute second-degree facial burn wounds that did not require surgical treatment were enrolled. Subjects
were divided equally and randomized to receive either topical chloramphenicol ointment twice daily (control) or
rhEGF ointment once daily and chloramphenicol ointment twice daily. Wounds were assessed at frequent intervals.
Wound size, complete healing day, pain score, infection, side eects, Vancouver Scar Scale and the cost of treatment
were recorded.
Results: Twenty-six wounds were enrolled. The mean wound size was similar in both groups (rhEGF plus
chloramphenicol treated group 38.5±18.2 cm
2
vs control group 42.1±19.4 cm
2
). Burn wounds in the rhEGF group
healed more rapidly, though the dierence was not statistically signicant (8.5 ± 3.4 days vs 9.3 ± 4.4 days, p=0.3).
No dierence was observed in the Vancouver scar scale. Mean post-treatment pain scores were the same in both
groups. ere were no infections or side eects in either group. e rhEGF-treated group was more expensive
(1463.1 ± 142.6 vs 19.2 ± 1.9 baht).
Conclusion: ere are no dierence in outcomes of acute facial burn wound healing treated with rhEGF and
chloramphenical ointment compared to chloramphenicol treatment alone. e further study to evaluate the eect
of rhEGF in patients who have some factors that delay wound healing should be done.
Keywords: Burn; wound healing; epidermal growth factor (Siriraj Med J 2019; 71: 446-449)
Corresponding author: Nantaporn Namviriyachote
E-mail: pmuangman@yahoo.com
Received 21 August 2019 Revised 15 October 2019 Accepted 18 October 2019
ORCID ID: http://orcid.org/0000-0001-9828-0060
http://dx.doi.org/10.33192/Smj.2019.66
INTRODUCTION
Facial burns represent 27-60% of all burn patients
1-3
and may result in physical and psychological morbidities.
Burn scar contracture can cause diculty with mastication,
breathing, facial expression, and pronunciation. Furthermore,
patients may develop keratopathy from ectropion,
making it dicult to breathe due to depression of the
nasal bridge and chondritis.
4
Facial appearance aects
many social outcomes such as personal identication,
interpersonal communication. Facial deformities may lower
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
447
self-condence and self-esteem that can lead to social
anxiety and mental illness.
Burn wound healing is a complex process. Previous
studies have described new drugs that can promote wound
healing without infection and decrease scar formation.
5-8
Epidermal growth factor (EGF) is secreted by platelets,
monocytes and broblasts, and stimulates epidermal
cells, broblasts, smooth muscle cells and endothelial
cells to promote wound granulation and angiogenesis.
Recombinant human epidermal growth factor (rhEGF) has
been reported to improve wound healing in diabetic foot
ulcer and radiation-related oral mucositis without major
side eects.
9-13
rhEGF can also decrease inammation
and scarring in the murine mode.
14
e chloramphenicol ointment is one of the topical
ointments that has been used as standard treatment of
facial burn wounds in some clinics because it provides
an optimal environment for eective wound healing and
reduces the risk of infection. Previous literatures reported
its ecacy for infection prophylaxis and infection treatment
purposes.
15
e chloramphenicol shows broad spectrum
which covers both gram positive and gram negative
bacteria.
16
Moreover, an ointment preparation mainly
contains oil base so it can adhere to the wound bed more
eective than cream and solution. It provides occlusive
property so it is a good option in topical antibiotics. e
combination of rhEGF with chloramphenicol might
improve healing ecacy. We aimed to compare the
eect of chloramphenicol ointment and recombinant
human epidermal growth factor (rhEGF) therapy with
chloramphenicol ointment alone for facial burn wound
healing. We hypothesized that the addition of rhEGF
would enhance the rate of wound healing, and decrease
scar formation without side eects or infection.
MATERIALS AND METHODS
A prospective randomized controlled trial was
conducted at the Department of Surgery, Faculty of
Medicine Siriraj Hospital, Bangkok, ailand. From
July of 2017 to September of 2018, patients aged 18 to 65
years with facial burn wounds were enrolled. According
to Guo et al (2010), the sample size was calculated based
on healing time between EGF-treated group and control
group.
17
Patients with an acute (< 48 hours) second-degree
facial burn wound that was expected to heal without
surgical treatment were eligible. Subjects with burns
involving the eyelids, those with compromised immunity,
renal disease, diabetes and pregnant or lactating women
were excluded. Informed consent was obtained. Data
were collected on sex, age, underlying diseases, cause
of burn, and burn wound characteristics.
e wound was cleaned and debrided to evaluate
the size and degree of burn by surgeons and experienced
burn unit nurses. Subjects were evaluated with the eKare
inSight 3D Wound Assessment (Daewoong Pharmaceutical
Co. Ltd, Seoul, South Korea) tool which measures the
area (cm
2
) of the wound.
Wounds were equally allocated by simple random
sampling method into a rhEGF-treated group (study
group) and chloramphenicol group (control group).
e control group dened as the wound was applied
with topical chloramphenicol ointment twice each day.
e study group or rhEGF-treated group was received
topical recombinant human epidermal growth factor
ointment (Easyef® Ointment Daewoong Pharmaceutical
Co. Ltd, Seoul, South Korea, 1g contains rhEGF 1 μg)
once daily in the morning combine with chloramphenicol
ointment twice each day. In case the area of wound could
not be divided to two groups, it would be assigned to
receive either rhEGF plus chloramphenicol ointment
or chloramphenicol ointment alone. Burn unit nurses
and patients dressed the wound and were blinded to the
intervention.
e wound was assessed at every other day by a
single, experienced burn unit nurse who has responsibility
to evaluate all wounds in this study. Size of the wound,
infection, pain score and side eects were recorded until
complete wound healing was achieved. Complete healing
was dened as full epithelialization of the wound with
absence of discharge. e pain score was evaluated using
a numericalscale. e Vancouver Scar Scale (VSS) was
used to evaluate the post-burn scar aer two months.
e costs of treatment including medicines and other
dressing materials were evaluated aer complete wound
healing.
Statistical analysis
Statistical analysis was performed using SPSS
version 21 statistical soware and Microso excel. Each
measurement is shown as mean ± standard deviation.
e dierences between group measurements such as
wound area, healing time and cost of treatments were
examined by T-test. A p-value of less than 0.05 was
considered statistically signicant.
RESULTS
Twenty-six wounds from fourteen patients (ve
men and nine women) who aged between 18 to 57 (mean
36.8±13.6 years) were enrolled in this study. No subject
had underlying medical illness. Nine cases were thermal
burns, three were scalds, and two were electrical burns.
e size of wounds in each group was similar. e average
Muangman et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
448
Original Article
SMJ
wound size in the rhEGF-treated group was 38.5±18.2
cm
2
compared to 42.1±19.4 cm
2
in the control group
(p = 0.31).
e results comparing between study and control
group were shown in Table 1. Wounds in rhEGF-treated
group healed faster, but the dierence was not statistically
signicant. e complete healing day in the rhEGF-
treated group was 8.5±3.4 days compared to 9.3±4.4
days in the control group (p Value = 0.3). Scar formation
aer two months was similar. e Vancouver scar scale
(VSS)
ranged from 0 to 3 with a mean in both groups was
1.6 ± 0.8.
e mean pre-treatment and 30-minutes post-
treatment pain scores were the same in both groups, and
decreasing post-treatment pain scores were observed
overall (pre-treatment 3.7±2.9; post-treatment 2.8±2.1).
Fig 1 showed the ecacy of rhEGF plus chloramphenicol
compare to chloramphenicol alone at 0, 2 weeks and
2 months aer injury.
All wounds in both groups healed without infection
or complications such as rash, swelling or urticaria.
e total cost of study group was higher than the control
group. e mean cost of EGF-treated group was 1463.1±142.6
baht and chloramphenicol group was 19.2±1.9 baht
(p ≤ 0.05). (Table 1)
DISCUSSION
Burn wound healing is a complex process. Due
to the ability of rhEGF that can stimulate epidermal cells,
broblast, smooth muscle cells and endothelial cells to
promote wound granulation and angiogenesis. is
growth factor was eective using in chronic wounds
such as diabetic foot ulcer, and radiation ulcer.
11-13
Complete healing occurred more rapidly in the rhEGF-
treated group; however, the rhEGF plus chloramphenicol
cannot accelerate facial burn wound healing eectively
compared with chloramphenicol alone. Without surgical
treatment, an acute second-degree facial burn wounds
usually heal spontaneously within two weeks.
18
Treatment with rhEGF plus chloramphenicol did
not increase post-burn hypertrophic scar formation or
TABLE 1. Comparison results between study group (rhEGF+Chloramphenicol group) and control group
(Chloramphenicol group).
Study group Control group P-value
Completehealingday(Days) 8.5±3.4 9.3±4.4 0.30
Vancouverscarscaleat2months 1.6±0.8 1.6±0.8 >0.05
Pre-treatmentpainscore 3.7±2.9 3.7±2.9 0.17
Post-treatmentpainscore 2.8±2.1 2.8±2.1 0.17
Sideeffects No No -
Totalcost(Baht) 1,463.1±142.6 19.2±1.9 <0.05
Fig 1. Facial burn wound treated with chloramphenicol (A) and rhEGF+ chloramphenicol (B).
Day 0 Day 14 Day 60
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
449
facial scar contracture compared with chloramphenicol
alone. e Vancouver Scar Scale at two months was the
same in both groups. Although there was no signicant
in scar reduction, previous animal study has already
proved the ecacy of rhEGF in visible external scars
formation over control group.
14
Moreover, the second
degree dermal layer burn wound usually create the scar
less than deep layer wound because it most likely to heal
by spontaneous epithelial wound healing.
Treatment with rhEGF and chloramphenicol did
not increase pain, infection or result in any side eects,
suggesting that this therapy is safe to use in facial burn
wounds. Our study showed, there was similar outcome
about decreasing pain score aer treatment of both rhEGF-
treated and chloramphenicol group but no statistical
signicance. It is possibly that the facial wound is a small
area which presents many sensory nerves, it is dicult
to dierentiate the pain feeling between these two areas.
Finally, the cost of rhEGF treatment is greater than
chloramphenicol. Due to insignicant eect to accelerate
burn wound healing, we might consider rhEGF just as
an alternative drug in facial burn wound or challenge in
patients who have some factors that might delay wound
healing.
Further studies are needed to evaluate the ecacy
of rhEGF in deeper facial burn wounds which might not
spontaneously heal, and in patients who have risk factors
that could delay wound healing. Deep layer facial burn
wounds cause more cellular dysfunction and post-burn
scar formation. rhEGF might demonstrate the benet
of treatment in those patients.
CONCLUSION
ere are no dierence in outcomes of acute facial burn
wound healing treated with rhEGF and chloramphenical
ointment compared to chloramphenicol treatment alone.
e further study to evaluate the eect of rhEGF in patients
who have some factors that delay wound healing or in
deeper burn wounds with less chance of spontaneous
epithelialization healing should be done. ere were no
increases in hypertrophic scar, scar contracture, infection
or other side eects aer using rhEGF. erefore, rhEGF
is safe to use just as an adjunct therapy in dermal facial
burn wounds.
ACKNOWLEDGMENTS
We thank Daewoong Pharmaceutical Company for
Support Easyef® Ointment and eKare inSight 3D Wound
Assessment in this study. e nancial relationships did
not pose a conict of interest to authors.
REFERENCES
1. Fatusi AO, Olabanji JK, Alatise OI. Management outcome and
associated factors in burn injuries with and without facial involvement
in a Nigerian population. J Burn Care Res 2006;27:869-76.
2. Chien WC, Pai L, Lin CC, Chen HC. Epidemiology of hospitalized
burns patients in Taiwan. Burns 2003;29:582-8.
3. Fraulin FOG, Illmayer SJ, Tredget EE. Assessment of cosmetic
and functional results of conservative versus surgical management
of facial burns. J Burn Care Rehab 1996;17:19-29.
4.
Cartotto R, Cicuto BJ, Kiwanuka HN. Common postburn deformities
and their management. Surg Clin North Am 2014;94:817-37.
5. Pyun G, Choi HJ, Yoon HS, ambi T, Lee DS. Polyurethane
foam containing rhEGF as a dressing material for healing
diabetic wounds: Synthesis, characterization, in vitro and in
vivo studies. Colloids Surf Biointerfaces 2015;135:699-706.
6. Khanbanha N, Atyabi F, Taheri A, Talaie F, Mahbod M,
Dinarvand R. Healing ecacy of an EGF impregnated triple
gel based wound dressing: In vitro and in vivo studies. BioMed
Research International 2014;2014:1-10.
7.
Xu J, Min D, Guo G, Liao X, Fu Z. Experimental study of epidermal
growth factor and acidic broblast growth factor in the treatment
of diabetic foot wounds. Exp er Med 2018;15:5365-70.
8. Gainza G, Bonafonte DC, Moreno B, Aguirre JJ, Gutierrez
FB, Villullas S, et al. e topical administration of rhEGF-loaded
nanostructured lipid carriers (rhEGF-NLC) improves healing
in a porcine full-thickness excisional wound model. J Control
Release 2015;197:41-7.
9. Hong JP, Jung HD, Kim YW. Recombinant Human Epidermal
Growth Factor (EGF) to Enhance Healing for Diabetic Foot
Ulcers. Ann Plast Surg 2006;56:394-8.
10. SW Lee et al. Eect of Epidermal Growth Factor Against
Radiotherapy-Induced Oral Mucositis in Rats. Int J Radiat
Oncol Biol Phys 2007;6:1172-8.
11. Lee SW, Moon SY, Kim YH, Hong JP. e use of recombinant
human epidermal growth factor to promote healing for chronic
radiation ulcer. Int Wound J 2007;4(3):216-20.
12. Tuyet HL,Nguyen Quynh TT,Vo Hoang Minh H,i Bich
DN,Do Dinh T,Le Tan D, et al. e ecacy and safety of
epidermal growth factor in treatment of diabetic foot ulcers:
the preliminary results. Int Wound J 2009;6:159-66.
13. Hong JP,Lee SW,Song SY,Ahn SD,Shin SS,Choi EK,et al.
Recombinant human epidermal growth factor treatment of
radiation-induced severe oral mucositis in patients with head
and neck malignancies. Eur J Cancer Care (Engl) 2009;18:636-41.
14. Kim YS, Lew DH, Tark KC, Rah DK, Hong JP. Effect of
recombinant human epidermal growth factor against cutaneous
scar formation in murine full-thickness wound healing. J
Korean Med Sci 2010;25:589-96.
15. Erel E, Goodyear S, Misra A. A survey of chloramphenicol
use in plastic surgery: a follow-up. J Plast Reconstr Aesthet
Surg 2010;63:e102-e3.
16. Dasgupta A. Chapter 3 - Advances in antibiotic measurement.
In: Makowski GS, editor. Advances in Clinical Chemistry.
Elsevier; 2012.p.75-104.
17. Guo X, Tan M, Guo L, Xiong A, Li Y, He X. Clinical study
on repair of burn wounds of degree II with recombinant
human epidermal growth factor in elderly patients. Zhongguo
Xiu Fu Chong Jian Wai Ke Za Xhi 2010;24:462-4.
18. Zatriqi V, Ari H, Zatriqi S, Duci S, Rrecaj Sh, Martinaj M.
Facial burns - our experience. Mater Sociomed 2013;25:26-27.
Muangman et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
450
Original Article
SMJ
Somphon Roeksomtawin, M.D.*, Palakorn Sontepa, M.D.*, Kanchana R. Kildegaard, Ph.D.**, Jatuporn Sirikun, M.D.*
*Department of Surgery, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok 10700, ailand, **e Novo Nordisk Foundation Center
for Bio sustainability, Technical University of Denmark, Lyngby, Denmark.
Decision-making Factors in Non-operative
Management of Zygomatic Fractures
ABSTRACT
Objective: A zygomatic fracture is one of the most common fractures of the maxillofacial bone. e treatment of
zygomatic fractures can be performed either by surgical or non-surgical management. However, few studies have
evaluated the reasons for choosing non-surgical treatment.
Methods: A retrospective observational study was performed of patients with zygomatic fractures that presented in
Siriraj Hospital, ailand, from January 1, 2010, to December 31, 2014. e factors associated with the non-operative
treatment of zygomatic fractures were evaluated. Moreover, we analyzed the patient data, such as their age, etiology,
type of fracture, complications, associated facial fractures, and associated other organ injuries.
Results: ere were 337 patients with a zygomatic fracture during this period. Most of the cases involved males. Trac
accidents represented a common cause of the fracture. e mean patient age was 36 years old. Trimalar fracture of
the zygoma was the most common type of fracture. e most common complication with zygoma fractures was
infraorbital nerve injury. Of the study population, 161 patients (47.8%) received non-operative treatment and 176
patients (52.2%) received operative treatment. Older age (adjusted odds ratio (95% CI); 1.02 (1.00-1.03), p-value =
0.049), no diplopia (adjusted odds ratio (95% CI); 12.30 (3.28-46.14), p-value < 0.001), no infraorbital nerve injury
(adjusted odds ratio (95% CI); 6.76 (3.81-11.99), p-value < 0.001), and no cosmetic concern (95% CI); 92.82 (11.97-
719.44), p-value < 0.001) were the only four factors related to non-operative management decisions.
Conclusion: Older age, no diplopia, no infraorbital nerve injury, and no cosmetic concern of the patient were the
factors associated with the non-operative treatment of zygomatic fractures.
Keywords: Zygomatic fracture; factors; non-operative treatment (Siriraj Med J 2019; 71: 450-456)
Corresponding author: Somphon Roeksomtawin
E-mail: somphon_r@yahoo.com
Received 20 August 2019 Revised 30 September 2019 Accepted 4 October 2019
ORCID ID: http://orcid.org/0000-0002-7214-1342
http://dx.doi.org/10.33192/Smj.2019.67
INTRODUCTION
A zygomatic fracture is a common fracture of the
facial skeleton. e incidence and etiology of zygomatic
fractures vary and dier among the reported studies.
For example, one study reported that the main causes of
zygomatic fractures were trac accidents (26%), assault
(20%), accidental falls (19%), sports injuries (10%),
home injuries (8%), and work accidents (6%)
1
, while
another reported retrospective review found that the most
common causes of zygomatic fractures were motor vehicle
accidents (40%), followed by assault (22%), motorcycle
accidents (12%), and all-terrain vehicle accidents (7%),
respectively.
2
e management of zygomatic fractures also vary,
depending on the institution or trauma center. Previous
research has reported that a high incidence of zygomatic
fractures (66%) could be managed non-operatively without
signicant complications.
2
Another report studied the
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
451
variables related to the surgical and non-surgical treatment
of zygomatic complex fractures and found that the presence
of other facial fractures, alteration of occlusion, presence
of comminuted fractures, and infraorbital nerve sensory
disturbances were statistically associated with the surgical
treatment of zygomatic fracture.
3
Nowadays, the non-operative treatment of a fracture
of the facial skeleton is interesting, but there are few reports
in the published literature on the conservative treatment
of facial fractures.
4
A previous study reported that most
patients with facial fractures in their study population
were managed with non-operative treatment.
4
ey set
some criteria for the non-operative treatment of facial
fractures and found that an undisplaced/minimally
displaced fracture (57%) or minimal/no symptoms
(24%) resulted in the patient undergoing conservative
treatment.
4
However, data about the factors associated
with the decision to apply a non-operative management
of zygomatic fractures are lacking. Consequently, we
conducted this research to identify the factors associated
with the non-operative treatment of zygomatic fractures.
Moreover, we evaluated the patients’ age, etiology, type
of fracture, complications, and time of hospital arrival
in a level I trauma center in ailand.
MATERIALS AND METHODS
A retrospective observational study was performed
of patients with a zygomatic fracture who presented to
the Facial Fracture Clinic, Division of Trauma Surgery,
Siriraj Hospital, Mahidol University, ailand, during a
period of 5 years, from January 1, 2010, to December 31,
2014. is study was approved by the ethical committee
for human research of Mahidol University (Si 641/2559).
e inclusion criteria for this study were patients who
were diagnosed with a zygomatic fracture (based on ICD-
10) and who were treated at the Facial Fracture Clinic
(FFC Clinic), Division of Trauma Surgery, Department
of Surgery, Siriraj Hospital. e exclusion criteria were:
(1) patients who refused treatment, (2) patients who died
at arrival or during hospital admission, and (3) patients
who had incomplete medical records.
In this study, we classied a zygomatic fracture
into ve types, based on Siriraj Hospital’s classication
of zygomatic fractures, namely: (1) classical sutural
fracture of the zygoma (trimalar fracture), (2) comminuted
fracture of the zygoma, (3) zygomaticomaxillary complex
fracture, (4) isolated fracture of the zygomatic arch, and
(5) isolated fracture of the frontal process of the zygoma.
For data analysis, we used the Pearson chi-square test for
discontinuous data analysis, and the Mann–Whitney U-test
for continuous data analysis (not a normal distribution).
Simple and multiple logistic regressions were used to
investigate the factors associated with the non-operative
treatment of zygomatic fractures. We used the soware
PASW Statistics 18.0 (SPSS Inc., Chicago, IL, USA) for
the data analysis.
RESULTS
In total, there were 337 patients included in this
study. e patients overall demographic data are shown
in Table 1. e patients were split into two groups based
on treatment, with 161 patients (47.8%) placed in the
non-operative treatment group and 176 patients (52.2%)
placed in the operative group. e majority of the patients
were males, representing 261 patients (77.4%), while the
females accounted for 76 patients (22.6%). e median
patient age (overall) was 30 years old (Min–Max.: 11-
92). In the non-operative treatment group, the median
age was 41.0 years old (Min–Max.: 14-92), while in the
operative treatment group, the median age was 30.5 years
old (Min–Max.: 11-91). e most common age group
that suered from a zygomatic fracture was 21-40 years
old (128 patients, 38.0%), while the least frequent age
group to suer such a fracture was those more than 80
years old (14 patients, 4.2%) (Fig 1).
e most common types of zygomatic fracture
(based on Siriraj Hospital’s classication of zygomatic
fracture) were a trimalar fracture (272 patients, 80.7%),
isolated zygomatic arch fracture (27 patients, 8.0%),
zygomaticomaxillary complex fracture (25 patients, 7.4%),
isolated frontal process fracture (9 patients, 2.7%), and
the least common of fracture was comminuted zygomatic
arch fracture (4 patients, 1.2%) (Table 1).
The most common etiologies of the zygomatic
fractures were trac accident (219 patients, 65%), assault
(54 patients, 16%), sport injuries (4 patients, 1.2%),
occupational accidents (3 patients, 0.9%), and other
etiologies (57 patients, 16.9%) (Fig 2).
e most frequent times of hospital arrival were
within 6 hours (54.6%), more than 24 hours (31.8%), 6
to 12 hours (5.0%), 18 to 24 hours (4.7%), and 12 to 18
hours (3.9%) (Fig 3).
Data were analyzed using multiple logistic regression
to identify the factors associated with the non-operative
treatment of zygomatic fracture and it was found that
older age, no underlying disease (crude odds ratio (95%
CI); 2.99 (1.64-5.14), p-value < 0.001), no other facial
fracture disease (crude odds ratio (95% CI); 1.66 (1.08-
2.56), p-value = 0.022), no diplopia (crude odds ratio (95%
CI); 8.32 (2.45-28.19), p-value < 0.001), no infraorbital
nerve injury (crude odds ratio (95% CI); 3.54 (2.18-5.74),
p-value < 0.001), and a time of hospital arrival 6.01-
Roeksomtawin et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
452
Original Article
SMJ
Demographic data Non-operative Operative Total (%) P-value
treatment (%) treatment (%)
Age(year)
Median(Min.-Max.) 41.0(14-92) 30.5(11-91) 36(11-92)
1-20 26(16.1) 44(25) 70(20.8)
21-40 53(32.9) 75(42.6) 128(38.0) <0.001
41-60 45(28.0) 46(26.1) 91(27.0)
61-80 25(15.5) 9(5.1) 34(10.1)
>80 12(7.5) 2(1.1) 14(4.2)
Gender
Male 131(81.4) 130(73.9) 261(77.4) 0.100
Female 30(18.6) 46(26.1) 76(22.6)
Etiology
Trafcaccident 99(61.5) 120(68.2) 219(65.0)
Assault 22(13.7) 32(18.2) 54(16.0)
Sport 2(1.2) 2(1.1) 4(1.2) 0.071
Occupational 1(0.6) 2(1.1) 3(0.9)
Others 37(23.0) 20(11.4) 57(16.9)
Typeoffracture(SirirajClassication)
Classicalsuturalfractureofthezygoma 133(82.6) 139(78.9) 272(80.7)
(Trimalarfracture)
Comminutedfractureofthezygoma 1(0.6) 3(1.7) 4(1.2)
Zygomaticomaxillarycomplexfracture 7(4.3) 18(10.3) 25(7.4) 0.085
Isolatedfractureofthezygomaticarch 17(10.6) 10(5.7) 27(8.0)
Isolatedfractureofthefrontalprocess 3(1.9) 6(3.4) 9(2.7)
ofthezygoma
Complications
Infra-orbitalnerve(ION)injury 33(20.5) 84(47.7) 117(34.7) <0.001
Diplopia 3(1.9) 24(13.6) 27(8.0) <0.001
Malocclusion 2(1.2) 5(2.8) 7(2.1) 0.304
Cosmeticconcern 1(0.6) 37(21.0) 38(11.3) <0.001
Timeofhospitalarrival
≤6hr. 101(62.7) 83(47.2) 184(54.6)
6.01-12hr. 10(6.2) 7(4.0) 17(5.0)
12.01-18hr. 5(3.1) 8(4.5) 13(3.9) 0.014
18.01-24hr. 8(5.0) 8(4.5) 16(4.7)
>24hr. 37(23.0) 70(39.8) 107(31.8)
Associatedfacialfracture
Nasalfracture 9(5.6) 14(8.0) 23(6.8) 0.390
Maxillaryfracture 22(13.7) 29(16.5) 51(15.1) 0.472
Mandibularfracture 2(1.2) 11(6.3) 13(3.9) 0.017
Orbitalfracture 38(23.6) 54(30.7) 92(27.3) 0.145
Otherfacialfracture 17(10.6) 21(11.9) 38(11.3) 0.691
Associatedotherorganinjury
Headinjury 61(37.9) 53(30.1) 114(33.8) 0.132
Orthopedicsfracture 23(14.3) 21(11.9) 44(13.1) 0.522
Chestinjury 10(6.2) 1(0.6) 11(3.3) 0.004
Abdominalinjury 5(3.1) 4(2.3) 9(2.7) 0.636
Eyeinjury 67(41.6) 82(46.9) 149(44.3) 0.334
Softtissueinjury 41(25.5) 45(25.7) 86(25.6) 0.958
Otherinjury 34(21.1) 35(19.9) 69(20.5) 0.780
Noneinjury 24(14.9) 31(17.6) 55(16.3) 0.502
TABLE 1. Demographic data of the patients with zygomatic fracture.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
453
Fig 1. Patients’ demographic data. e demographic
data of zygomatic fracture were demonstrated. e
numbers in the histogram represent age incidence
(years), the number of cases, the percentage of cases
from total 337 cases, respectively.
Fig 2. Etiology of injury. e etiology of zygomatic fracture was categorized in x axis, while the number of cases and the percentage of cases
from total 337 cases was demonstrated in each etiology.
Fig 3. Time of hospital arrival. e duration from time of injury to hospital visit of each patient with zygomatic fracture was recorded, and
categorized in x axis. e histogram demonstrated the time interval and the number and the percentage of cases in each interval.
