[Vitreal surgery and etiological diagnosis of bacterial endophthalmitis]

To evaluate the clinical effects of vitreal surgery and the importance of etiological diagnosis in the treatment for bacterial endophthalmitis.

Retrospective series case study. 20 cases (20 eyes) of bacterial endophthalmitis that were treated in Peking Union Hospital were enrolled. 14 eyes were post-traumatic endophthalmitis, and 6 eyes were postoperative endophthalmitis. Twenty cases of aged 3 to 83 years [mean (40.5 ± 23.9) years] were enrolled, including 12 male and 8 female patients. Pre-operative visual acuity: 2 cases were able to count fingers, 6 cases were able to perform hand movement, 11 cases had light perception, light projection were uncertain in all cases, and there was no light perception in 1 case. Hypopyon was seen in 13 eyes. Severe anterior chamber inflammatory reaction was seen in the other 7 eyes. The fundus could not be observed in all 20 eyes. B-Scan ultrasound examination indicated that all 20 eyes displayed moderate to severe vitreous opacity; proliferation and organization were apparent in 12 eyes, and retinal detachment in 2 eyes. Vitrectomy and intravitreal injection of antibiotics were performed in 18 eyes, and only intravitreal injection of antibiotics was administered in the other 2 eyes. At the beginning of operation, vitreous fluids were smeared and Gram stained. To eyes that showed a positive result in Gram staining, 1 mg of Vancomycin was injected into the vitreous cavity or added in the perfusion fluid (balanced salt solution, BSS) in the eyes. To eyes that showed a negative result in Gram staining, 2 mg or 4 mg of Ceftazidime was injected into the vitreous cavity or added in the perfusion fluid (BSS) in the eyes, respectively. Additionally, we examined the vitreous specimens and performed drug susceptibility testing of the bacteria cultured from the specimens. The antibiotics that the bacteria were susceptible to were chosen according to the drug sensitivity tests. The follow-up period is from 1 to 102 months (average 16.6 months).

Thirteen eyes presented with a positive Gram staining result, and 3 eyes presented a negative result; the other 4 eyes were not infected. Bacteria were cultured in 15 eyes. The detection rate of pathogen was 75%. The result for 11 eyes was consistent with the smear results. The bacteria detected were Staphylococcus aureus in 3 eyes, Staphylococcus epidermidis in 3 eyes, and Bacillus spp in 2 eyes. Streptococcus pneumoniae, Streptococcus mitis, Plesiomonas, Pseudomonas cepacia, Klebsiella oxytoca, Loffi Acinetobacter and Pseudomonas fluorescens were detected in 1 eye. The remaining 5 eyes did not have bacterial growth. The intraocular infection of all 20 eyes was controlled, and the intraocular inflammation was relieved. The visual acuity was significantly elevated. Postoperative visual acuity achieved were ≥ 0.3 in 4 eyes, 0.1 to 0.2 in 4 eyes, 0.02 to 0.09 in 6 eyes, CF in 2 eyes, HM in 3 eyes and LP in 1 eye. The retinas of 17 eyes were normal, but recurrent retinal detachment occurred in the other 3 eyes, postoperatively.

Vitrectomy combined with antibiotics and intravitreal injection of antibiotics were an effective treatment of bacterial endophthalmitis. We obtained the vitreous fluid smears at the beginning of surgery to quickly and accurately obtain etiological diagnoses by Gram staining. It is crucial to use etiological diagnosis to choose the susceptible antibiotics.