Abstract
Objective
Incidence of systemic lupus erythematosus (SLE) is disproportionately high in non-Whites vs. Whites. However, variation in mortality according to race/ethnicity has not been well studied. We examined all-cause mortality by race/ethnicity among SLE patients in Medicaid.
Methods
Within the Medicaid Analytic eXtract 2000–2006 from 47 U.S> states and D.C., we identified individuals aged 18–65 years, enrolled for ≥3 months, with ≥ 3 claims for SLE (ICD-9 710.0), each ≥ 30 days apart. Lupus nephritis (LN) was identified by ≥ 2 additional claims for glomerulonephritis, proteinuria, or renal failure. We calculated mortality rates (MR) per 1,000 person-years with 95% confidence intervals by race/ethnicity. Multivariable Cox regression models estimated mortality risks, adjusting for age, sex, demographics and comorbidities.
Results
Among 42,221 prevalent SLE patients, 8,191 had LN. Blacks represented 40.1%, Whites 38.4%, and Hispanics 15.3%. Overall SLE MRs per 1000 person-years were highest among Native Americans (27.52), Whites (20.17) and Blacks (24.13), and lower among Hispanic (7.12) or Asian SLE patients (5.18). After multivariable adjustment, Hispanic and Asian patients had lower mortality risks [HR 0.48 (95%CI 0.40–0.59) and 0.59 (95%CI 0.40-0.86)] vs. Whites. Conversely, risks for death were significantly higher among Native American (HR 1.40, 95%CI 1.04–1.90) and Black (HR 1.21, 95%CI 1.10–1.33) compared to White patients. Among LN patients, mortality risks were lower among Hispanic and Asian patients (by 56% and 40%) than among Whites.
Conclusion
After accounting for demographic and clinical factors, Asian and Hispanic SLE Medicaid patients had lower mortality than did Blacks, Whites or Native American patients.
Keywords: Systemic Lupus Erythematosus, Race, Ethnicity, Disparities, Mortality, Survival, Medicaid
Systemic lupus erythematosus (SLE) disproportionately affects non-White populations in the United States1,2. Black females, for example, have incidence rates three to four times higher than their White counterparts. Other racial and ethnic minorities, including Hispanics, Asians and Native Americans, are also at increased risk of developing SLE3. Past studies have reported poor outcomes including high rates of lupus nephritis (LN), end-stage renal disease, and SLE organ damage accrual among Blacks, Hispanics and Asians with SLE4–9. Academic cohort studies have suggested higher mortality among Black and Hispanic patients than among White patients with SLE10–13. For example, in the LUMINA study (Lupus in Minorities: Nature versus Nurture), both Blacks and Texan Hispanics with SLE had a lower five year survival than did non-Hispanic Whites in an unadjusted analysis, but after adjustment for age, socioeconomic status, disease activity and organ damage, differences in survival were no longer present14. Previous studies have been mainly based in academic centers with relatively few deaths. Moreover, race and socioeconomic status have historically been very hard to disentangle in their associations with poor outcomes in these SLE populations3,15,16.
Given the lack of studies with large populations of low income individuals affected by SLE, we investigated all-cause mortality and mortality rates, overall and by race and ethnicity, among SLE and LN patients enrolled in U.S. Medicaid from 2000 to 2006. We hypothesized that there would be significant variation in mortality rates and risks according to race and ethnicity among SLE and LN patients, with increased mortality among Hispanic and Black patients.
Patients and Methods
Study population
Medicaid is the U.S. health insurance program for individuals with low income and resources (low income children, pregnant women, mothers and people with disabilities) and provides coverage for medical expenses and prescription drugs. We employed data from the Medicaid Analytic eXtract (MAX) database, an administrative data system containing all billing claims for Medicaid enrollees in 47 U.S. states and Washington, D.C., from January 1, 2000 to December 31, 2006. (Arizona, Tennessee, and Maine do not contribute data to MAX.) We identified adults ages 18–65 years who were enrolled in Medicaid for at least three months between January 1, 2000 and December 31, 2006. The index date for SLE diagnosis was defined as occurring when individuals obtained ≥ 3 ICD-9 codes for SLE (710.0) at least 30 days apart, obtained from hospital discharge diagnoses or physician visit claims1. Among individuals with SLE, we identified those with LN, as having ≥ 2 additional ICD-9 hospital discharge diagnoses or physician billing claims for nephritis, proteinuria and/or renal failure, on or after the date of SLE diagnosis, and occurring at least 30 days apart1,17. This administrative definition has been found to have a 80% positive predictive value for LN in Medicaid claims data17.
Exposures
Race and ethnicity in MAX are categorized based mainly on self-report. We used the following previously defined combined race and ethnicity categories: White, Black, Hispanic or Latino, Asian (including Native Hawaiian or other Pacific Islander), Native American (including American Indian or Alaskan Native)18. Patients with missing or unclassifiable information for race or ethnicity (e.g. “other/unknown” or “more than one race” categories) were excluded from these analyses.
