Racial disparities in age at time of cardiovascular events and cardiovascular-related death in patients with systemic lupus erythematosus

Authors


Abstract

Objective

To determine whether racial disparities exist with regard to the age at which patients with systemic lupus erythematosus (SLE) experience cardiovascular disease (CVD) and CVD-associated death.

Methods

Using the 2003–2006 Nationwide Inpatient Sample, we calculated the age difference between patients with SLE and their race- and sex-matched controls at the time of hospitalization for a cardiovascular event and for CVD-associated death. In addition, we calculated the age difference between white patients with SLE and sex-matched controls for each minority group for the same outcomes.

Results

The mean age difference between women with and those without SLE at the time of admission for a CVD event was 10.5 years. All age differences between women with SLE (n = 3,627) and women without SLE admitted for CVD were significant (P < 0.0001). Among different racial groups with SLE, black women were the youngest to be admitted with CVD (53.9 years) and to have a CVD-associated in-hospital death (52.8 years; n = 218). Black women with SLE were 19.8 years younger than race- and sex-matched controls at the time of CVD-associated death. Admission trends for CVD were reversed for black women, such that the highest proportions of these patients were admitted before age 55 years, and then the proportions steadily decreased across age categories. Among the 805 men with SLE who were admitted with a CVD event, those who were black or Hispanic were youngest.

Conclusion

There are significant racial disparities with regard to age at the time of hospital admission for CVD events and CVD-related hospitalization resulting in death in patients with SLE.

Cardiovascular disease (CVD) is a significant cause of morbidity and mortality in patients with systemic lupus erythematosus (SLE). Recent data have demonstrated that ischemic heart disease is the leading cause of death in patients with SLE, representing 36% of all SLE-related deaths (1).

SLE is an independent risk factor for myocardial infarction (MI), after controlling for other known CVD risks (2). Given that SLE can be particularly severe in certain ethnic groups, including African Americans, and that inflammation is a key mechanism in the pathogenesis of CVD (3), it is logical to theorize that the epidemiology of CVD and CVD-related death may be altered in those racial groups. In addition, the mean age at the onset of SLE is younger in African Americans than in whites, and the duration of inflammation and disease may result in CVD and/or CVD-associated death at earlier ages in blacks compared with age-matched non-SLE subjects and age-matched white patients with SLE. The goal of our study was to examine the race-specific age differences in SLE patients experiencing CVD and CVD-associated death in a large national population-based in-patient population. We also examined the pattern of admissions for CVD and CVD-related death in SLE patients of different races and age groups.

PATIENTS AND METHODS

Data source.

Data from the Health Care Utilization Project (HCUP) Nationwide Inpatient Sample (NIS) database between 2003 and 2006 were analyzed (4). The NIS is the largest all-payer inpatient care database in the US. It contains data from ∼8 million hospital stays each year. The sampling frame for the 2006 NIS is a sample of hospitals that comprises ∼20% of all hospital discharges in the US. The NIS contains clinical and resource use information included in a typical discharge abstract, with safeguards to protect the privacy of individual patients, physicians, and hospitals. The 2006 NIS database contains all discharge data from 1,045 hospitals located in 38 states, approximating a 20% stratified sample of US community hospitals. The large sample size of the NIS database enables analyses of rare conditions, such as SLE.

Inclusion and exclusion criteria.

All adult patients (18 years of age or older) were included in the analysis. All patients with SLE were identified via the primary or secondary diagnosis codes, using the International Classification of Diseases, Ninth Revision (ICD-9) Clinical Modification (5) code 710.0 (SLE). Diagnoses were coded as either the primary diagnosis or as one of 14 secondary diagnoses. Patients were excluded if they had ICD-9 code 714.0 (rheumatoid arthritis) or code 710.1 (scleroderma), because both of these conditions may potentially overlap with SLE and/or be associated with CVD. We avoided these ICD-9 codes in order not to falsely include patients who may not truly have SLE. The HCUP also supplies a primary procedure code and up to 14 secondary procedure codes. Acute CVD included “MI” (ICD-9 410.0–411.9) and “angina” (ICD-9 413.0–413.9). We also examined all hospitalizations for which principal procedures or secondary procedures included cardiac revascularization (including angioplasty and stent placement and coronary artery bypass) (ICD-9 36.0–36.34). Any admission data that did not include sex, race, age, or median income by zip code were excluded from the analyses.

