If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Prevalence and outcomes of low ankle brachial index by atherosclerotic cardiovascular disease risk level: Insights from the National Health and Nutrition Examination Survey (NHANES)
1 Dr. Singla and Dr. Chobufo contributed equally to the manuscript.
Atul Singla
Correspondence
Corresponding author. Atul Singla, MD, FACC, Assistant Professor of Medicine, Department of Medicine, Tulane University School of Medicine and Southeast Louisiana Veterans Health Care System, New Orleans, LA 70112.
Ankle brachial index (ABI) as a risk-enhancing factor in addition to the pooled cohort equation (PCE) in assessing cardiovascular risk for primary prevention of atherosclerotic cardiovascular disease (ASCVD) is uncertain.
Methods
We analyzed data from the 1999–2004 National Health and Nutrition Examination Survey (NHANES), for 5130 participants, aged 40 and older, without known cardiovascular disease or diabetes, with available data on standard ASCVD risk and ABI. Prevalence of low ABI (ABI<0.9) and all-cause mortality in persons with low, borderline and intermediate ASCVD risk categories using PCE was assessed.
Results
The overall prevalence of low ABI was 3.1%. The participants with low ABI were predominantly clustered in the intermediate (33%) and high (33%) ASCVD risk categories while most participants with a normal ABI were in the low (56%) and intermediate (23%) risk categories. All-cause mortality was higher among participants with low ABI compared to those with a normal ABI in both the intermediate/borderline and high-risk categories, p<0.001 but not in the low-risk ASCVD category, p = 0.323.
Conclusions
Using the PCE, two-third of the participants with low ABI were classified as having a low, borderline or intermediate risk of ASCVD. Low ABI was associated with an increased all-cause mortality in the overall cohort and specifically among those with a borderline/intermediate or high risk of ASCVD but not in those with a low risk of ASCVD. Our study supports consideration of ABI as a risk enhancer for primary prevention among patients classified as borderline or intermediate risk of ASCVD.
Primary prevention begins with the assessment of ASCVD risk. All risk estimation tools have inherent limitations, and population-based risk equations must be interpreted in light of specific circumstances for individual patients. Use of the sex- and race-specific pooled cohort equations (PCE) to assess the 10-year risk of a first ASCVD event is recommended as an initial step in risk assessment for adults aged 40–79 years of age and in identifying individuals eligible for primary prevention therapies.
2019 ACC/AHA guideline on the primary prevention of cardiovascular disease: a report of the American College of Cardiology/American Heart Association task force on clinical practice guidelines.
The PCE have been shown to overestimate or underestimate ASCVD risk for certain subgroups. Thus, after calculation of the PCE, it is recommended to use additional risk-enhancing factors to guide shared decision making about preventive interventions, particularly for borderline- or intermediate-risk adults, one of which is the ankle brachial index (ABI). Low ABI (< 0.9) has been associated with cardiovascular events and mortality
and can be helpful for shared decision making regarding initiation of statin therapy. However, there are no data on the prevalence of low ABI among individuals with ASCVD in the borderline or intermediate risk categories. Furthermore, no data exist regarding mortality differences between normal and low ABI groups among individuals in the borderline and intermediate ASCVD risk categories. To address this gap in the literature, we set out to assess the prevalence of low ABI and all-cause mortality in persons with low, borderline and intermediate ASCVD risk using data collected from the National Health and Nutrition Examination Survey (NHANES) linked to mortality data.
Methods
Survey design
NHANES is a large, multicenter survey conducted by the National Center for Health Statistics of the Centers for Disease Control and Prevention, which collects nationally representative data on the health and nutritional status of the non-institutionalized US population. NHANES utilizes a multistage probability sampling design and enrolls approximately 5000 persons per year. Each survey cycle spans two years. The survey includes questionnaire, physical exam, and laboratory evaluation components. The survey oversamples certain minority groups to provide adequate statistical power to study these groups. NHANES protocols are approved by the National Center for Health Statistics Institutional Review Board, and informed consent was obtained from all participants. Detailed information on the survey design is available from the survey documentation. The survey instruments, physical examination, and laboratory measurements have been previously described in detail.
