Infectious endophthalmitis is one of the most feared complications of intraocular surgery with a high risk of severe visual loss, oftentimes despite timely diagnosis and treatment.1,2 Although commonly reported after anterior segment surgeries, infectious endophthalmitis is a serious complication after pars plana vitrectomy (PPV). The rate of postsurgical endophthalmitis over the years has reduced. In the late 1800s, the incidence of endophthalmitis after cataract surgery was as high as 10%, which decreased to 0.04% in 2000.3–5
Although postcataract surgery endophthalmitis is well known, there is very little information about endophthalmitis after PPV. As most vitrectomies in the modern era are performed via a sutureless transconjunctival approach, these cases have unsutured sclerotomies that close with tissue healing and the valvular effect of wound construction. A proposed etiology for endophthalmitis after sutureless vitrectomy could be postoperative hypotony, leading to a possible ingress of bacteria from the ocular surface.
Despite studies comparing the outcome of endophthalmitis in both 20G vitrectomy and the sutureless group, none of these proved potential risk factors and most found an equivocal difference between the 2 groups in terms of incidence.6–9 There are very few large case series of endophthalmitis after PPV,10 with none from the Indian subcontinent (PubMed search). In this study, we looked at in-house data of vitrectomies performed (both 20G and sutureless), the clinical presentations, risk factors, and outcomes of management of post-PPV endophthalmitis in our tertiary eye care center in South India.
MATERIALS AND METHODS
This was a retrospective, multicenter, observational study of all cases that underwent PPV across the 4 tertiary centers of our institute spread over 3 states (Telangana, Andhra Pradesh, and Odisha) in southern India from January 1990 to December 2014.
Cases that underwent primary vitrectomy were identified by the surgical coding assigned by the institute’s medical records department. Institutional review board approval for the study was taken as appropriate. Of these cases, those that underwent surgery for endophthalmitis, trauma, and those where simultaneous cataract surgery was performed were excluded. As per institute protocol, all cases were examined before surgery to rule out coexistent risk factors for postoperative infection like ocular surface disorders, dry eye, nasolacrimal duct obstruction, immunocompromised states, and uncontrolled diabetes mellitus. Wherever found, the particular risk factor was controlled first before surgery. Data collected included age, sex, indication for vitrectomy, time interval between primary surgery and the diagnosis of endophthalmitis, type of incision, experience of the operating surgeon, type of tamponade, microbiologic analysis, and final treatment outcome.
Clinically, endophthalmitis was diagnosed based on a combination of signs and symptoms observed from day 1 after surgery, which included decreased vision, pain, redness, circumcorneal congestion, corneal edema, anterior chamber cells, hypopyon, anterior chamber fibrin, and vitritis.
As per institute protocol, the surgical management of endophthalmitis consisted of vitreous lavage, microscopy and culture of undiluted vitreous, antimicrobial susceptibility testing of bacterial isolates, and intravitreal antibiotics [vancomycin (1 mg/0.01 mL) plus ceftazidime (2.25 mg/0.01 mL)] with or without dexamethasone (400 μg/0.01 mL). The medical treatment also included intensive topical antibiotics (ciprofloxacin 0.3% 1 hourly) and corticosteroid (prednisolone acetate 1% 1 hourly) and oral ciprofloxacin 750 mg 2 times per day for 7 to 10 days. Additional procedures such as repeat intravitreal antibiotics or repeat pars plana vitreous lavage mostly depended on the response to treatment and were left to the decision of the treating physicians.
All 20G procedures were performed using a standard 3-port technique using either the Accurus or Constellation vitrectomy system (Alcon Laboratories, Fort Worth, TX). Conjunctival peritomy was performed in the superotemporal, inferotemporal, and superonasal quadrants. Local cautery was done to achieve hemostasis. Twenty-gauge sclerotomies were created and infusion cannula tied in the inferotemporal quadrant. At the end of surgery, both the sclerotomies and the conjunctiva were sutured with 7–0 polyglactin interrupted sutures. The sutureless surgeries were performed using a 23G or 25G trocar cannula set on the Constellation vitrectomy system (Alcon Laboratories). The cannulas were inserted transconjunctivally by means of beveled trocars. The trocars were then removed. Topical povidone-iodine eye drops were instilled in the cul-de-sac in all cases at the end of surgery.
