The aftermath of childhood maltreatment, involving sexual, physical, emotional abuse, or neglect, can be devastating and result in serious lifelong psychological, interpersonal, and physical consequences (Burczycka & Conroy, Reference Burczycka2017; Gilbert et al., Reference Gilbert, Widom, Browne, Fergusson, Webb and Janson2009). For example, survivors of childhood maltreatment are at a higher risk of developing anxiety and mood disorders or symptoms (Li, D'Arcy, & Meng, Reference Li, D'Arcy and Meng2016), report more stress (Bossé, Stalder, & D'Antono, Reference Bossé, Stalder and D'Antono2018; Garami et al., Reference Garami, Valikhani, Parkes, Haber, Mahlberg, Misiak and Moustafa2019; LoPilato et al., Reference LoPilato, Addington, Bearden, Cadenhead, Cannon, Cornblatt and Walker2020), and are at an increased risk for developing chronic physical ailments such as coronary artery disease (CAD) (Jakubowski, Cundiff, & Matthews, Reference Jakubowski, Cundiff and Matthews2018). Mechanisms leading to increased risk of CAD or other chronic physical conditions are likely diverse and may involve emotional dysregulation and biological changes at the cellular level, such as in leukocyte telomere length (LTL).
Telomeres are non-coding DNA sequences on the ends of chromosomes that are meant to protect the coding sequences of DNA from potential damage that may occur during the cell division process (Blackburn, Reference Blackburn2001; Blasco, Reference Blasco2005). Because of their incomplete replication and progressive shortening over time (Blackburn, Reference Blackburn2000), telomere length (TL) has been suggested as a marker of biological aging (Terry, Nolan, Andersen, Perls, & Cawthon, Reference Terry, Nolan, Andersen, Perls and Cawthon2008; Zglinicki & Martin-Ruiz, Reference Zglinicki and Martin-Ruiz2005). Once telomeres reach a critical length, affected cells may fail to operate normally or die, contributing to disease development (Brouilette et al., Reference Brouilette, Moore, McMahon, Thompson, Ford, Shepherd and Samani2007; Nilsson, Dahlman, Roos, Nordfjäll, & Melander, Reference Nilsson, Dahlman, Roos, Nordfjäll and Melander2011; Samani & van der Harst, Reference Samani and van der Harst2008). Importantly, a growing body of research suggests that exposure to childhood maltreatment is associated with shorter TL.
Three meta-analyses reported small but significant relations between childhood trauma or adversity and TL in adulthood (Bürgin et al., Reference Bürgin, O'Donovan, d'Huart, di Gallo, Eckert, Fegert and Boonmann2019; Li, He, Wang, Tang, & Chen, Reference Li, He, Wang, Tang and Chen2017; Ridout et al., Reference Ridout, Levandowski, Ridout, Gantz, Goonan, Palermo and Tyrka2018). Considerable heterogeneity was observed, which the authors attributed to potential individual differences, including sex, health status, and characteristics of the studies, such as sample size and included covariates. For example, Bürgin et al. (Reference Bürgin, O'Donovan, d'Huart, di Gallo, Eckert, Fegert and Boonmann2019) observed that results from larger studies were more heterogenous, which they hypothesized was due to a greater statistical control of variables that may actually moderate or mediate the relationship. Controlling for them may thus conceal associations of childhood maltreatment with TL.
There is ample reason to believe that individual differences may exist. Men and women, for example, differ in the types of traumas they may have been exposed to (Burczycka & Conroy, Reference Burczycka2017; Gilbert et al., Reference Gilbert, Widom, Browne, Fergusson, Webb and Janson2009) and data suggests that men generally have shorter TL, and faster attrition rates compared to women (for a review see Barrett and Richardson, Reference Barrett and Richardson2011). Although Bürgin et al. (Reference Bürgin, O'Donovan, d'Huart, di Gallo, Eckert, Fegert and Boonmann2019) did not find sex differences in the relation between childhood maltreatment and TL, few studies included in their meta-analysis actually examined sex differences directly.
On the other hand, studies that included individuals with medical illnesses showed smaller effect sizes than those with healthy individuals, suggestive of a potential moderating effect of physical health status (Ridout et al., Reference Ridout, Levandowski, Ridout, Gantz, Goonan, Palermo and Tyrka2018). Given the poorer prognosis related to shorter LTL in individuals with chronic illnesses (Farzaneh-Far et al., Reference Farzaneh-Far, Cawthon, Na, Browner, Schiller and Whooley2008; Opstad, Kalstad, Pettersen, Arnesen, & Seljeflot, Reference Opstad, Kalstad, Pettersen, Arnesen and Seljeflot2019; Tian et al., Reference Tian, Zhang, Zhang, Pang, Chen, Shen and Zhang2017; Willeit et al., Reference Willeit, Willeit, Brandstätter, Ehrlenbach, Mayr, Gasperi and Kiechl2010), and the association between child maltreatment and the development of chronic illnesses (Hughes et al., Reference Hughes, Bellis, Hardcastle, Sethi, Butchart, Mikton and Dunne2017; Maschi, Baer, Morrissey, & Moreno, Reference Maschi, Baer, Morrissey and Moreno2013), examination of the impact of health status, especially CAD, on the relation between childhood maltreatment and LTL appears warranted. Indeed, there are high prevalence, morbidity, and mortality rates of CAD compared to most other chronic illnesses (Statistics Canada, 2020), as well as differences in psychological distress in CAD compared to non-CAD individuals (e.g. Vaillancourt, Busseuil, and D'Antono, Reference Vaillancourt, Busseuil and D'Antono2021). Thus, the relation between childhood maltreatment and LTL may very well differ as a function of presence of CAD v. other, less fatal chronic illnesses.
