Diabetes Mellitus and Cerebrovascular Disease - Diabetes Care

D

iabetes Mellitus and Cerebrovascular Disease: Prevalence of Carotid Artery Occlusive Disease and Associated Risk Factors in 482 Adult Diabetic Patients THOMAS W. KUEBLER, PHILLIP J. BENDICK, S. EDWIN FINEBERG, OMKAR N. MARKAND, JAMES A. NORTON, JR., FRANK N. VIN1COR, AND CHARLES M. CLARK, JR.

The prevalence of carotid artery occlusive disease (CAOD) in a population of adults with diabetes mellitus was determined, and factors associated with its presence were identified. By oculoplethysmography and phonoangiography, 20% of the population studied had detectable CAOD. Univariate analyses revealed that patients with CAOD had a higher prevalence of retinopathy proteinuria, a slower mean conduction velocity in two of three nerves tested, and more atherosclerosis in the leg arteries. Furthermore, increased age, higher cholesterol, lower fasting insulin, higher systolic blood pressure, and less adiposity were associated with the presence of CAOD. Sex, race, type of diabetes (type I or II), duration of diabetes, and measures of glycemia were not related to CAOD. When 10 variables were analyzed by multivariate methods, higher systolic blood pressure, higher cholesterol, and reduced adiposity were found to be the combination most significantly related to CAOD. DIABETES CARE 6.- 274278, MAY-JUNE 1983.

I

n general populations, diabetes mellitus has been reported to be a risk factor for cerebrovascular disease.1"3 Stroke patients with diabetes, especially those with thromboembolic stroke, appear to have a poorer prognosis.4 The purpose of this study was to identify which variables in an adult diabetic population best predicted who would have carotid artery occlusive disease (CAOD). To answer this question we examined the relationship between CAOD and a series of variables found in general populations to be associated with stroke, which included age, sex, race, blood pressure, cholesterol, and smoking history. We also assessed the relationship between CAOD and those variables associated with diabetes, which included duration of diabetes, neuropathy, retinopathy, proteinuria, peripheral vascular disease, plasma glucose, glycosylated hemoglobin, insulin, C-peptide, adiposity, amputations, type of diabetes, and antidiabetic treatment. METHODS

Cross-sectional data were collected by the Diabetes Research and Training Center of the Indiana University School of Medicine. Four hundred and eighty-two diabetic subjects were recruited from the General Medicine Clinic of Wishard Memorial Hospital. Informed consent was obtained. All pa-

274

tients met explicit glucose criteria, which included a fasting plasma glucose greater than 150 mg/dl or random glucose greater than 300 mg/dl. Some baseline characteristics of the study patients are listed in Table 1. The patients were classified as having either insulin-dependent diabetes mellitus (IDDM or type I diabetes) or non-insulin-dependent diabetes mellitus (NIDDM or type II diabetes), using criteria of the National Diabetes Data Group. s Each patient was assessed by oculoplethysmography and carotid phonoangiography. Oculoplethysmography indirectly assesses carotid arterial blood flow by detecting volumetric changes in each eye. A delay in the volume change in one eye relative to the other suggests a hemodynamically significant stenosis in the internal carotid artery on the side with the delay. In carotid phonoangiography, an oscilloscope records bruits detected by a microphone. Positive noninvasive tests indicate a stenosis significant enough to affect hemodynamic flow patterns; these have been shown to be at least a 40% diameter or 20% flow reduction.6 An estimate of the degree to which a stenosis exceeds this level can be made based on the amplitude of the pulse delay on plethysmography and the nature of the bruit recorded on the carotid phonoangiogram. These tests are associated with little risk and have been reported to have an overall accuracy of 90% compared with angiography.67 On the basis of the two tests,

DIABETES CARE, VOL. 6 NO. 1, MAY-JUNE 1983

DIABETES MELLITUS AND CEREBROVASCULAR DISEASE/THOMAS W. KUEBLER AND ASSOCIATES

TABLE 1 Baseline characteristics of diabetic patients in the sample Median ARC (yr)

Duration of diabetes (yr) Fasting plasma glucose (mg/dl) 2-h postprandial plasma glucose Glycosylated hemoglobin (%) Percent of ideal body weight

Sex Male Female Race Black White Type of diabetes N1DDM 1DDM NIDDM by treatment Oral hypoglycemic agent Diet Insulin NIDDM by obesity Obese Nonobese

;mg/dl)

Mean

± SD

12 57 8. 8 ± 12 220 94 339 + 113 10. 7 ± 3.1 143 31

58 6.3 210 334 10.4 139

•+•

No. of cases

482 480 458 443 472 476

Percent

No. of cases

21 79

102 380

71 29

344 138

95 5

447 22

11 20 69

53 96 333

81 19

358 85

patients were classified into two categories: having detectable CAOD or having no detectable CAOD. Patients having detectable CAOD were further classified as having mild, moderate, or severe stenosis. Blood flow was assessed noninvasively in the lower extremities using an ultrasound Doppler flowmeter, segmental systolic pressures, and plethysmography.8 Data relating to sex, race, age, duration of diabetes, smoking, and hypertension were collected. Percent ideal body weight (%IBW) was estimated from the (1959) Metropolitan Life Insurance Company Tables of Desirable Weights.9 Percent body fat was calculated from skinfold measurements.l0''' Ocular fundi were examined with an ophthalmoscope after mydriasis was induced by 1% cyclopentolate hydrochloride or 1% tropicamide. Duplicate plasma glucose determinations were obtained after an overnight fast and again 2 h after a 75-g glucose load. Plasma glucose was measured by the glucose-oxidase method using a Beckman Glucose Analyzer (Beckman Instruments, Fullerton, California). Duplicate samples were assayed for glycosylated hemoglobin (A la+b+c ) using an adaptation of the method of Trivelli.12 Serum creatinine was determined by the picric acid method, and cholesterol was determined by the cholesterol esterase method. Proteinuria was determined by Uristix. Serum insulin concentrations were measured in those individuals treated with diet and oral hypoglycemic agents.13 C-peptide concentrations were measured in those patients who were treated with insulin using a

