Abstract
Objectives
To evaluate result of early pars plana vitrectomy (PPV) within 24 hours of presentation for acute postoperative endophthalmitis after cataract operation, and to determine factors that predict visual outcome.
Methods
Consecutive patients who developed acute postoperative endophthalmitis within 6 weeks after cataract operation were reviewed. Patients were divided into two groups for analysis: (1) those receiving PPV within 24 hours of presentation (early PPV group), and (2) those receiving initial intravitreal antibiotics only without PPV within 24 hours of presentation (IVA group).
Results
Out of 41,411 cataract operations, 22 eyes developed acute postoperative endophthalmitis. Presenting VA was hand-movement or worse in 72.7%. The most common organisms were Staphylococcus (40.9%), Streptococcus (13.6%) and Enterococcus (13.6%). 22.7% of eyes had good final VA ≥ 20/30 and 27.3% had poor final VA < 20/400. Early PPV group had significantly lower rate of requiring additional treatments to control infection (25% versus 80%, P = 0.030), higher rate of retinal detachment (25% versus 0%, P = 0.221) and similar final logMAR VA (1.08 ± 1.08 versus 0.80 ± 0.80, P = 0.489) compared to IVA. Multivariate linear regression analysis showed that worse final VA was significantly associated with Streptococcus (ß = 1.92, P = 0.007) and retinal detachment (ß = 1.72, P = 0.005) but not with early PPV (P = 0.225).
Conclusion
Early PPV was superior to initial intravitreal antibiotics alone as it required fewer additional treatments to control infection. Visual outcome was similar between early PPV and initial intravitreal antibiotics alone despite high number of poor presenting VA of light-perception in early PPV group. Streptococcal infection and retinal detachment were major poor prognostic factors for vision.
Subject terms: Eye diseases, Infectious diseases
Introduction
Cataract operation is one of the most frequently performed intraocular surgeries nowadays. Postoperative endophthalmitis is a rare but serious complication that can lead to rapid, profound visual loss and blindness. The incidence of acute postoperative endophthalmitis after cataract operation was reported to range from 0.02 to 0.20% [1–6].
Although pars plana vitrectomy (PPV) was known to be effective for postoperative endophthalmitis [7–11], it is still controversial whether PPV should be performed immediately after presentation or reserved until failing initial intravitreal antibiotics [12, 13]. The best timing of PPV remains unknown. Endophthalmitis Vitrectomy Study (EVS) [14] suggested immediate vitrectomy within 6 hours was superior to intravitreal antibiotics only when the presenting visual acuity (VA) was light perception (LP) or worse. EVS was conducted around 30 years ago and it had not addressed the advancement of vitrectomy technology and change of cataract operation techniques from large-wound extracapsular cataract extraction to small-wound phacoemulsification over the past 3 decades. In addition, the criteria of EVS were very restricted that it had excluded eyes without sufficient corneal clarity for vitrectomy and those with VA ≥ 20/100. The vitrectomy technique employed by EVS was limited targeting at least 50% of vitreous gel only. Therefore, the recommendations of EVS for PPV based on vision alone may not be applicable for nowadays practice.
EVS mandated PPV to be performed immediately within 6 hours of presentation which could be difficult to achieve in many clinical settings. However, only a few other studies have compared the result of early PPV against intravitreal antibiotics in postoperative endophthalmitis and the results were conflicting [15–17]. While Negretti et al. [15] showed that early PPV within 7 days had better final visual outcome compared to delayed PPV in post-procedure endophthalmitis, Altan et al. [16] and the European Vitreo-Retinal Society (EVRS) Endophthalmitis Study [17] showed that the visual outcome was similar after early PPV and after intravitreal antibiotics. More studies are required to determine the optimal timing of PPV and to investigate the long-term outcome. The aim of this study was to evaluate the result and long-term visual outcome of early PPV within 24 hours of presentation for post-cataract surgery endophthalmitis, and to determine factors that predict visual outcome.
Materials and methods
Patients
This is a retrospective cohort study in which the medical records of all cases of acute postoperative endophthalmitis after cataract operation managed in Prince of Wales Hospital and Alice Ho Miu Ling Nethersole Hospital, Hong Kong from January 2010 to January 2021 were reviewed. This study was performed according to the tenets of Declaration of Helsinki and approved by the Cluster Research Ethics Committee/Institutional Review Board of New Territories East Cluster, Hong Kong.
The inclusion criteria were acute postoperative endophthalmitis that developed within 6 weeks after cataract operation. The exclusion criteria include: (1) chronic endophthalmitis that developed later than 6 weeks after cataract operation, (2) cataract operation performed in combination with other intraocular surgeries e.g. trabeculectomy, corneal graft operation, vitrectomy, microinvasive glaucoma surgery, etc., (3) other types of endophthalmitis e.g. endogenous, post-traumatic, post-intravitreal injection etc.
Data recorded include age, gender, types of cataract operation, clinical features, VA, types and number of treatments, microbiological diagnosis, treatment outcome and any complications. All VA were measured with Snellen chart and converted to logMAR for statistical analysis. VA of count fingers (CF), hand movement (HM), LP and no light perception (NLP) were assigned with values of 2.0 logMAR units, 2.3 logMAR units, 2.6 logMAR units and 2.9 logMAR units respectively for statistical analysis.
