Prophylactic Intracameral Antibiotics and Endophthalmitis After Cataract Surgery

Key Points

Question

What is the association between endophthalmitis after cataract surgery with disinfection in combination with selective antibiotic prophylaxis vs routine prophylactic antibiotic use?

Findings

In this cohort study, among 56 598 cataract surgical procedures performed at the Rotterdam Eye Hospital, postoperative endophthalmitis incidence was comparable when using 1% povidone iodine disinfection with selective antibiotic prophylaxis vs antibiotic prophylaxis use in 37 reported studies.

Meaning

The findings indicate comparable associations between cataract surgery with endophthalmitis after disinfection with 1% povidone iodine and selective antibiotic prophylaxis as was described in the literature after routine intraocular antibiotic use.

Thie cohort study compares postoperative endophthalmitis incidence following selective use of intracameral antibiotics with 1% povidine iodine vs routine use of intracameral antibiotics with 5% povidine iodine.

Abstract

Importance

Although the effectiveness of intracameral antibiotics to prevent postoperative endophthalmitis is described, selective use of antibiotics combined with 1% povidone iodine disinfection might be equally effective and could lead to cost reduction and avoidance of unnecessary use of antibiotics.

Objective

To compare the incidence of postoperative endophthalmitis when 1% povidone iodine disinfection is applied in combination with selective intracameral antibiotics with the incidence after routine use of intracameral antibiotics in combination with 5% povidone iodine.

Design, Setting, and Participant

This was a retrospective cohort study using incidence data from the ongoing endophthalmitis register of the Rotterdam Eye Hospital, a specialized hospital providing both secondary and tertiary ophthalmological care, when intracameral antibiotics were used only during cataract procedures with occurrence of a posterior capsular tear in comparison with results from cohorts described in the literature where routine antibiotics were used. All patients who had cataract (phacoemulsification) surgery at the Rotterdam Eye Hospital between 1993 and 2022 were included. No cataract surgical procedures combined with other intraocular procedures were included.

Exposure

Povidone iodine disinfection and intracameral antibiotics during cataract surgery either routinely or only in case of posterior capsular tears.

Main Outcome and Measure

Postoperative endophthalmitis incidence.

Results

Postoperative endophthalmitis incidence after 56 598 cataract (phacoemulsification) surgical procedures in the Rotterdam Eye Hospital between 2016 and 2022 was 0.0003 (95% CI, 0.0002-0.0004). A PubMed literature search until September 2023 with respect to the incidence of postoperative endophthalmitis after routine antibiotic prophylaxis yielded 37 publications with an overall postoperative endophthalmitis incidence of 0.0003 (95% CI, 0.0003-0.0004).

Conclusions and Relevance

No difference was observed between the postoperative endophthalmitis incidence during the last 7 years in the Rotterdam Eye Hospital and the overall postoperative endophthalmitis incidence after routine intracameral antibiotics prophylaxis as described in the literature. Disinfection with 1% povidone iodine in combination with selective antibiotic prophylaxis may be equally effective as routine antibiotic use and 5% povidone iodine.

Introduction

Endophthalmitis is a serious complication of intraocular surgery that can potentially lead to a substantial loss of vision in the affected eye. There is general consensus that the causative microorganisms leading to endophthalmitis frequently originate from the periocular surfaces and cavities^1,2 and enter the eye during or shortly after surgery.³ Therefore, strict adherence to preventive measures around intraocular procedures is essential.

