Article Text
Abstract
Aims To determine endophthalmitis rates and risk factors following intraocular surgeries in the Medicare population.
Methods Retrospective, cross-sectional study from 2016 to 2019 in Medicare fee-for-service and Medicare Advantage beneficiaries. 100% Medicare claims were used to identify eyes that underwent intraocular surgery. Exclusion criteria included eyes that had intraocular surgery 42 days before or after and eyes in patients with a history of endophthalmitis within 12 months.
Results 9 744 400 intraocular surgeries were performed among Medicare beneficiaries from 2016 to 2019. A 42-day postoperative endophthalmitis rate was 0.09%. Endophthalmitis rate by surgery type was 0.43% for corneal transplant, 0.36% for secondary intraocular lens (IOL), 0.24% for retina, 0.16% for glaucoma, 0.11% for cataract combined with other procedures and 0.08% for cataract surgeries alone. On multivariable analysis, the risk of endophthalmitis was increased for all intraocular surgery types when compared with cataract surgeries; adjusted OR (aOR) 5.30 (p<0.01) for corneal transplant, aOR 4.50 (p<0.01) for secondary IOL, aOR 3.00 (p<0.01) for retina, aOR 1.93 (p<0.01) for glaucoma, aOR 1.45 (p<0.01) for combined cataract surgeries. Increased risk of developing endophthalmitis was associated with older age (≥85 vs 65–75 years: aOR 1.36; p<0.01), male sex (aOR 1.20; p<0.001) and greater Charlson Comorbidity Index (≥7 vs 0: aOR 1.79; p<0.01).
Conclusions Postoperative endophthalmitis rate after intraocular surgeries was 0.09% between 2016 and 2019 for Medicare beneficiaries. Endophthalmitis rates were highest for corneal transplant surgeries followed by secondary IOL surgeries and lowest for cataract surgeries. Older age, male gender and higher comorbidity index were associated with risk of endophthalmitis.
- Epidemiology
Data availability statement
Data may be obtained from a third party and are not publicly available. Data used for the study cannot be shared publicly due to the rules from the Centers for Medicaid and Medicare Services.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Postoperative endophthalmitis rates after cataract surgeries have been reported from national databases to be between 0.06% and 0.20%. Additional studies have also examined endophthalmitis rates after combined cataract, corneal transplant, secondary intraocular lenses, glaucoma and retina surgeries.
WHAT THIS STUDY ADDS
Of nearly 10 million intraocular surgeries performed among Medicare beneficiaries from 2016 to 2019, the 42-day postoperative endophthalmitis rate was 0.09%. Endophthalmitis rates were highest for corneal transplants followed by secondary intraocular lenses. Patient risk factors include older age, male gender and a higher comorbidity index.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This study provides a better understanding of the epidemiology of postoperative endophthalmitis of intraocular surgeries that can guide future prevention strategies.
Introduction
Although a rare complication of intraocular surgery, postoperative endophthalmitis is a devastating complication. Given the relative rarity of postoperative endophthalmitis, large sample sizes are required to estimate rates and risk factors. Postcataract surgery endophthalmitis rates have been reported by studies to be between 0.06% and 0.20%.1–6 Literature for endophthalmitis following other intraocular surgeries has been reported for combined cataract,1 5 7 8 corneal transplant,1 9 secondary intraocular lenses (IOL),8 10 glaucoma11 12 and retina surgeries.13 14 To our knowledge, there have not been prior reports of endophthalmitis rates after all intraocular surgeries derived from a national source in the USA. Understanding endophthalmitis rates and risk factors is crucial to guide patient counselling and to serve as a foundation for prevention strategies.
Using the Medicare fee-for-service and Medicare Advantage database, the purpose of this study was to determine the rates and characteristics of endophthalmitis after intraocular surgeries and among different intraocular surgery types from 2016 to 2019. Intraocular surgeries examined included cataract, cataract combined with other procedures, corneal transplant, secondary IOL, glaucoma and retina surgeries.
