A clinical and serologic comparison of African American and Caucasian patients with systemic sclerosis

Authors


Abstract

Objective

Epidemiology studies suggest that systemic sclerosis (SSc) is more common, occurs at a younger age, and is more severe in African Americans than Caucasians. However, the scleroderma autoantibody profile is very different between these 2 ethnic groups. This study was undertaken to examine the demographic and disease features, frequency and severity of internal organ system involvement, and survival in African American patients compared to Caucasian patients with SSc, giving particular attention to their serum autoantibody profiles.

Methods

Demographic features, clinical characteristics, autoantibody profile, organ involvement, and survival were studied in consecutive African American and Caucasian patients with SSc whose visits were recorded between 1972 and 2007 as part of the Pittsburgh Scleroderma Database. The Medsger Severity Score for SSc was used to determine the severity of disease.

Results

African American patients were more likely to have anti–topoisomerase I (anti–topo I), anti–U1 RNP, and anti–U3 RNP autoantibodies. In comparing African American and Caucasian patients with these antibodies, pulmonary fibrosis was found to be more frequent and more severe, and the rate of survival was decreased, in African American patients with anti–topo I antibodies compared to Caucasian patients with anti–topo I. Pulmonary fibrosis was also more severe in the anti–U1 RNP–positive patients, but this was not associated with a difference in survival between African Americans and Caucasians. Anti–U3 RNP was associated with more severe gastrointestinal involvement in African Americans compared to Caucasians.

Conclusion

African Americans with SSc have more severe disease complications compared to Caucasians with SSc, and this is associated with both the type of autoantibody present and the severity of interstitial lung disease. Thus, it is hoped that early aggressive intervention in African Americans with interstitial lung disease will improve outcomes.

Several studies have suggested that African Americans have an increased incidence of systemic sclerosis (SSc) and a worse prognosis than has been observed in Caucasians with this disease (1, 2). Moreover, symptoms attributable to SSc occur at an earlier age in African Americans, and interstitial lung disease (3) and pulmonary arterial hypertension (PAH) (4) have both been shown to be more frequent and more severe in African Americans.

Serum autoantibodies in patients with SSc are strongly associated with clinical features. For example, the presence of anticentromere antibodies (ACAs) has been noted in association with PAH, anti–topoisomerase I (anti–topo I) antibodies in association with interstitial lung disease, and anti–RNA polymerase III (anti–RNAP III) antibodies in association with renal crisis (5). Ethnicity is also strongly associated with the scleroderma autoantibodies. Consistent with this finding, the frequency of serum autoantibodies differs between African American and Caucasian patients with SSc. Anti–topo I is somewhat more frequently detected in African American patients with SSc (6–10). Anti–U3 RNP antibodies are particularly common in African American patients with SSc (10–12). In contrast, ACAs are uncommon in African American patients (13).

There are genetic differences between African Americans and Caucasians. These differences may contribute to the varying autoantibody profiles among different patient populations (12). HLA–DRB1*08 is more frequent in African American patients with SSc than in healthy African Americans and Caucasian patients with SSc. The frequency of HLA–DRB1*1101 is increased in both African American and Caucasian patients with SSc whose serum is positive for anti–topo I antibodies. A recent study demonstrated a strong association of antifibrillarin (anti–U3 RNP) with HLA–DRB1*08:04 (14). These genetic influences could be important factors contributing to differences in disease outcome. Results from some studies have suggested that, similar to the observations in patients with systemic lupus erythematosus, socioeconomic differences between African American and Caucasian patients with SSc may influence survival (10, 12).