Other variables
Demographic variables extracted from MAX included age, sex, and region of residence, which was determined by zip code and categorized by U.S. Census region (Northeast, Midwest, South or West). For area-based socioeconomic status (SES), we employed a composite index of seven SES indicators at the zip code level using 2000 U.S. Census data19. These include median household income, proportion with income below 200% of the federal poverty level, median home value, median monthly rent, mean education level, proportion of people age ≥25 years who were college graduates, and proportion of employed persons with a professional occupation. Area-level SES was divided into quartiles as previously described1. We also employed a previously described “SLE-specific risk adjustment index”, which has been validated for prediction of in-hospital mortality among SLE patients20. The SLE-specific risk adjustment index developed by Ward includes comorbidities specific for SLE, including autoimmune hemolytic anemia, thrombocytopenia, pericarditis, seizures and psychosis. We also included validated ICD-9 codes for hypertension (401.1), diabetes mellitus (250.0), smoking (305.1), obesity (278.0), acute myocardial infarction (410.0), angina (413.× or 411.1), old myocardial infarction (412.0), percutaneous coronary intervention (00.66, 36.0×, 37.22, 37.23, and 88.5×, except 88.59), coronary atherosclerosis (414.00 and 414.9; not including 414.1x), and coronary artery bypass graft (3610 and 3619)21, 22, 23, 24, 25.
Outcomes
The outcome of our study was death from all causes. Subjects were followed from the index date through date of death, loss to follow-up (no further medical claims in the absence of documented death), or end of follow-up period of the study (December 31, 2006). Deaths were confirmed using the National Death Index.
Statistical analysis
We calculated crude annual mortality rate (MR) per 1,000 person-years with 95% confidence intervals for SLE patients by racial/ethnic group. We fit three Cox regression models to examine the association of race/ethnicity with mortality risk for both SLE and LN. Model A included age (continuous) and sex. Model B added potential confounding variables to model A, including residential region, calendar year, area-SES, baseline comorbidities collected from ICD-9 diagnoses from January 1, 2000 though the study index date (including history of angina, coronary artery bypass graft, coronary atherosclerosis, percutaneous coronary intervention, hypertension, smoking, obesity) and SLE specific risk-adjustment index. Finally, model C included model A, residential region, calendar year, area-SES, but comorbidities and the SLE-specific risk-adjustment index were included as time-varying covariates throughout the entire follow-up period. In addition, in model C, acute myocardial infarction at any time during follow-up was included as a comorbidity. We tested the proportional hazards assumption using Kaplan Meier curves, as well as time-varying covariates and observed no significant deviations in our models. In sensitivity analyses, we repeated models A, B and C using multivariable subdistribution proportional hazards models, with loss to follow-up from Medicaid as a competing event to investigate the hypothesis that loss to follow-up could account for some of the observed variation in mortality risks26.
All statistical analysis were conducted using SAS, version 9.3. Data were obtained from the Centers for Medicare and Medicaid Services through a data use agreement and are presented in accordance with CMS policies. The Partners Healthcare Institutional Review Board waived human subjects approval for this study.
Results
We identified 42,221 patients with prevalent SLE and, among them, 8,191 patients with prevalent LN. Baseline characteristics are summarized in Table 1. The mean age among SLE patients was 38.1 ± 12.3 years, and 93% were women, and the majority of the cohort resided in the South (38%). Black SLE patients resided predominantly in the South, had a higher prevalence of hypertension and a higher SLE-specific baseline risk-adjustment index compared to White patients. Hispanic patients resided predominantly in the West, and their prevalence of diabetes mellitus was similar to that in the Black population. The baseline prevalence of hypertension, heart failure, smoking, obesity were lower, as was the SLE-specific risk adjustment index, among Hispanics than among Blacks. Asian patients lived predominantly in the West, and had a higher prevalence of LN, but a lower prevalence of heart failure, diabetes mellitus, obesity and smoking and a lower SLE-specific risk adjustment index than Whites. Native Americans had a higher prevalence of diabetes mellitus, obesity and smoking than all other races/ethnicities, and had an SLE risk adjustment index higher than that of Whites, Asians or Hispanics. Finally, White patients had a higher prevalence of previous angina, coronary atherosclerosis, but lower prevalence of LN compared to other races/ethnicities.
Table 1.