Materials and methods
Study population
This study was a retrospective, cross-sectional analysis of the 2016–2019 100% Medicare fee-for-service and Medicare Advantage claims accessed via the Centers for Medicare and Medicaid Services Virtual Research Data Center.
All patients who underwent at least one intraocular surgery between 1 January 2016 and 19 November 2019 were identified using current procedural terminology (CPT) codes (online supplemental table 1). Eyes were excluded if the laterality of surgery was unspecified (n=1 378 553), had an endophthalmitis diagnosis within 12 months of the intraocular surgery (n=16 736), had any intraocular surgery (online supplemental table 1 and 2) on the same or unspecified eye 42 days before (n=76 143), had an intravitreal injection (CPT 67028) on the same or unspecified eye 42 days before (n=109 640), had any intraocular surgery (online supplemental table 1 and 2 except wound revisions (CPT 66250) and pars plana vitrectomy (PPV) alone (CPT 67036)) on the same eye or unspecified eye 42 days after or until diagnosis of endophthalmitis (n=703 805), or had less than 42 days follow-up after the intraocular surgery (including patients who died) (n=15 244). Additional study exclusion criteria were eyes in patients without 12 months of continuous enrolment in Medicare parts A and B prior to their intraocular surgery (n=451 378), younger than 65 years old (n=527 229) and with missing demographic information (n=2070). Online supplemental figure 1 shows complete exclusion criteria.
Supplemental material
Supplemental material
Additionally, PPV alone (CPT 67036) was not included since during preliminary review of the data, the endophthalmitis rate after PPV alone was abnormally elevated compared with the other PPV codes. We postulated that the endophthalmitis cases treated surgically with PPV alone were not appropriately coded for endophthalmitis until after the surgical date and thus we excluded these cases from the analysis.
Outcomes
Our primary outcome was diagnosis of endophthalmitis within 42 days of intraocular surgery. Endophthalmitis cases were identified using International Classification of Diseases, 10th Revision-Clinical Modification (ICD-10 CM) codes (online supplemental table 3) in the inpatient and outpatient setting. Endophthalmitis with unspecified laterality (H44.009, H44.019, H44.029) were included and if the patient had bilateral intraocular surgery, then endophthalmitis diagnosis was attributed to the most recent surgery. The analysis begins in 2016 since ICD-10 CM codes were first implemented in October 2015 and thus our results are not subject to variability from changes in coding. The analysis ended on 19 November 2019 since data for Medicare Advantage was available until 31 December 2019, 42 days before our study end date.
We calculated the overall postoperative endophthalmitis rate between 2016 and 2019 and by surgery type. Surgery types analysed were cataract, cataract combined with another intraocular surgery, corneal transplant, secondary IOL, glaucoma and retina surgeries. Corneal transplant surgeries included penetrating keratoplasty (PK), endothelial keratoplasty (EK), anterior lamellar keratoplasty (ALK) and keratoprosthesis. Glaucoma surgeries included trabeculectomy, tube shunt surgery and minimally invasive glaucoma surgery (MIGS). Secondary IOL surgeries included IOL insertion alone and IOL exchange. Retina surgeries included PPV with retinal detachment (RD) repair, laser or membrane peel and silicone oil removal.
Patient and surgery characteristics
For each intraocular surgery, we obtained the patients’ demographic information including age, sex and race/ethnicity from the Medicare Master Beneficiary Summary File. The calendar year of the intraocular surgery was derived from the claim date based on the date of surgery. We also reviewed each patient’s inpatient, outpatient, and carrier claims within 12 months before the intraocular surgery date to determine the Charlson Comorbidity Index (CCI). To define a comorbidity, we required at least one diagnosis from the inpatient claims or at least two diagnoses greater than 30 days apart from outpatient and carrier claims as used in prior literature.15
Statistical analysis
Two-sample t-tests were used to compare continuous variables and Pearson’s χ2 tests were used to compare categorical variables. We conducted a multivariable logistic regression model using generalised estimating equations to evaluate factors associated with occurrence of postoperative endophthalmitis. This model was clustered at the patient level and accounted for the correlation of eyes within the same patient for bilateral cases. Patient characteristics included demographic information (age, sex, race and CCI) and surgical characteristics (type of intraocular surgery and calendar year). All statistical analyses were performed by using SAS Enterprise V.7.1 (SAS). Significance level was set at p<0.05.
