The Prevalence of Peripheral Arterial Disease in Patients With Heart Failure by Race and Ethnicity
Kathy Hebert, MD, MMM, MPH, Miller School of Medicine, University of Miami, 1611 NW 12th Avenue, Jackson Memorial Hospital North Wing 210, Miami, FL 33136
More than 5 million people live with heart failure (HF) in the United States, and this number is expected to rise due to several factors including increased life expectancy brought about by medical therapy and the aging of the population. HF and peripheral arterial disease (PAD) share many risk factors. A review of the literature reveals several studies supporting a higher prevalence of HF in patients with PAD than in those without PAD. However, no study was found that estimates the prevalence of PAD in patients with HF. Moreover, the prevalence of PAD by US race/ethnic groups with HF has not been studied. The authors conducted a cross-sectional multicenter study of patients enrolled in an HF disease management program in Louisiana (n=330) and Florida (n=464). All patients with an ejection fraction ≤40% and a measured ankle-brachial index (ABI) were included in the study. PAD was defined as an ABI <0.9. The overall prevalence of PAD was 17.1%. The prevalence of PAD was 25.9% for white, 13.4% for Hispanic, and 13.7% for black patients. White patients had a higher prevalence of PAD than black or Hispanic patients (P<.001). Routine ABI measurements in these groups would enhance efforts to detect subclinical PAD. Congest Heart Fail. 2010;16:118–121. © 2010 Wiley Periodicals, Inc.
Congest Heart Fail. © 2010 Wiley Periodicals, Inc.
The expected prevalence of peripheral arterial disease (PAD) is 4.3% as estimated by the 1999–2000 National Health and Nutrition Examination Survey (NHANES), which included adults aged 40 and older,1 and the prevalence has been shown to vary by race, with a higher prevalence among blacks than whites.2–4 Results from studies calculating the prevalence of PAD by race and ethnicity report that the prevalence of PAD is higher in blacks than in whites after adjusting for conventional risk factors.2–4 Studies that have included Hispanic persons have had more mixed results. One study found that PAD is more prevalent in Hispanic persons (5.1%) compared to whites.4 These data from the 1999 to 2002 NHANES showed that the prevalence of PAD was 7.8% in blacks, 3.4% in whites, and 5.1% in Mexican Americans.1,4 Results of the prevalence of PAD in a racially diverse population study showed that the prevalence of PAD was 13.2% among whites, 22.8% among blacks, and 13.7% among Hispanic patients.5 Calculating the prevalence of PAD is important because many studies have shown that PAD is associated with an increased risk of cardiovascular death.6,7 After adjusting for conventional risk factors, the relative risk of dying for patients with PAD compared with patients without PAD was 3.1 from all-cause mortality, 5.9 from cardiovascular disease, and 6.6 from coronary artery disease (CAD).7
It has been shown that the presence of PAD is associated with a 2-fold increase in the prevalence of heart failure (HF).8 Given that HF and PAD share many risk factors, including increased age, diabetes, tobacco use, atherosclerosis, and poor renal function,1,9 it is not surprising that the prevalence of HF in patients with PAD ranges from 5.3% to 13.9%.8 The association between PAD and cardiovascular disease and the increasing prevalence of HF in industrialized countries suggests that PAD is an important medical concern in patients with HF.
Despite the strong association between PAD and HF, little information is known about the prevalence of PAD among patients with HF and no study was found that has calculated the prevalence of PAD in patients with HF by race and ethnicity. Given that the prevalence of PAD has been found to vary among patients of different ethnic minority populations, it is possible that ethnic/race variations exist in the prevalence of PAD in HF patients. The aim of this study is to determine the prevalence of PAD in HF patients across race and ethnicity. Specifically, we will compare the prevalence of PAD among white, black, and Hispanic patients enrolled in HF disease management programs.
Study Population. We conducted a cross-sectional multicenter study of patients enrolled in an HF disease management program (HFDMP) in Louisiana (n=330) and Florida (n=464). Inclusion criteria included age 18 or older and an ejection fraction ≤40% by echocardiography. Study recruitment took place from August 1999 to December 2007 at Leonard J. Chabert Medical Center in Houma, Louisiana, which is a rural safety net hospital that provides health care primarily to the uninsured and underinsured. In 2006, 7.9% of claims were from Medicare, 21.6% of claims were from Medicaid, and 50.5% were for uncompensated care.10
At the Florida site, enrollment was from October 2007 through November 2008 at the HFDMP in Jackson Memorial Hospital (JMH). In 2007, JMH spent $500 million on charity care. In 2005, 14.4% of its revenue was from Medicaid and 17.8% of its revenue was from Medicare.11
All patients at each site provided informed consent to enroll in an electronic data registry. The protocol was approved by the institutional review boards of the Miller School of Medicine at the University of Miami and Louisiana State University Health Sciences Center.