The present study was based on NHANES data collected from 1999 to 2004. Survey participants were interviewed in their homes to ascertain sociodemographic characteristics (age, gender, level of education, ethnicity, marital status, income, place of birth, health insurance, and smoking status) using a Computer-Assisted Personal Interviewing system (i.e., interviewer-administered). Participants who reported smoking ≥100 cigarettes including smoking during the preceding 30 days were classified as current smokers, those who reported smoking ≥100 cigarettes but no smoking during the preceding 30 days were classified as prior smokers, and together, current and prior smokers were classified as “ever smokers.” Body mass index (BMI) was computed from each participant's recorded height and weight obtained at their mobile examination center (MEC) examination by trained health technicians. The family poverty index ratio (PIR) was calculated by dividing the total family income by the contemporary poverty threshold, as defined by the US census bureau, with adjustment for family size at the time of interview. Family PIR was grouped into two categories (PIR <1.00–2.99 and PIR ≥3.00).
Use of antihypertensive or antihyperglycemic medications were self-reported. At MEC visits, up to four separate blood pressure measurements were obtained on two separate occasions using a standardized protocol, and the mean value was computed for analysis. Hypertension was defined as current use of antihypertensive medications, a mean diastolic blood pressure ≥ 90 mm Hg, or a mean systolic blood pressure ≥140 mm Hg. Persons with a self-reported diagnosis of diabetes or with a hemoglobin A1C ≥6.5% were classified as having diabetes. Other co-morbidities assessed included established cardiovascular disease. Respondents who replied “yes” to question “has a doctor or other health professional ever told you that you had coronary heart disease, angina, heart attack, or stroke?” were classified as having established cardiovascular disease. Laboratory tests of interest included hemoglobin A1c, high density lipoprotein cholesterol (HDL-C), total cholesterol, low-density lipoprotein cholesterol (LDL-C), and triglycerides. We computed the 10-year ASCVD risk using the PCE equations and categorized individuals as low-risk (<5%), borderline risk (5% to <7.5%), intermediate risk (7.5% to <20%), and high risk (≥20%).
2019 ACC/AHA guideline on the primary prevention of cardiovascular disease: a report of the American College of Cardiology/American Heart Association task force on clinical practice guidelines.
All participants 40 years of age and older were invited to undergo ABI examination. Systolic blood pressure was measured in the right arm (brachial artery) and both ankles (posterior tibial arteries) using a Doppler ultrasound. The ABI was calculated by dividing the systolic blood pressure in the right arm by the mean systolic blood pressure in the ankle. If the participant had a rash or open wound, dialysis shunt, right-sided radical mastectomy, or any other right arm condition that would interfere with accurate measurement or would cause discomfort to the participant, the left arm was used for the brachial pressure measurement. Systolic blood pressure was measured twice at each site for participants aged 40–59 years and once at each site for participants aged 60 years and older. ABI data for the right and left ankles were available in the NHANES dataset as pre-recorded variables, and the lower value of the two recordings was used for our analysis. We considered an ABI <0.9 in either leg to be abnormal/low.
Participants 40 −75 years old with an ABI ≤1.3 who were not pregnant at time of exam and had no history of established cardiovascular disease (stroke, myocardial infarction, and coronary heart disease) were included in the analysis.
Follow-up
Vital status (dead or alive) was determined by linkage to a record match of the National Death Index, conducted by the National Center for Health Statistics through December 31, 2015.
Center for Health Statistics N. The Linkage of National Center For Health Statistics Survey Data to the National Death Index –2015 Linked Mortality File (LMF): Methodology Overview and Analytic Considerations. n.d.
Duration of follow-up was calculated from the date of the examination to either the date of death or December 31, 2015, which ever came first.
Statistical analysis
Questionnaire data files with variables of interest were linked with demographic information and mortality files. Analytical weights for the combined dataset were computed in keeping with the stipulated guidelines to ensure the estimates would be representative of the non-institutionalized US population aged 40–75 years old. A composite code for computing the 10-year ASCVD risk scores using the PCE was developed and used for the analysis. The accuracy of the code was assessed by applying it to selected examples provided by authors of the PCE.
