Elsevier

Atherosclerosis

Volume 220, Issue 1, January 2012, Pages 160-167
Atherosclerosis

Ankle–brachial index and cardiovascular risk prediction: An analysis of 11,594 individuals with 10-year follow-up

https://doi.org/10.1016/j.atherosclerosis.2011.10.037Get rights and content

Abstract

Background

Low ankle–brachial index (ABI) is associated with increased risk of subsequent cardiovascular disease events, independent of Framingham risk factors, but its ability to improve risk prediction prospectively has not been examined.

Methods

We conducted post-hoc analysis of data from Atherosclerosis Risk in Communities Study (ARIC Study), a large prospective cohort study. 11,594 white and African American (24.2%) men and women, aged 45–64 years, with available Framingham Risk Score (FRS) variables and ABIs at baseline, and without known history of cardiovascular disease or diabetes mellitus or known peripheral arterial disease at baseline were assessed for hard cardiovascular events (hCVD; defined as heart attack, coronary death or stroke) over median follow-up of 10 years. Hazard ratios, C statistic, and net reclassification indexes were calculated to determine the independent predictive ability of ABI compared with FRS.

Results

659 hCVD events occurred. Standardized ABI was significantly associated with hCVD events but with a relatively small effect on events (hazard ratios of 0.85 per standard deviation (95% CI 0.79–0.91) (p-value < 0.0001)). The C statistic of FRS modified with ABI was only modestly improved (0.756–0.758). Net reclassification improvement, an indicator of prospective prediction performance, using an ABI threshold of 0.9 was small and statistically insignificant (0.8%, p = 0.50).

Conclusions

Although the 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 FRS was small in magnitude, and the FRS performed similarly with or without integration or supplementation with ABI. These findings do not provide strong evidence to support FRS modification to include ABI.

Introduction

Current guidelines call for assessment of cardiovascular risk using accepted risk variables (age, total cholesterol, HDL cholesterol, systolic blood pressure, treatment for hypertension, cigarette smoking, and family history of premature coronary heart disease) [1]. If two or more risk factors are present, an integrated risk prediction model, such as the Framingham Risk Score (FRS), is used to quantify short-term (10-year) risk [1]. Individuals with “high” risk (≥20% 10-year risk of “hard CHD” (heart attack or coronary-related death)) are considered “CHD equivalent” and are candidates for intensive medical risk reduction [1], and those at “intermediate” risk (≥10 < 20% 10-year risk) candidates for less intensive risk factor reduction. However the Framingham Risk Score does not fully explain cardiovascular risk [2] since 20% of MIs occur in those with no risk factors [3], [4], and at least 60–80% of MIs in those without known CHD or CHD equivalent occur in those at low or intermediate risk of CHD using Framingham Risk Scores [5], [6], [7]. Methods to improve risk prediction, especially by means that are noninvasive and inexpensive, are of considerable interest [8], [9] as they would allow more people at increased risk for cardiovascular events to receive intensive risk modification therapy and thereby reduce heart attacks, strokes, and related deaths.

The ankle–brachial index (ABI), which is the ratio of systolic pressure at the ankle to that in the arm, is inexpensive, widely available, and noninvasive. Recent data from National Health And Nutrition Examination Survey reveal that abnormal ABI is highly prevalent among individuals otherwise not considered at high risk of cardiovascular events [10]. Meta-analyses of many large observational studies with long-term follow-up have reported that ABI is associated with coronary heart events independent of traditional Framingham variables [11], and an abnormal ABI is accepted in current guidelines as a coronary-heart disease (CHD) equivalent [1]. Screening measurement of the ABI to supplement risk prediction has been recommended by American Heart Association and the American College of Cardiology [12], the Transatlantic Inter-Society Consensus Working Group [13], and the Fourth Joint European Task Force [14], among others [10], [15]. Despite this, the usefulness of the ABI as a risk prediction variable when Framingham-based variables are known is complicated. Both the magnitude of ABIs independent relationship with subsequent cardiovascular disease events and how ABI is integrated into a risk prediction model have strong implications for its usefulness.

In this study, we critically examine claims that ABI can improve risk prediction for hard cardiovascular disease events (hCVD; defined as heart attack, coronary death, or stroke) by evaluating its performance using a modified risk-prediction model with traditional Framingham Risk Scores, and also by examining how abnormal ABI performs when used to establish CHD equivalence per se, as indicated in current guidelines [1].

Section snippets

Methods

For this analysis we used data collected in the Atherosclerosis Risk in Communities (ARIC) Study [16], a prospective cohort study of middle-aged individuals with long-term follow-up. Our objective is to determine if ankle–brachial index provides information on risk of subsequent hCVD events independent of a standard risk factors’ model based on the Framingham Risk Score variables [1], by: (1) determining the incremental value of ABI when added in a standard FRS variables model; and (2)

Baseline characteristics

In our population of 11,594 individuals between the ages of 45 and 64, 270 (2.3%) had abnormal ABI. The average ankle brachial index for the entire population was 1.15 ± 0.12 (standard deviation). Table 1 presents the baseline characteristics of our sample, according to ABI. Of the seventeen variables examined, there were strong statistical differences between those with normal ABI compared those with abnormal ABI. Of note, slightly lower mean diastolic blood pressure was noted in abnormal ABI as

Discussion

Identification of individuals without known atherosclerotic coronary disease or diabetes mellitus who are at high risk of coronary heart events depends on traditional risk factors (age, total cholesterol, HDL cholesterol, systolic blood pressure, treatment for hypertension, cigarette smoking, and family history of premature coronary heart disease) [1]. However, these variables do not fully explain cardiovascular risk [2]. Most (60%) of acute MIs occur in those without known CHD [2] and at least

Contributors

Drs Murphy, Pencina and Dhangana had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Murphy, Dhangana, Pencina.

