ORIGINAL INVESTIGATION

The Ankle-Brachial Index in the Elderly and

Risk of Stroke, Coronary Disease, and Death
The Framingham Study
Joanne M. Murabito, MD, ScM; Jane C. Evans, DSc; Martin G. Larson, ScD;
Kenneth Nieto, MA; Daniel Levy, MD; Peter W. F. Wilson, MD

Background: A low ankle-brachial index (ABI) is associated with an increased risk of death and cardiovascular disease. Limited data exist regarding the relation
between a low ABI and stroke. We sought to examine the
relation between a low ABI and stroke, coronary heart
disease, and death in the elderly.

Only 18% of the participants with a low ABI reported claudication symptoms. One third of those with a normal ABI
and 55% of those with a low ABI had cardiovascular disease at baseline. Results of multivariable Cox proportional hazards analysis demonstrated a statistically significant increase in the risk of stroke or transient ischemic
attack in persons with a low ABI (hazards ratio, 2.0; 95%
confidence interval, 1.1-3.7). No significant relation between a low ABI and coronary heart disease (hazards ratio, 1.2; 95% confidence interval, 0.7-2.1) or death (hazards ratio, 1.4; 95% confidence interval, 0.9-2.1) was
observed.

Methods: We examined 251 men and 423 women with

a mean age of 80 years who had a Framingham Study examination from 1994 to 1995. A low ABI was defined as
less than 0.9. Persons were followed up for 4 years for
occurrence of stroke or transient ischemic attack, coronary disease, and death. Cox proportional hazards models were used to assess the relation between a low ABI
and each outcome after adjusting for age, sex, and prevalent cardiovascular disease.

Conclusions: A low ABI is associated with risk of stroke

or transient ischemic attack in the elderly. These results
need to be confirmed in larger studies.

Results: A low ABI was detected in 20% of our sample.

Arch Intern Med. 2003;163:1939-1942

From the National Heart, Lung,

and Blood Institutes
Framingham Heart Study,
National Institutes of Health,
Framingham, Mass
(Drs Murabito, Evans, Larson,
Levy, and Wilson, and
Mr Nieto); the Sections of
General Internal Medicine
(Dr Murabito), Preventive
Medicine (Drs Evans, Levy, and
Wilson), and Endocrinology
(Dr Wilson), Boston University
School of Medicine, Boston,
Mass; and the National Heart,
Lung, and Blood Institute,
Bethseda, Md (Dr Levy). The
authors have no relevant
financial interest in this article.

LOW ankle-brachial index

(ABI) is common in the elderly and the prevalence
exceeds 25% in persons
older than 85 years.1-4 In
middle-aged and older adults a low ABI has
been reported to be associated with an increased risk of death, total cardiovascular
disease (CVD), coronary heart disease
(CHD), congestive heart failure, and symptomatic peripheral arterial disease.5-12 Risk
for all-cause mortality is increased 2- to
4-fold in persons with a low ABI6-9 while risk
of death from coronary disease is increased more than 6-fold in middle-aged
adults6 and more than 3-fold in older
adults.7,8 Peripheral arterial disease that is
symptomatic and severe increased cardiovascular and CHD mortality 15-fold.6 Limited data are available on the relation between ABI and risk of stroke. Studies of men
with symptoms of lower extremity arterial
disease have reported an increased risk of
stroke death.13-15 Decreasing ABI levels have
been associated with a higher prevalence of
self-reported stroke or transient ischemic at-

(REPRINTED) ARCH INTERN MED/ VOL 163, SEP 8, 2003

1939

tack (TIA) in middle-aged adults.16 The Honolulu Heart Program recently reported an
increased risk of stroke with declining ABI
levels in elderly men.17 We had the opportunity to examine the relation between low
ABI and risk of stroke, CHD, and death in
elderly adults participating in the Framingham Heart Study.
METHODS
Our study sample included surviving members of the original cohort who had a Framingham Heart Study routine research examination between 1993 and 1995. Participants have
been examined every 2 years since the study
inception in 1948. Informed consent for this
study was obtained from participants at the time
of their examination. The institutional review
board of the Boston University School of Medicine approved the examination content, which
included a standardized medical history and
physical examination, an electrocardiogram,
noninvasive cardiovascular testing, and phlebotomy for measurement of blood lipids and
glucose levels.
Trained technicians obtained anklebrachial systolic blood pressure measure-

