榴莲视频

In this Expert Analysis, we review the paper "Associations of Combined Genetic and Lifestyle Risks With Incident Cardiovascular Disease and Diabetes in the U.K. Biobank Study," by Said, et al. The goal is to review the concepts, new ideas, and key questions underpinning the study and comment on the potential for gene scores to inform both individual and population-based risk modification for cardiovascular disease (CVD) and diabetes. Our fundamental point is that gene scores are likely to have limited utility in the prevention of CVD and diabetes.

Study Design

Data from ~300,000 generally healthy enrollees (~half women) in the U.K. Biobank program were included in the analysis. Their age ranged from 40 to 70 and based on survey data their smoking, physical activity, and dietary histories were evaluated. Their body mass index (BMI) and blood pressure were obtained via a physical exam and genotyping was performed. Outcomes of interest from 2006-2015 include:

"Risks of new-onset cardiovascular disease and diabetes associated with genetic risk and combined health behaviors and factors. Genetic risk was categorized as low (quintile 1), intermediate (quintiles 2-4), or high (quintile 5). Within each genetic risk group, the risks of incident events associated with ideal, intermediate, or poor combined health behaviors and factors were investigated and compared with low genetic risk and ideal lifestyle."

Foundational Concepts

A lifestyle which includes no smoking, moderate to high levels of physical activity, a healthy diet, and BMI less than 25 is associated with a vastly reduced risk of poor health and death from many causes. This is especially true for several forms of CVD and type 2 diabetes. Importantly, these lifestyle factors are measurable, meaningfully and mechanistically linked to health outcomes, potentially actionable, and can provide durable risk reduction with sustained adherence.

The story for population wide adherence to a healthy lifestyle is mixed with key evidence for broad based changes in risk factors as the incidence of CVD declined in the last four to five decades.¹ However, at least in the U.S., the incidence of type 2 diabetes is increasing steadily along with the associated morbidity and mortality.² There is also evidence over the last several years that deaths from CVD are on the uptick in some groups after a 5-decade long decline of some 75% in CVD death rates in the U.S. In both cases the obesity and inactivity epidemics are likely major contributors.³

In the midst of the success (and occasional failure) of lifestyle modification, enthusiasm for polygenic risk scores (PRS) as predictors of CVD and diabetes has recently emerged. At the completion of the Human Genome Project in the early 2000s, it was generally assumed via the "common disease/common variant hypothesis" that a limited number (~10) of gene variants with large effect sizes would explain much of the risk of common non-communicable diseases including coronary artery disease and diabetes.⁴ It was also assumed that communication of the risks associated with these elusive common variants could drive early intervention, behavior change, and drug development leading to improved individual and population health.

Unfortunately, the common disease/common variant hypothesis underperformed and the sequence of events outlined above has not occurred.⁵ Instead of a few large effect size gene variants associated with a disease, hundreds of small effects size variants were and are being identified. Some of these have clear or likely causal linkages to the biology of disease, but for many, it is unclear if they are in any way causal or merely casually associated. Additionally, the risk associated with key variants seems context and/or cohort dependent. For example, FTO gene variants have been linked to BMI in many cohorts, but less so in individuals born before the early 1940s and in groups with high levels of habitual physical activity.^6,7

The main New Idea as exemplified by Said et al. is to use some or all of these tiny effect size variants and one of several statistical approaches to generate a gene score.⁸ When this is done, typically a relationship between the number of "risky" gene variants and disease risk is seen. In other words, groups of individuals with gene scores in the top quartile or quintile of the distribution are at increased risk for the disease or condition of interest. A key point is that these risks are probabilistic and, as gene scores typically have low to moderate predictive ability, most people with a high gene score don't in fact develop the disease or condition of interest.