Roeksomtawin et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
454
Original Article
SMJ
TABLE 2. Factors associated with non-operative treatment of zygomatic fracture
Factors Crude odd ratio P-value Adjusted odd ratio P-value
(95% CI) (95% CI)
Age 1.02(1.00-1.03) 0.049
Noneunderlyingdisease 2.99(1.64-5.14) <0.001 1.80(0.79-4.14) 0.165
Nonotherfacialfracture 1.66(1.08-2.56) 0.022 1.46(0.83-2.54) 0.187
Nodiplopia 8.32(2.45-28.18) 0.001 12.30(3.28-46.14) <0.001
Noinfra-orbitalnerveinjury 3.54(2.18-5.74) <0.001 6.76(3.81-11.99) <0.001
Nocosmeticconcern 42.59(5.77-314.46) <0.001 92.82(11.97-719.44) <0.001
Timeofhospitalarrival
<=6Hr. 0.53 0.002 0.440
6.01-12Hr. 2.30(1.41-3.77) 0.001 1.07(0.32-3.58) 0.911
12.01-18Hr. 2.70(0.95-7.68) 0.062 0.53(0.12-2.31) 0.401
18.01-24Hr. 1.18(0.36-3.87) 0.782 0.98(0.26-3.65) 0.979
>24Hr. 1.89(0.66-5.45) 0.237 0.57(0.31-1.06) 0.077
12 hours (crude odds ratio (95% CI); 2.30 (1.41-3.77),
p-value < 0.001) were the factors associated with the
non-operative treatment of a zygomatic fracture when
the analysis was performed using univariable analysis,
but when we used multivariable analysis, we found that
older age (adjusted odds ratio (95% CI); 1.02 (1.00-1.03),
p-value = 0.049), no diplopia (adjusted odds ratio (95%
CI); 12.30 (3.28-46.14), p-value < 0.001), no infra-orbital
nerve injury (adjusted odds ratio (95% CI); 6.76 (3.81-
11.99), p-value < 0.001), and no cosmetic concern (95%
CI); 92.82 (11.97-719.44), p-value < 0.001) were the only
factors associated with the non-operative treatment of
zygomatic fracture (Table 2).
DISCUSSION
Due to the dierences in the nature of each injury in
trauma patients, the analysis of management modalities
is dicult.
4,5
Non-operative treatment may be used in
any individual patient, but there are no denite criteria
for deciding on such treatment in the management of
zygomatic fracture.
3,4
Consequently, realizing which
factors are associated with the non-operative treatment
of zygomatic fracture should help the decision-making
for the treatment modalities.
3
In our study, most of the zygomatic fracture cases,
which were most common in males, were caused by trac
accidents. e mean patient age was 36.0 years old. ese
represent similar results, albeit with a little dierence in
percentages, when compared with another study that
reported that zygomatic complex fractures were more
common in men (80.1%) than women (19.9%), and the
average patient age was 35.3 years old.
3
High energy and
moderate energy mechanisms are associated with more
comminution and displacement of the fractures, and
this may need more operative intervention.
6
Despite the
dierence in ratios of the etiologies of the injury between
a previous study, which reported that the most common
causes of the injuries were due to personal aggression
(19.5%), followed by falls (18.8%)
3
, and our data, which
found that most the injuries due to trac accidents
(65%), the data from both studies showed similar results
in that there was no statistically signicant association
between the etiology and the decision made between
the surgical treatment or non-operative treatment of
zygomatic fracture.
3
In this study, the most common type of zygomatic
fracture was trimalar fracture and the most common
complication of the fracture was infraorbital nerve injury,
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
455
respectively. Diplopia usually relates to an orbital wall
defect and intraorbital fat compromise and/or extraocular
muscles entrapment
7
and these patients need more
extensive evaluation and observation.
8
In our research,
27 patients (8.0 %) presented this sign and we found that
the lack of diplopia was statistically associated with the
non-operative treatment of zygomatic fractures (p-value
< 0.001). is was in contrast to the previous study,
which reported that only 8 cases out of 43 patients with
diplopia were in the operative treatment group, and
they concluded that there was no statistically signicant
association between diplopia and the operative treatment
of zygomatic fractures.
3
However, another study showed
that 7% of cases presented with diplopia, and all were in
the surgical group.
9
us, we concluded that the lack of
diplopia (no diplopia) was one of factors associated with
the non-operative treatment of zygomatic fracture.
e incidence of infraorbital nerve sensory disturbances
in zygomaticomaxillary complex (ZMC) fractures has
been reported to vary from 24% to 94%
10,11
, and it has
been found that infraorbital nerve sensory disturbances
were the most common sequelae aer ZMC fracture.
12
Previous studies reported that only 5.8–18% of cases
did not present infraorbital nerve sensory disturbances,
and most surgical patients showed this sign.
9,13
ese
previous studies concluded that patients with ZMC fracture
presenting infra-orbital nerve sensory disturbances were
highly correlated with operative treatment.
9,13
In this
study, we demonstrated that the presence of infraorbital
nerve injury was associated with operative treatment in
zygomatic fracture, and the lack of infraorbital nerve
injury was statistically signicantly associated with the
non-operative treatment of zygomatic fracture.
e presence of comorbidity may be correlated with
adverse peri-operative events and may increase the risk of
morbidity and mortality in surgical patients,
14
and here,
the ASA classications are the key predictors of major
complications in head and neck surgery.
15
In our study,
we found that the absence of comorbidity was correlated
with the non-operative treatment of zygomatic fracture
when we analyzed the data using univariable analysis,
but it was not statistically associated with non-operative
treatment when analyzed using multivariable analysis.
A previous study showed that the cosmetic or aesthetic
outcome was the dominant indicator for the treatment
of zygomatic fracture, and concluded that the aesthetic
result is predominantly the principal indication of surgical
xation in most zygomatic fractures.
16
Similarly, our
results demonstrated that having no cosmetic concern
was strongly statistically signicantly associated with the
non-operative treatment of zygomatic fracture. Many
studies have found that more complex facial fractures,
such as naso-orbito-ethmoidal fractures, maxillary
fractures, or mandibular fractures, are associated with
a high energy trauma mechanism.
3,6
Our data showed
some correlation, but not statistically signicant, between
other facial fractures and the non-operative treatment of
zygomatic fracture. Besides, other factors, such as time
of hospital arrival and the presence of associated other
organ injury, did not show any correlation with operative
management modalities. us, it may be assumed that
the etiology, time of hospital arrival, and the severity of
injury as well as the presence of other injuries and any
comorbidity did not aect directly the decision-making
in the management of zygomatic fractures. On the other
hand, some studies, which used radiographic ndings
of the facial anatomy in their classication, showed a
relationship between the etiology, or the severity of injury,
or other facial fractures, and the surgical management of
zygomatic complex fractures.
3,6,7
e dierences in these
relationships may be due to the dierent classications
of facial fractures. For example, when patients had a
radiographic diagnosis of a bilateral zygomatic complex
fracture and if presenting to our institute with fractures
around the midline components of the facial skeleton,
e.g. naso-orbital-ethmoid fractures, then the diagnosis
of Lefort fractures were given in these cases, and they
were not enrolled in this study. In summary, we could
summarize that patients who presented with zygomatic
fracture and had the following important factors, e.g.,
older age, no diplopia, no infraorbital nerve injury, and
no cosmetic concern, could be treated with non-operative
treatment, and in the future we may be set up denite
criteria for the non-operative treatment of zygomatic
fracture to ensure the best treatment for the patient
population.
ere were some limitations in this study to report:
(1) this study was a retrospective study and we did
not mention results concerning complications and the
duration of any complication that the patient suered
because of incomplete or a lack of data about this in this
study, (2) there were many surgeons in this study who
treated the patients and applied no standard guideline
treatment, leading to confounding factors and selection
bias that depended on the individualized treatment of the
surgeons. Further research is advised, especially regarding
the non-operative treatment of zygomatic fracture for
creating standard treatment guidelines and for setting
up criteria for this treatment group in the future.
Roeksomtawin et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
456
Original Article
SMJ
CONCLUSION
In our study, we found that the majority of the patients
were male, the most common type of zygomatic fracture
was a trimalar fracture, the most common etiologies of
zygomatic fracture were trac accidents and assault,
respectively, and the most frequent time of hospital
arrival was within 6 hours. Older age, no diplopia, no
infraorbital nerve injury, and no cosmetic concern were
the factors associated with the non-operative treatment
of zygomatic fracture.
REFERENCES
1. Ungari C, Filiaci F, Riccardi E, Rinna C, Iannetti G. Etiology
and incidence of zygomatic fracture: a retrospective study
related to a series of 642 patients. Eur Rev Med Pharmacol Sci
2012;16:1559-62.
2. Sargent LA, Fernandez JG. Incidence and management of zygomatic
fractures at a level I trauma center. Ann Plast Surg 2012;68:472-6.
3. Olate S, Lima SM, Jr., Sawazaki R, Moreira RW, de Moraes M.
Variables related to surgical and nonsurgical treatment of
zygomatic complex fracture. J Craniofac Surg 2011;22:1200-2.
4. Back CP, McLean NR, Anderson PJ, David DJ. e conservative
management of facial fractures: indications and outcomes. J
Plast Reconstr Aesthet Surg 2007;60:146-51.
5. Zachariades N, Mezitis M, Anagnostopoulos D. Changing trends
in the treatment of zygomaticomaxillary complex fractures: a
12-year evaluation of methods used. J Oral Maxillofac Surg
1998;56:1152-6.
6. Carr RM, Mathog RH. Early and delayed repair of orbitozygomatic
complex fractures. J Oral Maxillofac Surg 1997;55:253-8.
7. Ellis E, 3
rd
, Kittidumkerng W. Analysis of treatment for
isolated zygomaticomaxillary complex fractures. J Oral Maxillofac
Surg 1996;54:386-400.
8. He D, Blomquist PH, Ellis E, 3
rd
. Association between ocular
injuries and internal orbital fractures. J Oral Maxillofac Surg
2007;65:713-20.
9. Kovacs AF, Ghahremani M. Minimization of zygomatic complex
fracture treatment. Int J Oral Maxillofac Surg 2001;30:380-3.
10. Pedemonte C, Basili A. Predictive factors in infraorbital sensitivity
disturbances following zygomaticomaxillary fractures. Int J
Oral Maxillofac Surg 2005;34:503-6.
11. De Man K, Bax WA. e inuence of the mode of treatment
of zygomatic bone fractures on the healing process of the
infraorbital nerve. Br J Oral Maxillofac Surg 1988;26:419-25.
12. Boano P, Roccia F, Gallesio C, Karagozoglu KH, Forouzanfar
T. Infraorbital nerve posttraumatic decit and displaced zygomatic
fractures: a double-center study. J Craniofac Surg 2013;24:2044-6.
13. Jungell P, Lindqvist C. Paraesthesia of the infraorbital nerve
following fracture of the zygomatic complex. Int J Oral Maxillofac
Surg 1987;16:363-7.
14. Eyelade O, Sanusi A, Adigun T, Adejumo O. Outcome of
anesthesia in elective surgical patients with comorbidities.
Ann Afr Med 2016;15:78-82.
15. Ferrier MB, Spuesens EB, Le Cessie S, Baatenburg de Jong RJ.
Comorbidity as a major risk factor for mortality and complications
in head and neck surgery. Arch Otolaryngol Head Neck Surg
2005;131:27-32.
16. Mahmoud SM, Liao HT, Chen CT. Aesthetic and Functional
Outcome of Zygomatic Fractures Fixation Comparison
WithResorbable Versus Titanium Plates. Ann Plast Surg
2016;76 Suppl 1:S85-90.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
457
ammawat Parakonthun, M.D.*, Chawisa Nampoolsuksan, M.D.*, Jirawat Swangsri, M.D., Ph.D.*, Anusak
Yiengpruksawan, M.D.**, Asada Methasate, M.D., Ph.D.*
*Department of Surgery, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok 10700, ailand, **Capital Health Cancer Center,
Pennington, New Jersey, USA.
Retrospective Analysis of the Outcomes in Elderly
Patients with Adenocarcinoma of the Stomach and
Esophagogastric Junction Following Three
Different Treatments
ABSTRACT
Objective: To compare survival outcomes and complications in elderly patients who underwent three dierent
treatments.
Methods: e data of patients aged 70 years old or older diagnosed with adenocarcinoma of the stomach and
esophagogastric junction (Siewert types II and III) between January 2005 and December 2016 were reviewed. e
results of dierent treatments and risk factors for post-treatment morbidity and mortality were analyzed.
Results: In total, 220 elderly patients were included: 102 (46.4%) who underwent curative intended gastrectomy
with radical lymphadenectomy (Curative group), 62 (28.2%) who underwent non-curative surgery or endoscopy
(Non-curative group), and 56 (25.4%) who received best supportive care (BSC group). Mean ages were 76.6, 78.1,
and 78.7 years old, respectively (p=0.596). Median overall survival was 32, 5, and 3 months, respectively (p<0.001).
In the curative group, overall survival was associated with the pathological stage (p=0.017), but not the resection
category (p=0.298). Concerning the curative and non-curative groups, severe post-treatment complications occurred
in 14 (8.5%) patients (7.8%, curative group; 9.7%, non-curative group). Severe post-treatment complication was
associated with age ≥80 years (p=0.023) and coronary artery disease (p<0.001); however, multivariate analysis
identied coronary artery disease as an independent risk factor for severe post-treatment complication. 5 (3%)
patients had in-hospital mortality, associated with age ≥80 years old (p=0.039) and coronary artery disease (p=0.005).
Conclusion: Curative and non-curative procedures can be safely performed in elderly patients. However, caution
should be taken with extreme elderly patients, especially those with coronary artery disease. Best supportive care
should be considered only for unt patients.
Keywords: Stomach neoplasms; gastrectomy; aged; mortality; esophagogastric junction (Siriraj Med J 2019; 71: 457-465)
Corresponding author: Asada Methasate
E-mail: teamsster@gmail.com
Received 30 August 2019 Revised 24 September 2019 Accepted 1 October 2019
ORCID ID: http://orcid.org/0000-0002-8726-365X
http://dx.doi.org/10.33192/Smj.2019.68
INTRODUCTION
e number of elderly cancer patients has increased
worldwide during the past few decades. According to the
National Cancer Institute of ailand
1
, 20% of gastric
cancer patients are aged over 70 years old. Realizing that
the older a patient is the more oen they may experience
comorbidities or poor treatment performance status
could help decide the proper care for each patient.
In gastric cancer, including esophagogastric junction
cancer, surgery is the main modality of treatment, and
Parakonthun et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
458
Original Article
SMJ
can be classied into curative and non-curative surgery.
For patients who cannot tolerate surgery and who do not
have metastatic disease, palliative and best supportive
care are the preferable choices.
From a recent study in elderly gastric cancer patients
over 80 years old, curative resection showed a signicant
better overall survival compared to supportive care.
2,3
ose who received non-surgical care mostly died from
gastric cancer.
3
Cancer-specic survival was 100% in
clinical stage I. which can imply that this group died
from non-cancer diseases.
4
Surgery with reduced nodal
dissection in stage I disease also oered no signicant
dierence in survival compared to standard surgery.
5
Furthermore, complications in patients who underwent
curative surgery were more common in older age,
especially medical complications, occurring in 7.4% of
patients aged 80 years old or older compared to 2.2% of
patients aged 60-64 years old
6
, particularly cardiac events
and pulmonary complications, such as pneumonia and
respiratory failure.
7
With older age, the mortality rate is
signicantly higher than for younger patients.
6
is study aimed to measure the survival outcome
of these elderly patients receiving curative surgery, non-
curative treatment, and best supportive care. Postoperative
complications and mortalities were reviewed and analyzed
by statistics. e results from this study may suggest
treatment guidance for elderly patients. Old age may
not be a contraindication for curative surgery in surgical
resectable diseases. Hopefully, elderly patients will receive
the appropriate treatment concerning the stage of the
disease and their performance status.
MATERIALS AND METHODS
Patient population
We retrospectively reviewed the records of a total of
220 patients aged over 70 years old with pathologically
diagnosed adenocarcinoma of the stomach and Siewert–Stein
types II and III adenocarcinoma of the esophagogastric
junction (EGJ)
8
treated between January 2005 and December
2016 at the Faculty of Medicine Siriraj Hospital, Mahidol
University, Bangkok, ailand. Of these patients, 102
patients who had underwent curative surgical resection,
which was dened as a curative intent surgical procedure
with the extent of gastric resection and lymph node
dissection as per the Japanese classication and treatment
guidelines for gastric carcinoma,
9,10
were assigned to the
Curative group; while 62 patients who had underwent
non-curative treatment, which referred to any kind of
surgical or endoscopic management for relieving patients’
symptoms, were assigned to the Non-curative group;
and 56 patients who had underwent best supportive
care alone were assigned to the BSC group. BSC was
dened as any non-invasive treatments that aimed to
relieve symptoms such as analgesics, antibiotics and
blood transfusion, and did not involve any interventions
as in the two previous groups. None of patients in the
BSC group received palliative chemotherapy. Patients
who received neoadjuvant chemotherapy, underwent
endoscopic resection for curative intention, underwent
cytoreductive surgery with or without hyperthermic
intraperitoneal chemotherapy, and a transthoracic approach
for surgical resection were excluded from this study.
e clinicopathological characteristics were reviewed
from the patients’ medical records, including the patients’
age, gender, active smoking, comorbidities, such as
diabetes mellitus, hypertension, cerebrovascular accident,
chronic obstructive pulmonary disease, coronary artery
disease, valvular heart disease, and chronic kidney disease
≥ stage IV, patient’s performance status using the Eastern
Cooperative Oncology Group performance status score
(ECOG), body mass index, preoperative serum albumin
level, American Society of Anesthesiologists physical
status classication (ASA grade), preoperative tumor
location, macroscopic Borrmann classication, tumor
dierentiation, and preoperative clinical staging using
TNM staging according to the American Joint Committee
on Cancer.
Outcome measurements
e primary objective of this study was to evaluate
the overall survival between these three groups. Overall
survival was dened as the time from the date of conrmed
pathological diagnosis at our center to the date of death
from any causes. We obtained the date of death from
the Bureau of Registration Administration of ailand
and from family members of the patients. Postoperative
complications, which were classied by the Clavien–
Dindo classication system
11
, in-hospital mortalities,
and 30-days postoperative mortalities in curative surgery
and non-curative treatment groups, were also reviewed
and analyzed. is study protocol was approved by the
Siriraj Institutional Review Board (Si 024/2018 EC1).
Statistical methods
All the statistical analyses were performed using SPSS
statistical soware version 18.0 for Windows (SPSS Inc.,
Chicago, IL, USA). Categorical variables were analyzed
by using the chi-square test, and continuous variables
were compared using the Student’s t-test. Quantitative
data was conducted using n%, mean ± SD, or median
(min, max). Survival analysis was performed using the
Kaplan–Meier method and log-rank test. Univariate
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
459
Parakonthun et al.
and multivariate analysis were performed using the
logistic regression model. All the statistical results were
considered signicant when the p-value was less than
0.05.
RESULTS
Clinicopathological characteristics
All the patients in this study were over 70 years of
age and had pathologically diagnosed adenocarcinoma
of either the stomach or esophagogastric junction. eir
clinicopathological characteristics were analyzed and
the results are shown in Table 1. e average age of the
patients was 76.6 years old in the curative group, 78.1
years old in the non-curative group, and 78.7 years old
in the BSC group. ere was a higher proportion of only
hypertensive patients in the curative group, but no dierence
in the other comorbidities. A better performance status
according to the ECOG score and a higher preoperative
serum albumin level were found in the curative group.
e majority of patients in the curative group had an
ASA physical status grade 2, while the other two groups
had ASA physical status grade 3. In terms of the tumor
characteristics, there were similar preoperative tumor
locations, macroscopic Borrmann types, and tumor
dierentiations among the three groups. Preoperative
clinical metastasis was found to be positive in the non-
curative group and BSC group more oen than in the
curative group. Two positive clinical metastasis in the
curative group were suspected liver metastasis according
to preoperative computed tomography but were found
to be negative during operation.
In the curative group, there were 51 distal gastrectomy,
46 total gastrectomy including extended gastrectomy,
3 proximal gastrectomy, and 2 partial gastrectomy. All
the patients underwent standard lymph node dissection
according to the Japanese classication of gastric carcinoma.
9
In the non-curative group, the choice of treatments
was according to the clinical symptoms of the patients,
such as bleeding or obstruction. ere were a total of
40 operative surgical procedures and 22 endoscopic
treatments. For the surgical procedures, there were 15
gastric resections (9 distal gastrectomy, 5 total gastrectomy,
and 1 partial gastrectomy), 10 feeding jejunostomy, 6
surgical bypass with gastrojejunostomy, and 2 simple
sutures for perforated gastric cancer. e remaining 7
patient cases involved 6 laparoscopies and 1 laparotomy,
in which metastasis was found so the curative surgical
operations were terminated and changed to non-curative
intention instead. For the endoscopic treatments, there
were 10 stent placements, 8 nasojejunal tube insertions, 2
endoscopic treatments for bleeding control, 1 nasogastric
tube insertion, and 1 percutaneous endoscopic gastrostomy
by push technique.
erapeutic outcomes
Overall survival analysis was conducted using
Kaplan–Meier survival curves. Comparing all the treatment
groups, there was a signicant better overall survival
in the curative group (p < 0.001) and the 5-year OS
survival was 22.5%. e median survival times were 32
months (95%CI 10.7-53.3), 5 months (95%CI 3.2-6.8), and
3 months (95%CI 2.2-3.8) aer diagnosis for the curative
group, non-curative group, and BSC group, respectively
(Fig 1). Although the non-curative group was treated for
palliative intention, they also had a signicantly better
overall survival compared with best supportive care alone
(p = 0.020). For all three types of treatment, no statistically
signicant dierence in overall survival (p = 0.079) was
found for the extremely old patients, classied as 80 years
old or over, compared to the younger patients, although
the median survival time was better in the younger group
(5 months vs. 12 months). In addition, pathological
staging aer curative surgery showed signicantly better
survival for the earlier stage than more advanced stage
of the disease (p = 0.02) as the median survival times
were 82, 34, 26, and 24 months for stages I, II, III, and
IV, respectively (Fig 2). ere were ve patients who had
underwent macroscopically curative resection but had
a positive peritoneal cytology, so they were considered
as being in stage IV of the disease. However for the nal
resection category, either R0 or R1 resection, there was
no statistical dierence in overall survival (p = 0.298).
Postoperative and treatment complications from
the curative group and non-curative group were detected
and classied by Clavien–Dindo classication
11
for both
surgical-related and non-surgical-related complications.
e surgical-related complications were gastroparesis
(16.5%), superficial surgical site infection (7.3%),
pancreatic fistula (3.7%), intraabdominal collection
(3%), bleeding (2.4%), anastomotic leakage (1.8%), and
intestinal obstruction (0.6%). e non-surgical-related.
complications were renal (15.9%); acute kidney injury
and volume overload, respiratory (11.6%); atelectasis
and pneumonia, cardiovascular (5.5%); myocardial
infarction, cardiac arrhythmia and heart failure, and
infection (4.3%); urinary tract infection, catheter-related
blood stream infection and septicemia, and neurologic
complications (3%); ischemic stroke and delirium Severe
complication was dened as Clavien–Dindo grade 3b
or higher. ere were 14 (8.5%) patients with severe
complications from these two groups. e postoperative
mortality rate was 3%. e clinical details are shown in
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
460
Original Article
SMJ
Variables Curative Non-curative BSC P-value
n = 102 n = 62 n = 56
Age,(years) 76.63±4.93 78.05±5.92 78.73±5.39 0.596
Gender 0.075
Male 64(62.7) 29(46.8) 27(48.2)
Female 38(37.3) 33(53.2) 29(51.8)
Activesmoking 6(5.9) 6(9.7) 5(8.9) 0.628
Comorbidities
DM 24(23.5) 16(25.8) 16(28.6) 0.783
HT 70(68.6) 24(38.7) 27(48.2) <0.001
CVA 12(11.8) 2(3.2) 6(10.7) 0.162
COPD 6(5.9) 3(4.8) 2(3.6) 0.814
CAD 11(10.8) 7(11.3) 3(5.4) 0.464
 VHD 6(5.9) 1(1.6) 0(0) 0.093
CKD(≥stage4) 9(8.8) 3(4.8) 7(12.5) 0.333
ECOG <0.001
0-2 98(96.1) 49(79) 35(62.5)
3-4 4(3.9) 13(21) 21(37.5)
Bodymassindex(kg/m
2
) 0.872
 <18.5 23(22.5) 16(25.8) 8(14.3)
 18.5-24.9 54(52.9) 31(50.0) 16(28.6)
 25-29.9 17(16.7) 7(11.3) 2(3.6)
 ≥30 5(4.9) 2(3.2) 1(1.8)
 Unknown 3(2.9) 6(9.7) 29(51.8)
Albumin(g/dL) 3.61±0.47 3.23±0.65 2.94±0.67 <0.001
ASAgrade 0.344
I 1(1) 1(1.6) 1(1.8)
II 62(60.8) 27(43.5) 27(48.2)
III 38(37.3) 33(53.2) 28(50)
IV 1(1) 0(0) 0(0)
V 0(0) 1(1.6) 0(0)
Preoperative 0.828
Tumordifferentiation
Well 9(8.8) 6(9.7) 4(7.1)
 Moderately 32(31.4) 16(25.8) 20(35.7)
 Poorly 52(51.0) 37(59.7) 31(55.4)
 Unknown 9(8.8) 3(4.8) 1(1.8)
Preoperative 0.204
Tumorlocation
EGJ 17(16.7) 15(24.2) 17(30.4)
 Upper1/3 7(6.9) 2(3.2) 6(10.7)
 Middle1/3 15(14.7) 10(16.1) 10(17.9)
 Lower1/3 57(55.9) 30(48.4) 19(33.9)
 Diffuse 3(2.9) 4(6.5) 4(7.1)
 Anastomosis 2(2.0) 0(0) 0(0)
Borrmannclassication 0.039
I 11(10.8) 6(9.7) 10(17.9)
 II 35(34.3) 25(40.3) 15(26.8)
 III 43(42.2) 18(29.0) 15(26.8)
 IV 11(10.8) 11(17.7) 16(28.6)
 Unknown 2(2.0) 2(3.2) 0(0)
ClinicalTstaging <0.001
cT1 10(9.8) 2(3.2) 0(0)
 cT2 58(56.9) 18(29.0) 22(39.3)
 cT3 26(25.5) 26(41.9) 17(30.4)
 cT4 7(6.9) 14(22.6) 10(17.9)
 Unknown 1(1) 2(3.2) 7(12.5)
ClinicalNstaging <0.001
Negative 79(77.5) 32(51.6) 20(35.7)
Positive 21(20.6) 25(40.3) 24(42.9)
Unknown 2(2.0) 5(8.1) 12(21.4)
ClinicalMstaging <0.001
Negative 99(97.1) 33(53.2) 11(19.6)
Positive 2(2.0) 23(37.1) 38(67.9)
Unknown 1(1.0) 6(9.7) 7(12.5)
TABLE 1. Clinicopathological characteristics.
Abbreviations: DM = Diabetes mellitus; HT = Hypertension; CVA = Cerebrovascular accident; COPD = Chronic obstructive pulmonary
disease; CAD = Coronary artery disease; VHD = Valvular heart disease; CKD = Chronic kidney disease; ECOG = Eastern Cooperative
Oncology Group performance status score; ASA = American Society of Anesthesiologists physical status score.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
461
Parakonthun et al.
