Genes, Adversity and Adult Disease: Massive Study Maps the Genetic Legacy of Childhood Maltreatment

Childhood maltreatment leaves scars that reach far beyond the years in which it occurs. Decades of epidemiological research have linked early abuse and neglect to depression, heart disease, chronic pain, respiratory illness and a host of other conditions in adulthood. Yet the biological mechanisms that carry this early adversity forward into adult bodies have remained stubbornly opaque. A new study published in Translational Psychiatry takes one of the most comprehensive genetically informed looks at this question to date, mining the genomes and electronic health records of nearly a million people to ask which adult diseases are truly entangled with the genetic liability to childhood maltreatment, and which apparent links may be explained by something else entirely, such as the socioeconomic environments in which children grow up.

The research, led by Solveig Løkhammer and Renato Polimanti of Yale School of Medicine together with colleagues at the University of Bergen and partner institutions across Europe and the United States, harnessed three of the largest biomedical databases in existence: the All of Us Research Program with 98,551 participants, the FinnGen Project with 377,277, and the Million Veteran Program with 449,042. Together these cohorts provided both genome-wide genetic data and longitudinal electronic health records, allowing the team to connect inherited variation with real clinical diagnoses across the human phenome. The scale matters. Childhood maltreatment is difficult to study directly, both because retrospective reports are imperfect and because the environmental and genetic factors that shape families are deeply intertwined. Genetic approaches offer a way to disentangle some of that tangle.

The centerpiece of the study was a phenome-wide association study, or PheWAS, of genetic variants previously linked to childhood maltreatment. Rather than asking whether one variant relates to one disease, a PheWAS flips the design: the researchers took the set of genetic loci associated with maltreatment liability and systematically tested them against thousands of coded medical outcomes in the health records of the three cohorts. The signal was striking. The maltreatment-associated loci showed associations with 66 distinct medical outcomes at a genome-wide significance threshold of p less than 2.76 × 10⁻⁶. The affected domains were broad, spanning psychiatric conditions, musculoskeletal disorders, circulatory disease, neurologic conditions and respiratory illness, suggesting that the genetic architecture of maltreatment risk is woven into many different systems of the body rather than confined to the brain.

To probe those relationships more deeply, the team turned to genetic correlation analysis, which estimates the degree to which two traits share the same underlying genetic influences across the entire genome. Here the picture became even richer: 285 significant genetic correlations with childhood maltreatment emerged at a threshold of p less than 3.24 × 10⁻⁵, again covering a wide range of medical domains. Among all of these, one relationship towered above the rest. Emotional disorders with onset during childhood or adolescence showed the highest genetic overlap with maltreatment liability, with a genetic correlation of 0.75. That figure implies a very substantial sharing of genetic influences between the liability to experience or report maltreatment and the liability to develop early-onset emotional disorders, a finding that resonates with long-standing clinical observations that abuse, anxiety and depression cluster within the same families and the same developmental windows.

Genetic correlation, however, cannot by itself tell us the direction of causation or whether the shared signal reflects a single causal variant or many. The researchers therefore deployed latent causal variable analysis, a method that uses the observed genetic correlations and their standard errors to distinguish genuine causality from confounding due to shared ancestry or overlapping variant effects. The LCV results pointed to genetic relationships between childhood maltreatment and musculoskeletal, metabolic, respiratory and eye-related outcomes, extending the story beyond psychiatry into physical health domains that are rarely discussed in the same breath as early adversity. The suggestion that maltreatment genetic liability travels with conditions affecting bones, joints, metabolism and vision adds a provocative new dimension to the biology of early-life stress.

