OUP user menu

Chlamydia pneumoniae IgA titres and coronary heart disease: prospective study and meta-analysis

John Danesh, Peter Whincup, Mary Walker
DOI: http://dx.doi.org/10.1016/S0195-668X(02)00801-1 881 First published online: 1 May 2003

Our synthesis of the available evidence from 10 long-term prospective studies in approximately general populations, involving a total of 2283 cases of non-fatal myocardial infarction or coronary heart disease death and 7143 controls, yielded a combined odds ratio of 1.25 (1.03–1.53) for coronary heart disease in individuals with higher concentrations of Chlamydia pneumoniae IgA titres.1 This estimate was not significantly different from the odds ratio of 1.15 (0.97–1.36) observed in a previous synthesis of 15 long-term prospective studies of C. pneumoniae IgG titres and coronary heart disease (Math; Math).2 Hence, as stated in our report, the available prospective epidemiological information suggests that these markers of C. pneumoniae infection are not strongly predictive of coronary heart disease in general populations, but further data would be required to confirm or refute any weaker associations.

As with most other available prospective studies of C. pneumoniaeserological markers and coronary heart disease, data from the British Regional Heart Study (our contribution of new data to the meta-analysis) were analysed on a ‘nested’ case-control basis. In such analyses, individuals who develop disease (‘cases’) are compared with an appropriate subset of those who remain disease-free (‘controls’) in the same prospective cohort study. Such analyses are properly called ‘prospective’ studies because recruitment and collection of blood samples on which C. pneumoniaeexposure is based, as well as the ascertainment of potential confounding factors, have all occurred several years before the development of coronary heart disease outcomes. We measured C. pneumoniae IgA antibodies in more than 75% of the coronary heart disease cases known to emerge in British Regional Heart Study participants providing baseline blood samples. There was no important difference between these cases and those cases who did not have available blood samples for analysis, in terms of age at event, proportion of fatal vs. non-fataldisease, and other potentially relevant characteristics.

As stated in our report, to avoid potential biases arising from undue emphases on data from particular studies, we placed the C pneumoniae IgA findings from the British Regional Heart Study (which involved more coronary heart disease cases than all but one previous study) in the context of a meta-analysis of all other available long-term prospective studies. So, although data from the British Regional Heart Study were somewhat more positive than results from most previous prospective reports, there was no evidence of heterogeneity among the available studies (Math; Math).1 Similar considerations applied to our previous report of new data (and its accompanying meta-analysis) on C. pneumoniae IgG antibodies and coronary heart disease.2 Both instances reinforce the need to emphasise the totality of evidence rather than the results in one or another particular study. There is currently insufficient evidence to test Robert West's suggestion that serological markers of C. pneumoniae infection may be more strongly related to risk of coronary heart disease in women than in men.