Association of fibrinogen, C-reactive protein, albumin, or leukocyte count with coronary heart disease: meta-analyses of prospective studies.

CONTEXT A large number of epidemiologic studies have reported on associations between various "inflammatory" factors and coronary heart disease (CHD). OBJECTIVE To assess the associations of blood levels of fibrinogen, C-reactive protein (CRP), and albumin and leukocyte count with the subsequent risk of CHD. DATA SOURCES Meta-analyses of any long-term prospective studies of CHD published before 1998 on any of these 4 factors. Studies were identified by MEDLINE searches, scanning of relevant reference lists, hand searching of cardiology, epidemiology, and other relevant journals, and discussions with authors of relevant reports. STUDY SELECTION All relevant studies identified were included. DATA EXTRACTION The following information was abstracted from published reports (supplemented, in several cases, by the authors): size and type of cohort, mean age, mean duration of follow-up, assay methods, degree of adjustment for confounders, and relationship of CHD risk to the baseline assay results. DATA SYNTHESIS For fibrinogen, with 4018 CHD cases in 18 studies, comparison of individuals in the top third with those in the bottom third of the baseline measurements yielded a combined risk ratio of 1.8 (95% confidence interval [CI], 1.6-2.0) associated with a difference in long-term usual mean fibrinogen levels of 2.9 pmol/L (0.1 g/dL) between the top and bottom thirds (10.3 vs 7.4 pmol/L [0.35 vs 0.25 g/dL]). For CRP, with 1053 CHD cases in 7 studies, the combined risk ratio of 1.7 (95% CI, 1.4-2.1) was associated with a difference of 1.4 mg/L (2.4 vs 1.0 mg/L). For albumin, with 3770 CHD cases in 8 studies, the combined risk ratio of 1.5 (95% CI, 1.3-1.7) was associated with a difference of 4 g/L (38 vs 42 g/L, ie, an inverse association). For leukocyte count, with 5337 CHD cases in the 7 largest studies, the combined risk ratio of 1.4 (95% CI, 1.3-1.5) was associated with a difference of 2.8 x 10(9)/L (8.4 vs 5.6 x 10(9)/L). Each of these overall results was highly significant (P<.0001). CONCLUSIONS The published results from these prospective studies are remarkably consistent for each factor, indicating moderate but highly statistically significant associations with CHD. Hence, even though mechanisms that might account for these associations are not clear, further study of the relevance of these factors to the causation of CHD is warranted.