Cholesterol and carotid atherosclerosis in older persons: The framingham study

Cholesterol

and Carotid Atherosclerosis

in Older Persons:

The Framingham Study Daniel H. O’Leary, MD, Keaven M. Anderson, PhD, Philip A. Wolf, Jane C. Evans, MPH, and Harold W. Poehlman, MS

MD,

We studied the relationship between extracranial carotid atherosclerosis as measured by highresolution carotid sonography and serum total cholesterol and high-density-lipoprorein cholesterol (HDL-C) levels which were determined at the time of carotid sonography and 8 years previously in I I89 members of the Framingham cohort, aged 66 to 93 years. Among parricipants, no carotid disease was found in 30%; 1 to 49% stenosis, in 62%; 50 to 74% stenosis, in 5%; 75 to 99% stenosis, in 2%; and 100% stenosis, in 1%. Total cholesterol measured 8 years prior to the carotid examination showed a strong positive associarion with the occurrence of stenosis in both men and women. There was no association between concurrently measured cholesterol kvels and stenosis for either men or women. Fur women there was a strong association between both the g-year HDL-C level and the concurrently measured HDL-C level and the degree of carotid stenosis. For men, neither concurrent nor g-year HDL-C measurements were signifcanrly related to carotid stenosis. These results suggest that there is a time kzg between the observation of an elevated cholesterol level and its expression as an increased degree of carotid atherosclerosis. Ann Epidemiol 1992;2:147-153. KEY WORDS:

Carotid sonography,

cholesterol, elderly, Framingham

study, risk.

INTRODUCTION

An elevated blood cholesterol level is known to be associated with coronary and carotid atherosclerosis. The relationship between blood cholesterol and cardiovascular heart disease has been extensively studied, but far less attention has been focused on the impact of cholesterol levels on the development of cerebral vascular disease. Only one population-based study has investigated the relationship between extracranial carotid atherosclerosis as determined by high-resolution ultrasonography and concurrently measured serum cholesterol (l), and no study related carotid disease with cholesterol levels measured at an earlier age. Since the association between cholesterol level and cardiovascular disease weakens with increasing age (2, 3), it would be expected that cholesterol levels drawn at a younger age would be more likely to predict the eventual development of extracranial carotid stenosis than would those drawn late in life. The Framingham Study, a general population study that has been ongoing for 40 years, provided us with a means to investigate this hypothesis. Duplex ultrasound imaging, which incorporates Doppler and B-mode ultrasound, has gained wide acceptance over the past decade in the diagnosis of extracranial carotid atherosclerosis (4-6). As a valuable and reproducible alternative to angiography, its use From the Department of Radiology, Harvard Medical School and Brigham and Women’s Hospxal, Boston (D.H.O.); Framingham Heart Study, National Heart, Lung, and Blood Institute, Framingham (K.M.A., J.C.E.); and the Department of Neurology, Boston University School of Medicine, Boston (P.A.W., H.W.P.), MA. Address reprint requests to: Daniel H. O’Leary, MD, Departments of Radiology, Geisimger Medical Center, Danville, PA 17822. Received November 28, 1990; revised March 29, 1991. 0 1992 Elsevier Science Publishing

Co.,

Inc.

1047s2797/92/$03.50

148

O’Leary et al. CAROTID DISEASE AND CHOLESTEROL

in the clinical

setting

AEP Vol. 2, No. 112

LEVELS AT FRAMINGHAM

emphasizes

quantification

January/March 1992: 147-153

of flow-reducing

lesions

in evaluating

patients for surgical intervention. Because it is noninvasive and carries no risk, sonography can also be used to determine the extent of atherosclerotic disease in general populations.

With

its unique

ability

to visualize

the arterial

wall as well as the lumen,

sonography offers advantages over other imaging modalities, since atherosclerosis involves the arterial wall long before it leads to narrowing of the lumen. This is a significant strength because while the prevalence of symptomatic or severe atherosclerotic narrowing in a general population is low, milder degrees of wall thickening are not uncommon The carotid

and can be detected reproducibly with this technique artery serves as an obvious target of study on a population

it is easily imaged disease.

and is an important

We used sonography

ing members

of the original

site of atherosclerosis

to examine

the extracranial

Framingham

Study

related carotid

cohort

(7-9). basis because

to cerebrovascular

arteries

and classified

of the survivthe extent

of

atherosclerosis present. We related these findings to serum total cholesterol and highdensity-lipoprotein cholesterol (HDL-C) d rawn at the time of carotid sonography and 8 years previously.