Categorical variables are reported as percentages (95% CI) and continuous variables are reported as means ± standard deviations. Associations between participant characteristics and ABI categories were assessed using the equality of the mean for continuous variables and Chi square tests for categorical variables. Participants in the low, borderline and intermediate ASCVD risk categories with low ABI were reclassified as being at high risk and the relative percent change in high-risk reclassification computed. Hazard ratios (HRs) for all-cause mortality and mortality from cardiac or cerebrovascular disease were assessed using Cox proportional hazards models. The Cox models included variables with a significant association (p<0.05) with ABI categories in univariate analysis. Variables used in computing ASCVD risk scores were excluded from logistic analysis as these variables were already accounted for in the ASCVD scores. Kaplan-Meier (KM) curves for all-cause mortality, stratified by ABI categories, were generated for the general study population as well as for the different ASCVD risk categories. The log rank test was used to compare KM mortality curves. Individuals with missing data were excluded from relevant analyses. All analyses were conducted using STATA version 16 (StataCorp, College Station, TX) with p-values <0.05 considered statistically significant.
Results
A total of 31,126 individuals participated in the NHANES survey from 1999 to 2004, of whom 7898 were in the age group 40–75 years. After excluding participants with missing ABI values or an ABI >1.3, established cardiovascular disease, pregnancy at the time of their examination, or missing ASCVD risk scores, 5130 individuals remained eligible for analysis (weighted to represent 78 million Americans). A flowchart for the selection of the final study cohort is shown in Fig. 1.
Fig. 1Flowchart to show the derivation of study cohort.
Baseline characteristics of the study participants, stratified by ABI (normal vs. low), are presented in Table 1. The overall prevalence of low ABI was 3.1%. Participants with low ABI were older (mean age 61.4 years [95% CI 59.6–63.1]), in comparison to those with a normal ABI (mean age 52.9 years [95% CI 52.6–53.3], p = 0.001). Mean systolic blood pressure was higher (126.2 [95% CI 125.4–127.0]) vs 135.2 [95% CI 131.1–139.2)]) while mean diastolic blood pressure was lower (74.9 [95% CI 74.4–75.5] vs 69.8 [95% CI 67.3–72.4)]) among participants with low ABI compared to those with a normal ABI. The percentage of women and smokers was significantly higher in those with low ABI compared to those with normal ABI. Comparison of the two groups demonstrated no difference in mean total cholesterol, mg/dl (212 [95% CI 210–214] vs 208 [95% CI 202–215)]), LDL, mg/dl (128 [95% CI 125–130] vs 121 [95% CI 112–129)]), or HDL, mg/dl (54 [95% CI 53–54] vs 52 [95% CI 50–55]). The participants with low ABI were predominantly clustered in the intermediate (33%) and high (33%) ASCVD risk categories while most participants with a normal ABI were in the low (56%) and intermediate (23%) ASCVD risk categories (Table 1). Addition of ABI to PCE further reclassifies 38.5% of low, borderline and intermediate risk participants to high-risk category (Table 2).
Table 2Reclassification of cardiovascular risk with the addition of low ankle brachial index to atherosclerotic cardiovascular disease risk as estimated using pooled cohort equation.
The mean follow-up time was 156 months (95% CI 154 – 157) for the general cohort, 128 months (95% CI 121 – 136) for participants with abnormal ABI and 157 months (95% CI 155–159) for those with a normal ABI.
NHANES linkage to death files was successful for 99.8% of the study participants. All-cause mortality among participants with low ABI and a normal ABI was 43.1% (95% CI 35.3–51.3) and 12.2% (95% CI 10.8–13.6), respectively, p<0.001. Cardiac or cerebrovascular disease mortality was 8.4% (95% CI 5.1–13.7) and 1.9% (95% CI 1.6–2.4) in the low and normal ABI categories, respectively, p<0.001 (Table 3). After adjustment for potentially confounding variables in the Cox models, those with abnormal ABI were 2.22 (95% CI 1.57–3.14) times more likely to die from any cause and 3.02 (95% CI 1.60–5.69) times more likely to die from heart or cerebrovascular disease compared to those with a normal ABI (Table 4). Significant differences in all-cause mortality were also observed among participants with low ABI compared to those with a normal ABI in both the intermediate/borderline and high-risk categories of ASCVD, p<0.001 but not in the category of low ASCVD risk, p = 0.323 (Table 4). Kaplan-Meier survival curves for the overall cohort and ASCVD risk categories by PCE in participants stratified by ABI are shown in Fig. 2 (A-D).
