Low Sensitivity of the Van Herick Method for Detecting Gonioscopic Angle Closure Independent of Observer Expertise.

Abstract

To evaluate the diagnostic performance characteristics of the Van Herick assessment (VHA) for identifying angle closure compared to gold-standard gonioscopy, as conducted by trained observers of varying expertise. Reliability analysis. Patients (n = 131) from a glaucoma referral clinic aged ≥50 years without prior ocular surgery or iridotomy underwent unilateral VHA by 1 of 11 trained ophthalmic technicians, followed by VHA and indirect gonioscopy by 1 of 15 ophthalmology residents and 1 of 4 glaucoma specialist attending physicians. Observers were masked to others' gradings. Cohen's kappa (κ) assessed test reproducibility. VHA sensitivity and specificity for identifying gonioscopic angle closure were calculated. Mean patient age was 62.0 ± 8.7 years, 56% of patients were male, and 84% were African American. Angles were gonioscopically closed in 14.5% of eyes. Moderate agreement was observed comparing technician or resident VHA to attending VHA (κ = 0.48 and κ = 0.56, respectively). Resident and attending gonioscopy demonstrated excellent agreement (κ = 0.94). Sensitivities of technician, resident, and attending VHA for detecting angle closure were 57.9% (95% confidence interval: 34.0%-78.9%), 78.9% (53.9%-93.0%), and 68.4% (43.5%-86.4%), respectively. Specificities were 88.5% (80.3%-93.6%), 88.2% (80.3%-93.3%), and 87.5% (79.6%-92.8%), respectively. VHA, even when performed by experienced ophthalmologists, misses a substantial proportion of angle closure while incorrectly identifying roughly 1 in 8 open-angle eyes as closed. These results suggest that clinical assessment of anterior chamber angle configuration is best accomplished with gonioscopy.