Polygenic Childhood Obesity: Integrating Genetics and Environment for Early Intervention Open Access

Childhood obesity is a pressing global health issue, with prevalence rates reaching 8% among boys and 6% among girls aged 5–19 years globally [1, 2]. This epidemic carries far-reaching implications for morbidity and mortality as obesity has negative health consequences as a disease in itself [3] and is associated with a higher risk of developing comorbid conditions such as type 2 diabetes, cardiovascular diseases, and cancer [4‒7]. Moreover, childhood obesity is a strong predictor of adult obesity. According to a systematic review and meta-analysis, children and adolescents with obesity are about five times more likely to remain obese in adulthood compared to their non-obese peers; approximately 55% of children with obesity still have obesity in adolescence, and around 80% of adolescents with obesity continue to be obese as adults [8].

Genetic influences play a crucial role in determining obesity susceptibility, with genome-wide association studies (GWAS) identifying numerous significant loci, underscoring the highly polygenic nature of obesity [9]. To utilize the high volume of loci, polygenic risk scores (PRS) have emerged to distill these genetic signals into a unified measure of obesity risk.

However, genetic risk does not lead unavoidably to obesity [10]. Obesity often arises from a complex interaction between genetic predispositions and environmental factors (Fig. 1). For example, children carrying specific genetic variants demonstrate an increased sensitivity to environmental stimuli, such as high-fat diets, and reduced efficacy of traditional weight-loss strategies [11]. These environmental factors can either attenuate or amplify the impact of genetic predispositions, presenting a unique window for early prevention or treatment during childhood [12]. This also shows that ‘DNA is not destiny’, and a high polygenic risk does not make anyone predestined to develop obesity. For instance, a child with a high PRS, especially one with variants influencing satiety signals, might be recommended for a program that combines personalized dietary advice, behavioral strategies to manage cravings, and increased physical activity (PA), all started before significant weight gain occurs, to maximize the chances of prevention.

This review focuses on polygenic obesity in children, where the accumulation of numerous small genetic effects contributes to obesity risk. PRS provides a framework for quantifying this risk and can aid in identifying children who might benefit most from personalized interventions.

In the last decades, in GWAS numerous genes were found to be linked to childhood obesity [13]. The genetic architecture of childhood obesity differs from adult obesity, reflecting unique developmental, hormonal, and environmental interactions. Genetic variants often exert stronger effects in children due to the relative lack of environmental confounders compared to adults, whose obesity risk is influenced by longer-term lifestyle and environmental factors [14]. Moreover, childhood GWAS identifies developmental-stage-specific loci that are not significant in adults, reflecting the interplay between growth, metabolism, and genetic predispositions during early life.

Among the best studied genes are MC4R (melanocortin 4 receptor) [15], FTO (fat mass and obesity-associated gene) [16, 17], and other genes involved in the hypothalamic leptin-melanocortin pathway, such as LEP and LEPR, which regulate satiety and energy balance [18, 19]. GWAS variants, like FTO variants, are common single nucleotide polymorphisms (SNPs) often located in non-coding regulatory regions with small effects on obesity risk, while rare pathogenic monogenic MC4R variants affect protein function, with large effects directly contributing to monogenic obesity [20].

Among the possible applications of GWAS findings is the inference of causal links of obesity with other phenotypes. Correlation does not imply causation; the identification of GWAS genes represents associations rather than direct causal relationships. Insight in possible associations between obesity and environmental or disease phenotypes is important for defining which causal connections can be targeted by preventative strategies. Causality is established with post-GWAS analytical methods such as fine-mapping and Mendelian randomization (MR). Fine-mapping aims to identify the specific causal variants within associated loci, narrowing down thousands of candidate variants to a few that have functional significance [21]. This incorporates epigenomic data such as chromatin accessibility, transcription factor binding, and histone modification studies to pinpoint variants that affect gene regulation or expression. For example, a SNP in MAP2K5 likely alters adiposity through RNA-processing pathways [22].

MR uses genetic variants as instrumental variables (IVs) to infer causal relationships between risk factors and outcomes [23]. However, MR primarily identifies causal effects, which may include both direct effects and effects mediated through other pathways. There are other analyses available for this, like multivariable MR [24].

These approaches, combined with functional genomics and experimental validation, provide a robust framework for understanding how genetic variants contribute to obesity. Once a causal variant is identified, subsequent functional studies and clinical trials are needed to develop and test interventions.