Characteristics at Index Date of 42,221 Medicaid Recipients with SLE (2000–2006)
| All patients n (%) | White n (%) | Black n (%) | Hispanic n (%) | Asian n (%) | Native American n (%) | |
|---|---|---|---|---|---|---|
| Total number of patients | 42,221 | 16,219 | 16,956 | 6,489 | 1,880 | 677 |
| Female | 39,320 (93.13) | 15,040 (92.73) | 15,859 (93.53) | 6,058 (93.36) | 1,740 (92.55) | 623 (92.02) |
| Age, years (mean, SD) | 38.13 (12.29) | 40.23 (12.26) | 36.91 (11.86) | 36.32 (12.40) | 36.99 (13.23) | 38.66 (12.27) |
| Residential region | ||||||
| Midwest | 8,324 (19.72) | 4,011 (24.73) | 3,582 (21.13) | 466 (7.18) | 97 (5.16) | 168 (24.82) |
| Northeast | 8,296 (19.65) | 3,149 (19.42) | 2,937 (17.32) | 1,903(29.33) | 235 (12.50) | 72 (10.64) |
| South | 16,028 (37.96) | 5,363 (33.07) | 8,621 (50.84) | 1,715 (26.43) | 154 (8.19) | 175 (25.85) |
| West | 9,573 (22.67) | 3,696 (22.79) | 1,816 (10.71) | 2,405 (37.06) | 1,394 (74.15) | 262 (38.70) |
| Comorbidities* | ||||||
| Previous Angina | 5,116 (12.12) | 2,174 (13.40) | 1,965 (11.59) | 714 (11.00) | 197 (10.48) | 66 (9.75) |
| Previous CABG | 441 (1.04) | 142 (0.88) | 185 (1.09) | 77 (1.19) | 30 (1.60) | – |
| Previous CVA | 1,568 (3.71) | 547 (3.37) | 758 (4.47) | 182 (2.80) | 56 (2.98) | 25 (3.69) |
| Coronary atherosclerosis | 6,412 (15.19) | 2,737 (16.88) | 2,483 (14.64) | 886 (13.65) | 233 (12.39) | 73 (10.78) |
| Previous MI | 1327 (3.14) | 423 (2.61) | 505 (2.98) | 118 (1.82) | 24 (1.28) | 20 (2.95) |
| Previous PCI | 94 (0.22) | 660 (4.07) | 18 (0.11) | 16 (0.25) | – | – |
| Diabetes mellitus | 10,264 (24.31) | 3,792 (23.38) | 4,237 (24.99) | 1,623 (25.01) | 414 (22.02) | 198 (29.25) |
| Heart failure | 3,827 (9.06) | 1,299 (8.01) | 1,990 (11.74) | 386 (5.95) | 96 (5.11) | 56 (8.27) |
| Hypertension | 13,081 (30.98) | 4,644 (28.63) | 6,215 (36.65) | 1,616 (24.90) | 476 (25.32) | 130 (19.20) |
| Obesity | 5,857 (13.87) | 2,346 (14.46) | 2,467 (14.55) | 803 (12.37) | 138 (7.34) | 103 (15.21) |
| Smoking | 6,579 (15.58) | 3,646 (22.48) | 2,226 (13.13) | 440 (6.78) | 88 (4.68) | 179 (26.44) |
| Lupus nephritis | 8,191 (19.40) | 1,988 (4.71) | 4,099 (9.71) | 1,418 (3.36) | 545 (1.29) | 141 (0.33) |
| SLE risk adjustment index¥ (mean, SD) | 3.65 (4.19) | 3.52 (4.28) | 3.93 (4.22) | 3.31 (3.92) | 3.32 (3.86) | 3.81 (4.19) |
CABG: coronary artery bypass graft, CVA: cerebrovascular accident, MI: myocardial infarction, PCI: percutaneous coronary intervention
Comorbidities collected at any time during the follow-up, cell sizes under 11 suppressed per Centers for Medicare and Medicaid policy
SLE specific index ranges from 0–46.
Baseline demographic and clinical characteristics at index date of patients with and without LN are summarized in Supplementary Tables 1A and 1B. Compared to the prevalent SLE cohort, LN patients were younger, with a mean age of 34.5 ± 12.6 years, and had a higher prevalence of coronary heart disease (including angina, coronary atherosclerosis and previous MI), hypertension and diabetes. Furthermore, the SLE-specific risk adjustment index was almost two-fold higher in the LN patients compared to the SLE cohort.
The mean duration of follow-up for all SLE patients was 2.56 ± 1.99 years and in LN patients was 2.12 ± 1.77 years. During the follow up period, there were 2,058 deaths among all SLE patients and 774 deaths among LN patients. The overall unadjusted annual MR among SLE patients was 19.07 per 1,000 person-years (95%CI 18.36–19.91), while the MR among LN patients was more than two-fold higher 44.64 (95%CI 41.60–47.90) per 1,000 person-years (Table 2). Among SLE and LN patients, all-cause MRs were significantly lower and approximately one-third as high among Hispanic patients and one-quarter as high among Asian patients compared to Whites. Unadjusted MRs were highest among Native American and Black patients among SLE patients compared to all other race/ethnicities (MR ratios 1.36 and 1.19 respectively compared to Whites) (Table 2).
Table 2.
