Ankle-to-arm blood pressure index ≤ 0.9 was associated with increased mortality in hypertensive adults
ACP J Club. 1994 Mar-April;120:50. doi:10.7326/ACPJC-1994-120-2-050
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 Jul 28;270:487-9.
To evaluate the relation between the ankle-to-arm blood pressure index (AAI) and short-term mortality and morbidity in older adults with systolic hypertension.
Cohort analytic study with a mean follow-up of 16 months.
11 clinical centers in the United States.
1537 adults (mean age 76 y, 56% women) with systolic hypertension participating in the Systolic Hypertension in the Elderly Program (SHEP).
Assessment of risk factors
The AAI was measured at baseline (at the 1989 to 1990 clinic examination) by trained observers; participants with an AAI ≤ 0.9 in either leg were considered to have peripheral arterial disease. Intermittent claudication was assessed by the Rose questionnaire and defined as calf pain with exertion relieved by rest in ≤ 10 minutes. Myocardial infarction, angina, congestive heart failure, stroke, and transient ischemic attack were assessed by self-report and validated at annual clinic visits.
Main outcome measures
All-cause, coronary heart disease, and cardiovascular disease mortality; and coronary heart disease and cardiovascular disease morbidity.
The AAI was ≤ 0.9 in 26% of 1537 participants. 47 persons (3%) died during follow-up. After adjusting for age and sex, the relative risk (RR) for all-cause mortality when comparing those with an AAI ≤ 0.9 with those with an index > 0.9 was 3.8 (95% CI 2.1 to 6.9). The corresponding figures for coronary heart disease mortality and for cardiovascular disease mortality were 3.2 (CI 1.4 to 7.5) and 3.7 (CI 1.8 to 7.7), respectively. The age- and sex-adjusted RR for cardiovascular disease morbidity and mortality was 2.5 (CI 1.5 to 4.3). After adjustment for baseline cardiovascular disease and other cardiovascular risk factors, the RRs for all-cause mortality and for cardiovascular morbidity and mortality were 4.1 (CI 2.0 to 8.3) and 2.4 (CI 1.3 to 4.4), respectively. Excluding participants with cardiovascular disease at baseline did not change the results.
Among older men and women with systolic hypertension, the presence of lower-extremity arterial disease, defined as an ankle-to-arm blood pressure index of ≤ 0.9, was independently associated with increased mortality and cardiovascular morbidity in the entire group and among those free of cardiovascular disease at baseline.
Source of funding: National Institutes of Health.
For article reprint: Dr. A.B. Newman, Medical College of Pennsylvania, Allegheny General Hospital, 320 E. North Avenue, Pittsburgh, PA 15212, USA. FAX 412-359-3834.
Prevailing wisdom argues that persons at greatest risk for cardiovascular disease should be identified and managed more aggressively. Patients with peripheral arterial disease, just like those with transient ischemic attacks and carotid stenoses, are vulnerable to myocardial infarction and to death from cardiovascular disease. Because the relation between smoking and peripheral arterial disease is particularly strong, it is attractive to hypothesize an association between peripheral arterial disease (or an appropriate marker) and mortality from all causes, coronary disease, and smoking-related cancers. AAI offers the most sensitive noninvasive method for detecting important atherosclerotic stenoses in large, lower-limb arteries.
The Multicenter Study of Osteoporotic Fractures included a population of older white women, 6% of whom had an AAI of ≤ 0.9. The RR was 3.1 for all-cause mortality in this group after adjusting for other risk factors. Remarkably, however, mortality was lower in those with evidence of cardiovascular disease at entry, and patients with symptoms of claudication fared better than those who were asymptomatic.
In the SHEP substudy, 26% had AAI ≤ 0.9. On average, patients in this group were 3 years older and twice as likely to be nonwhite, to smoke, and to have diabetes as were the participants with AAI > 0.9. When corrected for baseline cardiovascular disease and other cardiovascular risk factors, the RR was 4.1, higher than for any other cardiovascular risk predictor or indicator in this age group. AAI ≤ 0.9 provides evidence of (mainly asymptomatic) large lower-limb artery atherosclerosis detected incompletely by either the Rose questionnaire for intermittent claudication or by physical examination. As well as indicating vascular disease, AAI is clearly a better independent indicator of total mortality, coronary heart disease mortality, and cardiovascular mortality in older women with osteoporosis and in older men and women with systolic hypertension than are conventional risk factors such as smoking, elevated BP, diabetes, or dyslipidemia.
A follow-up study of SHEP participants (1) provided confirmatory evidence of the prognostic value of AAI. In patients with and without evidence of preexistent cardiovascular disease, a reduced AAI predicted increased mortality, even when adjusted for other known cardiovascular risk factors. In those SHEP patients with no clinical cardiovascular disease, AAI ≤ 0.9 was associated with an adjusted odds ratio for total mortality of 2.76 (CI 2.33 to 3.20).
Can these results be applied to older patients with reduced AAI in the absence of isolated systolic hypertension? The pathogenesis of atheroma in aortoiliac or iliofemoral vessels is similar to that of coronary atherogenesis. In the setting of other known risk factors, the detection of reduced AAI probably augments our ability to determine a worse prognosis for heart disease and death and possibly for cancer as well. A group of Scottish investigators have confirmed the improvement in risk stratification offered by the AAI in a randomly selected group of 1592 men and women aged 55 to 74 years (2). After adjustment for age, sex, coronary disease, and diabetes at baseline, an AAI ≤ 0.9 predicted relative risks of 1.58 and 1.98 for all-cause mortality and stroke, respectively. Because a reduced AAI indicates existing disease, it seems unlikely that subgroups of patients exist in whom a reduced AAI heralds a lesser risk for coronary heart disease.
The editorial by Applegate on both cohort studies is a scholarly commentary on the clinical value of AAI (3). His conclusion that AAI should be considered a useful screen for older patients with any cardiovascular risk and middle-aged patients with intermediate risk has some merit. A Doppler stethoscope is relatively inexpensive ($200 to $300 U.S.), and experience in its use, leading to low interobserver variability, can be acquired with minimal training of 2 to 3 hours.
Nevertheless, health policy decisions that lead to the widespread application of any new diagnostic tests require more information than that provided by these and earlier reports on the prognostic implications of AAI. Before the acquisition of a Doppler stethoscope by every primary care office is advocated, we should know if using it will result in meaningful reductions in total coronary heart disease mortality. Until appropriate trials are done, the use of AAI to augment risk evaluation should remain (like echocardiography and ambulatory BP monitoring in this setting [4, 5]) a research technique rather than routine clinical practice.
S. George Carruthers, MD
Dalhousie UniversityHalifax, Nova Scotia, Canada