Measures. General demographic features of the study population were obtained, such as age, sex, comorbidities (hypertension, diabetes, tobacco use, ischemia), New York Heart Association (NYHA) functional classification, ejection fraction, and medications.
Systolic blood pressures were measured in the upper extremities and both ankles. The ankle brachial index (ABI) was calculated separately for each ankle. The ABI that was used was the lower of the right vs left leg. We considered PAD to be present if the ABI was ≤0.9. Studies that have evaluated the diagnostic accuracy of ABI have demonstrated that the ABI can differentiate arteriographically diseased vs normal limbs with a sensitivity of 97% and a specificity of 100%.12
All patients had the most likely etiology of HF determined by the site investigator’s opinion. The choices were ischemic cardiomyopathy (ICM) or dilated cardiomyopathy (DCM). Patients were determined to have ICM if they had (1) a prior cardiac catheterization that showed at least 1 vessel with at least 50% stenosis; (2) an abnormal result on noninvasive nuclear stress testing, dobutamine stress echocardiography, or stress echocardiography; or (3) a documented myocardial infarction. If patients did not meet these criteria, they were classified as having DCM.
Statistical Analysis. The overall PAD prevalence and prevalence by ethnicity/race were calculated. Ethnic and race differences in PAD prevalence were analyzed using chi-square test. Demographic and other clinical characteristics were compared between patients with PAD and those without PAD. For categorical variables, a chi-square test was used to evaluate group differences. For continuous variables, analysis of variance was used to evaluate group differences. A P value <.05 was considered statistically significant. All analyses were conducted using SPSS 16.0 (SPSS, Inc., Chicago, IL).
The PAD prevalence rates by race/ethnic groups are presented in Table I. Overall, 17.1% of patients had PAD. White patients had a significantly higher prevalence of PAD (25.9%) compared to black (13.75%) and Hispanic (13.4%) patients (P<.001).
Table I. Prevalence of PAD by Race and Ethnicity
|PAD, No. (%)||60 (25.9%)a,b||36 (13.4%)a||40 (13.7%)b||136 (17.1%)|
Demographic and clinical characteristics by PAD are presented in Table II. PAD patients were significantly older than non-PAD patients (P<.001). PAD patients had significantly higher ejection fractions (P=.004). PAD patients had significantly higher rates of ICM and lower rates of DCM (P<.001). PAD patients also had significantly lower diastolic blood pressure and body mass index than non-PAD patients (P=.006 and P=.022, respectively).
Table II. Demographic and Clinical Characteristics of Patients by PAD Status
| Female, No. (%)||218 (33.2)||48 (35.3)||266 (33.5)||.635|
| Male, No. (%)||439 (66.8)||88 (64.7)||527 (66.5)|
|Age, y||55.6 (11.5)||61.2 (10.9)||56.5 (11.6)||<.001|
|BMI, kg/m2||31.0 (7.9)||29.1 (5.9)||30.7 (7.7)||.022|
|Tobacco use (lifetime), No. (%)||183 (28.6)||40 (29.4)||223 (28.7)||.848|
|Medication, No. (%)|
| ACEI||562 (85.9)||110 (80.9)||672 (85.1)||.133|
| ARB||36 (5.5)||10 (7.4)||46 (5.8)||.393|
| β-Blocker||599 (91.7)||129 (94.9)||728 (92.3)||.215|
| Spironolactone||125 (28.3)||19 (29.2)||144 (28.5)||.882|
|NYHA class, No. (%)||.500|
| I||154 (24.6)||28 (21.4)||182 (24.0)|
| II||215 (34.3)||42 (32.1)||257 (33.9)|
| III||174 (27.8)||45 (34.4)||219 (28.9)|
| IV||84 (13.4)||16 (12.2)||100 (13.2)|
|ICM, No. (%)||195 (30.6)||65 (49.2)||260 (33.8)||<.001|
|DCM, No. (%)||411 (64.4)||54 (40.9)||465 (60.4)||<.001|
|Systolic BP, mm Hg||132.3 (25.8)||132.6 (24.5)||132.4 (25.5)||.903|
|Diastolic BP, mm Hg||82.3 (17.7)||77.8 (15.0)||81.5 (17.4)||.006|
|Diabetes, No. (%)||204 (31.0)||48 (35.3)||252 (31.7)||.328|
|HDL-C, mg/100 mL ||43.1 (19.0)||40.8 (13.2)||42.7 (18.1)||.288|
|LDL-C, mg/100 mL ||101.3 (38.9)||103.5 (37.8)||101.7 (38.7)||.634|
|Ejection fraction, %||26.6 (10.1)||29.4 (10.6)||27.1 (10.2)||.004|
|Atrial fibrillation, No. (%)||44 (6.7)||8 (5.9)||52 (6.5)||.730|
No significant group differences were found for rates of NYHA classification; use of angiotensin receptor blockers, angiotensin-converting enzyme inhibitors, β-blockers, and spironolactone medications; low-density lipoprotein cholesterol level, high-density lipoprotein cholesterol level, atrial fibrillation rate, diabetes rate, systolic blood pressure, and lifetime tobacco use rate.