Acquisition of data: Murphy, Dhangana.

Analysis and interpretation of data: Murphy, Pencina, Dhangana, D’Agostino.

Drafting of the manuscript: Murphy, Dhangana.

Critical revision of the manuscript for important intellectual content: Murphy, D’Agostino, Pencina, Dhangana.

Financial disclosure

None.

Disclaimer

ARIC Study is conducted and supported by National Heart, Lung, and Blood Institute (NHLBI) in collaboration with ARIC investigators. This research was conducted using de-identified data obtained from the NHLBI and does not necessarily reflect the opinions or views of the ARIC Study or NHLBI.

References (40)

  • P.M. Ridker et al.

    Established and emerging plasma biomarkers in the prediction of first atherothrombotic events

    Circulation

    (2004)
  • U.N. Khot et al.

    Prevalence of conventional risk factors in patients with coronary heart disease

    JAMA

    (2003)
  • P. Brindle et al.

    Predictive accuracy of the Framingham coronary risk score in British men: prospective cohort study

    BMJ

    (2003)
  • T.M. Reynolds et al.

    Concordance evaluation of coronary risk scores: implications for cardiovascular risk screening

    Curr Med Res Opin

    (2004)
  • P. Greenland et al.

    Improving coronary heart disease risk assessment in asymptomatic people: role of traditional risk factors and noninvasive cardiovascular tests

    Circulation

    (2001)
  • Ankle Brachial Index Collaboration et al.

    Ankle brachial index combined with Framingham risk score to predict cardiovascular events and mortality: a meta-analysis

    JAMA

    (2008)
  • A.T. Hirsch et al.

    ACC/AHA 2005 practice guidelines for the management of patients with peripheral arterial disease (lower extremity, renal, mesenteric, and abdominal aortic): a collaborative report from the American Association for Vascular Surgery/Society for Vascular Surgery Society for Cardiovascular Angiography and Interventions, Society for Vascular Medicine and Biology, Society of Interventional Radiology, and the ACC/AHA Task Force on Practice Guidelines (Writing Committee to Develop Guidelines for the Management of Patients With Peripheral Arterial Disease): endorsed by the American Association of Cardiovascular and Pulmonary Rehabilitation; National Heart, Lung, and Blood Institute; Society for Vascular Nursing; TransAtlantic Inter-Society Consensus; and Vascular Disease Foundation

    Circulation

    (2006)
  • I. Graham et al.

    European guidelines on cardiovascular disease prevention in clinical practice: full text Fourth Joint Task Force of the European Society of Cardiology and other societies on cardiovascular disease prevention in clinical practice (constituted by representatives of nine societies and by invited experts)

    Eur J Cardiovasc Prev Rehabil

    (2007)
  • J.A. Beckman et al.

    The United States preventive services task force recommendation statement on screening for peripheral arterial disease: more harm than benefit?

    Circulation

    (2006)
  • Anon.

    The Atherosclerosis Risk in Communities (ARIC) Study: design and objectives. The ARIC investigators

    Am J Epidemiol

    (1989)
  • Cited by (62)

    • Costs and effects of cardiovascular risk reclassification using the ankle-brachial index (ABI) in addition to the Framingham risk scoring in women

      2021, Atherosclerosis
      Citation Excerpt :

      In fact, the 2016 guidelines on cardiovascular disease prevention [38] simply do not support its use, while the 2018 guidelines on hypertension management [39] stated that it “should be considered in hypertensive particularly at intermediate risk”. As already mentioned, we choose to use the largest meta-analysis [13] to extract reclassification data of ABI in addition to FRS even because the analysis on gender differences has been rarely repeated in subsequent published studies [31–34]. Reported results in male/female reclassification could be determined by differences in population characteristics but also by the fact that FRS may perform differently between genders [40,41].

    • Association of Cardiovascular Mortality and Deep Learning-Funduscopic Atherosclerosis Score derived from Retinal Fundus Images

      2020, American Journal of Ophthalmology
      Citation Excerpt :

      Previous meta-analyses and cohort studies analyzed the added predictive value of C-reactive protein,25 CIMT,26 computed tomography coronary artery calcium score,27,28 and the ankle-brachial index,29,30 beyond FRS. Improvement of C-statistics above conventional risk factors and Framingham risk score, determined by the difference between concordance using conventional risk factors and concordance using conventional risk factors plus the additional measurement, were insignificant for CIMT (0.00-0.002),26,27 ankle-brachial index (0.00-0.002),27,30 and C-reactive protein (0.00),27 while improvements were mostly significant for computed tomography coronary artery calcium score (0.02-0.13).27,28 In our work, the improved concordance estimate of DL-FAS for cardiovascular mortality was small but significant (HR, 0.027; 95% CI, 0.004-0.049), and relative IDI and NRI measurements were also significant at 20.45% and 29.5%, respectively.

    View all citing articles on Scopus
    View full text