Regression dilution methods for meta-analysis: assessing long-term variability in plasma fibrinogen among 27,247 adults in 15 prospective studies

BACKGROUND: Within-person variability in measured values of a risk factor can bias its association with disease. The extent of this regression dilution bias for plasma fibrinogen was investigated using repeat measurement data collected at varying time intervals on 27 247 adults in 15 prospective studies. METHODS: Regression dilution ratios (RDRs) were estimated from a linear regression of repeat measurements on baseline values in each study and for each time interval, and pooled allowing for within- and between-study heterogeneity. RDRs were estimated both without and with adjustment for confounders, and factors were investigated that might influence the RDRs. RESULTS: The unadjusted overall RDR was 0.51 (95% CI: 0.47, 0.55), which decreased to 0.46 (95% CI: 0.42, 0.49) after adjustment for age, sex and measured values of other established vascular risk factors. The RDR did not vary materially by assay method, age, sex or smoking status, but decreased at higher levels of baseline fibrinogen. CONCLUSION: It is appropriate to use an RDR of 0.5 to correct approximately for regression dilution bias in plasma fibrinogen values; however, this correction factor may produce somewhat conservative hazard ratios in adjusted analyses, at higher fibrinogen concentrations and in follow-up beyond a decade. More generally, the methods described in this report have widespread applicability to quantifying regression dilution bias in repeatability data from multiple prospective studies.     

Oral glycoprotein IIb/IIIa antagonism in patients with coronary artery disease

Dose-finding studies and trials of interaction of oral glycoprotein IIb/IIIa antagonists with other antiplatelet agents have been limited. We hypothesized that these detailed assessments could be first performed in patients with stable coronary artery disease (CAD) and then extrapolated to the target population. To this end, we performed 2 sequential studies. The first study examined the dose-related effects on indexes of platelet and vascular function induced by the oral inhibitor RPR 109891, when given alone and in combination with aspirin, in patients (n = 100) with stable CAD. The second study (the Antagonism of the FIbrinogen Receptor after Myocardial Events trial) assessed the pharmacodynamics and safety of derived regimens in patients (n = 320) with unstable coronary syndromes. In patients with stable CAD, platelet aggregation was dose dependently inhibited by RPR 109891, and the dose-response relation was shifted to the right by the concomitant administration of aspirin (p = 0.0001). The degree of platelet inhibition induced by 3 doses of RPR 109891 (plus aspirin) was lower in patients with unstable than stable CAD. No drug-related major bleeding occurred in either study. RPR 109891 treatment was associated with acute and delayed thrombocytopenia. In conclusion, chronic treatment with an oral glycoprotein IIb/IIIa antagonist (1) induces antiplatelet effects that are potentiated by concomitant administration of aspirin, (2) may require dose adjustment in syndromes of platelet activation, (3) is associated with a low rate of clinically significant bleeding when doses inducing incomplete inhibition of platelet aggregation are used, and (4) requires frequent monitoring of platelet count unless reliable predictors of delayed thrombocytopenia become available.     

Measures to assess the prognostic ability of the stratified Cox proportional hazards model

Many measures have been proposed to summarize the prognostic ability of the Cox proportional hazards (CPH) survival model, although none is universally accepted for general use. By contrast, little work has been done to summarize the prognostic ability of the stratified CPH model; such measures would be useful in analyses of individual participant data from multiple studies, data from multi-centre studies, and in single study analysis where stratification is used to avoid making assumptions of proportional hazards. We have chosen three measures developed for the unstratified CPH model (Schemper and Henderson's V , Harrell's C-index and Royston and Sauerbrei's D), adapted them for use with the stratified CPH model and demonstrated how their values can be represented over time. Although each of these measures is promising in principle, we found the measure of explained variation V very difficult to apply when data are combined from several studies with differing durations of participant follow-up. The two other measures considered, D and the C-index, were more applicable under such circumstances. We illustrate the methods using individual participant data from several prospective epidemiological studies of chronic disease outcomes.     

Correcting for multivariate measurement error by regression calibration in meta‐analyses of epidemiological studies

Within‐person variability in measured values of multiple risk factors can bias their associations with disease. The multivariate regression calibration (RC) approach can correct for such measurement error and has been applied to studies in which true values or independent repeat measurements of the risk factors are observed on a subsample. We extend the multivariate RC techniques to a meta‐analysis framework where multiple studies provide independent repeat measurements and information on disease outcome. We consider the cases where some or all studies have repeat measurements, and compare study‐specific, averaged and empirical Bayes estimates of RC parameters. Additionally, we allow for binary covariates (e.g. smoking status) and for uncertainty and time trends in the measurement error corrections. Our methods are illustrated using a subset of individual participant data from prospective long‐term studies in the Fibrinogen Studies Collaboration to assess the relationship between usual levels of plasma fibrinogen and the risk of coronary heart disease, allowing for measurement error in plasma fibrinogen and several confounders. Copyright © 2009 John Wiley & Sons, Ltd.     

Associations of plasma fibrinogen levels with established cardiovascular disease risk factors, inflammatory markers, and other characteristics: individual participant meta-analysis of 154,211 adults in 31 prospective studies: the fibrinogen studies collaboration

Long-term increases in plasma fibrinogen levels of 1 g/liter are associated with an approximate doubling of risk of major cardiovascular disease outcomes, but causality remains uncertain. To quantify cross-sectional associations of fibrinogen levels with established risk factors and other characteristics, the investigators combined individual data on 154,211 apparently healthy adults from 31 prospective studies conducted between 1967 and 2003, using a linear mixed model that included random effects at the cohort level. Fibrinogen levels increased with age and showed continuous, approximately linear relations with several risk markers and slightly curvilinear associations with log triglycerides, albumin, and tobacco and alcohol consumption. Female sex, Black ethnicity, lower socioeconomic status, and alcohol abstinence were each associated with modestly higher fibrinogen levels. Approximately one third of the variation in fibrinogen levels was explained by cohort, age, and sex. An additional 7% was explained by established risk factors (notably, positive associations with smoking and body mass index and an inverse association with high density lipoprotein cholesterol), and a further 10% was explained by inflammatory markers (notably, a positive association with C-reactive protein). The association with body mass index was twice as strong in women as in men, whereas the association with smoking was much stronger in men. These findings substantially advance understanding of the correlates and possible determinants of fibrinogen levels.     

Systematically missing confounders in individual participant data meta‐analysis of observational cohort studies

One difficulty in performing meta‐analyses of observational cohort studies is that the availability of confounders may vary between cohorts, so that some cohorts provide fully adjusted analyses while others only provide partially adjusted analyses. Commonly, analyses of the association between an exposure and disease either are restricted to cohorts with full confounder information, or use all cohorts but do not fully adjust for confounding. We propose using a bivariate random‐effects meta‐analysis model to use information from all available cohorts while still adjusting for all the potential confounders. Our method uses both the fully adjusted and the partially adjusted estimated effects in the cohorts with full confounder information, together with an estimate of their within‐cohort correlation. The method is applied to estimate the association between fibrinogen level and coronary heart disease incidence using data from 154 012 participants in 31 cohorts.^? One hundred and ninety‐nine participants from the original 154 211 withdrew their consent and have been removed from this analysis. Copyright © 2009 John Wiley & Sons, Ltd.