The next question was whether these genetic relationships behave like causal effects, and the answer turned out to be complicated. Using bidirectional Mendelian randomization, a technique that uses genetic variants as natural instruments to test whether one trait influences another, the team found bidirectional and heterogeneous effects for the outcomes flagged by LCV. In other words, genetic instruments for maltreatment liability appeared to influence these health outcomes, but instruments for the outcomes also appeared to influence maltreatment liability, and the effects were not consistent across instruments. The authors interpret this pattern as possibly reflecting dynastic confounding, a subtle phenomenon in which the genotypes of parents and more distant ancestors shape the family environments into which children are born, blurring the line between inherited biology and inherited circumstance. In studies of parenting and childhood outcomes, dynastic effects are a serious methodological concern, and their presence here cautions against simple causal narratives.

Perhaps the most socially consequential finding came from the final layer of analysis. The researchers asked whether the genetic associations between childhood maltreatment and adult health might actually be carried, at least in part, by socioeconomic conditions, using household income as a proxy. Two complementary methods were applied: multi-trait conditional and joint analysis, which partitions genetic signal among correlated traits, and multivariate Mendelian randomization, which estimates the direct effect of one exposure while accounting for others that share its instruments. Both analyses converged on the same conclusion. When household income was taken into account, the effect of childhood maltreatment genetic liability on health outcomes was reduced. Socioeconomic factors, in other words, appear to explain a meaningful share of what had looked like a direct genetic relationship between early adversity and adult disease.

The authors are careful about what this does and does not mean. The finding does not suggest that maltreatment is benign, nor that its health consequences are illusory. Rather, it indicates that different pleiotropic mechanisms, in which the same genetic variants influence multiple traits, likely contribute to the observed relationships between maltreatment and several health domains, and that these relationships are at least partially explained by the socioeconomic contexts that shape both the risk of maltreatment and the trajectory of health across the life course. Genes, family income, neighborhood, and parenting behavior are not independent variables; they cascade through one another across generations. A study of this design cannot fully separate those strands, but it can measure how much of the apparent genetic signal survives when the socioeconomic strand is held steady, and here the answer is: less than the raw correlations implied.

For the field of trauma research, the study offers both a methodological template and a conceptual warning. The template is the integration of PheWAS, genetic correlation, latent causal variable analysis, and univariate and multivariate Mendelian randomization across multiple biobanks, a pipeline that interrogates the same question from several independent angles and reports where the answers agree and where they fracture. The warning is that observational associations between childhood adversity and adult disease, even those that appear in genetic data, can be shaped by dynastic and socioeconomic confounding in ways that standard adjustments do not capture. Interventions aimed at breaking the intergenerational cycle of maltreatment may need to target economic conditions as directly as they target parenting and mental health, because the two are genetically and environmentally entangled.

What emerges from this investigation is a portrait of childhood maltreatment as a genetically and socially embedded phenomenon whose fingerprints appear across psychiatric, musculoskeletal, circulatory, neurologic, respiratory, metabolic and even ocular health in adulthood. The strongest genetic overlap lies with emotional disorders beginning in childhood and adolescence, reinforcing the idea that the earliest years are a critical window in which liability to adversity and liability to emotional illness travel together. But the attenuation of maltreatment effects after accounting for household income is the finding with the clearest policy resonance: part of the long shadow of early adversity is cast not by biology alone but by the economic circumstances that surround a child. Disentangling those threads, the authors suggest, will require the continued growth of genetically informed cohorts and methods that can separate direct genetic effects from the inherited environments families pass down alongside their DNA.

Subject of Research: Genetic and socioeconomic mechanisms linking childhood maltreatment to adult health outcomes

Article Title: A genetically informed investigation into the relationship between childhood maltreatment and adult health

Article References: A genetically informed investigation into the relationship between childhood maltreatment and adult health. (n.d.). https://doi.org/10.1038/s41398-026-04498-1

Image Credits: AI Generated

DOI: 10.1038/s41398-026-04498-1

Keywords: childhood maltreatment, genetics, PheWAS, Mendelian randomization, genetic correlation, socioeconomic factors, household income, psychiatric disorders, Translational Psychiatry, All of Us, FinnGen, Million Veteran Program