Outcomes.

Age at the time of a CVD event or CVD-associated death was the primary outcome of interest. Our unadjusted and adjusted models used age (as a continuous variable) as the dependent variable. Cardiovascular disease was defined as any of the above-noted ICD-9 codes for MI and/or angina and/or the above-noted procedural codes. CVD-associated death was defined as any in-hospital death with the above-noted cardiac ICD-9 and/or procedural codes. We also examined CVD-associated death as a secondary outcome of interest.

Covariates.

Sex, age at the time of admission, and race were identified from the data set. Racial groups were identified as black, white, Hispanic, Asian/Pacific Islander, or other (which included American Indians). Data on race were available from 39 states for 2003–2004 and from 41 states for 2004–2006. The median income associated with the zip codes of the subjects was used as a surrogate of socioeconomic status, because individual income levels were not available. Zip code quartiles were defined as <$36,000, $36,000–$44,999, $45,000–$58,999, and ≥$59,000. We identified patients with specific concomitant comorbidities using the Elixhauser Comorbidity Index, which is incorporated in the HCUP software (6). Congestive heart failure, diabetes mellitus, neurologic disease, obesity, and renal failure were included in the comorbidities assessed in regression analyses.

Statistical analysis.

Sociodemographic differences between SLE and non-SLE patients were analyzed using analysis of variance and chi-square testing for continuous and ordinal values, respectively. To examine the age difference at the time of admission for a CVD event and CVD-related death between women and men of the same ethnicity, unadjusted regression models were used in each ethnic/sex group, using the diagnosis of SLE as the independent variable and age as the dependent variable. To examine the age differences between white patients with SLE and minority patients with SLE who have experienced a CVD event or CVD-related death, unadjusted regression analyses were used, with ethnicity as the independent variable and age as the dependent variable. An adjusted regression analysis of CVD-associated death in patients with SLE was also done, controlling for age, sex, race, female sex–race interactions, and the 5 abovementioned comorbidities. Due to colinearity between race and income by zip code, the income variable was not included in this regression analysis.

RESULTS

Sample characteristics.

There were 31,927,484 hospitalizations for non-SLE patients and 124,688 hospitalizations for SLE patients. Patients whose records did not include race, age, sex, and/or median income by zip code were eliminated from the analyses. This resulted in 18,939,916 hospitalizations for subjects without SLE and 90,444 hospitalizations for SLE patients (see Table 1). In the SLE population, 89% of the patients were women, compared with 61% of the non-SLE population. SLE affects women 9 times as frequently as men, matching the sex ratio of this population. The mean age of the female patients with SLE was 50.2 years, compared with 54.7 years for the female patients without SLE, and the mean ages of the men were 52.3 years and 60.4 years, respectively. The racial composition of the SLE group was as follows: 55% white, 28% black, 12% Hispanic, 2% Asian, and 3% other. The proportion of black patients was higher in the SLE group compared with the non-SLE group (28% versus 14%). Although it was statistically different due to the large sample size, the distribution of the median income by zip code for the SLE and non-SLE groups was somewhat similar. Due to the large population, these small differences were still significant (P < 0.0001).

Table 1. Sociodemographic characteristics of hospitalized SLE patients versus hospitalized non-SLE subjects*
 Non-SLE (n = 18,939,916)SLE (n = 90,444)P
  • *

    Except where indicated otherwise, values are the number (%). SLE = systemic lupus erythematosus.

