Atherosclerosis, 86 (1991) 31-37 0 1991 Elsevier Scientific Publishers ADONIS 0021915091000548

ATHERO

31 Ireland.

Ltd. 0021-9150/91/$03.50

04578

Extracranial

carotid artery stenosis: prevalence and associated factors in elderly stroke patients A.K. Admani

‘, D.M. Mangion

risk

2 and D.R. Naik 3

’Department of Medicine forthe Elderly, Northern General Hospital, Herries Road, Sheffield S5 7A lJ (il. K.) .’Department of Medicinefor the Elderly, Bamsley District General Hospital, Gawber Road, Barnsley S75 2EP (U.K.), and ’ Department of Radiology, Northern General Hospital, Herries Road, Sheffield S5 7A U (UK.) (Received 22 March, 1990) (Revised, received 31 August, 1990) (Accepted 13 September, 1990)

Summary The degree of arterial stenosis in both the right and left extracranial (common and internal) carotid arteries (ECAS) was assessed in 118 elderly (65 years and older; male 66) stroke patients by a Duplex Pulsed Wave ultrasound system (ATL 500). Of these, 33 (28%) had severe stenosis (755100%) 16 (14%) had moderate stenosis (30-74%) and 69 (58%) had no or minimal stenosis (O-29%). Overall, 34% (n = 40) of strokes were associated with moderate or severe stenosis in the ipsilateral extracranial carotid artery. Multivariate logistical regression analysis showed a significant (P < 0.05) positive correlation between ECAS (as assessed in either artery) and ischaemic heart disease, systolic blood pressure and male sex (multiple correlation coefficient, r = 0.240). The results were similar when the analysis was repeated for ECAS in the ‘clinically significant’ (maximal stenosis ipsilateral to the stroke) arteries only (multiple correlation coefficient, r = 0.276). The relationships, however, were weak, suggesting that other factors not identified in this study are more important aetiological factors for extracranial carotid artery atherosclerosis in the elderly.

Key words:

Extracranial

carotid

artery

stenosis;

Prevalence;

Introduction The risk of stroke increases with increasing severity of extracranial carotid artery stenosis (ECAS) [l-5] and large vessel disease was esti-

Correspondence to: Dr. D.M. Mangion, Department of Medicine for the Elderly, Bamsley District General Hospital. Gawber Road, Bamsley S75 2EP. U.K.

Risk factors;

Stroke;

Elderly

mated to cause 34% of all strokes in one study, on clinical and laboratory (angiography, computerised axial tomogrophy and post-mortem) findings [6]. Age is a recognised risk factor for cerebral infarction [7,8] and has been found to be an important risk factor for ECAS in most studies where it has been analysed as a risk factor [4,9-131. Despite the apparent importance of ECAS in the

32 aetiology of stroke in the elderly, little data is available on its contribution to stroke and associated risk factors in this age group. Studies in younger age groups suggest that hypertension [4,12-141, systolic blood pressure [4], diabetes mellitus [9], cigarette smoking [10,12,13,15], serum total cholesterol [4,15,16], low density lipoprotein cholesterol [ll] and triglycerides [16,17] are risk factors, while serum high density lipoprotein cholesterol [ll-131 has a protective effect on ECAS. Duplex pulsed wave ultrasonography is a safe, reproducible and sensitive method of visualising the carotid arteries [18]. The purpose of this study was to assess the prevalence of ECAS, and associated risk factors, in elderly (2 65 years old) stroke patients. Patients and methods One hundred and eighteen white Caucasian patients (age range 65-89 years, mean 74.5 years, SEM 6.9; males 66) admitted consecutively to the Stroke Unit, Northern General Hospital, over a 16month period with a diagnosis of first ever stroke in the carotid territory were studied. Patients had a full clinical history and examination, resting electrocardiograph, chest radiograph, routine haematological (Coulter-T890) and biochemical tests, including a post prandial venous plasma glucose (Parallel Systems, American Monitor); patients with a value L 6.5 mmol/l had an oral glucose tolerance test using a 75-g glucose load. Duplex pulsed wave ultrasonography (ATL 500) of both right and left common and extracranial internal carotid arteries, using a 7.5 MHz transducer with combined 2 dimensional real-time and Doppler capabilities, was carried out between 1 and 2 weeks after the event. All scans were performed by 2 experienced ultrasonographers and videotapes were reviewed by an experienced radiologist with a special interest in vascular imaging (D.R.N.). The assessment of the degree of stenosis was based on the Khz shift and the wave form. Real time imaging allowed accurate placement of sample volume and assessment of the degree of The degree of extracranial carotid narrowing. artery stenosis (ECAS) was assessed as (A) minimal O-29%, (B) moderate 30-74% and (C)

