Higher diastolic blood pressure but not total cholesterol level increases the risk for fatal stroke, particularly in middle age
ACP J Club. 1996 May-June;124:80. doi:10.7326/ACPJC-1996-124-3-080
Prospective Studies Collaboration. Cholesterol, diastolic blood pressure, and stroke: 13 000 strokes in 450 000 people in 45 prospective cohorts. Lancet. 1995 Dec 23/30;346:1647-53.
To determine, using meta-analysis, if an association exists between total blood cholesterol level and stroke and how the previously established association between diastolic blood pressure (DBP) and stroke varies with age.
Studies were identified by searching MEDLINE, tracing references from review articles, and discussion with study investigators.
Studies were selected if they were prospective, observational studies that included participants who were not selected on the basis of disease, with total cholesterol level and DBP measurement at recruitment, and with strokes recorded during follow-up.
Data were extracted on total cholesterol level, DBP, presence of coronary heart disease (CHD), sex, age, stroke events, and person-years at risk.
45 cohort studies met the selection criteria, involving 448 415 persons and 5- to 30-year follow-up. 13 397 participants had a stroke during follow-up. The results are mainly related to stroke mortality because most of the studies did not record nonfatal strokes. After adjustment for DBP, history of heart disease, sex, age at baseline, and within certain studies, ethnicity, no significant association existed between total cholesterol level and the risk for stroke.
After adjustment, a strong relation existed between DBP and risk for stroke. An 80% increase in the risk for stroke occurred with every 10 mm Hg increase in DBP. The association between DBP and the risk for stroke varied with age at baseline. Among those who were < 45 years of age when screened, a 10-fold difference in the risk for stroke existed between the lowest and highest categories of DBP. For participants aged 45 to 64 years, a 5-fold difference existed in the risk for stroke, and in participants aged > 65 years, a 2-fold difference existed in the risk for stroke between the lowest and highest categories of DBP. In all age groups up to at least 65 years of age when screened, the association between DBP and the risk for stroke was significant, positive, and continuous. Overall, the adjusted RR for stroke for each 10 mm Hg increase in DBP was 1.84 (CI 1.80 to 1.90).
No independent association exists between total blood cholesterol level and the risk for stroke. The association between diastolic blood pressure and the risk for stroke is strong, and the gradient is much steeper for the younger age groups.
Sources of funding: Merck Sharp and Dohme and the UK Medical Research Council.
For article reprint: PSC Secretariat, Department of Clinical Geratology, University of Oxford, Radcliffe Infirmary, Oxford 0X2 6HE, England, UK. FAX 44-1865-224-815.
The conclusions of this study relate mainly to fatal stroke. A much higher proportion of fatal strokes are hemorrhagic (1), and some evidence exists for a negative association between increasing blood cholesterol level and hemorrhagic stroke (2) and for a stronger positive association between increasing blood pressure and hemorrhagic stroke than with ischemic stroke. Therefore, it is possible that the proportional excess of hemorrhagic strokes among fatal strokes may have diluted any overall positive association between total blood cholesterol level and all stroke and enhanced the positive association between blood pressure and all stroke. Also, failure to adjust for a factor such as cigarette smoking that is more prevalent in persons with low blood cholesterol levels (3) and more specific for an etiologic stroke subtype (i.e., atherothrombotic vs cardioembolic) may have diluted any association with blood cholesterol levels. Nevertheless, the findings are consistent with those of a recent overview of 11 randomized trials that showed that reductions in cholesterol levels of 6% to 23% were not associated with changes in stroke risk (RR 1.0, CI 0.8 to 1.2) (4). 2 recent trials not included in this overview would not have altered the conclusions (5, 6).
The implications of this study for clinicians are 2-fold: 1) Lowering blood cholesterol is unlikely to have any major effect on mortality from stroke, but it should still reduce the patient's risk for a coronary event, which is a major cause of death in patients with stroke; and 2) lowering blood pressure is important to reduce the risk for stroke, especially in middle-aged and elderly patients.
Graeme J. Hankey, MBBS, MD
Royal Perth HospitalPerth, Western Australia