Sex   
 Female11,552,552 (61)80,744 (89)<0.0001
 Male7,387,364 (39)9,700 (11) 
Age, mean ± SD years   
 Women54.7 ± 22.550.2 ± 16.8<0.0001
 Men60.4 ± 18.152.3 ± 17.7 
Race   
 White13,261,642 (70)50,084 (55)<0.0001
 Black2,545,085 (14)25,170 (28) 
 Hispanic2,133,627 (11)10,997 (12) 
 Asian415,041 (2)1,944 (2) 
 Other584,521 (3)2,249 (3) 
Median annual income   
 <$35,0005,472,302 (29)28,841 (32)<0.0001
 $36,000–44,9994,722,823 (25)22,309 (25) 
 $45,000–58,9994,472,827 (24)20,418 (22) 
 ≥59,0004,271,964 (22)18,876 (21) 

Age at the onset of CVD.

We were particularly interested in determining whether racial disparities existed with regard to age at the time of hospitalization for CVD and age at the time of CVD-associated death. Table 2 shows the age at which hospitalized SLE patients and control subjects experienced a CVD event. Table 2 also shows the age differences between racially matched control subjects and SLE patients with CVD. Because the mean age of all control subjects was higher than that of the patients with SLE, the mean age of the SLE group was subtracted from the mean age of the control group to determine the age difference for each racial category. We also calculated the age difference between the white SLE group and each of the other racial groups. Because the mean age of the white patients with SLE was older than that of patients in each of the other racial groups, this difference was calculated as the mean age of the white SLE group minus the mean age of each of the racial groups with SLE.

Table 2. Mean age of the patients with systemic lupus erythematosus (SLE) and control subjects at the time of hospital admission for cardiovascular disease
 SLE patients*Control subjects*Controls versus SLE patientsWhite SLE patients versus minority SLE patients
Age difference, yearsPAge difference, yearsP
  • *

    Values are the mean age, years (number of subjects admitted) [%].

All women60.8 (3,627)71.3 (608,543)10.5<0.0001  
 White63.5 (2,352) [65]72.5 (466,056) [76]9.0<0.0001  
 Black53.9 (786) [21]65.8 (67,602) [11]11.9<0.00019.6<0.0001
 Hispanic57.5 (343) [9]68.6 (48,085) [8]11.1<0.00016.0<0.0001
 Asian60.6 (55) [2]71.5 (10,994) [2]10.8<0.00012.90.11
 Other57.0 (91) [3]68.6 (15,806) [3]11.6<0.00016.4<0.0001
All men60.4 (805)65.9 (828,608)5.5<0.0001  
 White63.0 (603) [75]66.7 (665,999) [80]3.7<0.0001  
 Black52.3 (114) [14]61.4 (59,851) [7]9.1<0.000110.7<0.0001
 Hispanic51.4 (58) [7]63.5 (59,556) [7]12.1<0.000111.6<0.0001
 Asian59.2 (8) [1]65.9 (16,914) [2]6.70.163.80.41
 Other55.4 (22) [3]62.6 (26,288) [3]7.20.0097.60.007

Women with CVD.

There were 3,627 admissions for acute CVD among female patients with SLE. The mean ± SD age of the female patients with SLE who were hospitalized with CVD was 60.8 ± 13.7 years, whereas the mean age of female patients without SLE who were admitted with CVD was 71.3 ± 13.4 years. The range of age differences between the female control subjects and female SLE patients was 9.0–11.9 years. Among the women with CVD, the group of black patients with SLE was the youngest (53.9 years versus 65.8 years for black women without SLE). This 11.9-year difference was the largest age difference between any SLE group and the control racial group. The mean age at the time of the CVD event for control subjects versus patients with SLE was examined via unadjusted regression models. All differences were significant (P < 0.0001). Although white patients with lupus were significantly younger than their non-SLE controls at the time of the CVD event and CVD-related death, they were significantly older than most of the other racial groups with SLE. When age at the time of the CVD hospitalization was compared between white and racial minority patients with SLE, significant age differences were demonstrated for the black (P < 0.0001), Hispanic (P < 0.0001), and other groups (P < 0.0001), but not for the Asian group. The largest age difference between white and minority groups was that for black patients with SLE. There was a 9.6-year difference for age at the time of hospitalization for CVD, with black patients being age 53.9 years and white patients with SLE being 63.5 years.

