creening for ight-threatening diabetic retinopathy: comparion of fundu photography with automated color contrat threhold tet

Purpoe To compare tritan contrat threhold (TCT) with fundu photography in creening for ight-threatening diabetic retinopathy (TDR), before ignificant viual lo. Deign Propective, comparative tudy. Method A total of 510 conenting diabetic patient attending a hopital-baed photographic creening clinic were recruited over a 2-year period. Excluion criteria included viual acuity of wore than 6/9, previou photocoagulation, and a hitory of previou eye dieae known to affect color viion. The automated TCT tet wa performed uing a computerized cathode ray tube-baed technique. Retinal photography wa performed uing a Polaroid mydriatic fundu camera with a 45 degree field. Grading of diabetic retinopathy wa carried out by an ophthalmologit uing lit-lamp biomicrocopy and a 78-diopter len. Aement of enitivity, pecificity, and predictive value for both fundu photography and the TCT tet were made. Reult Both the fundu photography and TCT tet correlated ignificantly with the preence of TDR (P < .0001, χ2 tet). The TCT tet yielded a enitivity of 94% (95% confidence interval [CI], 73%– 100%) and a pecificity of 95% (95% CI, 92%– 96%) for detection of TDR compared with a enitivity of 88% (95% CI, 66%–97%) and a pecificity of 95% (95% CI, 93%–97%) with fundu photography. Combined modality improved overall creening performance. Concluion The TCT aement i an effective and clinically viable technique, in comparion with fundu photography, to creen for TDR among a diabetic population. Additionally, our reult alo howed that combining the TCT tet with fundu photography greatly increae the performance of creening for TDR.