Clinical significance of a high ankle-brachial index: Insights from the Atherosclerosis Risk in Communities (ARIC) Study

Transcription

1 Atherosclerosis 190 (2007) Clinical significance of a high ankle-brachial index: Insights from the Atherosclerosis Risk in Communities (ARIC) Study Keattiyoat Wattanakit a, Aaron R. Folsom a, Daniel A. Duprez b, Beth D. Weatherley c, Alan T. Hirsch a,d, a Division of Epidemiology and Community Health, School of Public Health, University of Minnesota, Suite 300, 1300 South 2nd Street, Minneapolis, MN 55454, United States b Cardiovascular Division, University of Minnesota Medical School, MMC 508, 420 Delaware St. SE, Minneapolis, MN 55455, United States c Duke Clinical Research Institute, Duke University Medical Center, Room 0311 Terrace Level, 2400 Pratt Street, Durham, NC 22705, United States d Minneapolis Heart Institute Foundation, 920 East 28th Street, Suite 300, Minneapolis, MN 55407, United States Received 2 December 2005; received in revised form 22 January 2006; accepted 17 February 2006 Available online 29 March 2006 Abstract Background: The clinical significance of a high ankle-brachial index (ABI), defined by the associated risk factor burden and ischemic risk, is largely unknown. Methods: Using data from the Atherosclerosis Risk in Communities Study, we categorized 14,777 participants into normal (ABI between 0.9 and 1.3) and high ABI groups (ABI > 1.3, >1.4, and >1.5) and compared the risk factor profile and CVD event rates of the normal ABI group to each high ABI group. Results: The prevalence of high ABI was 5.5% for ABI > 1.3, 1.2% for ABI > 1.4, and 0.37% for ABI > 1.5. Compared with participants with a normal ABI, those with ABI > 1.3 had a lower prevalence of hypertension and current smoking. The ABI > 1.3 group had a greater mean body mass index, but was characterized by fewer pack years of smoking and lower systolic and diastolic blood pressures than the normal ABI group. The prevalence of diabetes, left ventricular hypertrophy, claudication, and coronary heart disease and mean values of fibrinogen, factor VIII activity, von Willebrand factor, lipoprotein (a), and carotid and popliteal intimal-medial thickness were similar between the two ABI groups. The risk factor profiles of the ABI > 1.4 and >1.5 groups were also not statistically significantly different from that of the normal ABI group. Over a mean follow-up time of 12.2 years, the age, sex, and race-adjusted CVD event rates per 1000 person years were 8.1 in the normal ABI group, 7.6 in the ABI > 1.3 group, 7.6 in the ABI > 1.4 group, and 7.4 in the ABI > 1.5 group. The CVD event rates of the high ABI groups were similar to that of the normal ABI group. Conclusion: Individuals with a high ABI are not characterized by a more adverse atherosclerosis risk factor profile and do not suffer greater CVD event rates than those with a normal ABI Elsevier Ireland Ltd. All rights reserved. Keywords: Ankle-brachial index; Risk factors; Cardiovascular disease 1. Introduction The ARIC study was supported by NHLBI contracts N01-HC-55015, N01-HC-55016, N01-HC-55018, N01-HC-55019, N01-HC-55020, N01- HC-55021, N01-HC Dr. Wattanakit was supported by NHLBI training grant T32-HL Corresponding author. Tel.: ; fax: address: hirsc005@umn.edu (A.T. Hirsch). An ankle-brachial index (ABI) less than 0.9 serves as the diagnostic tool most commonly used to define peripheral arterial disease (PAD) [1,2]. PAD is increasingly recognized as a clinically important marker of atherosclerotic disease due to its association with cardiovascular disease (CVD) incidence and mortality [3 5]. Decrements in ABI values are inversely associated with CVD risk factors, arterial disease /$ see front matter 2006 Elsevier Ireland Ltd. All rights reserved. doi: /j.atherosclerosis