We were interested in demonstrating any trends that differ between racial groups with regard not only to the mean age at presentation for CVD events, but also to the proportion of patients with CVD events in different decades of life. We chose age categories similar to those previously used in the general population: age younger than 55 years, 55–64 years, 65–74 years, and ≥75 years (7). The specific proportions of patients with and those without SLE in each racial group with either CVD events or CVD-related death are shown in Table 3. The proportion of women with SLE who were admitted with CVD at an age younger than 55 years was 2.4-fold to 3.3-fold higher than the proportion of women without SLE. Fifty-five percent of black women with SLE were admitted with CVD when they were in the youngest age group, compared with 41% of the Hispanic women, 33% of the Asian women, and 26% of the white women with SLE. Figure 1 demonstrates these trends visually. In women without SLE, the general tendency was that the proportion of those admitted with a CVD-related event increased with age. This trend was reversed in patients with SLE, with the effect being most dramatic in black and Hispanic women. This apparent reversal is likely in part attributable to the young age at which these patients with SLE experience their CVD events.

Table 3. Percentage of female and male patients without or with systemic lupus erythematosus (SLE) experiencing cardiovascular disease (CVD) or CVD-related death, according to race and age
 Age, years
<5555–6465–74≥75
Women    
 Without SLE and with CVD    
  White11162350
  Black23222530
  Hispanic16202836
  Asian10162945
 With SLE and CVD    
  White26262523
  Black5524156
  Hispanic4132207
  Asian33242023
 Without SLE and with CVD-related death    
  White371673
  Black12152350
  Hispanic7132357
  Asian582166
 With SLE and CVD-related death    
  White17242732
  Black5123179
  Hispanic23362912
  Asian22342222
Men    
 Without SLE and with CVD    
  White19242631
  Black33262318
  Hispanic26252623
  Asian21242728
 With SLE and CVD    
  White26262820
  Black5826124
  Hispanic6791410
  Asian252550
Without SLE and with CVD-related death    
  White6112261
  Black17202637
  Hispanic13152646
  Asian8102359
 With SLE and CVD-related death    
  White18252532
  Black6733
  Hispanic
  Asian100
Figure 1.

A and B, Proportion of cardiovascular disease (CVD) events and CVD-related death, respectively, in women with or without systemic lupus erythematosus (SLE), according to age and race. C and D, Proportion of CVD events and CVD-related death, respectively, in men with or without SLE, according to age and race.

Men with CVD.

Among men with SLE, there were 805 admissions for CVD. The mean age of the men with SLE who were admitted with CVD was 60.4 years, versus 65.9 years for men without SLE. The range for age differences between racially matched men without SLE and men with SLE was 3.7–12.1 years. The greatest difference, 12.1 years, was for Hispanic men with SLE, who had a mean age of 51.4 years versus 63.5 years of age for Hispanic men without SLE. The Hispanic men also had the greatest age difference when compared with the white men with SLE (11.6 years younger; P < 0.0001). The next youngest group was black men with SLE men, who were 9.1 years younger than their racially matched non-SLE controls (P < 0.0001) and 10.7 years younger than white men with SLE (P < 0.0001). The mean ages at the time of CVD-related admission in male control subjects compared with male patients with SLE as well as in white men with SLE and compared with minority men with SLE were all significantly different, except for the Asian group for each comparison (Table 2). Table 3 demonstrates that Hispanic and black men were admitted with CVD when they were younger than age 55 years in 67% and 58% of cases, respectively. In men without SLE, the general trend was that the number of those presenting with CVD slowly decreased over time and age categories. This effect was accentuated in men with SLE, especially Hispanics and blacks.

Age at CVD-associated death.

We performed similar analyses of age at the time of death associated with a CVD event. Table 4 shows the mean ages of women and men with and without SLE at the time of their CVD-associated deaths.