Fig 1. Kaplan-Meier overall survival
curves of the curative surgery, non-
curative treatment and best supportive
care groups.
Fig 2. Kaplan-Meier overall survival
curves of the dierent pathological stages
in the curative surgery group.
Table 2. e risk factors for severe complications in these
two groups were extremely old age over 80 years old
and the presence of coronary artery disease according
to the univariate analysis (Table 3). Multivariate analysis
showed that only the presence of coronary artery disease
(OR 7.65, 95%CI 2.2-26.6; p = 0.001) was an independent
risk factor. In the subgroup analysis of curative surgery
group, severe complications were found to be signicantly
higher in distal gastrectomy than in total gastrectomy
(11.8% vs. 4.3%; p = 0.033). We found that there was a
similar proportion of coronary artery disease patients
in both types of surgery, but a higher proportion of
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
462
Original Article
SMJ
Abbreviation: EGD = Esophagogastroduodenoscopy
TABLE 2. Clinical details of the fourteen patients with severe complications.
No Age Procedure Complications Grade Management Result
Surgical related Non-surgical related (IIIb–V)
Curative surgery
1 71 Distalgastrectomywith Gastroparesis Myocardialinfarction, IIIb Nasojejunaltubeplacement, Improved
wedgeliverresection acutekidneyinjury, medicaltreatment
volumeoverload
2 80 Distalgastrectomy Intestinalobstruction None IIIb Segmentaljejunalresection Improved
3 81 Distalgastrectomy Gastroparesis None IIIb DiagnosticEGD,intravenous Improved
prokinetics,parenteralnutrition
4 80 Distalgastrectomy Gastroparesis None IIIb Nasojejunaltubeplacement Improved
5 74 Robotic-assistedlaparoscopic Gastroparesis None IIIb Nasogastrictubeplacement Improved
distalgastrectomy underuoroscopy
6 79 Robotic-assistedlaparoscopic Accidentallyclipped Pneumonia,myocardial IVb Repairhepaticartery,hemodialysis, Improved
totalgastrectomy hepaticarteryproper infarction,acutekidneyinjury, mechanicalventilation,blood
volumeoverload transfusion,medicaltreatment
7 85 Extendedgastrectomy Anastomoticleakage, Pneumonia,cardiac V Abdominaltoiletwithfeeding Deathfrom
intraabdominalcollection arrhythmia,catheter-related jejunostomy,mechanicalventilation, severesepsis
bloodstreaminfection tracheostomy,intravenousantibiotics onday49
8 92 Distalgastrectomy Bleedingfromanterior Cardiacarrhythmia, V Reexplorelaparotomytostop Deathfrom
pancreaticoduodenal acutekidneyinjury, bleeding,massivebloodtransfusion, severemeta
arteryandanterior liverfailure,coagulopathy medicaltreatment bolicacidosis
surfaceofpancreas onday2
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
463
Parakonthun et al.
TABLE 2. Clinical details of the fourteen patients with severe complications.
No Age Procedure Complications Grade Management Result
Surgical related Non-surgical related (IIIb–V)
Non-curative treatment
1 76 Simplesuturewithomental Supercialsurgicalsite Atelectasis,volumeoverload, IVa Mechanicalventilation,medical Improved
graftwithfeedingjejunostomy infection delirium,septicshock treatment,intravenousantibiotics,
wounddressing
2 75 Simplesuturewithomental None Pneumonia,cardiacarrhythmia, IVa Mechanicalventilation, Improved
graft volumeoverload,catheter- tracheostomy,medicaltreatment,
relatedbloodstreaminfection intravenousantibiotics
3 86 Palliativedistalgastrectomy None Pneumonia,acutekidneyinjury, IVb Mechanicalventilation,hemodialysis, Improved
volumeoverload intravenousantibiotics
4 92 Laparoscopicfeeding Bowelileus Septicemia V Intravenousprokinetics, Deathfrom
jejunostomy  intravenousantibiotics septicemia
onday24
5 71 Diagnosticlaparoscopy None Myocardialinfarction, V Mechanicalventilation, Deathfrom
congestiveheartfailure, medicaltreatment myocardial
volumeoverload  infarctionon
day19
6 83 Diagnosticlaparoscopy None Pneumonia,volumeoverload, V Intravenousantibiotics, Deathfrom
withfeedingjejunostomy ischemicstroke oxygensupplement respiratory
failureon
day24
Abbreviation: EGD = Esophagogastroduodenoscopy
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
464
Original Article
SMJ
extremely old aged patients over 80 years underwent
distal gastrectomy surgery (66.6% vs. 50%). In contrast,
there was no statistically signicant dierence in the
severe complications in the subgroup analysis of the
various treatment options for the Non-curative group
(p = 0.137). ere were ve in-hospital mortalities and
four 30-days-postoperative mortalities, which are also
shown in Table 2. e signicant risk factors for in-hospital
mortalities were extremely old age over 80 years old
(OR 10.36, 95%CI 1.1-95.3; p = 0.039) and the presence
of coronary artery disease (OR 14.30, 95%CI 2.2-92.5;
p = 0.005). For the 30-days-postoperative mortalities, the
only signicant risk factor was the presence of coronary
artery disease (OR 9.00, 95%CI 1.2-68.3; p = 0.034).
e mean hospital length of stay from the curative and
non-curative groups was 12 days. e risk factors for
prolonged hospitalization were an extremely old age over
80 years old, the presence of hypertension, the presence
of coronary artery disease, and patients who had severe
complications according to the univariate analysis, but
severe complications was the only independent risk
factor (OR 22.84, 95%CI 4.9-107.4; p < 0.001) identied
from the multivariate analysis.
DISCUSSION
Life expectancy has been increasing globally, paralleling
the advancement in holistic medical care. e extent of
comorbidities and decline of functional status increase
with age, and these also aect morbidity and mortality
in cancer patients. A better understanding of frailty
in an oncology population may assist in identifying
older adults who are candidates for palliative therapy
as well as those patients who may benet from standard
treatment.
12
In 2006, Gretschel et al.
13
Reported the higher
mortality rate of curative surgery for gastric cancer in
a more elderly age group. With the improvement of
surgical techniques and postoperative care over time,
recent studies have revealed a signicant better survival
outcome in curative surgical resection regardless of older
age,
2
with no signicance dierences in postoperative
morbidity and mortality.
5,14
In the present study, curative surgery showed
signicantly better overall survival compared to non-
curative treatment and best supportive care despite
the more advanced age. is result corresponded to a
previous study by Choo et al.
2
e better survival outcomes
might be from unavoidable selection bias in choosing
curative surgery for patients who tend to have a better
performance status, relatively higher nutritional status,
and the absence of clinical metastasis at the time of the
treatment decision. Only four patients in our curative
surgery group had a poor performance status, and none of
them had severe postoperative complications. A previous
study at our center showed that positive microscopic
TABLE 3. Factors aecting severe postoperative complications.
Variables Univariate Multivariate
OR (95% CI) P-value OR (95% CI) P-value
Age≥80 3.667(1.19-11.22) 0.023 3.173(0.98-10.28) 0.054
Sex 1.344(0.45-4.02) 0.597
ASA≥3 1.744(0.58-5.27) 0.325
Comorbidities
DM 1.267(0.37-4.28) 0.704
HT 1.964(0.59-6.55) 0.272
CAD 8.625(2.57-28.98) <0.001 7.652(2.20-26.60) 0.001
CKD(≥stage4) 2.333(0.46-11.89) 0.308
ECOG≥3 1.500(0.31-7.35) 0.617
Albumin<3 0.378(0.05-3.07) 0.363
Abbreviations: ASA = American Society of Anesthesiologists physical status score; DM = Diabetes mellitus; HT = Hypertension; CAD =
Coronary artery disease; CKD = Chronic kidney disease; ECOG = Eastern Cooperative Oncology Group performance status score.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
465
Parakonthun et al.
residual tumor aer standard gastrectomy seemed to
have a downward tendency for the survival outcome
15
and this was relevant to the results from the present study.
For the remaining patients beyond this curative group,
they were treated as having incurable gastric cancer based
on physicians and patients’ reasons. Hsu et al.
16
found
that palliative gastrectomy still oered a better median
survival time than the non-resection procedure. In serious
symptomatic patients, stent placement for obstruction
has been shown to be clinically eective
17
, and endoscopic
hemostasis for bleeding associated gastric cancer achieved
an initial successful for 83% of patients.
18
ese evidences
support our various options for non-curative treatment,
which depended on the clinical symptoms and yielded
a signicantly better survival result compared to best
supportive care alone.
Due to the lack of complete follow-up data and
the recurrence rate from our retrospective study design,
disease-specic survival could not be clearly interpreted.
As the mortality rate in our study was relatively low,
the risk factors for postoperative mortality could not be
analyzed. Hopefully, further research and progression
in treatment would provide better survival outcomes
and less morbidities.
Severe complications in our series occurred in 7.8%
of patients in our curative surgery group and 9.7% in
the non-curative treatment group (p = 0.77). e only
predictive risk factor for severe complications was the
presence of coronary artery disease, which diered from
the previous study, which reported a low serum albumin
level and male sex as risk factors
19
From all the severe
complication cases in the non-curative group, there was
none from endoscopic management. is might be from
it being a less invasiveness and quicker procedure.
CONCLUSION
Elderly gastric cancer patients are still candidates
for curative surgical resection in spite of their more
advanced age and comorbidities. Surgery oers a better
survival outcome than best supportive care alone. e
possibility of severe postoperative complications should
be considered, especially in coronary artery disease
patients.
ACKNOWLEDGMENTS
We are thankful to Dr. Saowalak Hunnangkul for
her assistance with the statistical analysis.
REFERENCES
1. National Cancer Institute DoMS, Ministry of Public Heath of
ailand. Hospital-Based Cancer Registry 2016;2016:1-94.
2. Choo JW, Ju Y, Lim H, Youn SH, Soh JS, Park JW, et al. Is it
benecial to perform surgical resection in elderly patients more
than 80 years old with advanced gastric cancer? Scand J
Gastroenterol 2017;52:1057-64.
3. Endo S, Shimizu Y, Ikenaga M, Ohta K, Yamada T. Survival
benet of gastrectomy for gastric cancer in patients >/= 85
years old: A retrospective propensity score-matched analysis.
Surgery 2017;161:984-94.
4. Konishi H, Ichikawa D, Itoh H, Fukuda K, Kakihara N, Takemura
M, et al. Surgery for gastric cancer patients of age 85 and older:
Multicenter survey. World J Gastroenterol 2017;23:1215-23.
5. Mikami K, Hirano K, Futami K, Maekawa T. Gastrectomy
with limited surgery for elderly patients with gastric cancer.
Asian J Surg2018;41:65-72.
6. Yang JY, Lee HJ, Kim TH, Huh YJ, Son YG, Park JH, et al.
Short- and Long-Term Outcomes Aer Gastrectomy in Elderly
Gastric Cancer Patients. Ann Surg Oncol 2017;24:469-77.
7. Zhou CJ, Chen FF, Zhuang CL, Pang WY, Zhang FY, Huang
DD, et al. Feasibility of radical gastrectomy for elderly patients
with gastric cancer. Eur J Surg Oncol 2016;42:303-11.
8. Siewert JR, Stein HJ. Classication of adenocarcinoma of the
oesophagogastric junction. Br J Surg 1998;85:1457-9.
9. Japanese Gastric Cancer A. Japanese classication of gastric
carcinoma: 3
rd
English edition. Gastric Cancer 2011;14:101-12.
10. Japanese Gastric Cancer A. Japanese gastric cancer treatment
guidelines 2014 (ver. 4). Gastric Cancer 2017;20:1-19.
11. Dindo D, Demartines N, Clavien PA. Classication of surgical
complications: a new proposal with evaluation in a cohort of
6336 patients and results of a survey. Ann Surg 2004;240:205-13.
12. Pal SK, Katheria V, Hurria A. Evaluating the older patient
with cancer: understanding frailty and the geriatric assessment.
CA Cancer J Clin 2010;60:120-32.
13. Gretschel S, Estevez-Schwarz L, Hunerbein M, Schneider U,
Schlag PM. Gastric cancer surgery in elderly patients. World
J Surg 2006;30:1468-74.
14. Kim MS, Kim S. Outcome of Gastric Cancer Surgery in Elderly
Patients. J Gastric Cancer 2016;16:254-9.
15. Parakonthun T, Parichardsombat N, Paredes HSR, Phalanusittheph
C, Taweerutchana V, Trakarnsanga A, et al. Signicance of
Microscopic Residual Tumor in Adenocarcinoma of Stomach and
Esophagogastric Junction after Gastrectomy with D2
Lymphadenectomy. Siriraj Med J 2018;70:95-102.
16. Hsu JT, Liao JA, Chuang HC, Chen TD, Chen TH, Kuo CJ, et al.
Palliative gastrectomy is benecial in selected cases of metastatic
gastric cancer. BMC Palliat Care 2017;16:19.
17. Kim JH, Song HY, Shin JH, Hu HT, Lee SK, Jung HY, et al.
Metallic stent placement in the palliative treatment of malignant
gastric outlet obstructions: primary gastric carcinoma versus
pancreatic carcinoma. AJR Am J Roentgenol 2009;193:241-7.
18. Song IJ, Kim HJ, Lee JA, Park JC, Shin SK, Lee SK, et al. Clinical
Outcomes of Endoscopic Hemostasis for Bleeding in Patients with
Unresectable Advanced Gastric Cancer. J Gastric Cancer
2017;17:374-83.
19. Kang SC, Kim HI, Kim MG. Low Serum Albumin Level, Male
Sex, and Total Gastrectomy Are Risk Factors of Severe Postoperative
Complications in Elderly Gastric Cancer Patients. J Gastric
Cancer 2016;16:43-50.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
466
Original Article
SMJ
Pholasith Sangserestid, M.D., Chutwichai Tovikkai, M.D., Yongyut Sirivatanauksorn, M.D., Prawat Kositamongkol,
M.D., Prawej Mahawithitwong, M.D., Wethit Dumronggittigule, M.D., Somchai Limsrichamrern, M.D.
Department of Surgery, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok 10700, ailand.
The Malignant Potential in Adult Choledochal
Cysts; an Awareness Issue
ABSTRACT
Objective: Although still uncommon, choledochal cysts (CCs) are more frequently found in Asians than Caucasians.
e incidence of malignancy associated with these cysts and the surgical outcomes have not been reported in ailand.
Methods: A retrospective review was performed of 48 adult CC patients who underwent surgery January 2005–
December 2015. eir clinical data and outcomes were analyzed.
Results: ere were 31 (64.6%) female patients, with a female-to-male ratio of 1.8:1. e mean age ± SD at diagnosis
was 40.5 ± 17.4 years. Using the modied Todani classication, 32 patients (66.7%) had a type I CC, one (2.1%)
had type II, twelve (25.0%) had type IV, and three (6.3%) had type V. Twelve patients (25%) had a malignant
tumor: eleven had synchronous cancer at surgery, while one with CC type I developed metachronous intrahepatic
cholangiocarcinoma during the follow-up. e cholangiocarcinoma was inside the CC in 6 cases, intrahepatic in
three, perihilar in two, and distal in one. e 5-year overall survival of the cohort was 68%. e factor aecting the
overall survival was the coexisting cancer (p < 0.001). e 1-year and 5-year overall survival rates were 58.3% and
20%, respectively, for patients with cancer, but 96.7% and 88.4% for patients without cancer.
Conclusion: e adult choledochal cysts had a high incidence of associated malignancy. Factors predicting coexisting
malignancy were an age above 40 years and a signicant weight loss. If cancer occurred, the overall survival was
signicantly poor. Even aer denitive surgery, patients still need life-long surveillance for cancer.
Keywords: Cholangiocarcinoma; choledochal cyst; malignancy; prognosis factors; risk factors (Siriraj Med J 2019;
71: 466-471)
Corresponding author: Yongyut Sirivatanauksorn
E-mail: ysirivat@gmail.com
Received 31 May 2019 Revised 10 October 2019 Accepted 11 November 2019
ORCID ID: http://orcid.org/0000-0001-7520-8832
http://dx.doi.org/10.33192/Smj.2019.69
INTRODUCTION
Choledochal cysts (CCs) are an uncommon congenital
anomaly of the bile duct that present as an abnormal
cystic dilatation of the intra and/or extra hepatic bile
ducts. In 1959, Alonso-Lej et al. rst described and
proposed a classication of CCs into 3 types.
1
is was
modied in 1977 by Todani et al. by adding types IV and V.
2
e Todani system is currently the most widely used
classication scheme for CCs. e incidence of CCs is
approximately 1 in 100,000 to 150,000 live births in the
Western population, but it has been reported to be as
high as 1 in 13,500 live births in the United States and
1 in 15,000 in Australia.
3
In Asia, CCs are more commonly
reported, especially in Japan, with an incidence of up to
1 in 1,000 live births.
4
CCs are usually diagnosed in childhood; however
about 20% of patients present in adulthood.
5
In adult
patients, CCs commonly present with abdominal pain,
jaundice, and a palpable mass. e patients may also
present with severe complications, such as cholangitis,
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
467
Sangserestid et al.
pancreatitis, secondary biliary cirrhosis, and cancer. Cancer
associated with CC is one of the dire complications of CC
and drastically aects the outcome. Carcinoma arising in
CC can be found within the cyst or any part of the biliary
system. e incidence of cholangiocarcinoma arising
with CCs varies and depends on the CC subtype and the
patient’s age.
6-11
Moreover, the risk of malignancy is not
reduced in patients with inadequate treatment, such as
a cystoenterostomy and/or incomplete cyst excision.
10,11
In ailand, little data regarding CCs
12-16
have been
reported as case reports or small case series. erefore,
this study aimed to analyze the clinical data, operative
procedures, surgical outcomes, and risk factors associated
with cancer in a large series of CC patients.
MATERIALS AND METHODS
A total of 48 consecutive patients diagnosed with
CCs who underwent surgery between January 2005
and December 2015 at the Hepatopancreatobiliary and
Transplant Surgery Unit, Department of Surgery, Faculty
of Medicine Siriraj Hospital, Mahidol University, were
enrolled. We obtained the demographic data, medical
comorbidities, laboratory information, operative outcomes,
and long-term survival of the patients from a retrospective
chart review. All of the patients were above 18 years of
age at the time of denitive surgery. e type and extent
of the surgery were based on a clinical assessment by
a consultant surgeon. is study was approved by the
Institutional Review Board of the Faculty of Medicine
Siriraj Hospital, Mahidol University (Si 089/2016).
Statistical analysis
Continuous variables are expressed as mean and
standard deviation (± SD), and comparisons were made
using Student’s t-test. Categorical variables were compared
using a chi-squared or Fisher’s exact test, as appropriate.
A logistic regression model was used to identify factors
predicting a coexisting malignancy. e overall survival
rates were analyzed with the Kaplan–Meier method and
log-rank tests. All statistical analyses were performed using
Stata for Macintosh, version 15.0 (StataCorp, College
Station, TX, USA). e level of statistical signicance
was set at a p-value of ≤ 0.05.
RESULTS
Forty-eight consecutive patients diagnosed with
CCs underwent operations in the Hepatopancreatobiliary
and Transplant Surgery Unit at Siriraj Hospital between
January 1, 2005 and December 31, 2015. ere were 31
(64.6%) female and 17 (35.4%) male patients, with a
female-to-male ratio of 1.8:1. e mean age (± SD) of
the patients at diagnosis was 40.5 (± 17.4) years. Most
of the patients had at least one symptom. e most
common presentation was abdominal pain (93.8%),
followed by obstructive jaundice (45.8%) and cholangitis
(45.8%). Choledocholithiasis and hepatolithiasis were also
observed in 52.1% and 14.6% of patients, respectively. As
to the treatments, preoperative endoscopic retrograde
cholangiopancreatography and percutaneous transhepatic
biliary drainage were performed in 15 (31.3%) and 5
(10.4%) patients, respectively. Fourteen patients (29.2%)
had previously undergone a cholecystectomy before
definitive surgery. The mean time between the first
presenting symptoms and denite surgery was 49.7
months. e demographic data of the cohort are detailed
in Table 1.
Based on the modied Todani classication
2
, 32
patients (66.7%) had a type I cyst, one (2.1%) had type II,
twelve (25.0%) had type IV, and three (6.3%) had type V.
Of note, there was no CC type III in this cohort. CC type
IV was signicantly associated with hepatolithiasis (33% ;
p = 0.02). A CC excision with a hepaticojejunostomy was
the most frequently performed operation (24 patients;
50%). e detail of the operations and the CC types are
presented in Table 2.
In this series, there were 12 patients (25%) with a
malignant tumor. Eleven patients had cancer at the time
of the surgery. Of the eleven, ve underwent curative
surgery, while six only had a palliative bypass or a biopsy.
Thirty-six patients had no cancer at the beginning.
irty-one patients had no symptoms aer surgery,
with a mean follow-up time of 31.4 months. However,
2 patients died from liver abscess and massive upper
gastrointestinal bleeding, and another 3 patients had to
undergo reoperation. Moreover, one patient with CC
type I developed intrahepatic cholangiocarcinoma 18
months aer the cyst excision and hepaticojejunostomy.
e malignancy most oen associated with the CCs
was cholangiocarcinoma. Patients with CC type V had
the highest incidence of malignancy (66.7%), following
by types I and IV at 25% and 16.7%, respectively. One
patient with CC type II had no cancer at the beginning
or during the follow-up. e cholangiocarcinoma was
inside the CC in 6 cases, intrahepatic in three, perihilar
in two, and distal in one. e other cancer types that
occurred in our patients were pancreatic adenocarcinoma
in the pancreatic head and the body of the pancreas, and
adenocarcinoma in the sigmoid colon.
Three patients had 2 concomitant cancers, but
they were dierent types. e rst patient had distal
cholangiocarcinoma and pancreatic adenocarcinoma
in the body of the pancreas. e second had perihilar
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
468
Original Article
SMJ
TABLE 1. Patient characteristics and clinical manifestations of patients with and without malignancy.
Characteristics Total % Non-CA % CA % P
Number 48 100.0 36 75.00 12 25.00
Sex 0.60
Male 17 35.4 12 70.6 5 29.4
Female 31 64.6 24 77.4 7 22.6
Age(mean±SD;years) 40.5±17.4 35.8±15.7 54.9±14.2 0.001
Clinicalmanifestation
RUQpain,epigastricpain 45 93.8 33 91.7 12 100.0 0.40
Jaundice 22 45.8 17 47.2 5 41.7 0.74
Cholangitis 22 45.8 15 41.7 7 58.3 0.31
Signicantweightloss 6 12.5 2 5.6 4 33.3 0.01
Palpablemass 6 12.5 3 8.3 3 25.0 0.13
Pancreatitis 4 8.3 4 11.1 0 0.0 0.22
Cholecystitis 3 6.3 2 5.6 1 8.3 0.73
Liverabscess 1 2.1 1 2.8 0 0.0 0.56
Asymptomatic 2 4.2 2 5.6 0 0.0 0.40
ConcurrentCBDstone 25 52.1 18 50.0 7 58.3 0.62
ConcurrentIHDstone 7 14.6 5 13.9 2 16.7 0.81
Priorcholecystectomy 14 29.2 9 25.0 5 41.7 0.08
CA19-9(mean±SD;ng/ml) 422.8±1,573.9 47.1±70.1 736.0±2,119.0 0.32
CEA(mean±SD;U/ml) 4.00±5.3 2.3±3.6 5.9±6.3 0.13
Timetosurgery 49.7±71.1 48.0±76.3 54.6±54.7 0.80
(mean±SD;months)
Abbreviations: Non-CA: no malignancy; CA: malignancy; RUQ: right upper quadrant; CBD: common bile duct; IHD: intrahepatic duct
TABLE 2. Choledochal cyst type and operative procedure.
Operation Choledochal cyst type
I II IV V
CE+HJ 24 7
CE 1
CE+HJ+LR 1 3
CE+HJ+PD 2 1
LR 2
Palliative 4 1 1
Other 1
Total 32 1 12 3
Abbreviations: CE: cyst excision; HJ: hepaticojejunostomy; LR: liver resection; PD: pancreaticoduodenectomy; Palliative: palliative surgical
bypass
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
469
Sangserestid et al.
cholangiocarcinoma and sigmoid colon cancer. e
last patient had intrahepatic cholangiocarcinoma and
cholangiocarcinoma arising in the CC (type I).
Despite undergoing curative surgery, the patients
with cancer had a poor overall survival. Four patients
(80%) died at 2, 20, 35, and 40 months aer surgery
due to disease recurrence. Only one patient survived
without evidence of disease aer 67 months of follow-
up. Furthermore, all patients who had an unresectable
disease died within a year of surgery. In the case of the
patient who developed intrahepatic cholangiocarcinoma
during surveillance, she had not experienced any disease
recurrence as at 12 months aer a right hepatectomy.
e patients with cancer were signicantly older
than those without cancer (mean age ± SD: 54.9 ± 14.2
vs 35.8 ± 15.7 years; p < 0.001). e levels of CA 19-9
and CEA in the patients with cancer were also higher
than in those without cancer, but the dierences were not
statistically signicant. A univariable analysis identied
that the two risk factors associated with cancer in these
patients was an age above 40 years and a signicant
weight loss before surgery.
e operations were mostly performed in an elective
setting. However, six patients (12.5%) had urgent surgery
after improvement of acute cholangitis in the same
admission. ese patients had signicantly more blood
loss (1,064 ml vs 465 ml; p = 0.034) and usually required
more blood transfusion (1.2 units vs 0.2 units; p = 0.013).
On the other hand, the urgent operations did not aect
the patients’ post-operative morbidity or long-term
survival.
e overall post-operative complication rate was
around 54%, with half of the complications being minor
(such as surgical site infection, at 23%). A comparison
of the patients with and without cancer revealed that
the overall complications were quite similar, except that
there was a signicantly higher rate of intrabdominal
collection in the cancer group (33% vs 8%; p = 0.03). e
complication rates and surgical outcomes of the patients
with and without cancer are presented in Table 3.
e 5-year overall survival rate of the cohort was
68%. e factor aecting the overall survival was the
coexisting cancer (p < 0.001). In the case of the patients
with cancer, the 1-year and 5-year overall survival rates
were 58.3% and 20%, respectively, whereas for the patients
without cancer, the corresponding rates were 96.7% and
88.4% (Fig 1).
TABLE 3. Surgical outcomes and post-operative complications of patients with and without malignancy.
Parameters Total % Non-CA % CA % P
Operationtype 0.94
Elective 42 87.5 31 86.1 11 91.7
Urgent 6 12.5 5 13.9 1 8.3
Cystsize(mean±SD;cm) 4.6±3.6 4.8±4.1 4.1±2.2 0.55
Operativetime(mean±SD;min) 235±14 235±13 234±44 0.96
EBL(mean±SD;ml) 541±95 575±118 431±125 0.52
PRCtransfusion(mean±SD;units) 0.36±0.87 0.42±0.94 0.18±0.60 0.44
Lengthofstay(mean±SD;days) 11.5±9.0 10.5±6.4 14.7±14.5 0.18
Complications 26 54.2 6 50 8 66.7
Anastomosisstricture 4 8.3 4 11.1 0 0.0 0.23
Bileleakage 5 10.4 5 13.9 0 0.0 0.17
Vascularinjury 4 8.3 3 8.3 1 8.3 1.0
Woundinfection 11 22.9 8 22.2 3 25.0 0.84
Intra-abdominalcollection/abscess 7 14.6 3 8.3 4 33.3 0.03
Cholangitis 6 12.5 4 11.1 2 16.7 0.61
Pancreatitis 2 4.2 2 5.6 0 0.0 0.40
Pseudoaneurysm 1 2.1 0 0.0 1 8.3 0.08
Pancreaticstula 4 8.3 2 5.6 2 16.7 0.23
Abbreviations: Non-CA: no malignancy; CA: malignancy; EBL: estimated blood loss; PRC: packed red blood cells
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
470
Original Article
SMJ
DISCUSSION
In this single-center study, we describe a series of
CCs, which is not a common disease, especially in adults.
