Purchase this article with an account.
Afsana Karim, Igor Kozak, Dirk-Uwe G. Bartsch, Hector Lemus, Laurie Dustin, Jay Chhablani, Giulio Barteselli, Hyan Wang, Stanley P. Azen, William R. Freeman; The Best Functional Predictor of HIV Status in Relation to the Retinal Damage. Invest. Ophthalmol. Vis. Sci. 2012;53(14):6219. doi: https://doi.org/.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Recent studies have shown structural and functional subclinical damage in retinas of HIV+ patients without retinitis leading to definition of HIV-related neuroretinal disease as an entity in HIV disease. We attempted to determine which measure of vision dysfunction best correlates with HIV disease severity.
A prospective longitudinal study of 237 HIV+ patients and HIV- controls. Study participants were recruited from the UCSD HIV Neurobehavioral Research Center and UCSD Jacobs Retina Center. We performed ETDRS visual acuity (VA), Peli Robson contrast sensitivity score (CS), Farnsworth 100 hue color discrimination scores (CV) and Humphrey visual field global error scores (MD and PSD). We compared means of functional measures across three different groups and indentified which group is significantly different from others. For progression analysis we performed generalized estimation equations (GEE) model controlling for age, gender and HIV duration. We identified the best predictor of CD4 status (HIV burden) among these functional measurements by using polychotomous logistic regression where normal group is selected as reference group. We compared the Akaike's Information Criterion (AIC) between the regression models assuming the model with lowest AIC is best one predicting HIV status.
The study included 77 HIV+ patients with low CD4 count (100) as high group and 52 HIV- subjects. We took the second visit score to compare the mean across groups eliminate practice effect. For CV and VA the difference was significant (0.0005, 0.0179, respectively) across the groups. For CV the significant difference was between low and normal (0.000) and high and normal (0.000) groups. For VA the significant difference was between low and normal (0.0179). For VF PSD measure, there was a significant (0.007) difference across groups. The difference was between low and normal (0.0022). We found significant association between VF PSD measure and HIV status. For VF PSD measure the odds ratio for low vs. normal was 3.28 (95% CI 1.26, 8.6) and for high vs. normal it was 2.93 (95% CI 1.21, 7.78). AIC scores for VA, CS, CV, VF MD measure and VF PSD measure were 289.8, 247.8, 240.2, 187.9 187.2, respectively. PSD had the lowest AIC which showed this measure performed best among all in explaining the HIV status in relation to the retinal damage. For progression analysis we did not find any measures showing progressive loss of function.
The HIV+ group with low CD4 count had more functional deficits than the HIV+ group with high CD4 count and HIV negative group. There was no progressive pattern of any of the functional measures. PSD of Humphrey visual field test is the best functional measure in predicting HIV status in relation to the retinal damage.
This PDF is available to Subscribers Only