Crude Incidence Rates for Death among SLE and LN patients, stratified by Race/Ethnicity
| Adult SLE Patients, n= 42,221 | |||||
|---|---|---|---|---|---|
| Number of Patients | Number of Events | Person-years, mean (SD) | MR* (95% CI) | MR Ratio (95%CI) | |
| All patients | 42,221 | 2,058 | 2.56 (1.99) | 19.07 (18.36–19.91) | |
| White | 16,219 | 824 | 2.41 (1.97) | 20.17 (18.84–21.60) | 1.0 (Ref) |
| Black | 16,959 | 1,040 | 2.54 (1.98) | 24.13 (22.71–25.64) | 1.19 (1.09–1.31) |
| Hispanic | 6,489 | 119 | 2.58 (2.00) | 7.12 (5.95–8.52) | 0.35 (0.29–0.43) |
| Asian | 1,880 | 29 | 2.98 (2.15) | 5.18 (3.60–7.45) | 0.26 (0.18–0.37) |
| Native American | 677 | 46 | 2.47 (1.94) | 27.52 (20.61–36.74) | 1.36 (1.01–1.83) |
| Adult Lupus Nephritis Patients, n= 8,191 | |||||
|---|---|---|---|---|---|
| Number of Patients | Number of Events | Person-years, mean (SD) | MR* (95% CI) | MR Ratio (95%CI) | |
| All patients | 8,191 | 774 | 2.12(1.77) | 44.64 (41.60- 47.90) | |
| White | 1,988 | 230 | 2.16(1.80) | 53.49 (47.01–60.87) | 1.0 (Ref.) |
| Black | 4,099 | 461 | 2.05(1.72) | 54.95 (50.16–60.20) | 1.03 (0.88–1.20) |
| Hispanic | 1,418 | 52 | 2.12(1.76) | 17.33 (13.21–22.74) | 0.32 (0.24–0.44) |
| Asian | 545 | 18 | 2.46(1.95) | 13.45 (8.47–21.35) | 0.25 (0.16–0.41) |
| Native American | 141 | 13 | 2.19(1.72) | 42.05 (24.42–72.42) | 0.79 (0.45–1.37) |
MR = annual mortality rate per 1,000 person years, SD= Standard deviation
In multivariable-adjusted Cox regression models, racial/ethnic variation in survival persisted. The HRs for all-cause mortality by race/ethnicity for SLE and LN patients are presented in Table 3. In age and sex-adjusted models (model A), compared to White SLE patients, both Hispanic and Asian SLE patients had less than half the risk of death (HRs 0.41 (95%CI 0.34–0.50) and 0.30 (95%CI 0.21–43). This was also true among Hispanic and Asian LN patients [HRs 0.39 (95CI% 0.29–0.52) and 0.31 (95%CI 0.19–0.50)] compared to Whites. Conversely, Black SLE patients had a significantly increased mortality risk compared to White patients (HR 1.36 (95%CI 1.24–1.49), as did Native Americans (HR 1.43 95%CI 1.06–1.92). Black and Native American LN patients did not have statistically elevated HRs for all-cause mortality compared to Whites however, possibly due to smaller sample size of the LN patients.
Table 3.
Hazard Ratios (HR) for Death in Adult SLE patients and Lupus Nephritis patients
| SLE, n= 42,221 | Lupus Nephritis, n=8,191 | |||||
|---|---|---|---|---|---|---|
| Race/Ethnicity | Model A (HR, 95%CI) |
Model B* (HR, 95%CI) |
Model C** (HR, 95%CI) |
Model A (HR, 95%CI) |
Model B* (HR, 95%CI) |
Model C** (HR, 95%CI) |
| White | 1.0 (Ref) | 1.0 (Ref) | 1.0 (Ref) | 1.0 (Ref) | 1.0 (Ref) | 1.0 (Ref) |
| Black | 1.36 (1.24–1.49) | 1.25 (1.13–1.37) | 1.21 (1.10–1.33) | 1.18 (1.00–1.38) | 1.13 (0.95–1.34) | 1.04 (0.88–1.23) |
| Hispanic | 0.41 (0.34–0.50) | 0.50 (0.41–0.61) | 0.48 (0.40–0.59) | 0.39 (0.29–0.52) | 0.48 (0.35–0.65) | 0.44 (0.32–0.59) |
| Asian | 0.30 (0.21–0.43) | 0.62 (0.42–0.90) | 0.59 (0.40–0.86) | 0.31 (0.19–0.50) | 0.68 (0.41–1.13) | 0.60 (0.37–1.00) |
| Native American | 1.43 (1.06–1.92) | 1.51 (1.12–2.04) | 1.40 (1.04–1.90) | 0.85 (0.49–1.49) | 0.86 (0.49–1.52) | 0.85 (0.48–1.50) |
Model A: age (continuous) and sex
Model B: Model A + region of residence, calendar year, area-level socioeconomic status, comorbidities at study index date including history of angina, coronary artery bypass graft, coronary atherosclerosis, percutaneous coronary intervention, hypertension, smoking and obesity and SLE-specific index at study index date
Model C: Model A + region of residence, calendar year, area-level socioeconomic status, comorbidities at any point including previous history of angina, coronary artery bypass graft, coronary atherosclerosis, percutaneous coronary intervention, acute myocardial infarction, hypertension, smoking and obesity and SLE-specific index through last follow-up
Adjustment for baseline cardiovascular and SLE–related comorbidities (model B) and for these comorbidities throughout follow-up (model C), did attenuate racial/ethnic variation somewhat. In particular, the risk of death among Blacks with SLE decreased to 21% above that in Whites (HR 1.21 95%CI 1.01–1.33), and, among Blacks with LN, it was no longer statistically elevated when compared to that among Whites with LN (HR 1.04, 95% CI 0.88–1.23). We did not observe a large difference in the HRs when only baseline comorbidities were included compared to when covariates were updated throughout the follow-up period. The SLE-specific risk adjustment index increased in all groups from baseline through follow-up and a one unit increase in the risk index was itself associated with increased mortality risk in all racial/ethnic groups (model C HR 1.15, 95%CI 1.14–1.16 among all SLE patients). Increasing area-based SES (as a continuous variable) was inversely associated overall mortality HR 0.92 (95%CI 0.90, 0.96) among all SLE patients. In multivariable-adjusted subdistribution hazards models taking the competing risk of loss to follow-up into account, racial/ethnic variation in survival persisted and the results were extremely similar (data not shown).
