Author: Johannes Keller (United Kingdom)
Co-authors: Jordan Chervenkoff, Sidath E Liyanage
Purpose
To characterise cases of rhegmatogenous retinal detachment (RD) occurring after postoperative endophthalmitis (PE), describe outcomes and identify risk factors.
Setting/Venue
Tertiary referral centre.
Methods
Retrospective analysis of electronic patient and laboratory records from consecutive patients with RD occurring after PE, identified after cross-referencing clinical audit databases of PE and RD for the period between 01/01/2013 and 01/07/2020. Fungal, chronic and endogenous endophthalmitis or cases following trauma were excluded. The main outcome measure was best corrected visual acuity (BCVA) in logMAR units. Secondary measures were proportion of patients achieving 0.3 logMAR or better, 1.0 logMAR or worse, proportion requiring eye removal and proportion of phthisis. Putative univariate risk factors studied were bacterial taxonomy, initial mode of treatment, delay to antibiotic delivery, pars plana vitrectomy (PPV) surgery, inciting operation, and grade of surgeon performing PPV. Continuous variables were analysed for normal distribution with the Kolmogorov-Smirnov test and the Mann-Whitney rank-sum test was used to compare means. Discrete variables were analysed using the Fisher’s exact and Chi-squared tests as appropriate.
Results
During the study period, 103 cases of PE were treated at our centre. Two cases of fungal and one case of chronic endophthalmitis were excluded. Six patients were postoperatively managed at their distant area of residence and were excluded. Amongst the remaining 94 eyes, the frequency of RD was 21 (22%). The average age was 70 years and there were 14 male patients. The commonest inciting intervention leading to RD was intravitreal injection ([IVI] 8/21, 38%) followed by cataract surgery (7/21, 33%). Eight eyes (38%) presented with macula-involving RD and three had total RD (14%). Nine cases (43%) had RDs confined to one quadrant. Sixteen cases had one retinal break (76%) and in two cases no breaks were found (10%). Only 1 case (5%) had PVR at the time of presentation. Twenty cases underwent surgical repair with PPV and one patient declined surgery. In 16 cases (76%) silicone oil was used and 12 cases (57%) received a dose of antibiotics at the time of surgery. Seven cases (35%) experienced a recurrent RD. Eight eyes (40%) had one operation alone, with a mean number of operations of 2.1. In 13 cases (62%) the RD was identified during PPV where the indication was PE. In 6 cases (30%) the retinal break was iatrogenic. Seven eyes (35%) were left with permanent silicone oil fill. All 21 RD eyes had had PPV as part of the treatment for PE. None of the patients treated with intravitreal antibiotics developed RD (0%), whereas 13/38 (34%) patients receiving intravitreal antibiotics followed by PPV and 8/33 (24%) receiving only PPV did (p<0.01). Vitrectomy surgery carried out by trainee surgeons resulted in 14/36 (39%) retinal detachment (range:0%-100%) whereas if operated by consultant surgeons the rate was 7/35 (20%, range:11-50%, p<0.05). The commonest bacterium in RD cases was Staphylococcus epidermidis (6/21, 29%). All 3 cases with Enterococcus PE developed RD. There was a delay of a day or greater in antibiotic injection in 5% of cases resulting in RD and 8% of cases not having a RD (p=0.4). The BCVA at presentation was 2.1 for cases not resulting in RD and 2.2 for those complicated with RD (p=0.21). 30% of the IVI-related PE was accompanied by RD, 23% of PPV-related and 18% of cataract surgery related (p=0.57). Eyes with RD needed an average of 2.4±1.1 antibiotic injections whereas those with no RD were treated with 1.5±0.6 injections (p<0.01). Eyes with RD received 3.6±1.8 operations versus 1.0±0.8 for those without RD (p<0.01). After PE-RD the mean BVCA was 1.6±1.1, compared with 0.8±0.9 for PE not complicated with RD (p<0.01). Fifty-seven percent (12/21) of eyes with RD attained BCVA of 1.0 or worse versus 29% of PE without RD (p=0.01). Nineteen percent (4/21) of eyes with RD attained BCVA of 0.3 or better whether those without RD did so in 43% of cases (p=0.02). Five RD eyes (24%) resulted in phthisis and there were 2 eyes with phthisis without RD (3%,p<0.01). Three cases required removal of the eye, none had RD (4%,p=0.46).
Conlusions
The outcomes of retinal detachment after endophthalmitis are poor in many cases and worse than those of endophthalmitis cases without this complication. In our sample, risk factors for RD after endophthalmitis included: initial mode of antibiotic delivery, performance of vitrectomy, seniority of surgeon carrying out vitrectomy and potentially the virulence of the bacterium.
Financial Disclosure
University Hospitals Bristol and Weston NHS Foundation Trust, employment Newmedica Bristol, partner Alcon, educational grants received Bausch and Lomb, educational grants received Daybreak Medical, educational grants received DORC, educational grants received
Comments
-
