Childhood maltreatment leaves scars that often reach far beyond early life. Decades of research show that people who were maltreated as children are more likely to experience psychosocial difficulties, mental health problems, and physical illnesses as adults (see the previous Mental Elf blog post here in the context of childhood sexual abuse) (Goodwin & Stein, 2004). These problems often occur together; a phenomenon called multimorbidity, meaning two or more conditions occur at the same time (Wan et al., 2022).
In the context of maltreatment, multimorbidity often includes depression, heart disease, and type 2 diabetes. Recent studies suggest that maltreatment may not only raise the risk for each of these conditions individually, but also increases the risk of them occurring together (Baltramonaityte et al., 2023). However, how this happens remains unclear. One theory is that maltreatment causes long-term changes in the body, such as inflammation or metabolic disruption, which in turn make people more vulnerable to developing other diseases later in life (Li & Danese, 2018).
A new study by Baltramonaityte et al. (2025) aimed to explore this further by exploring these mechanisms using genetic data.
Childhood maltreatment may trigger some mechanisms that lead to the development of other diseases later in life.
Methods
The researchers focused on three types of biological markers that link genetic liability to childhood maltreatment with multimorbidity:
- Inflammatory proteins
- Metabolic markers
- Cortisol
For their analyses, they made use of a method called Mendelian Randomization. This approach uses genetic variants associated with an exposure (here: childhood maltreatment) to estimate whether that exposure might cause health outcomes. In their study, they combined Mendelian Randomization with a mediation analysis. This allowed them to test whether the biological markers mentioned above might be the middle step between genetic liability for maltreatment and multimorbidity.
They defined multimorbidity as a latent genetic factor (an unobserved mechanism that may contribute to a trait/disease, that isn’t directly measured, but can be identified by modelling other related traits) capturing risk for depression, coronary artery disease, and type 2 diabetes together. Data on childhood maltreatment liability, biological markers, and disease outcomes came from large, publicly available genome-wide association studies.
The analysis had three steps:
- Check associations – Were the biological markers linked to both genetic risk for childhood maltreatment and multimorbidity?
- Adjust for maltreatment – Did these links with multimorbidity remain even after accounting for maltreatment risk?
- Test mediation – Did the markers actually explain (mediate) part of the maltreatment–multimorbidity relationship?
Results
Step 1: Only a handful of markers were linked to both childhood maltreatment liability and multimorbidity:
- Inflammation marker: High C-reactive protein (a pro-inflammatory protein produced during an immune response)
- Metabolic makers: High glycated haemoglobin (average blood sugar levels), high triglycerides (blood fats), high non-HDL cholesterol (“bad” cholesterol), low HDL cholesterol (“good” cholesterol)
Cortisol was neither linked to childhood maltreatment liability nor to multimorbidity.
Step 2: When adjusting for childhood maltreatment liability, 19 markers were still associated with multimorbidity, including the five above. This suggests they may contribute to multimorbidity risk beyond the direct effect of childhood maltreatment liability.
Step 3: Mediation analysis showed that three markers significantly explained part of the link between childhood maltreatment liability and multimorbidity:
- Triglycerides (10.1 to 11.2% of the link)
- Glycated haemoglobin (7.3 to 8.4% of the link)
- HDL cholesterol (5.8 to 6.9% of the link)
Non-HDL cholesterol initially appeared to mediate the link, but this finding did not hold after correcting for multiple comparisons. Neither inflammation nor cortisol mediated the relationship. In other terms, people with a genetic risk for childhood maltreatment also tended to have higher blood fats and sugar, which in turn increased their risk for having depression, heart disease, and type 2 diabetes together.
Moreover, the authors explored whether the biological mediators also explained the increased risk of individual diseases after childhood maltreatment. This was only possible to test for depression as no association was found between childhood maltreatment liability and the other diseases. The findings showed that none of the markers explained the link between genetic liability for childhood maltreatment and depression alone.
Metabolic markers appear to link genetic liability for childhood maltreatment to multimorbidity.
Conclusions
The authors conclude that:
up to 11% of the effect of childhood maltreatment liability on multimorbidity was mediated by triglycerides, up to 8% by glycated haemoglobin, and up to 7% by HDL cholesterol.
This fits with the idea that early adversity can get under the skin and shape metabolic health, influencing the development of multiple conditions. Additionally, since these markers did not explain the link between childhood maltreatment liability and depression alone, biological pathways may be more easily identified for multimorbidity that for individual diseases.
By looking at multimorbidity, it may be easier to detect common biological pathways to diseases
Strengths and limitations
This is a well-conducted and innovative study with several notable strengths:
- Novel approach: It is one of the first attempts to use genetic methods to explore how childhood maltreatment might contribute to multimorbid diseases, rather than looking at each condition in isolation. Studying multimorbidity may make it easier to spot shared biological pathways, as the combined effect of multiple conditions could reflect stronger underlying dysregulations.
- Large-scale data: The analysis used summary statistics from very large genome-wide association studies, drawing on publicly available data to boost accuracy.
- Careful design: The team used non-overlapping samples for the exposure and outcome datasets, which helps to reduce bias and lower the risk of false positives.
As with any study, there are limitations to consider.
First, as mentioned by the authors, the results could be influenced by genetic confounding, so it’s important to confirm them with other methods and in populations of different ancestries (this study included participants of European ancestry only).
Second, the candidate biological mediators were analysed one at a time in separate (univariate) models. In reality, these markers often interact with one another, so future research should examine them together using multivariate models. Other influences, such as lifestyle factors, may also play a mediating role and deserve further investigation. Then, childhood maltreatment in this study was assessed using both retrospective reports (asking adults to recall childhood experiences) and prospective reports (gathering information during childhood from informants or the children themselves). These tools capture slightly different aspects of maltreatment, which can add heterogeneity. However, they were strongly genetically correlated, which increases confidence in the reliability of the findings.