Patients usually present with recurrent, self-limited,
abdominal discomfort at the beginning without any
other symptoms. erefore, they are oen misdiagnosed
with dyspepsia until they develop more obvious clinical
symptoms related to, for example, obstructive jaundice
or cholangitis. As abdominal ultrasonography sometimes
misses the identication of a cystic dilatation of bile duct,
there were some patients in our series who had undergone
a cholecystectomy before denitive surgery (29.2%).
Moreover, the average time to denite surgery was quite
long at about 49.7 months. Due to the late diagnosis,
the patients in the current study had a high incidence
of CBD and IHD stones; this was in concordance with
many previous studies comparing CCs in adults and
children.
8,9,18
Many previous studies have reported that the
malignancies occurring in CC patients are cholangiocarcinoma
and gallbladder cancer.
10,11,17-23
e reported incidence
rate of synchronous biliary tract malignancy ranges
from 7% to 23.5%.
6,10,11,17-23
Our series saw a similar high
incidence of concomitant cancer (25%), especially in
patients aged more than 40 years, which may result from
late diagnosis. Our results support previous studies that
found a higher age is associated with a higher cancer
incidence.
7,23
It is noteworthy that our series found no
coexisting gallbladder cancer, which may be due to the
high incidence of prior cholecystectomy.
Cholangiocarcinoma can subsequently develop in
about 0.6%-5.4% of cases, even aer curative surgery, as
described in numerous other reports.
6,10,11,17,18,20
is is
consistent with our nding that one patient (2%) with
a type I CC developed intrahepatic cholangiocarcinoma
18 months aer cyst excision and hepaticojejunostomy.
erefore, continued surveillance is advocated in every
patient aer denitive surgery. Interestingly, we also found
some CC patients presented with pancreatic adenocarcinoma
and colonic carcinoma. CCs may increase the risk for
other types of cancer without a clear pathogenesis. is
issue warrants further investigation.
e management of a CC type IVa is more complex
and still controversial. At our institution, it is managed
in two ways: cyst excision and hepaticojejunostomy,
with or without a hepatectomy. e indications for
hepatectomy in unilobar disease are hepatolithiasis,
atrophy, or suspected malignancy. A unilobar cystic
dilation only is not an indication for hepatectomy. As to
bilobar involvement, hepatectomy is performed only on
liver segments that have hepatolithiasis, lobar atrophy, or a
suspected malignancy. During surveillance, three patients
with additional liver resection were symptom-free aer
a mean follow-up of 63 months. Seven patients without
an indication for hepatectomy also had no symptoms
aer a mean follow-up of 15.6 months. Nevertheless, it
was dicult to compare the outcomes of the two groups
because of the markedly dierent mean follow-up times.
is study has some limitations due to its retrospective
nature and the small patient population. e follow-up
duration was not long enough to enable conclusions to be
drawn on the cumulative risk of cancer and hepatolithiasis
developing aer surgery. A larger series with a long-term
follow-up, such as a multi-center study or a national
registry, is therefore still needed.
Fig 1. Overall survival estimates of all patients (A), and a comparison between patients with and without cancer (B).
A B
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
471
Sangserestid et al.
CONCLUSION
e adult CCs had a high incidence of associated
malignancy. e factors predicting a coexisting malignancy
were an age above 40 years and a signicant weight loss.
If cancer occurred, the overall survival was signicantly
poor. Even aer denitive surgery, patients still need
life-long surveillance for cancer.
REFERENCES
1. Alonso-Lej F, Rever WB, Jr, Pessagno DJ. Congenital choledochal
cyst, with a report of 2, and an analysis of 94, cases. Int Abstr
Surg 1959;108:1-30.
2. Todani T, Watanabe Y, Narusue M, Tabuchi K, Okajima K.
Congenital bile duct cysts: Classication, operative procedures,
and review of thirty-seven cases including cancer arising from
choledochal cyst. Am J Surg 1977;134:263-9.
3. Gigot J, Nagorney D, Farnell M, Moir C, Ilstrup D. Bile duct
cysts: A changing spectrum of disease. J Hepatobiliary Pancreat
Surg 1996;3:405-11.
4. O’Neill JA Jr. Choledochal cyst. Curr Probl Surg 1992;29:361-
410.
5. Singham J, Yoshida EM, Scudamore CH. Choledochal cysts:
part 1 of 3: classication and pathogenesis. Can J Surg 2009;52:
434-40.
6. Watanabe Y, Toki A, Todani T. Bile duct cancer developed
aer cyst excision for choledochal cyst. J Hepatobiliary Pancreat
Surg 1999;6:207-12.
7. Voyles CR, Smadja C, Shands WC, Blumgart LH. Carcinoma in
choledochal cysts: age-related incidence. Arch Surg 1983;118:
986-8.
8. Chaudhary A, Dhar P, Sachdev A, Kumar N, Vij JC, Sarin SK,
et al. Choledochal cysts‐dierences in children and adults. Br
J Surg 1996;83:186-8.
9. Bhavsar MS, Vora HB, Giriyappa VH. Choledochal cysts: a
review of literature. Saudi J Gastroenterol 2012;18:230-6.
10. Ohashi T,Wakai T,Kubota M,Matsuda Y,Arai Y,Ohyama T,
et al. Risk of subsequent biliary malignancy in patients undergoing
cyst excision for congenital choledochal cysts. J Gastroenterol
Hepatol 2013;28:243-7.
11. Ten Hove A, de Meijer VE, Hulscher JB, de Kleine RH. Meta‐
analysis of risk of developing malignancy in congenital choledochal
malformation. Br J Surg 2018;105:482-90.
12. Khieocharoen S. Adult Choledochal Cyst in Buddhachinaraj
Phitsanulok Hospital. Buddhachinaraj Med J 2010;24:200-7.
13. Treepongkarunar S. Adult choledochal cyst: report of four
cases. Khon Kaen Med J 1993;17:27-38.
14. Akaraviputh T, Boonnuch W, Watanapa P, Lert-Akayamanee
N, Lohsiriwat D. Surgical management of adult choledochal
cysts. J Med Assoc ai 2005;88:939-43.
15. Singhavejsakul J, Ukarapol N. Choledochal cysts in children:
epidemiology and outcomes. World J Surg 2008;32:1385-8.
16. Boonyongsunchai P, Kongphanich C, Srikoch Y. Malignant
Choledochal Cyst: A Case Report and Reviews. ai J Surg
2011;32:80-83.
17. Lee SE, Jang JY, Lee YJ, Choi DW, Lee WJ, Cho BH, et al.
Choledochal cyst and associated malignant tumors in adults:
a multicenter survey in South Korea. Arch Surg 2011;146:1178-
84.
18. Edil BH, Cameron JL, Reddy S, Lum Y, Lipsett PA, Nathan H,
et al. Choledochal cyst disease in children and adults: a 30-
year single-institution experience. J Am Coll Surg 2008;206:
1000-5.
19. Tsuchiya RY, Harada NO, Ito TO, Furukawa MA, Yoshihiro
IT. Malignant tumors in choledochal cysts. Ann Surg 1977;186:
22-28.
20. Mizuguchi Y, Nakamura Y, Uchida E. Subsequent biliary
cancer originating from remnant intrapancreatic bile ducts
aer cyst excision: a literature review. Surg Today 2017;47:
660-7.
21. He XD, Wang L, Liu W, Liu Q, Qu Q, Li BL, et al. e risk of
carcinogenesis in congenital choledochal cyst patients: an
analysis of 214 cases. Ann Hepatol 2014;13:819-26.
22. Sastry AV, Abbadessa B, Wayne MG, Steele JG, Cooperman
AM. What is the incidence of biliary carcinoma in choledochal
cysts, when do they develop, and how should it aect management?
World J Surg 2015;39:487-92.
23. Feng JF, Chen WY, Chen DF, Zhou S, Liu J. Choledochal cysts
with malignancy in adult: A retrospective study with an experience
of twenty-two years. Pak J Med Sci 2011;27:6-10.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
472
Original Article
SMJ
Suparauk Geanphun, M.D., Pranya Sakiyalak, M.D.
Division of Cardiothoracic Surgery, Department of Surgery, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, ailand.
Survival and Factors Predictive of Survival in
Patients with Thymic Carcinoma
ABSTRACT
Objective: ymic carcinoma is a rare malignancy that has a poor prognosis and low 5-year survival. e rarity
of this disease and the variety of histologic subtypes has limited the evidence needed to establish disease-specic
staging and treatment guidelines. e aim of this study was to investigate overall survival, the factors that predict
survival, and the treatment modalities that inuence survival in patients with thymic carcinoma.
Methods: All thymic carcinoma cases diagnosed and treated at Siriraj Hospital (Bangkok, ailand) during 1997-
2013 were retrospectively reviewed. Univariate and multivariate analyses were performed to identify factors that
predict survival, and overall survival was estimated by Kaplan-Meier method.
Results: Of the 45 patients diagnosed during the study period, 8 were lost to follow-up. e remaining 37 patients
were included in our analysis. e median survival time was 2.5 years, with 5-year and 10-year survival of 40% and
16%, respectively. In univariate analysis, tumors >10 cm (p=0.017), high-grade histologic subtypes (p=0.014), and
high Masaoka stage (p=0.011) were signicant determinants of survival. Multivariate analysis revealed tumor size
(HR: 3.594, 95% CI: 1.103-11.714; p=0.034) and high-grade histologic subtypes (HR: 8.175, 95% CI: 1.689-39.566;
p=0.009) to be independent predictors of survival. Patients who underwent tumor removal had signicantly longer
median survival time than those who didn’t (3.85 vs. 0.63 years; p=0.005).
Conclusion: ymic carcinoma patients have low 5-year and 10-year survival. Tumor size and high-grade histologic
subtype were identied as independent predictors of survival, and surgical therapy was found to be signicantly
associated with longer median survival time.
Keywords: ymic carcinoma; malignant thymoma; invasive mediastinal tumor (Siriraj Med J 2019; 71: 472-479)
Corresponding author: Pranya Sakiyalak
E-mail: pranya.sak@mahidol.ac.th
Received 30 August 2019 Revised 9 October 2019 Accepted 15 October 2019
ORCID ID: http://orcid.org/0000-0003-1716-1225
http://dx.doi.org/10.33192/Smj.2019.70
INTRODUCTION
ymic tumors are the most common primary
neoplasms located at the anterior mediastinum, with
an incidence of less than 0.13 per 100,000 person-years
in the USA. However, thymic carcinoma is a very rare
malignancy that accounts for only 0.06% of all thymic
neoplasms.
1-5
ymic carcinoma is an invasive, aggressive,
metastasis-prone malignancy that has a poor prognosis
with an estimated 5-year survival of 30%.
7
e classication
and clinical staging of thymic carcinomas has not been
conclusively established due to the rarity of this condition
and the variety of histologic subtypes and their dissimilar
characteristics.
8,9,10
e clinical staging system that is
currently most appropriate for the evaluation of thymic
carcinomas is Masaoka staging due to its prognostic
value.
3
e aim of this study was to investigate overall
survival, the factors that predict survival, and the treatment
modalities that inuence survival in patients with thymic
carcinoma.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
473
Geanphun et al.
MATERIALS AND METHODS
is retrospective chart review included patients
diagnosed with and treated for thymic carcinoma at
the Division of Cardiothoracic Surgery, Department of
Surgery, Faculty of Medicine Siriraj Hospital, Mahidol
University, Bangkok, ailand during the January 1997
to December 2013 study period. e following patient
data were collected and analyzed: patient demographics;
clinical presentations; preoperative diagnosis; Masaoka
stage; tumor size; surgical procedure, including no surgery,
local therapy only, (including excisional biopsy), thymic
excision (including debulking and/or simple, total, and
radical excision); radiation therapy; histologic subtype;
recurrence; and, survival data. Patients with missing
treatment and/or follow-up data were excluded. In case
of the patients were deceased, they or their families would
be contacted by telephone to determine whether they
were alive or not.
e protocol for this study was approved by the
Siriraj Institutional Review Board (Si 609/2014(EC2)].
e customary requirement to obtain patient written
informed consent was waived due to the retrospective
nature of this study.
Statistical analysis
Continuous data variables were analyzed using Student’s
t-test. Nominal data were analyzed using crosstabs and
Pearson’s chi-square test. Data are reported as mean and
range or frequency and percentage. Kaplan-Meier survival
curves were generated, and they were compared using
log-rank test. A Cox proportional hazard model was used
to identify the variables that signicantly impact overall
survival. e proportionality of hazards was evaluated
using Cox regression analysis with time-dependent
covariates. e assumption of proportionality of hazards
was tested, and was not to be broken in any of the Cox
regression models. Statistical analysis was performed
with SPSS statistical soware package version 18.0 (SPSS,
Inc., Chicago, IL, USA). Statistical signicance was set
at a probability value less than 0.05.
RESULTS
Of the 45 cases diagnosed with thymic carcinoma
during the study period, 8 patients were excluded due
to loss to follow-up aer the rst hospital visit (Fig 1).
e remaining 37 patients were included. e gender
proportion breakdown was 24 males and 13 females, with
a mean age of onset of 47.59 years (range: 20-71). e
mean tumor size was 8.41 cm, and 25 patients (67.57%)
had tumor size less than 10 cm. e majority of patients
were in Masaoka Stage III and had high-grade histologic
subtypes. Half of all the patients (51.35%) had great vessel
tumor invasion, and 7 patients (18.92%) had metastasis,
which involved liver and/or bone in most cases (Table 1).
e number of patients that underwent radiation therapy
(RT) and chemotherapy (CMT) was 16 (43.24%) and
22 (59.46%), respectively. Eight patients (21.62%) had
recurrent tumor. e major histologic subtypes were
well-dierentiated squamous cell carcinoma (32.43%)
and undierentiated carcinoma (16.21%) (Fig 2).
Fig 1. Flow diagram describing enrollment of thymic carcinoma patients and the gender proportions of the 37 enrolled patients.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
474
Original Article
SMJ
Characteristics n (%)
Ageofonset(years),mean(range) 47.59(20-71)
Gender
Male 24(64.86%)
Female 13(35.14%)
Tumorsize(mean:8.41cm)
≤10cm 25(67.57%)
>10cm 12(32.43%)
MasaokaStaging
IandII 7(8.10%)
III 15(40.54%)
IV 12(32.43%)
Tumorremovalsurgeryperformed 28(75.68%)
Resection
R0 8(21.62%)
R1 10(20.02%)
R2 10(20.02%)
SVCinvasion 19(51.35%)
Histologicsubtypes
Low-grade 15(40.54%)
High-grade 22(59.46%)
Metastasis(liverand/orbone) 7(18.92%)
CMT 22(59.46%)
RT 16(43.24%)
Recurrence 8(21.62%)
TABLE 1. Demographic and clinical characteristics of 37 thymic carcinoma patients.
Abbreviations: SVC = superior vena cava; CMT = chemotherapy; RT = radiation therapy
Fig 2. Histologic subtypes
among thymic carcinoma
patients.
Geanphun et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
475
e median survival time among all thymic carcinoma
patients was 2.5 years (95% condence interval [CI]:
2.676-5.225), with 5-year survival of 40% and 10-year
survival of 16% (Fig 3). In univariate analysis (Table 2),
size of tumor greater than 10 cm (p=0.017), high-grade
histologic subtype (p=0.014), and Masaoka Stage IV
(p=0.011) were variables found to signicantly aect
survival (Fig 4). In multivariate analysis (Table 3), size
of tumor greater than 10 cm (p=0.034) and high-grade
histologic subtype (p=0.009) were identied as independent
predictors of survival in patients with thymic carcinoma.
Regarding treatment, patients that underwent surgical
therapy had signicantly better survival (p=0.005) (Fig 5)
and longer median survival time than those who did not
receive surgical therapy (3.85 vs. 0.63 years). Comparisons
between patients that did or did not undergo RT or CMT
revealed no signicant inuence on survival for either
treatment modality (p=0.187 and p=0.296, respectively).
N 2 years 5 years 10 years
Patients at risk 37 22 17 12
Fig 3. Overall survival in patients with thymic carcinomas. Median survival time was 2.5 years (95% condence interval: 2.676-5.225), with
5-year survival of 40% and 10-year survival of 16%.
TABLE 2. Univariate analysis for factors that signicantly inuence survival of patients with thymic carcinoma.
Factors P-value HR 95% CI
Age 0.147 0.978 0.945-1.008
Malegender 0.169 0.550 0.235-1.288
Size≥10cm 0.017 2.734 1.194-6.260
High-gradehistology 0.014 3.529 1.297-9.604
MasaokaStageIV 0.011 7.393 1.573-34.734
Surgery 0.372 1.631 0.558-4.770
SVCinvasion 0.956 1.022 0.463-2.259
CMT0.187 0.585 0.284-1.298
RT 0.296 0.648 0.287-1.462
Recurrence 0.489 1.416 0.533-3.732
A p-value<0.05 indicates statistical signicance
Abbreviations: HR= hazard ratio; CI= condence interval; SVC= superior vena cava; CMT= chemotherapy; RT= radiation therapy
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
476
Original Article
SMJ
Size of tumors 0 year 2 years 5 years 10 years
<10cm 25 17 15 11
≥10cm 11 4 1 1
0 year 2 years 5 years 10 years
Low-grade 15 11 11 8
High-grade 22 11 6 4
0 year 2 years 5 years 10 years
StageI-II 7 7 6 5
StageIII 15 10 8 6
StageIV 12 5 2 1
Fig 4. Overall survival curve according to (A) size of tumor >10 cm [p=0.013], (B) high-grade histologic subtype [p<0.001], and (C) Masaoka
Stage IV [p=0.014].
Geanphun et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
477
0 year 2 years 5 years 10 years
Surgery 28 19 15 11
Nosurgery 9 3 2 1
Fig 5. Overall survival compared between patients who were and were not treated with surgical tumor removal (p=0.005).
TABLE 3. Multivariate analysis for factors that signicantly inuence survival of patients with thymic carcinoma.
Factors P-value HR 95% CI
Size>10cm 0.034 3.594 1.103-11.714
High-gradehistology 0.009 8.175 1.689-39.566
Surgery 0.957 1.033 0.319-3.344
Recurrence 0.963 0.974 0.329-2.887
A p-value<0.05 indicates statistical signicance
Abbreviations: HR= hazard ratio; CI= condence interval
DISCUSSION
ymic carcinoma is a rare disease with variety
in classication and treatment modality. We analyzed
data to identify factors that impact prognosis and to
determine the treatment modality that confers the most
survival benet. e main ndings of our study were, as
follows: (1) tumor size, high-grade histologic subtype,
and Masaoka Stage IV were found to be signicant
predictors of survival in univariate analysis; (2) tumor
size and high-grade histologic subtype were identied as
independent predictors of survival in multivariate analysis;
and, (3) patients who underwent surgical intervention for
tumor removal had a median survival time signicantly
greater than that of patients who did not receive surgery
for tumor removal. e largest cohort study in thymic
carcinoma that included 290 patients also found surgical
tumor removal to be signicantly associated with longer
median survival time.
7
From previous studies – resectability, histology,
tumor stage, and tumor size were identied as prognostic
factors for survival in patients with thymic carcinoma.
7
In
this study, size of the tumor, Masaoka staging, histologic
subtype, SVC invasion, metastatic behavior, surgical
intervention, and other modalities, such as CMT and
RT, were investigated.
e 5-year cumulative survival rates reported in the
literature vary widely (14.5%
21
, 27.5%
7
, 30.0%
22
, 38.8%
23
,
and 61.1%
19
). In our study, the 5-year survival rate was
40%, and the 10-year survival rate was 16%, which is
higher than the previously reported rate of 12.1%.
24
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
478
Original Article
SMJ
e median survival times reported in the literature
also vary widely (15.6 months
21
, 35.6 months
23
, 48.0
months
22
, 53.0 months
7
, and 66.7 months.
19
e median
survival time in our study was 30 months. e observed
dierences among studies may be due to dierences in
patient selection, treatment modalities, and the number
of cases that were included in the study (some studies
had a very small study population).
24
Previous studies found Masaoka stage to be a factor
that signicantly predicts survival thymic carcinoma.
7,23,19
In the present study, Masaoka stage was not only a
prognostic factor, but it was also the strongest independent
predictor of survival in multivariate analysis.
18
e Tumor-
Nodes-Metastasis (TNM) staging system
25
was recently
proposed; however, that system has not yet been widely
accepted for general use in routine clinical practice. To
date, Masaoka staging is the system that is most commonly
used to stage patients with thymic carcinoma.
Regarding histologic subtypes, we found 12 dierent
subtypes in our study, all of which were previously dened
as being part of a subgroup of thymic carcinoma according
to a previous literature review6 and previous clinico-
pathological studies.
10-14
Hosaka, et al.
19
reported that the
prognostic values of histologic subtypes were dierent for
dierent combinations of histologic subtypes. Due to the
limited number of cases included in our study, we divided
the histology into low-grade and high-grade histologic
subtype as a denition of classication.
3,6,9-14
Levine and
Rosai
26
dened the histologic features of thymic tumors,
and the subsequent revisions culminated in the widely
accepted classication system that was proposed by the
World Health Organization (WHO) in 1999.
15
In our
study, univariate analysis showed histologic subtype
to be a prognostic factor for survival (p<0.001, Fig 4b).
Subsequent multivariate analysis revealed histologic
subtype to be an independent signicant prognostic
factor for overall survival (p=0.009, HR: 8.175, 95%
CI: 1.689-39.566), which is consistent with previous
reports.
13,15,19,21,27
Multivariate analysis also revealed tumor size to
be an independent predictor of survival (p=0.034, HR:
3.594, 95% CI: 1.103-11.714), which is consistent with
the findings of two previous studies.
7,27
Specifically,
tumor size less than or equal to 10 cm was associated
with signicantly longer overall survival when compared
to tumor size greater than 10 cm (Fig 4a). A previously
published review of the literature
6
suggested a cut point
of 8 cm for determining patient prognosis, but that value
applied to thymomas. Since not cut point currently exists
for thymic carcinoma, we included 8 cm, 9 cm, and 10
cm tumor size values in our statistical analysis, and we
found 10 cm to be the cut point that was associated
with a signicant dierence in survival. However and
given our small study population, additional study is
needed to conrm the tumor size cut point threshold
for determining patient prognosis.
All of the studies in thymic carcinoma that have
been reported to date found surgical treatment to be
signicantly associated with improved survival, which
explains why surgical therapy is the preferred treatment in
thymic carcinoma.
2,28
Consistent with previous studies, we
found the same survival benet of surgical intervention.
Patients who underwent tumor removal had a signicantly
longer median survival time than those who did not
receive surgical therapy (3.85 vs. 0.63 years; p=0.005)
(Fig 5). In comparison, an analysis of survival in 154
patients with thymic carcinoma by Kondo and Monden
16
revealed 5-year survival rates of 67% aer total resection,
30% aer subtotal resection, and 24% with no operation.
Complete resection was analyzed in our study; however,
no signicant dierences were observed among the in
the R0, R1, and R2 subgroups. is may be explained
by the small number of cases enrolled in our study.
However, our data support the widely adopted opinion
that surgery is the preferred modality of treatment for
patients with thymic carcinoma.
Limitations
is study has some mentionable limitations. First,
the retrospective nature of this study makes it vulnerable
to missing or incomplete data. Second, all of the data
included in this study was collected from a single center.
ird, we included a small number of patients due to the
relative rarity of thymic carcinoma. is small sample
size may have limited the strength of our study to identify
all signicant dierences and associations. Fourth, our
center is a national tertiary referral center, which means
that we are oen referred cases that are too complicated
to be treated in other care settings. is suggests that our
ndings may not be immediately generalizable to thymic
cancer cases in other care settings. e strength of this
study is that we included patients that were diagnosed,
treated, and followed at our center during a 17-year study
period. Moreover, the data from our study broaden our
understanding of this rare disorder, and will contribute
to the development of thymic carcinoma-specic staging
and treatment guidelines in the future.
CONCLUSION
ymic carcinoma is a rare and invasive mediastinal
tumor with low 5-year and 10-year survival. Univariate
analysis revealed size of tumor, high-grade histologic
Geanphun et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
479
subtype, and Masoaka Stage IV as factors that signicantly
inuence survival. Of those, tumor size and high-grade
histologic subtype were identied as independent predictors
of survival in multivariate analysis. Patients who underwent
surgical therapy had a signicantly longer median survival
time.
ACKNOWLEDGMENTS
e authors gratefully acknowledge to Ms. Julaporn
Pooliam of the Division of Clinical Epidemiology, Research
Department, Faculty of Medicine Siriraj Hospital, Mahidol
University for assistance with statistical analysis.
Conict of interest declaration: Both authors declare
no personal or professional conicts of interest relating
to any aspect of this study.
Funding disclosure: is was an unfunded study.
REFERENCES
1. Engels EA. Epidemiology of thymoma and associated malignancies.
J orac Oncol 2010;5(10 Suppl 4):S260-5.
2. Venuta F, Anile M,Diso D,Vitolo D,Rendina EA,De Giacomo
T,et al. ymoma and thymic carcinoma. Eur J Cardiothorac
Surg 2010;37:13-25.
3. Chen G, Marx A,Chen WH,Yong J,Puppe B,Stroebel P,et al.
New WHO histologic classication predicts prognosis of thymic
epithelial tumors: a clinicopathologic study of 200 thymoma
cases from China. Cancer 2002;95:420-9.
4. Wright CD, Wain JC,Wong DR,Donahue DM,Gaissert
HA,Grillo HC,Mathisen DJ. Predictors of recurrence in
thymic tumors: importance of invasion, World Health Organization
histology, and size. J orac Cardiovasc Surg 2005;130:1413-
21.
5. Marx A, Rieker R,Toker A,Länger F,Ströbel P. ymic
carcinoma: is it a separate entity? From molecular to clinical
evidence. orac Surg Clin 2011;21:25-31.
6. Syrios J, Diamantis N,Fergadis E,Katsaros L,Logothetis M,
Iakovidou I,et al. Advances in thymic carcinoma diagnosis
and treatment: a review of literature. Med Oncol 2014;31:44.
7. Weksler B, Dhupar R,Parikh V,Nason KS,Pennathur A,Ferson
PF. ymic carcinoma: a multivariate analysis of factors predictive
of survival in 290 patients. Ann orac Surg 2013;95:299-303.
8. Snover DC, Levine GD, Rosai J. Thymic carcinoma. Five
distinctive histological variants. Am J Surg Pathol 1982;6:451-
70.
9. Levine GD, Rosai J. ymic hyperplasia and neoplasia: a review
of current concepts. Hum Pathol 1978;9:495-515.
10. Bergh NP, Gatzinsky P, Larsson S, Lundin P, Ridell B. Tumors
of the thymus and thymic region: I. Clinicopathological studies
on thymomas. Ann orac Surg 1978;25:91-8.