Treatment regimen and study groups
In our centre, all cases of acute post-cataract surgery endophthalmitis received immediate PPV or immediate vitreous tap followed by intravitreal injection of antibiotics (vancomyin 1 mg/0.1 ml and either ceftazidime 2 mg/0.1 ml or amikacin 0.4 mg/0.1 ml) on presentation. If the patient received immediate vitreous tap and intravitreal antibiotics on presentation, prompt PPV would also be performed within 24 hours of presentation if: (1) presenting VA was LP or worse, (2) deterioration of VA to LP after initial intravitreal antibiotics, or (3) at the discretion of vitreoretinal surgeons to be in the best interests of patients.
In this study, we divided patients into two groups for analysis based on the initial treatment regimen: (1) early PPV group, consisting of patients who received PPV immediately or promptly within 24 hours of presentation, and (2) IVA group, consisting of patients who received initial intravitreal antibiotics only without PPV within 24 hours of presentation. Patients in both groups were monitored closely after initial treatment. If the clinical response was suboptimal within 48–72 hours, additional treatments such as repeated intravitreal antibiotics or PPV were performed. PPV performed after 24 hours of presentation was classified as delayed PPV in this study.
Outcome measures and statistical analysis
Primary outcome measure was VA at final visit. Secondary outcome measure was the incidence of retinal detachment (RD).
Statistical software R (R V.3.6.1) was used for statistical analysis. Demographic data was expressed as mean ± standard deviation. Paired sample T-test was used to evaluate change of VA from baseline. Comparison was performed between the early PPV group and IVA group. Two-sample T-test was used to evaluate the differences in VA and number of additional treatments between the two groups. Fisher’s exact test was used to evaluate the differences in categorical variables e.g. proportion of good final VA ≥ 20/30, poor final VA < 20/400, rate of RD and proportion requiring additional treatments between the two groups. Univariate and multivariate linear regression analyses were used to evaluate the relationship between prognostic factors and final VA. P value of <0.05 was considered as statistically significant. All tests were two-sided.
Results
Baseline characteristics
There were 41,411 cataract operations performed in the study period. No prophylactic intracameral antibiotics were used. A total of 22 patients developed acute postoperative endophthalmitis and met the inclusion criteria. The incidence was 0.053%. The mean age of patients was 75.6 ± 13.5 years (range 32–91 years). There were 10 males (45.5%) and 12 females (54.5%). In 21 cases (95.5%), the operations were uneventful with in-the-bag intraocular lens (IOL) implantation. One cataract operation (4.8%) was complicated with posterior capsular rupture, vitreous loss, conversion to extracapsular cataract extraction and implantation of IOL in anterior chamber. The overall mean follow-up time was 20.2 ± 18.3 months (range 1.1 month–64.1 months).
Presenting clinical features
Patients presented at a mean time of 6.5 ± 3.9 days after cataract operation (range 2–15 days) with eye pain, blurring of vision and/or increased floaters. The mean duration of symptoms prior to presentation was 2.1 ± 1.4 days (range 0–6 days). The mean logMAR VA on presentation was 2.17 ± 0.51 (median = HM, range 20/60 to LP). One patient did not have documented VA on presentation. The presenting VA was HM or worse in 16 patients (72.7%). There were significant hypopyon of ≥1 mm in 10 patients (45.5%) and absence of fundal view in 17 patients (77.3%). The clinical features of individual patients were shown in Table 1.
Table 1.
Clinical features and outcome of individual patients presented with acute postoperative endophthalmitis after cataract operation.
| Patient number | Age | Gender | Presenting VA | Final VA | Hypopyon ≥1 mm | Absence of fundal view | Bacterial growth | Initial treatments within 24 hours of presentation | Additional treatments after 24 hours of presentation | Reasons for early or immediate PPV | Complications | FU duration (months) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 79 | F | CF | 20/60 | y | y | No growth | IA | IA, delayed PPV | – | Nil | 49.4 |
| 2 | 91 | M | n/a | 20/120 | n/a | n/a | Staph. lugdunensis | IA | IA | – | Nil | 36.7 |
| 3 | 85 | F | 20/200 | 20/120 | n | n | No growth | IA | Nil | – | Nil | 1.1 |
| 4 | 73 | F | 20/60 | 20/30 | n | y | Staph. epidermidis | IA | Delayed PPV | – | Nil | 33.4 |
| 5 | 77 | F | HM | 20/30 | n | y | Coagulase -ve Staph. | IA | Delayed PPV | – | Nil | 36.2 |
| 6 | 81 | M | HM | 20/60 | y | y | Enterococcus sp. | IA, prompt PPV | Nil | Drop of VA to LP | Nil | 1.4 |