The data were arranged on an Excel spreadsheet. Relevant statistical analysis was done using MedCalc version 188.8.131.52 for statistical analysis. Vision was converted from Snellen to logMAR equivalents. Means and standard deviations were computed for all continuous variables. In case of nonparametric distribution, the median was calculated. Preoperative and postoperative data were compared using paired t test for parametric data and Wilcoxon rank-sum test for nonparametric data. Odds ratios were computed for the possible risk factors for endophthalmitis with appropriate confidence intervals (CIs). A P value of less than 0.05 was assigned as statistically significant.
A total of 38,591 cases of PPV satisfied the inclusion criteria, and 20 cases were diagnosed as endophthalmitis clinically. Thus, the incidence of clinical endophthalmitis was 0.052%. There were 13 men and 7 women (P = 0.11). The mean age at presentation was 44.4 ± 14.64 years (median, 45 years; range, 8–67 years). Nine cases occurred in the right eye and 11 in the left eye (P = 0.75). The mean presenting vision at the time of diagnosing endophthalmitis was 2.16 ± 1.51 logMAR (Snellen equivalent 20/2890); median, 1.82 logMAR (Snellen equivalent 20/1321). The best previous recorded vision was mean 1.18 ± 1.05 logMAR (Snellen equivalent 20/302); median, 0.89 (Snellen equivalent 20/155). Table 1 lists the detailed clinical features of all cases. Three cases had 20G sclerotomies and were sutured, 15 cases had 23G sutureless vitrectomy, and 2 cases had 25G sutureless vitrectomy (P < 0.0001). In all, 14 cases were operated on by retina faculty and 6 cases by retina fellows-in-training (P = 0.02). Four cases had silicone oil tamponade, 3 cases had gas tamponade, and 13 cases had Ringer lactate (RL). The mean time between PPV and the diagnosis of endophthalmitis was 4 ± 6.89 days; median, 1.5 days. The mean posttreatment vision at the last follow-up (mean, 6.72 ± 9.56 months; median, 3.5 months) was recorded as 1.7 ± 1.36 logMAR (Snellen equivalent 20/1002); median, 1.26 logMAR. Six eyes retained reading vision, and phthisis bulbi developed in 4 eyes. Eleven (55%) cases were culture positive; 5 grew Staphylococcus species, 4 Acinetobacter baumannii, 2 Leclercia adecarboxylata, and 1 Haemophilus influenzae.
Of the total 38,591 cases of PPV, 31,481 cases were sutured (20G) and 7110 cases were sutureless (25G or 23G). The number of endophthalmitis cases after PPV in the 2 groups were 3 (0.009%) and 17 (0.24%), respectively. The difference was statistically significant (P < 0.0001) (95% CI for the difference, 0.13%–0.37%). Bivariate analysis was done to substantiate the effect of sutureless surgery and RL tamponade on the outcome of endophthalmitis (Tables 2 and 3). The Kaplan–Meier survival analysis curve depicts overall survival trends from presentation to resolution of endophthalmitis (Fig. 1).
