The importance of blood lipids in the risk of ischemic heart disease in older people is unclear; as a result, cholesterol-lowering drug therapy is not widely prescribed for older individuals without diagnosed cardiovascular disease. We conducted a study to determine the relationship between death from ischemic heart disease and the level of cholesterol, cholesterol fractions, and apolipoproteins
The value of cholesterol measurements for prediction of ischemic heart disease risk in older individuals is unclear; consequently, cholesterol-lowering drug therapy is not widely prescribed for older individuals who do not have previously diagnosed cardiovascular disease. Much of the uncertainty relates to misinterpretation of the observational epidemiologic studies of ischemic heart disease in relation to total cholesterol and cholesterol fractions.1-9 A positive correlation between ischemic heart disease mortality has been shown with total cholesterol levels in older individuals according to some studies.1-7 Inverse or no associations, however, have been shown in other studies.8,9
Individual epidemiologic studies of older individuals differ in the age at blood lipid measurement, inclusion of individuals with a prior diagnosis of cardiovascular disease, and follow-up period, and some studies have failed to take into account cholesterol fractions or to correct for regression dilution bias10 associated with a single measurement of blood lipids.
Total cholesterol is made up of low-density lipoprotein (LDL) cholesterol, which increases the risk of ischemic heart disease, and high-density lipoprotein (HDL) cholesterol, which decreases the risk of ischemic heart disease. The major surface proteins on LDL cholesterol, apolipoprotein B (apo B), and on HDL cholesterol, apolipoprotein A1 (apo A1), also increase and decrease, respectively, the risk of ischemic heart disease. We conducted a study among older men with and without cardiovascular disease to assess the relationship between death from ischemic heart disease and the level of total cholesterol, cholesterol fractions, and apolipoproteins, after correction for regression dilution bias.11
Subjects and methods
The study population consisted of 8,448 survivors in a sample of 19,019 men aged 40 to 69 years who participated in the Whitehall Study of London Civil Servants recruited between 1967 and 1970.12 In 1997 and 1998, all surviving participants were asked to complete a questionnaire regarding diagnoses of prior myocardial infarction, angina, or stroke (referred to as prior diagnosed cardiovascular disease), medications taken in the last month, and smoking status.13 Each of the 7,044 individuals who returned a completed questionnaire was sent a blood collection kit and was asked to visit their own physician to have a blood sample drawn and to have blood pressure, height, and weight measurements taken. From the Clinical Trial Service Unit and Epidemiological Studies Unit, University of Oxford, Oxford, England.
Nonfasting blood samples were obtained for 5,434 men, and blood lipid levels were obtained for 5,355 men. Complete questionnaire data and blood samples were obtained for 5,344 of these men. The chief comparisons in the present study involved 1,369 men (26%) with previously diagnosed cardiovascular disease or history of statin use and 3,975 men (74%) without a history of cardiovascular disease or statin use. Subjects were linked with national mortality statistics coordinated by the Office for National Statistics, from which we obtained the date and cause of mortality through September 2005.
We used Cox proportional hazards regression model to estimate smoking and age-adjusted hazard ratios (HRs) of ischemic heart disease. For each third of the lipid distributions, the HRs and 95% confidence intervals (CIs) were estimated. We also estimated the HRs associated with 2-standard deviation (SD) differences in measured levels. A difference of 2 SDs in lipid levels is roughly equal to the mean difference between the top and bottom thirds of a normal distribution and roughly equal to the changes that can be achieved with statin treatment for LDL cholesterol. All analyses were corrected for regression dilution bias.10,11
The mean SD levels of blood lipids for all men and for those with previously diagnosed cardiovascular disease (26%) and without prior cardiovascular disease (74%) are shown in Table 1. One hundred twenty-five (9.1%) of the 1,369 men in the prior cardiovascular disease group used statins (32 men without cardiovascular disease and 93 men with cardiovascular disease who were taking statins were included in the same group). The average patient age at the time of the re-survey was 76.9 (SD ± 4.9) years. During the follow-up period of approximately 7 years, 447 of the 5,344 men (8.4%) died from ischemic heart disease. Death from ischemic heart disease occurred in 232 (16.9%) of the 1,369 men with prior cardiovascular disease (27.8 per 1,000 person-years) and in 215 (5.4%) of the 3,975 men with no prior cardiovascular disease or statin use (7.7 per 1,000 person-years).
Table 2 shows the smoking and age-adjusted HRs associated with 2-SD differences in usual levels of total cholesterol in all men and in men with and without prior cardiovascular disease after correction for regression dilution bias.10,11 These correlations are shown by thirds of the lipid distributions for total cholesterol in the Figure.
Figure. Associations of ischemic heart disease mortality with usual plasma
cholesterol measurements in all men (left panel) and in subsets with (n =
232) and without (n = 215) prior cardiovascular disease (CVD; right panel).
