To the Editor:
In a case-control study, Dr Mathiesen and colleagues1 compared the blood levels of high-density lipoprotein cholesterol (HDL-C) in 216 patients with carotid stenosis with those in 223 subjects who were free from the disease. They found an inverse association between HDL-C levels and the risk of carotid stenosis, a precursor of ischemic stroke.1 The authors acknowledged that the main problem with their study was its use of cross-sectional data—the HDL-C levels among the cases were measured at the time of the diagnosis of carotid stenosis. Consequently, the postdisease levels of HDL-C in cases were compared with the predisease levels in control subjects. If the study risk factor had a value that could not change, such as a blood group type, the case-control comparisons would not have been biased. However, blood cholesterol is a variable whose value can change due to medications or it can physiologically change over time. This case-control study could not establish that the HDL-C levels in cases at the time of diagnosis represented their HDL-C levels before atherosclerotic disease of carotid arteries had developed. Consequently, the time lag between measuring the baseline attributes and ascertaining the disease was not possible.
It could be argued that a relative increase or decrease in blood cholesterol among cases in this study could have been caused by the disease. This problem with cross-sectional data was addressed by an analysis of the data on cholesterol in the Framingham Heart Study. Although atherosclerotic carotid stenosis was associated with total serum cholesterol and HDL-C levels 8 years prior to the diagnosis,2,3 there was no association between carotid stenosis and total blood cholesterol at the time of the diagnosis.3 These data clearly show that the total cholesterol levels at the time of the diagnosis were not representative of the levels 8 years prior to the diagnosis. The investigators concluded that “there is a time lag between the observation of an elevated cholesterol level and its expression as an increased degree of carotid atherosclerosis.”3
The HDL-C very likely protects against carotid artery disease and ischemic stroke. However, this association should be established in more than 1 population. Existing data sets of large cohort studies, such as the Framingham Heart Study, would be an ideal setting for this investigation. Expensive experimental studies, as called for by Mathiesen et al,1 are not necessary at this stage.
DISCLAIMER: The views expressed by Dr. Sheikh do not represent the views of the Centers for Medicare & Medicaid Services or of the United States.
In our population-based study on risk factors for plaque echogenicity in carotid stenosis, we found that HDL cholesterol was independently associated with plaque echogenicity.1 Within the stenosis group, subjects with echolucent plaques had significantly lower HDL cholesterol levels than subjects with more echogenic plaques, also when corrected for age, sex, degree of stenosis, and other cardiovascular risk factors. Echolucent plaques have previously been found to be associated with higher risk of ischemic cerebrovascular events and stroke.2,3 We did not, as indicated in the letter from Dr. Sheikh, report a significant association between HDL cholesterol and the presence of carotid stenosis. Although HDL levels were lower in 216 subjects with carotid stenosis than in 223 age- and sex-matched controls, this was not statistically significant (1.46 versus 1.56 mmol/L, P=0.08).
We agree with Dr. Sheikh that inference about causality cannot be done in a cross-sectional study. However, we doubt that data from the Framingham study can be used to study the issue of concern in our article, namely the relationship between risk factors and plaque echogenicity in carotid stenosis. As far as we know, the Framingham study has not published data concerning plaque echogenicity. Apart from our own study, we are not aware of any population-based, prospective studies on plaque echogenicity that have collected data on cardiovascular risk factors several years prior to the ultrasound examination. The Tromsø study is a prospective study on cardiovascular disease, with repeated population-based health surveys. The first survey was done in 1974, and ultrasound examination of the carotid arteries was included for the first time at the fourth survey, in 1994 to 1995. In 95 of the 216 subjects with carotid stenosis, we have information on HDL cholesterol levels and other risk factors collected 8 years prior to the ultrasound examination. We have therefore examined the association between ultrasound plaque morphology and risk factors measured 8 years prior to the ultrasound examination in these subjects. We found that subjects with echolucent and predominantly echolucent plaques had significantly lower mean level of HDL cholesterol than subjects with echogenic and predominantly echogenic plaques (1.38 mmol/L versus 1.56 mmol/L, P=0.04). In multivariate analysis with 4 categories of plaque echogenicity and adjustment for age, sex, degree of stenosis, triglycerides, and systolic blood pressure, the odds ratio for being in a lower plaque echogenicity category was 0.53 (95% CI 0.33 to 0.87; P=0.01) for a 1-SD increase in HDL (SD=0.39 mmol/L). We think these results add evidence to the assumption that HDL is an important risk factor for echolucent, rupture-prone stenotic plaques.
Mathiesen EB, Bønaa KH, Joakimsen O. Low high-density lipoprotein is associated with echolucent, soft carotid artery plaques: the Tromsø Study. Stroke. 2001; 32: 1960–1965.
Mathiesen EB, Bønaa KH, Joakimsen O. Echolucent plaques are associated with high risk of ischemic cerebrovascular events in carotid stenosis: the Tromsø study. Circulation. 2001; 103: 2171–2175.
Grønholdt M-LM, Nordestgaard BG, Schroeder TV, Vorstrup S, Sillesen H. Ultrasonic echolucent carotid plaques predict future strokes. Circulation. 2001; 104: 68–73.