Table 4. Mean age of the patients with systemic lupus erythematosus (SLE) and control subjects at the time of cardiovascular disease–related death*
 SLE patients*Control subjects*Controls versus SLE patientsWhite SLE patients versus minority SLE patients
Age difference, yearsPAge difference, yearsP
  • *

    Values are the mean age, years (number of subjects admitted) [%]. NA = not applicable.

All women63.4 (218)78.2 (35,003)14.8<0.0001  
 White67.1 (141) [65]79.3 (27,142) [78]12.2<0.0001  
 Black52.8 (47) [21]72.6 (3,674) [10]19.8<0.000114.3<0.0001
 Hispanic62.0 (17) [8]74.9 (2,472) [7]12.9<0.00015.10.09
 Asian63.8 (9) [4]77.5 (825) [2]13.70.00043.30.41
 Other63.5 (4) [2]75.4 (890) [3]11.90.073.60.54
All men63.7 (38)74.7 (35,501)11.0<0.0001  
 White66.2 (33) [86]75.7 (28,178) [79]9.5<0.0001  
 Black52.4 (3) [8]68.4 (2,785) [8]16.00.0513.90.06
 Hispanic071.0 (2,588) [7]NANANANA
 Asian45.0 (1) [3]75.4 (965) [3]30.40.0221.20.09
 Other36.0 (1) [3]71.7 (985) [3]35.70.00530.20.02

Women with CVD-associated death.

Among women with SLE who experienced a CVD-associated in-hospital death, there were 218 admissions; the mean ± SD age of women with SLE who had a CVD-related death was 63.4 ± 15.5 years. The range of age differences between women without SLE compared with those with SLE was 11.9–19.8 years. Black women with SLE were the youngest group to have an in-hospital CVD-associated death. Among this group, the mean age at the time of death was 52.8 years, which is 19.8 years younger (P < 0.0001) than the age at the time of death in black women without SLE. Black women with SLE who died were 14.3 years younger than white women with SLE who died during their CVD-related hospitalization (P < 0.0001). The other racial groups with SLE were all much younger than the matched non-SLE group, with the highest to lowest age differences in the following order: Asian (13.7 years younger; P = 0.0004), Hispanic (12.9 years younger; P < 0.0001), whites (12.2 years younger; P < 0.0001), and other (11.9 years younger; P = 0.07). Among the racial minorities with SLE, black women were the youngest to have CVD-associated death. The mean ages of the other minorities who had CVD-associated death, as compared with those of white women with SLE, were not significant in unadjusted regression models. Table 3 demonstrates that more than half of the black women with SLE who died of CVD were younger than age 55 years. Black women were the only group of women with SLE who had a consistent decline in mortality sequentially through the age categories. As compared with women without SLE, women with SLE in the other 3 racial groups had attenuated increases in the percentages of patients dying of CVD with advancing age.

Men with CVD-associated death.

Among men with SLE, 38 experienced CVD-associated death (Table 4). The mean ± SD age of men with SLE having a CVD-related in-hospital death was 63.7 ± 14.8 years, almost exactly the same age as the women with SLE and CVD-related death. Thirty-three of the 38 men were white, 3 were black, 1 was Asian, and 1 was classified as other. No Hispanic men with SLE were identified. There was a 16-year age difference between black men with and those without SLE (P = 0.05), but interpretation of this result is limited by the few men included in the analysis. There was a 9.5-year age difference between white men with and those without SLE who died during their CVD-related hospitalization (P < 0.0001).

CVD-related mortality regression.

An adjusted logistic regression model controlling for age, sex, race, sex–race interactions, congestive heart failure, diabetes mellitus, neurologic disease, obesity, and renal failure revealed that age, congestive heart failure, obesity, and renal failure were significant variables in determining CVD-related mortality among patients with SLE (Table 5). Increasing age overall was significantly related to CVD-related mortality. The presence of concomitant congestive heart failure and renal failure increased the odds of CVD-related mortality in patients with SLE by 2.2-fold and 3-fold, respectively. Obesity was not a risk factor for dying of a CVD-related event; rather, those patients who were not obese were at a slightly higher risk of dying of CVD while hospitalized.

