Risk prediction models for cardiovascular disease and overall mortality
Author: Ganna, Andrea
Date: 2015-01-16
Location: Atrium lecture hall, Nobels väg 12B, Karolinska Institutet, Campus Solna
Time: 09.00
Department: Inst för medicinsk epidemiologi och biostatistik / Dept of Medical Epidemiology and Biostatistics
View/ Open:
Thesis (1.950Mb)
Abstract
Prediction or prognostication is at the core of modern evidence-based medicine. Prediction of overall mortality and cardiovascular disease can be improved by a systematic evaluation of measurements from large-scale epidemiological studies or by using nested sampling designs to discover new markers from omics technologies.
In study I, we investigated if prediction measures such as calibration, discrimination and reclassification could be calculated within traditional sampling designs and which of these designs were the most efficient. We found that is possible to calculate prediction measures by using a proper weighting system and that a stratified casecohort design is a reasonable choice both in terms of efficiency and simplicity.
In study II, we investigated the clinical utility of several genetic scores for incident coronary heart disease. We found that genetic information could be of clinical value in improving the allocation of patients to correct risk strata and that the assessment of a genetic risk score among intermediate risk subjects could help to prevent about one coronary heart disease event every 318 people screened.
In study III, we explored the association between circulating metabolites and incident coronary heart disease. We found four new metabolites associated with coronary heart disease independently of established cardiovascular risk factors and with evidence of clinical utility. By using genetic information we determined a potential causal effect on coronary heart disease of one of these novel metabolites.
In study IV, we compared a large number of demographics, health and lifestyle measurements for association with all-cause and cause-specific mortality. By ranking measurements in terms of their predictive abilities we could provide new insights about their relative importance, as well as reveal some unexpected associations. Moreover we developed and validated a prediction score for five-year mortality with good discrimination ability and calibrated it for the entire UK population.
In conclusion, we applied a translational approach spanning from the discovery of novel biomarkers to their evaluation in terms of clinical utility. We combined this effort with methodological improvements aimed to expand prediction measures in settings that were not previously explored. We identified promising novel metabolomics markers for cardiovascular disease and supported the potential clinical utility of a genetic score in primary prevention. Our results might fuel future studies aimed to implement these findings in clinical practice.
In study I, we investigated if prediction measures such as calibration, discrimination and reclassification could be calculated within traditional sampling designs and which of these designs were the most efficient. We found that is possible to calculate prediction measures by using a proper weighting system and that a stratified casecohort design is a reasonable choice both in terms of efficiency and simplicity.
In study II, we investigated the clinical utility of several genetic scores for incident coronary heart disease. We found that genetic information could be of clinical value in improving the allocation of patients to correct risk strata and that the assessment of a genetic risk score among intermediate risk subjects could help to prevent about one coronary heart disease event every 318 people screened.
In study III, we explored the association between circulating metabolites and incident coronary heart disease. We found four new metabolites associated with coronary heart disease independently of established cardiovascular risk factors and with evidence of clinical utility. By using genetic information we determined a potential causal effect on coronary heart disease of one of these novel metabolites.
In study IV, we compared a large number of demographics, health and lifestyle measurements for association with all-cause and cause-specific mortality. By ranking measurements in terms of their predictive abilities we could provide new insights about their relative importance, as well as reveal some unexpected associations. Moreover we developed and validated a prediction score for five-year mortality with good discrimination ability and calibrated it for the entire UK population.
In conclusion, we applied a translational approach spanning from the discovery of novel biomarkers to their evaluation in terms of clinical utility. We combined this effort with methodological improvements aimed to expand prediction measures in settings that were not previously explored. We identified promising novel metabolomics markers for cardiovascular disease and supported the potential clinical utility of a genetic score in primary prevention. Our results might fuel future studies aimed to implement these findings in clinical practice.