Results
Study population demographics
A total of 9 744 400 intraocular surgeries were performed among Medicare fee-for-service and Medicare Advantage beneficiaries from 2016 to 2019. A total of 5 557 760 (57.04%) were on eyes for Medicare fee-for-service beneficiaries and 4 186 640 (42.96%) were Medicare Advantage beneficiaries. The mean age of the patient population at the time of surgery was 75.13±6.04 years. Most surgeries were performed on females (59.75%) and whites (83.99%) (table 1). Cataract surgery alone was the most common intraocular surgery (n=8 837 521; 90.70%). A total of 334 830 (3.44%) cases were retina surgeries; 325 032 (3.34%) were cataract surgeries combined with another intraocular surgery; 138 020 (1.41%) were glaucoma surgeries; 73 183 (0.75%) were corneal transplant surgeries and 35 814 (0.37%) were secondary IOL surgeries.
Postoperative endophthalmitis rates
Among Medicare beneficiaries between 2016 and 2019, there were 8655 cases of postoperative endophthalmitis, yielding a 42-day postoperative endophthalmitis rate of 0.09% for intraocular surgery. Intraocular surgeries performed on females had a lower rate of endophthalmitis at 0.08% compared with males at 0.10% (online supplemental table 4). With regard to race/ethnicity, endophthalmitis rates were highest for surgeries on Hispanics (0.11%).
For surgery types, cataract surgeries had the lowest rate of endophthalmitis at 0.08% (table 2). The endophthalmitis rate for combined cataract surgeries was 0.11%. Corneal transplant surgeries, including PK, EK, ALK and keratoprosthesis, had the highest postoperative endophthalmitis rate of 0.43%. The endophthalmitis rate was 0.36% for secondary IOL surgeries, similar between IOL insertion alone (0.39%) and secondary IOL exchange (0.35%) (p=0.58). For all glaucoma surgeries, the endophthalmitis rate was 0.16%. Following trabeculectomies, the rate was 0.16% vs 0.20% for tube shunt surgeries (p=0.19). MIGS alone had an endophthalmitis rate of 0.10% compared with 0.08% for cataract surgeries alone and 0.10% for cataract surgeries combined with MIGS (p<0.001). The endophthalmitis rate after intraocular retina surgeries was 0.24%.
Overall, endophthalmitis was diagnosed at a median of 7 days (IQR 4–14) and mean of 10.35±9.86 days after surgery. Glaucoma surgeries had the latest median day of endophthalmitis diagnosis at 15 days (IQR 6–28).
Risk factors for endophthalmitis
On multivariable analysis (table 3), the risk of endophthalmitis following intraocular surgeries, when compared with cataract surgeries, was increased for corneal transplant surgeries (adjusted OR (aOR) 5.30; 95% CI 4.73 to 5.95), secondary IOL surgeries (aOR 4.50; 95% CI 3.78 to 5.37), retina surgeries (aOR 3.00; 95% CI 2.78 to 3.23), glaucoma surgeries (aOR 1.93; 95% CI 1.68 to 2.20) and cataract surgeries combined with other procedures (aOR 1.45; 95% CI 1.30 to 1.61). Intraocular surgeries for patients aged 75–84 years (aOR 1.09; 95% CI 1.04 to 1.14) and 85 years or older (aOR 1.36; 95% CI 1.26 to 1.46) were more likely to have endophthalmitis compared with patients 65–74 years old. Surgeries on males (aOR 1.20 vs females; 95% CI 1.15 to 1.25) had a higher risk of postoperative endophthalmitis. Intraocular surgeries on patients with a higher CCI also had increased likelihood of endophthalmitis. Intraocular surgeries performed in the year 2019 had a lower risk of endophthalmitis (aOR 0.91; 95% CI 0.85 to 0.97). There was no race with statistically significant increased risk of endophthalmitis when compared with whites.