11. Weissferdt A, Moran CA. Thymic carcinoma, part 1: a
clinicopathologic and immunohistochemical study of 65 cases.
Am J Clin Pathol 2012;138:103-14.
12. Weissferdt A, Moran CA. Thymic carcinoma, part 2: a
clinicopathologic correlation of 33 cases with a proposed
staging system. Am J Clin Pathol 2012;138:115-21.
13. omas de Montpréville V, Ghigna MR,Lacroix L,Besse B,Broet
P,Dartevelle P,et al. ymic carcinomas: clinicopathologic study
of 37 cases from a single institution. Virchows Arch 2013;462:
307-13.
14. Wang H, Sima CS, Beasley MB,Illei P,Saqi A,Nonaka D,et al.
Classication of thymic epithelial neoplasms is still a challenge to
thoracic pathologists: a reproducibility study using digital
microscopy. Arch Pathol Lab Med 2014;138:658-63.
15. Marx A,Ströbel P,Badve SS,Chalabreysse L,Chan JK,Chen G,
et al. ITMIG consensus statement on the use of the WHO
histological classication of thymoma and thymic carcinoma:
rened denitions, histological criteria, and reporting. J orac
Oncol 2014;9:596-611.
16. Kondo K, Monden Y. erapy for thymic epithelial tumors:
a clinical study of 1,320 patients from Japan. Ann orac Surg
2003;76:878-85.
17. Berardi R, De Lisa M,Pagliaretta S,Onofri A,Morgese F,Savini
A,et al. ymic neoplasms: an update on the use of chemotherapy
and new targeted therapies. A literature review. Cancer Treat
Rev 2014;40:495-506.
18. Masaoka A. Staging system of thymoma. J Thorac Oncol
2010;5(10 Suppl 4):S304-12.
19.
Hosaka Y, Tsuchida M,Toyabe S,Umezu H,Eimoto T,Hayashi J.
Masaoka stage and histologic grade predict prognosis in patients
with thymic carcinoma. Ann orac Surg 2010;89:912-7.
20. Yen YT, Lai WW, Lai WW,Chang KW,Chang KC,Lee SC,et al.
Factors predicting recurrence and postrecurrence survival in
completely resected thymic carcinoma. Ann Thorac Surg
2014;97:1169-75.
21. Chalabreysse L,Etienne-Mastroianni B,Adeleine P,Cordier
JF,Greenland T,Thivolet-Bejui F. Thymic carcinoma: a
clinicopathological and immunohistological study of 19 cases.
Histopathology 2004;44:367-74.
22. Liu HC, Hsu WH, Chen YJ,Chan YJ,Wu YC,Huang BS,et al.
Primary thymic carcinoma. Ann orac Surg 2002;73:1076-81.
23. Lee CY,Bae MK,Park IK,Kim DJ,Lee JG,Chung KY. Early
Masaoka stage and complete resection is important for prognosis
of thymic carcinoma: a 20-year experience at a single institution.
Eur J Cardiothorac Surg 2009;36:159-63.
24. Wang S, Wang Z,Liu X,Wang D,Liu F. Prognostic Factors
of Patients with ymic Carcinoma Aer Surgery: A Retrospective
Analysis of 58 Cases. World J Surg 2014;38:2032-8.
25. Travis WD,Brambilla E,Nicholson AG,Yatabe Y,Austin
JHM,Beasley MB, et al. e 2015 World Health Organization
Classication of Lung Tumors: Impact of Genetic, Clinical
and Radiologic Advances Since the 2004 Classication. J orac
Oncol2015;10:1243-60.
26. Levine GD, Rosai J. ymic hyperplasia and neoplasia: a review
of current concepts. Hum Pathol 1978;9:495-515.
27. Blumberg D, Port JL, Weksler B, Delgado R,Rosai J,Bains
MS,et al. ymoma: a multivariate analysis of factors predicting
survival. Ann orac Surg 1995;60:908-14.
28. Chung DA. Thymic carcinoma analysis of nineteen
clinicopathological studies. orac Cardiovasc Surg 2000;48:
114-9.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
480
Original Article
SMJ
Vutthipong Sanphasitvong, M.D.*, Loh Yee Jim, M.D.**, Kriangkrai Tantiwonglkosri, M.D.*
*Department of Surgery, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok 10700, ailand, **Cardiothoracic Surgery Service, KK
Women’s and Children’s Hospital, 100 Bukit, Singapore.
Impact of Accuracy of Preoperative Transthoracic
Echocardiography on Complex Congenital Heart
Surgery in Pediatrics
ABSTRACT
Objective: In simple congenital heart disease, transthoracic echocardiography for the diagnosis and preoperative
evaluation is an eective investigation. But its role in complex congenital heart disease is yet to be proven. e aim
of this study was to nd an accuracy ratio of preoperative echocardiography in complex congenital heart disease
and the eect of non-equivalent ndings to surgical plan and outcomes.
Methods: Medical records of complex congenital heart disease patients who underwent open heart surgery during
2013 -2015, including echocardiographic reports, operative notes, in-patient and out-patient data were retrospectively
reviewed. Patients who underwent palliative shunt procedure, age > 15 year and missing data were excluded. A
total of 300 patients were included in the study. We used intraoperative ndings as a gold standard to compare with
preoperative transthoracic echocardiography. en, we analyzed data comparing between groups of patients who
had concordant and non-concordant echocardiographic nding. Aer that we compared between groups of patients
who had surgery changed from the preoperative plan and those who had not. SPSS version 18 was used to analyze.
Results: e concordance of preoperative echocardiography was 77.7%. In the non-concordant data group, approximately
one third had to change the plan of operation and about one tenth had to change cannulation technique. Most
of non-concordance were that of systemic venous drainage and coronary artery pattern. But it had no signicant
eect to surgical outcomes including operative time, cardiopulmonary bypass time, morbidity and mortality rate.
Conclusion: Transthoracic echocardiography is valuable tool that give a lot of details on anatomical and physiological
data in congenital heart disease. However, some complex congenital heart patients may need additional investigations
such as cardiac catheterization, CT scan, Cardiac MRI in order to obtain adequate anatomical and physiological
data for surgical planning.
Keywords: Preoperative; transthoracic echocardiography; complex congenital heart surgery (Siriraj Med J 2019;
71: 480-485)
Corresponding author: Kriangkrai Tantiwongkosri
E-mail: kae814@yahoo.com
Received 13 May 2019 Revised 16 September 2019 Accepted 25 September 2019
ORCID ID: http://orcid.org/0000-0002-8926-6715
http://dx.doi.org/10.33192/Smj.2019.71
INTRODUCTION
In congenital heart surgery, anatomical and
physiological data are important in planning the surgical
procedure. In the past, cardiac catheterization was used to
obtain the information. However, cardiac catheterization
is an invasive procedure with some procedural risks.
Nowadays, echocardiography as an imaging modality has
improved tremendously. We can obtain many anatomical
and physiological data from this non-invasive procedure.
In simple congenital heart diseases such as atrial septal
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
481
defect (ASD), ventricular septal defect (VSD) and patent
ductus arteriosus (PDA), we can do surgery based solely
on preoperative echocardiography.
1-6
During recent years,
there has been many studies suggested that we can also
perform complex congenital heart surgery based on
preoperative echocardiographic nding alone without
increasing the risk of the complication and overall
outcome.
7-13
However, this had to be proven and is
still not a standard practice in our institute. We usually
have both preoperative echocardiography and cardiac
catheterization in complex congenital heart patients.
e aim of this study is to nd out the proportion
of complex congenital heart patients with concordant
data between preoperative echocardiogram and surgical
ndings and its impact on surgical plan and outcomes.
Complex congenital heart disease in this study
includes all congenital heart disease except atrial septal
defect (ASD), ventricular septal defect (VSD) and patent
ductus arteriosus (PDA).
Objectives
Primary Objective
1. To assess the proportion of patients with concordant
data between preoperative echocardiography and intra-
operative ndings.
Secondary Objectives
2. To study the impact of non-concordance ndings on
surgical outcome (morbidity, mortality rate).
3. To study the impact of non-concordance ndings on
surgical plan.
MATERIALS AND METHODS
is is a descriptive study. Data of complex congenital
heart patients who had undergone heart surgery during
2013-2015 at Siriraj Hospital were retrospectively reviewed.
We excluded palliative surgery cases (systemic to pulmonary
shunt, PA banding), patients over 15 years old and missing
data patients. A total of 300 patients were reviewed. e
rst stage palliative surgery cases were excluded because
this procedure underwent without cardiopulmonary
bypass support and we performed systemic to pulmonary
shunt through thoracotomy incision which might limit
intraoperative nding. e adult complex congenital heart
patients were excluded because this group of patients
have higher risk due to delay in surgery.
We reviewed inpatient and outpatient documents,
operative notes and preoperative transthoracic
echocardiographic reports. Preoperative echocardiogram
was performed by pediatric cardiologist and/or fellow
in pediatric cardiologist under the supervision of a
certied pediatric cardiologist. Preoperative transthoracic
echocardiographic reports were assessed on their
concordance based on the intraoperative ndings. Data
that is deemed concordant must have all the data as
listed in Fig 1 and correlate with intraoperative ndings.
In the non-concordance data group, the percentages of
procedural change from preoperative plan due to the new
ndings were calculated. Procedural change is dened
as change of cannulation technique from preoperative
planning or having additional procedure performed.
For example, persistent le sided superior-vena cava
(SVC) that was found intra-operatively and additional
le SVC cannulation that had to be performed. is was
considered a change of cannulation technique. For change
of intraoperative decisions that were not the result of
dierent preoperative echocardiographic report, these
were not included in this study. e concordance data
and non-concordance data groups were analyzed to nd
the dierences between demographic data, morbidity
and mortality rate, cardiopulmonary bypass time and
aortic cross clamp time.
Quantitative data was presented in median and
range. Qualitative data was presented in number and
percentage. In inferential statistics, Chi-square test was
used for qualitative data and t-test/Mann Whitney U-test
were used for quantitative data. P-value < 0.05 was used
as a cut-o point of statistical signicance. SPSS
V18.0 was used to analyze data.
Ethical approval for our clinical study was obtained
by human research protection unit of Siriraj Hospital.
(Si 248/2016)
RESULTS
300 cases were included in the study. Median age
of patients is 21.5 months ranging from 2 days to 15
years old. Most cases were Tetralogy of Fallot (TOF),
Univentricular heart disease and double outlet of right
ventricle as shown in Table 1.
Among these patients, 77.7% (233 patients) of
preoperative transthoracic echocardiography correlated
to the intra-operative ndings. In 22.3% of patients
(n= 67), transthoracic echocardiographic ndings diered
from intraoperative ndings. Systemic venous drainage
(i.e. missing le superior vena cava (SVC), abnormal
SVC drainage) were found to be the most common
non-concordant ndings. Second most common non-
concordant ndings were that of the coronary artery
pattern (Table 2) which was found in all our patients
with transposition of great arteries (TGA). So in TGA
patients, echocardiography had accurately report type
Sanphasitvong et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
482
Original Article
SMJ
Total Echo concordant Echo non- p-value
(n=300) group concordant
(n=233) group (n=67)
Age,months 21.5(.07,180) 22(.07,175) 20(.27,180) 0.878
Median(Min,Max)
PreoperativeDiagnosis,n(%)
0.293
TOF 86(28.7%) 68(29.2%) 18(26.9%)
TGA 21(7%) 13(5.6%) 8(11.9%)
DORV 25(8.3%) 16(6.9%) 9(13.4%)
PA/IVS 3(1%) 3(1.3%) 0(0%)
Truncus 15(5%) 10(4.3%) 5(7.5%)
TAPVR 14(4.7%) 11(4.7%) 3(4.5%)
PAPVR 7(2.3%) 5(2.1%) 2(3.0%)
AVcanel 27(9%) 17(7.3%) 10(14.9%)
Ebstein 1(0.3%) 1(4%) 0(0%)
APWindow 1(0.3%) 1(4%) 0(0%)
CoA 14(4.7%) 12(5.2%) 2(3.0%)
ALCAPA 3(1%) 3(1.3%) 0(0%)
HLHS 8(2.7%) 8(3.4%) 0(0%)
TA 17(5.7%) 16(6.9%) 1(1.5%)
UVH 40(13.3%) 33(14.2%) 7(11.4%)
Others 18(6%) 16(6.9%) 2(3.0%) 0.731
Urgentsurgery,n(%) (n=294) (n=228) (n=66)
Elective 263(89.5%) 204(89.5%) 59(89.4%)
Urgent 29(9.9%) 22(9.6%) 7(10.6%)
Emergency 2(0.7%) 2(0.9%) 0(0%)
Redooperation (n=300) (n=233) (n=67) 0.013
32(10.7%) 31(13.3%) 1(1.5%)
TABLE 1. Patient characteristics.
Abbreviations: TOF=Tetralogy of Fallot, TGA=Transposition of great artery, DORV=Double outlet right ventricle, PA/IVS=pulmonary
atresia/intact ventricular septum, Truncus=Truncus arteriosus, TAPVR=total anomalous pulmonary venous return, PAPVR=partial
anomalous pulmonary venous return, AV canal=Common AV canal, AP window=Aortopulmonary window, CoA=Coarctation of aorta,
ALCAPA=anomalous le coronary origin from pulmonary artery, HLHS=hypoplastic le heart syndrome,TA=Tricuspid atresia,
UVH=univentricular heart disease
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
483
TABLE 2. Preoperative TTE anatomical non-concordance.
TABLE 3. Non-concordance by pre-operative TTE.
Types of non-concordance N = 54 Percent
Systemicveindrainage 19 35.19%
Coronaryarterypattern 17 31.48%
Pulmonaryveindrainage 6 11.11%
RVOT 3 5.56%
ValvularPSGoodsizepulmonicvalve 2
SubvalvePSNosubvalvularobstruction 1
Others 9 16.67%
MissedASDsecundum 1
43leaetstruncalvalve 1
Missedcortiratum 1
Left-sidedRight-sidedarch 1
MildSeverehypoplasticLV 1
MissedAPwindow 1
MissedvegetationinRVOT 2
MildSeverMR,TR 1
of coronary artery 62 % (13 patients from total 21).
In non-concordant data group, 24 from 67 patients
(35.8%) had their operative procedures changed from the
preoperative plan. Cannulation technique was changed
in 6 patients (9%).
Most common non-concordant diagnosis between
preoperative echocardiographic reports and intraoperative
ndings was complete atrioventricular canal defects
(complete AV canal) and followed by TOF as show in
Table 3.
Demographic data between the two groups were
similar, No statistical dierences were found except for
rates of redo-operation which was higher in concordant
data group (13.3% VS 1.5%) as shown in Table 1. Redo-
operation in this study mean that patient who underwent
operation through previous incision.
For the analysis of intraoperative data, cardiopulmonary
bypass time was found to be longer in non-concordant
data group but p-value did not meet statistical signicance.
Morbidity, 30-day mortality, overall mortality were not
dierent between two groups (Table 4).
Wrong diagnosis Intraoperative diagnosis N = 14 Percent
(Preoperative diagnosis)
5
TOF DORV 4 35.71%
VSD/PS 1
CompleteAVcanal 6
PartialAVcanal 4 42.86%
TypeCTypeA 2
TransitionalAVcanal 1 7.14%
Truncusarteriosus 2 14.29%
Sanphasitvong et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
484
Original Article
SMJ
TABLE 4. Clinical outcome between echo groups.
Echo concordant group Echo non-concordant group p-value
(n=233) (n=67)
Optime,min (n=229) (n=65)
Median(Min,Max) 175(55,565) 190(90,410) 0.095
CPBtime (n=222) (n=67)
Median(Min,Max) 95(23,394) 110(41,215) 0.056
AoXtime (n=187) (n=63)
Median(Min,Max) 62(7,237) 74(19,164) 0.141
Complicationrate,n(%) 90(38.8%) 25(37.3%) 0.939
Deathrate,n(%) 13(5.6%) 5(7.5%) 0.565
30daysmortality,n(%) 10(4.3%) 3(4.5%) 1.000
1. Situsposition/Atrial-ventriclerelationship/ventricle–greatvesselrelationship
2. Systemicveindrainage
3. Pulmonaryveindrainage
4. Atrioventricularvalveanatomyandfunction
5. Aorticandpulmonicvalveanatomyandfunction
6. Truncalvalveanatomyandfunction(inTruncusarteriosus)
7. Shuntinatriallevel/ventricularlevel/Greatvessellevel
8. RVoutowtract
9. McGoonratio(inTOF,PA/VSD,Univentricularheart)
10. LVoutowtract
11. Coronaryanatomy(inTGA,TOFandPA/VSD)
12. Greatvesselsrelationship
13. Aorticarchanatomyandposition
14. LVandRVfunction
15. Assessmentofpulmonaryhypertension
16. Proximaldescendingaortaanatomyandgradientacrossstenosislesion(inCoarctationofaorta)
Abbreviations: TOF=Tetralogy of Fallot, TGA=Transposition of great artery
Fig 1. Criteria for completeness of echocardiography report.
DISCUSSION
Preoperative transthoracic echocardiography provided
concordant data in 77.7% of our study group. Approximately
one third of patients (32.3%) who had non-concordant
data, did not undergo surgery as planned preoperative and
approximately one tenth of patients (9%) had to change
the cannulation technique. e most common diering
data was information on systemic venous drainage such
as Lt SVC. Second most common non-concordant data
were details of the coronary artery pattern. All non-
concordant data found in our TGA patients were error in
coronary artery pattern. However, diering preoperative
data from transthoracic echocardiography did not aect
the morbidity and mortality rates of patients.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
485
Although no statistical signicant eect of non-
concordant preoperative data was found, there was
one Tetralogy Of Fallot (TOF) patient who underwent
total repair and preoperative echocardiography did
not demonstrate any le ventricular hypoplasia. Aer
the operation, he needed extracorporeal membrane
oxygenation (ECMO) insertion and reoperation.
Compared to other studies, Our study showed
higher ratio of non-concordant data (32.3 % VS 2-7%).
11-14
is may be because our study was a retrospective study
and we used only 2D transthoracic echocardiography.
Furthermore, some of primary investigators were fellows
who have less experience compared to certied pediatric
cardiologist. With respect to non-concordance data to
surgical procedure,
11-14
our result was comparable with
other studies with the rate of changed procedure about
30-40% without eect to surgical outcome, such as study
by Tworetzky W which was discovered in 42.1% (8 in
19 cases)
12
or study by Pfammatter JP in 33.3% (4 in 12
cases).
11
Our study found a high rate of non-concordant
data in coronary artery pattern dierent from other
studies which suggested that coronary anatomy could be
accurately assessed by preoperative echocardiogram (62%
VS 86%).
15
is nding may be aected by retrospective
design. Improvement of technology in echocardiography
machine may be ameliorate this problem.
We also noticed that patients who required staged
procedures tend to have a higher rate of concordant
preoperative echocardiographic data. is may be due
to the fact that these cases have had some data from
previous intraoperative ndings and may have also gotten
multiple echocardiographies before surgery. Cardiologists
might also pay more attention to those redo cases due
to awareness of its complexity in surgery.
Limitations of this study were its retrospective
design and high distribution of diseases. So it is dicult to
compare between groups of patients. Secondly, changing
surgical procedures were also decided subjectively by
individual surgeons. Lastly, Interpersonal variability
between surgeons performing the operation is another
uncontrolled factor in our study. Future study should
be more disease specic and prospective design.
CONCLUSION
Transthoracic echocardiography is valuable tool that
provides a lot of details on anatomical and physiological
data. However, in some complex congenital heart diseases
especially in TGA patients whom we not sure about
coronary artery anatomy, patients may still need additional
investigations such as cardiac catheterization, CT scan,
cardiac MRI in order to get adequate anatomical and
physiological data for planning the surgical procedure
and counselling purposes.
REFERENCES
1. Huhta JC, Glasow P, Murphy DJ, Jr., Gutgesell HP, Ott DA,
McNamara DG, et al. Surgery without catheterization for
congenital heart defects: management of 100 patients. J Am
Coll Cardiol 1987;9:823-9.
2. Carotti A, Marino B, Bevilacqua M, Marcelletti C, Rossi E,
Santoro G, et al. Primary repair of isolated ventricular septal defect
in infancy guided by echocardiography. Am J Cardiol 1997;79:1498-
501.
3. Shub C, Tajik AJ, Seward JB, Hagler DJ, Danielson GK. Surgical
repair of uncomplicated atrial septal defect without “routine”
preoperative cardiac catheterization. J Am Coll Cardiol 1985;6:49-54.
4. Freed MD, Nadas AS, Norwood WI, Castaneda AR. Is routine
preoperative cardiac catheterization necessary before repair of
secundum and sinus venosus atrial septal defects? J Am Coll
Cardiol 1984;4:333-6.
5. Lipshultz SE, Sanders SP, Mayer JE, Colan SD, Lock JE. Are
routine preoperative cardiac catheterization and angiography
necessary before repair of ostium primum atrial septal defect?
J Am Coll Cardiol 1988;11:373-8.
6. Gutgesell HP, Huhta JC, Latson LA, Hunes D, McNamara
DG. Accuracy of two-dimensional echocardiography in the
diagnosis of congenital heart disease. Am J Cardiol 1985;55:514-8.
7. Santoro G, Marino B, Di Carlo D, Formigari R, de Zorzi A,
Mazzera E, et al. Echocardiographically guided repair of tetralogy
of Fallot. Am J Cardiol 1994;73:808-11.
8. Zellers TM, Zehr R, Weinstein E, Leonard S, Ring WS, Nikaidoh
H. Two-dimensional and Doppler echocardiography alone can
adequately dene preoperative anatomy and hemodynamic
status before repair of complete atrioventricular septal defect
in infants < 1 year old. J Am Coll Cardiol 1994;24:1565-70.
9. Marek J, Skovranek J, Hucin B, Chaloupecky V, Tax P, Reich O,
et al. Seven-year experience of noninvasive preoperative diagnostics
in children with congenital heart defects: comprehensive
analysis of 2,788 consecutive patients. Cardiology 1995;86:488-95.
10. Pfammatter JP, Berdat PA, Carrel TP, Stocker FP. Pediatric
open heart operations without diagnostic cardiac catheterization.
Ann orac Surg 1999;68:532-6.
11. Pfammatter JP, Berdat P, Hammerli M, Carrel T. Pediatric
cardiac surgery after exclusively echocardiography-based
diagnostic work-up. Int J Cardiol 2000;74:185-90.
12. Tworetzky W, McElhinney DB, Brook MM, Reddy VM,
Hanley FL, Silverman NH. Echocardiographic diagnosis alone
for the complete repair of major congenital heart defects. J
Am Coll Cardiol 1999;33:228-33.
13. Sohn S, Kim HS, Han JJ. Pediatric cardiac surgery with
echocardiographic diagnosis alone. J Korean Med Sci 2002;17:463-7.
14. Lopes LM, Damiano AP, Moreira GN, Azevedo TJ, Nagamatsu
CT, Tavares GM, et al. e role of echocardiography as an
isolated method for indicating surgery in patients with congenital
heart disease. Arq Bras Cardiol 2005;84:381-6.
15. Fundora MP, Aregullin EO, Wernovsky G, Welch EM, Muniz
JC, Sasaki N, et al. Echocardiographic and Surgical Correlation
of Coronary Artery Patterns in Transposition of the Great
Arteries. Congenit Heart Dis 2016;11:570-7.
Sanphasitvong et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
486
Original Article
SMJ
KaninPruekprasert, M.D.*, Chanean Ruangsetakit, M.D.*, Chumpol Wongwanit, M.D.*, Khamin Chinsakchai,
M.D.*, Kiattisak Hongku, M.D.*, Nattawut Puangpanngam, M.D.*, Jutapim Khayankit, M.D.*, Tze Tec Chong,
M.D.**, Nuttawut Sermsathanasawadi, M.D.*
*Department of Surgery, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok 10700, ailand, **Department of Vascular Surgery,
Singapore General Hospital, Singapore.
Presence of Residual Venous Thrombus at Warfarin
Withdrawal: a Predictor for Recurrence after a
First Episode of Symptomatic Provoked Proximal
Deep Venous Thrombosis in Thai Population?
ABSTRACT
Objective: To assess the risk for venous thromboembolism (VTE) recurrence by presence of residual venous
thrombus (RVT) at warfarin withdrawal following symptomatic rst provoked proximal DVT.
Methods: Medical records of 45 consecutive patients with symptomatic rst provoked proximal DVTs who had
undergone warfarin surveillance for ≥ 3 months were reviewed retrospectively. Altogether, 22 patients discontinued
anticoagulation aer ≥ 3 months regardless of duplex ultrasonography results of RVT diagnosed by compression
ultrasonography. Another 23 patients discontinued anticoagulation aer the RVT disappeared. Primary outcome
was recurrent VTE.
Results: Four of the 45 patients experienced recurrent VTE (8.89%), including 2 (9.00%) of 22 patients who
had discontinued anticoagulant regardless of duplex ultrasonography results and 2 (8.70%) of 23 patients who
discontinued anticoagulation aer RVT disappearance (p = 0.963). All of the recurrent VTE were recurrent DVT.
Conclusion: RVT at warfarin withdrawal was not a predictor for recurrence VTE following a rst symptomatic
provoked proximal DVT.
Keywords: Deep vein thrombosis; duplex ultrasound; VTE management (Siriraj Med J 2019; 71: 486-490)
Corresponding author: Nuttawut Sermsathanasawadi
E-mail: nuttawut@gmail.com
Received 1 July 2019 Revised 6 November 2019 Accepted 8 November 2019
ORCID ID: http://orcid.org/0000-0002-0262-1438
http://dx.doi.org/10.33192/Smj.2019.72
INTRODUCTION
Because the recurrent rate of a proximal deep venous
thrombosis (DVT) of the leg that had been provoked by
surgery or by nonsurgical transient risk factors is high
during the rst 3 months, the American College of Chest
Physicians recommends treatment with anticoagulation for
3 months aer the diagnosis.
1
In patients with provoked
proximal DVTs, the estimated 5-year cumulative risk
of recurrent venous thromboembolism (VTE) after
discontinuing anticoagulant treatment was reported to
be 3% for surgery-provoked VTEs and 15% for VTEs
provoked by nonsurgical, reversible risk factors.
2
e
risk of recurrence of a provoked DVT is considered
low enough not to warrant prolonged anticoagulation,
although the risk of recurrence is not negligible in the
long term. In addition, the risk of recurrence associated
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
487
with provoked VTE varies among reports. Patients
with VTE due to reversible risk factors have a 10-year
cumulative recurrence rate of 22.5%: surgically provoked
VTE 11.4%, patients with medical illness 31.8%, women
with pregnancy-related or hormonal therapy 20.3%.
3
e presence of residual venous thrombus (RVT)
has been associated with an increased risk of thrombotic
recurrence in patients with provoked venous thrombosis.
Recurrent DVT occurs not only in previously thrombosed
veins but also at other sites.
4
RVT may indicate an underlying
prothrombotic state for the initial DVT.
4
The role of RVT as a predictor for a recurrent
proximal provoked DVT remains unclear. e DACUS
study showed that RVT was predictive of recurrence of
a provoked DVT. e AESOPUS study showed that the
advantage of prolonging warfarin treatment on the basis
of persistent RVT was more signicant aer unprovoked
DVTs than provoked DVTs.
5
Methods for measuring
RVTs in the AESOPUS and DACUS studies diered,
however, which might have aected the results of those
studies.