Discussion
In the present study with more than 40,000 adult SLE patients enrolled in Medicaid between 2000–2006, we found marked variation in all-cause mortality rates by race/ethnicity. Not unexpectedly, mortality rates were higher among Black patients, although adjustment for comorbidities and sociodemographic factors did attenuate this risk to 21% higher than that of Whites. Strikingly, Native Americans had 40% higher adjusted mortality risks than did Whites with SLE, and adjustment for comorbidities and sociodemographic factors did not substantially affect the risk estimates. The relationships of race/ethnicity to socioeconomic status, lifestyle factors, and comorbidities, such as obesity, smoking, and diabetes, all of which are related to SLE outcomes as well, are extremely complex. These factors are highly correlated and are likely mediators of health outcomes, as well as confounders of observed relationships between race/ethnicity and outcomes. Mortality rates among Hispanic and Asian patients with SLE were lower by 52% and 41% respectively compared to White adults, even after adjusting multiple demographic and clinical factors. To the best of our knowledge, this is the first large-population study demonstrating lower mortality rates among Hispanic and Asian patients with SLE than among Black and White patients.
Previous studies in North American academic centers have reported worse prognoses, including higher mortality rates, among Black and Hispanic patients compared to Whites4,12,14,27. Our finding of decreased mortality rates and adjusted risks among Hispanic adults with SLE was thus surprising14,28. There may be several potential explanations. First, as Medicaid provides health care coverage for low-income populations in the U.S., the SES divide between Hispanic and non-Hispanic patients was likely smaller than in past academic center-based cohort studies. Second, this cohort included all Medicaid patients meeting our administrative definition of SLE, whereas academic medical centers, in particular referral centers, may tend to capture the most severe SLE cases with the worst outcomes. Third, our finding is not unique to SLE. Several past epidemiologic studies of the U.S. Hispanic population have demonstrated that, after adjusting for age and annual family income, Hispanics have a lower all-cause mortality and lower mortality due to cardiovascular disease compared with non-Hispanics, despite having a higher burden of cardiovascular risk factors29. This epidemiologic observation, first termed the “Hispanic paradox” two decades ago, has since been demonstrated in vital statistics, nationally representative surveys, systematic reviews and a recent meta-analysis30,31. To our knowledge this is the first epidemiologic study to suggest the existence of the “Hispanic paradox” among SLE patients. This seeming paradox has recently been reported for survival among rheumatoid arthritis patients in Texas as well32. Environmental, cultural, and social factors likely play roles. It is possible that higher neighborhood social cohesion and family and social support may improve health outcomes in Asian and Hispanic communities33,34. The Hispanic paradox has also been attributed in part to the “salmon effect”—the phenomenon that Hispanics may return to their home countries at the end of life, becoming statistically “immortal” and leading to an artificially low denominator35. For this reason, we performed competing risk analyses, accounting for the competing risk of loss to follow-up from Medicaid (no further claims or encounters of any kind without documented death). Despite accounting for potential differential follow-up, we still observed a lower mortality risk among Hispanics. Moreover, the follow-up time in Medicaid was very similar in Hispanics, Asians and other racial/ethnic groups in this study. Lastly, we acknowledge that racial and ethnic categorization by self-report is a very imperfect measure of genetic ancestry and, with time and growing genetic admixture in our society, classifying individuals in these limited categories is increasingly challenging36.
The risks of death among Native Americans with SLE were the highest of any racial/ethnic group in the US Medicaid population during the years of study. Prior data about mortality among Native American SLE patients have been limited. In a past study of Canadian Native Americans with SLE by Peschken and colleagues, SLE prevalence, severity and mortality were all reported to be elevated compared to non-native patients37. Fifty-nine Native Americans were included in that study of patients seen in a regional arthritis center, and were found to have more vasculitis, proteinuria, cellular casts, receive more immunosuppressants and prednisone and have higher damage scores and fatality rates. Our study allows an examination of a much larger Native American population and highlights that after adjustment for multiple comorbidities and socioeconomic factors, mortality risks were 20% higher among Native Americans than among African Americans and 40% higher than among White patients. Clearly the reasons for these striking disparities deserve further study. Among individuals of Asian origin, SLE has been reported to be increased in both incidence and severity38. Past population-based studies of SLE mortality from the 1970s and 1980s reported that Asians had three to six times higher mortality rates than Whites39,40. However, in more recent studies in other countries, comparable mortality has been seen among Asian and White patients with SLE11,41. In the current study, U.S. Asian SLE patients enrolled in Medicaid had lower mortality rates and adjusted risks compared to White patients. Variation in mortality across racial and ethnic groups likely depends on both genetic and environmental factors, such as poverty, education, health care access, cultural and health behaviors, and it is interesting to see that, in the U.S. Medicaid system, Asians with SLE have increased survival compared to Whites.