In our cumulative experience, the antibodies observed most frequently in Caucasian patients with SSc are ACAs, anti–topo I, and anti–RNAP III, while the most frequently found antibodies in African American patients with SSc are anti–topo I, anti–U3 RNP, and anti–U1 RNP. Between 1984 and 1989, 462 consecutive patients with SSc in our series underwent testing for all serum antibodies (15). The frequency of these autoantibodies in African American and Caucasian patients, respectively, was as follows: 7% versus 25% with ACAs, 27% versus 21% with anti–topo I, 3% versus 31% with anti–RNAP III, 16% versus 7% with anti–U1 RNP, 40% versus 2% with anti–U3 RNP, and 1% versus 4% with anti-Th/To. More than 75% of Caucasian patients with SSc are positive for ACA, anti–topo I, or anti–RNAP III antibodies, while >80% of African American patients with SSc are positive for anti–U1 RNP, anti–U3 RNP, or anti–topo I antibodies. For these reasons, comparisons of cohorts of African American and Caucasian patients with SSc, with regard to their clinical manifestations, organ system involvement, and long-term outcomes, should also consider the autoantibody prevalence in the groups being studied.

This study examines the demographic and disease classification features, frequency and severity of internal organ system involvement, and survival in African American and Caucasian patients with SSc, with particular attention given to their serum autoantibody profiles.

PATIENTS AND METHODS

Patients.

The Pittsburgh Scleroderma Databank is a prospective natural history study of consecutive patients with SSc who were first evaluated beginning on January 1, 1972. Patient outcomes, including organ system involvement and survival, were determined for all patients based on data obtained from the initial visit, followup visits, managing physicians' office records, annual/biannual patient-completed questionnaires, and telephone interviews (16). We used the level of education as a surrogate for socioeconomic status. Deaths were confirmed through the Social Security Death Index as of December 31, 2009, and causes of death were determined from hospital and physician records and death certificates (17).

An accurate evaluation of disease involvement in organ systems is always challenging, but we have worked hard to objectively characterize our patients and to standardize our approach as best as possible. We have previously validated these methods of assessing severe internal organ involvement and causes of death in patients with SSc (16). In this study, we included all African American and Caucasian patients who were first evaluated during 1972–2007.

Evaluation of organ system involvement.

The severity of organ system involvement in SSc was defined as previously described (16). Severe organ involvement was defined as either grade 3 (severe) or grade 4 (end stage) based on the revised Medsger Severity Score for SSc (18).

Different features were evaluated for severity of disease in each organ system. Severe involvement of the peripheral vascular system was defined as the presence of digital tip ulcers or gangrene. In the skin, severe disease was defined as a modified Rodnan total skin thickness score (19) of >30. In the joints and tendons, a fingertip-to-palm distance of ≥4.0 cm, which is indicative of significant loss of fist closure, was required for designation as severe. For the skeletal muscles, severe proximal muscle weakness had to be identified on physical examination or assistance with ambulation aids was required. Severe disease involvement in the gastrointestinal (GI) tract included malabsorption syndrome, episodes of pseudoobstruction, hyperalimentation, or death due to intestinal involvement. For severe lung disease, pulmonary fibrosis had to be detectable on radiographs, in addition to at least one of the following features: forced vital capacity <50% of predicted, requirement for oxygen in those with interstitial lung disease, lung transplantation performed, or death due to SSc-related interstitial lung disease. Severe pulmonary involvement could also include the presence of PAH (not secondary to pulmonary fibrosis), defined as a pulmonary artery systolic pressure of ≥65 mm Hg estimated on echocardiography or measured at right-sided heart catheterization, a requirement for oxygen due to SSc-related PAH, or death due to SSc-related PAH. In the heart, severe disease was defined as left-sided congestive heart failure due to SSc (not secondary to renal crisis), arrhythmia requiring treatment, a left ventricular ejection fraction of <40%, heart transplantation performed, or death due to SSc-related heart disease. Severe involvement of the kidney included development of renal crisis, defined as new malignant hypertension with a serum creatinine level of ≥3.0 mg/dl at any time, requirement for dialysis, renal transplantation performed for SSc-related renal disease, or death due to SSc-related renal disease.

Autoantibody studies.

All patients had an antinuclear antibody (ANA) test performed in the same laboratory, using HEp-2 cells as substrate. Serum tests for SSc-specific antibodies, including ACAs, anti–topo I, anti–U1 RNP, anti–RNAP III, anti–U3 RNP, anti-Ku, and anti-Th/To, were performed in serum according to previously described methods (6, 11, 20, 21). SSc-specific autoantibody testing was not performed on all patients in the Databank, since patient entry spanned >3 decades. The few patients with >1 SSc-associated serum autoantibody were not included in this study.

