High-sensitivity C-reactive protein, LDL cholesterol, lipoprotein(a) and 30-year risk of stroke in healthy women: a prospective, longitudinal cohort study.

BACKGROUND: Primary stroke prevention guidelines recommend routine screening of individuals for elevated LDL cholesterol from the age of 40 years, but recommendations are ambiguous for high-sensitivity C-reactive protein (hsCRP) and lipoprotein(a). We aimed to examine correlations between hsCRP, LDL cholesterol, and lipoprotein(a) and 30-year risk of stroke in healthy women. METHODS: In this prospective, longitudinal cohort study, participants who were enrolled in the Women's Health Study (a randomised controlled trial of aspirin and vitamin E for the prevention of cardiovascular disease and cancer in women in the USA that completed in 2004) were prospectively followed for up to 30 years via annual questionnaires. Healthy women (ie, without cardiovascular disease and cancer) aged 45 years and older were eligible for the original trial. Individuals with available baseline measurements for hsCRP, LDL cholesterol, or lipoprotein(a) were included in our analyses. We constructed cumulative incidence curves and calculated hazard ratios (HRs) for total, ischaemic, and haemorrhagic stroke across biomarker quintiles and for combined elevation of all biomarkers from age-adjusted and multivariable-adjusted cause-specific Cox models. FINDINGS: Between September, 1992 and May, 1995, 39 876 women were enrolled in the Women's Health Study, of whom 28 345 consented to participate in further follow-up and provided a baseline blood sample. Baseline concentrations of hsCRP and LDL cholesterol were available for 27 939 participants, with lipoprotein(a) measurements also reported for most participants. At enrolment, median age was 53 years (IQR 49-59), median BMI was 25 kg/m2 (22-28), 3252 (12%) were current smokers, 7026 (25%) had hypertension, and 685 (2%) had a history of diabetes. 1345 stroke events accrued during a median of 27·7 years (IQR 23·1-29·0) of follow-up. Baseline hsCRP concentrations increasing from the lowest quintile (<0·7 mg/L) to the highest quintile (≥5·2 mg/L) were associated with an increasing cumulative incidence of total stroke. Only individuals in the highest LDL cholesterol (≥3·4 mmol/L) and lipoprotein(a) quintiles (≥44·1 mg/dL) had higher cumulative incidences of total stroke than those in the lowest quintiles (<2·5 mmol/L and <3·6 mg/dL, respectively). Multivariable-adjusted hazard ratios (HRs) for total and ischaemic stroke for quintile five versus quintile one were 1·32 (95% CI 1·07-1·61) and 1·56 (1·22-1·99), respectively, for hsCRP; 1·05 (0·88-1·25) and 1·17 (0·95-1·45), respectively, for LDL cholesterol; and 1·23 (1·04-1·45) and 1·27 (1·05-1·55), respectively, for lipoprotein(a). Women with three versus no biomarkers in the fifth quintile had HRs of 1·60 (1·10-2·34) for total stroke and 1·79 (1·23-2·61) for ischaemic stroke. None of the biomarkers correlated with risk of haemorrhagic stroke. INTERPRETATION: Elevated plasma concentrations of hsCRP, LDL cholesterol, and lipoprotein(a), individually and in combination, are associated with 30-year risk of ischaemic stroke. Early screening for these risk factors might facilitate improved lifestyle interventions for the primary prevention of stroke. FUNDING: The US National Heart, Lung, and Blood and Cancer Institutes and the Independent Research Fund Denmark.