| 7 | 87 | F | HM | 20/60 | y | y | Coagulase -ve Staph. | IA | Delayed PPV | – | Nil | 3.7 |
| 8 | 80 | F | 3/200 | 20/80 | n | n | No growth | IA | Nil | – | Nil | 7.8 |
| 9 | 78 | F | HM | 20/80 | y | y | No growth | IA | Delayed PPV | – | Nil | 64.1 |
| 10 | 80 | M | HM | NLP | y | y | Group G Strep. | IA | Delayed PPV | – | Evisceration | 30.7 |
| 11 | 72 | F | LP | 20/120 | y | y | Staph. lugdunensis | Immediate PPV | Nil | Presenting VA = LP | Nil | 50.2 |
| 12 | 82 | M | HM | 20/30 | n | y | Pseudo. aeruginosa | IA, prompt PPV | Nil | Drop of VA to LP | Nil | 4.8 |
| 13 | 67 | M | HM | 20/16 | n | y | Serratia | IA, prompt PPV | IA | Drop of VA to LP | Nil | 20.2 |
| 14 | 86 | M | HM | 3/80 | y | n | Staph. lugdunensis | IA | Delayed PPV | – | Nil | 24.4 |
| 15 | 57 | M | HM | 20/40 | n | y | Coagulase -ve Staph. | IA, prompt PPV | Nil | At discretion of surgeon | Nil | 6.4 |
| 16 | 32 | M | HM | NLP | n | y | Group G Strep. | Immediate PPV | IA | At discretion of surgeon | Phthisis bulbi | 17.7 |
| 17 | 90 | F | LP | 1/200 | n | y | Staph. aureus | IA, prompt PPV | Nil | Presenting VA = LP | Intraop RD | 16.7 |
| 18 | 65 | F | LP | 18/200 | y | y | Strep. pneumoniae | IA, prompt PPV | Nil | Presenting VA = LP | Nil | 18.1 |
| 19 | 85 | F | HM | 20/60 | n | y | Enterococcus faecalis | IA, prompt PPV | IA, IA | At discretion of surgeon | Nil | 8.7 |
| 20 | 80 | F | HM | LP | y | y | Enterococcus faecalis | IA, prompt PPV | Nil | Drop of VA to LP | Postop RD | 7.3 |
| 21 | 57 | M | LP | HM | y | y | No growth | IA, prompt PPV | Nil | Presenting VA = LP | Postop RD | 3.1 |
| 22 | 80 | M | 2/200 | 20/20 | n | n | Staph. lugdunensis | IA, prompt PPV | Nil | At discretion of surgeon | Nil | 2.6 |
CF count fingers, F female, FU follow-up, HM hand movement, IA intravitreal antibiotics, intraop intraoperative, LP light perception, M male, n no, n/a not available, NLP no light perception, postop postoperative, PPV pars plana vitrectomy, Pseudo. Pseudomonas, RD retinal detachment, sp. species, Staph. Staphylococcus, Strept. Streptococcus, VA visual acuity, y yes, -ve negative.
Treatment group
Early PPV group consisted of 12 patients, among which 2 had immediate PPV within 6 hours of presentation (9.1%) and 10 had immediate intravitreal antibiotics followed by prompt PPV within 24 hours of presentation (45.5%). The IVA group consisted of 10 patients, among which 3 had intravitreal antibiotics only without any PPV (13.6%) and 7 had delayed PPV after 24 hours of presentation (31.8%). The reasons for early PPV in each case were shown in Table 1. In 3 patients, the endophthalmitis resolved with intravitreal antibiotics only and PPV was not required (13.6%). All PPV was performed along with intraoperative intravitreal antibiotics without any endotamponade except in one case where silicone oil tamponade was used because there was RD intraoperatively. The comparison of clinical features and outcome between the two groups was shown in Table 2.
Table 2.
Comparison of clinical features and outcome between early PPV group and IVA group.
| Early PPV group (n = 12) | IVA group (n = 10) | P value | |
|---|---|---|---|
| Baseline characteristics | |||
| Age ≥ 65 years (n, %) | 9 (75.0%) | 10 (100.0%) | 0.221 |
| Male (n, %) | 7 (58.3%) | 3 (30.0%) | 0.231 |
| Time from cataract operation to presentation (days) | 5.2 ± 3.9 | 8.0 ± 3.4 | 0.084 |
| Duration of symptoms prior to presentation (days) | 2.0 ± 1.5 | 2.2 ± 1.5 | 0.755 |
| Presenting VA | 2.38 ± 0.19 | 1.89 ± 0.68 | 0.064 |
| • >20/60 | 0 | 0 | |
| • >20/200–20/60 | 0 | 1 (10.0%) | |
| • >CF – 20/200 | 0 | 1 (10.0%) | |
| • CF | 1 (8.3%) | 2 (20.0%) | |
| • HM | 7 (58.3%) | 5 (50.0%) | |
| • LP | 4 (33.3%) | 0 | |
| • NLP | 0 | 0 | |
| • Not documented | 0 | 1 (10.0%) | |
| Significant hypopyon ≥1 mm (n, %) | 5 (41.7%) | 5 (50.0%) | 0.670 |
| Absence of fundal view (n, %) | 11 (91.7%) | 6 (60.0%) | 0.272 |
| Causative organisms (n, %) | 0.068 | ||
| • No growth | 1 (8.3%) | 4 (40.0%) | |
| • Staphylococcus | 4 (33.3%) | 5 (50.0%) | |
| • Streptococcus | 2 (16.7%) | 1 (10.0%) | |
| • Non-staphylococcus, non-Streptococcus | 5 (41.7%) | 0 (0.0%) | |
| Follow-up duration (months) | 13.1 ± 13.5 | 28.8 ± 20.2 | 0.054 |
| Visual outcome and rate of retinal detachment | |||
| Final VA (logMAR units) | 1.08 ± 1.08 | 0.80 ± 0.80 | 0.489 |
| Proportion of good final VA ≥ 20/30 (n, %) | 3 (25.0%) | 2 (20.0%) | 1.000 |
| Proportion of poor final VA < 20/400 (n, %) | 4 (33.3%) | 2 (20.0%) | 0.646 |
| Development of retinal detachment (n, %) | 3 (25.0%) | 0 (0.0%) | 0.221 |
CF count fingers, HM hand movement, IVA initial intravitreal antibiotics, LP light perception, n number, NLP no light perception, PPV pars plana vitrectomy, VA visual acuity.