Infection prevention generally consists of a complex combination of measures, including personnel training, preoperative preventive measures, proper patient selection, disinfection and, in selected subgroups, prophylaxis with antibiotics. To avoid postoperative endophthalmitis after cataract surgery 2 major preventive measures have been recognized through the last decades. First, disinfection with povidone iodine had a significant effect on endophthalmitis incidence in a prospective trial,⁴ and for a long period was considered to be the most relevant prophylactic measure.⁵ After publication of the results of the prospective European Society of Cataract and Refractive Surgeons (ESCRS) multicenter study of the prophylaxis of endophthalmitis after cataract surgery on endophthalmitis prophylaxis in cataract surgery,⁶ much attention has been given to prophylaxis with intraocular antibiotics at the end of surgery. Nevertheless, some concerns were raised about the high incidence of endophthalmitis in the control group of the ESCRS trial.^7,8 Subsequent studies on this subject were usually of a case-control design, where large cohorts of operations performed in the same setting were compared during episodes before and after introduction of an intracameral antibiotic prophylaxis measure. Most of these studies confirmed the prophylactic efficacy of intraocular antibiotics. Disinfection measures during these former investigations (including the ESCRS trial) generally include the application of 5% povidone iodine. Traditionally, much attention is paid in the Rotterdam Eye Hospital (REH) to a careful preoperative disinfection with 1% povidone iodine without routine use of antibiotics. Here, an analysis of the ongoing endophthalmitis registration at the REH is presented and compared with literature on endophthalmitis prevention in cataract surgery.

Methods

From 1993 onward, all postoperative cases of endophthalmitis occurring weeks after intraocular surgery in the REH have been registered. Two input sources are used: all individuals treated for endophthalmitis at the vitreoretinal surgery department at the REH were collected together with data from the microbiology department on vitreous and anterior chamber samples. As a large tertiary center for eye surgery with a 24-hour service, it is improbable that patients with severe postoperative complications attend other hospitals for this purpose. (Personal communication with the head of the Department of Ophthalmology of the nearby tertiary Erasmus University Medical Center confirmed that patients with endophthalmitis following surgery in the REH were not treated there.)

For our analysis, we identified 3 periods with distinct perioperative measures around cataract operations: 1993 to 1999, when 5% povidone iodine and no intraocular, peribulbar, or postoperative antibiotics were applied; 2000 to 2010, when 1% povidone iodine was applied preoperatively on the periocular skin and ocular surface and no prophylactic intraocular, peribulbar, or postoperative antibiotics; and 2016 to 2022, when 1% povidone iodine was applied as before and, in cases of posterior capsular rupture, ceftazidime/vancomycin was injected in the anterior chamber at the end of surgery. To avoid potential bias by suboptimal adherence (caused by an ongoing national debate about the efficacy of antibiotics prophylaxis) after introducing the selective intracameral antibiotic regimen (ceftazidime/vancomycin) in the REH, the period from 2011 to 2015 was excluded from this analysis.

As phacoemulsification was introduced early in the REH, most cataract procedures were performed with this technique in all three periods. All surgical procedures were performed in daycare and the average duration of the procedures did not change substantially over time. Phacoemulsification equipment was provided by Alcon and consisted of the Legacy (up to 2008), Infinity (2008-2017), and Centurion (from 2018) models. Incisions were mainly corneal and changed from 2.7 mm to 2.2 mm in 2016. From 1993 to 1999, mainly polymethyl methacrylate 5-mm intraocular lenses were implanted through a 5-mm scleral incision. After 1999, foldable hydrophobic acrylate implant lenses were used (Acrysof and from 2018 onward Technis ICB00/GIB00) where incisions were mainly corneal and changed from 2.7 mm to 2.2 mm in 2016. Apart from the disinfection measures mentioned above, there were no major changes in air management, operation theater design or attention toward patient related risk factors, such as lacrimal sac infections and blepharitis. No cataract surgical procedures combined with other intraocular surgery or immediately sequential bilateral cataract surgery procedures were included in this study. Vitreous samples of all individuals with endophthalmitis presenting at the REH are obtained at the operating theater after 1% povidone iodine disinfection and with the aid of a vitrectome introduced through a trocar.