4,5
ere are no reported clinical studies on the relation
between RVT and the risk of recurrent DVT aer a rst
provoked proximal DVT in an Asian population. e
specic objective of the study was to assess the risk for
DVT recurrence conferred by RVT aer anticoagulant
withdrawal in a cohort of consecutive Asian patients
with symptomatic provoked proximal DVT.
MATERIALS AND METHODS
is study is based on a retrospective review of cohort
data. Consecutive patients attending a vascular surgery
clinic because of a rst episode of symptomatic proximal
provoked DVT of a lower extremity who were on warfarin
anticoagulation for at least 3 months between January
2011 and December 2013 were included. Patients with
unprovoked DVT requiring indenite anticoagulation,
patients with antiphospholipid syndrome, and those
with cancer were excluded.
Provoked DVT was dened as a DVT caused by
reversible risk factors, including immobilization for >
30 min during surgery with general anesthesia, history
of trauma during the previous 3 months, current use of
oral contraception or hormonal therapy, recent fracture
or plaster casting of a leg, connement to bed for 3 days
during the previous 3 months, pregnancy or puerperium,
and/or prolonged travel (> 4 h) during the previous 4
weeks.
Aer at least 3 months of treatment, some of the
vascular surgeons stopped warfarin following the American
College of Chest Physicians Guideline, neither duplex
ultrasonography (DUS) nor the D-dimer test was performed
in these patients. Although, some vascular surgeons prefer
to evaluate the patients with compression ultrasonography
and stopped warfarin aer the patients were free of residual
vein thrombus (RVT). e examination was performed
with the patient in supine position and the aected leg
externally rotated and slightly exed at the knee. CUS
images were obtained in transverse sections. Lumen
compressibility was then evaluated by gentle pressure
exerted using the ultrasound probe. e RVT diameter
was determined by measuring the distance between the
anterior and posterior walls of the vein before and aer
compression with the ultrasound probe.
Measurements of the common femoral vein diameter
were obtained 1 cm below the inguinal ligament. ose
of the popliteal vein were obtained at the most prominent
crease in the mid-popliteal fossa.
4
RVT was evaluated
following the techniques described in the DACUS and
AESOPUS studies.
4,5
In the DACUS study, the percentage
of the RVT was the ratio between the vein’s diameter before
and during compression.
4
Patients were considered not
to have an RVT when a persisting thrombus was ≤ 40%
of the venous measurement. Patients were considered
to have an RVT when a persisting thrombus of > 40%
was present in at least one of the two examined venous
segments.
4
In the AESOPUS study, RVT was dened as
being present if the vein’s transverse diameter was > 2
mm at maximum compression or absent if the transverse
diameter was ≤ 2 mm.
5
In this study, an RVT was considered present when
the RVT measured by at least one technique was deemed
positive. An RVT was considered absent when both
techniques produced negative results.
In patients who performed to measure RVT, warfarin
was stopped aer no RVT was found. Patients in whom
RVT was considered present continued to receive warfarin.
ese patients underwent CUS evaluation every 3 months
until the RVT had disappeared. CUS was performed
whenever a DVT was suspected to have recurred.
e study outcome was recurrent VTE. In cases of
recurrence, CUS results were compared with those of the
previous examination. Recurrent DVT was diagnosed if
a previously fully compressible segment (contralateral
or ipsilateral) was no longer compressible or there was a
≥ 4 mm increase in the diameter of the residual thrombus
during compression. In dicult cases, CUS was repeated
(aer 5-7 days) or computed tomography venography was
performed. e lower limb veins were evaluated using
a GE LOGIC 9 system (GE Healthcare, Milwaukee, WI,
USA) with 5- to 10-MHz linear transducers. Recurrent
PE (pulmonary embolism) was diagnosed if clinical
Pruekprasert et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
488
Original Article
SMJ
suspicious, with conrmation by computed tomography
pulmonary angiography.
e ethics committee of the Siriraj Institutional
Review Board approved this study (Si 002/2019).
For the statistical analyses, continuous variables are
described as the mean and standard deviation (SD) or the
median and range, as appropriate. Categorical variables
are described as the number and percentage. e t-test or
Mann-Whitney U test was used to compare continuous
variables, as appropriate. e χ
2
test or Fisher’s exact test
was used to compare categorical variables. e Kaplan-
Meier method was used to estimate the recurrence-free
survival curve and the recurrence rate. Cox regression
analysis was used to investigate the association between
RVT status (assessed by DUS) and recurrence of DVT.
ose associations were evaluated using the hazard ratio
and corresponding 95% condence intervals. All p values
were two-tailed with a signicance level of 0.05. Data
were recorded and analyzed using PASW Statistics 18.0
soware (SPSS Inc., Chicago, IL, USA).
RESULTS
ere were 45 patients with provoked DVT in this
study. ere were 35 (77.78%) women and 10 (22.22%) men.
Twenty-two patients with provoked DVTs discontinued
the anticoagulant aer ≥3 months of treatment without
DUS evaluation. e other 23 patients discontinued it aer
the RVT had disappeared. e mean age was 60.0 ± 18.9
years in patients who discontinued anticoagulant without
DUS and 63.9 ± 19.4 years in patients who discontinued
it aer RVT disappearance. e mean follow-up time
was 121.2 ± 40.2 weeks (range 50.0-207.0 weeks) for
patients who discontinued the anticoagulant without
DUS and 118.3 ± 38.8 weeks (range 52-243 weeks) for
patients who discontinued it aer RVT disappearance.
Factors that provoked the DVT are shown in Table 1.
During the follow-up period, four recurrent DVTs
(8.89%) were recorded from among 45 patients with
a provoked DVT. Two of these patients (9.00%) had
discontinued anticoagulant without DUS and two (8.70%)
had discontinued anticoagulant aer RVT disappearance
(p =0.963) (Fig 1). ere was no recurrent PE in this
study.
TABLE 1. Demographic data from Asian patients with their rst symptomatic provoked DVT who discontinued
anticoagulation.
Patients discontinued anticoagulant
Without DUS After no RVT P-value
(n = 22) (n = 23)
Age
Mean±SD,year 60.0±18.9 63.9±19.4 0.502
Sex
Female,n(%) 18(81.8) 17(79.9) 0.722
Anticoagulantduration,week
Median(range) 30.5(13.4-87.0) 32.9(13.0-75.7) 0.901
Mean±SD 36.9±21.4 36.6±18.8
Provokedfactors
Surgery/Trauma 11(50) 11(47.8)
Immobilization/Illness 7(31.8) 9(39.1)
Hormonalused 2(9.1) 2(8.7) 0.520
Prolongedtravel 0(0) 1(4.3)
Pregnancy 2(9.1) 0(0)
Abbreviations: DVT = deep vein thrombosis; DUS = Duplex ultrasonography; RVT = residual vein thrombus
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
489
Fig 1. Kaplan-Meier curves for recurrent deep vein thrombosis (DVT) in patients who stopped taking an anticoagulant (warfarin) aer no
residual vein thrombosis (RVT) was detected and those who stopped the anticoagulant without duplex ultrasonography (DUS) conrmation.
DISCUSSION
Patients who develop a DVT associated with a
transient surgical procedure (surgically provoked DVT)
have a low rate of recurrence, estimated at 0.7% per year
for the 2 years aer stopping anticoagulation therapy.
6
Patients who develop a non-surgically provoked DVT
(e.g., due to immobilization, pregnancy, use of estrogen-
containing contraception) have a slightly higher risk of
recurrence, estimated at 4.2% per year for the same 2-year
period aer discontinuing anticoagulation therapy.
6
e
risk of recurrence aer the appearance of a provoked
DVT is considered low enough not to warrant prolonged
anticoagulation, although the risk of recurrence is not
negligible in Asian populations.
2
In our study of ai patients with provoked DVTs,
the annual risk of recurrence aer warfarin withdrawal was
6.77%. is risk was somewhat higher than the reported
annual risk of recurrence in subjects with provoked
DVTs in Caucasian populations.
2
DVTs provoked by a major reversible risk factor,
such as recent surgery, had a low risk of recurrence.
Although the risk of recurrence in patients with DVTs
provoked by a nonsurgical trigger was higher than that
for patients with DVTs provoked by surgery, the risk
was still low.
2
Provoked DVTs in a ai population were at a
suciently low risk of recurrence to recommend stopping
anticoagulants at 3 months. If the provoking factor is
incompletely resolved, however, it is appropriate to treat
the patients longer than 3 months.
9
Our results were in line with those of the AESOPUS
study and a study by Cosmi et al. in which an RVT was
not predictive for recurrence aer provoked DVT.
5,7
Our results, however, were dierent from those of the
DACUS study in which RVT was predictive of recurrence
following a provoked DVT.
4
e reproducibility of RVT measurements could
be an issue across studies because of the lack of widely
accepted criteria for dening vein recanalization.
8
For
example, the AESOPUS study adopted the method of
Prandoni et al.,
3,5
whereas the RVT in the DACUS study
was dened as a residual thrombus occupying more
than 40% of the vein’s area calculated in the absence of
compression.
4,5
Although measurements of RVT in the
AESOPUS and DACUS studies were dierent, which
could aect the results of the study, our center uses
Pruekprasert et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
490
Original Article
SMJ
both techniques to dene the presence of an RVT, with
a positive result from at least one of them indicating the
need to continue anticoagulation.
Even though the presence of an RVT was not a risk
factor for recurrence in ai patients experiencing their
rst symptomatic provoked proximal DVT at the time
of warfarin withdrawal, DUS should be performed to
dene the baseline characteristics of the vein’s lumen
for detecting a new thrombus in the future.
e small number of patients limits the statistical
power of this study. Larger series are needed to clarify the
relation between the presence of an RVT and recurrence of
a DVT in Asian patients with a rst episode of symptomatic
provoked proximal DVT.
CONCLUSION
e presence of an RVT at the time of warfarin
withdrawal was not a predictor for recurrence of the
VTE in ai patients with a rst episode of symptomatic
provoked proximal DVT.
ACKNOWLEDGMENTS
e authors thank Miss Supaporn Tunpornpituk
and Dr. Sasima Tongsai for statistical assistance and
Miss Phakawan Phutthakunphithak for data collection.
Declaration of Conicting Interests: e authors declared
no potential conicts of interest with respect to the
research, authorship, and/or publication of this article.
Funding: is study was supported by the Faculty of Medicine
Siriraj Hospital, Mahidol University. KaninPruekprasert,
Chanean Ruangsetakit, Chumpol Wongwanit, Khamin
Chinsakchai, Kiattisak Hongku, Nattawut Puangpanngam,
and Nuttawut Sermsathanasawadi are supported by
Chalermphrakiat grants, Faculty of Medicine Siriraj
Hospital, Mahidol University.
REFERENCES
1. Kearon C,Akl EA,Ornelas J,Blaivas A,Jimenez D,Bounameaux
H,et al. Antithrombotic erapy for VTE Disease: CHEST
Guideline and Expert Panel Report. Chest 2016;149:315-52.
2. Kearon C,Akl EA,Comerota AJ,Prandoni P,Bounameaux
H,Goldhaber SZ,et al. Antithrombotic therapy for VTE
disease: Antithrombotic erapy and Prevention of rombosis,
9
th
ed: American College of Chest Physicians Evidence-Based
Clinical Practice Guidelines. Chest 2012;141:e419S-94S.
3. Prandoni P,Noventa F,Ghirarduzzi A,Pengo V,Bernardi E,
Pesavento R,et al. e risk of recurrent venous thromboembolism
aer discontinuing anticoagulation in patients with acute
proximal deep vein thrombosis or pulmonary embolism. A
prospective cohort study in 1,626 patients. Haematologica
2007;92:199-205.
4. Siragusa S,Malato A,Anastasio R,Cigna V,Milio G,Amato C,
et al. Residual vein thrombosis to establish duration of
anticoagulation aer a rst episode of deep vein thrombosis: the
Duration of Anticoagulation based on Compression
UltraSonography (DACUS) study. Blood 2008;112:511-5.
5. Prandoni P,Prins MH,Lensing AW,Ghirarduzzi A,Ageno
W,Imberti D,et al. Residual thrombosis on ultrasonography
to guide the duration of anticoagulation in patients with deep
venous thrombosis: a randomized trial. Ann Intern Med 2009;
150:577-85.
6. Iorio A,Kearon C,Filippucci E,Marcucci M,Macura A,Pengo
V,et al. Risk of recurrence aer a rst episode of symptomatic
venous thromboembolism provoked by a transient risk factor:
a systematic review. Arch Intern Med 2010;170:1710-6.
7. Cosmi B,Legnani C,Cini M,Guazzaloca G,Palareti G. D-dimer
and residual vein obstruction as risk factors for recurrence
during and aer anticoagulation withdrawal in patients with
a rst episode of provoked deep-vein thrombosis. romb
Haemost 2011;105:837-45.
8. Linkins LA, Stretton R, Probyn L, Kearon C. Interobserver
agreement on ultrasound measurements of residual vein
diameter, thrombus echogenicity and Doppler venous ow in
patients with previous venous thrombosis. romb Res 2006;
117:241-7.
9. Kearon C, Akl EA. Duration of anticoagulant therapy for deep
vein thrombosis and pulmonary embolism. Blood 2014;123:
1794-801.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
491
INTRODUCTION
Arteriovenous hemodialysis access facilitates a
connection between the patient and the hemodialysis
system in end-stage renal disease patients. Current
Nattawut Puangpunngam, M.D.*, Nathakorn Supokaivanich, M.D.*, Chanean Ruangsetakit, M.D.*, Chumpol
Wongwanit, M.D.*, Nuttawut Sermsathanasawadi, M.D., Ph.D.*, Khamin Chinsakchai, M.D.*, Suteekhanit
Hahtapornsawan, M.D.*, Kiattisak Hongku, M.D.*, Pramook Mutirangura, M.D.*, Somrach amtorawat,
M.D.**, Satit Rojwatcharapibarn, M.D.**, Walailak Chaiyasoot, M.D.**, Jirawadee Yodying, M.D.**, Trongtum
Tongdee, M.D.**
*Department of Surgery, **Department of Radiology, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, ailand
Endovascular Thrombectomy versus Open
Surgical Thrombectomy for Thrombosed
Arteriovenous Hemodialysis Graft
ABSTRACT
Objective: e aim of this study was to investigate the procedure success rate, one-year primary patency rate,
one-year secondary patency rate, and complications compared between endovascular therapy and open surgical
thrombectomy for treatment of thrombosed arteriovenous hemodialysis gra.
Methods: is retrospective chart review included patients with thrombosed arteriovenous hemodialysis gra
who were treated at the Division of Vascular Surgery, Department of Surgery, Faculty of Medicine Siriraj Hospital,
Mahidol University, Bangkok, ailand during January 2012 to December 2016. Demographic, gra type, time
before treatment, thrombus removal technique, additional technique, procedure success rate, operative, primary
and secondary patency, follow-up time, and complication data were collected.
Results: Seventy-four thrombosed dialysis gras were included. Twenty-ve and 49 gras underwent endovascular
therapy and open surgical thrombectomy, respectively. ere was no signicant dierence in demographic data,
gra type, or adjunct procedure between groups. e procedure success rate was 92% and 98% in the endovascular
group and thrombectomy group, respectively (p=0.262). e one-year primary patency rate was 26% in the
endovascular group, and 33% in the thrombectomy group (p=0.054). One-year secondary patency rate was 82.6%
in the endovascular group, and 56.3% in the thrombectomy group (p=0.122).
Conclusion: No signicant dierences were observed between groups for procedure success rate or 1-year primary
patency rate; however, the one-year secondary patency rate in the endovascular group was signicantly better than
in the thrombectomy group. No dierence in complications was observed between groups.
Keywords: Endovascular thrombectomy; open surgical thrombectomy; thrombosed arteriovenous hemodialysis
gra (Siriraj Med J 2019; 71: 491-498)
Corresponding author: Nattawut Puangpunngam
E-mail: nattawut.pua@mahidol.ac.th
Received 22 May 2019 Revised 21 November 2019 Accepted 22 November 2019
ORCID ID: http://orcid.org/0000-0002-4874-5285
http://dx.doi.org/10.33192/Smj.2019.73
guideline
1
recommends the use of native arteriovenous
stula (AVF) as the rst choice for hemodialysis access.
However, arteriovenous hemodialysis gra (AVG) is
still preferred in many patients due to vessel-related
Puangpunngam et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
492
Original Article
SMJ
limitations, and the fact that some patients in need of
urgent hemodialysis cannot wait for AVF maturation
aer surgery. Although AVG does not require time for
maturation, it is easier to puncture than AVF and it has
a higher likelihood of complications. One of the most
important complications of AVG is acute thrombosis
that leads to gra blockage that results in loss of function
and access.
Restoration of thrombosed hemodialysis AVG can be
managed by either endovascular therapy or open surgical
thrombectomy. Several studies have compared outcome
of treatment between endovascular therapy and open
surgical thrombectomy. A 2002 meta-analysis review by
Green, et al.
2
found that open surgical thrombectomy
had a signicantly better outcome relative to both failure
rate and 90-day patency rate.
Another meta-analysis review by Tordior, et al.
3
in 2009 analyzed 78 studies that included 8 randomized
controlled trials and 1 meta-analysis. at group found
results similar to those reported by Green, et al. except
that a comparison between studies conducted before
and after 2002 revealed no significant difference in
outcome between the two treatment groups in studies
conducted aer 2002. is dierence may be explained
by improvements in endovascular techniques over time.
e most recent retrospective study in ailand was
published in 2015 by Hongsakul, et al.
4
at group also
found no signicant dierence in outcomes between
the two treatment options.
Objective
e aim of this study was to investigate the procedure
success rate, one-year primary patency rate, one-year
secondary patency rate, and complications compared between
endovascular therapy and open surgical thrombectomy
for treatment of thrombosed arteriovenous hemodialysis
gra.
MATERIALS AND METHODS
Population
is retrospective chart review included patients
(age >18 years) with acute thrombosed arteriovenous
hemodialysis gra (onset 0-14 days) who were treated
at the Division of Vascular Surgery, Department of
Surgery, Faculty of Medicine Siriraj Hospital, Mahidol
University, Bangkok, ailand during the January 2012 to
December 2016 study period. Demographic, gra type, time
before treatment, thrombus removal technique, adjunct
procedure, procedure success rate, operative, primary
and secondary patency, follow-up time, and complication
data were collected, recorded, and analyzed. Included
data were retrieved from the databases of the Division of
Vascular Surgery of the Department of Surgery and the
Division of Interventional Radiology of the Department
of Radiology at our center. Only patients who were
follow-up for at least one year were included. Patients
who had thrombosis of AVG related to infected gra
or aneurysm were excluded. e treatment method was
determined by the patient’s attending vascular surgeon
or nephrologist. e protocol for this study was approved
by the Siriraj Institutional Review Board (Si 603/2559).
Open surgical thrombectomy
e procedure was performed by vascular surgeons
under general or local anesthesia. e only thrombus
removal technique used in this group was surgical
thrombectomy with Fogarty catheter. In the operating
room, the hemodialysis gra was opened directly at
the venous limb and thrombectomy was performed
by advancing the Fogarty thrombectomy catheter
through the gra into the native venous outow. e
same thrombectomy technique was also performed on
the arterial limb. Angiography was then performed
to evaluate for residual thrombus, and to investigate
for underlying stenotic lesions of arteries and veins in
every cases. Surgical thrombectomy would be repeated
if residual thrombus was observed in the gra. Aer
adequate thrombus removal was conrmed, then any
needed adjunct procedure was selected according to
angiographic nding. If underlying lesion was revealed,
balloon angioplasty was performed to correct luminal
stenosis of vessels. Stent was deployed if there was residual
stenosis greater than 30% aer balloon angioplasty. In
case of recurrent stenosis of venous outow, gra revision
using a short bypass gra to more central venous outow
was performed in selected cases. Final angiogram was
performed to assess the patency of gra and anastomosis.
Fig 1. Open surgical thrombectomy
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
493
Fig 2. (A) Angiogram aer thrombus removal showing long stenosis at venous outow. (B, C) Angioplasty with stent was performed and
nal angiogram revealed improvement in lumen stenosis.
A B C
Endovascular therapy
All endovascular procedures were performed by
interventional radiologists. e majority of these procedures
were performed under local anesthesia. A needle was used
to puncture the gra, aer which an introducer sheath
was inserted into the gra. In this group, three dierent
endovascular thrombus removal techniques were used, as
follows. Method 1: Catheter aspiration technique using
5-6 Fr. catheter to aspirate thrombus from the AVG
until the thrombus was removed. Method 2: Mechanical
thrombectomy technique that uses a thrombectomy device
that generates mechanical force to dissolve the thrombus
and aspirate clot fragments into a catheter. Method 3:
Catheter-directed thrombolysis (CDT) technique that
uses a catheter to directly infuse a thrombolytic agent
(recombinant tissue plasminogen activator [rt-PA]) into
the thrombus. For all methods, angiography was performed
aer thrombus removal to check for residual thrombus
and to reveal underlying stenotic lesions of arteries and
veins. Adjunct procedure was selected according to
angiographic nding. If residual thrombus was found,
the same thrombus removal technique could be repeated,
or another of the 3 techniques could be used to ensure
adequate thrombus removal. en underlying lesion was
revealed, balloon angioplasty was performed to correct
luminal stenosis of vessels. Stent was deployed if there
was residual stenosis greater than 30% aer balloon
angioplasty. Aer the aforementioned procedures were
completed, nal angiogram was performed to conrm
the patency of gra and anastomosis.
Follow-up
All patients were clinically followed-up at the outpatient
vascular surgery clinic or nephrology clinic at 1 month,
3 months, and then every 3 months thereaer. If we found
any problem in dialysis or physical examination, duplex
ultrasound was used to evaluate gra function. Gra
dysfunction was conrmed before further correction
by duplex ultrasound in all cases. Decision to abandon
gra was maded by vascular surgeon who performed
ultrasound and we abandoned it only very poor patency
aer restoration of ow (less than 1 months).
Denitions
According to standard practice guideline published
by the Society of Intervention Radiology
5
, procedural
success is dened as restoration of ow in the dialysis
gra and palpable thrill. Primary patency is dened as
the time interval aer the procedure until the next access
thrombosis or rst subsequent intervention. Secondary
patency is dened as the time interval aer the procedure
until the access was surgically revised or abandoned.
Major complication is dened as complication that
requires additional treatment or that results in permanent
sequelae or death.
Statistical analysis
We compared demographic data, clinical data,
procedure-related data, complications, and outcomes
of procedures between groups. Paired-samples t-test
was used to compare continuous variables, and chi-
square test was used to compare categorical variables.
Normally and non-normally distributed continuous
data are reported at mean ± standard deviation and
median and range, respectively. Categorical data are
shown as frequency and proportion. Survival analysis
was performed using Kaplan-Meier method and log
rank test was used to compare patency. A p-value less
than 0.05 indicates statistical signicance. SPSS Statistics
data analysis program (SPSS, Inc., Chicago, IL, USA)
was used to perform all data analyses.
Puangpunngam et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
494
Original Article
SMJ
RESULTS
Of the 74 patients that were included in this study,
25 were in the endovascular group, and 49 were in the
open thrombectomy group. e mean age of patients was
60.68±14.37 years in the endovascular therapy group,
and 64.33±14.81 years in the open thrombectomy group.
All patients had hypertension as a comorbidity, and
diabetes was the second most common comorbidity in
both groups. Loop forearm and loop upper arm AVG
were the most common gra types in the endovascular
and open thrombectomy groups, respectively.
ere was no signicant dierence between groups
for time of gra function aer creation, time since last
gra revision, or number of prior gra thrombectomies.
Eighty percent of patients in the endovascular group and
55% of patients in the open thrombectomy group that
developed failure of hemodialysis access presented with
clinical signs and symptoms of thrombosis. Mostly of
patients had total gra thrombosis, only 11 of 25 patients
in the endovascular group presented with clinical partial
gra thrombosis. Median time from gra thrombosis to
treatment was 2 days (range: 1-14) in the endovascular
group, and 1 day (range: 1-14) in the open thrombectomy
group (p=0.078). A summary of patient and gra data
is shown in Table 1.
All endovascular procedures were performed under
local anesthesia. Eighty percent of the cases used catheter
aspiration as the thrombus removal technique. Mechanical
thrombectomy and catheter-directed thrombolysis were
also used as single or combined thrombus removal
technique. e vast majority (96%) of procedures required
angioplasty for stenosis lesion correction. Stent was
required in 1 case.
In the open thrombectomy group, 73.5% of procedures
were performed under local anesthesia, and additional
angioplasty was needed in 90% of cases. Aer angioplasty,
stent was required in 2 cases, and jump gra procedure
was required in 3 cases due to failure of angioplasty
(Table 2). e most common gra stenotic lesions in both
groups were venous anastomosis and venous outow.
ere was no signicant dierence between the
endovascular and open thrombectomy groups relative to
procedure success rate (92% vs. 98%; p=0.262). Median
operative time was signicantly shorter in the endovascular
group than in the open thrombectomy group (17 vs. 120
minutes; p=0.001). ere was no mortality in this study,
and both groups had a median hospital stay of 3 days.
ere was no signicant dierence between groups for
minor procedure-related complications (Table 3).
Data Endovascular Open P value
(n=25) (n=49)
Age (years), mean±SD 60.68±14.37 64.33±14.81 0.315
Gender, n (%) 0.083
Male 8(32.0%) 26(53.1%)
Female 17(68.0%) 23(46.9%)
Graft type, n (%) 0.207
Loopforearm 10(43.5%) 13(27.1%)
Loopupperarm 11(47.8%) 28(58.3%)
Straightupperarm 2(8.7%) 2(4.2%)
Femoral 0(0.0%) 5(10.4%)
Time before treatment (days), 2(1-14) 1(1-7) 0.078
median (min-max)
TABLE 1. Patient demographic and gra data.
A p-value<0.05 indicates statistical signicance
Abbreviation: SD = standard deviation
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
495
TABLE 2. Procedure-related data.
TABLE 3. Operative, follow-up, patency, and complication data.
Data Endovascular Open P value
(n=25), n (%) (n=49), n (%)
Thrombus removal technique
Openthrombectomy 0(0.0%) 49(100%)
Mechanicalthrombectomy 8(34.8%) 0(0.0%)
CDT 3(13.0%) 0(0.0%)
Catheteraspiration 20(80.0%) 0(0.0%)
Adjunct procedure after thrombus removal
Angioplasty 25(100%) 44(89.8%) 0.098
Stent 1(4.0%) 2(4.1%) 0.987
rt-PA 0(0.0%) 0(0.0%)
Jumpgraft 0(0.0%) 3(6.1%) 0.207
A p-value <0.05 indicates statistical signicance
Abbreviations: CDT = catheter-directed thrombolysis, rt-PA = recombinant tissue plasminogen activator
Operative data Endovascular Open P value
(n=25) (n=49)
Proceduresuccessrate,n(%) 23(92.0%) 48(98.0%) 0.262
Medianoperativetime(minutes) 17 120 0.001
Follow-uptime(months),mean±SD 24.0±3.9 17.6±4.0 0.111
Primarypatencytime(months),median(min-max) 6.37(0.79-54.47) 6.37(0.13-43.04) 0.703
Secondarypatencytime(months),median(min-max) 25.75(1.58-54.47) 14.48(0.13-57.56) 0.037
Minorcomplication,n(%) 2(8.6%) 5(9.5%) 0.546
A p-value <0.05 indicates statistical signicance
Abbreviation: SD = standard deviation
As shown in Table 3, the median average primary
patency times in the endovascular and open thrombectomy
groups were the same at 6.37 months (p=0.703). In contrast,
the median secondary patency time was signicantly
longer in the endovascular group (25.75 months) than in
the open thrombectomy group (14.48 months) (p=0.037).
e 1-year primary patency rate was 26.1% in the
endovascular group, and 33.3% in the open thrombectomy
group (p=0.875) (Fig 3). e 1-year secondary patency
rate was 82.6% in the endovascular group, and 56.3% in
the open thrombectomy group (p=0.028) (Fig 4). e
mean follow-up time was 24 months and 17.6 months
in the endovascular group and open thrombectomy
group, respectively (p=0.111).
