Measures

Statistical analyses

All analyses were undertaken using SPSS v. 29. Univariate and bivariate statistics (i.e., frequency/percentages and chi-square tests) are presented to describe the sample and the relationships between each of the study variables and cancer. To address our research questions, two binary logistic regression analyses with cancer as the outcome were conducted, the first included all types of childhood adversities, the second included all types of childhood adversities and all the control variables (described above). Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated by exponentiating the logistic regression β-coefficients. To assess model fit, Nagelkerke Pseudo R² values were reported, reflecting the percentage of variance explained. Data from all the analyses were weighted to adjust for the probability of selection and non-response. Sample sizes are reported in their unweighted form. Statistical significance was defined as p < 0.05.

Childhood sexual abuse and cancer risk in older adulthood

The prevalence of cancer was significantly higher among respondents who experienced CSA compared to those without CSA (CSVC: 35.5%; Unwanted touch/fondling: 25.5%; no CSA: 20%). In the logistic regression analyses, after adjusting for demographic, socioeconomic, behavioral, and psychosocial covariates, CSVC remained significantly associated with approximately double the odds of cancer compared to individuals without CSA (OR = 2.05, 95% CI = 1.41–2.98). Unwanted touch/fondling, however, was not significantly associated with cancer risk after covariate adjustments.

The current study’s findings on older Canadians align with Hovdestad et al. [12], who found that CSA was significantly associated with increased cancer risk, particularly among adult Canadian women. Their study identified a dose-response relationship between CSA severity and cancer prevalence. While they used a tiered model based on frequency and severity, our dichotomous model focusing on CSVC still yielded a similarly robust association. This suggests that even a single experience of sexual violence involving coercion may be associated with long-lasting physiological consequences relevant to cancer risk, particularly in older adulthood.

Earlier research suggests that survivors of CSVC show greater long-term dysregulation of stress-response systems, particularly chronic alterations in cortisol production and heightened inflammation [29,30]. Such biological disruptions have been implicated in the development and progression of cancer through mechanisms involving chronic inflammation and immune dysregulation [9,31,32]. Thus, CSVC may exert independent physiological effects, increasing vulnerability to cancer in ways distinct from other forms of childhood adversity.

Childhood Physical abuse, parental domestic violence, and cancer risk

Individuals who experienced CPA in the current study had a significantly higher prevalence of cancer compared to those who did not (28% vs. 20.7%). Similarly, respondents exposed to PDV had an elevated cancer prevalence relative to unexposed individuals (26.5% vs. 20.6%). In logistic regression analyses, however, after adjusting for all forms of childhood adversity, neither CPA nor PDV significantly predicted cancer risk. Further adjustment for demographic, socioeconomic, behavioral, and psychosocial covariates failed to further attenuate the odds.

Hovdestad et al's. study [12] had a much larger sample size than was available in the MHACS, which provided them with adequate statistical power to conduct sex-specific analyses. Their findings indicated that CPA and PDV were significantly associated with cancer for women, but not for men, in their 2012 nationally representative sample of adults. Among women, inclusion of a wide range of potential mediating factors (e.g., smoking, stress, depression and substance use disorders) provided moderate attenuation of the direct relationship with cancer. It is unclear why adjustment for other potential risk factors was associated with a much greater attenuation of the childhood adversity-cancer relationship in the 2012 sample of women than in our current 2022 mixed sex sample. It is important to note that they used a broader definition of CPA, including less severe behaviors like slapping and pushing, whereas our measure focused exclusively on more severe acts such as kicking or choking.

Hovdestad and colleague’s findings align with previous literature, indicating that the CPA and PDV association with cancer often diminish or become non-significant after adjusting for broader psychosocial and environmental factors [15]. One possible explanation for this attenuation is that CPA and PDV commonly co-occur within broader contexts of familial dysfunction, socioeconomic disadvantage, and chronic stress – which are associated with long-term health outcomes [33]. Furthermore, the initially observed associations between these adversities and cancer risk in bivariate analyses may reflect indirect pathways through shared risk factors – such as childhood poverty [34], and lifetime cumulative stress exposure [35] – rather than the direct effects of CPA or PDV themselves. Future research should explore how these cumulative exposures, alongside intervening stressors across the life course, influence cancer susceptibility, potentially through chronic physiological stress responses and prolonged inflammatory states.