The incidence of post-PPV clinical endophthalmitis was 0.052%, and culture-proven endophthalmitis was 0.031%. We have previously reported the incidence of acute endophthalmitis after cataract surgery in our centers.11 This was 0.13% for clinical endophthalmitis and 0.07% for culture-proven endophthalmitis. Thus, the incidences of both clinical and culture-proven endophthalmitis in the current study were less than those after cataract surgery. Our observed rate of postvitrectomy endophthalmitis is similar to the one reported in the UK nationwide surveillance study at 0.058%.12
Various risk factors have been proposed to be responsible for endophthalmitis after PPV. One important factor is inadequate wound closure in sutureless PPV surgeries due to a propensity for wound leak and subsequent intraocular ingress of surface microorganisms.13–16 Compared with sutured PPV, a relatively higher risk of endophthalmitis in sutureless PPV has been previously reported.17–19 However, later studies have demonstrated no significant difference in endophthalmitis in the 2 groups.7,9,20 In our series, 17 of 20 cases were sutureless and 3 were 20G sutured surgeries; this difference was statistically significant (P < 0.001). The balanced salt solution is supposedly a poor tamponading agent compared with air, gas, or silicone oil due to different surface tension properties. This was proposed by Chiang et al21 in their study. Gupta et al22,23 also proposed that an eye filled with air or gas is more likely to avoid hypotony in the postoperative period. Higher surface tension across the oil, gas, or air interface may provide better protection from leakage. Studies using anterior segment optical coherence tomography have shown that gas- or air-filled eyes have a quicker wound closure and possible lesser risk for endophthalmitis.24
In our study, the results (as depicted in Tables 1–3) show that the risk of developing endophthalmitis after PPV in sutureless surgery was 25 times that in sutured surgery. In the subset where the final tamponade was RL, the risk of endophthalmitis was 19.5 times greater in sutureless surgery as compared with sutured surgery (Table 4). In sutureless surgeries, the risk of endophthalmitis was 3.5 times higher if the final tamponade was RL as opposed to oil/gas/air. In the largest published study, 16 of 18 cases were culture positive, and coagulase-negative Staphylococci was the commonest isolated organism.25 Recently, Shi et al26 in China reported a postvitrectomy endophthalmitis rate of 0.05% over the decade from 2002 to 2012. As shown in Table 1, case numbers 1 and 2 occurred at a close interval in the same operating room and showed the same organism (L. adecarboxylata) on culture. Similarly, case numbers 3, 4, 5, and 6 also occurred within a short duration and had the same organism on culture (A. baumannii). This could raise the possibility of cluster endophthalmitis in both situations. A detailed and exhaustive microbiologic evaluation, however, did not reveal any source of contamination in either situation.
In our series, the commonest organisms were Staphylococci (n = 5) and Acenitobacter (n = 4). Although theoretically vision improved with appropriate repeated intravitreal antibiotic injections, the overall visual improvement from a preoperative mean logMAR of 1.82 to a postoperative logMAR of 1.26 was not statistically or clinically significant (P = 0.31). Cataract was the commonest secondary complication seen in 6 cases, and 4 eyes developed phthisis. This result was no different than observations in other centers.10 Our study does have inherent limitations. As the event occurrence of interest is very rare, the numbers in the case arm are very low. Various limitations of a retrospective study such as a heterogeneous indication for surgery, varied age of presentation, and different operating surgeons and conditions were present which could not be accounted for.
In conclusion, endophthalmitis after PPV is a rare and potentially serious clinical condition. Despite prompt and appropriate treatment, the visual and anatomic outcome is often poor. As the event is relatively rare, it is difficult to study large numbers in one center to get a correct estimation of the potential risk factors. As per the outcome in our series and past literature, it may be intuitive that sutureless PPV surgeries could have higher odds of developing endophthalmitis and hence need a meticulous sclerotomy preparation. We also propose that the risk of endophthalmitis can be reduced by opting for a non-RL tamponade (opting for either air or gas, and silicone oil when absolutely necessary) at the end of sutureless vitrectomies. In cases where there is no choice of tamponade other than RL, one could consider suturing the ports to reduce the risk of endophthalmitis. An analysis of pooled data from various centers around the world might help us substantiate these results.
1. Lemley CA, Han DP. Endophthalmitis
: a review of current evaluation and management. Retina
. 2007; 27: 662–680.
Vitrectomy Study Group. Results of Endophthalmitis
Vitrectomy Study. A randomized trial of immediate vitrectomy and of intravenous antibiotics for the treatment of postoperative bacterial endophthalmitis
. Arch Ophthalmol
. 1995; 113: 1479–1496.
3. Kattan HM, Flynn HW Jr, Pflugfelder SC, et al. Nosocomia endophthalmitis
survey: current incidence of infection after intraocular surgery. Ophthalmology
. 1991; 98: 227–238.
4. Axenfeld T. The Bacteriology of the Eye
. [Translated by A. Macnab]. London, England: Bailliere, Tindall & Cox; 1908: 77–107.
5. Eifrig CWG, Flynn HW Jr, Scott IU, et al. Acute-onset postoperative endophthalmitis
: review of incidence and visual outcome (1995–2001). Ophthalmic Surg Lasers
. 2002; 33: 373–378.
6. Shaikh S, Ho S, Richmond PP, et al. Untoward outcomes in 25-gauge versus 20-gauge vitreoretinal surgery. Retina
. 2007; 27: 1048–1053.