WWW.ARCHINTERNMED.COM

Downloaded from www.archinternmed.com on March 5, 2012

2003 American Medical Association. All rights reserved.

Table 1. Baseline Characteristics of the Study Sample

by Ankle-Brachial Index (ABI)*

Characteristic

ABI0.9
(n = 533)

ABI0.9
(n = 141)

Age, mean, y
Men
Intermittent claudication
Coronary heart disease
Stroke or transient ischemic attack
Total cardiovascular disease
Hypertension
Diabetes mellitus
Current smoker
Former smoker
Cholesterol, mg/dL (mmol/L)

80
37
2
22
9
36
69
10
6
54
204 (5.27)

81
40
18
33
15
55
86
18
13
58
210 (5.43)

*Data are percentages unless otherwise specified.

Table 2. Four-Year Outcome Events by

Ankle-Brachial Index (ABI)
No. (%) of Events
Outcome
Coronary heart disease
Stroke or transient ischemic
attack
Death

ABI0.9

ABI0.9

46 (9)
28 (5)

17 (12)
19 (13)

74 (14)

33 (23)

ments according to a standardized protocol. Blood pressure cuffs

were applied to both arms and both ankles and the systolic blood
pressure was measured twice at each site using an 8-MHz Doppler pen probe and an ultrasonic Doppler flow detector (Parks
Medical Electronics Inc, Shaw Aloha, Ore). A third measurement was taken when the initial and second blood pressure measurements differed by more than 10 mm Hg at any site. If the
posterior tibial pulse could not be located by palpation or with
the Doppler probe, measurement was taken from the dorsalis
pedis artery.
To calculate the ABI ratio, the average systolic blood pressure measurement in the ankle was divided by the average systolic blood pressure measurement in the arm. The mean pressure of the higher arm was used to calculate the ABI separately
for each leg. A low ABI was defined as less than 0.9 in either
leg. We excluded 11 participants with insufficient blood pressure data. An additional 18 participants with an ABI greater than
1.5 were also excluded.
Intermittent claudication was ascertained using a standardized physician-administered questionnaire. The specific
questions were about exertional leg discomfort that was related to ground steepness or rapidity of walking and relieved
with rest. Participants were queried at each examination about
cigarette smoking. The examining physician measured resting
blood pressure twice and the average of the 2 readings was used
to determine the presence of hypertension. Hypertension was
defined as a blood pressure of 140/90 mm Hg or greater or the
use of antihypertensive medications.3 Diabetes mellitus was considered present if a random glucose reading was 140 mg/dL (7.77
mmol/L) or greater, or if the participant took insulin or oral
hypoglycemic medications.
Participants were followed up for 4 years for the occurrence of stroke, TIA, fatal and nonfatal CHD (angina pectoris;
coronary insufficiency; myocardial infarction; and CHD death,
(REPRINTED) ARCH INTERN MED/ VOL 163, SEP 8, 2003
1940

Table 3. Relation of a Low Ankle-Brachial Index

to Adverse Outcomes*
Hazards Ratio
(95% Confidence Interval)

Outcome
Stroke or transient ischemic attack
Coronary heart disease
Death

2.0 (1.1-3.7)
1.2 (0.7-2.1)
1.4 (0.9-2.1)

*Results of multivariable Cox proportional hazards regression analysis.

Covariates for coronary heart disease and death: age, sex, and prevalent
cardiovascular disease; covariates for stroke or transient ischemic attack:
age, sex, prevalent cardiovascular disease, systolic blood pressure, and atrial
fibrillation.