2 460 K. Wattanakit et al. / Atherosclerosis 190 (2007) risk markers (e.g., carotid intima-media thickness (IMT)) and structural heart disease (e.g., left ventricular hypertrophy), in a step-wise fashion [6]. Furthermore, a low ABI value identifies patients at risk for impaired lower extremity physical function, including poor standing balance, walking distance, and walking velocity [7]. In contrast, the clinical significance of a high ABI, often arbitrarily defined by values greater than 1.3 [2], has not been as completely evaluated. Past investigations, usually derived from referral populations with overt PAD, have suggested that a high ABI is associated with lower extremity medial arterial calcification [8], which has been shown to be independently associated with increased mortality in individuals with a long duration of diabetes [9]. Moreover, Newman and colleagues have suggested that individuals with ABI > 1.5 had a higher prevalence of clinical CVD than those with ABI between 1.0 and 1.5 [6]. From these populations, the prognosis of individuals with a high ABI is believed to be less favorable than those with a normal ABI, but there are few data to support this belief [10]. Recently, the Strong Heart Study characterized the risk factor profile of American Indians with high ABI values and reported a dose response relationship of high ABI with greater all-cause and CVD mortality [11]. However, data from the general population are not available. We therefore investigated whether the atherosclerosis risk factor profile, subclinical markers of arterial and heart disease, and the incidence of CVD events in adults with a high ABI (ABI > 1.3, >1.4, and >1.5) were different from those with a normal ABI ( ) in a non-referral, community-derived population. 2. Methods 2.1. Study population The Atherosclerosis Risk in Communities Study (ARIC) is a prospective investigation of the etiology and natural history of atherosclerosis. The study cohort comprised 15,972 white and black men and women ages years at baseline in , who recruited from four US communities: Forsyth County, NC, Jackson, MS, suburbs of Minneapolis, MN, and Washington County, MD. The detailed descriptions of the ARIC study design and objectives have been published elsewhere [12] Measurement of baseline risk factors and subclinical markers The ARIC participants underwent a standardized medical history and examination that included interviews, a fasting venipuncture, ABI, electrocardiogram, and carotid and popliteal IMT. Trained interviewers ascertained basic demographic data, medical history, and personal habits, including smoking and alcohol drinking, physical activity, and medication use. Anthropometric measurements, including weight and height, were obtained while the participant was wearing a scrub unit. Body mass index (BMI) was calculated as weight (kg) divided by the square of height (m). Fasting blood samples were drawn from an antecubital vein for measurement of plasma lipids [13 15], fibrinogen [16], lipoprotein(a) [17], factor VIII (George King Biomedical, Inc., Overland Park, KS), and von Willebrand factor (American Bioproducts Co., Diagnostica, France). Serum glucose was measured using the hexokinase method. Serum creatinine was measured by an alkaline picrate colorimetric assay. The ABI was measured on nearly all participants at baseline. The ABI was computed by dividing the average of ankle systolic blood pressure measurements by the average of brachial systolic blood pressure measurements. Using the Dinamap 1846SX (Criticon Inc., Tampa, FL), trained technicians measured two ankle blood pressures at the posterior tibial artery in a randomly selected leg while the participant was prone. This automated blood pressure measurement device has high validity compared to the standard Doppler ultrasound measurement, and high repeatability [18]. Two brachial artery blood pressures were measured, usually in the right arm, with the participant supine. When measured in both legs, an ABI < 0.91 has 79% sensitivity and 96% specificity for detecting angiogram positive PAD ( 50% or more stenosis) [19]. B-mode carotid and popliteal ultrasound (Biosound 2000 II SA; Biosound Inc., Indianapolis, IN) evaluations were completed on bilateral segments of the extracranial carotid arteries and one randomly selected popliteal artery segment using standardized scanning and reading protocols [20,21]. The average value of the mean far wall thickness was calculated and used in analysis. If the participants had missing IMT information from any carotid artery site, values were imputed for missing sites based on sex and race. Mean far wall IMT was also adjusted for reader differences and downward drift of readers. The intrareader and interreader agreements for maximum far wall thickness at the carotid bifurcation were 0.81 and 0.93, respectively [20,21]. For the popliteal ultrasound evaluation, the technician first identified the horizontal crease behind the knee and then placed the center of transducer so that the image of near and far wall of popliteal artery