Table 3All-cause and CVD-specific mortality in participants with a normal or low ABI, overall and in those with a low, borderline/intermediate, or high risk of ASCVD.
Fig. 2Kaplan-Meier survival curves for the general population and ASCVD risk categories by PCE in participants stratified by ABI category. Panel A: Overall cohort stratified as normal and low ABI. Panel B: Low ASCVD risk category with normal and low ABI (p = 0.277). Panel C: Borderline/intermediate ASCVD risk category with normal and low ABI p<0.001. Panel D: High ASCVD risk category with normal and low ABI p<0.001.
Among 5130 NHANES participants, we determined the prevalence of low ABI and of all-cause mortality in persons who had a low, borderline, or intermediate risk of ASCVD. We report two major findings. First, 4% of the participants with borderline/intermediate ASCVD risk had low ABI, and conversely, 43% of those with low ABI were categorized as having a borderline/intermediate risk of ASCVD, based on estimation using the PCE. Second, there was a significant difference in all–cause mortality between the normal and abnormal ABI groups in both the borderline/intermediate and high ASCVD risk categories but not in the low ASCVD risk category. Overall, those with low ABI had a 2-fold higher mortality compared with participants who had a normal ABI.
ABI in borderline/intermediate ASCVD risk
We found a substantial rate of abnormal ABI among participants with borderline/intermediate ASCVD risk. ASCVD remains the leading cause of morbidity and mortality globally.
Assessment of ASCVD risk remains the foundation of primary prevention. Estimating an individual's 10-year absolute ASCVD risk utilizing the PCE enables decision-making for many preventive interventions, including lipid and blood pressure management, in US adults. Statin treatment is recommended in US adults with clinical ASCVD or a high risk of ASCVD risk, whereas this is not recommended in those without known ASCVD and a low, borderline or intermediate risk of ASCVD. However, guidelines suggest using additional risk enhancing markers to improve cardiovascular risk estimation in adults at borderline- or intermediate-risk of ASCVD.
2019 ACC/AHA guideline on the primary prevention of cardiovascular disease: a report of the American College of Cardiology/American Heart Association task force on clinical practice guidelines.
U.S. Preventive Services Task Force Screening for peripheral artery disease and cardiovascular disease risk assessment with the ankle-brachial index in adults: U.S. Preventive Services Task Force recommendation statement.
A meta-analysis of sixteen cohort studies showed that inclusion of ABI in cardiovascular risk stratification using the FRS would result in reclassification of the risk category and modification of treatment recommendations in approximately 19% of men and 36% of women. A low ABI increased the 10-year hard coronary heart disease and cardiovascular mortality event rate by twofold in each FRS risk category.
Ankle Brachial Index Collaboration F Ankle brachial index combined with Framingham Risk Score to predict cardiovascular events and mortality: a meta-analysis.
In contrast, a sub-study of the Multi-Ethnic Study of Atherosclerosis (MESA) which assessed the value of adding nontraditional risk markers including coronary artery calcium, ABI, high-sensitivity C- reactive protein, and family history of ASCVD to ASCVD risk estimation using the PCE found each marker to be an independent predictor of ASCVD events.
The study also showed that coronary artery calcium scoring provided a modest improvement in discrimination, but the addition of ABI did not improve discrimination when used in addition to PCE. Similarly, an analysis of the Atherosclerosis Risk in Communities (ARIC) study concluded that although ABI adjusted for Framingham risk variables was independently associated with subsequent events in terms of hazard ratios, the independent effect of ABI when adjusted for Framingham Risk Score (FRS) was relatively small, and the FRS performed similarly with or without supplementation with ABI.
In our study, 24% and 43% of the participants in the low ABI cohort were categorized as having a low or borderline/intermediate risk of ASCVD by PCE, respectively. Addition of ABI to the PCE resulted in 2.5% absolute and 38.5% relative reclassification of the participants to higher risk category. Thus, our findings suggest that ABI can add precision to ASCVD risk classification and guide appropriate initiation of statin therapy.
ABI and mortality
We also found that ABI was associated with increased mortality in the participants with a high or borderline/intermediate risk of ASCVD but not in the low ASCVD risk participants. Several prior studies have identified an independent association between abnormal ABI and increased risk of cardiovascular events, including coronary artery disease, death, and all-cause mortality.