The Questions from the Paper are Several-fold

• First, what effects does lifestyle have on risk? In the case of coronary artery disease (CAD), an intermediate lifestyle was associated with about 1.9 times the risk of CAD compared with an ideal lifestyle. The value was closer to 3.5 for a poor lifestyle. So there was a lifestyle risk dose-response curve and the relationships between lifestyle and the risk for stroke, hypertension, and atrial fibrillation were in similar ranges. For diabetes, the effect of a poor lifestyle on risk was ~10 compared with an ideal lifestyle. In general these effects of lifestyle on risk are consistent with prior observations

• Second, what effects do gene scores have on risk? In the case of CAD, in those with ideal lifestyles, intermediate genetic risk was associated with about ~1.3 times the risk compared with low genetic risk and the value was ~1.7 for high genetic risk. So, as was the case for lifestyle, there was a dose-response curve for genetic risk. Additionally, the dose-response curves for genetic risk were in similar ranges for stroke, hypertension, atrial fibrillation, and diabetes. Of note, the effects of lifestyle on the risk of diabetes were roughly five times greater than the effects of genes

• Third, do genetic risk and lifestyle risk interact? In general the answer is no and genetic risk can be seen as simply causing a modest increase in baseline risk with fairly similar and much more powerful effects of lifestyle superimposed on that modest risk. Importantly, for every condition except atrial fibrillation, the combination of an ideal lifestyle and high genetic risk carried an overall risk similar to low genetic risk and an intermediate lifestyle, and was superior to that of high genetic risk with a poor lifestyle. Thus lifestyle can clearly overcome much of the impact of genetic risk

Gene Scores and Lifestyle: Individual and Population Health?

With the failure of large population genomics studies to find "a few" gene variants that explain "most" of the risk for lifestyle-related non-communicable diseases, genomics enthusiasts have shifted their focus towards the idea that screening individuals and entire populations early in life for genetic risk would facilitate interventions that reduce the individual and population wide burden of these diseases. There are several problems with this approach:

• First, the predictive power of gene scores is well below generally accepted thresholds for clinical utility. Most cases will occur in people at less than high genetic risk and most people at high genetic risk will not become cases.⁹

• Second, the idea that communicating genetic risk will evoke long-term behavior change and adherence to either a lifestyle or medical intervention is not supported by existing evidence.^10-12

• Third, there is a possibility that those at either low or high genetic risk will adopt a fatalistic approach to lifestyle-based risk management and consider themselves to be either genetically "immune" or "doomed."

• Fourth, simply encouraging individual members of the general public to improve their lifestyles in the absence of strong public health policies is likely to be futile. For the recent uptick in CVD and the obesity related diabetes epidemic to be reversed, broad based and multi-faceted public (health) policies similar to those adopted for tobacco control will be needed. This will require rich countries to align food and transportation policies to promote population-wide healthy lifestyles. Policies that include schools without soda machines and communities with sidewalks and useable bicycle lanes could have a greater impact on health than genetic screening.

References

1. Ford ES, et al. Explaining the decrease in U.S. deaths from coronary disease, 1980-2000. N Engl J Med 2007; 356: 2388-2398.

2. Mensah GA, et al. Decline in cardiovascular mortality: possible causes and Implications. Circ Res 2017; 120: 366-380.

3. Bhupathiraju SN, Hu FB. Epidemiology of obesity and diabetes and their cardiovascular complications. Circ Res 2016; 118: 1723-1735.

4. Regalado A. New Genetic Tools May Reveal Roots of Everyday Ills. The Wall Street Journal. Apr 14 2006.

5. Shields R. Common disease: are causative alleles common or rare? PLoS Biol 2011; e1001009.

6. Rosenquist JN, et al. Cohort of birth modifies the association between FTO genotype and BMI. Proc Natl Acad Scu USA 2015; 112: 354-359.

7. Rampersaud E, et al. Physical activity and the association of common FTO gene variants with body mass index and obesity. Arch Intern Med 2008; 168: 1791-1797.

8. Khera AV, et al. Genome-wide polygenic scores for common diseases identify individuals with risk equivalent to monogenic mutations. Nat Genet 2018; 50: 1219-1224.

9. Saracci R. Epidemiology in wonderland: big data and precision medicine. Eur J Epidemiol 2018; 33: 245-257.

10. Austin J. The effect of genetic test-based risk information on behavioral outcomes: a critical examination of failed trials and a call to action. Am J Med Genet A 2015; 167A: 2913-2915.

11. French DP, et al. Can communicating personalized disease risk promote healthy behavior change? A systematic review of systematic reviews. Ann Behav Med 2017; 51: 718-729.

12. Holladns GJ, et al. The impact of communicating genetic risks of disease on risk-reducing health behavior: systematic review with meta-analysis. BMJ 2016; 352: i1102.

Comment