Statistical analysis.

The Pittsburgh Scleroderma Databank uses the Medlog database management system. Baseline characteristics and organ system involvement were compared between groups using the t-test, chi-square test, or Wilcoxon test, as appropriate. Survival was assessed by Kaplan-Meier analysis, and associations with mortality were determined using Cox proportional hazards models. Predictors of severe lung involvement were assessed by multivariable logistic regression.

RESULTS

Baseline characteristics.

There were 2,945 Caucasian patients with SSc and 203 African American patients with SSc included in the study. The baseline characteristics of all patients at first visit are presented in Table 1. African American patients were significantly younger at disease onset, at first SSc diagnosis, and at first Pittsburgh visit. Although disease duration was shorter in the African American patients, functional scores on the Health Assessment Questionnaire disability index (HAQ DI) (22) at presentation indicated that African Americans had worse function compared to Caucasians (median HAQ DI score 1.33 versus 0.86; P = 0.01). This finding may be attributable, in part, to an increased frequency of diffuse cutaneous SSc (dcSSc) involvement in African American patients compared to Caucasian patients (50% versus 43%; P = 0.03).

Table 1. Baseline characteristics at first visit among 3,148 African American and Caucasian patients with SSc between 1972 and 2007*
FeatureAfrican American (n = 203)Caucasian (n = 2,945)P
  • *

    Except where indicated otherwise, values are the number (%) of patients. NS = not significant; IQR = interquartile range; CTD = connective tissue disease.

  • In systemic sclerosis (SSc)–specific autoantibody testing, denominators were different because not all patients in the Databank underwent testing for these autoantibodies over the 3 decades.

  • Health Assessment Questionaire (HAQ) disability index scores were available for 23% of the African American and Caucasian patients.

Sex, female155 (76)2,374 (81)NS
Education, completed high school or higher105 (52)1,698 (58)NS
Residence within 100 miles of Pittsburgh133 (66)1,428 (48)<0.0001
Age at symptom onset, mean ± SD years38.4 ± 14.342.9 ± 14.9<0.001
Age at first visit, mean ± SD years44.8 ± 12.350.6 ± 14.1<0.001
Disease duration from onset to first visit, median (IQR) years2.97 (1.27–8.04)3.46 (1.64–10.84)NS
Tobacco use  0.002
 Never107 (53)1,618 (55) 
 Prior44 (22)827 (28) 
 Current52 (26)482 (16) 
Diffuse cutaneous involvement102 (50)1,264 (43)0.03
Overlap syndrome with another CTD32 (16)237 (8)0.0002
Body mass index, mean ± SD kg/m224.1 ± 4.924.1 ± 4.9NS
Autoantibodies   
 Anticentromere12 (7)597 (22)<0.0001
 Anti–topoisomerase I48 (26)490 (17)0.01
 Anti–RNA polymerase III20 (11)578 (21)0.001
 Anti–U1 RNP33 (18)137 (5)<0.0001
 Anti–U3 RNP34 (19)80 (3)<0.0001
 Anti-Th/To7 (4)168 (6)NS
HAQ disability index, median (IQR)1.33 (0.50–1.81)0.86 (0.30–1.38)0.01

Autoantibody frequencies in these patients had a similar distribution as that seen in our smaller subset of patients who underwent testing for all serum autoantibodies. African American patients, as compared to Caucasian patients, had significantly higher frequencies of anti–U3 RNP (19% versus 3%), anti–U1 RNP (18% versus 5%), and anti–topo I (26% versus 17%).

Organ involvement and survival.