METHODS The Framingham Massachusetts,

Study of a general has been

underway

sample of the residents since

1948 (10).

of the town of Framingham,

Noninvasive

carotid

sonography

was used to assess the presence and degree of carotid stenosis in the surviving ham cohort beginning in 1988 at the start of biennial examination 20. At Framingham, resolution, real-time (PW)

Doppler

projections

Framing-

carotid sonography was performed with an Ultrasonix, highscanner with a 7.5MHz imaging transducer, a ~-MHZ pulse wave

transducer,

and a ~-MHZ

of the carotid

1 cm of the common

bifurcation

carotid

continuous

were obtained

artery,

the carotid

wave

(CW)

transducer.

on each side to include

bulb,

and the proximal

Three the distal

1 cm of the

internal carotid artery. Frozen images were captured on a Mitsubishi page printer and both the frozen image and a short segment of real-time scanning to demonstrate motion were

recorded

external exclusive bance

on videotape

for later

interpretation.

CW

Doppler

recordings

of the

carotid artery and both PW and CW recordings of the carotid bifurcation of the external carotid artery were obtained at the site of maximum distur-

of flow. Studies

Measurements

were

analyzed

of plaque

in a blinded

thickness

fashion,

were obtained

identified

only

by a study

for the near and far wall,

code.

exclusive

of the external carotid artery, at the site of maximum disease in each view using handheld calipers. Total plaque thickness was calculated by combining near- and far-wall measurements. Residual lumen was measured at the site of maximum luminal narrowing and unobstructed lumen was measured just distal to the site of any wall abnormality. Codes were utilized to identify the quality of each image and to identify reasons where no measurements could be obtained. When no plaque was identified, a determination was made as to whether far-wall thickness exceeded 1.2 mm. Peak systolic velocities and frequencies were recorded from the sites of maximum flow disturbance. An estimate of vascular stenosis was arrived at by a composite of both Doppler spectral criteria and assessment of gray-scale images. For the purposes of this article, we used the maximum percent stenosis of the two arteries. During the 1980 to 1982 examination, exam 15, and the 1988 to 1989 examination, exam 20, total serum cholesterol was measured using the Abell-Kendall method

AEP Vol. 2, No. l/2 January/March 1992: 147-1.53

TABLE

O’Leary et al. CAROTID DISEASE AND CHOLESTEROL LEVELS AT FRAMINGHAM

1

149

Tertiles of total and HDL cholesterol (HDL-C) (mmol/L) 1980-1981 Tertile

Total

1 2 3 1 2 3

Men

Women

and HDL-C

HDL-C

3.10- 5.30 5.31- 6.13 6.14- 9.80 3.18- 5.72 5.73- 6.57 6.58-10.21

and HDL-C

Statistical

Methods between categories

HDL-C

2.25-4.81 4.82-5.61 5.62-8.61 2.56-5.33 5.34-6.10 6.11-9.78

chloride

.44- .88 .89-1.11 1.12-2.48 .47-1.14 1.15-1.47 1.48-3.26

precipitation

of fresh

of the Lipid Research Clinics Program protocol.

were measured from a nonfasting

Associations measurements

Total

.41- .98 .99-1.22 1.23-2.69 .49-1.24 1.25-1.53 1.54-3.23

was measured after heparin-manganese

plasma, following a modification cholesterol

1988-1989

of stenosis and categories

Total

blood sample.

of tertiles of lipid protein

were carried out. Carotid stenosis was divided into five categories

for

the purpose of these analyses: none; 1 to 24% stenosis; 25 to 49% stenosis; 50 to 74% stenosis; and 75 to 100% stenosis. Subjects were grouped by age into decades. Tests of association

were sex-specific,

and were done using the Cochran-Mantel-Haenszel

statistic blocking by decade. Age-adjusted rates of disease were computed for both sexes simultaneously using the direct method so that rate comparisons between sexes are reasonable.

Stacked bar charts are presented to show the amount of the four higher

categories of stenosis.

RESULTS Duplex imaging studies were performed on 1189 subjects from 66 to 93 years old, to assess patterns of extracranial carotid disease. No significant disease was found in 360 subjects (30%); mild disease to (1 to 49% stenosis), in 739 (62%); moderate disease (50 to 74% stenosis), occlusion,

in 62 (5%);

severe disease (75 to 99%),

in 19 (2%);

and

in 9 (1%).