severe 75-100% [19]. Identification of the common, external and internal carotid arteries was possible in all cases. Cerebral angiography confirmed a Duplex ultrasound diagnosis of severe stenosis in all 5 patients referred for carotid endarterectomy following minor stroke in this series. Patients had a fasting serum lipid estimation at least 3 months (median 3.3 months) after the episode, in the absence of acute ongoing illness [20,21]. Serum total cholesterol (TC) was measured according to the method of Allain et al. [22] (Gilford Systems) and high density lipoprotein cholesterol (HDL-C) by the method of Finley [23] (Gilford Systems). Serum triglycerides (TG) were measured using a modification of the method of Wahlefeldof [24] (Gilford Systems). Low density lipoprotein cholesterol (LDL-C) was estimated by the Friedewald equation (LDL-C = TC-TG/ 2.21-HDL-C) [25]. Of 142 eligible patients admitted during the period of study, 24 (17%) were excluded on the following criteria: (a) unable or unwilling to give informed consent (n = 16); (b) a history of recent weight loss or on a low fat diet or lipid lowering agents at the time of the stroke or serum lipid estimation (n = 6); (c) previous carotid artery surgery (n = 2).

Definitions Stroke - sudden or rapid development of a neurological deficit of presumed vascular origin lasting more than 24 h [26]; for the purposes of this study, ‘stroke’ refers to the cerebral lesion. Computerised axial tomography (CAT) was performed infrequently and was done primarily to exclude tumour or sub dural haematoma, since obtaining scans at a standard time after the stroke was difficult. Ischaemic heart disease (IHD) - (a) a history of myocardial infarction with a documented diagnosis on WHO criteria [27]; or (b) codable Q/QS pattern on a 12-lead electrocardiogram (Minnesota codes l-l-l to l-3-6 inclusive) [28]; or (c) a history of angina elicited by the Rose questionnaire [28]. Peripheral vascular disease - pain or discomfort in the calves on exertion and relieved by rest ]291.

33 Systolic (SBP) and diastolic (DBP) (Korotkoff phase V) blood pressures were measured in the sitting position during the first 2 weeks after the stroke using a standard mercury sphygmomanometer, before the initiation or change of antihypertensive treatment. The mean of 3 readings on successive days was used for analyses. History of previous hypertension - documented history from hospital or general practitioner records and evidence of antihypertensive treatment. Diabetes mellitus and impaired glucose tolerance (IGT) - a history of diabetes and evidence of treatment with insulin or oral hypoglycaemic agents and/or diet; or when diabetes or IGT was diagnosed after stroke on WHO criteria [30]. Current cigarette smoking was expressed in cigarette years. Ex-smokers were classified as non smokers for logistic regression analyses if they had stopped smoking for at least 1 year: otherwise they were classified as smokers. Body mass index (BMI) was calculated as weight (kg)/height (m)’ [31]; males > 27.8 kg/m2 and females 2 27.3 kg/m2 were defined as obese ]321. Statistical analysis was by Student’s t test, chisquare, linear regression, univariate (ULR) and multivariate (MLR) logistic regression analyses [32] using the Statistical Analysis System program [33]. MLR was carried out using a maximum likelihood stepwise backward elimination procedure. Because of the non-normal distribution of serum TG, the logarithm value (log TG) was used for ULR and MLR. Since TC correlated strongly with HDL-C and LDL-C on multiple linear regression analysis (multiple correlation coefficient, r = 0.922, P < O.OOOl), TC was included in one model while HDL-C and LDL-C were included in another model for MLR. Peripheral vascular disease was not included in the logistical analyses as the number was small. Independent variables analysed as risk factors were age, sex, TC, HDL-C, LDL-C, log TG, IHD, SBP, DBP; history of hypertension, diabetes mellitus/ IGT, cigarette years and BMI. Logistical regression analyses were performed twice. To relate risk factors to ECAS, the artery with the more severe side of stenosis (either side) was used in one analysis. To further assess risk