simultaneously appeared on the screen. Prevalent PAD was defined, for exclusion, as ABI < The Rose Questionnaire [22] was used to identify intermittent claudication, which was defined as exertional leg pain relieved by resting within 10 min. Prevalent coronary heart disease (CHD) was defined as a reported history of physician-diagnosed myocardial infarction, a prior myocardial infarction identified on the baseline electrocardiogram, coronary angioplasty, or cardiovascular surgery. Prevalent left ventricular hypertrophy (LVH) was defined by the sum of S wave in lead V3 and R wave in lead AVL being greater than 28 mm for men and 22 mm for women [23]. Prevalent hypertension was defined as seated diastolic blood pressure 90 mmhg, systolic blood pressure 140 mmhg,

3 or use of anti-hypertensive medications within the past 2 weeks. Prevalent diabetes mellitus was defined as a fasting serum glucose level 7.0 mmol/l (126 mg/dl), nonfasting glucose level 11.1 mmol/l (200 mg/l), participant report of a physician diagnosis of diabetes, or current use of diabetes medication Ascertainment of incident events K. Wattanakit et al. / Atherosclerosis 190 (2007) The ARIC Study ascertained cardiovascular events (CHD or stroke) after baseline by several methods [12,24]. Interviewers contacted participants annually by telephone to identify all hospitalizations. ARIC Study staff also surveyed discharge lists from local hospitals to detect additional cardiovascular events. Technicians visually coded up to three 12-lead electrocardiograms using the Minnesota Code [23]. CHD incidence was defined as a definite or probable hospitalized myocardial infarction or definite fatal CHD. CHD events were validated by a committee of physicians using standardized criteria [24]. Hospitalized strokes were identified and classified according to published criteria based on the occurrence and duration of neurologic signs and symptoms, the results of neuroimaging and other diagnostic procedures, and treatments provided [25]. Strokes secondary to trauma, neoplasm, hematologic abnormality, infection, or vasculitis were not counted, and a focal deficit <24 h was not considered a stroke Statistical analysis Of the 15,792 ARIC participants, 15,231 had baseline ABI measurement (96.4%). We excluded 355 participants who had baseline PAD (ABI < 0.9) and 99 participants because of small numbers in ethnic groups in their field centers, leaving a total of 14,777 participants for the analysis. We categorized ABI into normal (ABI between 0.9 and 1.3) and high ABI groups (at three cut points of ABI > 1.3, >1.4, and >1.5) and compared the baseline traditional and non-traditional risk factors and arterial disease risk markers of the normal ABI group with each of the high ABI groups. Proportions and mean values were calculated using analysis of covariance, adjusted for age, sex, race, center, and use of antihypertensive and cholesterol medication. To determine the relation of high ABI with CVD incidence, age, sex, and race adjusted CVD event rates were calculated for each ABI group using Poisson regression. Length of follow-up was calculated for clinically recognized cases as the time elapsed between the baseline examination and the first CVD event. For non-cases, follow-up ended on the date of death, date of last contact (if lost), or on 31 December 2001, whichever occurred first. We also performed a calculation to determine whether the power was adequate to test a reasonable difference in the proportion of CVD events between the normal ABI and the ABI > 1.4 groups. All statistical analyses were conducted using SAS software version 8.2 (SAS Institute, Cary, NC). Fig. 1. Age, sex, and race adjusted cardiovascular event rate per 1000 personyears according to baseline ankle-brachial index (ABI) level, ARIC Study, : (a) p-value = 0.54 comparing ABI vs. ABI > 1.3; (b) p-value = 0.79 comparing ABI vs. ABI > 1.4; (c) p-value = 0.84 comparing ABI vs. ABI > Results In this sample of 14,777 participants, 13,969 participants had a normal ABI ( ). Of those with a high ABI value, 808 had ABI > 1.3 (5.5%), 184 had ABI > 1.4 (1.2%), and 54 had ABI > 1.5 (0.37%). This prevalence was not different in the 1703 participants with diabetes (5.3%) and the 12,949 participants without diabetes (5.5%) (p-value = 0.84). Compared with participants with normal ABI, those with ABI > 1.3 were more likely to be male, white, and older, and to have a lower prevalence of