Comparison of polymerase chain reaction and standard microbiological techniques in presumed bacterial corneal ulcers

To assess utility of PCR in the diagnosis of bacterial corneal ulcer and to compare sensitivity and specificity of this technique with conventional laboratory methods. A prospective nonrandomized investigative study conducted on 122 eyes of presumed bacterial keratitis. Samples were collected for bacterial and fungal culture and Gram stain smear. A separate sample was taken for PCR with 26 gauge needle and was dipped directly into Eppendorff tube with lysis buffer in it. Diagnosis of culture proven bacterial keratitis was established in 53 (43.4 %) and most common bacterial isolate was staphylococcal sp. (83 %). Direct microscopic examination of Gram stained smear revealed presence of bacteria in 24 (23.9 %) specimens and PCR positivity was evident in 56 (45.5 %). In preantibiotic treated eyes culture was positive in 15 (30 %), Gram stain in 9 (18 %), and PCR in 18 (36 %). The same for untreated (fresh) eyes, positivity of culture as well as PCR was noted in 38 (52.7 %) and that of Gram stain was noted in 20 (27.7 %). Sensitivity of Gram stain and PCR was 45.28 and 88.68 % respectively; whereas specificity was 92.75 % for Gram stain and 86.96 % for PCR. The average time taken for PCR reaction was 4-8 h while culture reporting took at least 24-48 h. Our findings suggest that PCR is a good adjunct modality to the "Gold Standard" technique in the diagnosis of bacterial corneal ulcer.