Another important use of GWAS findings is the ability to identify individuals most at risk (i.e., those who carry most risk loci identified by GWAS). PRS estimates an individual’s genetic predisposition to a specific trait or disease. Unlike monogenic diseases caused by a single pathogenic gene variant, complex diseases like obesity are influenced by an array of genetic variants acting together with the environment (Fig. 1) [20].

PRS calculation involves summing the number of risk alleles for each SNP, weighted by their effect sizes (i.e., the SNP effect in the GWAS; Fig. 2, step 1), which represent the strength of the association with the trait (for obesity: with BMI), as depicted in Figure 2, step 2 and 3. The resulting score is a relative risk assessment, comparing an individual’s risk to that of the sample population, as shown in Figure 2, step 4. This cumulative approach captures the collective effect of many genes that individually may have minimal impact on obesity risk [20].

PRS can be used for risk stratification, identifying individuals at different levels of genetic predisposition to a condition. This enables personalized approaches tailored to an individual’s genetic risk. Additionally, PRS can be a useful supplement to existing diagnostic tools, providing a broader genetic context for clinical decision-making [25].

Although PRS provide a relative quantitative measure of polygenic risk, understanding the clinical relevance and phenotypic presentation of this risk requires insight into modifying factors. Sex and age influence the effects of genetic variants on obesity. Research shows that many obesity-associated loci show sexually dimorphic effects [26]. GWAS of waist-to-hip ratio adjusted for BMI (WHRadjBMI) show that over one-third of the loci have stronger associations in women, often linked to visceral and subcutaneous fat distribution patterns [26]. Hormonal differences such as the role of estrogen in fat storage and testosterone in fat mobilization contribute to these sex-specific effects and their metabolic consequences, such as differential risks for cardiovascular disease and type 2 diabetes [26, 27].

Age further modifies the influence of genetic variants as the relative contribution of genetics to obesity changes across the life course [28]. Childhood PRS (46 genome-wide significant variants) often have stronger predictive power during early life stages as environmental confounders are relatively limited compared to adulthood [9, 29]. Conversely, adult PRS (941 ‘near’ genome-wide significant variants) tend to better predict obesity in later life, reflecting cumulative lifestyle exposures and age-related physiological changes, such as hormonal shifts during puberty, pregnancy or menopause [30].

Studies comparing these two PRS reveal dynamic genetic influences [31]. Childhood PRS is more associated with BMI during early life (<8 years), at which point the adult PRS becomes increasingly relevant.

As children age, diet, PA, stress, and socioeconomic conditions increasingly shape obesity risk. The following subsections will explore these gene-environment interactions (GxEs).

A recent meta-analysis [32] highlights how adherence to healthy dietary patterns, like those measured by the Healthy Eating Index (HEI), can mitigate PRS-related obesity risk. Higher HEI scores correlated with reduced BMI and waist circumference in individuals with high PRS. Plant-based diets rich in fruits, vegetables, and whole grains also consistently moderate PRS-related obesity risks [32].

Diets that are high in fat, low in carbohydrates, or include fried foods and sugar-sweetened beverages can increase the obesity risk associated with certain genetic variants [32]. For example, children carrying specific SNPs in genes like TMEM18 and FTO show increased sensitivity to these diets, leading to a higher BMI compared to those without these risk alleles [33, 34]. In contrast, when energy intake is restricted – with an adequate protein supply and maintained exercise – short-term changes in body composition occur similarly in both FTOrisk allele carriers and non-carriers, as demonstrated in a 4-week hypocaloric intervention in exercise-trained individuals [34].

Not all dietary patterns show consistent interactions. Several studies found no significant interactions between PRS and diets characterized by specific patterns, such as Western-style diets [32]. This may be – at least partially – explained by methodological limitations. Many of the studies examined by the meta-analysis relied on self-reported dietary data, and most research focuses on European-ancestry-based cohorts [32, 35].

PA has been shown to reduce the impact of specific obesity-related genetic risk factors [36, 37]. Regular exercise may mitigate genetic risk by enhancing metabolic efficiency and altering body composition. Notably, individuals with higher genetic predisposition to obesity often experience greater benefits from PA compared to those with lower genetic risk, highlighting the potential for targeted interventions [38].

However, studies on the interaction between PA and genetics have yielded variable results. While many demonstrate PA attenuate genetic risk, others report conflicting findings – particularly for metabolic outcomes like type 2 diabetes – which may depend on the specific PRS used [39, 40]. These inconsistencies may stem from differences in study design, population characteristics, measurement methods, and reliance on self-reported PA data, which can introduce potential biases and reduce statistical power [41]. Overall, these findings suggest that promoting fitness and activity early in life may effectively counteract the genetic predisposition to obesity over the long term, although further research is needed to clarify the underlying mechanisms and optimize intervention strategies.