The main strength of this study is the use of nationwide data including over 40,000 SLE patients, providing robust information concerning all-cause mortality over a seven-year period. We fit several models to adjust for potential confounders that might contribute to a high risk of mortality in SLE patients and performed sensitivity analyses accounting for competing risks of loss to follow-up. Additionally, although the SLE-specific risk adjustment index was developed as a specific risk-adjustment index for in-hospital mortality, we successfully applied the SLE-specific risk adjustment index to the Medicaid population and found that it captured SLE-related comorbidity and was itself a predictor for mortality.
A limitation of our study is that it is a prevalent cohort that by definition includes both incident cases, and prevalent cases who have survived. Our estimates cannot address variation by race/ethnicity in mortality anchored at diagnosis and the follow-up period was relatively short at less than four years on average. We did not have access to specific causes of death. Moreover, while we did adjust for many demographic and clinical factors, we were unable to account for SLE disease duration or activity, manifestations or SLE organ damage. Based on work by others, we have developed and used this definition of SLE in past studies, and it is very similar to administrative definitions used in other cohorts, and stringent in that it requires three diagnosis each 30 days apart to eliminate “rule out SLE” and subsequent follow-up visits1,42. While these data have not been directly validated in Medicaid, our estimates of SLE prevalence, overall and in demographic strata, are very close to those published in the CDC-funded epidemiology projects in Michigan and Georgia, providing external validation of our methods2,43. (For example, among Black women in Medicaid we found the prevalence of SLE to be 286/100,0001. Among Black women unrestricted by medical insurance type, in Georgia it was reported to be 241/100,000 and in Michigan it was 181/100,000)2,43. Again, we also acknowledge that self-reported race/ethnicity is imperfect and we were bound by the system of reporting race/ethnicity category used by the Medicaid system, which does not correspond to that of the U.S. Census. Finally, results from this U.S. Medicaid population may not be generalizable to populations abroad or higher SES groups.
In conclusion, even within a relatively short time window of less than three years of average follow-up, we have found marked variation in mortality rates and adjusted hazards ratios among Medicaid patients with SLE according to patient race/ethnicity. Documenting and understanding this variation is important for determining prognoses for individual patients, as well as for further investigation into the root causes of such variation in mortality, including genetic and environmental factors. Further research is needed to identify the mechanisms mediating observed variation among SLE patients. Ultimately, the goal is understand the factors contributing to increased mortality in SLE, in order to modify risk factors and provide tailored therapies to enhance survival.
Supplementary Material
Acknowledgments
The authors would like to acknowledge Alexander Fine, BS for his technical assistance.
Funding statement: This study was supported by NIAMS R01 AR057327 and K24 AR066109A. Dr. Gómez Puerta was supported by Fundación Alfonso Martin Escudero Grant. Dr. Barbhaiya received support from NIAMS T32 AR055885. Dr. Feldman received support from the Lupus Foundation of America Career Development Award.
References
- 1.Feldman CH, Hiraki LT, Liu J, Fischer MA, Solomon DH, Alarcon GS, et al. Epidemiology and sociodemographics of systemic lupus erythematosus and lupus nephritis among US adults with Medicaid coverage, 2000–2004. Arthritis Rheum. 2013;65:753–63. doi: 10.1002/art.37795. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Lim SS, Bayakly AR, Helmick CG, Gordon C, Easley KA, Drenkard C. The incidence and prevalence of systemic lupus erythematosus, 2002–2004: The Georgia Lupus Registry. Arthritis Rheumatol. 2014;66:357–68. doi: 10.1002/art.38239. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Demas KL, Costenbader KH. Disparities in lupus care and outcomes. Curr Opin Rheumatol. 2009;21:102–9. doi: 10.1097/BOR.0b013e328323daad. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Scalzi LV, Hollenbeak CS, Wang L. Racial disparities in age at time of cardiovascular events and cardiovascular-related death in patients with systemic lupus erythematosus. Arthritis Rheum. 2010;62:2767–75. doi: 10.1002/art.27551. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Cooper GS, Treadwell EL, St Clair EW, Gilkeson GS, Dooley MA. Sociodemographic associations with early disease damage in patients with systemic lupus erythematosus. Arthritis Rheum. 2007;57:993–9. doi: 10.1002/art.22894. [DOI] [PubMed] [Google Scholar]