Bacterial characteristics
Bacterial cultures were positive in 17 cases (77.3%) and the majority of them were Gram-positive bacteria (68.2%) followed by no growth (22.7%) and Gram-negative bacteria (9.1%). The 3 most common organisms were Staphylococcus (40.9%), Streptococcus (13.6%) and Enterococcus (13.6%).
Visual outcome
The overall mean logMAR VA increased from baseline of 2.17 ± 0.51 (≈HM) to 1.92 ± 0.62 (≈20/1600) at 1 week (P = 0.09), 1.27 ± 0.86 (≈20/400) at 1 month (P < 0.001), 1.03 ± 0.91 (≈20/200) at 3 months (P < 0.001), 1.08 ± 0.96 (≈20/250) at 6 months (P < 0.001) and 1.06 ± 0.97 (≈20/250) at 1 year (P = 0.003). The mean final logMAR VA was 0.95 ± 0.95 (≈20/180, range 20/16 to NLP) which was significantly better than baseline (P < 0.001). A total of 5 eyes had good final VA ≥ 20/30 (22.7%), 6 eyes had final VA ≥ 20/60 (27.3%) and 6 eyes had poor final VA < 20/400 (27.3%). Seventeen eyes (81.0%) had visual improvement of at least 3 Snellen lines after treatment. The complications included RD in 3 eyes (13.6%), evisceration in 1 eye (4.5%) and phthisis bulbi in 1 eye (4.5%).
All 3 cases of RD occurred in those where early PPV were performed. One case was noted to have superior RD from a superotemporal retinal hole intraoperatively during PPV. Silicone oil was used and the retina remained attached after operation. VA improved from LP at baseline to CF. One case developed RD at 30 days after PPV with proliferative vitreoretinopathy (PVR) and retinal hole at the parafoveal region. Operation was performed with vitrectomy, PVR membrane removal, endolaser retinopexy and silicone oil infusion. The retina was attached. VA dropped from HM at baseline to LP. One case developed RD at 40 days after PPV due to an inferior peripheral retinal hole. Operation was performed with vitrectomy, endolaser retinopexy and silicone oil infusion. VA improved from LP at baseline to HM.
Early PPV group versus IVA group
The eyes in early PPV group and IVA group had similar baseline characteristics and visual outcomes in terms of final logMAR VA (1.08 ± 1.08 versus 0.80 ± 0.80, P = 0.489), proportion of good final VA ≥ 20/30 (25.0% versus 20.0%, P = 1.000) and proportion of poor final VA < 20/400 (33.3% versus 20.0%, P = 0.646). (Table 2) The complication rate of RD was higher in early PPV group than IVA group but the difference was not statistically significant (25.0% versus 0%, P = 0.221). Only 25% of eyes in the early PPV group required additional treatments after 24 hours of presentation to control infection compared to 80% in IVA group (P = 0.030). On average, each eye in the early PPV group required 0.33 additional treatments compared to 0.90 additional treatments in the IVA group (P = 0.018).
Prognostic factors for visual outcome
Multivariate linear regression analysis showed that the risk factors significantly associated with worse final VA in logMAR were causative organism of Streptococcus (β coefficient = 1.92, P = 0.007) and development of RD (β coefficient = 1.72, P = 0.005). (Table 3) Final VA was not associated with age ≥ 65 years, gender, complicated cataract operation, duration of symptoms prior to presentation, presenting VA, significant hypopyon ≥1 mm, absence of fundal view and use of early PPV. The conclusion remained true when early PPV was defined as those performed within 24 hours of presentation (P = 0.225, Table 3), within 48 hours of presentation (P = 0.840, data not shown) or within 72 hours of presentation (P = 0.840, data not shown). The conclusion also remained true when the timing of early PPV was defined with respect to onset of symptoms instead of presentation as performed within 1 day of symptom onset (P = 0.093, data not shown), within 2 days of symptom onset (P = 0.458, data not shown), within 3 days of symptom onset (P = 0.136, data not shown), within 5 days of symptom onset (P = 0.972, data not shown) or within 7 days of symptom onset (P = 0.822, data not shown).
Table 3.