Tertiles of total and HDL-C are presented in Table 1. For both men and women, total cholesterol increased as HDL-C increased. With a single exception, total cholesterol and HDL-C obtained

levels determined

from the same participants

prior to carotid examination

during 1980 to 1982 were higher in 1988 to 1989. Cholesterol

than those

measured 8 years

showed a strong positive association with the occurrence

of stenosis in both men and women (Figure 1). For instance,

in the lowest tertile of

cholesterol, only 7.0% of the men later had 50% or more stenosis, while this figure was 16.2% for men in the upper tertile. In women, the corresponding figures were 2.6% and 9.8%. This association was not seen for concurrently measured cholesterol for either sex (Figure 2). The results for HDL-C were similar. For women, HDL-C levels measured 8 years before and concurrently were inversely related to the degree of measured stenosis (Figures 3 and 4). For women it was significant for the 8-year HDL-C value and the concurrent HDL-C measurement. For men neither the 8-year nor the concurrent HDLC measurements

were significantly

related to carotid stenosis (see Figures 3 and 4).

150

O’Leary et al. CAROTID DISEASE AND CHOLESTEROL LEVELS AT FRAMINGHAM

AEP Vol. 2, No. l/2 lanuaryllvlarch 1992: 147-153

x 8c 7c w60 s? 50 2i Y iE 40 &? 30 20 10 0 TERTILE 1 TERTILE 2

TERTILE 3

TERTILE 1 TERTILE 2

TERTILE 3

WOMEN (P-c ,001)

MEN (P = .006) DEGREE OF STENOSIS

m

75 -100%

I

25-49%

m B

50-74% l-24%

FIGURE

1

Age-adjusted

prevalence

of stenosis

by tertile

of cholesterol

level at exam 15.

FIGURE

2

Age-adjusted

prevalence

of stenosis by tertile

of cholesterol

level at exam 20.

TEFITlLE1 TERTILE 2

TERTILE 3

TERTILE 1 TERTILE 2

MEN (P = .43) DEGREE OF STENOSIS

TERTILE 3

WOMEN (P= .14) -

75-100%

r

25-49%

= B

50-74% l-24%

TERTILE1 TERllLE2

MEN

EFrlllE3

ERTlLEl

(P= .86) I

3

Age-adjusted

prevalence

ERTLE2

TERTlLE3

WOMEN (P= .003)

DEGREE OF STENOSIS w

FIGURE

151

O’Leary et al. CAROTID DISEASE AND CHOLESTEROL LEVELS AT FRAMINGHAM

AEP Vol. 2, No. 112 .Janua~lMarch 1992: 147-l 53

75-100% 25-49%

of stenosis by tertile

m m

of HDL

50 - 74% l-24% cholesterol

level

at

exam 15.

DISCUSSION indicates that there is a strong association between total cholesterol measured 8 years previously and the occurrence of carotid stenosis in elderly men and women. For concurrently measured cholesterol, no association was found in either sex. HDLC levels measured both 8 years previously and concurrently were strongly associated with carotid stenosis in women but not in men. Our findings are compatible with the relationship found by Kannel and associates between coronary heart disease incidence and measured blood lipids using Framingham data (11). That is, the association weakens in advanced age. Their data demonstrate that for men, cholesterol measures from younger ages are more strongly associated with disease incidence than are cholesterol measures obtained at varying intervals later in life. The implication of this is that the lack of an association of concurrent cholesterol with atherosclerosis is not completely a reflection of the harvesting of all cholesterol-associated coronary heart disease prior to the advanced ages studied at Framingham. This finding is in concordance with our results, and both suggest that there is a time lag between the observation of an elevated cholesterol level and its expression as an increased degree of atherosclerosis. In a similar study of Finnish men aged 42 to 60 years carried out by Salonen and associates who used carotid sonography, both low-density-lipoprotein cholesterol (LDL-C) (positive) and HDL-C ( mverse) were strongly associated with concurrently measured carotid stenosis [l]. This LDL-C result is in contrast to our negative finding for concurrent measurements of total cholesterol in men. However, the ages of the

Our study

152

O’Leary et al. CAROTID DISEASE AND CHOLESTEROL

TERTILE 1

TERTILE 1

TERTILE 2 TERTILE 3

MEN (P = .84)

Age-adjusted

412 men studied Framingham

prevalence

by Salonen

cohort

may be explained

0

75-100% 25-49%

of stenosis

TERTILE 3

and her coworkers

of

50-74% l-24%

HDL

cholesterol

level

were 42 to 60 years, while those of the

at least in part by age. Other and carotid populations

= m

by tertile

were 66 to 93. Thus the difference

total cholesterol and HDL-C rather than studies of general

TERTllE 2

WOMEN (P = .028)

DEGREE OF STENOSIS

FIGURE 4 at exam 20.