Table 5. Multivariate logistic regression analysis of characteristics associated with cardiovascular disease–related death in patients with systemic lupus erythematosus
CovariateOR (95% CI)*P
  • *

    OR = odds ratio; 95% CI = 95% confidence interval.

Age1.044 (1.034–1.055)<0.0001
Congestive heart failure2.220 (1.606–3.070)<0.0001
Renal failure3.024 (2.240–4.083)<0.0001
Obesity0.382 (0.157–0.932)0.03

DISCUSSION

We have demonstrated, by reviewing a large national database of inpatient hospitalizations, that there are significant racial disparities with regard to age at the time of hospital admission for CVD and for CVD-related in-hospital death in patients with SLE. The mean ages at the time of hospital admission for CVD for men and women in our study are similar to those that have been previously reported for patients in the general population (8). Women with SLE were, on average, 10.5 years younger than women without SLE at the time of admission for CVD and 14.8 years younger at the time of their CVD-related death. Manzi et al previously showed that age-specific incidence rates for MI and angina are much higher in women with SLE compared with participants in the Framingham Offspring Study (9). In that study, the mean age of their 33 patients with SLE at the time of the first CVD event was 48 years (range 22–72 years). A more recent study identified 53 British patients with SLE with a mean ± SD age of 53 ± 10 years (range 33–73 years) at the time of their first CVD event (10).

In our study, the mean age of women with SLE at the time of hospital admission for CVD was 60.8 years, which is older than the ages reported in these 2 other studies. This difference may be attributable to the larger number of patients and/or the wider age range (18–98 years) of the identified patients with SLE in our study. Alternatively, we were unable to determine whether a CVD-related admission was a first event, such that if an individual had experienced ≥1 previous non–life-threatening CVD event, any subsequent events occurring at an older age were included in our analysis. Very little information about minority men with SLE is available for outcomes such as CVD, given that most cohorts do not include a meaningful number of men with this outcome. To our knowledge, this study is the first to compare age at the time of CVD and CVD-related death in men with and those without SLE. On average, men with SLE were 5.5 years younger than men without SLE at the time of a CVD-related admission and 11.0 years younger at the time of a CVD-related in-hospital death.

Hospitalizations due to acute MI have been shown to be 2.3 times more common in women with SLE ages 18–44 years than in women without SLE in the same age group. Younger women with SLE had a particularly high risk of dying during a CVD-related hospitalization (11). Race differences were not reported in those studies. Our data show that in specific racial groups, the mean ages at the time of CVD and CVD-related death in women with SLE versus women without SLE were all statistically significantly lower when compared with racially matched control subjects. In the National Registry of Myocardial Infarction 2, the characteristics of >500,000 black patients versus white patients in the general population experiencing acute MI and MI-related death were compared. The proportion of black women was higher than the proportion of white women, especially in younger age groups. There was also a higher proportion of black patients younger than age 65 years at the time of MI as compared with white patients. In all age groups younger than age 65 years, black patients had higher mortality than white patients, whereas there was a trend toward lower mortality compared with whites in patients ≥80 years of age (7).

Our study illustrates similar results, with a larger proportion of black female subjects in the general population with CVD and CVD-associated death in younger age groups as compared with white patients, with a similar trend also seen in Hispanic women. As age increased, the proportion of women without SLE in each of the individual racial categories also increased. The opposite trend was observed in female patients with lupus who were black or Hispanic, and to a lesser extent in those who were Asian. The proportion of white women with SLE admitted with CVD events remained fairly constant across age groups.

In all racial and sex groups, the presence of SLE either reversed or attenuated the proportions of patients admitted with CVD and CVD-related death across the age categories. The proportion of women admitted for CVD-related death increased across age categories in patients without SLE, whereas the highest proportion of black women with SLE (55%) was in the group younger than age 55 years, and this proportion progressively decreased as age increased. Some of the most startling results from this study include the vast age difference between patients with SLE and their age-matched controls at the time of CVD-related death: black women with SLE were almost 20 years younger than black women without SLE (52.8 years versus 72.6 years of age). Reasons for such significant age differences are not clear but may be related to specific race-related and SLE-related causes.