sudden and nonsudden), and death from all causes. Using previously established criteria, a panel of neurologists made the
final determination of the occurrence of stroke or TIA. A separate panel of senior physician investigators determined all CHD
disease end points and causes of death.18-21 A stroke occurred
when there was documentation that a focal neurologic deficit
was of abrupt onset and lasted more than 24 hours. In this report all stroke subtypes were included (atherothrombotic brain
infarction, cerebral embolus, and intracranial hemorrhage). A
diagnosis of a TIA was made when there was documentation
that a focal neurologic deficit was of abrupt onset and lasted
less than 24 hours. Cox proportional hazards regression analysis was used to assess the relation between a low ABI level and
each outcome after adjusting for age, sex, and prevalent CVD.
Additional covariates in the Cox model assessing the relation
between a low ABI level and stroke or TIA were systolic blood
pressure and atrial fibrillation. All statistical analyses were performed using SAS statistical software.22
RESULTS

Our sample of elderly persons had a mean age of 80 years

(range, 74-95 years), and about 60% of them were women.
A low ABI occurred in 20% (141) of subjects, of whom
only 18% reported symptoms of intermittent claudication at the baseline examination. The baseline prevalence
of CHD, stroke or TIA, and CVD was substantial (Table 1).
Among those with a low ABI, one third had CHD, 15%
had experienced a stroke or TIA, and a few more than half
had CVD (CHD, stroke or TIA, or intermittent
claudication). Nearly 70% of those with a normal ABI and
86% of those with a low ABI had hypertension. Although
the prevalence of current smoking was low, more than half
of the participants were former smokers.
During 4 years of follow-up, 12% of subjects with a
low ABI had a CHD event, 13% had a stroke or TIA, and
23% died (Table 2). Corresponding events in subjects
with a normal ABI were 9%, 5%, and 14%. Of the 47 stroke
or TIA events that occurred during follow-up, 17 were
atherothrombotic brain infarctions, 14 were TIAs, 12 were
cerebral emboli, and 4 were intracerebral hemorrhages.
There was no significant relation between a low ABI and
CHD (hazards ratio [HR], 1.2; 95% confidence interval
[CI], 0.7-2.1) or death (HR, 1.4; 95% CI, 0.9-2.1)
(Table 3). There was, however, a significant 2-fold risk
for stroke or TIA in persons with a low ABI (HR, 2.0; 95%
CI, 1.1-3.7). The relation between ABI level and stroke
or TIA revealed an increasing hazard as the ABI level decreased (Figure 1). Compared with the referent group
WWW.ARCHINTERNMED.COM

Downloaded from www.archinternmed.com on March 5, 2012

2003 American Medical Association. All rights reserved.

Hazards Ratios and 95% Cl

of persons with an ABI greater than 1.0, persons with an

ABI of 0.9 to 1.0 had an HR of 1.5 while persons with an
ABI less than 0.9 had an HR of 2.2 (test for trend across
ABI levels, P=.02).
The relation between a low ABI and risk for stroke
or TIA was consistent in persons with and without baseline CVD (Figure 2). The HR was 2.2 (95% CI, 1.14.4) when persons with stroke or TIA at baseline were
excluded. In persons free of baseline CVD or atrial fibrillation the HR was also 2.2 but the relation was no
longer significant owing to the small number of stroke
events (P=.11). Similarly, in persons with CVD at baseline, the HR was 1.8; again, the number of events was
small, which limited our power to detect a significant relation (P=.15).

5
4
3
2.2

1.5
1
0

(REPRINTED) ARCH INTERN MED/ VOL 163, SEP 8, 2003

1941

>1.0

Ankle-Brachial Index

Figure 1. Ankle-brachial index (ABI) and risk of stroke or transient ischemic

attack: results of multivariable Cox proportional hazards model. The referent
group had an ABI greater than 1.0; test for trend across ABI levels, P=.02.
CI indicates confidence interval.