Discussion
In our analysis of nearly 10 million intraocular surgeries between 2016 and 2019, the 42-day postoperative rate of endophthalmitis was 0.09%. The highest rate of endophthalmitis was among corneal transplant surgeries followed by secondary IOL surgeries. While studies have examined endophthalmitis rates after specific types of intraocular surgeries, particularly for cataract surgeries, to our knowledge, this is the first study to examine endophthalmitis rates after the most common intraocular surgeries using the Medicare database.
Aaberg et al. reported postoperative endophthalmitis following intraocular surgeries at a single institution from 1984 to 1994 and found an endophthalmitis rate of 0.09% within 6 weeks of intraocular surgery with a median postoperative day of diagnosis at 5 days.8 Further stratification in that study found that secondary IOL surgeries had the highest incidence of postoperative endophthalmitis (0.37%) compared with cataract surgeries combined with PK (0.19%), PK (0.18%), glaucoma surgeries (0.12%), cataract surgeries combined with trabeculectomy (0.11%), cataract surgeries (0.08%) and PPV (0.04%).8 The overall endophthalmitis rates in our study were similar; however, our rates of endophthalmitis were higher for PK (1.18%) and intraocular retina (0.24%). Our median postoperative day of diagnosis at 7 days was comparable as well. Another recent study from a national French database from 2009 to 2018 reported overall endophthalmitis rates after intraocular surgeries as 0.07% when not including intravitreal injections and open globe injuries.16
Patient characteristics associated with increased risk of postoperative endophthalmitis included increased age, male sex and more medical comorbidities. Males have consistently been identified at having increased risk of postcataract endophthalmitis.1–3 Medical comorbidity as measured in our study by the CCI is associated with postoperative endophthalmitis. Previous studies have found that diabetes17 18 and immunosuppression1 to be an endophthalmitis risk factor after cataract surgery. Although prior studies have found racial differences for cataract surgery,3 our study did not find significant racial differences for endophthalmitis rates after intraocular surgeries.
We also observed a decrease in overall endophthalmitis rates over from 2016 to 2019. However, 2016 and 2017 had increased rates of endophthalmitis compared with 2016 which were not statistically significant. Given the fluctuating endophthalmitis rates, the decrease in 2019 may only be statistically significant due to the large sample size. Future studies should follow the endophthalmitis trend long-term to determine whether there is a true pattern.
Cataract and combined cataract surgeries
Cataract surgery was the most common intraocular surgery performed but had the lowest rate of endophthalmitis at 0.08%. Previous studies of Medicare beneficiaries have reported similar rates including 0.12% of surgeries from 2003 to 2004 at 42 days,3 0.16% from 1994 to 2006 at 1 year,2 0.06% from 2006 to 2011 at 42 days,1 0.12% from 2010 to 2014 at 3 months4 and 0.14% from 2011 to 2019 at 90 days.6 Another study, using the American Academy of Ophthalmology Intelligent Research in Sight (IRIS) Registry, published postoperative endophthalmitis rates of over 5.5 million cataract surgeries to be 0.056% at 4 weeks after surgery from 2013 to 2018.19 This study identified cases of endophthalmitis from ICD CM codes and confirmed with clinical data with decreased visual acuity or an intraocular tap and inject procedure. The lower endophthalmitis rates from the IRIS Registry compared with our study may be due to our longer postoperative period of 42 days or overestimation from the billing codes since confirmatory clinical data was not available in the Medicare database.