Sleep duration plays an important role in modulating the effect of genetic predisposition to obesity. Twin studies show that shorter sleep durations amplify genetic influences on BMI, while longer durations suppress these effects. Moreover, shared environmental factors – such as healthy lifestyle habits – tend to exert a stronger influence with increased sleep [42].

In children and adolescents, maintaining healthy sleep duration along with limited screen time may offset genetic risk. For instance, a study involving 1,338 participants (6–17 years) found that with normative sleep and screen time, genetic predisposition had no significant impact on waist circumference [43]. However, the reliance on self-reported data and cross-sectional designs of the study limit the ability to draw causal inferences.

Furthermore, there is a modest genetic overlap between obesity and sleep traits, with shared genes like FTO playing a role. Research revealed modest genetic correlations were found, with the strongest for snoring, and U-shaped correlations with BMI and sleep durations, where both short and long sleep durations are linked to higher BMI. MR studies further suggest bidirectional effects: insomnia may contribute to an increase in BMI, while a higher BMI linked to snoring and daytime sleepiness [44].

Sun et al. [45] investigated the interplay between genetic predisposition and environmental stress by examining the relationship between PRS for BMI, cumulative stress exposure, and childhood obesity in a cohort of Chinese children. The researchers analyzed genetic data, stress biomarker (hair cortisol), and BMI, revealing a significant interaction effect. Specifically, children with a high PRS for obesity exhibited higher BMIs when exposed to elevated cumulative stress. Conversely, in the absence of significant stress, this genetic predisposition was less apparent and, in some cases, an inverse association was observed. Further research is needed to further grasp these complex interactions and their effects.

Another significant environmental modifier of genetic risk for childhood obesity with parental education emerged as the most reliable proxy [46]. Although income and neighborhood deprivation also contribute, they are less stable over time.

Socioeconomic position (SEP) affects PRS through both indirect and direct pathways. Higher SEP can mitigate genetic risk by promoting healthier diets, increased PA, and ensuring better access to resources [47, 48]. Data from the IDEFICS cohort indicates that high-PRS children with well-educated parents have a 32% lower obesity risk compared to peers with also a high-PRS but lower parental education [46]. In contrast, low SEP appears to amplify genetic risk. Stress-induced epigenetic changes in these environments have been associated with a two-fold increase in DNA methylation patterns linked to metabolic dysfunction [49].

Early life environments, largely shaped by SEP, are also critical. Exclusive breastfeeding for 5 months has been associated with notable reductions in BMI among high-PRS children [50]. Rapid infant weight gain – which is more prevalent in low-SEP settings – increases obesity risk, while neighborhood deprivation limits access to healthy food and safe play areas [51].

Long-term studies indicate that SEP and genetic risk have independent effects on obesity. The highest risk group is observed in children with both high PRS and low SEP. An Australian study confirmed these findings and further investigated “selection effects” – the hypothesis that families with higher PRS might have come to live in certain neighborhoods. However, the post hoc correlation analysis rejected that hypothesis [46, 52].

Understanding the interaction of genetic predisposition and environmental factors in childhood obesity is crucial for developing effective interventions. Moving beyond a one-size-fits-all therapy, personalized approaches based on individual genetics and environmental contexts show increasing promise [25]. Our overview of the literature shows that (modifiable) environmental risk factors need to be taken into account when assessing polygenic risk for developing obesity at young age.

There is an increase in adoption of personalized interventions, which use genetic information – particularly PRS – to customize approaches [53, 54], partly because high-density microarray genotyping has fallen from ≈€300 per sample a decade ago to as little as €35 when processed in large batches (>500 samples). Informing parents about their child's genetic predisposition to obesity, such as being in the top 5% compared to the general population, can serve as a powerful motivator for adopting an even healthier family lifestyle [55]. It is crucial to avoid inducing anxiety or a sense of fatalism by emphasizing the modifiable nature of obesity risk and providing actionable recommendations alongside genetic risk information, potentially complemented by genetic counseling and psychological support [25].