Potential mechanisms linking ACEs to cancer risk

Although this observational study prohibits identification of causal mechanisms, we hypothesize the association between childhood adversities— CSVC, unwanted touch/fondling, CPA, and PDV—and elevated cancer prevalence in later life may be partially explained by biological embedding, a process by which early-life stressors disrupt stress regulation and immune function, influencing long-term health [36]. ACEs such as CSVC, unwanted touch/fondling, CPA, and PDV, have been linked to dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis, chronic inflammation, and immune suppression—all of which are implicated in cancer pathogenesis through mechanisms like oxidative stress, DNA damage, and impaired tumor surveillance [37–39].

There is evidence that childhood adversity may contribute to long-term physiological changes that could be relevant to cancer risk. Studies have shown that adults with histories of childhood maltreatment exhibit higher levels of systemic inflammation than those without such experiences [32]. A recent meta-analysis similarly found that childhood adversity is associated with elevated inflammatory activity across developmental stages [40]. Although our study cannot establish causality, these findings suggest that early stress may become biologically embedded over time, contributing to chronic low-grade inflammation implicated in the development of chronic diseases, including cancer [32].

While only CSVC emerged as a significant predictor of cancer after controlling for a wide range of factors, the ORs for CPA, PDV, and unwanted touch/fondling were similar in magnitude (approximately 20% higher odds than their peers without these early childhood adversities). This pattern suggests that common physiological disruptions may underlie all ACEs, even if statistical significance was not reached when all these adversities were considered simultaneously. Future research with larger studies are needed to explore this hypothesis, and studies should examine whether the cumulative impact or severity of adversity differentially affects biological pathways linked to cancer development.

Unexpected findings & covariate adjustments

It was expected that the association between CSVC and cancer risk would be substantially attenuated after adjusting for health behaviors, SES, and psychosocial factors. When all three ACEs were modeled together in Block 1 (see Table 2) of our regression, only CSVC remained statistically significant. The addition of further covariates in Block 2 (See Table 2) did not substantially alter the ORs. This stability highlights the independent contribution of CSVC to cancer risk—an insight echoed by Hovdestad et al. [12], who similarly found that CSA (particularly among women) remained associated with cancer after accounting for co-occurring maltreatment and risk behaviors.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Logistic Regression of Factors Associated with Lifetime Cancer Among Canadians Aged 65 and Older (N = 2,636).

https://doi.org/10.1371/journal.pone.0345411.t002

An unexpected finding in our fully adjusted model was that White respondents had higher odds of reporting a cancer diagnosis compared to those identifying as a visible minority. This pattern contrasts with some prior work suggesting that racialized populations may experience greater cumulative adversity and stress exposure across the life course. Several explanations are possible. Visible minority groups in Canada may under-report cancer diagnoses due to barriers in healthcare access, lower screening participation, or cultural differences in disclosure. It is also possible that older White Canadians in this sample had higher historical exposure to well-established cancer risk factors, such as smoking or occupational hazards, which may contribute to this pattern. Furthermore, the majority of visible minority older Canadians are immigrants. There is a well-established healthy immigrant effect among immigrants to Canada, which is in keeping with the current findings.

We also observed significant associations for several “missing” categories (e.g., race/ethnicity, alcohol abuse). These should be interpreted with caution, as they likely reflect systematic patterns of non-response rather than true underlying relationships with cancer. Missingness in national surveys is rarely random; individuals who skip sensitive items may differ in health status, trust in institutions, or comfort disclosing personal information. As such, the elevated odds among respondents with missing data likely indicate measurement or reporting error rather than meaningful associations. This highlights the importance of cautious interpretation when working with self-reported survey data and underscores the need for improved measurement of key constructs in future research.