Table 2 shows the frequency of organ system involvement in African American and Caucasian patients with SSc. The overall frequency of organ involvement, as well as the frequency of severe organ system involvement, is presented. Overall, skeletal muscle involvement was more common in African American patients than in Caucasian patients with SSc (27% versus 12%; P < 0.0001). Severe skeletal muscle involvement was also more often encountered in African Americans compared to Caucasians (6% versus 2%; P = 0.001). The same differences were observed for the overall frequency of pulmonary fibrosis between African American and Caucasian patients (54% versus 38%; P < 0.0001), and for the frequency of severe pulmonary fibrosis (32% versus 13%; P = 0.0001). Slightly more African American patients than Caucasian patients developed GI involvement, although there was no significant difference in the frequency of severe GI disease. There were no notable differences in the frequency of PAH, cardiac, renal, or joint/tendon involvement.

Table 2. Frequency of organ system involvement and severe organ system involvement in African American and Caucasian patients with systemic sclerosis*
Organ systemAfrican American (n = 203)Caucasian (n = 2,945)P
OverallSevereOverallSevereOverallSevere
  • *

    Values are the number (%) of patients. NS = not significant; GI = gastrointestinal; PAH = pulmonary arterial hypertension.

  • Values for these organ systems are the number/total number of patients, representing the proportion of patients with severe organ system involvement among those for whom there were objective data that permitted assessment of severity. For all other organ systems, the denominator was all patients.

Vascular201 (99)97 (48)2,880 (98)1,296 (44)NSNS
Skin185 (91)47 (23)2,798 (95)618 (21)NSNS
Joints/tendons172 (85)45 (22)2,331 (79)589 (20)0.06NS
Skeletal muscles54 (27)12 (6)367 (12)59 (2)<0.00010.001
GI tract143/169 (85)20/169 (12)1,826/2,312 (79)254/2,312 (11)0.02NS
Pulmonary fibrosis100/186 (54)60/186 (32)973/2,562 (38)337/2,562 (13)<0.00010.0001
PAH23/149 (15)17/148 (11)317/2,005 (16)243/2,005 (12)NSNS
Heart43/173 (25)31/173 (18)529/2,411 (22)382/2,411 (16)NSNS
Kidney25 (12)20 (10)282 (10)236 (8)NSNS

The overall survival curves were significantly different, with African American patients having worse survival than Caucasian patients (survival rate 66% versus 75% at 5 years, and 51% versus 60% at 10 years; P = 0.0063). After adjustment for age, sex, and diffuse disease, African American patients were 1.68 times more likely to have died (95% confidence interval [95% CI] 1.30–2.16) at 5 years of followup when compared to Caucasian patients (Table 3). Adjustment for tobacco use and the presence of interstitial lung disease did not change the overall results. The proportional hazards assumption was met for all models. We also assessed the frequency of use of immunosuppressive agents among these patients, just to verify that access to care and use of these agents was not an easy explanation for the differences in survival. There was no difference in the use of immunosuppressive agents either at first visit or ever between African Americans and Caucasians (frequency of use 27% versus 24%, respectively). When this factor was included in our proportional hazards model, African American race was still an independent predictor of death.

Table 3. Risk factors for mortality within 5 years from the first Pittsburgh visit
 Hazard ratio95% confidence intervalP
Age at first Pittsburgh visit1.041.03–1.04<0.0001
Male1.351.15–1.600.003
Diffuse cutaneous involvement1.341.16–1.54<0.0001
African American1.681.30–2.16<0.0001

If we were to publish these data without qualifications, we would postulate that African American patients had worse survival than Caucasian patients because they more often had diffuse scleroderma and severe pulmonary fibrosis, both of which are recognized risk factors in SSc. In order to determine whether there were specific ethnic differences, we focused on the SSc-associated serum antibodies most frequently detected in African Americans. We specifically compared African American and Caucasian patients who had anti–topo I, anti–U3 RNP, and anti–U1 RNP antibodies, which are the antibodies most frequently detected in African American patients with SSc.

Autoantibody patterns.

Anti–topo I antibody–positive patients.

We compared the demographic characteristics, HAQ DI scores, and frequency of organ involvement between 48 African American patients with SSc and 490 Caucasian patients with SSc who were positive for anti–topo I antibodies (Table 4). African American patients with anti–topo I antibodies were younger than Caucasian patients with these antibodies at the onset of their disease (mean age 37 years versus 43 years; P = 0.005), but there was no difference in the frequency of diffuse cutaneous scleroderma in these patients (71% of African Americans versus 65% of Caucasians). In addition, the peak modified Rodnan total skin thickness score was similar in both groups.