Univariate and multivariate linear regression analyses showing the relationship between risk factors and final visual acuity in logMAR units.
| Univariate regression analysis | Multivariate regression analysis | |||||
|---|---|---|---|---|---|---|
| ß coefficient | 95% CI | P value | ß coefficient | 95% CI | P value | |
| Age ≥ 65 years | −1.02 | −2.19 to 0.15 | 0.084 | −1.46 | −3.18 to 0.27 | 0.087 |
| Gender (reference = female) | 0.22 | −0.64 to 1.09 | 0.594 | −0.43 | −1.40 to 0.54 | 0.335 |
| Complicated cataract operation | 2.04 | 0.19 to 3.89 | 0.032a | 1.11 | −0.81 to 3.03 | 0.220 |
| Symptoms duration prior to presentation (days) | 0.04 | −0.27 to 0.34 | 0.801 | 0.08 | −0.23 to 0.39 | 0.589 |
| Presenting visual acuity (logMAR) | 0.56 | −0.31 to 1.42 | 0.195 | 0.04 | −0.65 to 0.72 | 0.908 |
| Significant hypopyon ≥1 mm | 0.61 | −0.25 to 1.47 | 0.151 | 0.30 | −0.33 to 0.93 | 0.303 |
| Absence of fundal view | 0.37 | −0.77 to 1.52 | 0.501 | −0.84 | −1.91 to 0.23 | 0.107 |
| Early PPV | 0.28 | −0.58 to 1.14 | 0.501 | −0.54 | −1.50 to 0.41 | 0.225 |
| Causative organisms | ||||||
| • No growth (reference) | ||||||
| • Staphylococcus | −0.32 | −1.30 to 0.66 | 0.501 | 0.31 | −0.47 to 1.09 | 0.389 |
| • Streptococcus | 1.33 | 0.05 to 2.61 | 0.043a | 1.92 | 0.68 to 3.15 | 0.007a |
| • Non-Staphylococcus, non-Streptococcus | −0.23 | −1.33 to 0.88 | 0.675 | 1.09 | −0.42 to 2.60 | 0.135 |
| Retinal detachment | 1.56 | 0.53 to 2.59 | 0.005a | 1.72 | 0.68 to 2.76 | 0.005a |
95% CI 95% confidence interval, PPV pars plana vitrectomy.
aStatistically significant.
Discussion
Our study showed that early PPV within 24 hours of presentation required significantly fewer additional treatments to control infection compared to initial intravitreal antibiotics alone. Visual outcome was similar between the two groups. RD was the most frequent complication. Streptococcal infection and development of RD were major poor prognostic factors for vision.
Although PPV had been shown to be effective in treating postoperative endophthalmitis in previous studies [7–11], there is still controversy about the best timing of PPV and whether or not it should be performed immediately after presentation [12, 13]. Early PPV allows rapid removal of infective and inflammatory load in vitreous, thereby reduces further inflammatory damage to the retina [13]. The EVS [14] required immediate vitrectomy to be performed within 6 hours of presentation which may not be feasible for many clinical settings. Therefore, in this study we defined early PPV as those performed within 24 hours of presentation which is more attainable and yet should be early enough to control the intraocular inflammation and retinal damage. In EVRS Endophthalmitis Study [17], early PPV within 7 days of presentation was not found to be associated with better visual outcome compared with intravitreal antibiotics alone. The rates of favourable VA ≥ 20/60 and poor VA ≤ CF were similar between the two groups when early PPV was defined as within 7 days or within 24 hours of presentation [17]. Altan et al. [16] evaluated the effect of immediate PPV within hours of presentation and showed it had a higher rate of VA ≥ 20/40 compared to intravitreal antibiotics (52.9% versus 29.4%) but the difference was not statistically significant. Similarly, our study did not show any significant differences between early PPV and initial intravitreal antibiotics in terms of final VA, proportion of good VA ≥ 20/30 and proportion of poor VA < 20/400. There are several possible reasons to explain this finding: [18] (1) The beneficial effect of early PPV could have been offset by the higher complication rate of RD due to severe intraocular inflammation and poor fundal view at the acute phase of infection; (2) PPV performed at around 24 hours of presentation might already be too late to control the infective and inflammatory damage; and (3) the visual outcome of early PPV could have been biased unfavourably in this study due to large number of eyes having poor VA of LP before receiving early PPV.
The visual outcome of our patients was generally poor compared to those in previous studies. Only 27% of our patients achieved final VA ≥ 20/60. In EVS (year 1990–1994), 53% had VA ≥ 20/40 [14]. In Endophthalmitis Population Study of Western Australia (EPSWA) (year 1980–2000), 66% had VA ≥ 20/60 [19]. In the prospective French Institutional Endophthalmitis Study (FRIENDS) (year 2004–2005), 45% had VA ≥ 20/40 [20]. In the EVRS Endophthalmitis Study (year 2016–2017), 41% had VA ≥ 20/60 [17]. We believe the poor visual outcome of our patients was related to their poor baseline vision and presence of virulent organisms. Around 73% of our patients presented with VA of HM or worse and 64% were caused by bacteria other than coagulase-negative Staphylococcus. Poor baseline vision and virulent organism were known poor prognostic factors for vision [17, 19, 20]. The EVRS Endophthalmitis Study showed that baseline vision of LP had an odd ratio of 12.2 for poor final VA ≤ CF [17]. The EPSWA study showed that organism other than coagulase-negative Staphylococci had an odd ratio of 9.84 for poor final VA < 6/18 [19]. The FRIENDS study showed that organism other than coagulase-negative Staphylococci had an odd ratio of 14.0 for poor VA < 20/100 [20].
