Comparison of clinico-microbiological profile and treatment outcome of in-house and referred post cataract surgery endophthalmitis in a tertiary care center in South India

Patients with post cataract surgery endophthalmitis treated by us at a tertiary eye care facility in Hyderabad, India, are a pool of “in-house” and “referred” cases for further management. A decade and half ago, we reported the microbiological profile of these pooled patients [1] in which the gram-positive cocci accounted for less than 50% eyes and gram-negative bacilli were detected in more than 25% eyes. This was contrary to the report of the Endophthalmitis Vitrectomy Study (EVS) [2] results published before our report and also is contrary to the European Society of Cataract and Refractive Surgery (ESCRS) [3] study results published after our report. In view of these differences, we reexamined the microbiological profile and antibiotic sensitivity between two distinct groups of patients, in-house and referred, managed at this facility. We hypothesized that the in-house endophthalmitis profile (both clinical and microbiological) and outcome after treatment could be different.

The clinical records of all culture-positive cases of post cataract surgery endophthalmitis managed at our institute between January 2010 and December 2014 were reviewed. There were 50 cases of in-house (group A) endophthalmitis. These 50 cases were matched with the first 50 consecutive culture-positive referred cases in the same period (group B; beginning December 2014 backwards and ending March 2010). Two cluster endophthalmitis cases, one in-house (reported earlier) [4] and one referred (reported by another treating center) [5] were excluded from both groups. Institutional review board approval (LEC 09-15-110) was obtained for retrospective data collection and analysis.

A detailed ocular history (cataract surgery, event to onset of symptoms, time to presentation, and previous treatment history in referred cases), the demography (age, gender), laterality, systemic factors (diabetes and immune status) were recorded. Management of all these patients was as per the institutional protocol. Briefly, this consisted of a comprehensive clinical examination (uncorrected and best corrected visual acuity, slit lamp biomicroscopy, applanation/digital tonometry, indirect ophthalmoscopy) and ultrasonography. The severity of inflammation in presentation was measured in all cases by our earlier published inflammatory score [6]. Briefly, this consisted of slit lamp and indirect ophthalmoscopic features of the cornea, anterior chamber, iris, and vitreous in a scale of 0 (no involvement) to 4 (gross involvement) and additional allowance given for any opaque ocular tissues not allowing further examination. The treatment consisted of intraocular antibiotics (vancomycin 1 mg in 0.1 ml normal saline; ceftazidime 2.25 mg in 0.1 ml normal saline) with either vitreous biopsy or core vitrectomy. This was followed with topical antibiotics (every 2 h), topical prednisolone acetate 1% (every 4 h) except in the eyes with significant corneal infiltrate and cycloplegic (every 8 h). The use of intravitreal corticosteroid (typically, dexamethasone 0.4 mg in 0.1 ml) and systemic antibiotics (typically, oral ciprofloxacin 1500 mg/day in two divided doses, in adult) were left to the discretion of the treating faculty. The vitreous sample was plated on the same day for culture. Microbiological work-up of undiluted vitreous included microscopy (after vital dye staining), culture (aerobic, anaerobic, and fungus), and antibiotic sensitivity (Kirby-Bauer disc diffusion). Unlike the current institutional standard, the polymerase chain reaction (PCR) was not done routinely in this period. Any further administration of intravitreal antibiotics with or without corticosteroid was guided by the culture report of the vitreous sample and the sensitivity of the cultured microorganism.

Endophthalmitis, the most dreaded complication following cataract surgery, has an incidence from 0.04% to 0.13% [7‐11]. Gram-positive bacteria were the predominant isolates in the EVS [2] and the ESCRS [3] studies. We and others in India have reported different microbiological profile of infection where the gram-positive cocci infection was relatively less, and gram-negative bacilli infection was relatively more common [1, 12]. These reports included both in-house and referred cases of endophthalmitis. Two Indian reports have published the in-house endophthalmitis data [13, 14]. In our report [14], culture-positive endophthalmitis was 0.07%; in this cohort 64.8% (n = 24/37) grew gram-positive cocci (n = 16/37, 43.2% were Staphylococcus epidermidis) and 24.3% (n = 9/37) grew gram-negative bacilli (n = 5/37, 13.5% were P. aeruginosa). In the other report, [13] clinical endophthalmitis was 0.09%, but the microbial profile was not part of this report.

Our hypothesis that the profiles of the two groups of endophthalmitis, in-house and referred could be different, was only partly true. There were differences in two areas—one, in the reporting time and two, in the microbiological spectrum. The delay in reporting time of the referred patients was likely because it depended on the referring ophthalmologist. The detection of more gram-positive cocci in the in-house cases and more gram-negative bacilli in referred cases was indeed a surprise (Table 2). The microbial spectrum of in-house endophthalmitis was closer to the EVS [2] and ECRS [3] reports. It is possible that only more fulminant cases that did not respond to initial treatment, invariably caused by gram-negative or similar virulent organisms, were referred to our facility.

The antibiotic sensitivity profile was nearly similar to our earlier report. Gram-positive bacilli had good sensitivity to vancomycin in either group, similar to the EVS report [2], and the gram-negative bacilli had poor sensitivity to ceftazidime in either group, similar to our earlier report [1]. All ceftazidime resistant gram-negative bacilli in group A (in-house cases) were sensitive to imipenem. Additionally, 77.78% of resistant isolates from referred cases were sensitive to amikacin. Imipenem appears to be good choice in gram-negative organisms resistant to ceftazidime and amikacin [15‐17]. Amikacin could still be an alternative choice, though we have already reported emergence of gram-negative bacteria resistant to both ceftazidime and amikacin [18]. Hence it is prudent to always do a culture-sensitivity test for appropriate management.

The limitations of this study included small number of cases (50 in either group), small number of individual isolates tested for sensitivity, and non-inclusion of PCR as routine basis. The strength of the study is the use of uniform management protocol in both groups. The study clearly demonstrates that the post cataract surgery infection pattern in India may not be significantly different from the ones reported by the EVS and ESCRS reports, though one must employ wiser discretion in treating the in-house and referred patients in a referral facility such as ours.