Purchase this article with an account.
Oliver L. Yeh, Roshni Singh, Anna K Junk, Kara Marie Cavuoto, Ta Chen Chang; Preliminary Assessment of Humphrey Visual Field Automated Perimetry Reliability Indices as a Predictor of Mortality Risk. Invest. Ophthalmol. Vis. Sci. 2016;57(12):3914.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Humphrey Visual Field (HVF) testing requires patients to have the ability to follow a set of instructions, the visual and mental consciousness to notice a visual stimulus, and the psychomotor skill to press a button upon seeing the stimulus. HVF reliability indices such as fixation loss (FL), false positives (FP), and false negatives (FN) have traditionally been used to assess the usefulness of a given test. Unreliable fields have been generally discarded as not clinically useful. We ask the novel question as to whether increasing unreliability in a HVF test correlates to increased mortality risk.
All patients carrying a diagnosis of glaucoma at the Miami Veterans Affairs Healthcare System between 2001-2013 were included in a retrospective case-controlled study. The study group consisted of 27 patients with at least 4 HVF of stimulus size III studies of the same type who died within 6 months of their most recent visual field. These patients were matched with controls (n=27) who lived at least one year past their last visual field with the same adjusted age (age at last HVF + 3 months) ± 3 years, visual field severity, and visual field testing type. Visual field severity was graded using Hodapp-Anderson-Parrish criteria. In each of the study and control groups, there were 10 patients with early, 12 with moderate, and 5 with severe visual field loss at baseline measurement.
There was no significant difference in the rate of change for visual field reliability metrics (FL/FP/FN) between the study and control groups. For the study group vs controls, change per month for fixation losses was 0.89% ± 1.23 vs 0.33% ± 0.69 (p=0.370), for false positives -0.31% ± 0.26 vs 0.04% ± 0.51 (p=0.212), and for false negatives 0.16% ± 0.38 vs 0.06% ± 0.20 (p=0.612). Furthermore, when stratified by severity of baseline visual field defect either within or between the groups, there was no difference in the rate of change for any reliability metric.
Our findings suggest that unreliability in visual field testing does not portend impending mortality. Limitations to the study include possible insufficient statistical power as well as its retrospective nature; in particular, patients who were the most sick may have refused visual field testing, and therefore could not be analyzed.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.
This PDF is available to Subscribers Only