hypertension and current smoking (Table 1). The ABI > 1.3 group had a greater mean BMI, but lower mean systolic and diastolic blood pressures and cigarette pack years among smokers than the normal ABI group. The prevalence of diabetes, LVH, intermittent claudication, and CHD was quite similar between the normal and high ABI groups. Nontraditional risk factors were also not significantly different between the two groups. Similar findings were observed when comparing the ABI > 1.4 and the ABI > 1.5 groups with the normal ABI group. Overall, the prevalence of traditional risk factors in the high ABI groups was not statistically significantly different from the normal ABI group. The exceptions were the prevalence of diabetes and BMI, which tended to increase as ABI increased. The prevalence of nontraditional risk factors and arterial disease markers also was not significantly different between the high ABI groups and the normal ABI group. The number of CVD events was 1555, 96, 22, and 6 for ABI between 0.9 and 1.3, ABI > 1.3, >1.4, and >1.5, respectively. The mean follow-up time was 12.2 years. The age, sex, and race adjusted CVD event rates per 1000 person years were 8.1 in the normal ABI group, 7.6 in the ABI > 1.3 group, 7.6 in the ABI > 1.4 group, and 7.4 in the ABI > 1.5 group (Fig. 1). One previous study has suggested that ABI > 1.4 was associated with an adjusted CVD mortality risk of 2.09 relative to a normal ABI (defined by ABI between 1.10 and 1.30) [11]. With 11% of CVD events (1555 CVD events/13,969) in the normal ABI group in our study, the relative risk of 2.09 would imply that 22% of CVD events would occur in

4 462 K. Wattanakit et al. / Atherosclerosis 190 (2007) Table 1 Baseline characteristics of men and women according to ankle-brachial index (ABI) level in the Atherosclerosis Risk in Communities (ARIC) Study a Normal ABI High ABI p-value b p-value c p-value d ABI ( ) (N = 13,969) ABI > 1.3 (N = 808) ABI > 1.4 (N = 184) ABI > 1.5 (N = 54) Baseline No. of CVD events Male (%) < White (%) Age (year) 54 (5.7) 55 (5.8) 56 (5.5) 55 (5.6) < Traditional risk factors e Hypertension (%) Diabetes (%) Current smokers (%) < Pack years in smokers 27.9 (22.1) 24.7 (22.1) 26.1 (24.8) 31.8 (21.2) SBP (mmhg) (18.9) (16.4) (16.9) (18.7) < DBP (mmhg) 73.3 (11.3) 72.7 (10.8) 72.2 (10.1) 71.3 (9.5) Total cholesterol (mg/dl) 215 (42) 213 (43) 214 (44) 207 (44) Triglycerides (mg/dl) 131 (88) 127 (77) 128 (85) 138 (89) LDL cholesterol (mg/dl) 137 (39) 137 (40) 134 (41) 129 (44) HDL cholesterol (mg/dl) 52 (17) 52 (16) 52 (16) 51 (19) BMI (kg/m 2 ) 27.6 (5.3) 28.7 (5.5) 29.9 (6.4) 31.6 (8.1) < < < Non-traditional risk factors e Fibrinogen (mg/dl) 302 (65) 302 (65) 305 (65) 318 (75) Factor VIII activity (%) 131 (39) 130 (39) 133 (38) 132 (35) Von Willebrand factor (%) 118 (48) 117 (45) 120 (44) 118 (36) Lipoprotein(a) ( g/ml) 101 (107) 102 (106) 107 (117) 115 (154) Serum creatinine (mg/dl) 1.1 (0.4) 1.1 (0.3) 1.08 (0.2) 1.06 (0.2) Serum albumin (mg/dl) 3.9 (0.3) 3.9 (0.3) 3.9 (0.3) 3.8 (0.3) Arterial disease risk markers e Left ventricular hypertrophy (%) Intermittent claudication (%) NA Coronary heart disease (%) Carotid IMT (mm) 0.72 (0.2) 0.72 (0.2) 0.74 (0.3) 0.71 (0.2) Far-wall popliteal IMT (mm) 0.62 (0.2) 0.62 (0.2) 0.61 (0.2) 0.62 (0.3) ABI: ankle-brachial index; ARIC: Atherosclerosis Risk in Communities; DBP: diastolic blood pressure; IMT: intima-media thickness; SBP: systolic blood pressure. a Values are percentages or means (S.D.). b p-value comparing ABI vs. ABI > 1.3. c p-value comparing ABI vs. ABI > 1.4. d p-value comparing ABI vs. ABI > 1.5. e Adjusted for age, sex, race, center, and use of antihypertensive and cholesterol medication. the ABI > 1.4 group. Using the Gaussian approximation in the Z-test for independent proportions, we found that this difference in proportion of CVD events corresponded to a Z statistic of Power to detect such a difference exceeded 95%. Given the observed proportion and having 184 persons in the ABI > 1.4 group, we had an excellent power to detect a relative risk of 2.0. We also evaluated varying definitions of the normal ABI in this analysis. Subjects with ABI values between 0.9 and 1.0 might have evidence a great burden of atherosclerosis than those with ABI values between 1.0 and 1.3. Thus, inclusion of subjects from the former group into the reference group could theoretically decrease the ability to detect any risk factor profile or