Gene Score Predicts Obesity, Even in Young Children

Despite the uptick in usage of weight management medications such as the GLP-1 inhibitor drugs, obesity statistics remain grim. In 2022, one in eight people worldwide were living with obesity, a rate which has doubled since 1990.¹ Early intervention is key to reducing these figures.

Polygenic risk scores (PRS) predict the likelihood of developing a disease based on genetic information. PRS for obesity would be a valuable tool for identifying at-risk individuals at the earliest possible age. As obesity is a complex disease, and genetic factors only make up a fraction of the overall causes of the condition, an accurate PRS eluded researchers for years.

Now, a huge international consortium of researchers has developed an enhanced PRS for obesity. The score predicted increased risk of adulthood obesity in children as young as five. The study was published in Nature Medicine.²

Building Biomarkers with Genetic Information

The study team used powerful genetic analyses known as genome-wide association studies (GWAS) to develop their PRS, said Ruth Loos, an obesity researcher at the University of Copenhagen and coauthor of the study. GWAS identify genes that contribute to body weight, allowing researchers to tease out how they influence appetite or metabolism. This genetic information is then packaged up into a clinically useful marker for obesity: the PRS. "I sometimes compare it with blood pressure as a biomarker for hypertension," said Loos. Unlike blood pressure, our genes are fixed from conception, meaning PRS can be predictive long before any other symptoms manifest.

In their study, the researchers drew from more than 200 studies to build a data resource totaling over five million participants. The study was at least five times larger than any previous GWAS for obesity. The study also had a more diverse sample, which improved PRS accuracy for ancestry populations other than Europeans.

Overall, the PRS identified genetic contributions that accounted for approximately 14 percent of the overall variation in body weight across all analyzed participants. The level of variance predicted in this study was more than twice that of previous PRS efforts for obesity, suggesting previously missed genetic links being identified. Loos said that heritability studies, which look at how traits vary between people with different levels of relatedness, suggest that the total genetic contribution to body mass index (BMI) is likely to be roughly 50 percent, with the other half of the variation attributed to other factors, such as age. Even more powerful GWAS studies will be needed to identify the remaining gene contributors.

Predicting Obesity Decades in Advance

To examine how well their PRS could predict obesity, the authors applied the score to data taken from the Avon Longitudinal Study of Parents and Children­­—a long-term study of nearly 15,000 pregnancies in southwest England. In children younger than five, BMI was a poor predictor of future BMI in early adulthood. However, when combined with the PRS, BMI became a significantly stronger predictor.

At age five, BMI alone explained 22 percent of the variance in early adulthood BMI. At age one, BMI explained just eight percent of the variation. When the PRS was added, these values jumped to 35 percent and 26 percent respectively.

The scores weren't perfect, and were notably worse at explaining variance in BMI among people with African ancestry, who are consistently underrepresented in large-scale genetics studies.³ Loos said that further work is needed to improve representation. Obesity is not just a single disease, said Loos. "It's a very heterogeneous condition," she added. "Some people have obesity since childhood, whereas for other people, it developed later on in life. Some people have obesity with type 2 diabetes and cardiovascular disease, whereas other people have obesity, but without these comorbidities.” Loos says that her future work will calculate separate PRS for all these subtypes.

Obesity Scores Aren’t a Destiny

In the future, Loos believes that PRS will be part of routine health readouts. Genome data will be stored alongside blood pressure data, allowing clinicians to predict health outcomes. Maryam Shoai, a statistical geneticist at University College London who was not involved with the study, said that PRS were originally designed with the ultimate aim of being used as diagnostic tools. “If they are to be used in these contexts for common diseases like obesity, robust methodology with a representative population is a necessity,” she said.

Loos is careful to reiterate that a high PRS is not a guarantee of future poor health. "Just like your blood pressure, the score is not your destiny," she said. "It's particularly people who are genetically susceptible who will respond badly to an obesogenic environment. It's an interaction between the two and the score is just a marker on how you might respond to an unhealthy environment.”