Cataract surgeries are often performed with additional intraocular procedures including concurrent cornea, glaucoma and retina surgeries. Studies have reported that endophthalmitis rates of cataract surgery are higher when combined with another surgery. In an analysis of over 6 million surgeries, Creuzot-Garcher et al. found that the incidence of endophthalmitis increased from 0.102% in cataract surgery alone to 0.149% in cataract surgery combined with cornea, glaucoma and vitreoretinal surgery.7 The IRIS Registry analysis found that rates of endophthalmitis were four times higher in cataract surgeries combined with another intraocular surgery compared with cataract surgery alone.5 For combined cataract and corneal transplants, including PK and EK, postsurgical endophthalmitis rates have been reported to be 0.16% and 0.68% at 6 weeks and 6 months, respectively.1 Our findings of endophthalmitis rates of 0.08% and 0.11%, for cataract surgeries alone and combined with other procedures respectively, are consistent with previously published literature.
Secondary IOL surgeries
For the 35 814 cases of secondary IOL surgeries in this study, the postoperative endophthalmitis rate was 0.36%, with no significant difference (p=0.58) between IOL insertion alone and IOL exchange. Secondary IOL surgeries are performed in complex eyes, often with exposure to vitreous, and involve many different techniques including implantation in the sulcus, iris or scleral fixation, or anterior chamber IOLs. The prior literature on endophthalmitis rates for such procedures is limited. Aaberg et al. reported 5 cases of endophthalmitis after 1367 secondary IOL surgeries (0.37%) over a 10-year period.8 In a retrospective analysis of 119 scleral fixated IOLs, 1 eye (0.80%) developed endophthalmitis 3 months after implantation.10 Two case series of 10 patients from 1995 and 15 patients from 2021 have been published for endophthalmitis related to secondary IOLs.20 21 The series from 1995, identified risk factors of postsecondary IOL endophthalmitis as transscleral suture fixation, polypropylene haptics and preoperative eyelid abnormalities such as blepharitis.20 A limitation of our secondary IOL surgeries analysis includes that the CPT codes do not specify fixation technique and future studies can examine whether scleral fixation increases the risk of postoperative endophthalmitis.
Corneal transplant surgeries
The 42-day postoperative endophthalmitis rate after corneal transplant surgeries was 0.43% which included PK (1.18%), EK (0.19%) and other corneal transplants including ALK and keratoprosthesis (0.15%). Endophthalmitis rates among Medicare beneficiaries have been published by Du et al. as 0.11% at 6 weeks from 2006 to 2011 for PK and EK combined.1 Borkar et al, another study of large academic private practices, found 30-day endophthalmitis rates of 0.7% for PK and 0.2% for EK.9 The lower rates of endophthalmitis for PK may be due to age since the mean age of patients in Borkar et al. was 58.1 years. PK had higher rates of endophthalmitis than EK, as confirmed in our study, likely due to the inherent differences in surgical techniques including the larger circumferential wound and exposure of sutures in PK. Anterior lamellar keratoplasty and keratoprosthesis endophthalmitis rates were not able to be reported individually in our study due to the Centers for Medicare & Medicaid Services Cell Suppression Policy.