COMMENT

A low ABI was associated with risk of stroke or TIA in

this sample of elderly persons but not with risk of coronary disease or death. The ABI was both independently
and inversely related to risk of stroke or TIA. The association between low ABI and stroke or TIA was consistent in persons with and without CVD at baseline. Our
findings are remarkably similar to those recently reported by the Honolulu Heart Program Study,17 the first
study to describe an independent association between ABI
and stroke in the elderly. In elderly Japanese men free of
stroke, CHD, and history of lower extremity vascular surgery at baseline, a low ABI was associated with twice the
hazard for total and thromboembolic stroke.17 Furthermore, in that study, a low ABI was related to an increased stroke risk both before and after risk factor adjustment.
Studies of the relation between ABI and incident
stroke in middle-aged populations have suggested a positive association.23-25 In middle-aged adults (45-64 years)
a low ABI was found to be strongly associated with ischemic stroke in men and women as well as in blacks
and whites, but the relationship was attenuated and not
significant after risk factor adjustment.23 In that study,
the HR for stroke in the group with the lowest ABI (0.8),
however, was nearly 2.0 (95% CI, 0.8-4.7), which is consistent with our findings and those of the Honolulu Heart
Program Study. The Edinburgh Artery Study24 reported
an increased relative risk of stroke in persons with a low
ABI (relative risk, 1.98; 95% CI, 1.05-3.77), as did a Swedish study25 of 68-year-old men with asymptomatic carotid stenosis (relative risk, 2.0; 95% CI, 1.1-3.7). Although the Cardiovascular Health Study9 did not find a
significant relation between a low ABI and stroke in persons free of baseline CVD, the incidence of stroke was
higher in persons with a low ABI. The incidence of stroke
at 6 years was 21.2 per 1000 person-years in those with
a low ABI compared with 9.1 per 1000 person-years in
those with a normal ABI.9 These findings from the Cardiovascular Health Study are consistent with results from
the Honolulu Heart Program Study26 and our study.
Low ABI levels have been shown to be an important marker of generalized atherosclerosis through a positive association with CVD risk factors and outcomes as
well as measures of subclinical atherosclerotic disease.

0.9 -1.0

<0.9

Entire Study Sample

Events

673

47

Subjects Without Prevalent

Stroke

606

39

CVD and AF

404

22

269

25

Subjects With Prevalent

CVD or AF

1
2
3
4
5
6
Hazards Ratios and 95% Cl

Figure 2. Low ankle-brachial index increases risk of stroke or transient

ischemic attack in persons with and without cardiovascular disease at
baseline. AF indicates atrial fibrillation; CI, confidence interval; and CVD,
cardiovascular disease.

Population-based studies of middle-aged and older adults

have demonstrated a positive correlation between low ABI
and ultrasound-detected carotid atherosclerosis, the latter being defined by measurement of carotid stenosis or
carotid intima-media thickness.27,28 In addition, elderly
persons with peripheral arterial disease have increased
aortic and common carotid artery stiffness compared with
persons without peripheral arterial disease.29 Atherosclerosis and vascular stiffness may explain in part the underlying pathologic mechanisms mediating the relation
between ABI and stroke risk.
Many studies have reported a 2- to 4-fold increase
in mortality associated with a low ABI.6-9,30 In contrast,
we did not find a relation between a low ABI and mortality in our sample. The disparity in findings may be age
related, as the average age of our subjects, which exceeded 80 years, was much higher than that of subjects
in most prior reports. In addition, most deaths in our
sample (62%) were from noncardiovascular causes. On
the other hand, the SHEP (Systolic Hypertension in the
Elderly Program) study,8 with participants only slightly
younger than those in our sample, reported a positive association between mortality and a low ABI. Most of the
deaths that occurred in the SHEP study, a trial of blood
pressure treatment in older adults with isolated systolic
hypertension, resulted from CVD. Other populationbased studies have reported that deaths from causes other
than CVD were not significantly increased in persons with
WWW.ARCHINTERNMED.COM