factors in those arteries thought to be ‘clinically significant’, only those arteries with maximal stenosis ipsilateral to the stroke were used in a second analysis. The level of statistical significance was taken as P 2 0.05throughout. Results Of the 118 stroke patients, 33 (28%) had severe ECAS, 16 (14%) had moderate ECAS and 69 (58%) had no or minimal ECAS. The side of stroke was associated with the side of maximal stenosis in 29 (88%) in the group with severe stenosis and 11 (69%) in the group with moderate stenosis (P > 0.05). Twenty-two patients (19%) (males = 12) had a TC level > 6.5 mmol/l), 48 (41%) (males = 31) had an HDL-C level < 1.0 mmol/l, 14 (12%) (males = 6) had an LDL-C level > 4.5 mmol/l, 35 (30%) (males = 21) had a TG level > 2.0.mmol/l, 26 (22%) had a history of IHD (males = 16), 8 (5%) (males = 6) had a peripheral vascular disease, 24 (20%) (males = 14) had a history of hypertension, 26 (22%) (males = 11) were diabetic or had IGT (n = 4) 43 (36%) (males = 29) were cigarette smokers, 18 (males = 15) were ex-cigarette smokers while 27 (23%) were obese (males = 9). There was no significant (P < 0.05) sex difference in the prevalence of these conditions, except obesity (males = 14%, females = 35%; P = 0.01) and ex cigarette smokers (males = 23%, females = 6%; P = 0.01). There was no significant (P > 0.05) sex difference in the prevalence of cholesterol levels > 6.5 mmol/l when patients 65-75 years of age and 76 years and older were analysed separately (65575 years males 5/24, females 7/33; 2 76 years males 7/42, females 3/19). Mean levels of serum TC, HDL-C, LDL-C and log TG (expressed as antilog t- SEM), mean SBP and DBP and patient characteristics in the different groups are shown in Table 1. In the first analysis, relating ECAS in either artery to risk factors, positive association was found between ECAS and IHD, SBP and log TG on ULR (Table 2). None of the other variables showed significant association (P > 0.05). On MLR the two models (using TC in one and HDL-C and LDL-C in the other) showed identical results. ECAS was positively associated with IHD, SBP

34 TABLE

1

MEAN VALUES (*ISEM) FOR TC, HDL-C, LDL-C, TG, SBP AND DBP, AND PREVALENCE OF CIGARETTE SMOKING AND HISTORY OF HYPERTENSION IN DIFFERENT DEGREES OF ECAS Degree ECAS

Minimal

No. (males) TC (mmol/I) (3.4-6.5) HDL-C (mmol/l) ( > 1.0) LDL-C (mmol,‘l) ( < 4.5) Log TG (antilog + SEM)

140.80 83.75

Diabetic (X) Cigarette smoking

Moderate

69 (34) 5.30 (0.15) 1.12 (0.04) 3.40 (0.13) 1.60 (1.05)

SBP (mm Hg) DBP (mm Hg) History of hypertension IHD (W)

(o-294;)

,

(W)

l l-20/day 21 +/day Ex-smokers No. obese (%) * P < 0.05 compared

33 (22) 5.72 (0.27) 1.08 (0.05) 3.50 (0.20) 2.07 (1.1) * 148.80 (3.42) 86.51 (2.28)

12 9

(17) (13)

5 4

(31) (25)

7 13

11

(16)

2

(13)

9

to ‘minimal’

8

(12)

6

(37) *

9

(13)

1: 17

$; (25)

1 0 1 3

(6) (0) (6) (19)

ECAS. Values in parenthesis

male sex (Table 2). To assess the influence of age on the relationship between ECAS and IHD, age was reintroduced in the final model; IHD remained as a significant independent variable and

(21) (39) * (27)

are normal

ranges

11 4 1 3 7

(33) * (12) (3) (9) (21)

for this laboratory.

