LAUSR

OBJECTIVE To determine which patient reported symptoms best distinguishes patients with and without glaucoma and explains the most variance in visual field (VF) damage, and compare the amount of variance that can be explained by symptoms vs. retinal nerve fiber layer (RNFL) thickness. DESIGN Cross-sectional study. PARTICIPANTS Adults diagnosed with glaucoma or suspicion of glaucoma (controls). METHODS Worse-eye VF damage was defined based on perimetric testing. RNFL thickness was defined by OCT imaging. Patients rated their visual symptoms on questions collated from several published questionnaires, rating the frequency and severity of 28 symptoms on a scale of 1 (never/not at all) to 4 (very often/severe). Multivariable regression models identified patient reported symptoms that contributed the highest variance in VF damage. MAIN OUTCOME MEASURES Patient reported symptoms that explained the most variance in VF damage; amount of variance in VF damage explained by patient reported symptoms and RNFL RESULTS: A total of 170 patients (mean age= 64; 58% female; 47% employed) completed testing, including 95 glaucoma suspects and 75 glaucoma patients. In glaucoma patients, median mean deviation of VF damage in the worse eye was -19.3 and ranged from -5.3 to -34.7 dB. Symptoms more common amongst glaucoma patients compared to glaucoma suspects included better vision in one eye, blurry vision, glare, sensitivity to light, cloudy vision, and little peripheral vision. Worse severity ratings for the symptom 'little peripheral vision' explained the most variance in VF damage (43%). A multivariable model including the frequency of cloudy vision, severity of having little peripheral vision, missing patches, one eye having better vision, and vision worsening, plus sociodemographic features explained 62% of the variance in VF damage. Comparatively, a multivariable model of worse-eye RNFL thickness and sociodemographic features explained 42% of the variance in VF damage, while a model including only sociodemographic features explained 8% of the variance in VF damage. CONCLUSIONS Five patient reported symptoms explain a significant amount of the variance in VF damage. Asking patients about their symptoms may optimize patient-physician communication and be a useful adjunct to clinical testing in some patients to estimate disease severity.