Our finding that low ABI is associated with both increased all-cause mortality and increased cardiovascular mortality in those with a borderline/intermediate risk of ASCVD could have significant implications for initiating primary prevention interventions at an earlier stage than the current practice in such individuals. Our findings are, in part, consistent with a study that showed increased mortality with abnormal ABI in each risk category using the FRS.
Ankle Brachial Index Collaboration F Ankle brachial index combined with Framingham Risk Score to predict cardiovascular events and mortality: a meta-analysis.
Contemporary US clinical practice guidelines recommend using the PCE rather than the FRS for estimation of ASCVD risk, because the PCE includes more variables, predicts both fatal and non-fatal cardiac and cerebrovascular events and has been validated in both US White and Black adults.
The findings from our study also support the current recommendation to perform ABI in adults who are at borderline or intermediate risk ASCVD as an additional risk enhancer.
The ABI is an inexpensive and simple test that requires no specialized equipment and can be performed in just a few minutes.
In our study, the mean age of the population with low ABI was 68 years and were older compared to normal ABI group, which is similar to that reported from other studies.
The impact of low ABI in asymptomatic individuals to detect early peripheral artery disease on either limb or cardiovascular outcomes or on survival has not yet been evaluated in prospective trials. A randomized controlled trial that evaluated the role of aspirin for prevention of CVD in adults with asymptomatic atherosclerosis did not document any MI, stroke, or coronary artery revascularization benefit during a mean follow-up of 8 years.
The Population Screening and Intervention for Vascular Disease in Danish Men (VIVA) randomized trial showed a 7% relative risk reduction in mortality in a population who were screened for abdominal aortic aneurysm, hypertension, or peripheral artery disease. This reduction was attributed to the initiation of antihypertensives, lipid lowering agents and antithrombotic therapy.
Whether a strategy of initiating preventive intervention therapies in adults classified as being at high risk for ASCVD because they have low ABI, improves outcomes remains to be studied.
Limitations and strengths
Major limitations of our study include its retrospective design and missing data. Although the multicenter data were collected during a community-based, national PAD screening program, the dataset was not designed to answer the primary hypothesis of our study, and hence, no definitive conclusion can be drawn regarding the added contribution of the ABI to the PCE for ASCVD risk prediction. Data for variables used for calculation of 10-year risk of events using PCE was missing for number of participants that resulted in small sample size. Major strengths include use of NHANES data, which represents the general population in the United States, and linking all-cause mortality as an outcome.
Conclusions
Using the pooled cohort equations, two-third of the participants with low ABI were classified as having a low, borderline or intermediate risk of ASCVD. Low ABI was associated with an increased all-cause mortality in the overall cohort and specifically among those with a borderline/intermediate or high risk of ASCVD but not in those with a low risk of ASCVD. Our study supports consideration of ABI in risk stratification for primary prevention among patients classified as having a borderline or intermediate risk of ASCVD using the PCE.
Author contributions
Atul Singla: Conceptualization, Supervision, Visualization, Writing, Reviewing. Muchi Ditah Chobufo: Software, Formal analysis, Validation. Abbas Ali: Conceptualization, Methodology, Writing. Amir Meesum: Data Curation, Investigation. Wilbert Aronow: Writing- Reviewing and Editing. Andrew M Goldsweig: Writing- Reviewing and Editing. Sudarshan Balla: Conceptualization, Original writing draft. Paul K Whelton: Supervision, Writing- Reviewing and Editing.
Disclosures
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this study.
Acknowledgements
We thank Dr. Heather Gornik for critical review of the paper.
References
Virani S.S.
Alonso A.
Benjamin E.J.
et al.
Heart disease and stroke statistics—2020 update: a Report From the American Heart Association.
2019 ACC/AHA guideline on the primary prevention of cardiovascular disease: a report of the American College of Cardiology/American Heart Association task force on clinical practice guidelines.
Center for Health Statistics N. The Linkage of National Center For Health Statistics Survey Data to the National Death Index –2015 Linked Mortality File (LMF): Methodology Overview and Analytic Considerations. n.d.
Screening for peripheral artery disease and cardiovascular disease risk assessment with the ankle-brachial index in adults: U.S. Preventive Services Task Force recommendation statement.