- 6.Alarcon GS, Friedman AW, Straaton KV, Moulds JM, Lisse J, Bastian HM, et al. Systemic lupus erythematosus in three ethnic groups: III. A comparison of characteristics early in the natural history of the LUMINA cohort. LUpus in MInority populations: NAture vs. Nurture. Lupus. 1999;8:197–209. doi: 10.1191/096120399678847704. [DOI] [PubMed] [Google Scholar]
- 7.Alarcon GS, Bastian HM, Beasley TM, Roseman JM, Tan FK, Fessler BJ, et al. Systemic lupus erythematosus in a multi-ethnic cohort (LUMINA) XXXII: [corrected] contributions of admixture and socioeconomic status to renal involvement. Lupus. 2006;15:26–31. doi: 10.1191/0961203306lu2260oa. [DOI] [PubMed] [Google Scholar]
- 8.Jakes RW, Bae SC, Louthrenoo W, Mok CC, Navarra SV, Kwon N. Systematic review of the epidemiology of systemic lupus erythematosus in the Asia-Pacific region: prevalence, incidence, clinical features, and mortality. Arthritis Care Res (Hoboken) 2012;64:159–68. doi: 10.1002/acr.20683. [DOI] [PubMed] [Google Scholar]
- 9.Kuan WP, Li EK, Tam LS. Lupus organ damage: what is damaged in Asian patients? Lupus. 2010;19:1436–41. doi: 10.1177/0961203310370050. [DOI] [PubMed] [Google Scholar]
- 10.Alarcon GS, McGwin G, Jr, Bastian HM, Roseman J, Lisse J, Fessler BJ, et al. Systemic lupus erythematosus in three ethnic groups. VII [correction of VIII]. Predictors of early mortality in the LUMINA cohort. LUMINA Study Group. Arthritis Rheum. 2001;45:191–202. doi: 10.1002/1529-0131(200104)45:2<191::AID-ANR173>3.0.CO;2-2. [DOI] [PubMed] [Google Scholar]
- 11.Johnson SR, Urowitz MB, Ibanez D, Gladman DD. Ethnic variation in disease patterns and health outcomes in systemic lupus erythematosus. J Rheumatol. 2006;33:1990–5. [PubMed] [Google Scholar]
- 12.Bernatsky S, Boivin JF, Joseph L, Manzi S, Ginzler E, Gladman DD, et al. Mortality in systemic lupus erythematosus. Arthritis Rheum. 2006;54:2550–7. doi: 10.1002/art.21955. [DOI] [PubMed] [Google Scholar]
- 13.Tan TC, Fang H, Magder LS, Petri MA. Differences between male and female systemic lupus erythematosus in a multiethnic population. J Rheumatol. 2012;39:759–69. doi: 10.3899/jrheum.111061. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Fernandez M, Alarcon GS, Calvo-Alen J, Andrade R, McGwin G, Jr, Vila LM, et al. A multiethnic, multicenter cohort of patients with systemic lupus erythematosus (SLE) as a model for the study of ethnic disparities in SLE. Arthritis Rheum. 2007;57:576–84. doi: 10.1002/art.22672. [DOI] [PubMed] [Google Scholar]
- 15.Alarcon GS, McGwin G, Jr, Sanchez ML, Bastian HM, Fessler BJ, Friedman AW, et al. Systemic lupus erythematosus in three ethnic groups. XIV. Poverty, wealth, and their influence on disease activity. Arthritis Rheum. 2004;51:73–7. doi: 10.1002/art.20085. [DOI] [PubMed] [Google Scholar]
- 16.Trupin L, Tonner MC, Yazdany J, Julian LJ, Criswell LA, Katz PP, et al. The role of neighborhood and individual socioeconomic status in outcomes of systemic lupus erythematosus. J Rheumatol. 2008;35:1782–8. [PMC free article] [PubMed] [Google Scholar]
- 17.Chibnik LB, Massarotti EM, Costenbader KH. Identification and validation of lupus nephritis cases using administrative data. Lupus. 2010;19:741–3. doi: 10.1177/0961203309356289. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.at https://www.ccwdata.org/web/guest/medicaid-charts#a2_race_2009.)
- 19.Ward MM. Socioeconomic status and the incidence of ESRD. Am J Kidney Dis. 2008;51:563–72. doi: 10.1053/j.ajkd.2007.11.023. [DOI] [PubMed] [Google Scholar]
- 20.Ward MM. Development and testing of a systemic lupus-specific risk adjustment index for in-hospital mortality. J Rheumatol. 2000;27:1408–13. [PubMed] [Google Scholar]
- 21.Quan H, Khan N, Hemmelgarn BR, Tu K, Chen G, Campbell N, et al. Validation of a case definition to define hypertension using administrative data. Hypertension. 2009;54:1423–8. doi: 10.1161/HYPERTENSIONAHA.109.139279. [DOI] [PubMed] [Google Scholar]
- 22.Klompas M, Eggleston E, McVetta J, Lazarus R, Li L, Platt R. Automated detection and classification of type 1 versus type 2 diabetes using electronic health record data. Diabetes Care. 2013;36:914–21. doi: 10.2337/dc12-0964. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Wiley LK, Shah A, Xu H, Bush WS. ICD-9 tobacco use codes are effective identifiers of smoking status. J Am Med Inform Assoc. 2013;20:652–8. doi: 10.1136/amiajnl-2012-001557. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.McCormick N, Lacaille D, Bhole V, Avina-Zubieta JA. Validity of myocardial infarction diagnoses in administrative databases: a systematic review. PLoS One. 2014;9:e92286. doi: 10.1371/journal.pone.0092286. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Davis LAMA, Cannon GW, Mikuls TR, Reimold AM, Caplan L. Validation of diagnostic and procedural codes for identification of acute cardiovascular events in US veterans with rheumatoid arthritis. eGEMs (Generating Evidence & Methods to improve patient outcomes) 2013:3. doi: 10.13063/2327-9214.1023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Fine J, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94:496–509. [Google Scholar]