Estimates are adjusted for age and cigarette smoking. The prior CVD group
included 32 men without prior CVD but who were taking statins. The relative
risk differences associated with 2-standard deviation (SD) differences in usual
levels of each lipid measure are shown in Table 2. CI indicates confidence
A positive correlation between death from ischemic heart disease and total cholesterol in men without prior cardiovascular disease and a nonsignificant inverse correlation in men with cardiovascular disease (HR = 1.47 vs 0.84; heterogeneity, = .01) was concealed by the lack of a correlation between ischemic heart disease mortality and total cholesterol in all men (HR = 1.05; 95% CI, 0.85-1.30). The pattern was similar for LDL cholesterol and apo B, as there was a significant positive correlation between ischemic heart disease mortality in men without prior cardiovascular disease but not in those with cardiovascular disease for LDL cholesterol (HR = 1.50 vs 0.98; = .06) and for apo B (HR = 1.68 vs 0.93; = .01). For all men, there was an inverse association between death from ischemic heart disease and HDL cholesterol (HR = 0.60; 95% CI, 0.48-0.75) and apo A1 (HR = 0.61; 95% CI, 0.48-0.78). There were also inverse associations between death from ischemic heart disease and HDL cholesterol (HR = 0.74 vs 0.69; = .71) and apo A1 (HR = 0.87 vs 0.62; = .16) for men with and without cardiovascular disease. The ratios of total/HDL cholesterol or apo B/apo A1 combined the positive effects on ischemic heart disease risk of increasing levels of total cholesterol or apo B with the inverse effects on ischemic heart disease risk of increasing levels of HDL cholesterol or apo A1. These ratios significantly predicted ischemic heart disease mortality in all men, with a HR of ischemic heart disease of 1.57 (95% CI, 1.32-1.86) for total/HDL cholesterol and of 1.54 (95% CI, 1.27-1.87) for apo B/apo A1. Among men without cardiovascular disease, these 2 lipid ratios had the greatest ability to predict death from ischemic heart disease. With increasing age, the strength of these HRs for any of these lipids did not decrease significantly. A 2-SD higher than usual level of apo B/apo A1 was associated with an HR of 1.58 (95% CI, 1.09-2.28) for men younger than 75 years of age, 1.47 (95% CI, 1.06-2.04) for men aged 75 to 79 years, and 1.57 (95% CI, 1.14-2.16) for men 80 years of age or older. These associations did not change significantly after adjustment for other cardiovascular risks (data not shown).
Results of our study showed that for all men, ischemic heart disease mortality was not associated with total cholesterol levels. The risk of death from ischemic heart disease, however, was associated with total cholesterol levels for men who had not had a previous cardiovascular disease diagnosis, and there was a nonsignificant inverse relationship for men with previously diagnosed cardiovascular disease or history of statin use.11 Among men without cardiovascular disease, the risk of death from ischemic heart disease was correlated with LDL cholesterol and apo B, whereas there was a nonsignificant inverse relationship among those with cardiovascular disease. Regardless of the presence or absence of cardiovascular disease, HDL cholesterol and apo A were inversely associated with ischemic heart disease mortality. Among men with prior cardiovascular disease, "reverse causality bias" (caused by high-risk individuals being treated, resulting in a low LDL-cholesterol level and lower risk, but not as low as those who always had a low LDL-cholesterol level) is most likely the reason there were no associations between ischemic heart disease mortality and LDL cholesterol, apo B, or total cholesterol. LDL-lowering treatment has been shown to protect against the risk of death from ischemic heart disease among older individuals.14
Despite total cholesterol level being used extensively for prediction of death from ischemic heart disease in the general population, our findings indicate that when used alone, it has the lowest predictive value. Regardless of age, better predictors of death from ischemic heart disease were the ratios of total/HDL cholesterol or apo B/apo A1.
The importance of blood lipid levels in portending the risk of death from ischemic heart disease in older individuals is not clear. Therefore, older patients are rarely prescribed cholesterol-lowering treatment. Only 2% of the subjects in our study (8% of those with prior cardiovascular disease) were receiving statin therapy when they were re-surveyed in 1997.11
Another study showed that only 19% of 77,000 older individuals with cardiovascular disease were receiving statin therapy.15 The results of our study indicate that the differences in LDL-cholesterol levels that can be reached with statin therapy14 may be associated with decreases in ischemic heart disease risk, regardless of age.
I am grateful to all the participants in the Whitehall Study of London Civil Servants. I would like to thank Jonathan Emberson, Sarah Parish, Martin Shipley, Pamela Linksted, Paul Sherliker, Sarah Clark, Jane Armitage, Astrid Fletcher, and Rory Collins for their input into this work. The study was supported by grants from the British Heart Foundation and Medical Research Council.