Serum Biomarkers in the MORGAM Populations

Lead Research Organisation: Queen's University of Belfast
Department Name: Medicine Dentistry and Biomedical Sci

Abstract

The role of risk factors, such as raised blood cholesterol and blood pressure to heart disease, to determining future heart disease and stroke has been known for more than thirty years. More recently the level of a protein (C reactive protein)in the blood has been shown to improve prediction but better prediction is needed to identify patients at increased risk, on whom preventive efforts can be concentrated. To do this it is necessary to follow up large groups of people to see if many novel new proteins (biomarkers) can improve our ability to identify persons at higher risk of developing heart disease or stroke. We need to study human populations as the reevance of studing these new factors in animals is dubious. MORGAM is a large European studt which has included 120,000 men and women in 9 countries. A sample of serum is available for all these persons and we will use the latest tecchnology to measure a battery of 19 Biomarkers, some of which are highly novel. The results obtained in the United Kingdom samples will be verified in other European poulations, in men and women separately. In this fashion we plan to derive a scoring system to quantify the cahace of developing a heart attack in the future so that steps can be taken to avert them. The Team is based at the Queen s University of Belfast and in Helsinki, Finland, and in Mainz, Germany.

Technical Summary

Background: Cardiovascular disease is the leading cause of morbidity and mortality in the industrialised world. Biomarkers can substantially improve the early identification of high-risk patients for future cardiovascular events.
Aim of the Study: The primary aim is to produce a British/European risk score that allows improved risk prediction beyond that obtained from current scoring systems. The objective of the present proposal is to identify a single biomarker or a group of biomarkers which, in the primary prevention setting, are highly predictive of future cardiovascular disease.
Methods:
MORGAM (MOnica, Risk, Genetics, Archiving, and Monography, www.ktl.fi/morgam/) is a multinational study (53 cohorts in 9 countries) of the epidemiology and genetics of cardiovascular disease morbidity and mortality. These harmonized data are completely available at the MORGAM data centre. MORGAM is sufficiently large to permit the development of a biomarker instrument in the UK cohorts like Belfast (PRIME), Scotland (SHHS), Wales (Caerphilly), and London (UKWHS) which will then be compared to European populations with an overall sample size of 120,000 subjects. A panel of 21 easily detectable biomarkers covering distinct pathophysiological pathways has been selected. For the inflammatory pathways hs-CRP, PLA2 (PAF-AH), IL-18, sTNFR-1, hs-IL-10, and sICAM-1. To assess oxidative stress, MPO and PON will be determined. Hemodynamics and vascular remodelling are represented by (Nt-pro)BNP, (pro)ANP and TIMP-1, MMPs, and metabolic pathways include TSH, adiponectin, APO A-I and B 100. In additiion, renal function parameter will be assessed including cystatin C and neutrophil gelatinase associated lipocalin. The risk estimates will be derived primarily using time-to-event models. The predictive performance of the biomarkers will be evaluated in the UK training data-set and assessed in the overall European cohort.
Overall, the study concept provides unique opportunity to identify single biomarkers or a discrete panel of biomarkers of clinical relevance for early cardiovascular risk prediction in primary prevention.

Publications

10 25 50

publication icon
Hughes MF (2012) A multiple biomarker risk score for guiding clinical decisions using a decision curve approach. in European journal of preventive cardiology

publication icon
Hägg S (2015) Adiposity as a cause of cardiovascular disease: a Mendelian randomization study. in International journal of epidemiology

 
Title ValidStats an R package of tools 
Description ValidStats is an R package of statistical tools to facilitate model prediction and validation for analysing biomarkers. Validstats is freely available online through the R interface and contains various prediction model validation statistics (ROC, AUC, c-index, reclassification/NRI, decision curve, calibration) all adapted for censored survival data and allowing for weights. This package was written by Olli Saarela, primary MORGAM biomarker statistician based at the National Institute of Health and Welfare, Helsinki, Finland. 
Type Of Material Data analysis technique 
Year Produced 2010 
Provided To Others? Yes  
Impact These techniques were applied in Blankenberg et al. 2010 and are currently being used in other manuscripts using genetic and biomarker data being prepared by our group. 
 
