Racial/Ethnic Differences in Rates of Penetrating or Endothelial Keratoplasty for Fuchs Endothelial Corneal Dystrophy Among US Medicare Beneficiaries

Importance  Fuchs endothelial corneal dystrophy (FECD) is the most common indication for corneal transplant in the United States. The association between race/ethnicity and incidence of advanced FECD, defined by a need for endothelial or penetrating keratoplasty, has not been investigated.

Observations  The 2014 US Medicare Limited Data Set (5% sample of 27 163 740 fee-for-service Medicare patients) was analyzed for rate of keratoplasty performed for FECD (International Classification of Diseases, Ninth Edition code 371.57), stratified by race/ethnicity. Among all Medicare beneficiaries 65 years or older, a diagnosis code for FECD was used in 1.55% (95% CI, 1.51%-1.59%) of white and 1.38% (95% CI, 1.26%-1.50%) of African American beneficiaries who had an ophthalmologist eye examination in 2014 (P = .01). Among beneficiaries who obtained medical care for FECD, keratoplasty was 1.9 times more likely in white than African American patients (4.7%; 95% CI, 4.2%-5.2% vs 2.5%; 95% CI, 1.1%-3.9%; P < .001) among approximately 6500 patients undergoing 8420 procedures.

Conclusions and Relevance  In 2014, keratoplasty was 1.9 times more likely in US Medicare fee-for-service white patients than African American patients with FECD. This might be caused by racial/ethnic differences in the biology of FECD, access to care, or other unidentified factors.

As the most common cause for endothelial and penetrating keratoplasty in the United States, Fuchs endothelial corneal dystrophy (FECD) is likely to have varied phenotypic expressions.^1,2 National claims research shows that African American patients are as likely or more likely to have a diagnosis of FECD relative to white patients.³ A regional study also showed that African American patients and white patients receive diagnoses of guttae at similar rates, but white patients are more likely to have more severe disease in the form of confluent guttae.^2,4 Previous attempts to genetically characterize FECD in African American individuals have failed to discern a genetic association relative to a broader population.⁵ Our aim was to use a national cross-sectional Medicare data set to test for racial/ethnic differences in the rates of penetrating or endothelial keratoplasty in FECD.

The Mayo Clinic Institutional Review Board approved this study. Data were derived from the 2014 Medicare Limited Data Set Carrier Standard Analytical File. These data contain a deidentified 5% sample of all claims for approximately 27 163 740 US Medicare fee-for-service enrollees after filtering to include age 65 years or older and race/ethnicity of either African American or white. Informed consent was not required for this retrospective use of a deidentified Medicare database. Excluded from this data set are the approximately 28% of Medicare patients enrolled in non–fee-for-service Medicare Advantage plans. Analysis was performed using Excel 2013 software with PowerPivot and Power Query add-ins (Microsoft) during March 2016 and April 2016. After filtering by provider specialty code, we tabulated the number of patients who received any eye care from an ophthalmologist at least once during 2014. By filtering the primary diagnosis code for FECD (International Classification of Diseases, Ninth Revision 371.57), we determined the number of patients having an FECD diagnosis in the context of any medical care in 2014. Finally, we summed the number of distinct patients who received at least 1 penetrating (Current Procedural Terminology codes 65730, 65750, and 65755) or endothelial (Current Procedural Terminology code 65756) keratoplasties for an FECD diagnosis specifically. All data were stratified by race/ethnicity (white vs African American), which is self-reported by Medicare patients at the time of Social Security enrollment using form SS-5. Percentages were used as a summary for each group, and 95% CIs were calculated for the percentages within each group. Overall comparisons between groups were completed using χ² tests.