Recent epigenome-wide association studies, synthesized in a 2025 pediatric review, reveal that dietary modifications and increased PA can alter the expression of obesity-related genes; differential DNA methylation at loci such as ABCG1, CPT1A, SREBF1, SBNO2, and SOCS3 rises or falls with diet quality and activity level, highlighting a dynamic interplay between lifestyle interventions and gene regulation [56]. These observations suggest that environmental factors may trigger epigenetic mechanisms, such as DNA methylation, to modulate gene expression; although longitudinal data in children are accumulating, more research is still needed to confirm long-term implications. In addition, genetic differences may underlie responses to these interventions; for example, ADRB2 gene variants, which play a role in energy metabolism, have been linked to improvements in the effectiveness of exercise programs compared to patients without these variants [38]. While still under investigation, these findings open the possibility of personalized exercise prescriptions based on genetic profiles. The increasing use of medications like GLP-1 receptor agonists (e.g., semaglutide) for treating childhood obesity highlights the need for a more nuanced understanding of individual responses, and PRS may eventually play a role in predicting these responses for these new treatment types, similar to pharmacogenomics in other medical fields [57‒59].

Targeted intervention and prevention strategies are particularly needed for populations facing socioeconomic disparities. Children from low socioeconomic backgrounds bear a disproportionately higher risk of obesity, possibly due to a confluence of factors including limited access to nutritious foods, fewer opportunities for PA, and increased exposure to stress. Interventions specifically designed to address the challenges faced by these communities, such as community- and school-based programs, are essential to mitigate these disparities [60]. Although PA is advantageous for everyone, research suggests it may be particularly effective in reducing obesity among those with a high genetic predisposition. This suggests that people with elevated PRS may derive greater benefits from consistent exercise, even though everyone gains from being active [37]. Finally, looking ahead, as the ability to deconstruct PRS into biological pathways gets more refined, personalized interventions may be developed that account for specific GxEs. This means that even among individuals with identical overall PRS, those with different underlying genetic profiles might eventually receive personal recommendations that more precisely address their unique risk factors.

However, there are limitations that need to be addressed in future studies. First, many current risk scores rely on predominantly European-ancestry datasets – favored historically due to large biobanks – which limits their transferability. Differences in allele frequencies and genetic architectures can lead to inaccurate risk estimates in non-European populations, while GxEs may also vary across ethnic groups. Expanding research to include diverse populations and developing multi-ancestry PRS methods are essential for improving accuracy and ensuring equitable risk assessment. Second, implementing widespread population screening presents substantial financial and health-economic obstacles. While individual chip costs are EUR 35–40, the total expense for nationwide pediatric testing results into tens of millions of euros when factoring in additional necessities like sample collection, data processing, genetic counseling, and follow-up appointments. More focused strategies, such as screening only children with a family history of severe obesity or those residing in high-risk neighborhoods, seem financially more practical. Finally, while the predictive power of PRS is steadily improving, it should be viewed as just one piece of a larger risk assessment strategy.

GWAS have revealed numerous genes involved in childhood obesity – such as those in the leptin-melanocortin pathway (e.g., MC4R, FTO). Integrating these into PRS, offers a promising yet still modest tool for risk prediction. However, the clinical utility of PRS is limited by their small-effect sizes and reliance on predominantly European-ancestry data, underscoring the necessity of incorporating environmental factors like diet, PA, sleep, stress, and socioeconomic status into risk assessments. Emerging evidence suggests that lifestyle interventions can even modulate gene expression through epigenetic mechanisms, further supporting the potential for personalized, holistic approaches. In addition, ethical considerations – such as the risk of inducing anxiety or inequities in health care – must be addressed, especially when delivering sensitive genetic feedback. Overall, a comprehensive strategy that blends genetic insights with targeted public health initiatives is essential to effectively tailor early interventions and combat the complex challenge of childhood obesity.

We acknowledge the use of large language models for assistance with spelling, grammar, and phrasing during the preparation of this manuscript. All large language model-generated text was carefully reviewed, edited, and verified for accuracy by the authors.

The authors have no conflicts of interest to declare.

J.U.’s PhD is funded by the OBCT project (www.obct.nl), which has received funding from the European Union’s Horizon Europe Research and Innovation Programme under grant agreement No. 101080250. The views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the European Union. Neither the European Union nor the granting authority can be held responsible for them. The funder had no role in the design, data collection, data analysis, and reporting of this study.

Jorrit van Uhm performed writing – original draft, conceptualization, and visualization. Elisabeth F.C. van Rossum and Mieke M. van Haelst performed writing – review and editing. Philip R. Jansen performed writing – review and editing and visualization. Erica L.T. van den Akker performed writing – review and editing and supervision.