Description MORGAM project 
Organisation National Institute for Health and Welfare
Department Department of Chronic Disease Prevention
Country Finland 
Sector Academic/University 
PI Contribution The Principal Investigators have made a significant intellectual contribution to the project and faciliated access to the datasets.
Collaborator Contribution The MORGAM Data Management Centre (MDC) is based at the National Institute of Health and Welfare. The MDC harmonizes the data from the various MORGAM participating centres and facilitates access to the database. The statistical analysis of the project is carried out here by MORGAM statisticians, datamanagers from this unit prepare datasets for analysis by other MORGAM collaborating Centres. Biomarker identification, selection and measurement is carried out at the Medical Clinic II laboratories at the Department of Cardiology in Mainz which moved to University Heart Centre, Hamburg, Germany in 2011.
Impact Publications in the development phase are entitled; 'Sensitive Troponin as long term predictor for cardiovascular and cardiac disease risk' along with a systematic review 'Is cardiac Tropinin a useful prognostic marker in healthy populations and those with stable coronary heart disease?' 'Vitamin D and risk of Cardiovascular disease in MORGAM' Based on this research we have been invited to publish a review paper entitled 'Clinical Utility of Multiple Biomarker Panels for Cardiovascular Disease Risk Prediction' accepted in Current Cardiovascular Risk Reports (Hughes, Kee and Salomaa).
Start Year 2006
 
Description MORGAM project 
Organisation University Heart Centre Hamburg
Department Department of Cardiology, Electrophysiology
Country Germany 
Sector Hospitals 
PI Contribution The Principal Investigators have made a significant intellectual contribution to the project and faciliated access to the datasets.
Collaborator Contribution The MORGAM Data Management Centre (MDC) is based at the National Institute of Health and Welfare. The MDC harmonizes the data from the various MORGAM participating centres and facilitates access to the database. The statistical analysis of the project is carried out here by MORGAM statisticians, datamanagers from this unit prepare datasets for analysis by other MORGAM collaborating Centres. Biomarker identification, selection and measurement is carried out at the Medical Clinic II laboratories at the Department of Cardiology in Mainz which moved to University Heart Centre, Hamburg, Germany in 2011.
Impact Publications in the development phase are entitled; 'Sensitive Troponin as long term predictor for cardiovascular and cardiac disease risk' along with a systematic review 'Is cardiac Tropinin a useful prognostic marker in healthy populations and those with stable coronary heart disease?' 'Vitamin D and risk of Cardiovascular disease in MORGAM' Based on this research we have been invited to publish a review paper entitled 'Clinical Utility of Multiple Biomarker Panels for Cardiovascular Disease Risk Prediction' accepted in Current Cardiovascular Risk Reports (Hughes, Kee and Salomaa).
Start Year 2006
 
Title Multiple biomarker risk score 
Description We evaluated the usefulness of 30 novel biomarkers to improve risk estimation for incident fatal and non-fatal cardiovascular disease events in the prospective community based samples from FINRISK and Belfast PRIME within the MORGAM project. The 30 novel biomarkers were chosen to represent distinct pathophysiological pathways, for example lipid metabolism, inflammation and myocardial necrosis. These were screened in the population based FINRISK97 cohort (7915 participants, 538 cases of incident CVD at 10 years follow up) and the Belfast PRIME cohort (2551 participants, 260 cases at 10 years of follow up). Of the 30 biomarkers the strongest associations with incident CVD events among FINRISK97 men and women and PRIME men were found between N-terminal pro-B-type natriuretic peptide (hazard ratio per increment standard deviation in FINRISK97 men cohort 1.23), C-reactive protein (1.23), B-type natriuretic peptide (1.19) and sensitive troponin I (1.18), but no single biomarker consistently improved risk estimation in these three cohorts. Statistical regression and lasso (penalized regression shrinkage) methods selected these three biomarkers; C-reactive protein, B-type natriuretic peptide and sensitive troponin I in combination with conventional risk factors for CVD including blood pressure, cholesterol level, body mass index and history of diabetes to establish a biomarker risk score in FINRISK97 cohort. Testing the biomarker score in the Belfast PRIME validation cohort consistently led to an improved risk estimation of cardiovascular disease. We are investigating potential sources of funding to move into the next stage of product development i.e. refinement non-clinical and clinical. We are in the process of testing the assay parameters of the multiple biomarker score and assessing modality in additional prospective populations. 
Type Diagnostic Tool - Non-Imaging
Current Stage Of Development Initial development
Year Development Stage Completed 2010
Development Status Under active development/distribution
Impact We have demonstrated that the biomarker risk score improves risk prediction of fatal and non-fatal cardiovascular events in two European cohorts. The biomarker risk score improves model discriminative accuracy (c-statistic P = 0.0022 in Belfast-PRIME) and also improves the Integrated Discrimination Index (P < 0.0001 in Belfast PRIME). The model improves classification of individuals across clinically defined strata of low (0-5%), intermediate (5-20%) and high (>20%) risk categories (P = 0.0007 in Belfast PRIME) compared to a traditional risk factor model. Indeed, 10 in every 100 patients would benefit from modified management due to this reclassification. The biomarker score has been internally and externally validated in two different European cohorts with high absolute risk of cardiovascular disease highlighting the potential generalizability of the score across several European regions.