double-antibody technique and monoiodinated tyrosylated synthetic human C-peptide and synthetic human C-peptide standards. Plasma was extracted with polyethylene glycol before C-peptide determinations.H The maximum motor conduction velocities along the right median nerve (elbow to wrist segment) and the right common peroneal nerve (knee to ankle segment) were determined in the conventional manner. The orthodromic sensory conduction velocity along the distal part of the right median nerve was determined by a technique similar to that described by Cohen and Brumlik15 and Buchthal and Rosenfalck.16 The research design was that of a cross-sectional prevalence study designed to estimate the prevalence of CAOD and to identify variables associated with the presence of detectable CAOD. The data were initially analyzed by chisquare or Fisher exact test for 2 by 2 contingency tables or with Student's t test for differences between means. To detect which variables in combination best differentiate patients with CAOD from those without, stepwise multiple discriminant analysis was performed. RESULTS

Prevalence of C A O D and univariate findings. Using the defi-

nition of detectable CAOD previously described, 94 (20%) adult diabetic patients out of 482 had CAOD in either or both carotid arteries. Thirty-three patients (7%) had bilateral CAOD and 61 patients (13%) had unilateral disease, 31 with


275

DIABETES MELLITUS AND CEREBROVASCULAR DISEASE/THOMAS W. KUEBLER AND ASSOCIATES

TABLE 2 Associations between CAOD and other discontinuous variables Variable

CAOD

No CAOD

P

Total cases

Sex (% women) Race (% black) Prevalence of history of hypertension Prevalence of cases who never smoked cigarettes Prevalence of retinopathy Prevalence of proteinuria Prevalence of amputations Prevalence of aortoiliac ASO Prevalence of leg ASO Prevalence of trifurcation ASO Prevalence of obesity (> 120% IBW, N1DDM only)

76% (71/94)' 67% (63/94)

80% (308/387) 72% (280/387)

NSt NS

481 481

69% (65/94)

67% (259/387)

NS

481

65% 60% 16% 3% 18% 40% 11%

63% 47% 7% 2% 12% 22% 6%

NS

NS

482 478 426 471 471 432 344

0.004

443

(61/94) (56/93) (14/86) (3/91) (17/93) (33/83) (6/56)

70% (60/86)

(246/388) (181/385) (25/340) (7/381) (45/378) (77/349) (16/288)

0.03 0.009

NS NS (0.08) 0.001

84% (298/357)

'Number of cases with attribute/total number of cases. tNS: not statistically significant.

right carotid lesions and 30 with left carotid lesions. Seventy' six patients were classified as having mild stenosis; 14, moderate; and 4, severe. The prevalence of CAOD was not significantly related to sex, race, or smoking history (Table 2). As seen in Table 3 the prevalence of CAOD was significantly related to increased age. The mean systolic pressure was 148 mm Hg among those with CAOD and 138 mm Hg

among those without (P < 0.001). However, diastolic blood pressure and history of hypertension did not differ significantly between the two groups. Patients with CAOD had a significantly higher mean cholesterol (P = 0.04)The prevalence of retinopathy was 60% among those with CAOD and 47% among those without CAOD (P = 0.03); the prevalence of proteinuria was 16% among those with

TABLE 3 Associations between CAOD and other continuous variables

CAOD

Variable Age (yr) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Creatinine (mg/dl) Cholesterol (mg/dl) Median nerve motor conduction velocity (m/s) Common peroneal nerve motor conduction velocity (m/s) Median nerve sensory conduction velocity (m/s) Fasting insulin (fxU) Postprandial insulin (jiU) Fasting plasma glucose (mg/dl) Postprandial plasma glucose (mg/dl) Fasting C-peptide (ng/ml) Postprandial C-peptide (ng/ml) Glycosylated hemoglobin (%) Percent ideal body weight (%) Percent fat (%) Duration of diabetes (yr)

60 148 82

No CAOD

0.04

11.0(356)

0.05

441

±

7.9 (334)

NS

418

± ± ± ± ± ± ± ± ± ± ±

8.4 (340) 16.3 (113) 48.1 (111) 93 (365) 113 (352) 1.4(191) 2.4(182) 3.1 (378) 31 (383) 10 (387) 89 (387)

0.009 0.05

424 143 150 457 442 232 223 471 455 481

10' (94)t 28(94) 13 (94) 0.4(81) 70 (86)

56 138 83 1.07 230

42.3 ±

11.7(85)

44.9

±

9.0 (84)

42.8

43.5 ± 8.0 (84) 25.9 ± 14.5 (30) 61.1 ± 44.7 (39) ± 95 (92) 225 ± 115 (90) 345 1.9(41) 2.8 ± 2.4(41) 4.6 ± 10.6 ± 3.1 (93) ± 31 (72) 134 35 ± 10 (94) 9.8 ± 9.5 (93)

46.1 31.8 60.9 218 337 2.9 4.4 10.7 144 39 8.8

±

± 12 (387) ± 18 (387) ± 10 (387) ± 0.4 (307) ± 59 (367)

' Mean ± SD. tNumber of cases summarized. : not statistically significant.

276

Total cases

481 481 481 388 453

± ± ± 1.15 ± ± 247

43.1

P


0.001