This study examined the prevalence of PAD across 3 race/ethnic groups of HF patients. We found that the prevalence of PAD was 17.1% in the total sample. To our knowledge, this is the first study to report the prevalence of PAD in HF patients. Moreover, to our knowledge, this is the first study to report the prevalence of PAD in HF patients across race/ethnic groups.
The high prevalence of PAD in our sample of HF patients is expected given that similar risk factors (increased age, diabetes, tobacco use, atherosclerosis, and poor renal function) cluster in patients with HF and PAD. In addition, prior studies8,13,14 have shown an increased prevalence (5.3%–13.9%) of HF in patients with PAD. Our results further support an association between PAD and HF, by being the first to report the high prevalence of PAD in an HF population. This suggests that obtaining routine ABI measurements in HF patients may enhance efforts to detect PAD in this population.
In contrast to previous studies,2–5 we found that the prevalence rate of PVD in white patients was higher than in black and Hispanic patients in our sample, although this result was only statistically significant between white and Hispanic patients. The lack of concordance between our results and those of previous studies may be attributed to differences in ICM rates between the patient groups in our study. In this study, only 20.0% of black patients had ICM, while 38.7% of white patients and 44.7% of Hispanic patients had ICM. While no study has been done to evaluate the prevalence of PAD in patients with ICM, the correlation between CAD and PAD has been well-established.9,15 Past research has shown that patients with CAD have a PAD prevalence of 24%, higher than the general population.9 Because by definition16 patients with ICM have CAD, a similarly high prevalence of PAD would be expected in this population. Furthermore, the studies8,13,14 that have shown an increased prevalence of HF in patients with PAD did not report the etiology of HF (ICM or DCM). Thus, it is unknown whether the association between HF and PAD is attributed to the prevalence of CAD in patients with ICM or whether patients with DCM also have an increased prevalence of PAD. Further studies are needed to determine how the prevalence of PVD in HF patients differs by etiology of HF (ICM or DCM).
Past studies have shown that the difference in prevalence of PAD between white and black patients was partially attenuated when adjusted for conventional risk factors, such as smoking, diabetes, hypertension, hyperlipidemia, and increasing age.2,17 When novel risk factors such as higher levels of C-reactive protein,1 fibrinogen,1 homocysteine,18 and lipoprotein(a)19 were taken into account, the different prevalences between black and white patients were further attenuated. In our sample, white patients had a significantly higher prevalence of diabetes (80.0%) than did black (29.3%) and Hispanic (47.6%) patients. In addition, the black patients were slightly younger. Thus, the increased prevalence of PAD in our sample of white patients with HF when compared to black and Hispanic HF patients could be partially attributed to the significantly different prevalence of ICM, prevalence of diabetes, and age.
The research on the prevalence of PAD in Hispanic populations has shown mixed results. While some studies have shown that Hispanic persons have a higher prevalence of PAD than whites,4,5 others have shown prevalence rates either lower than or similar to the prevalence rates of PAD in whites.1,7,20 Prior general population studies have reported prevalence of PAD in Hispanic patients as 1.8% to 13.7%.5,20 Moreover, one study showed that Hispanic persons had lower odds (0.45) of PAD when compared to whites when “novel” and traditional risk factors were adjusted.21 This result is consistent with our result of a higher prevalence of PAD in white patients than in Hispanic patients with HF.
The limitations of this study are that the populations were drawn from large safety net hospitals that serve primarily indigent patients and patients from low socioeconomic backgrounds, so it is unclear whether results from this study would generalize to white, black, and Hispanic persons in the general population.
Second, Hispanic patients in this study were primarily of Cuban descent. Although Cubans represent a majority of the Hispanic population in Miami, Mexican Americans represent the majority of Hispanic persons in the United States. This may represent a threat to the generalizability of these results to other Hispanic populations. As a result, replication of these results in a Mexican American population is suggested.
Third, we did not assess C-reactive protein,1 fibrinogen,1 homocysteine,18 or lipoprotein(a)19 in this sample. Given the importance of these novel risk factors for increasing risk for PAD, it would be important to assess their role across race and ethnicity.
The prevalence of PAD in HF patients is high among patients of Hispanic, black, and white ethnicity/race. Routine ABI measurements in these populations would enhance efforts to detect PAD.