(beta coefficient = 1.2962; SE 0.4596; correlation coefficient, r = 0.163; P = 0.005), as did SBP and male sex (P < 0.05). When the analysis was repeated without including IHD as an independent

2

UNIVARIATE AND MULTIVARIATE REGRESSION EITHER SIDE AND (B) ARTERIES WITH MAXIMAL Significant (1) Univariate (a)

(b)

Severe (75-1008)

16 (10) 5.31 (0.39) 1.09 (0.07) 3.53 (0.35) 1.42 (1.1) 155.00 (6.83) * 87.50 (2.99)

(2.82) (1.64)

DIABETES,

(Sg)

l-lo/day

TABLE

(30-74%)

IHD,

variables

results IHD Log TG SBP IHD SBP Log TG male sex

(2) Multivariate results (a) IHD SBP Male sex Multiple correlation coefficient, (b) IHD SBP Male sex Multiple correlation coefficient,

ANALYSES FOR (A) ARTERIES WITH MAXIMAL STENOSIS IPSILATERAL TO THE STROKE

Beta coefficient

Standard

1.2728 2.3516 0.0149 1.3764 0.0179 2.2159 0.8861

0.4333 1.0106 0.0076 0.4838 0.0089 1.0656 0.4139

1.3372 0.0224 0.8615 r = 0.240, P = 0.00@3 1.6461 0.0257 1.0745 r = 0.276; P = 0.0001.

0.4529 0.0076 0.4091 0.5226 0.0097 0.4494

error

_

r

P

0.173 0.124 0.087 0.179 0.104 0.111 0.116

0.003 0.02 0.05 0.004 0.04 0.04 0.03

0.174 0.151 0.105

0.003 0.008 0.03

0.204 0.164 0.140

0.002 0.008 0.02

STENOSIS,

35 variable, log TG (beta coefficient = 2.3614, SE = 1.0558, r = 0.116, P = 0.02) SBP (beta coefficient = 0.0189, SE = 0.0083, r = 0.120, P = 0.02) and male sex (beta coefficient = 0.8900, SE = 0.4029, r = 0.114, P = 0.03) were independent variables on MLR (multiple correlation coefficient, Y= 0.192, P = 0.003). When the analyses were repeated for ‘clinically significant’ arteries only, the results were similar except that male sex was a significant variable on ULR (Table 2). IHD remained as a significant variable when age was reintroduced in the final model (beta coefficient 1.604, SE = 0.532, r = 0.193; P = 0.003), as did SBP and male sex (P -c 0.05). Discussion This study, in unselected elderly stroke patients, demonstrates a high prevalence (41%; 95% CL 32-50%) of carotid artery disease (2 30% stenosis) compared to that in asymptomatic elderly (11%; 95%CL 9-13%), none of whom had severe (2 80%) stenosis [34]. In the present study, 28% (95% CL 20-369;;) of patients had severe stenosis, representing a 50% reduction in blood flow [35] and commonly held to be associated with an increased risk of stroke in asymptomatic patients [2-51. Moreover, 34% (95% CL 26-42) of strokes in this study were associated with moderate or severe stenosis in the ipsilateral carotid artery, comparable to the prevalence (28%; 95%CL 1441%) demonstrated by angiography in younger stroke patients [36]. In the present study, ECAS, as assessed in either artery, correlated positively with IHD, SBP and male sex on multivariate analyses. The results were similar when the analysis was repeated for only those arteries with maximal stenosis ipsilateral to the stroke, suggesting that these may be risk factors for cerebral infarction due to large vessel disease in the elderly. However, CAT scanning was not routinely employed for diagnosis of the type of stroke. A recent study [37] in a predominantly asymptomatic [75%) elderly (mean age = 82 years) population suggested that diabetes, hypertension (SBP 2 160 mm Hg, DBP 2 90 mm Hg) and cigarette smoking were significant risk factors for ECAS on univariate analyses but no