Table 4. Demographic characteristics, disease classification features, and frequency of severe organ system involvement in patients with systemic sclerosis who were anti–topoisomerase I antibody positive and were first evaluated during 1972–2007*
FeatureAfrican American (n = 48)Caucasian (n = 490)P
  • *

    Except where indicated otherwise, values are the number (%) of patients. NS = not significant; HAQ = Health Assessment Questionnaire; TSS = total skin thickness score; GI = gastrointestinal; PAH = pulmonary arterial hypertension.

  • Values for these organ systems are the number/total number of patients, representing the proportion of patients with severe organ system involvement among those for whom there were objective data that permitted assessment of severity. For all other organ systems, the denominator was all patients.

Sex, female36 (75)366 (75)NS
Age at symptom onset, mean ± SD years36.6 ± 14.143.3 ± 15.90.005
Disease duration from onset to diagnosis, mean ± SD years3.0 ± 5.23.6 ± 6.0NS
HAQ disability index, mean ± SD maximum score1.6 ± 1.01.3 ± 0.90.097
Diffuse cutaneous involvement34 (71)317 (65)NS
Organ system with severe involvement   
 Skin12 (25)127 (26)NS
  Mean ± SD maximum modified Rodnan TSS20.8 ± 12.620.0 ± 13.2NS
 Vascular25 (52)250 (51)NS
 Joints/tendons19 (40)164 (33)NS
 Skeletal muscles1 (2)8 (2)NS
 GI tract5/38 (13)35/400 (9)NS
 Pulmonary fibrosis19/43 (44)82/450 (18)0.0001
 PAH2/39 (5)10/332 (3)NS
 Heart10/38 (26)89/403 (22)NS
 Kidney4 (8)36 (7)NS

Pulmonary fibrosis was more frequent in African American patients who were positive for anti–topo I antibodies than in Caucasian patients with these antibodies (72% versus 52%; P = 0.0135), with a significantly greater frequency of severe pulmonary fibrosis in African Americans compared with Caucasians (44% versus 18%; P = 0.0001), despite the fact that there was no difference in disease duration at presentation. There were no other differences in the frequencies of severe organ system involvement between the 2 groups.

After adjustment for age, sex, and diffuse disease, African American anti–topo I–positive patients were 75% more likely to have died at 5 years of followup than were Caucasian anti–topo I–positive patients (hazard ratio 1.75 [95% CI 1.06–2.88], P = 0.03), a finding that is consistent with the presence of more severe lung disease in the African American patients.

Anti–U3 RNP–antibody–positive patients.

Twenty-four African American patients with SSc and 61 Caucasian patients with SSc who were positive for anti–U3 RNP antibodies were compared (Table 5). There were no age, sex, or disease duration differences between the 2 groups. African American anti–U3 RNP antibody–positive patients more frequently had dcSSc compared to Caucasian patients with these antibodies (71% versus 44%; P = 0.05), and, in accordance with this finding, the maximum modified Rodnan total skin thickness score was higher in African Americans, although the difference was not statistically significant. Severe GI involvement was more common in African Americans than Caucasians with anti–U3 RNP antibodies. This latter finding is consistent with the findings from a prior study of Japanese patients, and the anti–U3 RNP antibody is the only serum antibody marker that has been found to be associated with GI disease (23).

Table 5. Demographic characteristics, disease classification features, and frequency of severe organ system involvement in patients with systemic sclerosis who were anti–U3 RNP antibody positive and were first evaluated during 1972–2007*
FeatureAfrican American (n = 24)Caucasian (n = 61)P
  • *

    Except where indicated otherwise, values are the number (%) of patients. NS = not significant; HAQ = Health Assessment Questionnaire; TSS = total skin thickness score; GI = gastrointestinal; PAH = pulmonary arterial hypertension.