CVD event differences between the normal and high ABI groups. We therefore performed a supplemental analysis that redefined the reference group using ABI values between 1.0 and 1.3 and repeated the same analysis. The results of all analyses were similar to those presented in Table 1. Because diabetic subjects may have high ABI as a result of medial arterial calcification, we performed a supplemental analysis in the subgroup with diabetes. Among diabetic subjects, those with high ABI (n = 91) had a significantly lower prevalence of hypertension and mean systolic blood pressure and greater mean BMI and fibrinogen values than did those with a normal ABI (n = 1612). Smoking and pack-years of tobacco use, however, did not differ significantly between the ABI groups. 4. Discussion The ARIC Study provided a unique epidemiologic opportunity to examine the traditional and non-traditional risk

5 K. Wattanakit et al. / Atherosclerosis 190 (2007) factor profile, arterial and heart structure, and CVD rates of individuals with a high ABI. These data are clinically important as use of the ABI measurement is increasingly deployed as a diagnostic test in clinical practice [2,26]. The clinical utility of the ABI to diagnose PAD has also been recognized via surveys of primary care clinicians [27], international consensus statements [28], diabetes care guidelines [29], and national PAD public awareness initiatives [30]. Thus, as this diagnostic method becomes more widely used, these data aid clinical interpretation of high ABI values. Most prior epidemiologic studies have focused on individuals with a low ABI (<0.9). Rarely have the risk factor profile and subclinical markers of individuals with high ABI been described. In our study, the 5.5% prevalence of ABI > 1.3 in the general population was notably lower than in the American Indian cohort of the Strong Heart Study (9.2%) [11] and in other selected populations with a high prevalence of preexisting diabetes (18%) [31] and chronic renal failure (42%) [32]. The low prevalence in ARIC may be partially explained by the low prevalence of diabetes and renal disease and the measurement of the ABI from a single leg in a predominantly middle-aged, community-derived sample. Contrary to a common perception, our data suggest that the risk factor profile of individuals with a high ABI in the general population is similar to that of individuals with a normal ABI. Hence, it is not unexpected to observe that CVD event rates were comparable for all ABI groups. In a national context in which ABI measurements will be more widely used and in which the interpretation of a high ABI value will be of increasing importance, high ABI values should not automatically be considered as an adverse marker of CVD in the general population. The Strong Heart Study represents the only study to date that has prospectively evaluated the clinical significance of a high ABI (>1.40) in detail [11]. In contrast with our findings, individuals in the Strong Heart Study with high ABI values had higher levels of several risk factors than did those with normal ABI values ( ). Older age, hypertension, and higher levels of triglycerides, total and LDL cholesterol, fibrinogen, hemoglobin A 1c, fibrinogen, and fasting glucose were observed in the Strong Heart Study s high ABI group. As in ARIC, the prevalence of diabetes in the Strong Heart Study was not different between the high and normal ABI groups. High ABI was associated with multivariable adjusted rate ratios for all-cause mortality and CVD mortality of 1.77 (95% CI: ) and 2.09 (95% CI: ), respectively, compared with the normal ABI group. Because the magnitude of mortality risk associated with high ABI was comparable to that of low ABI, the authors concluded that the prognosis of individuals with a high ABI was poor. Nevertheless, the Strong Heart Study comprised only American Indians, who were characterized on study entry by high rates of type 2 diabetes, microalbuminuria, and CVD. We also note that the high ABI group in the Strong Heart Study contained a substantial proportion with incompressible pedal arteries, and this subgroup appears to have contributed to the high CVD event rates in their high ABI group. The ABI distribution in ARIC, including a population that has a low prevalence of pre-existing PAD, diabetes, and renal disease, is probably more representative of the normal distribution of ABI values and outcomes that might