7. Hu AYH, Bourges JL, Shah SP, et al. Endophthalmitis
after pars plana vitrectomy
. 2009; 116: 1360–1365.
8. Parolini B, Romanelli F, Prigione G, et al. Incidence of endophthalmitis
in a large series of 23-gauge and 20-gauge transconjunctival pars plana vitrectomy
. Graefes Arch Clin Exp Ophthalmol
. 2009; 247: 895–898.
9. Scott IU, Flynn HW Jr, Acar N, et al. Incidence of endophthalmitis
after 20-gauge vs 23-gauge vs 25-gauge pars plana vitrectomy
. Graefes Arch Clin Exp Ophthalmol
. 2011; 249: 377–380.
10. Dave VP, Pathengay A, Schwartz SG, et al. Endophthalmitis
following pars plana vitrectomy
: a literature review of incidence, causative organisms, and treatment outcomes. Clin Ophthalmol
. 2014; 8: 2183–2188.
11. Das T, Hussain A, Naduvilath T, et al. Case control analyses of acute endophthalmitis
after cataract surgery in South India associated with technique, patient care and socioeconomic status. J Ophthalmol
. 2012; 7: 1360–1365.
12. Park JC, Ramasamy B, Shaw S, et al. A prospective and nationwide study investigating endophthalmitis
following pars plana vitrectomy
: incidence and risk factors
. Br J Ophthalmol
. 2014; 94: 529–533.
13. Taylor SR, Aylward GW. Endophthalmitis
following 25-gauge vitrectomy. Eye (Lond)
. 2005; 19: 1228–1229.
14. Aylward GW. Sutureless
. 2011; 225: 67–75.
15. Fujii GY, De Juan E, Humayun MS, et al. Initial experience using the transconjunctival sutureless
vitrectomy system for vitreoretinal surgery. Ophthalmology
. 2002; 109: 1814–1820.
16. Acar N, Kapran Z, Unver Y, et al. Early postoperative hypotony after 25-gauge sutureless
vitrectomy with straight incisions. Retina
. 2008; 28: 545–552.
17. Scott IU, Flynn HW Jr, Dev S, et al. Endophthalmitis
after 25 gauge and 20 gauge pars plana vitrectomy
: incidence and outcomes. Retina
. 2008; 28: 138–142.
18. Chen JK, Khurana RN, Nguyen QD, et al. The incidence of endophthalmitis
following transconjunctival sutureless
25 vs 20 gauge vitrectomy. Eye (Lond)
. 2009; 23: 780–784.
19. Kunimoto DY, Kaiser RS, Wills Retina Service. Incidence of endophthalmitis
after 20 and 25 gauge vitrectomy. Ophthalmology
. 2007; 114: 2133–2137.
20. Shimada H, Nakashizuka H, Hattori T, et al. Incidence of endophthalmitis
after 20 and 25 gauge vitrectomy: causes and prevention. Ophthalmology
. 2008; 115: 2215–2220.
21. Chiang A, Kaiser R, Avery R, et al. Endophthalmitis
in microincision vitrectomy: outcomes of gas-filled eyes. Retina
. 2011; 31: 1513–1517.
22. Gupta OP, Weichel ED, Regillo CD, et al. Postoperative complications associated with 25-gauge pars plana vitrectomy
. Ophthalmic Surg Lasers Imaging
. 2007; 38: 270–275.
23. Gupta OP, Ho AC, Kaiser PK, et al. Short-term outcomes of 23-gauge pars plana vitrectomy
. Am J Ophthalmol
. 2008; 146: 193–197.
24. Yamane S, Kadonosono K, Inoue M, et al. Effect of intravitreal gas tamponade for sutureless
vitrectomy wounds: three dimensional corneal and anterior segment optical coherence tomography study. Retina
. 2011; 31: 702–706.
25. Cohen SM, Flynn HW Jr, Murray TG, et al. Endophthalmitis
after pars plana vitrectomy
. The Postvitrectomy Endophthalmitis
Study Group. Ophthalmology
. 1995; 102: 702–712.
26. Shi XY, Zhao HS, Wei WB. Analysis of post-operative endophthalmitis
after pars plana vitrectomy
: a 10-year experience at a single center. Chin Med J (Engl)
. 2013; 126: 2890–2893.