A PubMed search was conducted until September 2023 and confirmed in Embase with the following combination of terms: cataract AND intracameral antibiotic AND (prophylactic OR prophylaxis OR prevention). From the results, studies were selected that reported the numbers of surgical procedures and postoperative endophthalmitis for a cohort receiving intracameral antibiotics, as well as for a control cohort that did not. After inspection for and exclusion of overlapping publications, the overall incidences were calculated for the intracameral antibiotics cohorts and for the control cohorts.⁹

As this was a retrospective study, ethical committee approval and informed consent were not necessary. This was confirmed by the review board of the REH.

Results

From 1993 to 1999, 36 cases of endophthalmitis occurred after 27 114 phacoemulsification procedures; from 2000 to 2010, 62 cases were recorded after 68 335 surgical procedures and from 2016 to 2022, 17 endophthalmitis cases were recorded after 56 598 phacoemulsification procedures (Table 1). Only cases occurring within 6 weeks after cataract surgery were included. The difference in incidence between the periods from 1993 to 1999 and from 2000 to 2010 was χ², 0.0004 (95% CI, 0.000X-0.001X; P = .07), the difference between the periods from 2000 to 2010 and from 2016 to 2022 was χ², 0.0006 (95% CI, 0.0003-0.0009; P < .001). Among the 17 endophthalmitis cases from 2016 to 2022, cultures yielded no growth in 5 cases, coagulase negative staphylococci in 8 cases and a single case each of Staphylococcus aureus, Streptococcus pneumoniae, Streptococcus mitis, and Enterococcus faecalis. Fifteen cases occurred within 1 week after the cataract surgery and the remaining 2 within 2 postoperative weeks. All but 1 of the 17 endophthalmitis cases were uncomplicated surgical procedures where the intraocular lens was inserted in the capsular bag. The single complicated case needed anterior vitrectomy because of a posterior capsular tear with vitreous loss, and it was the only case in the endophthalmitis group where prophylactic antibiotics were applied; culture of the subsequent vitreous sample did not yield any bacterial growth.

Table 1. Incidence of Endophthalmitis After Cataract Surgery in the Rotterdam Eye Hospital.

	1993-1999	2000-2010	2016-2022
Incidence (95% CI)	0.001 (0.001-0.002)	0.001 (0.001-0.002)	0.0003 (0.0002-0.0004)

The initial literature search yielded 281 hits. After selection, 37 studies remained (Table 2) for analysis and comparison with the incidence in the REH.^{6,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45} In the Figure, the incidences of both cohorts of each study are shown, together with their corresponding 95% CIs. At the bottom of the Figure, the overall incidence calculated from the literature studies is shown for intracameral antibiotics vs no intracameral antibiotics, and the incidence in the REH from 2016 to 2022.

Table 2. Summary of Included Reference Studies.