In 2014, approximately 24 817 940 white and 2 345 800 African American patients 65 years or older were enrolled in Medicare fee-for-service plans (Table 1). Of these, 34.9% (95% CI, 34.8%-35.0%) of white patients (n = 8 661 420) and 29.6% (95% CI, 29.4%-29.8%; P < .001) of African American patients (n = 694 740) received medical care from an ophthalmologist at least once during that year (Table 2). A diagnosis of FECD as defined by International Classification of Diseases, Ninth Revision diagnosis code 371.57, paired with medical claims of any type, was made for an estimated 134 580 white patients and 9560 African American patients, representing 1.55% (95% CI, 1.51%-1.59%) of white and 1.38% (95% CI, 1.26%-1.50%) of African American Medicare beneficiaries who had an eye examination by an ophthalmologist in 2014 (P = .01). Among these patients who obtained medical care for FECD, a penetrating or endothelial keratoplasty was 1.9 times more likely in white patients (4.7%; 95% CI, 4.2%-5.2%) than African American patients (2.5%; 95% CI, 1.1%-3.9%; P < .001). Across all categories, African American patients were more likely to be women (Table 3; P < .001). Among all corneal grafts for FECD, 95% were endothelial keratoplasties (95% CI, 93%-97%; P < .001) and 5% were penetrating keratoplasties (95% CI, 2.6%-7.4%; P < .001), representing approximately 8420 procedures in 6500 patients.

The study findings suggest that in white fee-for-service Medicare patients, the rate of Medicare claims-based diagnoses of FECD is slightly higher, but clinically similar, to the rate in African American patients after correcting for the generally lower overall rate at which African American Medicare patients saw ophthalmologists. The 1.9-fold higher rate of endothelial or penetrating keratoplasties in white patients than African American patients suggests that racial/ethnic differences may be present in the manifestation or treatment, or both, of more severe FECD. Whether this difference represents underlying genetic influence vs environmental influences, access to care, or other unknown or confounding factors merits further investigation.

This study represents, to our knowledge, the largest cross-sectional analysis of racial/ethnic differences in rate of keratoplasties for FECD. The nature of our Current Procedural Terminology and diagnosis code pairing may tend to underestimate the total number of keratoplasties in eyes with FECD because some procedures may have been coded with other, less specific diagnostic codes, such as corneal edema or pseudophakic bullous keratopathy. We do not expect this factor to influence relative racial/ethnic differences in the reported corneal graft rates because it is less plausible that race-based differences in physician coding behavior took place. Although disparities in the availability of ophthalmologic surgical care may potentially explain some of this observation, underlying racial/ethnic differences in the biology of FECD may be present. Fuchs endothelial corneal dystrophy is a genetically heterogeneous phenotype. Racial/ethnic differences in the genetic basis of FECD are known to occur, with trinucleotide repeat expansion in the transcription factor 4 gene as the genetic variant found in most US and European populations but in only a small percentage of Asian and Indian populations.^6-11 Furthermore, rates of disease progression and severity may reflect the associated genetic variant.

Because FECD is known to also be present in patients younger than 65 years, our Medicare study has an age-based sampling bias. General limitations of Medicare claims data in evaluating disease frequency and expression are well known. Potential confounding factors include access to care, coding behavior, potential coding errors, and lack of further clinical detail or data in the Medicare claims data. According to the 2010 US Census data, 8.5% of the US population older than 65 years was African Amercian,¹² which is consistent with the 8.6% Medicare proportion, suggesting that there is not a race/ethnicity-based difference in Medicare enrollment rates. At birth in 2014, US African American individuals had a life expectancy that was 3.4 years less than that for white individuals,¹³ which introduces the possibility of a race-based survivor bias in the Medicare data set for persons 65 years or older.

Future additional insights and geographic or sex data stratifications may be possible by expanding this work to encompass more patients over a longer period or by soliciting Medicare 100% Research Identifiable File data sets, which are more difficult to obtain. Elucidation of the genetic variants associated with FECD in US white individuals vs African American individuals may also shed light on the pathogenic mechanism of this genetically heterogeneous trait.

Corresponding Author: Michael A. Mahr, MD, Department of Ophthalmology, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (mahr.michael@mayo.edu).

Accepted for Publication: June 15, 2016.

Published Online: August 11, 2016. doi:10.1001/jamaophthalmol.2016.2735

Author Contributions: Dr Mahr had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Mahr, Erie.

Acquisition, analysis, or interpretation of data: Mahr, Baratz, Hodge.

Critical revision of the manuscript for important intellectual content: Mahr, Baratz, Hodge, Erie.

Statistical analysis: Mahr, Hodge.

Administrative, technical, or material support: Mahr, Baratz.

Study supervision: Mahr, Erie.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.

Funding/Support: This study was made possible using the resources of the Mayo Clinic, Rochester, Minnesota; National Institutes of Health grant EY25071 (Dr Baratz); and Research to Prevent Blindness Inc.

Role of the Funder/Sponsor: The sponsors or funding organizations had no role in the design or conduct of this study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.