The accuracy of binocular fixation preference for the diagnosis of strabismic amblyopia

PurposeTo determine the accuracy of different scales of the binocular fixation preference test and 10Δ fixation test in diagnosing amblyopia in patients with strabismus.MethodsA prospective and masked diagnostic study was undertaken. We compared 3 binocular fixation preference test scales with interocular Early Treatment Diabetic Retinopathy Study (ETDRS) acuity differences. The 10Δ fixation test was used for patients with strabismus <10Δ. Sensitivity, specificity, and likelihood ratios for amblyopia with an interocular difference of ≥2 lines were determined. Intra- and interexaminer agreements were measured.ResultsThe study included 221 literate strabismic patients. The accuracy of the binocular fixation preference test was similar for all scales, with no advantage in combining them (Cronbach's α = 0.99). The sensitivity and specificity of binocular fixation preference among patients with strabismus ≥10Δ were 72.8% (95% CI, 59.7%-83.6%) and 77.6% (95% CI, 66.6%-86.3%), respectively; among patients with deviations <10Δ, these were 89.6% (95% CI, 72.6%-97.8%) and 64.2% (95% CI, 44.0%-81.3%), respectively. The 10Δ fixation test did not alter this accuracy. Likelihood ratios were stronger for extreme grades of the binocular fixation preference test; however, in intermediate grades, they only changed by approximately 15% of the pretest probability of amblyopia. Intra- and interexaminer agreements were 76% (95% CI, 51%-100%) and 73% (95% CI, 48%-97%), respectively.ConclusionsThe binocular fixation preference test is more useful for diagnosing amblyopia when the results indicate either a very strong preference or no fixation preference. The intermediate grades of the test were less accurate in our study, accounting for most of the false positives and negatives results.