Finally and, in my opinion, most importantly, it is difficult to draw strong conclusions about how childhood maltreatment itself causes multimorbid diseases. Mendelian randomization is designed to estimate causal associations by reducing confounding (Smith & Ebrahim, 2005), but it relies on several assumptions that are harder to meet for complex, non-biological exposures such as maltreatment, as acknowledged by the authors. Genetic liability to childhood maltreatment – carrying variants linked to a higher likelihood of being maltreated (Warrier et al., 2021) – may reflect not just maltreatment itself but also related behavioural traits, family environments, or socioeconomic factors. When a genetic variant influences multiple traits at once (horizontal pleiotropy), it becomes difficult to know which trait is truly responsible for the health outcome. Another assumption of Mendelian randomization, gene–environment equivalence (the idea that genetic variants are randomly distributed across environments), may also be violated. Exposure to maltreatment is shaped by environmental factors that could independently affect disease risk. The authors addressed these issues with sensitivity analyses, which showed modest evidence of bias. This warrants a careful interpretation of the findings as they may not signify a direct causal pathway from childhood maltreatment to biological mediators to multimorbidity.
Genetic liability for childhood maltreatment likely reflects a combination of inherited traits and environmental factors.
Implications for practice
The effects of childhood maltreatment often echo far beyond childhood – but in the future, we may be able to soften that echo. I believe prevention is the ultimate goal, but for people who have experienced maltreatment, we should also focus on minimising the long-term health burden.
The study of Baltramonaityte et al. (2025) suggests that higher blood lipids and blood sugar may be part of a biological chain linking genetic vulnerability to childhood maltreatment and the development of multimorbid depression, coronary heart disease, and type 2 diabetes. If confirmed, these pathways could offer opportunities for prevention and early intervention. For example, genetic testing (already used in some areas of medicine) could one day help identify people at higher risk of developing multimorbidity. In such cases, clinicians might monitor blood lipids and sugar more closely, stepping in early to correct any abnormalities before disease develops.
But we are not there yet. These results need replication, their clinical significance needs to be established, and intervention trials will be necessary to test whether targeting these biological mediators actually reduces the incidence of multimorbidity.
As an early-career researcher and clinician who recently started working in a medical psychology unit, I have become increasingly aware of how mental and somatic health are intertwined. When they are both compromised, suffering can feel exponential, making it all the more important to prevent, detect, and treat these problems wherever they (co-)occur.
Mental and somatic diseases often co-occur and are linked to a heavy individual suffering and societal burden.
Statement of interests
From 2020 until 2024, I participated in the EarlyCause consortium (earlycause.eu) together with the authors of this study. I have collaborated with some of them on other research papers, but was not involved in the study presented or its peer-review evaluation.
I used ChatGPT to improve the readability and flow of the text I originally wrote.
Links
Primary paper
Baltramonaityte, V., Karhunen, V., Felix, J. F., Penninx, B. W. J. H., Cecil, C. A. M., Fairchild, G., Milaneschi, Y., & Walton, E. (2025). Biological pathways underlying the relationship between childhood maltreatment and Multimorbidity: A two-step, multivariable Mendelian randomisation study. Brain, Behavior, and Immunity, 126, 59–69. https://doi.org/10.1016/j.bbi.2025.01.024
Other references
Baltramonaityte, V., Pingault, J.-B., Cecil, C. A. M., Choudhary, P., Järvelin, M.-R., Penninx, B. W. J. H., Felix, J., Sebert, S., Milaneschi, Y., Walton, E., & Consortium, on behalf of the E. (2023). A multivariate genome-wide association study of psycho-cardiometabolic multimorbidity. PLOS Genetics, 19(6), e1010508-. https://doi.org/10.1371/journal.pgen.1010508
Goodwin, R. D., & Stein, M. B. (2004). Association between childhood trauma and physical disorders among adults in the United States. Psychological Medicine, 34(3), 509–520. https://doi.org/10.1017/S003329170300134X
Li, J. C., & Danese, A. (2018). Biological Embedding of Child Maltreatment Through Inflammation. In J. G. Noll & I. Shalev (Eds.), The Biology of Early Life Stress: Understanding Child Maltreatment and Trauma (pp. 1–14). Springer International Publishing. https://doi.org/10.1007/978-3-319-72589-5_1
Smith, G. D., & Ebrahim, S. (2005). What can mendelian randomisation tell us about modifiable behavioural and environmental exposures? BMJ, 330(7499), 1076. https://doi.org/10.1136/bmj.330.7499.1076
Wan, A., Bernstein, C. N., Graff, L. A., Patten, S. B., Sareen, J., Fisk, J. D., Bolton, J. M., Hitchon, C., Marriott, J. J., Marrie, R. A., & Disease, for the C. T. in D. the B. and M. the E. of I. I. (2022). Childhood Maltreatment and Psychiatric Comorbidity in Immune-Mediated Inflammatory Disorders. Biopsychosocial Science and Medicine, 84(1).
Warrier, V., Kwong, A. S. F., Luo, M., Dalvie, S., Croft, J., Sallis, H. M., Baldwin, J., Munafò, M. R., Nievergelt, C. M., Grant, A. J., Burgess, S., Moore, T. M., Barzilay, R., McIntosh, A., van IJzendoorn, M. H., & Cecil, C. A. M. (2021). Gene–environment correlations and causal effects of childhood maltreatment on physical and mental health: a genetically informed approach. The Lancet Psychiatry, 8(5), 373–386. https://doi.org/10.1016/S2215-0366(20)30569-1
