Microbial Keratitis: A 5-Year Experience From Bolton Eye Unit

Purpose – Microbial keratitis remains one of the commonest acute eye conditions presenting to Eye Casualty. Corneal scrapes are often performed to guide treatment but the practice of when and how varies between units. At our district general hospital, Levofloxacin, a 3^rd generation fluoroquinolone (FQs) is the treatment of choice. We aimed to determine the yield of positive cultures from corneal scrapes and determine the influence of lesion size on a positive yield in addition to the effect of pre treatment with chloramphenicol. We also aimed to assess the antibiotic susceptibility of the organisms grown to determine if empirical treatment with topical Levofloxacin offers sufficient coverage.

Setting - A district general hospital in the North West of England

Methods – We conducted a 5-year retrospective study from January 2014 until March 2019. We used a combination of patient notes and electronic patient record (EPR) to identify a total of 218 eyes with a diagnosis of microbial keratitis. The mean age was 49.3 ± 2.672 (95% CI) with a male to female ratio: 1:1.33. Corneal epithelial defects (ED) were classified according to size where small was ≤ 1 x 1mm, medium was 1.1 x 1.1 – 4 x 4mm and large was > 4 x 4mm. Antibiotic sensitivities from culture were documented in addition to use of chloramphenicol (CPL) prior to corneal scraping.

Results – Contact lens use was associated with 42.7% of all cases (n=93). The overall yield of positive culture were 43.6% (n=95). The distribution of corneal EDs was small 54%, medium 34% and large 12%. Yield of positive culture according to size of ED was: small 37.6%, medium 47.3% and large 59.3%. Two-thirds of organisms were gram positive (65.3%) compared to one-third which were gram negative (34.7%). Seven percent had superadded infection with fungal species. FQ sensitivity was identified in 73.5% of Gram-positive and 88.5% of Gram-negative organisms. Of these, 92.7% (n=202) were commenced on a FQs as the initial empirical therapy. Twenty one cases reported CPL usage prior to scraping with 52.4% being found to be culture positive.

Conclusion – Small ulcers have a lower yield suggesting that empirical treatment with FQs without scrapes can be considered. Levofloxacin remains an excellent initial empirical antibiotic therapy. Prior Chloramphenicol use does not appear to have a significant influence on a positive yield so corneal scrapes should still be performed if clinically indicated.