  • Values for these organ systems are the number/total number of patients, representing the proportion of patients with severe organ system involvement among those for whom there were objective data that permitted assessment of severity. For all other organ systems, the denominator was all patients.

Sex, female17 (71)42 (69)NS
Age at symptom onset, mean ± SD years38.5 ± 14.639.9 ± 15.4NS
Disease duration from onset to diagnosis, mean ± SD years2.8 ± 4.22.7 ± 4.3NS
HAQ disability index, mean ± SD maximum score1.6 ± 0.71.2 ± 0.8NS
Diffuse cutaneous involvement17 (71)27 (44)0.05
Organ system with severe involvement   
 Skin5 (21)6 (10)NS
  Mean ± SD maximum modified Rodnan TSS17.4 ± 10.413.3 ± 11.5NS
 Vascular12 (50)26 (43)NS
 Joints/tendons6 (25)8 (13)NS
 Skeletal muscles2 (8)2 (3)NS
 GI tract6/19 (32)4/46 (9)0.05
 Pulmonary fibrosis3/21 (14)5/54 (9)NS
 PAH4/18 (22)14/47 (30)NS
 Heart4/22 (18)9/54 (17)NS
 Kidney4 (17)0 (0)0.001

Unlike that in anti–topo I antibody–positive patients, the frequency of any pulmonary fibrosis or of severe pulmonary fibrosis was not different between African American and Caucasian patients. PAH was increased in frequency in both African American anti–U3 RNP antibody–positive patients (22%) and Caucasian anti–U3 RNP antibody–positive patients (30%) when compared to all patients (11%). This is particularly important since a majority of the African American patients had diffuse cutaneous scleroderma.

Renal crisis was more frequent overall among African Americans than among Caucasians with anti–U3 RNP antibodies (17% versus 0%), but this is likely related to the higher frequency of dcSSc among African American patients. After adjustment for age, sex, and diffuse disease, there was no increased risk of death at 5 years of followup between African American and Caucasian anti–U3 RNP–positive patients.

Anti–U1 RNP antibody–positive patients.

In 30 African American patients with SSc and 148 Caucasian patients with SSc who were positive for anti–U1 RNP antibodies, there were no demographic differences between the groups (Table 6). Most patients had limited cutaneous disease (67% of African Americans versus 80% of Caucasians). Inflammatory muscle involvement was common in both African Americans and Caucasians (each 30%). Similar to those in the anti–topo I antibody–positive patients, the frequencies of any pulmonary fibrosis (74% versus 31%; P = 0.0175) and severe pulmonary fibrosis (52% versus 11%; P = 0.0001) were significantly higher in African American anti–U1 RNP–positive patients than in Caucasian anti–U1 RNP–positive patients (Table 6). Five-year survival was no different between the groups after adjustment for age, sex, and diffuse disease.

Table 6. Demographic characteristics, disease classification features, and frequency of severe organ system involvement in patients with systemic sclerosis who were anti–U1 RNP antibody–positive and were first evaluated during 1972–2007*
FeatureAfrican American (n = 30)Caucasian (n = 148)P
  • *

    Except where indicated otherwise, values are the number (%) of patients. NS = not significant; HAQ = Health Assessment Questionnaire; TSS = total skin thickness score; GI = gastrointestinal; PAH = pulmonary arterial hypertension.

  • Values for these organ systems are the number/total number of patients, representing the proportion of patients with severe organ system involvement among those for whom there were objective data that permitted assessment of severity. For all other organ systems, the denominator was all patients.

Sex, female23 (77)122 (82)NS
Age at onset, mean ± SD years34.2 ± 12.336.7 ± 14.1NS
Disease duration from onset to diagnosis, mean ± SD years3.0 ± 6.93.9 ± 5.7NS
HAQ disability index, mean ± SD maximum score1.1 ± 0.91.2 ± 0.8NS
Diffuse cutaneous involvement10 (33)29 (20)NS
Organ system with severe involvement   
 Skin1 (3)8 (5)NS
  Mean ± SD maximum modified Rodnan TSS8.5 ± 9.08.5 ± 9.9NS
 Vascular10 (33)60 (40)NS
 Joints/tendons3 (10)21 (14)NS
 Skeletal muscles1 (3)4 (3)NS
 GI tract1/24 (4)14/110 (13)NS
 Pulmonary fibrosis12/23 (52)14/122 (11)0.0001
 PAH1/21 (5)11/87 (13)NS
 Heart4/24 (17)20/121 (16)NS
 Kidney0 (0)4 (3)NS

Effect of tobacco use.

