An Agreement of Two Tonometers: Goldmann Applanation and Non-Contact Scheimflug Technology in Healthy, Ocular Hypertension and Open-angle Glaucoma Patients

Abstract

Objective: The primary objective was to find an agreement of intraocular pressure (IOP) assessed by Goldmann applanation tonometry (GAT) and Corvis in healthy, ocular hypertension (OHT) and primary open-angle glaucoma (POAG). The secondary objective was to find a reliability of intra-examiner and inter-examiner IOP measurement by GAT and Corvis.
Methods: Fifty three eyes from 53 participants were included and were divided into healthy (N=20), OHT (N=13) and POAG group (N=20). Only right eyes were selected for further statistical analysis except one patient with only left eye eligible. The eyes with corneal pathologies, greater than 2.5 diopters astigmatism, or recent ocular surgery were excluded. Randomized examining sequence between GAT and Corvis was applied. To minimize an after measurement IOP fluctuation, five minutes and two minutes gap between measurements were strictly applied for Corvis and GAT respectively. The first ten patients had 3 measurements per measurer and two measurers were assigned per machine to evaluate intra-examiner and inter-examiner reliability. Intraclass correlation coefficient was used to analyze the reliability of the IOP measuring machine. Bland & Altman plot was used to analyze an agreement between the machines.
Results: High ICCs were found in both measurers using GAT (ICC of measurer 1 = 0.954, measurer 2 = 0.977) and Corvis (ICC of measurer 1 = 0.920, measurer 2 = 0.927) which indicated excellent intra-examiner reliability. High ICCs were found when comparing IOP between 2 measurers who used the same machine (GAT ICC = 0.928, Corvis ICC = 0.915) which indicated excellent inter-examiner reliability. GAT tends to yield higher IOP reading. The mean IOP were 13.93±3.849 by GAT and 12.15±4.030 by Corvis. The mean IOP differences were 1.8, 1.7, 1.4 and 2.2 mmHg in total, healthy, OHT and POAG group respectively. POAG had highest mean difference and widest standard deviation which might result from poor agreement between 2 machines. According to Bland & Altman plot the values were scattered and no trend was found indicating higher or lower average IOP would result in higher or lower difference between the two machines. From the clinical point of view, 71.7% and 47.2% fall into IOP difference range of ±3 and ±2 mmHg respectively.
Conclusion: Corvis-IOP is a good parameter with excellent intra-examiner and inter-examiner reliability. In clinical practice, the usefulness of Corvis-IOP is limited especially in POAG patients according to the poor agreement with gold standard GAT-IOP.