assessment of the possible influence of IHD was made [38]. Multivariate analysis allows the assessment of the independent contribution of the various variables to ECAS. In this study, IHD, of all the independent variables, showed the strongest correlation with ECAS, and the relationship was independent of age. In younger patients, Crouse et al. [13] demonstrated a positive correlation between ECAS and coronary stenosis diagnosed on coronary angiography, though not with a history of IHD using similar diagnostic criteria. It is well recognised that the major cause of death in patients with ECAS is ischaemic heart disease [3-51. The correlation between IHD and ECAS suggests that carotid and coronary atherosclerosis share common risk factors, unidentified in this study. Hyperinsulinaemia [39] and hyperfibrinogenemia [40], which may be related to atherosclerosis, were not assessed in this study. Although it has been suggested that hypertension is a more important risk factor for intracranial small vessel atherosclerosis [17], positive correlation was found on multivariate analysis between ECAS and SBP, though not with a history of hypertension, in this study. A history of hypertension was not related to progression of the severity of ECAS [5]. Hypertension and a history of hypertension have been considered together in other studies using multivariate analysis [10,12, 13,151, with inconsistent results. However, none of these studies [10,12,13,15] mentions the degree of antihypertensive control and different definitions for hypertension were used. Epidemiological studies [8,41-431 have demonstrated the importance of systolic blood pressure as a risk factor for cerebral infarction, especially in the elderly [8]. In younger patients, male sex was a risk factor for ECAS [3-5,16,44]. While the female protection against cerebral infarction declines with age [8], this study suggests that male sex is an independent risk factor for ECAS in the elderly. Females, however, may be more susceptible to intracranial arterial occlusive disease [44]. It has been suggested that lipoprotein abnormalities may be a more important risk factor for ECAS in the elderly than in younger subjects [38]. In this study no correlation was found between ECAS and TC, HDL-C or LDL-C in contrast to

36 findings in younger patients [ll-161. Even for JHD, where the relationship with lipid abnormalities is stronger than for cerebral infarction, the effect with lipids declines with age [8]. The results are consistent with those from prospective epidemiological studies [8,41-43,45,46]. In this study, log TG were not a significant risk factor for ECAS when IHD was included in the analysis. Triglycerides may be a risk factor for IHD in patients over the age of 50 years [47] but have been inconsistently related to cerebral infarction [8,41]; log TG were not a risk factor for ECAS in a recent study [13]. While transient changes in lipid levels are well known to occur after stroke [20,21], making comparison between different studies difficult, care was taken to obtain blood samples for lipid levels at a standard time in all patients. Unlike other studies [4,9-131, age was not a significant risk factor in the present study, probably reflecting the narrow age range studied. Other non significant risk factors for ECAS on multivariate analysis were diabetes/IGT, BMI and cigarette smoking. While diabetes increased the risk of cerebral infarction in prospective epidemiological studies [8,41,43], other studies using multivariate analysis have also not found diabetes to be a risk factor for ECAS [10,12,13,16]. Although not all patients had a formal glucose tolerance test, it is unlikely that our method of screening would have missed a significant number of diabetics in this age group [48]. Studies in younger patients have noted an association between cigarette smoking and ECAS on multivariate analysis [10,12,13, 161. However, the importance of cigarette smoking as a risk factor for cerebral infarction declines with age [49]. Obesity is related to hypertension, diabetes and lipid abnormalities but has not been found to be an independent risk factor for atherosclerosis [50], ECAS [13,15] or cerebral infarction [8,42,45]. This study demonstrates that carotid and coronary atherosclerosis share common risk factors, while SBP and male sex are independent risk factor for ECAS. However, the relationships demonstrated in this study are weak suggesting that other unidentified factors, possibly including intrinsic cellular ageing processes and genetic determinants [51] may be more important.

References 1

Harrison, M.I.G. and Marshall, J., Prognostic significance of of severity of carotid atheroma in early manifestations

