Elsevier

The Lancet

Volume 370, Issue 9602, 1–7 December 2007, Pages 1829-1839
The Lancet

Articles
Blood cholesterol and vascular mortality by age, sex, and blood pressure: a meta-analysis of individual data from 61 prospective studies with 55 000 vascular deaths

https://doi.org/10.1016/S0140-6736(07)61778-4Get rights and content

Summary

Background

Age, sex, and blood pressure could modify the associations of total cholesterol (and its main two fractions, HDL and LDL cholesterol) with vascular mortality. This meta-analysis combined prospective studies of vascular mortality that recorded both blood pressure and total cholesterol at baseline, to determine the joint relevance of these two risk factors.

Methods

Information was obtained from 61 prospective observational studies, mostly in western Europe or North America, consisting of almost 900 000 adults without previous disease and with baseline measurements of total cholesterol and blood pressure. During nearly 12 million person years at risk between the ages of 40 and 89 years, there were more than 55 000 vascular deaths (34 000 ischaemic heart disease [IHD], 12 000 stroke, 10 000 other). Information about HDL cholesterol was available for 150 000 participants, among whom there were 5000 vascular deaths (3000 IHD, 1000 stroke, 1000 other). Reported associations are with usual cholesterol levels (ie, corrected for the regression dilution bias).

Findings

1 mmol/L lower total cholesterol was associated with about a half (hazard ratio 0·44 [95% CI 0·42–0·48]), a third (0·66 [0·65–0·68]), and a sixth (0·83 [0·81–0·85]) lower IHD mortality in both sexes at ages 40–49, 50–69, and 70–89 years, respectively, throughout the main range of cholesterol in most developed countries, with no apparent threshold. The proportional risk reduction decreased with increasing blood pressure, since the absolute effects of cholesterol and blood pressure were approximately additive. Of various simple indices involving HDL cholesterol, the ratio total/HDL cholesterol was the strongest predictor of IHD mortality (40% more informative than non-HDL cholesterol and more than twice as informative as total cholesterol). Total cholesterol was weakly positively related to ischaemic and total stroke mortality in early middle age (40–59 years), but this finding could be largely or wholly accounted for by the association of cholesterol with blood pressure. Moreover, a positive relation was seen only in middle age and only in those with below-average blood pressure; at older ages (70–89 years) and, particularly, for those with systolic blood pressure over about 145 mm Hg, total cholesterol was negatively related to haemorrhagic and total stroke mortality. The results for other vascular mortality were intermediate between those for IHD and stroke.

Interpretation

Total cholesterol was positively associated with IHD mortality in both middle and old age and at all blood pressure levels. The absence of an independent positive association of cholesterol with stroke mortality, especially at older ages or higher blood pressures, is unexplained, and invites further research. Nevertheless, there is conclusive evidence from randomised trials that statins substantially reduce not only coronary event rates but also total stroke rates in patients with a wide range of ages and blood pressures.

Introduction

The effects of other vascular risk factors—–particularly blood pressure—on the epidemiological associations of cholesterol with ischaemic heart disease (IHD) and stroke remain uncertain. Although blood levels of total cholesterol are used widely to predict IHD, the relative risk per unit change in cholesterol decreases with age1, 2 and, perhaps, blood pressure,3, 4 and it is unclear whether an importantly positive association persists into old age. Furthermore, total cholesterol consists largely of the cholesterol in low-density lipoprotein particles (LDL cholesterol) plus the cholesterol in high-density lipoprotein particles (HDL cholesterol), which have opposite associations with IHD risk. Results from randomised trials have shown that treatment with a statin, which lowers LDL cholesterol, substantially reduces the incidence of IHD.5 These trials have also shown a substantial reduction in the incidence of ischaemic stroke (without any apparent increase in haemorrhagic stroke).5 The definite reduction in total stroke in the statin trials contrasts strongly with the weakness of the epidemiological association between blood cholesterol and stroke,1, 6, 7, 8, 9, 10, 11, 12, 13 and that epidemiological association needs further exploration.

The results from retrospective epidemiological studies of IHD or stroke can be distorted by reverse causality (since vascular disease can itself directly or indirectly affect both blood cholesterol and blood pressure). In people with no previous history of vascular disease, however, prospective epidemiological studies have to be very large to assess reliably the extent to which one risk factor affects the relevance of another. The Prospective Studies Collaboration (PSC) has brought together evidence from many individual prospective studies of vascular mortality that recorded both blood pressure and total cholesterol at baseline, to undertake collaborative meta-analyses of the joint relevance of these two risk factors.

The present collaboration differs from previous meta-analyses in several ways that increase its reliability and precision: it is large, involving 55 262 vascular deaths in 892 337 apparently healthy adults in 61 cohorts (and, additionally, provides parallel analyses of the Multiple Risk Factor Intervention Trial [MRFIT] observational study that involve a further 34 242 vascular deaths); HDL cholesterol measurements at baseline are available for 153 798 of these participants, in whom there were 4966 vascular deaths (but, HDL cholesterol was not measured at baseline in MRFIT); and individual records are available for every participant in every study (except MRFIT), allowing detailed analyses of cause-specific mortality with respect to age, sex, blood pressure, and some other factors. Moreover, repeat measurements of HDL cholesterol in 40 313 participants allow quantitative correction for the regression dilution bias.14 Results for blood pressure have already been published,15 and the present report characterises, with greater precision and better control of some biases than has previously been possible, the age-specific relevance of total and HDL cholesterol to vascular mortality, and the extent to which this relation is modified by sex, blood pressure, and other risk factors.

Section snippets

Study design

Details of study selection, data collection, and statistical methods have all been described previously,15, 16 and are available in full in the webappendix (which includes webtables 1–6 and webfigures 1–11).

Cause-specific mortality was sought in the greatest detail available, using a three-digit International Classification of Diseases coding (ICD-6 to ICD-10), with vascular causes categorised as before15 (webtable 1). In most studies the cause of death was initially obtained from the death

Results

Individual records for all 892 337 eligible participants (without previous vascular disease recorded) in 61 studies were included in this meta-analysis: 70% from Europe, 20% from the USA or Australia, and 10% from Japan or China (webtable 5). During 11·6 million person-years at risk between the ages of 40 and 89 years (mean follow-up 13 [SD 6] years; mean time to death in those who died was 12 [7] years), there were 33 744 deaths attributed to IHD, 11 663 to stroke, and 9855 to other vascular

Discussion

This collaborative meta-analysis of almost 900 000 individuals in 61 prospective observational studies, with 55 000 vascular deaths during nearly 12 million person-years of follow-up, has characterised reliably the age-specific associations of total cholesterol with IHD, stroke, and other vascular mortality, and has assessed the quantitative and qualitative relevance of other risk factors to these associations. For IHD mortality, age and blood pressure substantially affected the strength of the

References (25)

  • H Iso et al.

    Serum cholesterol levels and six-year mortality from stroke in 350,977 men screened for the multiple risk factor intervention trial

    N Engl J Med

    (1989)
  • MR Law et al.

    Assessing possible hazards of reducing serum cholesterol

    BMJ

    (1994)
  • Cited by (0)

    Collaborators listed in full at end of paper

    View full text