Clinical outcomes of keratitis.

OBJECTION

To analyse the patient, clinical and microbiological variables associated with poor outcomes from keratitis in patients presenting to a major public hospital in Australia.

METHODS

A retrospective audit of the records of all patients who had a corneal scraping in 5 years at Princess Alexandra Hospital (Brisbane, Australia) was carried out. The outcome of a patient's episode of keratitis was classified as poor if they had final visual acuity of 6/60 or worse; had vision loss during treatment; or a complication of keratitis; or needed surgical intervention.

RESULTS

A final outcome was established in 207 cases during the 5-year period. Final vision of 6/12 or better was found in 48% (100) of cases while a poor outcome was seen in 28% (58). Linear regression showed poor outcomes were directly associated with age (P < 0.001) and disease severity (P < 0.001). Univariate analysis indicated that poor outcomes were more likely in patients who had had prior ocular surgery (P = 0.005) or ocular surface disease (P = 0.01) and were also associated with presenting visual acuity of worse than 6/60 (P < 0.001) and isolation of Streptococcus pneumoniae (P = 0.002). While patients with traumatic keratitis, contact lens-related keratitis or negative corneal cultures (P = 0.009) were more likely to have good outcomes. Multivariate analysis showed that the relative risk of a patient having a poor outcome was 4.3x (CI 2.0-9.5) if they had severe keratitis, 4.1x (CI 1.8-9.5) if they had keratitis related to ocular surface disease and 3.8x (CI 1.8-8.3) if they were over 50 years old.

CONCLUSIONS

An outcome of poor vision, vision loss during treatment, surgical intervention or complication of keratitis is more likely in patients with severe keratitis, keratitis related to prior ocular surface disease or older age.