Purpose To determine trends in the microbial spectrum of endophthalmitis over the past 25 years and to review its antibiotic susceptibility patterns over the last 10 years. Methods Microbiology records of culture-positive endophthalmitis cases from 1991 to 2015 were reviewed. Additionally, data between 2005 and 2015 was also analyzed for trends in antibiotic susceptibility. Results Of the total of 9278 patients, 3319 (35.7%) were culture positive and included bacteria (2840/3319, 85.56%), fungi (387/3319, 11.66%), and mixed cultures (92/3319, 2.7%). Gram-positive bacteria accounted for 67.68% (1922/2840) of the total bacteria seen, with the most prevalent pathogen being Streptococcus pneumoniae and Staphylococcus epidermidis. Among the gram-negative organisms Pseudomonas aeruginosa was the most prevalent while. Aspergillus flavus was the most common fungus isolated and Candida sp. accounted for 6.9% of the total fungi isolated. There was no significant change in the trends of bacteria isolated during the study period. Overall susceptibility patterns showed that gram-positive bacteria were most susceptible to vancomycin (96%) and fluoroquinolones (89%). The resistance to ceftazidime increased from 31% in 2005 to 62% in 2015 (P = 0.006) and amikacin decreased from 36% in 2005 to 33% in 2015 (P = 0.782). Although a significant trend (P < 0.001) toward increasing microbial resistance against cephalosporins and fluoroquinolones was observed, decreasing microbial resistance against glycopeptides and aminoglycosides was also detected. Conclusion The spectrum of pathogens causing endophthalmitis at our institute remained similar over the study period. These findings impact the empiric treatment and choice of antibiotics in patients with endophthalmitis.

To evaluate the clinical utility of high-throughput sequencing (HTS) approach-based analysis of the bacterial and fungal genome in vitreous fluids from patients clinically diagnosed as endophthalmitis, we subjected 75 vitreous fluids from clinically presumed infectious endophthalmitis patients to high-throughput sequencing (Illumina HiSeq 2500) after DNA extraction and amplification of the 16S rRNA for the detection of bacteria, and ITS 2 region for detection of fungal pathogens. As controls, we included vitreous biopsies from 70 patients diagnosed with other non-infectious retinal disorders. Following the construction of the curated microbial genome database and filtering steps to reduce ambiguousness/contaminants from the environment, the paired reads were analyzed. Our HTS reads revealed in almost all cases the same organism that was grown in culture (bacterial-14/15, fungal 3/3) by conventional microbiological workup. HTS additionally diagnosed the presence of microbes in 42/57 (73.7%) patients who were conventionally negative (fungal pathogens in 36/57, bacterial pathogens in 11/57, including five cases that showed the presence of both bacterial and fungal organisms). Aspergillus sp., Fusarium sp., Exserohilum sp., and Candida sp. were the most predominant genera in our cohort of culture-negative endophthalmitis cases. Heat map based microbial clustering analysis revealed that these organisms were taxonomically similar to the species identified by conventional culture methods. Interestingly, 4/70 control samples also showed the presence of bacterial reads, although their clinical significance is uncertain. HTS is useful in detecting pathogens in endophthalmitis cases that elude conventional attempts at diagnosis and can provide actionable information relevant to management, especially where there is a high index of suspicion of fungal endophthalmitis, particularly in tropical countries. Outcome analyses and clinical trials addressing the success and cost savings of HTS for the diagnosis of endophthalmitis are recommended.

No abstract

No abstract