- 27.Reveille JD, Bartolucci A, Alarcon GS. Prognosis in systemic lupus erythematosus. Negative impact of increasing age at onset, black race, and thrombocytopenia, as well as causes of death. Arthritis Rheum. 1990;33:37–48. doi: 10.1002/art.1780330105. [DOI] [PubMed] [Google Scholar]
- 28.Pons-Estel GJ, Alarcon GS. Lupus in Hispanics: a matter of serious concern. Cleve Clin J Med. 2012;79:824–34. doi: 10.3949/ccjm.79a.12048. [DOI] [PubMed] [Google Scholar]
- 29.Sorlie PD, Backlund E, Johnson NJ, Rogot E. Mortality by Hispanic status in the United States. Jama. 1993;270:2464–8. [PubMed] [Google Scholar]
- 30.Borrell LN, Lancet EA. Race/ethnicity and all-cause mortality in US adults: revisiting the Hispanic paradox. Am J Public Health. 2012;102:836–43. doi: 10.2105/AJPH.2011.300345. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Cortes-Bergoderi M, Goel K, Murad MH, Allison T, Somers VK, Erwin PJ, et al. Cardiovascular mortality in Hispanics compared to non-Hispanic whites: a systematic review and meta-analysis of the Hispanic paradox. Eur J Intern Med. 2013;24:791–9. doi: 10.1016/j.ejim.2013.09.003. [DOI] [PubMed] [Google Scholar]
- 32.Molina E, Haas R, del Rincon I, Battafarano DF, Restrepo JF, Escalante A. Does the “Hispanic paradox” occur in rheumatoid arthritis? Survival data from a multiethnic cohort. Arthritis Care Res (Hoboken) 2014;66:972–9. doi: 10.1002/acr.22254. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Rios R, Aiken LS, Zautra AJ. Neighborhood contexts and the mediating role of neighborhood social cohesion on health and psychological distress among Hispanic and non-Hispanic residents. Ann Behav Med. 2012;43:50–61. doi: 10.1007/s12160-011-9306-9. [DOI] [PubMed] [Google Scholar]
- 34.Patel MI, Schupp CW, Gomez SL, Chang ET, Wakelee HA. How do social factors explain outcomes in non-small-cell lung cancer among Hispanics in California? Explaining the Hispanic paradox. J Clin Oncol. 2013;31:3572–8. doi: 10.1200/JCO.2012.48.6217. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Abraido-Lanza AF, Dohrenwend BP, Ng-Mak DS, Turner JB. The Latino mortality paradox: a test of the “salmon bias” and healthy migrant hypotheses. Am J Public Health. 1999;89:1543–8. doi: 10.2105/ajph.89.10.1543. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Klimentidis YC, Miller GF, Shriver MD. Genetic admixture, self-reported ethnicity, self-estimated admixture, and skin pigmentation among Hispanics and Native Americans. Am J Phys Anthropol. 2009;138:375–83. doi: 10.1002/ajpa.20945. [DOI] [PubMed] [Google Scholar]
- 37.Peschken CA, Esdaile JM. Systemic lupus erythematosus in North American Indians: a population based study. J Rheumatol. 2000;27:1884–91. [PubMed] [Google Scholar]
- 38.Mok MY, Li WL. Do Asian patients have worse lupus? Lupus. 2010;19:1384–90. doi: 10.1177/0961203310375832. [DOI] [PubMed] [Google Scholar]
- 39.Kaslow RA. High rate of death caused by systemic lupus erythematosus among U. S. residents of Asian descent. Arthritis Rheum. 1982;25:414–8. doi: 10.1002/art.1780250409. [DOI] [PubMed] [Google Scholar]
- 40.Serdula MK, Rhoads GG. Frequency of systemic lupus erythematosus in different ethnic groups in Hawaii. Arthritis Rheum. 1979;22:328–33. doi: 10.1002/art.1780220403. [DOI] [PubMed] [Google Scholar]
- 41.Mok CC, Lee KW, Ho CT, Lau CS, Wong RW. A prospective study of survival and prognostic indicators of systemic lupus erythematosus in a southern Chinese population. Rheumatology (Oxford) 2000;39:399–406. doi: 10.1093/rheumatology/39.4.399. [DOI] [PubMed] [Google Scholar]
- 42.Bernatsky S, Joseph L, Pineau CA, Tamblyn R, Feldman DE, Clarke AE. A population-based assessment of systemic lupus erythematosus incidence and prevalence–results and implications of using administrative data for epidemiological studies. Rheumatology (Oxford) 2007;46:1814–8. doi: 10.1093/rheumatology/kem233. [DOI] [PubMed] [Google Scholar]
- 43.Somers EC, Marder W, Cagnoli P, Lewis EE, DeGuire P, Gordon C, et al. Population-based incidence and prevalence of systemic lupus erythematosus: the Michigan Lupus Epidemiology and Surveillance program. Arthritis Rheumatol. 2014;66:369–78. doi: 10.1002/art.38238. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.