Abstract
Purpose: :
To evaluate the utility of ring ratios in detecting hydroxychloroquine related retinal toxicity using the 103-hexagon multifocal electroretinogram (mfERG).
Methods: :
Twenty-three patients taking hydroxychloroquine who were consecutively evaluated for retinal toxicity were divided into those without evidence of such (HCQ-nontoxic group) and those with documented visual field loss (HCQ-toxic group). A control group of patients without retinal disease and not on hydroxychloroquine was used as a comparison group. Response densities for 103-hexagon P1 mfERG amplitudes were analyzed by averaging the 103 responses into six concentric rings [R1 - R6], calculating standard ring ratios [R1/R2, R1/R3, etc.], and auxiliary ring ratios [(R1+R2)/R3, (R1+R2)/R4, etc.]. Receiver Operating Characteristic (ROC) curves were utilized to ascertain the best test for detecting toxicity.
Results: :
The HCQ-toxic group had statistically lower absolute ring responses compared to either controls or HCQ-nontoxic groups (P < 0.043, post-hoc Games-Howell ANOVA). Low absolute ring values and ring ratios were a more powerful discriminant between the HCQ-toxic group and controls than high values. ROC analysis revealed that the most powerful discriminant was the absolute ring value R2 (AUC of 0.829). At a 90% sensitivity threshold of the R2 test, the associated specificity was 44%. Standard and auxiliary ring ratios were less powerful discriminants than the absolute R2 ring values.
Conclusions: :
Absolute 103-hexagon central foveal and pericentral mfERG responses were more useful indicators of hydroxychoroquine toxicity than ring ratios in our patient population. ROC analysis suggested that quantitative ring amplitude analyses for the 103-hexagon test may be more useful as a confirmatory test for detection of hydroxychloroquine toxicity than as a screening test.
Keywords: retina • drug toxicity/drug effects • electroretinography: clinical