Source	Study design	No.	Antibiotics	Disinfection
Montan et al,10 2002	Historic cohort as comparator	66 282	Cefuroxime	Chlorhexidine
Endophthalmitis Study Group,6 2007	Randomized clinical trial	16 211	Cefuroxime	5% Povidone iodine
Lundstrom et al,11 2007	Parallel cohort	225 471	Cefuroxime	Chlorhexidine
Yu-Wai-Man et al,12 2008	Historic cohort as comparator	36 743	Cefuroxime	5% Povidone iodine
Anijeet et al,13 2010	Historic cohort as comparator	16 606	Vancomycin	5% Povidone iodine
Arshinoff and Bastianelli14 2011	Parallel cohort	112 536	Cefuroxime/moxifloxacin/vancomycin	Unclear
Romero-Aroca et al,15 2012	Historic cohort as comparator	25 001	Cefazolin	5% Povidone iodine
Barreau et al,16 2012	Historic cohort as comparator	5115	Cefuroxime	5% Povidone iodine
Tan et al,17 2012	Historic cohort as comparator	50 177	Cefazolin	5% Povidone iodine
van der Merwe et al,18 2012	Parallel cohort	8190	Cefuroxime	2.5% Povidone iodine
Rodriguez-Caravaca et al,19 2013	Historic cohort as comparator	19 463	Cefuroxime	5% Povidone iodine
Myneni et al,20 2013	Historic cohort as comparator	25 296	Cefuroxime	5% Povidone iodine
Matsuura et al,21 2013	Historic cohort as comparator	34 752	Moxifloxacin	Unclear
Friling et al,22 2013	Parallel cohort	464 755	Cefuroxime	Chlorhexidine
Beselga et al,23 2014	Historic cohort as comparator	15 689	Cefuroxime	5% Povidone iodine
Rudnisky et al,24 2014	Parallel cohort	75 318	Moxifloxacin/vancomycin	Unclear
Röck et al,25 2014	Historic cohort as comparator	31 752	Cefuroxime	5% Povidone iodine
Sharma et al,26 2015	Historic cohort as comparator	15 122	Cefuroxime	5% Povidone iodine
Rahman and Murphy,27 2015	Historic cohort as comparator	14 043	Cefuroxime	5% Povidone iodine
Katz et al,28 2015	Historic cohort as comparator	56 094	Cefuroxime	Unclear
Rush et al,29 2015	Historic cohort as comparator	20 719	Vancomycin	5% Povidone iodine
Daien et al,30 2016	Parallel cohort	2 432 067	Cefuroxime	Unclear
Au et al,31 2016	Historic cohort as comparator	14 805	Vancomycin	5% Povidone iodine
Ng et al,32 2016	Historic cohort as comparator	30 428	Cefuroxime	5% Povidone iodine
Herrinton et al,33 2017	Parallel cohort	300 950	Cefuroxime/moxifloxacin	Unclear
Vieira et al,34 2017	Historic cohort as comparator	7195	Moxifloxacin	5% Povidone iodine
Haripriya et al,35 2019	Historic cohort as comparator	2 062 643	Moxifloxacin	5% Povidone iodine
Li et al,36 2019	Historic cohort as comparator	32 526	Cefuroxime/moxifloxacin	Unclear
Moser et al,37 2019	Historic cohort as comparator	29 275	Cefazolin	5% Povidone iodine
Melega et al,38 2019	Randomized clinical trial	3640	Moxifloxacin	5% Povidone iodine
Rathi et al,39 2020	Historic cohort as comparator	92 010	Cefuroxime/moxifloxacin	5% Povidone iodine
Ma et al,40 2020	Historic cohort as comparator	61 299	Cefuroxime	5% Povidone iodine
Shenoy et al,41 2021	Historic cohort as comparator	214 782	Moxifloxacin	5% Povidone iodine
Bhatta et al,42 2021	Historic cohort as comparator	111 983	Moxifloxacin	5% Povidone iodine
Dave et al,43 2021	Historic cohort as comparator	66 967	Moxifloxacin	5% Povidone iodine
Porwal et al,44 2021	Historic cohort as comparator	40 392	Moxifloxacin	Unclear
da Silva Conci et al,45 2022	Historic cohort as comparator	16 902	Cefuroxime	5% Povidone iodine

Figure. Incidence of Endophthalmitis After Cataract Surgery With and Without Intracameral Antibiotics.

Blue indicates with antibiotics; orange, without, as reported in 37 studies, and in the Rotterdam Eye Hospital from 2016 to 2022. Postoperative endophthalmitis incidence in the Rotterdam Eye Hospital without standard prophylactic use of intracameral antibiotics did not differ from the overall mean incidence of 37 studies in which prophylactic intracameral antibiotics were used. In Rathi et al,³⁹ reported incidence and study period were used to estimate totals of surgical procedures and postoperative endophthalmitis cases for the no antibiotic group assuming that number of surgical procedures per month was constant. For analytical purposes, in Beselga et al,²³ Rush et al,²⁹ and Ng et al,³² the incidence of 0 in the antibiotics group was replaced by 0.5.