Puangpunngam et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
496
Original Article
SMJ
Fig 3. Kaplan-Meier curve for primary patency rate
Fig 4. Kaplan-Meier curve for secondary patency rate
DISCUSSION
AVG is preferred as permanent hemodialysis access
in end-stage renal disease patients with limited time to
rst hemodialysis, but access thrombosis is a common
complication. Endovascular therapy for thrombosed
hemodialysis gra is now used worldwide as an alternative
to open surgical thrombectomy. Early studies found
open surgical thrombectomy to have better outcome
than endovascular therapy. Green, et al.
2
reviewed 7
randomized controlled trials and found that open surgical
thrombectomy had signicantly better outcome for both
failure rate (risk ratio [RR]: 1.90, 95% condence interval
[CI]: 1.32-2.73; p=0.0005) and 90-day patency rate (RR:
1.22, 95% CI: 1.05-1.40; p=0.007) than endovascular
therapy. However, a 2009 meta-analysis review by Tordior,
et al.
3
and a 2015 retrospective study by Hongsakul, et al.
4
both reported equivalent outcome between groups.
e 92% and 98% procedure success rates in the
endovascular and open thrombectomy groups, respectively,
in our study were comparable to the rates reported in the
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
497
aforementioned meta-analysis (92-95% and 79-100%,
respectively) and retrospective study (94% and 93.8%,
respectively).
e time interval from gra thrombosis to intervention
is one of the factors that signicantly inuences the
procedure success rate. In the present study, the median
duration before treatment was longer in the endovascular
group than in the open thrombectomy group (2 days vs.
1 day), and there was some dierence in the procedure
success rate (92% vs. 98%), but the dierence for both
comparisons failed to achieve statistical signicance.
Our comparative analysis revealed operative time
to be one of the variables most dierent between groups.
Open thrombectomy had a signicantly longer operative
time than endovascular method. is may be explained
by additional gra open and closure steps, and the need
for general or regional anesthesia in some patients.
Although, longer operative time was required, open
thrombectomy still had similar procedural success rate
and primary gra patency compared to endovascular
procedure. Moreover, the rate of complications was
non-signicantly dierent between groups.
Mechanical thrombectomy and catheter-directed
thrombolysis were reported to be effective and safe
techniques for endovascular gra thrombectomy.
4,6-10
Catheter aspiration, which accounted for 80% of the
endovascular procedures performed in our study, is a
simple and eective technique that requires no expensive
surgical devices, but it yields comparable results relative
to procedure success rate and gra patency.
We found signicant lesions aer open thrombectomy
that need intervention to correct lesions more than 90%
of cases. ese data near the result in previous report
(78%) from Kuma, et al.
11
Regarding one-year primary patency rate, the
26% and 33% rates of one-year primary patency in the
endovascular and open thrombectomy groups in our
study were near the upper end of the range reported by
Tordoir, et al. (8-29% and 10-30%), and similar to the
rate reported by Hongsakul, et al. (28% and 30%).
Concerning the one-year secondary patency rate,
the 82.6% and 56.3% rates of one-year secondary patency
in the endovascular and open thrombectomy groups in
the present study are in some ways dierent and in some
ways comparable to the results reported by Tordoir, et al.
(23-62% and 27-65%) and Hongsakul, et al. (54.3% and
57%). Recent study from Koraen-Smith, et al.
12
show the
same results that thrombolysis group had better assist
primary patency.
Our study showed markedly better one-year
secondary patency compared to previous studies due
to selection bias and dierent severity of thrombosis in
the endovascular group. We found that 11 of 25 patients
in the endovascular group presented with clinical partial
gra thrombosis, so lower severity of gra thrombosis
may lead to better secondary patency rate. Furthermore,
early follow-up at 1 and 3 months in our study resulted
in early additional endovascular therapy for partial gra
stenosis before complete gra thrombosis occurred.
To summarize, our study found advantage of
endovascular therapy to open thrombectomy in thrombosed
AVBG, with signicantly better one-year secondary
patency, signicantly shorter operative time, and all
endovascular procedures could be performed under
local anesthesia.
Limitations
is study has some limitations. First, the retrospective
design of our study suggests the potential for missing or
incomplete data. Second, our center is a national tertiary
referral hospital that is oen referred complex cases. As
such, our results may not reect or be generalizable to
other care settings. ird, our overall study population was
relatively small, and we had a small number of patients
in the endovascular group. ese factors suggest the
possibility that our study may have lacked the statistical
power to identify all signicant dierence and associations
between groups. Fourth, we were not able to include detail
relating to improvement in lumen stenosis from post-
procedure angiogram, which could inuence gra patency.
Further randomized controlled study is recommended
to determine the advantage of endovascular therapy
over open thrombectomy in patients with thrombosed
AVBG.
CONCLUSION
No signicant dierences were observed between
groups for procedure success rate or 1-year primary patency
rate; however, the one-year secondary patency rate in the
endovascular group was signicantly better than in the
thrombectomy group. No dierence in complications
was observed between groups. ese results suggest
endovascular treatment as a safe and ecacious treatment
alternative in patients with thrombosed arteriovenous
hemodialysis gra.
ACKNOWLEDGMENTS
e authors gratefully acknowledge Dr. Sasima
Tongsai of the Division of Clinical Epidemiology,
Department of Research, Faculty of Medicine Siriraj
Hospital, Mahidol University, Bangkok, ailand for
assistance with statistical analysis.
Puangpunngam et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
498
Original Article
SMJ
Conict of interest declaration
All authors declare no personal or professional
conicts of interest, and no nancial support from the
companies that produce and/or distribute the drugs,
devices, or materials described in this report.
Funding disclosure
is was an unfunded study.
REFERENCES
1. Gilmore J. KDOQI clinical practice guidelines and clinical practice
recommendations--2006 updates. Nephrol Nurs J 2006;33:
487-8.
2. Green LD, Lee DS, Kucey DS. A metaanalysis comparing surgical
thrombectomy, mechanical thrombectomy, and pharmacomechanical
thrombolysis for thrombosed dialysis gras. J Vasc Surg 2002;
36:939-45.
3. Tordoir JH, Bode AS, Peppelenbosch N, van der Sande FM, de
Haan MW. Surgical or endovascular repair of thrombosed dialysis
vascular access: is there any evidence? J Vasc Surg 2009;50:
953-6.
4. Hongsakul K, Rookkapan S, Sungsiri J, Boonsrirat U, Kritpracha
B. Pharmacomechanical Thrombolysis versus Surgical
rombectomy for the Treatment of rombosed Haemodialysis
Gras. Annals of the Academy of Medicine, Singapore. 2015;
44(2):66-70.
5. Aruny JE, Lewis CA, Cardella JF, Cole PE, Davis A, Drooz AT,
et al. Quality improvement guidelines for percutaneous
management of the thrombosed or dysfunctional dialysis
access. J Vasc Interv Radiol 2003;14(9 Pt 2):S247-53.
6. Dougherty MJ, Calligaro KD, Schindler N, Raviola CA, Ntoso A.
Endovascular versus surgical treatment for thrombosed
hemodialysis gras: A prospective, randomized study. J Vasc
Surg 1999;30:1016-23.
7. Sofocleous CT, Hinrichs CR, Weiss SH, Contractor D, Barone
A, Bahramipour P, et al. Alteplase for hemodialysis access
gra thrombosis. J Vas Interv Radiol 2002;13:775-84.
8. Karatepe C, Aldemir M, Cinar B, Onalan A, Issever H, Goksel OS.
Midterm Results Following Percutaneous Rotational rombectomy
for Acute rombotic Occlusions of Prosthetic Arteriovenous
Access Gras. Int Surg 2015;100:1249-54.
9. Chan PG, Goh GS. Safety and ecacy of the AngioJet device
in the treatment of thrombosed arteriovenous stula and gras:
A systematic review. J Vasc Access 2018;19:243-51.
10. Marcelin C, D’Souza S, Le Bras Y, Petitpierre F, Grenier N, van
den Berg JC, et al. Mechanical Thrombectomy in Acute
rombosis of Dialysis Arteriovenous Fistulae and Gras
Using a Vacuum-Assisted rombectomy Catheter: A Multicenter
Study. J Vasc Interv Radiol 2018;29:993-7.
11. Kumar SD, Jain KM, Jain S, Munn JS, Rummel MC. Fistulogram
aer arteriovenous dialysis gra thrombectomy should be
mandatory. Am Surg 2008;74:1154-8.
12. Koraen-Smith L, Krasun M, Bottai M, Hedin U, Wahlgren
CM, Gillgren P. Haemodialysis access thrombosis: Outcomes
aer surgical thrombectomy versus catheter-directed thrombolytic
infusion. J Vasc Access 2018;19:535-41.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
499
Banjerd Praditsuktavorn, M.D., Chumpol Wongwanit, M.D., Tiwa Chaisongrit, M.D., Chanean Ruangsetakit,
M.D., Kiattisak Hongku, M.D., Nattawut Puangpunngam, M.D.
Department of Surgery, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok 10700, ailand.
Outcomes of Autogenous Snuffbox Radiocephalic
Arteriovenous Fistula-First Strategy for
Hemodialysis Access
ABSTRACT
Objective: An autogenous arteriovenous stula (AVF) has been recommended as the rst-line hemodialysis access
option. A distal radiocephalic (snuox) arteriovenous stula (SBAVF), the most distally located AVF, provides an
extended vascular access area, allows future AVF correction in the proximal part, and oers relatively easy surgical
access. is study aimed to evaluate the outcomes of patients who had SBAVFs as their rst option for dialysis access.
Methods: e medical-record electronic database of Siriraj Hospital, Bangkok, ailand, was retrospectively reviewed
to evaluate the outcomes of patients receiving SBAVFs July 2013-December 2016 with over 12 months of follow-up.
Results: Of 39 patients, SBAVFs were successfully created in 33 patients (84.6%). Early thrombosis was observed in
1 patient. Steal syndrome, distal thrombosis, and complications of high-venous ow were not detected. e primary,
primary-assisted, and secondary patency rates at 12 months were 60.6%, 81.8%, and 100%, respectively, and at 24 months
were 51.5%, 72.7%, and 97.0%, respectively. Diabetes showed a signicant correlation with failed AVF maturation, and
venous diameters under 2.5 mm signicantly reduced the 1-year primary patency. SBAVFs were successfully created in 9/12
patients who had a borderline venous size (< 2.5 mm), resulting in the AVF creation rate improving from 66.7% to 89.7%.
Conclusion: Autogenous distal radiocephalic (snuox) arteriovenous stula is a feasible rst-line option, especially
for young, non-diabetic patients who are not in urgent need of hemodialysis. is strategy could also enhance the
possibility of autogenous AVF creation in patients with a borderline venous size.
Keywords: Arteriovenous stula; snuox arteriovenous stula; hemodialysis; vascular access for hemodialysis;
AVF (Siriraj Med J 2019; 71: 499-505)
Corresponding author: Nattawut Puangpunngam
E-mail: nattawut.pua@mahidol.edu
Received 14 August 2019 Revised 30 August 2019 Accepted 27 September 2019
ORCID ID: http://orcid.org/0000-0002-4874-5285
http://dx.doi.org/10.33192/Smj.2019.74
INTRODUCTION
According to a report on renal replacement therapy
in ailand, the number of patients requiring treatment
for renal replacement therapy has been increasing. ere
were twice as many patients in 2015 as in 2010, and
hemodialysis was implemented in the majority of cases
(65.2%).
1
As to the hemodialysis modalities, an autogenous
arteriovenous stula (AVF) has been recommended as
the rst-line approach over arteriovenous gra (AVG)
and central venous catheter-based hemodialysis (CVC)
by both the National Kidney Foundation Kidney Disease
Outcomes Quality Initiative (NKF-KDOQI) guideline
recommendations of 2006
2
and the Society for Vascular
Surgery clinical practice guidelines of 2008.
3
is was
because of AVF’s comparatively lower complication rates;
superior long-term functions; and favorable survival
Praditsuktavorn et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
500
Original Article
SMJ
outcomes
4-5
, evidenced by fewer high-ow or ischemic
complications (especially in the distal-location). It also
oers the possibility of a new, proximal, AVF creation.
Distal radiocephalic arteriovenous fistula for
hemodialysis at the anatomical snuox (SBAVF) is
considered the most distally located AVF. It is created
by an anastomosis between the dorsal branch of the
radial artery in the anatomical snuox region and the
cephalic vein at the base of the thumb. Because the two
vessels lie in parallel and are in close proximity, SBAVF
creation provides a favorable angle and allows for ease
of surgical access. Further advantages of SBAVF are its
wider vascular access area and its potential for the later
creation of a wrist radiocephalic AVF.
6–7
Despite the
advantages of an AVF in the wrist or snuox region,
many patients are not feasible candidates for SBAVF
creation. is is mainly because of the relatively small
vessel size in this area, which is challenging for a distal
AVF-rst strategy. However, several authors have reported
favorable outcomes for SBAVFs in patients with borderline
vessel size.
7-9
e aim of this study was to evaluate the outcomes
of SBAVFs, including factors aecting maturation and
patency in patients with borderline vessel size, in order
to enhance the creation of an autogenous arteriovenous
stula-rst protocol for patients requiring long-term
hemodialysis.
MATERIALS AND METHODS
e medical-record electronic database of Siriraj
Hospital was retrospectively reviewed to identify all
patients who had undergone an autogenous distal
radiocephalic (snuox) arterio-venous stula creation
for hemodialysis between July 2013 and December 2016.
Excluded were patients who had incomplete data, were
unable to attend follow-ups, or had an abandoned
arterio-venous stula within 12 months of the stula
creation. Kidney transplantations, elective changes to
dierent renal replacement therapies, or death were
considered to be censored events. Demographic data
(such as age, gender, and co-morbidities like diabetes
mellitus and coronary artery disease), factors related
to the hemodialysis management (including the serum
creatinine level), and the status of the hemodialysis were
recorded. Patient management and follow-ups were
conducted as per the Division of Vascular Surgery’s
protocols, which encompassed a preoperative assessment
with duplex ultrasonography, the operative procedures,
and the postoperative follow-ups (comprising an initial
1- to 2-week postoperative period for wound care and
subsequent 1- to 3-monthly assessments of the stula
maturation and complications). Preoperative duplex
ultrasonography was used to assess both the arterial and
venous size (using the tourniquet technique), continuity
of the vessel, and potential sites for a proximal AVF and
AVG. Patients who had a history of coronary artery disease
were evaluated by cardiologists before undergoing the
AVF procedure.
Surgical procedures: All arteriovenous stulas were
created through a 2-cm longitudinal incision under local,
regional, or general anesthesia by vascular surgeons.
Radiocephalic stulas were accomplished in a side-to-
end fashion, and 6-8 mm radial artery arteriotomies were
anastomosed with cephalic veins using polypropylene
no. 6/0 or 7/0, according to the vessel size (Fig 1).
Outcome measures: e primary outcome was
stula maturation, which was determined by vascular
surgeons based on physical examinations and evidence of
adequate venous ow in duplex ultrasonography (Fig 2).
Primary maturation was dened as an arteriovenous stula
maturing without any additional intervention. Assisted
primary maturation was dened as an arteriovenous stula
needing additional surgical or endovascular intervention
to mature. Secondary (cumulative) maturation was dened
as an arteriovenous stula maturing regardless of a prior
thrombosis or the need for a surgical or endovascular
intervention.
Fig 1. e photographs of (A) incision and cephalic vein dissection of le distal radio-cephalic arteriovenous stula and (B) completion of
le snuox radio-cephalic anastomosis.
A B
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
501
Fig 2. Postoperative assessment of anastomosis size.
e secondary outcomes were the time when the
stulas were accessed and the patency of the arteriovenous
stulas. In detail, primary patency was dened as the time
to achieve a functioning arteriovenous stula without
any additional intervention. Primary-assisted patency
was dened as the time to achieve an uninterrupted
functioning arteriovenous in the case of those stulas
needing a surgical or endovascular intervention to maintain
patency. Secondary (cumulative) patency was dened as
the time to achieve a functioning arteriovenous stula
regardless of interventions or thrombosis. is study was
approved by the Ethics Committee of the Institutional
Review Board of the Faculty of Medicine Siriraj Hospital,
Mahidol University, ailand (Si 783/2559).
Statistical analysis: All data were analyzed using
IBM SPSS Statistics for Windows, version 24 (IBM
Corp., Armonk, NY, USA). e analyses comprised
descriptive statistics, outcomes analysis by Chi-square,
logistic regression, and survival analysis by Kaplan–Meier
survival plots for patency rates.
A total of 39 distal radiocephalic arteriovenous
stula creations were included in this study. e patients’
median age was 58 years (range, 25–80 years). Twenty-
ve patients (64.1%) were male; 21 (53.8%) had diabetes
mellitus; 8 (20.5%) had received a coronary artery
intervention; while one (2.6%) had lower extremity
arterial occlusive disease (Table 1). Eighteen patients
(46.2%) were receiving catheter-based hemodialysis,
and of those without hemodialysis, the mean serum
creatinine level was 5.84 ± 1.97 mg/dL.
e preoperative vascular assessment found that
the mean radial artery diameter was 2.41 mm (standard
deviation, ± 0.46 mm) and the mean cephalic vein diameter
was 2.29 mm (standard deviation, ± 0.50 mm).
Fistula creations were performed under local, regional,
and general anesthesia on 32 (82.0%), 4 (10.3%), and
3 (7.7%) patients, respectively. Anastomoses were performed
by running polypropylene number 7/0 in 34 patients
(87%) and number 6/0 in 5 patients (13%). ere were
no reports of wound complications (infections and
hematomas), steal syndrome, or distal ischemia. One
patient developed thrombosis within the rst 24 hours,
while another two complained of reversible numbness at
the le thumbs, both of which were completely recovered
within 3 and 6 months post operation.
TABLE 1. Demographic data and chronic kidney disease parameters.
Clinical variable No. of patients (%)
Total n = 39
Age(years)
Median 58
Range 25-80
<50years 14(35.9%)
50-70years 16(41.0%)
>70years 9(23.1%)
Gender
Male 25(64.1%)
Female 14(35.9%)
Comorbidities
Diabetesmellitus 21(53.8%)
Coronaryarterydisease 8(20.5%)
Lowerextremityarterialocclusivedisease 1(2.6%)
Praditsuktavorn et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
502
Original Article
SMJ
The primary, primary-assisted, and secondary
(cumulative) maturation responses were 59%, 64.1%, and
84.6%, respectively (Table 2). All 10 patients who failed
to achieve primary maturation required transposition
to wrist radiocephalic stula creations. e maturation
times ranged from 2 to 12 months (median, 3 months).
Diabetic patients had a significant correlation with
maturation failure. Male patients and patients younger
than 50 years of age demonstrated a tendency to higher
maturation rates aer correlations with diabetes were
adjusted. On the other hand, venous size and a history
of coronary artery disease did not display signicant
correlations with maturation rates.
e follow-up periods ranged from 12 to 50 months
(median, 29 months). e patency results are presented
as Kaplan–Meier Plots in Fig 2. It was found that the
primary, primary-assisted, and secondary patency rates
at 12 months were 60.6%, 81.8%, and 100%, respectively,
and at 24 months were 51.5%, 72.7%, 97.0%, respectively.
A Cox regression analysis evaluating the clinical factors
(age group, gender, diabetes, and vein size) and the
1-year primary patency rates revealed signicantly lower
primary patency rates in patients whose cephalic venous
size was less than 2.5 mm.
DISCUSSION
A distal radiocephalic arteriovenous stula at the
anatomical snuox (SBAVF) has been recommended as
the rst-line approach over other types of AVF, including
wrist AVF, by several authors.
6–10
However, its popularity
and the number of reported cases have been low relative
to the other AVF types. is could be the result of limited
vessel-size feasibility and varied outcomes. Wolowczyk
et al. reported that 14% (30/208) of stulas failed within
6 weeks and that the 1-year patency rate was 65%7; by
comparison, other authors have reported 1-year patency
rates ranging from 61%
7–8
to over 90%.
6,11
TABLE 2. Correlation between cephalic vein size and maturation of arterio-venous stulas.
Clinical factors No. of patients Maturation (%) P-value
(%) Primary vs primary-assisted
vs secondary
Gender 0.09
Male 25(64.1%) 68.0%vs72.0%vs92.0%
Female 14(35.9%) 42.9%vs50.0%vs71.4%
Agegroup 0.05
<50years 14(35.9%) 78.6%vs78.6%vs100%
50-70years 16(41.0%) 43.8%vs56.3%vs75.0%
>70years 9(23.1%) 55.5%vs55.5%vs77.7%
Diabetesmellitus <0.05
DM 21(53.8%) 47.6%vs47.6%vs71.4%
Non-DM 18(46.2%) 72.2%vs83.3%vs100%
CAD n.s.
CAD 8(20.5%) 50%vs50%vs75%
Non-CAD 31(79.5%) 61.3%vs67.7%vs87.1%
Veinsize(mm) n.s.
<2.0 9(23.1%) 44%vs55%vs77%
2-2.4 16(41.0%) 63%vs69%vs94%
2.5-2.9 10(25.6%) 60%vs60%vs80%
>3.0 4(10.3%) 75%vs75%vs75%
Abbreviation: n.s.= no statistical signicance
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
503
Our autogenous snuox AVF-rst approach resulted
in an 84.6% success rate and a 1-year patency of 60.6%,
both of which were comparable to other studies
12–13
as
well as the outcomes of wrist radiocephalic arteriovenous
stula creations at our center.
14
Of the 6 patients (15.4%)
who had a failed AVF, two underwent an AVG, two
were given brachiocephalic arteriovenous stulas, and
two received catheter-based hemodialysis due to having
borderline cardiac functions. However, of those who had
a failed AVF, three had no suitable site for an AVF as
they had a small venous diameter (< 2.5 mm); a small
venous size was also observed in 9 patients for whom
autogenous AVFs were successfully created.
Regarding vascular size as a determining factor for
autogenous AVF creation, the results of the present study
indicated that the SBAVF-rst approach enhanced the
autogenous AVF creation from 66.7% (26/39) to 89.7%
(35/39).
e major disadvantages of this approach were its
relatively long maturation time (mean, 3 months) and
a high incidence of juxta-anastomosis stenosis at the
wrist level, which frequently required transposition to a
wrist radiocephalic arteriovenous stula as a secondary
procedure. Nonetheless, that procedure was able to be
performed simply under local anesthesia due to the
vessels being enlarged aer the rst stula creation
(Fig 3).
Diabetes had a signicant association with failed
SBAVF maturations, while being female and a patient age
of over 50 years tended towards lower maturation rates.
Although venous size was not signicantly correlated with
the maturation rate, a venous diameter of 3 mm or more
did not require any intervention to assist maturation.
A venous size of less than 2.5 mm signicantly correlated
with a reduction in the 1-year patency rate. However,
our sample size was relatively small and contained only
a few patients who had a large venous diameter.
ere are several measures to determine the factors
aecting snuox AVF outcomes. In 2012, Twine et al.
developed a DISTAL score by combining 6 factors (each
factor representing 1 score) aecting stula failure, namely,
diabetes, ischemic heart disease, stroke, two snuox
procedures, an age > 70 years, and a venous diameter of
< 2 mm.
8
is DISTAL scoring system was used in our
study (Table 4), and it revealed a signicant correlation
with the maturation rates but not with the patency rates.
is can be explained by the DISTAL scoring system’s
use of a dierent cut-point for signicant venous size
(< 2 mm versus < 2.5 mm in the current study) and
a patient age of > 70, which tended to lead to higher
patency rates in our study (Fig 4).
Fig 3. Transposition of le radio-cephalic arteriovenous stula (Le)
and Transposed le radio-cephalic arteriovenous stula (Right).
Fig 4. Outcomes of autogenous snuox radiocephalic AVF (Le)
Early postoperative period, (Middle) in a 47 year-old male patient,
and (Right) a 79 year-old female patient.
CONCLUSION
e study results suggest that an autogenous snuox
radiocephalic arteriovenous stula (SBAVF)-rst strategy
is feasible because of its relative ease of surgical access,
its comparatively extended venous access area, and the
provision of options for future AVF correction, especially
in the case of young patients without diabetes and not
in urgent need of hemodialysis. SBAVF could also be
considered in cases of borderline venous size in order
to enhance the rate of autogenous arteriovenous stula
for hemodialysis.
Praditsuktavorn et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
504
Original Article
SMJ
TABLE 3. Correlation between clinical factors and 1-year primary patency rates.
Clinical factors No. of patients Patency rates (%) P-value
(n = 33)*
Gender n.s.
Male 23(69.7%) 65.2%
Female 10(30.3%) 50.0%
Agegroup 0.056
<50years 14(42.4%) 57.1%
50-70years 12(36.4%) 50.0%
>70years 7(21.2%) 85.7%
Diabetesmellitus n.s.
DM 15(45.4%) 73.3%
Non-DM 18(54.5%) 50.0%
CAD n.s.
CAD 6(18.2%) 83.3%
Non-CAD 27(81.8%) 55.6%
Venoussize(mm) <0.05
<2mm 7(21.2%) 28.6%
 2-2.4mm 15(45.4%) 53.3%
 2.5-2.9mm 8(24.3%) 87.5%
 >3mm 3(9.1%) 100%
*Patients in whom the AVF failed to mature were excluded.
Abbreviation: n.s.= no statistical signicance
e follow-up period range was 12–50
months (median, 29 months). e patency
results are demonstrated in Kaplan–Meier
plots in Fig 2; the primary, primary-assisted,
and secondary patency rates at 12 months
were 60.6%, 81.8%, and 100%, respectively;
and at 24 months were 51.5%, 72.7%, and
97.0%, respectively. A Cox regression analysis
evaluating the clinical factors (age group,
gender, diabetes, and vein size) and 1-year
primary patency rates demonstrated signicantly
lower primary patency rates in patients whose
cephalic venous size was < 2.5 mm.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
505
REFERENCES
1. Nephrology Society of ailand. Nephrothaiorg. [Online].
Available from: http://www.nephrothai.org/images/Final_
TRT_report_2015_edited.pdf [Accessed 20 Nov 2018].
2. NKF-K/DOQI Clinical Practice Guidelines for Vascular Access:
Update 2006. Guideline 2: Selection and placement of hemodialysis
access, 2.1 e order of preference for placement of stulae,
2.1.1 Preferred: stulae.
3. Sidawy AN,Spergel LM, Besarab A, Allon M,Jennings
WC,Padberg FT Jr,et al. e Society for Vascular Surgery:
clinical practice guidelines for the surgical placement and
maintenance of arteriovenous hemodialysis access. J Vasc Surg
2008; 48:2S-25S.
4. Xue JL, Dahl D, Ebben JP, Collins AJ. e association of initial
hemodialysis access type with mortality outcomes in elderly
Medicare ESRD patients. Am J Kidney Dis 2003;42:1013-9.