To determine whether there was any effect of cigarette smoking on the severity of pulmonary fibrosis, we examined smoking histories in the African American and Caucasian patients. African Americans were more likely to be current smokers (26% versus 16%; P < 0.0002). However, among the smokers, 30% of African Americans had severe pulmonary fibrosis, compared with 13% of Caucasians (P = 0.0001). Similarly, 29% of African American lifelong nonsmokers compared to 10% of Caucasian lifelong nonsmokers had severe pulmonary fibrosis (P = 0.0001). In multivariable logistic analyses, after adjustment for age, sex, and tobacco use, African American patients were more likely than Caucasian patients to develop pulmonary fibrosis (odds ratio 3.04 [95% CI 2.23–4.13], P = 0.008).

DISCUSSION

Both the incidence and the prevalence of SSc are greater in African Americans than in Caucasians (1, 2). In a comprehensive epidemiologic study of women who were residents of Michigan between 1980 and 1991, the annual incidence of SSc was 22.5 per million population at risk in African Americans compared with 12.8 per million in Caucasians (2). Prior studies have shown that African American patients with SSc are younger at disease onset, are diagnosed at an earlier age, and more frequently have diffuse scleroderma (2, 7–9, 24) and interstitial lung disease (3, 10).

Although in the current study the group of all African American patients combined was younger at the onset of disease than was the group of all Caucasian patients, the age difference was only significant for 1 of the 3 autoantibody subsets, anti–topo I. Among all of the patients, African Americans had a slightly shorter disease duration, suggesting that delay in diagnosis is unlikely to be a cause for differences in outcomes. Also, the highest-attained education level was not different between our 2 patient groups. Therefore, socioeconomic status was not a factor in this cohort. Of note, however, the mean number of years of education reported by our African American patients with SSc (13 years) was considerably higher than has been reported in some studies of patients with systemic lupus erythematosus in the inner city, where socioeconomic status seemed to play a greater role in outcome (25). Within the specific autoantibody analyses, there were no differences in either of these features.

Even though the socioeconomic status may have been similar among these patients, there may have been differences in their access to care and in their ability or desire to receive aggressive treatment. This study was not attempting to define response to therapy. However, there were no differences in the frequency of the use of cyclophosphamide (or other immunosuppressive medications) or in the use of D-penicillamine between the African American and Caucasian patients. This finding at least suggests that both groups had a similar pattern of use of aggressive treatments.

The proportion of patients with diffuse skin changes was greater among African American patients than among Caucasian patients (50% versus 43%; P = 0.03). However, within the specific antibody subsets, there was only a marginally significant difference in the frequency of dcSSc between groups, and the only significant difference was found in the anti–U3 RNP–positive patients. The mean maximum modified Rodnan total skin thickness score and the frequency of severe skin involvement were similar in all 3 antibody groups. This suggests that the increased frequency of diffuse disease primarily reflects the difference in antibody profiles between the races.

Among SSc patients overall, the frequency of skeletal muscle involvement was significantly greater in African American patients compared to Caucasian patients, but severe skeletal muscle involvement was infrequent in both groups. In addition, the frequency of any GI tract involvement was slightly higher overall in African Americans, but this is likely due to the differences in GI involvement among patients with anti–U3 RNP antibodies. Severe GI involvement was strikingly more common in African Americans than Caucasians with anti–U3 RNP antibodies (32% versus 9%; P = 0.05). This subset of anti–U3 RNP–positive SSc patients with severe GI disease has been identified previously in a study of Japanese patients with SSc (23), as well as in a study of African American scleroderma patients with antifibrillarin antibodies compared to African American patients without antifibrillarin antibodies (14) and in male African American patients with SSc who had severe small bowel involvement and ANAs showing a nucleolar staining pattern (26). It is likely that many of the latter patients had anti–U3 RNP antibodies, but this testing was not available to the authors.