AEP Vol. 2, No. l/2 January/March 1992: 147-153

LEVELS AT FRAMINGHAM

between

studies

our findings

found

atherosclerosis, ( 12, 13).

a strong

but these

and theirs

association

are clinical

of

series

REFERENCES 1.

Salonen

R, Seppanen

sis and serum cholesterol 2.

K, Rauramaa R, Salonen JT. Prevalence

levels in eastern

Stamler J, Wentworth

and risk of premature

death

Finland,

Atherosclerosis.

D, Neaton JD, et al. Is relationship

from coronary

heart

disease

ofcarotid

atherosclero-

1988;8:788-92. between serum cholesterol

continuous

and graded?,

JAMA.

1986;256:2823-8. 3. Coronary JAMA.

Castelli

WP,

Garrison

heart disease incidence

RJ, Wilson

PWF,

and lipoprotein

Abbott

RD,

cholesterol

Kalousdian

levels:

S, Kannel

the Framingham

WB. study,

1986;256:2835-8. 4.

Jacobs NM, Grant EC, Schellinger

carotid sonography: 5. Carotid 6.

Blackshear artery velocity Ricotta

endarterectomy?, 7.

Criteria WM,

for stenosis, Phillips

patterns

DJ, Chikos

Poli A, Tremoli

and pitfalls,

JD, Cohan

Radiology.

PM, Harley JD, Thiele

in normal and stenotic

J, Holen J, Schenk J Vast Surg.

D, Byrd MC, Richardson

accuracy,

vessels,

Stroke.

E, et al. Is routine angiography

SL. Duplex

1985;154:385-91. BL, Strandness

DE.

1980;11:67-71.

necessary prior to carotid

1984;1:96-102.

E, Colombo

A, Sirtori

M, Pignoli

P, Paoletti

R. Ultrasonographic

AEP Vol. 2, No. l/2 JanuqlMarch 1992: 147-153

CAROTID

measurement

of the common

Atherosclerosis. 8.

DISEASE AND CHOLESTEROL

carotid

Pignoli P, Tremoli

9.

evaluations

Circulation. 10.

patients,

with ultrasound

from normal

imaging,

Circulation.

M, Benassi A, L’Abbate

and atherosclerotic

arterial

walls:

of the

1986;6:1399-1406. A. Time domain A study in vitro,

1988;3:654-9. Cupples

cardiovascular

LA,

D’Agostino

disease and death

Heart Study, 30-year follow-up. An Epidemiological

National

in hypercholesterolemic

E, Poli A, Oreste P, Paoletti R. lntimal plus medial thickness

Picano E, Landini L, Lattanzi F, Salvadori

pattern

Study:

wall thickness

1988;70:253-261.

arterial wall: A direct measurement echo

arterial

153

O’Leary et al. LEVELS AT FRAMINGHAM

Heart,

RB.

Some risk factors related

using pooled repeated

biennial

to the annual

In: Kannel WB, Wolf PA, Garrison

Investigation

Lung and Blood Institute;

of Cardiovascular 1987. National

of

Framingham

RJ, eds. The Framingham

Disease.

Institutes

incidence

measurements: Section

of Health

34.

Bethesda:

Publication

No.

87.2703. 11.

Kannel WB,

Framingham 12.

Study,

13.

KM, Evans JC. Relevence Risk Factors

Crouse JR, Toole JF, McKinney

atherosclerosis, nial carotid

Anderson

Int J Cardiovasc

Stroke.

of blood lipids in the elderly. The

(in press).

WM, et al. Risk factors for extracranial

carotid artery

1987;18:990-6.

Rubens J, Espeland MA, Ryu J, et al. Individual variation atherosclerosis,

Arteriosclerosis.

1988;8:389-97.

in susceptibility

to extracra-