Prior research has emphasized smoking, obesity, and physical inactivity as key behavioral pathways linking ACEs to long-term health outcomes, including cancer [41,42]. However, the present study found that adjusting for these factors did not significantly weaken the relationship between CSVC and cancer risk, suggesting that association between ACEs and problematic health behaviors do not fully account for the observed association.

Inclusion of education level and household income, which are well-documented protective factors against cancer [43,44], did not substantially attenuate the association between CSVC and cancer. While childhood adversity is strongly associated with lower adult SES, the persistence of the CSVC–cancer link, even after accounting for these factors, suggests that socioeconomic disadvantage does not fully mediate this relationship. As was discussed above, one possible explanation is that early-life sexual violence involving coercion may exert direct biological and psychological effects on long-term health, independent of socioeconomic conditions [45]. Additionally, individuals with a history of CSA may develop maladaptive coping mechanisms, such as substance use [46], and disordered eating [47], that can contribute to cancer risk through behavioral pathways. We were able to control for substance use disorders in our logistic regression analyses, but not disordered eating.

This study also statistically adjusted for psychosocial variables, including social support and spirituality. However, the association between CSVC and cancer risk remained virtually unchanged, reinforcing the idea that early-life trauma may exert lasting biological effects that persist even in the presence of protective psychosocial resources. These findings highlight the need for future research to examine stress-related epigenetic changes, chronic inflammation, and other physiological pathways that may underlie the link between CSA and cancer in older adulthood.

Implications for research and practice

The findings of this study emphasize the association between later-life health conditions and early-life trauma, particularly the independent association between CSVC and increased cancer risk in older adulthood. Given that this relationship persisted even after adjusting for SES, mental health conditions, health-risk behaviors, and social support, it is crucial for future research to further explore the biopsychosocial pathways linking childhood adversity to cancer in older adulthood [48].

Our findings, alongside those of Hovdestad et al. [12], highlight the importance of incorporating trauma history into preventative cancer care. Both studies found strong associations between CSVC and cancer risk, reinforcing the need for trauma-informed outreach and interventions to promote equitable access to early detection and support services [49]. Emerging research on trauma-exposed populations suggests that psychosocial barriers—such as those associated with serious mental illness and adverse childhood experiences—may inhibit engagement in cancer screening, with studies showing lower uptake of breast and cervical screening among women with serious mental illness [50] and complex, ACE-specific associations with screening behaviors [51]. Future research should explore these barriers further, as well as potential biological markers—such as inflammation and dysregulated cortisol—to inform integrated, multidisciplinary approaches to cancer survivorship care.

It has been proposed that healthcare providers should implement standardized screening protocols for childhood trauma in routine clinical practice to better identify individuals at risk for long-term health complications, including chronic disease and mental illness [52]. While our study focused on physical health outcomes, the link between CSA and psychological distress—particularly depression and post-traumatic stress disorder—has been well-established in longitudinal studies [53]. As depression has also been identified as a potential risk factor for cancer [54], future studies should consider including it as a covariate to better elucidate the pathways through which early-life trauma may influence later cancer risk. Incorporating trauma screening into cancer care could also support more holistic assessment and intervention efforts, especially when paired with access to evidence-based mental health supports. While the evidence linking ACEs and cancer is growing, findings across studies are not always consistent—likely due to methodological differences in how ACEs are measured, the types and timing of outcomes assessed, and whether studies adjust for specific psychosocial and behavioral mediators, such as chronic stress [55], psychological distress [56], and health behaviors like smoking, alcohol consumption, and obesity [57].