Our study showed that early PPV was superior to initial intravitreal antibiotics with a lower rate of requiring additional treatments (25% versus 80%), suggesting early PPV was more effective and faster in controlling the infection. Our finding concurs with the study by Altan et al. [16] which showed that the rate of early re-intervention within 60 hours was lower with initial PPV than initial intravitreal antibiotics (16.2% versus 54.9%).
In developed countries, Gram-positive organisms accounted for around 70–97% cases of postoperative endophthalmitis [13, 21–25]. Streptococcus is the second most common causative organism following Staphylococcus [13]. Streptococcal endophthalmitis often presents with severe intraocular inflammation with hypopyon, dense media opacities and absence of fundal view [26]. Our study showed that Streptococcal infection was an independent prognostic factor for poor vision. Similarly, Kuriyan et al. showed that 75% of Streptococcal endophthalmitis had final VA < 20/400 [26]. Kurniawan et al. showed that 76.7% had final VA < 6/36 where 19.8% was NLP and 24.5% resulted in evisceration/enucleation [18]. The guarded visual prognosis was probably related to the bacteria biofilm-producing properties which withstand phagocytosis by macrophages and action of antibiotics [13]. Therefore, PPV might be a better therapeutic option for Streptococcal infection compared to intravitreal antibiotics. One previous study evaluated the use of early PPV within 48 hours of presentation in Streptococcal endophthalmitis but it did not show a better visual outcome with early PPV [18]. The authors postulated that an even earlier PPV might be necessary as the maximal inflammation occurred as early as 12 hours after bacterial inoculation [18]. Earlier and more aggressive treatment may be necessary for Streptococcal endophthalmitis.
RD was common in postoperative endophthalmitis. The incidence was 8.3% in EVS [27], 5.9% in EVRS Endophthalmitis Study [17], 13% in FRIENDS [28], and 15% in the study by Ho et al. [11]. In our study, the incidence of RD was 13.6% and it was an independent prognostic factor for poor vision. Similarly, Altan et al. showed that RD was significantly associated with poor VA < 20/800 [16]. In EVRS Endophthalmitis Study, RD had an odd ratio of 7.7 for poor VA ≤ CF [17]. In contrast to EVS which showed that the risk of RD was similar between those receiving immediate PPV and intravitreal antibiotics (7.8% versus 9.0%) [27], our study showed that RD was more common after early PPV than intravitreal antibiotics (25% versus 0%). This discrepancy could be explained by the large proportion of virulent organism and poor baseline vision of LP in our early PPV group. RD was known to be associated with organism virulence and baseline VA of LP [27]. In EVS, RD was more common if the endophthalmitis was caused by Gram-positive organism other than coagulase-negative micrococci (23%) and Gram-negative organism (12%) compared to Gram-positive coagulase-negative organism (5%) or no growth (5%) [27]. RD was significantly more common when the presenting VA was LP (16%) than that of HM or better (6%) [27]. Regarding the time for RD development, FRIENDS [20] showed that RD occurred in 56% of cases in the first month, 31% in the second month and 6% in the third month, at a mean time of 47 days after PPV. Similarly, the 2 cases of post-PPV RD in our study occurred at 30 days and 40 days respectively, suggesting it is important to closely monitor patients for RD especially in the first 2 months after PPV.
This study had several limitations. First, the sample size was small. Hence, the statistical power to determine significant differences between the two treatment groups was low with high probability of type II error. Second, due to retrospective design we were not able to evaluate other risk factors which might predict visual outcome such as antibiotics sensitivity of causative organisms, baseline corneal status, systemic comorbidities of diabetes mellitus etc. Third, the type of intravitreal antibiotics and the time of starting steroid were not standardised and therefore the outcome may not reflect the real effect of PPV. Fourth, the duration of symptom before presentation was variable between 0 and 6 days and therefore early PPV performed within 24 hours of presentation might not reflect the true duration of infection onset, making it difficult to determine how promptly PPV should be offered to patients. Fifth, Streptococcal infection was a poor prognostic factor but the infectious organism could not be known until after treatment. Last, we followed EVS recommendations that all endophthalmitis with poor VA of LP were treated with early PPV. Therefore, the result of early PPV could have been biased unfavourably due to selection bias. Despite these limitations, our study evaluated the long-term outcome of early PPV, identified risk factors for poor vision, and described features of RD after PPV for endophthalmitis which had not been well reported in the literature.
Conclusion
In conclusion, Streptococcal infection and development of RD were major poor prognostic factors for vision in acute post-cataract surgery endophthalmitis. Early PPV was superior to initial intravitreal antibiotics alone as it required fewer additional treatments to control infection. Visual outcome was similar between early PPV and initial intravitreal antibiotics alone despite high number of poor presenting VA of LP in the early PPV group which was known to be associated with poorer outcome such as greater RD risk, more severe inflammation and more virulent organisms. Close monitoring of patients for development of RD was necessary after early PPV.