List of papers:
I. Ganna A, Reilly M, De Faire U, Pedersen NL, Magnusson KE, Ingelsson E. Risk prediction measures for case-cohort and nested case-control designs: an application to cardiovascular disease. American Journal of Epidemiology. 2012; 175(7):715-24
Fulltext (DOI)
Pubmed
View record in Web of Science®
II. Ganna A, Magnusson PK, Pedersen NL, de Faire U, Reilly M, Arnlöv J, Sundström J, Hamsten A, Ingelsson E. Multilocus Genetic Risk Scores for Coronary Heart Disease Prediction. Arteriosclerosis, Thrombosis, and Vascular Biology. 2013; 33(9):2267-72
Fulltext (DOI)
Pubmed
View record in Web of Science®
III. Ganna A, Salihovic S, Sundström J, Broeckling CD, Hedman ÅK, Magnusson PKE, Pedersen NL, Larsson A, Siegbahn A, Zilmer M, Prenni J, Ärnlöv J, Lind L, Fall T, Ingelsson E. Large-scale metabolomic profiling identifies novel biomarkers for incident coronary heart disease. PLoS Genetics. 2014; 10(12): e1004801
Fulltext (DOI)
Pubmed
IV. Ganna A, Ingelsson E. Five-year mortality predictors: a prospective study of ~500,000 UK Biobank participants. [Manuscript]
I. Ganna A, Reilly M, De Faire U, Pedersen NL, Magnusson KE, Ingelsson E. Risk prediction measures for case-cohort and nested case-control designs: an application to cardiovascular disease. American Journal of Epidemiology. 2012; 175(7):715-24
Fulltext (DOI)
Pubmed
View record in Web of Science®
II. Ganna A, Magnusson PK, Pedersen NL, de Faire U, Reilly M, Arnlöv J, Sundström J, Hamsten A, Ingelsson E. Multilocus Genetic Risk Scores for Coronary Heart Disease Prediction. Arteriosclerosis, Thrombosis, and Vascular Biology. 2013; 33(9):2267-72
Fulltext (DOI)
Pubmed
View record in Web of Science®
III. Ganna A, Salihovic S, Sundström J, Broeckling CD, Hedman ÅK, Magnusson PKE, Pedersen NL, Larsson A, Siegbahn A, Zilmer M, Prenni J, Ärnlöv J, Lind L, Fall T, Ingelsson E. Large-scale metabolomic profiling identifies novel biomarkers for incident coronary heart disease. PLoS Genetics. 2014; 10(12): e1004801
Fulltext (DOI)
Pubmed
IV. Ganna A, Ingelsson E. Five-year mortality predictors: a prospective study of ~500,000 UK Biobank participants. [Manuscript]
Institution: Karolinska Institutet
Supervisor: Ingelsson, Erik
Issue date: 2014-12-19
Rights:
Publication year: 2014
ISBN: 978-91-7549-798-3
Statistics
Total Visits
Views | |
---|---|
Risk ...(legacy) | 1451 |
Risk ... | 409 |
Total Visits Per Month
October 2023 | November 2023 | December 2023 | January 2024 | February 2024 | March 2024 | April 2024 | |
---|---|---|---|---|---|---|---|
Risk ... | 29 | 13 | 16 | 20 | 13 | 24 | 10 |
File Visits
Views | |
---|---|
Thesis_Andrea_Ganna.pdf(legacy) | 1544 |
Thesis_Andrea_Ganna.pdf | 559 |
Final_kappa_8DEC2014.pdf(legacy) | 1 |
Top country views
Views | |
---|---|
China | 482 |
United States | 299 |
Sweden | 226 |
Germany | 102 |
United Kingdom | 96 |
Ireland | 76 |
Egypt | 28 |
South Korea | 24 |
Russia | 21 |
Canada | 19 |
Top cities views
Views | |
---|---|
Shenzhen | 280 |
Dublin | 76 |
Stockholm | 62 |
Ashburn | 53 |
Sunnyvale | 50 |
Nürnberg | 34 |
Kiez | 33 |
Alexandria | 23 |
Seoul | 21 |
Beijing | 20 |