Overall, black race is independently associated with a worsened probability of survival. In the general population, after adjusting for CVD risk factors, associated comorbidities, insurance, medications, and MI characteristics, black race is associated with higher mortality as age decreases; this is true for men and women, although it is more pronounced in men (7). American black women have been shown to have an early age at onset of SLE, aggressive disease at disease onset, and a high frequency of renal involvement (12, 13). Our adjusted model demonstrated that concomitant comorbidities are associated with CVD-associated mortality. Patients with SLE were 2.2-fold more likely to experience CVD-associated death with concomitant congestive heart failure and 3-fold more likely to experience CVD-associated death with concomitant renal failure. It may be that the higher risk of CVD and CVD-related death in certain racial groups at younger ages is a result of higher SLE disease activity, including renal and central nervous system complications. Data from the Third National Health and Nutrition Examination Survey demonstrated that an increased number of CVD risk factors cluster among Americans with lower socioeconomic status, particularly among non-Hispanic blacks (14). Shah et al showed that patients with SLE who are hospitalized with an MI have higher mortality with associated congestive heart failure (15). In addition to SLE-related comorbidities, clustering of CVD risks may be contributing to earlier CVD events and CVD-related death.

Given that many of the traditional CVD risk factors are modifiable, active risk factor management should be targeted to women with SLE. In particular, identification of SLE patients at particular risk, including young black and Hispanic patients, should be a high priority. Risk factor management and recognition in SLE remain a pertinent issue. Urowitz et al demonstrated that even though hypertension and hyperlipidemia had improved in a closely monitored cohort, there was still room for tighter control. Over a 12-year period of observation, only 22% of patients with hypercholesterolemia achieved adequate control (16). The definition of hypercholesterolemia varies greatly between studies, but its prevalence is 4–75% in SLE cohorts (9, 16, 17). Studies in ethnically diverse lupus populations are warranted to examine the success in addressing risk factor management and barriers that may exist in optimizing risk factor management, and whether there are specific racial disparities in terms of physicians' prescribing habits, adherence to prescribed medications, optimization of risk factor management, awareness of heart attack warning signs, and appropriate management of acute CVD events. In the general population, all of these issues have been shown to be different between the races, with poorer management achieved in ethnic minorities (18–23).

The strengths of this study include the fact that it is based in the NIS, a large database that is well suited for diseases with a low incidence, such as SLE. SLE is a disease that affects predominantly women (9:1 ratio of women to men), and the large data set that includes men with lupus has provided us with unique insight about the incidence of CVD in men with lupus. Due to the rigorous nature of the sampling techniques, the NIS is free of sampling and referral bias. The NIS does not contain information about SLE severity, SLE disease duration, medications, or whether a CVD event was a first event. This information would be helpful to determine associations of SLE-related indices with CVD events, biases that could exist with regard to SLE and/or race, and whether the age noted at the time of admission may be different between first CVD events and all CVD events. SLE may have been underdiagnosed, such that not all patients with SLE were captured in the database. We also cannot confirm that all patients with SLE met the clinical criteria of a diagnosis of SLE, but all diagnoses were made by physicians. The colinearity of race and the measure of income by zip code prevented us from reliably including zip code in the regression analysis for CVD mortality.

We have shown significant differences between white patients and ethnic minority patients, including blacks and Hispanics, in the age at which they are admitted to the hospital with CVD and CVD-associated death. Large multicenter and multiethnic studies are needed to examine whether SLE-specific indices may be contributing to these age disparities. Physician awareness and patient education should be considerations in all patients with SLE in the third and fourth decades of life, especially for African Americans and Hispanics. In this way, primary and secondary preventative measures can be undertaken at the most favorable time.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Scalzi 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.

Study conception and design. Scalzi, Hollenbeak, Wang.

Acquisition of data. Scalzi, Wang.

Analysis and interpretation of data. Scalzi, Hollenbeak, Wang.

Ancillary