Summary
What was known before
Postoperative endophthalmitis is a serious complication of cataract operation
Intravitreal antibiotics and PPV are effective treatments for postoperative endophthalmitis
What this study adds
Early PPV within 24 hours of presentation is superior to initial intravitreal antibiotics alone as it required fewer additional treatments to control infection
Streptococcal infection and RD are major poor prognostic factors for vision
Author contributions
LPLI had substantial contribution to the conception and design of the work; acquisition and analysis of data; drafting of the paper. HYC, GKHL, ACYM and PPYW had substantial contribution to the acquisition and analysis of data; critical review and revision of the paper. MH, KWK, LJC, MB and ALY had substantial contribution to the conception and design of the work; acquisition of data; critical review and revision of the paper.
Data availability
Data are available on reasonable request. Data are available from the corresponding author.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Carrim ZI, Richardson J, Wykes WN. Incidence and visual outcome of acute endophthalmitis after cataract surgery-the experience of an eye department in Scotland. Br J Ophthalmol. 2009;93:721–5. doi: 10.1136/bjo.2008.147520. [DOI] [PubMed] [Google Scholar]
- 2.Sheng Y, Sun W, Gu Y, Lou J, Liu W. Endophthalmitis after cataract surgery in China, 1995-2009. J Cataract Refract Surg. 2011;37:1715–22. doi: 10.1016/j.jcrs.2011.06.019. [DOI] [PubMed] [Google Scholar]
- 3.Creuzot-Garcher C, Benzenine E, Mariet AS, de Lazzer A, Chiquet C, Bron AM, et al. Incidence of Acute Postoperative Endophthalmitis after Cataract Surgery: A Nationwide Study in France from 2005 to 2014. Ophthalmology. 2016;123:1414–20. doi: 10.1016/j.ophtha.2016.02.019. [DOI] [PubMed] [Google Scholar]
- 4.Inoue T, Uno T, Usui N, Kobayakawa S, Ichihara K, Ohashi Y. Incidence of endophthalmitis and the perioperative practices of cataract surgery in Japan: Japanese Prospective Multicenter Study for Postoperative Endophthalmitis after Cataract Surgery. Jpn J Ophthalmol. 2018;62:24–30. doi: 10.1007/s10384-017-0545-6. [DOI] [PubMed] [Google Scholar]
- 5.Kim SH, Yu MH, Lee JH, Kim SW, Rah SH. Endophthalmitis after Cataract Surgery in Korea: A Nationwide Study Evaluating Incidence and Risk Factors in a Korean Population. Yonsei Med J. 2019;60:467–73. doi: 10.3349/ymj.2019.60.5.467. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Pershing S, Lum F, Hsu S, Kelly S, Chiang MF, Rich WL, 3rd, et al. Endophthalmitis after Cataract Surgery in the United States: A Report from the Intelligent Research in Sight Registry, 2013-2017. Ophthalmology. 2020;127:151–8. doi: 10.1016/j.ophtha.2019.08.026. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Kuhn F, Gini G. Ten years after… are findings of the Endophthalmitis Vitrectomy Study still relevant today? Graefes Arch Clin Exp Ophthalmol. 2005;243:1197–9. doi: 10.1007/s00417-005-0082-8. [DOI] [PubMed] [Google Scholar]
- 8.Tan CS, Wong HK, Yang FP, Lee JJ. Outcome of 23-gauge sutureless transconjunctival vitrectomy for endophthalmitis. Eye (Lond) 2008;22:150–1. doi: 10.1038/sj.eye.6702987. [DOI] [PubMed] [Google Scholar]
- 9.Almanjoumi AM, Combey A, Romanet JP, Chiquet C. 23-gauge transconjunctival sutureless vitrectomy in treatment of post-operative endophthalmitis. Graefes Arch Clin Exp Ophthalmol. 2012;250:1367–71. doi: 10.1007/s00417-012-1926-7. [DOI] [PubMed] [Google Scholar]
- 10.Hsu CM, Chen SC, Wu TT, Sheu SJ. Outcomes of 23-gauge transconjunctival sutureless vitrectomy for acute postoperative endophthalmitis. J Chin Med Assoc. 2017;80:503–7. doi: 10.1016/j.jcma.2017.05.001. [DOI] [PubMed] [Google Scholar]
- 11.Ho IV, Fernandez-Sanz G, Levasseur S, Ting E, Liew G, Playfair J, et al. Early Pars Plana Vitrectomy for Treatment of Acute Infective Endophthalmitis. Asia Pac J Ophthalmol (Philos) 2019;8:3–7. doi: 10.22608/APO.2018414. [DOI] [PubMed] [Google Scholar]
- 12.Peck TJ, Patel SN, Ho AC. Endophthalmitis after cataract surgery: an update on recent advances. Curr Opin Ophthalmol. 2021;32:62–8. doi: 10.1097/ICU.0000000000000727. [DOI] [PubMed] [Google Scholar]
- 13.Clarke B, Williamson TH, Gini G, Gupta B. Management of bacterial postoperative endophthalmitis and the role of vitrectomy. Surv Ophthalmol. 2018;63:677–93.. doi: 10.1016/j.survophthal.2018.02.003. [DOI] [PubMed] [Google Scholar]