cerebrovascular disease. Stroke, 13 (1982) 567. 2 Roederer, G.O., Langlois, Y.E., Jager, K.A. et al., The natural history of carotid arterial disease in asymptomatic patients with cervical bruits. Stroke, 15 (1984) 605. 3 Chambers, B.R. and Norriss, J.W., Outcome in patients with asymptomatic neck bruits. NFJM, 315 (1986) 860. 4 Autret, A., Saudeau, D., Bertrand, P., Pourcelot, L., Marchal, C. and De Boisvilliers, S., Stroke risk in patients with carotid stenosis. Lancet, 1 (1987) 888. 5 Hennerici, M., Hulsbomer, H.B., Hefter, H., Lasmmerts, D. and Rautenberg. W., Natural history of asymptomatic extracranial arterial disease. Brain, 110 (1987) 777. 6 Mohr, J.P., Caplan, L.R., Melski, J.W. et al., The Harvard co-operative stroke registry: a prospective study. Neurology. 28 (1978) 754. 7 Louis, S. and McDowell, F., Age: its significance in nonembolic cerebral infarction. Stroke, 1 (1970) 449. 8 Wolf, P.A., Kannel, W.B. and McGee, D.L., Prevention of ischaemic stroke: disk factors. In: Bamett, J.H.M., Stein, B.M., Mohr, J.P. and Yatsu, F.M. (Eds.), Stroke: Pathophysiology, Diagnosis and Management, Churchill Livingstone, Edinburgh, 1987, pp. 967-988. 9 Ghan, A., Beach, K.W., Martin, D.C. and Standness, D.E., Carotid artery disease in NIDDM Diabetes. Diabetes Care, 6 (1983) 562. 10 Candelise, L., Bianchi, F., Galligoni, F. et al., Italian multicentre study on reversible cerebral ischaemic attacks. III. Influence of age and risk factors on cerebral atherosclerosis. Stroke, 15 (1984) 379. 11 Postiglione, A., Rubba, P., De Simone, B., Patti, L., Cicerano, U. and Mancini, M., Carotid atherosclerosis in familial hypercholesterolemia. Stroke, 16 (1985) 658. 12 Ford, SC., Crouse, J.R., Howard, G., Toole, J.F., Ball, M.R. and Frye, J., The role of plasma lipids in carotid bifurcation atherosclerosis. Ann. Neurol., 17 (1985) 301. 13 Crouse, J.R., Toole, J.F., McKinney, W.M. et al., Risk factors for extracranial carotid artery atherosclerosis. Stroke, 18 (1987) 990. 14 Van Merode, T., Hick, P., Hoeks, A.P.G. and Reneman, R.S., Serum HDL/Total Cholesterol ratio and blood pressure in asymptomatic atherosclerotic lesions of the cervical carotid arteries in men. Stroke, 16 (1985) 34. 15 Passero, S., Rossi, G., Nardini, M. et al., Italian multicentre study on reversible cerebral ischaemic attacks. Part 5. Risk factors and cerebral atherosclerosis. Atherosclerosis, 63 (1987) 211. 16 Bansal, B.C., Gupta, R.R., Bansal, M.R. and P&ash, C., Serum lipids and uric acid relationship in ischaemic thrombotic vascular disease. Stroke, 6 (1975) 304. K., 17 Mathew, N.T., Davis, D., Meyer, J.S. and Chandar, Hyperlipoproteinemia in occlusive cerebrovascular disease. JAMA, 232 (1975) 262. 18 Feussner, J.R. and Matchar, D.B., When and how to study the carotid arteries. Ann. Intern. Med., 109 (1988) 805.

31 19 Jacobs, N., Grant, E., Schlesinger, D. et al., Duplex carotid sonography: criteria for stenosis, diagnosis and pitfalls. Radiology, 154 (1985) 385. 20 Hollanders, F.D., Shafar, J. and Burton, P., Serum lipid changes following the completed stroke syndrome. Post. Grad. Med. J.. 51 (1975) 386. 21 Mendez, I.. Hachinski, V. and Wolfe, B., Serum lipids after stroke. Neurology, 37 (1987) 507. 22 Allain, C.C., Poon, L.S., Chan, S.C.G., Richmond, W. and Fu, P.C.. Enzymatic determination of total serum cholesterol. Clin. Chem., 20 (1974) 470. 23 Finley, P.R., Schifman, R.B., Williams, R.J. and Lichti, D.A., Cholesterol in HDL: use of Mg/Dextran sulphate in its enzymatic determination. Clin Chem., 24 (1978) 931-933. 24 Wahlefeld. A.W.. Triglycerides: Determination after enzymatic hydrolysis. In: Bergemeyer, H.U. (Ed.), Methods of Enzymatic Analysis, Academic Press, New York, 1974, pp. 1831-1835. 25 Friedewald. W.T., Levy, RI. and Fredrickson, D.S., Estimation of the concentration of low density lipoprotein in plasma without the use of preparative ultracentrifuge. Clin. Chem., 18 (1972) 499. 26 Hatano, S., Experience from a multi centre stroke register: a preliminary report. Bull. WHO, 54 (1976) 541. 27 World Health Organisation Regional Office for Europe: Public Health in Europe: 5. Myocardial Infarction community register. WHO Annexe, Copenhagen, 1 (1976). 28 Rose. G.A. and Blackburn, H., Cardiovascular survey methods. WHO, (1978) 137. 29 Hale, W.E., Marks, R.G.. May, F.E., Moore, M.T. and Stewart, R.B., Epidemiology of intermittent claudication: evaluation of risk factors. Age Ageing, 17 (1988) 57. 30 World Health Organisation, Expert Committee on Diabetes. Technical Report Series No. 646, WHO, Geneva, 1978. 31 Garrow. J.S. and Webster, J., Quetelet’s index as a measure of fatness. Int. J. Obes., 9 (1985) 147. 32 National Institute of Health. Health implications of obesity: consensus conference statement. Ann. Intern. Med., 103 (1985) (pt 2) 1073. 32 Armitage, P. and Berry, G., Statistical Methods in Medical Research. 2nd edn., Blackwell Scientific Publications, Oxford, 1987. 33 SAS Institute Inc., Sas User’s Guide: Basics Version 5. SAS Institute Inc., Gary, NC, 1985. 34 Schoenfeld. M.R., Aronow, W.S. and Paul, P., Are strokes more likely to result from severe carotid atherosclerotic stenosis? If not, why not? Angiology, 38 (1988) 720. 35 Benetos, A., Simon, A., Levenson, J., Lagneau, P., Bouthier, J. and Safar, M., Pulsed Doppler: an evaluation of diameter. blood velocity blood flow of the common carotid artery in patients with isolated unilateral stenosis of the internal carotid artery. Stroke, 16 (1985) 969.