Discussion

In this cohort study, between 2016 and 2022, when we conducted disinfection with 1% povidone iodine before cataract surgery and administered prophylactic intracameral antibiotics in complicated surgical procedures only, postoperative endophthalmitis incidence was 0.0003. This did not differ from the overall incidence calculated for the intracameral antibiotics cohorts of the 37 studies from literature (0.0003). The reliability of the reported postoperative endophthalmitis incidences of these studies was not always described in detail and, thus it cannot be excluded that some incidences were underestimated.

The introduction over time of additional preventive measures in the REH would potentially constitute a confounding factor of this study when comparing the 3 episodes. However, the only other substantial change was the replacement of HEPA filters by lamellar flow in our daycare center (ie, 2 of the 6 operating theaters) in 2013, which made no noticeable difference. As the REH has always had a very active infection prevention committee, we are confident that its postoperative endophthalmitis registry system is sound and that hardly any endophthalmitis case was missed. The reduction of postoperative endophthalmitis incidence we achieved (from 0.001 to 0.0003) by modifying the perioperative regimen between the periods from 1993 to 1999 and from 2016 to 2022 is comparable with the overall effect computed from the reports of 37 studies by the introduction of intracameral antibiotics (from 0.001 to 0.0003). Results similar to ours by means of changing perioperative disinfection over time, but without intracameral antibiotics, were reported before.⁴⁶ The difference in postoperative endophthalmitis incidence we detected between the period from 2000 and 2010 and the period from 2016 to 2022 (Table 2) is an indication that intracameral antibiotics were associated with effective results in case of a posterior capsular tear during phacoemulsification surgery.

After the first randomized clinical trial on the effect of prophylactic antibiotics (cefuroxime),⁶ only 1 more randomized clinical trial has been performed (moxifloxacin).³⁸ All other studies were cohort studies, 7 of which had a parallel control group and 28 of which had a historic control group. Most of the 37 studies confirm the prophylactic effect of intracameral antibiotics on postoperative endophthalmitis after cataract surgery (Figure). In most studies, either intracameral cefuroxime or moxifloxacin was administered.

Preoperative disinfection of the periorbital skin, eyelashes, and conjunctival sac with 5% povidone iodine was used in 25 studies and probably also in another 3 studies; 1 study used 2.5% povidone iodine, 3 studies used chlorhexidine, and 5 studies did not report their disinfection method. In this respect it seems noteworthy that the postoperative endophthalmitis incidence in the cohort without antibiotics of some studies was rather high (see Figure), which was also the case for the 2 randomized clinical trials. The frequent occurrence of staphylococci (20% to 66%) in endophthalmitis cases in studies providing details on causative strains^{10,11,20,47,48,49,50} appears to indicate a suboptimal preoperative disinfection of the periocular surface, which may have led to an unnecessary high incidence of postoperative endophthalmitis in some control cohorts.

More effective methods for preoperative disinfection with diluted povidone iodine have been described previously.^51,52 A study on the effect of intracameral moxifloxacin in 2 million surgical procedures demonstrated a significant effect of moxifloxacin on the occurrence of coagulase negative staphylococci, but not on S aureus in culture-positive postoperative endophthalmitis cases.³⁵ In our current study, where 1% povidone iodine was applied during the 7 years of the study period and in which 56 598 cataract surgical procedures were performed, S aureus was detected in only 1 endophthalmitis case; in 13 of the 17 cases, either no microorganism or a coagulase negative staphylococci was cultured. This is an indication that most virulent microorganisms were eradicated with our disinfection method, and the occurrence of only 1 staphylococcal endophthalmitis in this large cohort can be regarded as an indicator for effective skin disinfection. These clinical findings are in agreement with laboratory studies indicating a substantially enhanced antibacterial effect of diluted povidone iodine solutions.^53,54

Previously, results have been presented from experiments subjecting some endophthalmitis strains to lower povidone iodine concentrations which showed a reduced susceptibility.⁵⁵However, it is questionable whether the bacterial load used in such experiments adequately represents the exposure of the interior eye during perioperative contamination, and to what extent the organic substances present in such in vitro experiments inactivate the lower povidone iodine concentrations.⁵⁶

