Meta-analysis of sero-epidemiologic studies of the relation between Chlamydia pneumoniae and atherosclerosis: Does study design influence results?

Background The results of sero-epidemiologic studies of the relation between Chlamydia pneumoniae and atherogenesis vary considerably. A meta-analysis was performed to calculate the overall odds ratio (OR) and to identify potential sources of variability. Methods The MEDLINE database was searched from January 1997 to December 2000 for studies describing the seroprevalence of IgG antibodies to C pneumoniae in relation to clinically manifest atherosclerosis. Data extracted included 1) the study design used, 2) the number of included patients, 3) the type of disease (or end point), 4) the number of control subjects, 5) the type of control subjects, 6) the follow-up duration, 7) the crude and adjusted ORs (and 95% CI), and 8) the extent of adjustment for potential confounders. Results Thirty-eight studies were included in the meta-analysis. The overall OR was 1.6 (95% CI 1.3-2.0). The overall OR of cross-sectional case-control studies (2.0, 95% CI 1.5-2.6) was higher than that of prospective (nested) case-control studies (1.1, 95% CI 0.8-1.4). Moreover, the duration of follow-up in prospective studies was inversely related to the strength of the relation (β = −0.04, 95% CI −0.08, −0.01). Adjustment of risk estimates for potentially confounding factors did not influence the strength of the relation. Conclusions The relation between C pneumoniae and clinically manifest atherosclerosis is not durable. This probably is due to a lack of causality, although results may have been influenced by the design of the component studies. (Am Heart J 2003;145:409-17.)