Multiple studies have consistently shown that African American patients with scleroderma have more frequent and more severe pulmonary fibrosis than has been seen in Caucasians. This study is the first to document that the frequency of severe pulmonary fibrosis was higher in African Americans than in Caucasians even within some autoantibody subsets, particularly the subsets of patients who were positive for anti–topo I and anti–U1 RNP, both of which are intrinsically associated with pulmonary fibrosis. Interestingly, this was not seen in patients with anti–U3 RNP, an autoantibody profile in which severe pulmonary fibrosis is less common.

An increased frequency of PAH among anti– U3 RNP–positive patients was found in both African Americans and Caucasians. Greidinger et al previously found more severe lung disease in African American and anti–topo I–positive patients (3). In our analysis, African American race appeared to be independently associated with the development of severe pulmonary fibrosis. Thus, African Americans more frequently have antibodies associated with lung disease, but also are at greater risk for severe lung disease, possibly because of genetic or unrecognized environmental factors. Recent studies have shown that there are ethnic differences in antifibrotic hepatocyte growth factor expression in lung fibroblasts between Caucasian and African American subjects. Reduced levels of hepatocyte growth factor as well as a deficiency in the c-Met receptor function appear to be present in African American patients with SSc (27). These findings may explain, in part, the greater disease severity and worse prognosis observed in African Americans with SSc.

Results of several studies have suggested that African Americans with SSc have a worse survival rate, potentially related to ethnic disparities or quality of care (4, 28). Age, SSc subtype, autoantibodies, and socioeconomic status significantly influence these findings. A study by Laing et al showed an increase in mortality in African Americans with SSc, which was significant after adjustment for age and disease subtype (2). Although we revealed a significant decrease in survival when all African Americans were compared to all Caucasians, this was found to be present only in the anti–topo I–positive patients, in whom increased severity of pulmonary fibrosis was so prevalent. There was no difference in survival between African American and Caucasian patients with anti–U3 RNP antibodies or anti–U1 RNP antibodies. Thus, the overall prognosis seems to be most closely related to autoantibody type and the presence of pulmonary fibrosis, rather than to a more generalized racial effect. Survival differences seen in other studies may be attributed to the fact that African Americans have a higher frequency of these 3 antibodies compared to Caucasians; the frequency was 83% in African Americans compared to 30% in Caucasians in our 1984–1989 series of patients with all antibodies tested (15).

We have shown that there are major differences between African Americans and Caucasians with respect to serologic and clinical phenotypes of SSc. Although this study was not specifically designed to carefully investigate socioeconomic status or delay-of-treatment differences, overall there were no differences in education level (an often-used surrogate for socioeconomic status) or in disease duration at the time of first evaluation or first diagnosis in these different ethnic groups. We also showed that there were no differences in the overall frequency of use of immunosuppressive agents between African Americans and Caucasians, and therefore access to care seems to be a less likely explanation for our results.

We believe that most of the differences in organ system outcomes appear to be attributable to the different patterns of SSc-associated autoantibodies observed in African American patients. However, there were also differences even within the antibody subsets. African American patients with anti–topo I and anti–U1 RNP antibodies clearly had more severe pulmonary fibrosis than did Caucasian patients with these same antibodies, and severe GI disease was more frequent in African Americans with anti–U3 RNP.

Our study is limited in that we could only observe differences in disease severity and outcomes. Whether these are related to differences in environmental exposures or to varying genetic backgrounds remain to be seen. These findings have important implications for both clinical interventions and future studies of the pathogenesis of SSc.

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. Steen 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. Steen, Medsger.

Acquisition of data. Steen, Lucas, Fertig, Medsger.

Analysis and interpretation of data. Steen, Domsic, Lucas, Fertig, Medsger.

Ancillary