be observed in general populations. The physiological mechanism(s) that contribute to calculation of a high ABI value, and that thus might subserve the observed dissociation between arm and ankle pressures is unknown. Traditionally, a high ABI value has been considered a consequence of an artifactual elevation in the ankle systolic pressure due to non-compliant pedal arteries. This might be true in individuals with concomitant diseases that are known to mediate such a local arterial effect (e.g., longstanding diabetes, advanced age, etc.). However, our data suggest that this mechanism does not contribute to a high ABI when measured in a community-derived population. In healthy individuals, a high ABI value is derived by a relatively low arm systolic pressure (i.e., a dissociation of the arm and ankle systolic pressures). In this dissociation of arm and ankle pressures, we note that the lower arm systolic pressure must be more closely associated with central aortic pressure, as there are no physiological mechanisms to artifactually lower arm pressure. The higher ankle pressure could be derived from either a less compliant lower extremity arterial system or a lower extremity arterial system that has greater numbers of bifurcations; both of these structural changes could occur without evident increases in IMT and would lead to augmentation of leg systolic pressure via increased reflected waves. Regardless of the mechanism of the arm-ankle dissociation, this finding could have clinical significance. Central aortic pressure is likely to mediate atherosclerosis progression in carotid and coronary arteries, and thus the low arm systolic pressure in individuals with a high ABI might confer a relative protective effect on ischemic events in these critical circulations. Although our population sample has broad generalizability, we acknowledge a series of study limitations. An important limitation is that this study did not directly assess medial arterial calcification. However, we believe it unlikely that medial calcification was common in this study because diabetes prevalence and serum creatinine did not differ in the high ABI group compared with the normal ABI group. A second limitation is that we measured ABI in only one randomly selected leg, and therefore some participants may have been misclassified. With a single measurement of ABI, high ABI values may be subject to regression to the mean due to intraindividual variation. However, when the ABI is measured in clinical settings, often only a single ABI measurement is performed. In clinical practice, as in clinical research, appreciation that the ABI might regress to the mean for both high and low ABI values in asymptomatic subjects suggests that ABI measurements might best be repeated in order to obtain the true ABI values. We conclude that, in this mostly asymptomatic, middleaged general population with a low prevalence of PAD, people with a high ABI are not characterized by a more adverse

6 464 K. Wattanakit et al. / Atherosclerosis 190 (2007) atherosclerosis risk factor profile and do not have greater CVD event rates than those with a normal ABI. Until more data are available, a high ABI value should not be considered a definitive marker of increased CVD risk in low risk populations. Acknowledgement The authors thank the participants and staff of the ARIC Study for their important contributions. References [1] Hirsch AT, Criqui MH, Treat-Jacobson D, et al. Peripheral arterial disease detection, awareness, and treatment in primary care. JAMA 2001;286: [2] Hiatt WR. Medical treatment of peripheral arterial disease and claudication. N Engl J Med 2001;344: [3] Zheng ZJ, Sharrett AR, Chambless LE, et al. Associations of anklebrachial index with clinical coronary heart disease, stroke and preclinical carotid and popliteal atherosclerosis: the Atherosclerosis Risk in Communities (ARIC) Study. Atherosclerosis 1997;131: [4] Newman AB, Shemanski L, Manolio TA, et al. Ankle-arm index as a predictor of cardiovascular disease and mortality in the Cardiovascular Health Study. The Cardiovascular Health Study Group. Arterioscler Thromb Vasc Biol 1999;19: [5] Leng GC, Fowkes FG, Lee AJ, Dunbar J, Housley E, Ruckley CV. Use of ankle brachial pressure index to predict cardiovascular events and death: a cohort study. BMJ 1996;313: [6] Newman AB, Siscovick DS, Manolio TA, et al. Ankle-arm index as a marker of atherosclerosis in the Cardiovascular