Furthermore, when older adults are diagnosed with cancer, incorporating evidence-based mental health interventions, such as Cognitive Behavioral Therapy (CBT) [58], or Emotion-Focused Therapy [59] may play a critical role in comprehensively addressing psychological distress. Although no studies have yet demonstrated that psychological interventions such as CBT directly reduce cancer incidence, emerging evidence suggests that these interventions can modulate systemic inflammation and stress-related biological pathways and improve health behaviors that are relevant to cancer risk and survivorship [60]. Furthermore, Managing Cancer and Living Meaningfully (CALM) therapy is a psychotherapeutic intervention designed to support patients with advanced cancer by providing emotional support, alleviating distress, and helping them navigate the complexities of living with a serious illness [61]. CALM therapy has been widely studied in palliative care, demonstrating its effectiveness in reducing existential distress and improving quality of life for individuals with advanced cancer. Given these benefits, there is growing interest in expanding CALM beyond palliative care into primary healthcare settings across Canada to enhance psychosocial support for cancer patients. Its adoption has also gained international traction, with successful implementations in the Netherlands [62], Japan [63], and Italy [64], underlining its global potential as a standard component of oncology care.

Limitations

There are several important limitations to this research that must be considered when interpreting the results. This study examines the associations between childhood adversities—specifically CPA, CSA (unwanted touch/fondling and CSVC), and PDV—and cancer risk in older adulthood but was constrained by the use of secondary data. Although this study adjusts for several key covariates—including SES, social support, and substance use —residual confounding remains a concern. For example, malignancies involving inflammatory or immune-related pathways may be more strongly linked to early-life trauma. Investigating these relationships by cancer type would provide more precise insights into potential associative mechanisms of risk.

The MHACS data did not include several factors that have been implicated in cancer risk including body mass index (BMI) [65], information on the duration of the childhood abuse nor the perpetrator [66], genetic predisposition [67], early-life health conditions [68], poverty in childhood [69], or other childhood stressors, such as neglect, exposure to community violence, or peer victimization, all of which could contribute to long-term health vulnerabilities [5,15,70]. Future research would benefit from inclusion of anthropometric data to examine whether BMI plays a mediating role in the CSA-cancer link. Another limitation of this study is that the MHACS Public Use Dataset provides only a single aggregated age category for respondents aged 65 and older. It would be preferable in future research to be able to take into account the well-established relationship between age within 10 year cohorts and cancer risk.

Another limitation is that the study does not differentiate between cancer types. Given that cancers have distinct etiologies and risk factors, previous research suggests associations between childhood adversity and cancer may vary depending on cancer subtype [8–10]. The cross-sectional design also prohibits the determination of causality between ACEs and cancer risk, as unmeasured factors may contribute to both. Additionally, reliance on self-reported data introduces the potential for recall bias and misclassification, particularly with retrospective reporting of trauma, which may be influenced by memory limitations, stigma, or reinterpretation over time [71]. Furthermore, the cancer diagnoses in the MHACS were based on self-reports rather than medical records, potentially leading to misclassification. However, while prior research suggests self-reports of cancer diagnoses are generally accurate in large-scale epidemiological surveys [72], correctly identifying approximately 73% of true cases and accurately ruling out 96% of non-cases [73,74], they are not validated by chart review in the current study and should be in future research.

Selection bias may further affect the findings, as the MHACS is a community-based survey and does not include individuals who are hospitalized or institutionalized. Hospitalized individuals may have higher rates of both trauma exposure and serious illness, which could have resulted in an underestimation of the association between childhood adversities and cancer. Additionally, national differences in healthcare access and sociocultural factors across populations may restrict the generalizability of findings beyond Canada. Furthermore, the response rate in the MHACS was 30%, which is less than ideal. This may have led to bias in the conclusions drawn.

Despite these limitations, the current study provides a Canadian population-based study of older adults. We found that the odds of cancer among older adult survivors of CSVC were double that of their peers without a history of CSA. This association persisted even after controlling for multiple demographic, socioeconomic, behavioral, and psychosocial factors. Future longitudinal research using chart-reviews of cancer diagnoses is needed to clarify the specific biopsychosocial mechanisms underlying the childhood adversity-cancer’s relationship, incorporating factors such as chronic inflammation, immune dysregulation, and epigenetic changes that were not assessed in this study. Expanding research to explore these pathways may inform targeted prevention and intervention strategies aimed at improving the long-term health of child abuse survivors.