Downloaded from www.archinternmed.com on March 5, 2012

2003 American Medical Association. All rights reserved.

a low ABI6,7 or in persons with a low ABI and baseline

CVD.9 Our findings therefore are not inconsistent with
prior reports.
A low ABI has been shown to increase risk for CHD
mortality.6-8 Criqui et al6 were the first to demonstrate a
significant increase in CHD deaths in both men and
women with large-vessel peripheral arterial disease. In
that study of predominantly white upper-middle-class
adults with a mean age of 66 years, CHD deaths were increased more than 6-fold in those with peripheral arterial disease. Subsequently, the SHEP study8 and an ancillary study to the Multicenter Study of Osteoporotic
Fractures7 demonstrated a more than 3-fold risk for coronary disease mortality in older adults with a low ABI. There
were too few CHD deaths in our sample to evaluate this
relationship (only 9 deaths during follow-up were attributable to coronary disease). However, although the
occurrence of coronary disease was higher in persons with
a low ABI (12% vs 9%), we did not observe a relation between a low ABI and a composite of nonfatal and fatal
CHD events. Others have reported a nonsignificant increase in risk of nonfatal myocardial infarction24 and total
myocardial infarction and angina in persons with a low
ABI.9 In the Cardiovascular Health Study9 a low ABI was
positively related to risk of total myocardial infarction
and angina; however, after multivariate adjustment, the
risk ratios were reduced and no longer significant. The
lack of an association between a low ABI and coronary
disease in our study may in part be related to the small
number of coronary events.
In conclusion, a low ABI is associated with risk of
stroke or TIA in the elderly. The ankle-brachial blood pressure measurement is a simple test that can be used to identify stroke-prone elderly persons, who can then be considered for more aggressive preventive interventions.
Others have shown that peripheral arterial disease often
goes unrecognized in primary care practice, which may
present a barrier to secondary prevention of cardiovascular events.31 These findings need to be confirmed in
larger studies.
Accepted for publication October 31, 2002.
The National Heart, Lung, and Blood Institutes
(NHLBI) Framingham Heart Study is supported by National Institutes of Health/NHLBI contract NO1-HC38083.
Corresponding author and reprints: Joanne M. Murabito, MD, ScM, The Framingham Heart Study, 73 Mount
Wayte Ave, Suite 2, Framingham, MA 01702-5827.
REFERENCES
1. Newman AB, Siscovick DS, Manolio TA, et al. Ankle-arm index as a marker of
atherosclerosis in the Cardiovascular Health Study. Circulation. 1993;88:837845.
2. Vogt MT, Wolfson SK, Kuller LH. Lower extremity arterial disease and the aging
process: a review. J Clin Epidemiol. 1992;45:529-542.
3. Meijer WT, Hoes AW, Rutgers D, Bots ML, Hofman A, Grobbee DE. Peripheral
arterial disease in the elderly: the Rotterdam Study. Arterioscler Thromb Vasc
Biol. 1998;18:185-192.