Health Study. Cardiovascular Heart Study (CHS) Collaborative Research Group. Circulation 1993;88: [7] McDermott MM, Fried L, Simonsick E, Ling S, Guralnik JM. Asymptomatic peripheral arterial disease is independently associated with impaired lower extremity functioning: the women s health and aging study. Circulation 2000;101: [8] Young MJ, Adams JE, Anderson GF, Boulton AJ, Cavanagh PR. Medial arterial calcification in the feet of diabetic patients and matched nondiabetic control subjects. Diabetologia 1993;36: [9] Everhart JE, Pettitt DJ, Knowler WC, Rose FA, Bennett PH. Medial arterial calcification and its association with mortality and complications of diabetes. Diabetologia 1988;31: [10] Emanuele MA, Buchanan BJ, Abraira C. Elevated leg systolic pressures and arterial calcification in diabetic occlusive vascular disease. Diabetes Care 1981;4: [11] Resnick HE, Lindsay RS, McDermott MM, et al. Relationship of high and low ankle brachial index to all-cause and cardiovascular disease mortality: the Strong Heart Study. Circulation 2004;109: [12] The Atherosclerosis Risk in Communities (ARIC) Study: design and objectives. The ARIC investigators. Am J Epidemiol 1989;129: [13] Siedel J, Hagele EO, Ziegenhorn J, Wahlefeld AW. Reagent for the enzymatic determination of serum total cholesterol with improved lipolytic efficiency. Clin Chem 1983;29: [14] Nagele U, Hagele EO, Sauer G, et al. Reagent for the enzymatic determination of serum total triglycerides with improved lipolytic efficiency. J Clin Chem Clin Biochem 1984;22: [15] Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972;18: [16] Clauss A. Rapid physiological coagulation method in determination of fibrinogen. Acta Haematol 1957;17: [17] ARIC: Atherosclerosis: Risk in Communities Study: Manual 8:Lipid and Lipoprotein Determinations. Chapel Hill, NC: National Heart, Lung, and Blood Institute; [18] Mundt KA, Chambless LE, Burnham CB, Heiss G. Measuring ankle systolic blood pressure: validation of the Dinamap 1846 SX. Angiology 1992;43: [19] Lijmer JG, Hunink MG, van den Dungen JJ, Loonstra J, Smit AJ. ROC analysis of noninvasive tests for peripheral arterial disease. Ultrasound Med Biol 1996;22: [20] The ARIC Study Group. High-resolution B-mode ultrasound scanning methods in the Atherosclerosis Risk in Communities Study (ARIC). J Neuroimag 1991;1: [21] The ARIC Study Group. High-resolution B-mode ultrasound reading methods in the Atherosclerosis Risk in Communities (ARIC) cohort. J Neuroimag 1991;1: [22] Rose GABH, Gillum RF, Prineas RJ. Cardiovascular survey methods. Geneva, Switzerland: World Health Organization; [23] Casale PN, Devereux RB, Alonso DR, Campo E, Kligfield P. Improved sex-specific criteria of left ventricular hypertrophy for clinical and computer interpretation of electrocardiograms: validation with autopsy findings. Circulation 1987;75: [24] White AD, Folsom AR, Chambless LE, et al. Community surveillance of coronary heart disease in the Atherosclerosis Risk in Communities (ARIC) Study: methods and initial two years experience. J Clin Epidemiol 1996;49: [25] Rosamond WD, Folsom AR, Chambless LE, et al. Stroke incidence and survival among middle-aged adults: 9-year follow-up of the Atherosclerosis Risk in Communities (ARIC) cohort. Stroke 1999;30: [26] Hirsch AT, Hiatt WR. PAD awareness, risk, and treatment: new resources for survival the USA PARTNERS program. Vasc Med 2001;6:9 12. [27] Mohler 3rd ER, Treat-Jacobson D, Reilly MP, et al. Utility and barriers to performance of the ankle-brachial index in primary care practice. Vasc Med 2004;9: [28] Belch JJ, Topol EJ, Agnelli G, et al. Critical issues in peripheral arterial disease detection and management: a call to action. Arch Intern Med 2003;163: [29] Peripheral arterial disease in people with diabetes. Diabetes Care 2003;26: [30] Hirsch AT, Gloviczki P, Drooz A, Lovell M, Creager MA. Mandate for creation of a national peripheral arterial disease public awareness program: an opportunity to improve cardiovascular health. J Vasc Surg 2004;39: [31] Goss DE, de Trafford J, Roberts VC, Flynn MD, Edmonds ME, Watkins PJ. Raised ankle/brachial pressure index in insulin-treated diabetic patients. Diabet Med 1989;6: [32] Leskinen Y, Salenius JP, Lehtimaki T, Huhtala H, Saha H. The prevalence of peripheral arterial disease and medial arterial calcification in patients with chronic renal failure: requirements for diagnostics. Am J Kidney Dis 2002;40:472 9.