Glaucoma surgeries
In our Medicare population, 42-day postoperative endophthalmitis rates after glaucoma surgeries were 0.16%. For trabeculectomy, the endophthalmitis rate was 0.16% with a median onset of 15 days. A large study from a single centre in Sweden found endophthalmitis rates of 0.19% within the first 6 weeks of trabeculectomy with a median onset of 10 days.12 However, the risk of endophthalmitis after glaucoma surgery extends beyond the acute postoperative period. Studies that included delayed onset endophthalmitis rates for trabeculectomies ranged from 0.1% to 0.2% after 1 year.11 22 Risk factors of delayed endophthalmitis after trabeculectomies include inferior blebs and use of antifibrotic agents.12 23
For glaucoma tube shunts, the rate of endophthalmitis in our population was 0.20%, with no significant difference (p=0.19) from trabeculectomies. In the Medicare population from 1994 to 2003, Stein et al. reported an endophthalmitis rate of 0.4% 1 year after glaucoma drainage devices.11 A higher rate of endophthalmitis after tube shunts compared with trabeculectomies was observed by Stein et al. who hypothesised that patients receiving tube shunts may be at higher risk since they are more likely to have failed prior surgeries.11
For MIGS-related endophthalmitis, a recent study at a single institution found that postoperative endophthalmitis rates after MIGS-related procedures were 0.13% (13 out of 3055 surgeries).24 Identifiable risk factors in these 13 cases included intraoperative vitreous loss, anterior chamber tap from high postoperative pressure, and exposure of the implant. In our study, endophthalmitis rates after MIGS alone was 0.10%, after cataract surgeries combined with MIGS was 0.10% and cataract alone was 0.08%.
Retina surgeries
PPV is the primary surgical technique for many vitreoretinal pathologies including RD, vitreous haemorrhage, macular hole and epiretinal membrane. Our study found that the endophthalmitis rate after 334 830 retina surgeries was 0.24%. Recently, Chen et al. performed a meta-analysis of 31 studies which found 199 postvitrectomy endophthalmitis cases out of 363 544 participants (0.05%).13 The endophthalmitis rates after retina surgeries in our Medicare population may be higher due to the age distribution of the population, since older age is an independent risk factor.
Strengths and limitations
The major strength of this study is the large sample size that includes 100% of Medicare fee-for-service and Medicare Advantage beneficiaries. Medicare Advantage accounts for 31%–39% of total Medicare beneficiaries from 2016 to 2020,25 and prior studies which examined endophthalmitis rates in Medicare beneficiaries have not included patients enrolled in Medicare Advantage.1 6 26 Additionally, nearly 94% of US citizens and residents who are 65 years or older are enrolled in Medicare.27
A limitation of this study is that the Medicare database is based on billing records which could have miscoded endophthalmitis and we did not use any confirmatory codes such as use of intravitreal injections. The Medicare database also does not have any clinical data such as visual acuity which could be used to verify endophthalmitis diagnoses. Additionally, we limited our patient population to those 65 years or older. Rates of endophthalmitis may be different for younger patients.
Conclusion
In conclusion, we found that the overall 42-day postoperative endophthalmitis rate after intraocular surgery was 0.09% for the Medicare population between 2016 and 2019. For surgery types, endophthalmitis rates were highest for corneal transplant surgeries (0.43%) followed by secondary IOL (0.36%), retina (0.24%), glaucoma (0.16%), combined cataract (0.11%) and cataract (0.08%) surgeries. Older age, male gender and higher comorbidity index were associated with risk of endophthalmitis. Future research should focus on identifying modifiable surgical strategies to reduce risk.
Data availability statement
Data may be obtained from a third party and are not publicly available. Data used for the study cannot be shared publicly due to the rules from the Centers for Medicaid and Medicare Services.
Ethics statements
Patient consent for publication
Ethics approval
The study was approved by the Institutional Review Board of the Johns Hopkins University School of Medicine and adhered to the Declaration of Helsinki (Reference: IRB00085313).
Acknowledgments
The authors wish to thank Dr. Ravi R Pandit, MD, MPH and Dr. Thomas V Johnson III, MD, PHD for their assistance in reviewing the CPT codes for retina and glaucoma surgeries.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors The authors confirm contribution to the paper as follows - guarantor: FW; study conception and design: AC, CD, ODS, DS, SZ, MM and FW; data collection: CD and MM; analysis and interpretation of results: AC, CD, ODS, DS, SZ and FW; draft manuscriptpreparation: AC, CD, ODS and FW. All authors reviewed the results and approved the final version of the manuscript.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
Linked Articles
- Commentary
- Highlights from this issue