Prognostic effect size of cardiovascular biomarkers in datasets from observational studies versus randomised trials: meta-epidemiology study

Department of Hygiene and Epidemiology, University of Ioannina School of Medicine, Ioannina, Greece.
BMJ (online) (Impact Factor: 16.38). 11/2011; 343:d6829. DOI: 10.1136/bmj.d6829
Source: PubMed

ABSTRACT To compare the reported effect sizes of cardiovascular biomarkers in datasets from observational studies with those in datasets from randomised controlled trials.
Review of meta-analyses.
Meta-analyses of emerging cardiovascular biomarkers (not part of the Framingham risk score) that included datasets from at least one observational study and at least one randomised controlled trial were identified through Medline (last update, January 2011).
Study-specific risk ratios were extracted from all identified meta-analyses and synthesised with random effects for (a) all studies, and (b) separately for observational and for randomised controlled trial populations for comparison.
31 eligible meta-analyses were identified. For seven major biomarkers (C reactive protein, non-HDL cholesterol, lipoprotein(a), post-load glucose, fibrinogen, B-type natriuretic peptide, and troponins), the prognostic effect was significantly stronger in datasets from observational studies than in datasets from randomised controlled trials. For five of the biomarkers the effect was less than half as strong in the randomised controlled trial datasets. Across all 31 meta-analyses, on average datasets from observational studies suggested larger prognostic effects than those from randomised controlled trials; from a random effects meta-analysis, the estimated average difference in the effect size was 24% (95% CI 7% to 40%) of the overall biomarker effect.
Cardiovascular biomarkers often have less promising results in the evidence derived from randomised controlled trials than from observational studies.


Available from: John Ioannidis, Jun 14, 2015
  • [Show abstract] [Hide abstract]
    ABSTRACT: The cause of multiple sclerosis is believed to involve environmental exposure and genetic susceptibility. We aimed to summarise the environmental risk factors that have been studied in relation to onset of multiple sclerosis, assess whether there is evidence for diverse biases in this literature, and identify risk factors without evidence of biases. We searched PubMed from inception to Nov 22, 2014, to identify systematic reviews and meta-analyses of observational studies that examined associations between environmental factors and multiple sclerosis. For each meta-analysis we estimated the summary effect size by use of random-effects and fixed-effects models, the 95% CI, and the 95% prediction interval. We estimated the between-study heterogeneity expressed by I(2) (defined as large for I(2)≥50%), evidence of small-study effects (ie, large studies had significantly more conservative results than smaller studies), and evidence of excess significance bias (ie, more studies than expected with significant results). Overall, 44 unique meta-analyses including 416 primary studies of different risk factors and multiple sclerosis were examined, covering a wide range of risk factors: vaccinations, comorbid diseases, surgeries, traumatic events and accidents, exposure to environmental agents, and biochemical, infectious, and musculoskeletal biomarkers. 23 of 44 meta-analyses had results that were significant at p values less than 0·05 and 11 at p values less than 0·001 under the random-effects model. Only three of the 11 significant meta-analyses (p<0·001) included more than 1000 cases, had 95% prediction intervals excluding the null value, and were not suggestive of large heterogeneity (I(2)<50%), small-study effects (p for Egger's test >0·10), or excess significance (p>0·05). These were IgG seropositivity to Epstein-Barr virus nuclear antigen (EBNA) (random effects odds ratio [OR] 4·46, 95% CI 3·26-6·09; p for effect size=1·5 × 10(-19); I(2)=43%), infectious mononucleosis (2·17, 1·97-2·39; p=3·1 × 10(-50); I(2)=0%), and smoking (1·52, 1·39-1·66; p=1·7 × 10(-18;)I(2)=0%). Many studies on environmental factors associated with multiple sclerosis have caveats casting doubts on their validity. Data from more and better-designed studies are needed to establish robust evidence. A biomarker of Epstein-Barr virus (anti-EBNA IgG seropositivity), infectious mononucleosis, and smoking showed the strongest consistent evidence of an association. None. Copyright © 2015 Elsevier Ltd. All rights reserved.
    The Lancet Neurology 02/2015; 14(3). DOI:10.1016/S1474-4422(14)70267-4 · 21.82 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Objectives: This investigation explored the short-term prognostic value of pentraxin-3 (PTX3) levels in patients with non-ST-segment elevation myocardial infarction (NSTEMI) treated by percutaneous coronary intervention (PCI). Methods: We measured plasma levels of PTX3 and other biomarkers in 525 PCI-treated NSTEMI patients (mean age, 57.7 years; 328 males). The associations of PTX3 levels with cardiac events and cardiac deaths occurring within 30 days of discharge were evaluated with multivariable Cox proportional hazard models. Results: Renal function, diabetes prevalence, systolic blood pressure, heart rate and ejection fraction differed significantly in the high PTX3 (≥3.0 ng/ml, n = 107) and low PTX3 (<3.0 ng/ml, n = 418) groups (all p < 0.05). Plasma PTX3 levels were correlated with high-sensitivity C-reactive protein, troponin T and N-terminal pro-B-type natriuretic peptide in NSTEMI patients (all p < 0.05). Kaplan-Meier analysis showed in-hospital and 30-day cardiac events and deaths were higher in the high PTX3 group (both p < 0.01). Elevated PTX3 was an independent predictor of 30-day cardiac events (95% CI 1.09-1.68; p = 0.006) and mortality (95% CI 1.18-2.15; p = 0.002). Conclusions: An elevated plasma level of PTX3 predicts 30-day cardiac events and mortality in PCI-treated NSTEMI patients. © 2014 S. Karger AG, Basel.
    Cardiology 10/2014; 129(3):178-188. DOI:10.1159/000364996 · 2.04 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Background Elevated glycated hemoglobin (HbA1c) is associated with increased risk of cardiovascular disease (CVD) and mortality but little is known about potential mechanisms underlying the reported associations. Methods We used data from 1,798 non-diabetic participants from the population-based cohort Study of Health in Pomerania (SHIP) to investigate cross-sectional and longitudinal associations of HbA1c with subclinical atherosclerosis (common carotid artery intima-media thickness [CCA-IMT]), cardiac structure (left ventricular mass [LVM]), and cardiac function (fractional shortening). Results Cross-sectional analyses revealed a positive association between HbA1c and mean CCA-IMT with a 0.02 mm (95% confidence interval: 0.01–0.04) increase in CCA-IMT per 1% increase in HbA1c, and a similar positive trend across HbA1c quartiles (overall p-value <0.01). We also observed a graded association between HbA1c and high CCA-IMT (>75th percentile) with an odds ratio of 1.42 (95% CI: 1.11–1.81) per 1% increase in HbA1c. Longitudinal analyses showed no consistent associations of baseline HbA1c with mean follow-up CCA-IMT. There were no consistent associations of HbA1c with cardiac remodeling in cross-sectional and longitudinal analyses, respectively. Conclusions The association between HbA1c and CCA-IMT in non-diabetic adults may be a crucial link between high-normal HbA1c levels and an increased risk of CVD and mortality.
    Diabetes Research and Clinical Practice 09/2014; 105(3). DOI:10.1016/j.diabres.2014.05.004 · 2.54 Impact Factor