- 14.Results of the Endophthalmitis Vitrectomy Study. A randomized trial of immediate vitrectomy and of intravenous antibiotics for the treatment of postoperative bacterial endophthalmitis. Endophthalmitis Vitrectomy Study Group. Arch Ophthalmol. 1995;113:1479–96. doi: 10.1001/archopht.1995.01100120009001. [DOI] [PubMed] [Google Scholar]
- 15.Negretti GS, Chan W, Pavesio C, Muqit MMK. Vitrectomy for endophthalmitis: 5-year study of outcomes and complications. BMJ Open Ophthalmol. 2020;5:e000423. doi: 10.1136/bmjophth-2019-000423. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Altan T, Acar N, Kapran Z, Unver YB, Yurttaser S, Küçüksümer Y, et al. Acute-onset endophthalmitis after cataract surgery: success of initial therapy, visual outcomes, and related factors. Retina. 2009;29:606–12. doi: 10.1097/IAE.0b013e3181953a31. [DOI] [PubMed] [Google Scholar]
- 17.Soliman MK, Gini G, Kuhn F, Parolini B, Ozdek S, Adelman RA, et al. Visual outcome of early vitrectomy and intravitreal antibiotics in acute postsurgical and postintravitreal injection endophthalmitis: European Vitreo-Retinal Society Endophthalmitis Study Report Two. Retina. 2021;41:423–30. doi: 10.1097/IAE.0000000000002856. [DOI] [PubMed] [Google Scholar]
- 18.Kurniawan ED, Rocke JR, Sandhu SS, Allen PJ. Predictors of visual outcome and the role of early vitrectomy in streptococcal endophthalmitis. Clin Exp Ophthalmol. 2018;46:424–31. doi: 10.1111/ceo.13077. [DOI] [PubMed] [Google Scholar]
- 19.Ng JQ, Morlet N, Pearman JW, Constable IJ, McAllister IL, Kennedy CJ, et al. Management and outcomes of postoperative endophthalmitis since the endophthalmitis vitrectomy study: the Endophthalmitis Population Study of Western Australia (EPSWA)‘s fifth report. Ophthalmology. 2005;112:1199–206. doi: 10.1016/j.ophtha.2005.01.050. [DOI] [PubMed] [Google Scholar]
- 20.Combey de Lambert A, Campolmi N, Cornut PL, Aptel F, Creuzot-Garcher C, Chiquet C. Baseline factors predictive of visual prognosis in acute postoperative bacterial endophthalmitis in patients undergoing cataract surgery. JAMA Ophthalmol. 2013;131:1159–66. doi: 10.1001/jamaophthalmol.2013.4242. [DOI] [PubMed] [Google Scholar]
- 21.Han DP, Wisniewski SR, Wilson LA, Barza M, Vine AK, Doft BH, et al. Spectrum and susceptibilities of microbiologic isolates in the Endophthalmitis Vitrectomy Study. Am J Ophthalmol. 1996;122:1–17. doi: 10.1016/S0002-9394(14)71959-2. [DOI] [PubMed] [Google Scholar]
- 22.Lalwani GA, Flynn HW, Jr, Scott IU, Quinn CM, Berrocal AM, Davis JL, et al. Acute-onset endophthalmitis after clear corneal cataract surgery (1996-2005). Clinical features, causative organisms, and visual acuity outcomes. Ophthalmology. 2008;115:473–6. doi: 10.1016/j.ophtha.2007.06.006. [DOI] [PubMed] [Google Scholar]
- 23.Mollan SP, Gao A, Lockwood A, Durrani OM, Butler L. Postcataract endophthalmitis: incidence and microbial isolates in a United Kingdom region from 1996 through 2004. J Cataract Refract Surg. 2007;33:265–8. doi: 10.1016/j.jcrs.2006.10.022. [DOI] [PubMed] [Google Scholar]
- 24.Friling E, Lundström M, Stenevi U, Montan P. Six-year incidence of endophthalmitis after cataract surgery: Swedish national study. J Cataract Refract Surg. 2013;39:15–21. doi: 10.1016/j.jcrs.2012.10.037. [DOI] [PubMed] [Google Scholar]
- 25.Yannuzzi NA, Si N, Relhan N, Kuriyan AE, Albini TA, Berrocal AM, et al. Endophthalmitis After Clear Corneal Cataract Surgery: Outcomes Over Two Decades. Am J Ophthalmol. 2017;174:155–9. doi: 10.1016/j.ajo.2016.11.006. [DOI] [PubMed] [Google Scholar]
- 26.Kuriyan AE, Weiss KD, Flynn HW, Jr, Smiddy WE, Berrocal AM, Albini TA, et al. Endophthalmitis caused by streptococcal species: clinical settings, microbiology, management, and outcomes. Am J Ophthalmol. 2014;157:774–80. doi: 10.1016/j.ajo.2013.12.026. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Doft BM, Kelsey SF, Wisniewski SR. Retinal detachment in the endophthalmitis vitrectomy study. Arch Ophthalmol. 2000;118:1661–5. doi: 10.1001/archopht.118.12.1661. [DOI] [PubMed] [Google Scholar]
- 28.Chiquet C, Aptel F, Combey-de Lambert A, Bron AM, Campolmi N, Palombi K, et al. Occurrence and risk factors for retinal detachment after pars plana vitrectomy in acute postcataract bacterial endophthalmitis. Br J Ophthalmol. 2016;100:1388–9. doi: 10.1136/bjophthalmol-2015-307359. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data are available on reasonable request. Data are available from the corresponding author.