Childhood Adversity Linked to Modest Rise in Heart Disease Risk Across Four Ageing Cohorts

What happens to a child can echo through the cardiovascular system for decades. A new analysis published in BMC Public Health suggests that adults who experienced more domains of adversity in childhood — with hunger emerging as a particularly stark marker — face a modestly higher risk of developing cardiovascular disease as they age. The study, led by AnNa Dai and colleagues at Wujin Hospital Affiliated with Jiangsu University, drew on four of the world’s largest longitudinal studies of ageing, spanning China, the United States, England, and continental Europe, and followed tens of thousands of middle-aged and older adults for incident cardiovascular disease.

The research team analyzed data from the China Health and Retirement Longitudinal Study (CHARLS), the Health and Retirement Study (HRS), the English Longitudinal Study of Ageing (ELSA), and the Survey of Health, Ageing and Retirement in Europe (SHARE). These four harmonized cohort studies interview representative samples of older adults, tracking their health, economic circumstances, and life histories over time. The complete-case primary samples were substantial: 9,336 participants in CHARLS, 2,142 in HRS, 6,224 in ELSA, and 13,912 in SHARE, giving the investigators a combined analytical population of more than 31,000 people across very different social and epidemiological settings.

Methodologically, the study is notable for how carefully it handles a thorny measurement problem. Childhood adversity cannot be measured identically across cohorts, because the retrospective questions asked of participants differ from survey to survey and from birth cohort to birth cohort. Rather than pretending that a single questionnaire captures adversity uniformly everywhere, the researchers constructed a cohort-specific observed-domain childhood adversity index (CAI) ranging from 0 to 4, in which the domains actually measured varied by participant and by study. Their primary analysis related each additional observed adverse domain to the risk of incident cardiovascular disease, pooling cohort-specific estimates using restricted maximum likelihood — a statistical technique that estimates the distribution of true effects across studies while accounting for sampling error.

The headline finding was a pooled hazard ratio of 1.07 per additional observed adverse domain, with a 95 percent confidence interval of 1.03 to 1.10. In plain terms, each additional measured domain of childhood adversity was associated with roughly a 7 percent higher rate of new cardiovascular disease during follow-up, after adjustment for age, sex, birth cohort, smoking, alcohol consumption, and body mass index. The inconsistency statistic I² was 0 percent, indicating that the cohort-specific estimates were statistically homogeneous — the four studies produced strikingly similar point estimates despite their different settings.

But the authors are emphatic that this apparent consistency should not be over-interpreted. They stress that the pooled figure is a descriptive average across non-equivalent, cohort-specific measures, not a common-scale or population-average effect. Statistical homogeneity, they note, is not the same thing as measurement equivalence: an I² of 0 percent tells you the numbers line up, not that the underlying constructs are identical. This distinction matters because retrospective adversity measures are shaped by the historical conditions in which respondents grew up — a person reporting childhood hunger in 1950s rural China and one reporting it in postwar Europe may be describing materially different experiences with different physiological consequences.

The results were not uniformly robust across every sensitivity check. In the conventional HRS analysis, the Model 2 estimate was not statistically significant, with a hazard ratio of 1.04 and a confidence interval spanning 0.93 to 1.16. Time-varying estimates also indicated that the association attenuated after five years of follow-up, raising the possibility that some of the observed excess risk reflects earlier-onset disease or reverse-causation pathways rather than a persistent life-course effect. When the researchers applied multiple imputation to selected covariates, the pooled hazard ratio was 1.06 with a Hartung–Knapp confidence interval of 1.02 to 1.09, remaining statistically significant. These methodological choices — Hartung–Knapp intervals, imputation, sequential adjustment — reflect a deliberate effort to quantify how fragile or durable the association really is.

Childhood hunger received its own dedicated analysis. Direct hunger measures were available in CHARLS and SHARE, and the cohort-specific estimates were positive in both, but when pooled across just those two cohorts the Hartung–Knapp confidence interval included the null value, ranging from 0.95 to 1.41. In other words, while both studies pointed in the same direction, the evidence was too imprecise to rule out no effect. An exploratory analysis examined whether childhood hunger interacted additively with later-life loneliness or social isolation — testing the idea that early deprivation and late-life psychosocial stress might compound each other’s cardiovascular toll. That five-year additive interaction was supported only in CHARLS, while the SHARE estimate was too imprecise to be informative, leaving the interaction hypothesis open rather than settled.

The team also constructed a three-domain life-course burden measure, which showed a positive average association with incident cardiovascular disease, though the categorical patterns across burden levels were non-uniform, meaning the risk did not simply climb in neat steps with each additional domain. Separately, in CHARLS alone, the researchers applied an adapted version of the American Heart Association’s cardiovascular-kidney-metabolic (CKM) staging framework, a recent scheme for classifying the progression of combined heart, kidney, and metabolic dysfunction. The authors are careful to flag that these CKM results are cross-sectional, CHARLS-only, and analytically separate from the primary longitudinal findings — a snapshot, not a trajectory.

Perhaps the most striking feature of the paper is the discipline of its conclusions. The authors explicitly state that the findings support further life-course risk research, not causal attribution, not claims about intervention efficacy, and not preventive targeting. They note that no absolute-risk difference was estimated for the primary CAI contrast, and that the results should not be read as population-average effects or as a validated common-scale measure of adversity. In an era when early-life adversity is frequently invoked in public-health advocacy, this restraint is a model of how observational epidemiology should be communicated: an association observed under cohort-specific measures, with honest uncertainty attached at every step.

Even with those caveats, the study adds meaningful weight to a growing body of evidence that cardiovascular risk is seeded across the entire life course, not merely accumulated in middle age through diet, smoking, and inactivity. The consistency of the direction of association across four cohorts on three continents — China, the United States, England, and continental Europe — despite heterogeneous measurement, suggests that childhood adversity deserves a place in longitudinal cardiovascular risk research alongside established behavioral and metabolic factors. What remains to be established is mechanism and magnitude: whether early deprivation acts through persistent metabolic and inflammatory changes, through the clustering of adult risk behaviors, through socioeconomic pathways, or through some combination of all three. The authors received no specific grant funding for the research, and the secondary analysis used de-identified data from cohorts approved by their respective ethics committees, including Peking University’s Biomedical Ethics Review Committee for CHARLS, the University of Michigan’s institutional review board for HRS, London Multi-Centre Research Ethics Committee for ELSA, and the ethics bodies of the University of Mannheim and the Max Planck Society for SHARE. The work, published open access on 5 October 2026, offers researchers a carefully calibrated foundation for the next generation of life-course cardiovascular studies — and a reminder that the heart keeps a record of childhood that science is only beginning to read.

Subject of Research: The association between cohort-specific childhood adversity measures and incident cardiovascular disease in four longitudinal ageing studies

Article Title: Cohort-specific childhood adversity measures and incident cardiovascular disease across four ageing cohorts: analyses of childhood hunger, later-life psychosocial indicators, and life-course burden

Article References: Dai, A., Yin, J., Zheng, X., & Gong, Y. (2026). Cohort-specific childhood adversity measures and incident cardiovascular disease across four ageing cohorts: analyses of childhood hunger, later-life psychosocial indicators, and life-course burden. BMC Public Health. https://doi.org/10.1186/s12889-026-29777-w

Image Credits: AI Generated

DOI: 10.1186/s12889-026-29777-w

Keywords: childhood adversity, cardiovascular disease, childhood hunger, life course, cohort study, CHARLS, HRS, ELSA, SHARE, loneliness, epidemiology, ageing