Abstract
Purpose :
Though automated perimetry using the 10-2 protocol is one of the main screening tests for hydroxychloroquine (HCQ) retinopathy, its subjective nature makes it liable to ambiguous results. Studies on the expected range of this test’s results in patients taking HCQ but without retinopathy are limited. We performed a retrospective cohort study to characterize this “normal” range of 10-2 automated perimetry results in patients screened for HCQ retinopathy who lack confounding ocular disease or any evidence of retinopathy on spectral-domain optical coherence tomography (SD-OCT) or multifocal electroretinography (mfERG).
Methods :
We reviewed medical records of 669 adult patients seen in two tertiary-care ophthalmology clinics for HCQ retinopathy screening and found 80 patients meeting the following criteria: visual acuity Snellen 20/25 or better in each eye, at least two Humphrey Field Analyzer 10-2 visual field tests (HVF 10-2) with white targets performed on different dates, at least one mfERG or SD-OCT without evidence of HCQ toxicity performed no more than 30 days prior to the patient’s latest HVF, and no eye disease besides mild dry eye or non-visually significant cataract. One eye from each qualifying patient was randomly selected, and for each study eye we analyzed up to four HVFs (the first two and the last two consecutive HVFs). HVF results from all study eyes were analyzed together.
Results :
We found that 54/80 (68%), 58/80 (73%), 24/42 (57%), and 15/18 (83%) of first through fourth HVFs were reliable, respectively. Only 62% of eyes with unreliable first HVFs had reliable second HVFs, and 22% of eyes with reliable first HVFs had unreliable second HVFs. In reliable HVFs, the median MD ranged from -0.86 to -1.79, and the median PSD ranged from 1.17 to 1.25. About 25% of reliable HVFs had MD less than -2.00, while up to 20% had PSD greater than 2.00. The proportions of reliable HVFs with 2 or more contiguous points on the pattern deviation plot with sensitivities depressed below that in 98% of normal patients were 43%, 26%, 38%, and 33% for first through fourth HVFs, respectively.
Conclusions :
There is considerable variation of results even from reliable HVFs, even with repeat testing, and some MD values overlap with those reported in the literature for patients with early HCQ retinopathy. We caution against decisions to discontinue HCQ use based solely on nonspecifically abnormal HVF results.
This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.