How Long Will You Live?

The Scientific Statement by O'Sullivan et al. reviews the latest word on genetic predisposition to five cardiometabolic diseases: coronary artery disease; hypercholesterolemia; type 2 diabetes; atrial fibrillation; and venous thromboembolic disease. Genetic predisposition to these and other common disorders is believed to be due to polygenes, or large numbers of genes with small individual effects. Polygenic risk scores represent the combined effects of the genetic variants that add risk for a trait in a person. The authors give background information on polygenic risk scores (PRSs; also known as polygenic scores) and discuss potential uses.

The study of polygenic inheritance is >100 years old, even though polygenes were invisible for nearly a century (reviewed in Rotter and Lin 2020). Polygenes were first recognized only through complex statistical analyses, due to their small individual effects. Genome-wide association study (GWAS) analyses identified the genes (or loci) and measured their effects (called beta values). The current method for combining beta values into weighted polygenic scores was proposed by Horne et al. (2005). Other approaches also appeared (Morrison et al. 2007).

Conceivably, polygenic scores may have different uses at different stages of life, even though germline DNA does not change appreciably over the life span. For example, in infancy through young adulthood, PRSs may identify individuals with the highest genetic risks for developing a condition. Preventive measures may be considered, while the person is healthy. In maturity -- as risk factors appear -- partitioned polygenic scores that reflect particular biological pathways may give clues to disease mechanisms and preferred treatments (Udler et al. 2019). PRSs will likely be too late for disease prediction in old age, but the scores may be useful for future generations -- as in cascade testing for single gene disorders (Reid et al. 2021).

Below are edited excerpts from the recap section (following the Conclusions) on the five cardiometabolic disorders. Readers should become familiar with the entire Scientific Statement, to better understand current thinking on clinical use of polygenic scores.

Atrial fibrillation. Polygenic risk scores for AF have consistently shown incremental predictive abilities, when added to clinical risk factors. A proposed use has been to refine identification of individuals who would benefit from close surveillance for atrial fibrillation.

Coronary artery disease. Coronary artery disease is perhaps the cardiovascular condition most studied in relation to polygenic risk scores. Among middle-aged adults, coronary artery disease PRSs perform in a manner comparable to conventional risk factors and provide added prognostic information. But the clinical significance of the improvement is contentious. Evidence suggests that coronary artery disease PRSs may help guide pharmacological management (particularly for lowering LDL cholesterol).

Hypercholesterolemia. LDL cholesterol PRSs have been shown to predict LDL-C levels, including severe hypercholesterolemia. LDL-C PRSs are predictive of atherosclerotic cardiovascular events independent of LDL-C levels. Whether they should be used to help allocate novel LDL-C-lowering medicines (as with familial hypercholesterolemia variants) requires further study.

Type 2 diabetes mellitus. Early research suggested that polygenic risk scores and clinical risk factors had similar predictive abilities for type 2 diabetes. More recent evidence suggests that PRS information may add to clinical risk estimates. However, identification of individuals at high risk for type 2 diabetes currently has unclear value, because lifestyle modification and metformin for diabetes prevention did not appear to have different effects across genetic risks. Nevertheless, type 2 diabetes PRSs may help guide management, via both sulfonylurea responsiveness and intensity of glucose management.

Venous thromboembolic disease. A PRS is associated with risk of venous thromboembolic disease (VTE). Because the clinical utility of identifying inherited thrombophilia is unknown, the usefulness of VTE PRSs is also unknown. Further study is needed on use of clinical and genetic factors to best balance benefits and risks of prophylactic anticoagulation.

Regarding criteria for when to implement polygenic scores, the authors advise: Benefits overall or for subgroups -- estimated through observational datasets -- “are likely to be key driving forces.” Use of PRSs may be appropriate when the scores: (1) substantially improve accuracy of clinical risk tools (e.g., pooled cohort equations), or (2) identify individuals with risks as high as single-gene risks (as in familial hypercholesterolemia caused by LDLR variants).

Potential harms of PRSs are also mentioned. One concern is that use of existing PRSs could worsen ethnic disparities, because scores are based on mostly European ancestry data. Ongoing work includes non-Europeans, and development of new statistical methods may help.

Finally, the authors address the business of polygenic scores, with practical input for healthcare systems, commercial genetics organizations, and payors. One gets the feeling that polygenic scores are heading for “prime time” (Levin and Rader 2020).

Obviously, the Scientific Statement will not be the last word on how to use polygenic scores. Areas for research are noted throughout the document. Discoveries will happen. For example, protein associations may be combined with polygenic scores (Ferkingstad et al. 2021; Ritchie et al. 2021). We hope the new developments will help us all live longer and healthier.

Acknowledgements

Supported in part by the National Center for Advancing Translational Sciences, CTSI grant UL1TR001881, and the National Institute of Diabetes and Digestive and Kidney Disease Diabetes Research Center (DRC) grant DK063491 to the Southern California Diabetes Endocrinology Research Center. Also supported in part by the National Institutes for Diabetes and Digestive and Kidney Diseases contract R01HL151855-01 and contract R01HL146860. Infrastructure for the CHARGE Consortium is supported in part by the National Heart, Lung, and Blood Institute (NHLBI) grant R01HL105756.