5. Ravani P,Palmer SC,Oliver MJ,Quinn RR,MacRae JM,Tai
DJ,et al. Associations between hemodialysis access type and
clinical outcomes: a systematic review. J Am Soc Nephrol
2013;24:465-73.
6. Bonalumi U, Civalleri D, Rovida S, Adami GF, Gianetta E,
Grianti-Bartoli F. Nine years’ experience with end-to-end
arteriovenous stula at the ‘anatomical snuox’ for maintenance
haemodialysis. Br J Surg 1982;69:486-8.
7. Wolowczyk L, Williams AJ, Donovan KL, Gibbons CP. e
snuox arteriovenous stula for vascular access. Eur J Vasc
Endovasc Surg 2000;19:70-6.
8. Twine CP, Haidermota M, Woolgar JD, Gibbons CP, Davies
CG. A scoring system (DISTAL) for predicting failure of snuox
arteriovenous stulas. Eur J Vasc Endovasc Surg 2012;44:
88-91.
9. Letachowicz K, Gołębiowski T, Kusztal M, Letachowicz W,
Weyde W, Klinger M. e snuox stula should be preferred
over the wrist arteriovenous stula. J Vasc Surg 2016;63:436-
40.
10. Mehigan LT, McAlexander RA. Snuox arteriovenous stula
for haemodialysis. Am J Surg 1982;143:252-3.
11. Horimi H, Kusano E, Hasegawa T, Fuse K, Asano Y. Clinical
experience with an anatomic snu box arteriovenous stula
in hemodialysis patients. ASAIO J 1996;42:177-80.
12. Al-Jaishi AA, Oliver MJ, omas SM, Lok CE, Zhang JC, Garg
AX, et al. Patency rates of the arteriovenous stula for hemodialysis:
a systematic review and meta-analysis. Am J Kidney Dis
2014;63:464-78.
13. Almasri J, Alsawas M, Mainou M, Mustafa RA, Wang Z, Woo
K, et al. Outcomes of vascular access for hemodialysis: A
systematic review and meta-analysis. J Vasc Surg 2016;64:236-
43.
14. Ruangsetakit C, Chuemor P, Hongku K, Hahtapornsawan S,
Chinsakchai K, Sermsathanasawadi N, et al. Outcomes of
Autogenous Hemodialysis Access at Siriraj Hospital. J Med
Assoc ai 2017;100(Suppl 3):S193-8.
TABLE 4. Correlations between distal scores, maturation rates, and 1-year patency rates.
Distal score No. of patients Maturation rate* No of AVFs 1-year primary
(n=39) (n=33) patency rate
n.s.
0 10(25.6%) 100% 10 50%
1 17(43.6%) 88.2% 15 50.0%
2 4(10.3%) 75% 3 66.7%
3 8(20.5%) 62.5% 5 80.0%
*p < 0.05,
Abbreviation: n.s.= no statistical signicance
Praditsuktavorn et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
506
Original Article
SMJ
Nutthawut Akaranuchat, M.D., Jongdee Aojanepong, M.D.
Department of Surgery, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok 10700, ailand.
Formulaic Prediction of Z-Plasty Outcomes Relative
to Z Scar Lengthening, Z Flap Tension, and Area
of Distortion, and Determination of the Multiple
Z-Plasty Congurations that Optimally Increase Z
Scar Length, Decrease Z Flap Tension, and
Decrease Area of Distortion
ABSTRACT
Objective: To generate mathematical formulas to predict Z-plasty outcomes relative to Z scar lengthening, Z ap
tension, and area of distortion, and to determine the multiple Z-plasty congurations that optimally increase Z scar
length, decrease Z ap tension, and decrease area of distortion.
Methods: Each Z-plasty conguration was evaluated for highest percent Z scar lengthening, lowest tension to
close Z ap, and lowest area of distortion. For part 1 of the study, conventional one-ap, two-ap, four-ap, and
eight-ap Z-plasties were created, aer which relocation of the Z aps was performed. e outcomes were analyzed
and formulas were generated to predict Z-plasty outcomes. In part 2, the following modications to the four-ap
Z-plasty were made: modication of Z ap angle, modication of Z ap limb size, and addition of a gap between
each Z ap. Outcomes were evaluated to identify the conguration that produced the optimal outcome for each
outcome measure.
Results: irty-six pig skins were incised and sutured under controlled condition. For part 1, one-ap Z-plasty was
best for Z scar lengthening, and 8-ap Z-plasty was best for reducing both tension to close and area of distortion.
ree formulas were generated to predict the studied outcomes of Z-plasty. For part 2, the conguration that added
a gap of 75% of the central limb length yielded the best outcome for scar length and also work very well for reducing
the tension to close. e conventional design was the best for minimizing area of distortion.
Conclusion: Mathematical formulas were generated to predict Z scar percent lengthening, Z ap tension to close, and
area of distortion. Addition of a gap 75% of the central limb length was the most ecacious Z-plasty conguration
for increasing Z scar length.
Keywords: Mathematical prediction; Z-plasty outcomes; Z scar lengthening; Z ap tension; area of distortion;
multiple Z-plasty congurations (Siriraj Med J 2019; 71: 506-514)
Corresponding author: Nutthawut Akaranuchat
E-mail: nutthawut.joe@gmail.com
Received 27 May 2019 Revised 22 September 2019 Accepted 25 September 2019
ORCID ID: http://orcid.org/0000-0003-1798-8484
http://dx.doi.org/10.33192/Smj.2019.75
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
507
INTRODUCTION
Z-plasty is a valuable technique for plastic surgeons.
Z-plasty is an eective surgical treatment to manage scar
contracture
2,3,5,6
, and it is incorporated in many plastic
surgery procedures. e four fundamental functions of
Z-plasty are: 1) To lengthen a scar or release a contracture;
2) To realign a scar within relaxed skin tension lines
(RSTLs); 3) To move tissue from one area to another; and,
4) To obliterate or create a web or cle.
7
Modications of
Z-plasty were developed to serve dierent purposes
8
, and
one of the most popular variations is multiple Z-plasty
or compound Z-plasty technique. Multiple Z-plasty is
dened as the linkage of two or more Z aps. Multiple
Z-plasty is an eective alternative for the management
of large scars that would not respond well to the use
of a single (one-ap) Z-plasty, because the large aps
of a single Z ap would eectuate more distortion and
tension to the adjacent area than multiple smaller Z
aps. Many studies were conducted to understand and
predict the eect of Z-plasty on adjacent tissue in both
animal models1 and computer simulations.
4,9,10
However,
methods to predict the surgical outcomes of Z-plasty,
and data relating to the Z-plasty congurations that
yield the best surgical outcomes are scarce. Accordingly,
the aims of this study were to generate mathematical
formulas to predict Z-plasty outcomes relative to Z scar
lengthening, Z ap tension, and area of distortion, and
to determine the multiple Z-plasty congurations that
optimally increase Z scar length, decrease Z ap tension,
and decrease area of distortion.
MATERIALS AND METHODS
Fresh pig skins, including the epidermis, dermis, and
subcutaneous layer were prepared. All specimens were
cut to 16.0 x 16.0 cm in size, and to between 1.3 to 1.5
cm in thickness. Specimen temperature was maintained
within the range of 25-30 degrees Celsius using an infrared
thermometer. Skin surface tension was set within 0.07-
0.08 newton/m (N/m)
12
by using digital manometer
before commencement of the experiment.
Specimens were marked with colored dots located
0.5 cm apart, aer which Z-plasty and multiple Z-plasty
aps were drawn then marked the reference point of
both end of Z flap as point A and point B (After Z
ap was relocated these points will be point A’ and
point B’ respectively). Aer that incisions were made.
Part 1
e conventional design of 1-ap, 2-ap, 4-ap,
and 8-ap Z-plasties was performed, which consisted of
60 degree angle size and equality among all Z ap limb
lengths (Fig 1A-D).
Fig 1. (A) One-ap Z-plasty, (B) Two-ap Z-plasty, (C) Four-ap
Z-plasty, (D) Eight-ap Z-plasty.
Part 2
Modications to the conventional 4-ap Z-plasty
were designed. First, we modied the angle of each Z
ap to 75 and 45 degrees (Fig 2A), and to 90 and 30
degrees (Fig 2B). Second, we modied the size of Z ap
limb length to 2.5 cm and 1.5 cm (Fig 2C), and also to
alternating unequal Z ap peripheral limb size of 2.5
cm and 1.5 cm, and central limb size 2.0 cm (Fig 2D).
ird, we added a gap between each Z ap, as follows:
1. Gap 0.5 cm (25% of central limb length; Fig 2E)
2. Gap 1.0 cm (50% of central limb length; Fig 2F)
3. Gap 1.5 cm (75% of central limb length; Fig 2G)
4. Gap 2.0 cm (100% of central limb length; Fig 2H)
Z aps were then relocated to their nal position
and the following outcomes were measured: 1) actual
scar length (centimeter, cm); tension to close (newton,
N); and, area of distortion (square centimeter, cm2).
A descriptive analysis was performed, and the results
are presented as mean, and as frequency and percentage
for categorical variables. SPSS Statistics version 18 (SPSS,
Inc., Chicago, IL, USA) was used to analyze the data.
Akaranuchat et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
508
Original Article
SMJ
RESULTS
Part 1
is experimental study was performed on 36 pig
skin specimens (12 designs, 3 per each design), and post-
procedure measurements were taken and recorded.
For one-ap Z plasty, it increased the length about
46.875%, need tension to close the Z ap about 5.283 N,
and create area of distortion about 208.0 cm
2
. For two-
ap Z plasty, it increased the length about 31.25%, need
tension to close the Z ap about 5.913 N, and create area
of distortion about 122.375 cm
2
. For four-ap Z plasty, it
increased the length about 18.75%, need tension to close
the Z ap about 6.433 N, and create area of distortion
about 84.5 cm
2
. For eight-ap Z plasty, it increased the
length about 17.5%, need tension to close the Z ap
about 3.255 N, and create area of distortion about 37.5
cm
2
.
e results of each design were compared (Table 1),
and we found that 1-ap Z-plasty yielded the best result
in terms of scar lengthening, followed by 2-ap, 4-ap,
and 8-ap Z-plasty.
Regarding the tension needed to close the Z ap
and the area of distortion, the authors found that 8-ap
Z-plasty produced both the least tension and the least
area of distortion aer Z ap relocation, followed by
4-ap, 2-ap, and 1-ap Z-plasty.
From the data we found, the curve estimation models
were match with inverse equation; Y = b0 + (b1 / t),
(Y = result from equation, b0 = constant, b1 = constant,
t = number of Z plasty) (Fig 3A-C).
1
Finally, the mathematical models that were generated
to predict percent scar lengthening, tension to close Z
ap, and area of distortion outcomes aer Z-plasty are,
as follows:
Percent lengthening = 12.10 + (35.74 / t)
Tension to close Z ap = 0.05 + (5.36 / t)
Area of distortion = 26.54 + (184.65 / t)
Part 2
In this part, the authors set forth to evaluate the
eect of each modied Z-plasty technique. e studied
Fig 2. (A) Modication of the angles of each Z ap; 75 and 45 degrees,
(B) Modication of the angles of each Z ap; 90 and 30 degrees, (C)
Modication of the limb size of each Z ap; Z ap size 2.5 cm and
1.5 cm, (D) Unequal peripheral limb of Z ap size 2.5 cm, and 1.5
cm and central limb size 2.0 cm, (E) Addition of a gap between each
Z ap; Gap of 0.5 cm (25% of limb size), (F) Addition of a gap between
each Z ap; Gap of 1.0 cm (50% of limb size), (G) Addition of a gap
between each Z ap; Gap of 1.5 cm (75% of limb size), (H) Addition
of a gap between each Z ap; Gap of 2.0 cm (100% of limb size).
Fig 3. (A) Model summary of percent scar lengthening; percent scar lengthening: b0 (constant) = 12.10, b1 = 35.74, (B) Model summary
of tension to close Z ap; tension to close Z ap: b0 (constant) = 0.05, b1 = 5.36, (C) Model summary of area of distortion; area of distortion:
b0 (constant) = 26.54, b1 = 184.65.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
509
Type Original length Actual length Lengthening Percent Total tension Area of
(cm) (cm) (cm) lengthening to close (N) distortion
(%) (cm
2
)
1-ap Z-plasty 8.0 11.8 3.8 46.875 5.283 208.0
2-ap Z-plasty 8.0 10.5 2.5 31.25 5.913 122.375
(2.957each)
4-ap Z-plasty 8.0 9.5 1.5 18.75 6.433 84.5
(1.608each)
8-ap Z-plasty 8.0 9.4 1.4 17.5 3.255
(0.407each) 37.5
TABLE 1. Measurement parameters compared among the 4 conventional Z-plasty designs.
modications included modifying the angle of each Z
ap (Fig 2A and 2B), modifying the limb size of Z ap
(Fig 2C and 2D), and adding a gap between each Z ap
(Fig 2E, 2F, 2G, and 2H). e results were then compared,
including with the conventional 4-ap Z-plasty design
(Fig 1C). e outcomes of each modication are shown
in Tables 2 to 9 (Supplement data), with comparative
data shown in Table 10.
In Table 2 and Table 3, modifying the angle of
each Z ap yield the results that less eective than the
conventional design in all 3 circumstances (Percent
lengthening, Tension to close Z ap, and Area of distortion).
In Table 4, modifying the limb size of Z flap from
2.5 cm to 1.5 cm yield the results that less eective than
the conventional design in all 3 circumstances. For
the modifying the limb size of Z ap to 2.0 cm for the
central limb size, and 2.5 cm and 1.5 cm for the sizes of
the peripheral limbs (Table 5) shown comparable result
to the conventional one in term of percent lengthening,
better in tension to close Z ap, but signicantly worse in
area of distortion. In Table 6 to Table 9, every design of
adding a gap between each Z ap can signicantly reduce
the tension to close Z ap better than the conventional
design. Especially for the design that adding a gap of
0.5 cm (25% of central limb length) that yield the best
result. And for percent lengthening, adding a gap between
each Z ap of 1.5 cm (75% of central limb length) shown
very interest outcome, because it yielded the better result
than the conventional design (21.60% vs 18.75%).
e results revealed that the addition of a gap of
1.5 cm (or 75% of central limb length) was the best design
for lengthening the Z scar, followed by the conventional
design and then the design that alternates the limb size of
the Z ap from 2.5 cm to 1.5 cm. Regarding the lowering
of tension, addition of a gap of 0.5 cm (or 25% of central
limb length) yielded the best result, followed by addition
of a gap of 1.5 cm (or 75% of central limb length). For
area of distortion, the conventional technique yielded
the best outcomes (Table 10).
DISCUSSION
In part 1, the authors generated the following
mathematical formulas to predict surgical outcomes
aer Z-plasty:
Percent scar lengthening = 12.10 + (35.74 / t)
Tension to close Z ap = 0.05 + (5.36 / t)
Area of distortion = 26.54 + (184.65 / t)
Compare to the previous studies
1, 11
, our new version
of mathematical models are easier to use, to remember
and can estimate the nal outcomes of multiple Z plasty
not only the percentage of scar lengthening but also the
ap tension and area of distortion.
In part 2 of this study, the authors found that addition
of a gap of 1.5 cm (or 75% of the central limb length) was
the best design to optimally lengthen the Z scar (2.85%
better than the conventional design (21.6% vs 18.75%)).
is new knowledge will signicantly help us to combat
with the contracted scar. e design that added a gap of
0.5 cm (or 25% of the central limb length) yielded the
best outcome relative to lowering the tension needed
to close the Z ap. e conventional Z-plasty design
was found to deliver the best outcome relative to area
of distortion.
Akaranuchat et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
510
Original Article
SMJ
TABLE 2. Outcomes of modifying the angle of each Z ap from 75 degrees to 45 degrees.
Parameter Actual length Lengthening Percent Result
(cm) (cm) lengthening (%)
Actual length 8.0
A’B’(AB=8.0) 9.2 1.2 15.00
A’G’(AG=2.0) 2.2 0.2 10.00
 G’J’(GJ=2.0) 2.2 0.2 10.00
 J’M’(JM=2.0) 2.3 0.3 15.00
 M’B’(MB=2.0) 2.5 0.5 25.00
Actual length
C’D”(CD=3.80) 1.9 -1.9 -50.00
E’F’(EF=2.83) 2.2 -0.63 -22.26
H’I’(HI=3.80) 2.1 -1.7 -44.74
K’L’(KL=2.83) 2.1 -0.73 -25.80
Tension to close (N) 10.086
FirstZ-plasty 2.850
SecondZ-plasty 1.578
ThirdZ-plasty 3.499
FourthZ-plasty 2.159
Area of distortion (cm
2
) 96.0
TABLE 3. Outcomes of modifying the angle of each Z ap from 90 degrees to 30 degrees.
Parameter Actual length Lengthening Percent Result
(cm) (cm) lengthening (%)
Actual length 8.0
A’B’(AB=8.0) 9.2 1.2 15.00
A’G’(AG=2.0) 2.4 0.4 20.00
G’J’(GJ=2.0) 2.1 0.1 5.00
J’M’(JM=2.0) 2.5 0.5 25.00
M’B’(MB=2.0) 2.2 0.2 10.00
Actual length
C’D”(CD=4.0) 2.3 -1.7 -42.50
E’F’(EF=2.0) 1.8 -0.2 -10.00
H’I’(HI=4.0) 2.4 -1.6 -40.00
K’L’(KL=2.0) 1.7 -0.3 -15.00
Tension to close (N) 15.726
FirstZ-plasty 5.643
SecondZ-plasty 2.960
ThirdZ-plasty 5.288
FourthZ-plasty 1.835
Area of distortion (cm
2
) 99.375
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
511
TABLE 4. Outcomes of modifying the limb size of the Z ap from 2.5 cm to 1.5 cm.
TABLE 5. Outcomes of modifying the limb size of Z ap to 2.0 cm for the central limb size, and 2.5 cm and 1.5
cm for the sizes of the peripheral limbs.
Parameter Actual length Lengthening Percent Result
(cm) (cm) lengthening (%)
Actual length 8.0
A’B’(AB=8.0) 9.4 1.4 17.50
A’G’(AG=2.5) 3.0 0.5 20.00
G’J’(GJ=1.5) 1.7 0.2 13.33
J’M’(JM=2.5) 2.9 0.4 16.00
M’B’(MB=1.5) 1.8 0.3 20.00
Actual length
C’D”(CD=4.4) 2.6 -1.8 -40.91
E’F’(EF=2.6) 1.8 -0.8 -30.77
H’I’(HI=4.4) 2.6 -1.8 -40.91
K’L’(KL=2.6) 1.8 -0.8 -30.77
Tension to close (N) 8.425
FirstZ-plasty 1.481
SecondZ-plasty 1.877
ThirdZ-plasty 3.022
FourthZ-plasty 2.045
Area of distortion (cm
2
) 95.25
Parameter Actual length Lengthening Percent Result
(cm) (cm) lengthening (%)
Actual length 8.0
A’B’(AB=8.0) 9.5 1.5 18.75
A’G’(AG=2.0) 2.3 0.3 15.00
G’J’(GJ=2.0) 2.3 0.3 15.00
J’M’(JM=2.0) 2.5 0.5 25.00
M’B’(MB=2.0) 2.4 0.4 20.00
Actual length
C’D”(CD=3.5) 2.4 -1.1 -31.43
E’F’(EF=3.5) 2.5 -1.0 -28.57
H’I’(HI=3.5) 2.6 -0.9 -25.71
K’L’(KL=3.5) 2.5 -1.0 -28.57
Tension to close (N) 5.620
FirstZ-plasty 2.212
SecondZ-plasty 0.703
ThirdZ-plasty 1.937
FourthZ-plasty 0.768
Area of distortion (cm
2
) 120.25
Akaranuchat et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
512
Original Article
SMJ
TABLE 6. Outcomes of adding a gap between each Z ap of 0.5 cm (25% of central limb length).
TABLE 7. Outcomes of adding a gap between each Z ap of 1.0 cm (50% of central limb length).
Parameter Actual length Lengthening Percent Result
(cm) (cm) lengthening (%)
Actual length 9.5
A’B’(AB=9.5) 10.9 1.4 14.74
A’G’(AG=2.25) 2.6 0.35 15.56
G’J’(GJ=2.5) 2.8 0.3 12.00
J’M’(JM=2.5) 2.8 0.3 12.00
M’B’(MB=2.25) 2.7 0.45 20.00
Actual length
C’D”(CD=3.46) 2.3 -1.16 -33.53
E’F’(EF=3.46) 2.4 -1.06 -30.64
H’I’(HI=3.46) 2.4 -1.06 -30.64
K’L’(KL=3.46) 2.4 -1.06 -30.64
Tension to close (N) 1.154
FirstZ-plasty 0.214
SecondZ-plasty 0.204
ThirdZ-plasty 0.191
FourthZ-plasty 0.545
Parameter Actual length Lengthening Percent Result
(cm) (cm) lengthening (%)
Actual length 11
A’B’(AB=11.0) 12.5 1.5 13.64
A’G’(AG=2.5) 3.1 0.6 24.00
G’J’(GJ=3.0) 3.2 0.2 6.67
J’M’(JM=3.0) 3.2 0.2 6.67
M’B’(MB=2.5) 3.0 0.5 20.00
Actual length
C’D”(CD=3.46) 2.2 -1.26 -36.42
E’F’(EF=3.46) 2.3 -1.16 -33.53
H’I’(HI=3.46) 2.3 -1.16 -33.53
K’L’(KL=3.46) 2.2 -1.26 -36.42
Tension to close (N) 3.171
FirstZ-plasty 0.372
SecondZ-plasty 0.639
ThirdZ-plasty 1.318
FourthZ-plasty 0.842
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
513
TABLE 8. Outcomes of adding a gap between each Z ap of 1.5 cm (75% of central limb length).
TABLE 9. Outcomes of adding a gap between each Z ap of 2.0 cm (100% of central limb length).
Parameter Actual length Lengthening Percent Result
(cm) (cm) lengthening (%)
Actual length 12.5
A’B’(AB=12.5) 15.2 2.7 21.60
A’G’(AG=2.75) 3.6 0.85 30.91
G’J’(GJ=3.5) 4 0.5 14.29
J’M’(JM=3.5) 4 0.5 14.29
M’B’(MB=2.75) 3.6 0.85 30.91
Actual length
C’D”(CD=3.46) 2.2 -1.26 -36.42
E’F’(EF=3.46) 2.2 -1.26 -36.42
H’I’(HI=3.46) 2.3 -1.16 -33.53
K’L’(KL=3.46) 2.2 -1.26 -36.42
Tension to close (N) 1.867
FirstZ-plasty 0.665
SecondZ-plasty 0.47
ThirdZ-plasty 0.381
FourthZ-plasty 0.351
Parameter Actual length Lengthening Percent Result
(cm) (cm) lengthening (%)
Actual length 14.0
A’B’(AB=14.0) 15.3 1.3 9.29
A’G’(AG=3.0) 3.6 0.6 20.00
G’J’(GJ=4.0) 4.1 0.1 2.50
J’M’(JM=4.0) 4.1 0.1 2.50
M’B’(MB=3.0) 3.5 0.5 16.67
Actual length
C’D”(CD=3.46) 2.2 -1.26 -36.42
E’F’(EF=3.46) 2.2 -1.26 -36.42
H’I’(HI=3.46) 2.2 -1.26 -36.42
K’L’(KL=3.46) 2.2 -1.26 -36.42
Tension to close (N) total 2.954
FirstZ-plasty 0.713
SecondZ-plasty 0.680
ThirdZ-plasty 0.887
FourthZ-plasty 0.674
Akaranuchat et al.
Volume 71, No.6: 2019 Siriraj Medical Journal
www.smj.si.mahidol.ac.th
514
Original Article
SMJ
Limitations
Even though this experimental study was performed
using a pig skin model, the authors propose that these
ndings may be extrapolatable to single and multiple
Z-plasty procedures performed in human patients.
CONCLUSION
Using a pig skin model, mathematical formulas
were generated to predict Z ap percent lengthening,
Z ap tension to close, and area of distortion. Addition
of a gap 75% of the central limb length was the most
ecacious Z-plasty conguration for increasing Z ap
length and also better than many designs to reduce Z
ap tension to close.
ACKNOWLEDGMENTS
e authors gratefully acknowledge Miss Julaporn
Pooliamof the Division of Clinical Epidemiology,
Department of Research and Development, Faculty
of Medicine Siriraj Hospital, Mahidol University for
assistance with statistical analysis.
Conict of interest declaration: Both authors declare
no personal or professional conicts of interest, and no
nancial support from the companies that produce and/
or distribute the drugs, devices, or materials described
in this report.
TABLE 10. Measured parameters compared among the studied modied Z-plasty designs.
Type Original length Actual length Lengthening Percent Total tension Area of
(cm) (cm) (cm) lengthening to close (N) distortion
(%) (cm
2
)
Conventional design 8.0 9.5 1.5 18.75 6.433 84.50
Angle 75 : 45 8.0 9.2 1.2 15.00 10.086 96.00
Angle 90 : 30 8.0 9.2 1.2 15.00 15.726 99.375
Size 2.5 switching 1.5 8.0 9.4 1.4 17.50 8.425 95.25
Limb size 2.5 &1.5 8.0 9.5 1.5 18.75 5.620 120.25
Add gap of 0.5 cm 9.5 10.9 1.4 14.74 1.154
Add gap of 1.0 cm 11.0 12.5 1.5 13.64 3.171
Add gap of 1.5 cm 12.5 15.2 2.7 21.60 1.867
Add gap of 2.0 cm 14.0 15.3 1.3 9.29 2.954
Funding disclosure: is was an unfunded study.
REFERENCES
1. Furnas DW, Fischer GW. e Z-plasty: biomechanics and
mathematics. Br J Plast Surg 1971;24:144-60.
2. Borges AF, Gibson T. e original Z-plasty. Br J Plast Surg
1973;26:237-46.
3. Gibson T, Kenedi RM. Biomechanical properties of skin. Surg
Clin North Am 1967;47:279-94.
4. Kawabata H, Kawai H, Masada K, Ono K. Computer- aided
analysis of Z-plasties. Plast Reconstr Surg 1989;83:319-25.
5. Rohrich, RJ, Zbar, RI. A simplied algorithm for the use of
Z-plasty. Plast Reconstr Surg 1999;103:1513.
6. Donald AH. Some oughts on Choosing a Z-Plasty: e Z
made simple. Plast Reconstr Surg 2000;106:665-70.
7. Christopher RH, Edwin FW, Bret JR. Z-Plasty: A Concise
Review. Facial Plast Surg 2001; 17:289-93.
8. Dávila MA. Elongation through Z-plasty: an experimental
study comparing 11 Z-plasty variations on nonbiological tissue.
Plast Reconstr Surg 2012;129:890e-92e.
9. Erika K, Masataka A. Biomechanics and Computer Simulation
of the Z-Plasty. J Nippon Med Sch 2013;80:218-23.
10. Sunderraj E, Norman LG. A mathematical model to predict
the change in direction of the common limb in Z plasty. Indian
J Plast Surg 2009;42:82-84.
11. Ellur S. A mathematical model to predict the requirement for
multiple Z plasties. Indian J Plast Surg2008;41:99-101.
12. Randall WC, Calman C. e surface tension of human sweat; its
determination and its signicance. J Invest Dermatol1954;23:
113-8.