When intravitreal injections for macular degeneration were introduced in our center, we managed to cut down an initial rise in endophthalmitis rate by means of standardization of the antiseptic protocol with 1% povidone iodine (unpublished data). Others have also advocated that endophthalmitis prophylaxis for intravitreal injections should focus on iodine disinfection rather than topical antibiotics.^57,58 A comprehensive review of povidone iodine in cataract surgery that discusses the effect of different concentrations and dosing for disinfection also appears to support the use of diluted povidone iodine.⁵⁹ Indications that disinfection measures with lower concentrations of povidone iodine may contribute to the avoidance of postoperative endophthalmitis are (1) theoretical considerations and experimental research,^53,54 (2) previous clinical results from the REH,⁶⁰ (3) results of investigations with diluted povidone iodine in Japan,^61,62,63 and (4) the results of this study.

Due to the no nocere concept of medical interventions, (potential) adverse reactions should always be addressed. The dosage of intraocular antibiotics used in endophthalmitis prophylaxis is such that levels reached in the eye are well above the minimum inhibitory concentration of the targeted microorganisms and, therefore, adverse reactions are not hypothetical. In their review article on the relative efficacy of 3 intracameral antibiotics, Bowen et al⁵⁰ also evaluate toxicity. Toxicity affecting visual acuity was detected in nearly 3% of 503 cases analyzed for cefuroxime, mostly attributed to local dilution errors. For moxifloxacin, such adverse events were only detected in patients with a penetrating corneal graft, and for vancomycin, in a single case series of a severe condition known as hemorrhagic occlusive retinal vasculitis. In addition, anaphylactic reactions to cefuroxime have been described in at least 4 patients,⁶⁴ which may only reflect a small proportion of all cases occurring, considering the vast numbers of patients exposed to the compound and also because cataract registers do not keep track of this adverse event. On the other hand, after repeated application of povidone iodine during cataract surgery, no effect on corneal endothelial cell density was detected⁵² and the experimental data strongly suggest that endothelial or retinal damage after topical application of povidone iodine is improbable.⁶⁵

Questions on widespread routine use of antibiotics in cataract surgery have been raised before,⁷ 1 of the arguments being the accelerated inducement of bacterial resistance. The detection of prophylaxis escaping strains leading to endophthalmitis after cataract procedures with intracameral cefuroxime has been reported by several authors.^6,11,66,67 Therefore, eradicating pathogens with an effective disinfectant before surgery appears to be a more rational approach than letting them enter the eye and subsequently applying potentially deleterious or only partially effective prophylactic antibiotics. Only in cases of complicated cataract surgery known to be associated with an increased risk of postoperative endophthalmitis^6,35 would antibiotic prophylaxis be required.

Limitations

The main limitation of this study design is that it can only show associations, not causation. Furthermore, its retrospective design has a potential for selection and information bias. However, this is also the case for the retrospective studies of 35 of the studies we used for the comparison, and the 2 randomized trials^6,38 from the literature showed a relatively high endophthalmitis incidence. The REH is highly specialized and the largest provider of secondary and tertiary eye care in the Netherlands, but not an academic hospital. The patient group, level of training, internal protocols, and facilities, might not be fully comparable to general hospitals nor academic centers.

Conclusions

In conclusion, the findings indicate comparable outcomes with antibiotic prophylaxis limited to complicated cataract surgery. Such indication may be preferred above routinely exposing large numbers of patients undergoing the most frequently performed surgical procedure in the world to prophylactic antibiotics. The REH data show that a low level of postoperative endophthalmitis was associated with 1% povidone iodine disinfection in combination with selective prophylactic antibiotic use and that postoperative endophthalmitis incidence did not substantially differ from routine antibiotic use. Although the routine use of prophylactic antibiotics in modern cataract surgery may still be a matter of ongoing debate, preoperative antiseptic preparation of the surgical field with diluted povidone iodine supports that discussion.

Supplement.

Data sharing statement