(REPRINTED) ARCH INTERN MED/ VOL 163, SEP 8, 2003

1942

4. Curb JD, Masaki K, Rodriguez BL, et al. Peripheral artery disease and cardiovascular risk factors in the elderly: the Honolulu Heart Program. Arterioscler Thromb
Vasc Biol. 1996;16:1495-1500.
5. McKenna M, Wolfson S, Kuller L. The ratio of ankle and arm arterial pressure as
an independent predictor of mortality. Atherosclerosis. 1991;87:119-128.
6. Criqui MH, Langer RD, Fronek A, et al. Mortality over a period of 10 years in patients with peripheral arterial disease. N Engl J Med. 1992;326:381-386.
7. Vogt MT, Cauley JA, Newman AB, Kuller LH, Hulley SB. Decreased ankle/arm
blood pressure index and mortality in elderly women. JAMA. 1993;270:465469.
8. Newman AB, Sutton-Tyrrell K, Vogt MT, Kuller LH. Morbidity and mortality in
hypertensive adults with a low ankle/arm blood pressure index. JAMA. 1993;
270:487-489.
9. Newman AB, Shemanski L, Manoli TA, et al. Ankle-arm index as a predictor of
cardiovascular disease and mortality in the Cardiovascular Health Study. Arterioscler Thromb Vasc Biol. 1999;19:538-545.
10. Newman AB. Peripheral arterial disease: insights from population studies of older
adults. J Am Geriatr Soc. 2000;48:1157-1162.
11. Jager A, Kostense PJ, Ruhe HG, et al. Microalbuminuria and peripheral arterial
disease are independent predictors of cardiovascular and all-cause mortality, especially among hypertensive subjects: five-year follow-up of the Hoorn Study.
Arterioscler Thromb Vasc Biol. 1999;19:617-624.
12. McDermott MM. Ankle brachial index as a predictor of outcomes in peripheral
arterial disease. J Lab Clin Med. 1999;133:33-40.
13. Smith GD, Shipley MJ, Rose G. Intemittent claudication, heart disease risk factors, and mortality: the Whitehall Study. Circulation. 1990;82:1925-1931.
14. Ingolfsson IO, Sigurdsson G, Sigvaldason H, Thorgeirsson G, Sigfusson N. A
marked decline in the prevalence and incidence of intermittent claudication in
Icelandic men 1968-1986: a strong relationship to smoking and cholesterol: the
Reykjavik Study. J Clin Epidemiol. 1994;47:1237-1243.
15. Bowlin SJ, Medalie JH, Flocke SA, Zyzanski SJ, Yaari S, Goldbourt U. Intermittent claudication in 8343 men and 21-year specific mortality follow-up. Ann Epidemiol. 1997;7:180-187.
16. Zheng Z-J, Sharrett AR, Chambless LE, et al. Associations of ankle-brachial index with clinical CHD, stroke, and preclinical carotid and popliteal atherosclerosis: the Atherosclerosis Risk in Communities (ARIC) Study. Atherosclerosis. 1997;
131:115-125.
17. Abbott RD, Rodriguez BL, Petrovitch H, et al. Ankle-brachial blood pressure in
elderly men and the risk of stroke: the Honolulu Heart Program. J Clin Epidemiol. 2001;54:973-978.
18. Kannel WB, Wolf PA, Veter J, et al. Epidemiologic assessment of the role of blood
pressure in stroke: the Framingham Study. JAMA. 1970;214:301-310.
19. Seshadri S, Wolf PA, Beiser A, et al. Elevated midlife blood pressure increases
stroke risk in elderly persons: the Framingham Study. Arch Intern Med. 2001;
161:2343-2350.
20. Dawber TR, Meadors GF, Moore FE. Epidemiologic approaches to heart disease:
the Framingham Study. Am J Public Health. 1951;41:279-286.
21. Dawber TR, Kannel WB, Lyell LP. An approach to longitudinal studies in a community: the Framingham Study. Ann N Y Acad Sci. 1963;107:539-556.
22. SAS Institute SAS/STAT Software: Changes and Enhancements Through Release 6.11. Cary, NC: SAS Institute Inc; 1996:809-884.
23. Tsai AW, Folsom AR, Rosamond WD, Jones DW. Ankle-brachial index and 7-year
ischemic stroke incidence: the ARIC study. Stroke. 2001;32:1721-1724.
24. Leng GC, Fowkes FGR, Lee AJ, Dunbar J, Housley E, Ruckley CV. Use of ankle
brachial pressure index to predict cardovascular events and death: a cohort study.
BMJ. 1996;313:1440-1444.
25. Ogren M, Hedblad B, Isacsson S, Janzon L, Jungquist G, Lindell S. Ten year cerebrovascular morbidity and mortality in 68 year old men with asymptomatic carotid stenosis. BMJ. 1995;310:1294-1298.
26. Kuller LH. Is ankle-brachial pressure measurement of clinical utility for asymptomatic elderly? J Clin Epidemiol. 2001;54:971-972.
27. Sutton KC, Wolfson SK, Kuller LH. Carotid and lower extremity arterial disease
in elderly adults with isolated systolic hypertension. Stroke. 1987;18:817-822.
28. Allan PL, Mowbray PI, Lee AJ, Fowkes FGR. Relationship between carotid intimamedia thickness and symptomatic and asymptomatic peripheral arterial disease: the Edinburgh Artery Study. Stroke. 1997;28:348-353.
29. Van Popele NM, Grobbee DE, Bots ML, et al. Association between arterial stiffness and atherosclerosis: the Rotterdam Study. Stroke. 2001;32:454-460.
30. Newman AB, Tyrrell KS, Kuller LH. Mortality over four years in SHEP participants with a low ankle-arm index. J Am Geriatr Soc. 1997;45:1472-1478.
31. Hirsch AT, Criqui MH, Treat-Jacobson D, et al. Peripheral arterial disease detection, awareness, and treatment in primary care. JAMA. 2001;286:1317-1324.

WWW.ARCHINTERNMED.COM

Downloaded from www.archinternmed.com on March 5, 2012

2003 American Medical Association. All rights reserved.