36 Harrison, M.J.G. and Marshall, J., Angiographic appearance of carotid bifurcation in patients with completed stroke, transient ischaemic attacks, and cerebral tumour. Br. Med. J., 1 (1976) 205. 37 Aronow, AS., Schoenfeld, M.R. and Paul, P.. Risk factors for extracranial internal or common carotid arterial disease in persons aged 60 years and older. Am. J. Cardiol., 63 (1989) 881. 38 Tell, G.S., Crouse, J.R. and Furberg, C.D.. Relation between blood lipids, lipoproteins, and cerebral atherosclerosis. Stroke, 19 (1988) 423. 39 Stout, R.W., Insulin and atheroma: 20 year perspective. Diabetes Care, 13 (1990) 631. 40 Welin. L., Svardsudd, K., Wilhelmsen, L., Larsson, B. and Tibblin, G., Analysis of risk factors for stroke in a cohort of men born in 1913. NEJM, 317 (1987) 521. 41 Salonen, J.T., Puska, P., Tuomilehetto, J. and Homan, K., Relation of blood pressure, serum lipids. and smoking to the risk of cerebral stroke - longitudinal study in Eastern Finland. Stroke, 13 (1982) 327. 42 Tanaka, H., Ueda, Y., Hayashi, M. et al.. Risk factors for cerebral haemorrhage and cerebral infarction in a Japanese rural community. Stroke, 13 (1982) 62. 43 Kagan, A., Popper, J.S., Rhoads, M.D. and Yano. K., Dietary and other risk factors for stroke in Hawaiin Japanese men. Stroke, 16 (1985) 390. 44 Caplan, L.R., Gorelick, P.B. and Hier. D.B., Race, Sex and occlusive cerebrovascular disease: a review. Stroke, 17 (1986) 648. 45 Ostfeld, A.M., Shekelle, R.B., Klawans, H. and Tufo, H.M., Epidemiology of stroke in an elderly welfare population. Am. J. Publ. Health. 64 (1974) 450. 46 Reed, D., Yano, K. and Kagan. A., Lipids and lipoproteins as predictors of coronary heart disease, stroke and cancer in the Honolulu Heart Program. Am. J. Med., 80 (1986) 871. 47 Castelli, W.P., The triglyceride issue: a view from Framingham. Am. Heart J., 112 (1986) 432. 48 Marble, A., Diabetes in the Elderly. In: Marble, A., Krall, L.P.. Bradley, R.F., Chriestlieb, A.R., Soeldner, J.S. (Eds.), Joslin’s Diabetes Mellitus, 12th edn., Lea and Febiger, Philadelphia, PA, 1985, pp. 520-525. 49 Shinton. R. and Beevers, G., Meta-analysis of relation between cigarette smoking and stroke. Br. Med. J., 298 (1989) 789. 50 Barrett-Connor, E.L., Obesity, atherosclerosis and coronary artery disease. Ann. Intern. Med., 103 (1985) 1010. 51 Bierman, E.. Ageing and atherosclerosis In: Stout, R.W. (Ed), Arterial Disease in the Elderly, Churchill Livingstone, Edinburgh, 1984, pp. 17-31.