The relation between childhood maltreatment and LTL was also weaker in studies whose samples were composed of individuals suffering from psychiatric conditions compared to studies whose samples were non-clinical (Bürgin et al., Reference Bürgin, O'Donovan, d'Huart, di Gallo, Eckert, Fegert and Boonmann2019; Ridout et al., Reference Ridout, Levandowski, Ridout, Gantz, Goonan, Palermo and Tyrka2018). To our knowledge, the only two studies that investigated whether depression influenced the relation between childhood maltreatment and LTL found no evidence of moderation in 92 young Colombians (Jiménez, Pereira-Morales, Adan, & Forero, Reference Jiménez, Pereira-Morales, Adan and Forero2019) nor in 180 middle-aged Ukrainians (Vincent et al., Reference Vincent, Hovatta, Frissa, Goodwin, Hotopf, Hatch and Powell2017). Data concerning moderation by other markers of psychological distress, such as anxiety or stress is lacking.
While psychological distress may relate to the individuals' recent life situations, it may also reflect emotional dysregulation resulting from childhood maltreatment experiences (Bowlby, Reference Bowlby1979; Neacsiu, Bohus, & Linehan, Reference Neacsiu, Bohus, Linehan and Gross2014; Porges, Reference Porges2009; Thayer & Lane, Reference Thayer and Lane2000), and associated symptoms of complex post-traumatic stress disorder (Herman, Reference Herman1992). These children may develop emotional reactivity and difficulties implementing strategies to regulate it (i.e. emotional dysregulation; Gross and Jazaieri, Reference Gross and Jazaieri2014), including heightened fight-flight or freeze responses and behaviors (e.g. substance use, overeating, social withdrawal) that may persist over a lifetime, and increase their risk for premature aging and illness (Ellis, Bianchi, Griskevicius, & Frankenhuis, Reference Ellis, Bianchi, Griskevicius and Frankenhuis2017; McEwen & Stellar, Reference McEwen and Stellar1993; Porges, Reference Porges2009; Thayer & Lane, Reference Thayer and Lane2000). While shorter TL have been observed in individuals with symptoms or disorders of depression (Ridout, Ridout, Price, Sen, & Tyrka, Reference Ridout, Ridout, Price, Sen and Tyrka2016), anxiety (Malouff & Schutte, Reference Malouff and Schutte2017), and stress (Mathur et al., Reference Mathur, Epel, Kind, Desai, Parks, Sandler and Khazeni2016), research on whether psychological distress mediates the childhood maltreatment-LTL relation is extremely limited. One study found that the self-control aspect of self-regulation (the ability to manage emotions in order to conduct behavior) partially mediated the relationship between adverse childhood adversity and TL in a sample of 2527 children with unwed parents, though the effect was very small (b = −0.002) (Sosnowski et al., Reference Sosnowski, Kliewer, Valrie, Winter, Serpell and Amstadter2020). In another large investigation in middle-aged men, 16% of the effect of stressful childhood events on LTL was mediated by depressive mood (Osler, Bendix, Rask, & Rod, Reference Osler, Bendix, Rask and Rod2016).
Thus, evidence suggests that childhood maltreatment is associated with shorter LTL, and preliminary data suggests that characteristics of individuals may moderate or mediate this relation. The objectives of the study are therefore to replicate and extend findings regarding childhood maltreatment and LTL in older men and women with CAD or other chronic non-cardiovascular disease (CVD) and determine whether sex and CAD status moderate the observed relations. Finally, we seek to determine whether correlates of emotional dysregulation (anxiety, depression, stress) moderate or cross-sectionally mediate this relationship. It is hypothesized that self-reported childhood maltreatment will be associated with shorter LTL. Given the paucity of information on potential moderation or mediation of this relation, no directional hypotheses are presented at this time.
Methods
This study is part of an ongoing prospective research project (BEL-AGE) examining the role of psychological burden in pathological aging. This study was approved by the Research Ethics and New Technology Development Committee of the Montreal Heart Institute (#11-1313).
Participants
A total of 1325 (40% women) participants with CAD or other non-CVD illnesses (mean age = 65.4 ± 7.2 years) were recruited from the André and France Desmarais Hospital Cohort of the Montreal Heart Institute (the MHI cohort). Participants were eligible for the BEL-AGE study if they met the following criteria when they first enrolled into the MHI cohort: they were (a) aged between 30 and 70 years old, (b) living in the greater Montreal area, (c) spoke English or French, (d) had no previous or current diagnosis of serious psychological disorders (e.g. bipolar disorder, schizophrenia, dementia) or major cognitive impairment that could hinder the participants understanding of any aspect of the study, (e) had no previous or current diagnosis of major life-threatening diseases other than CAD (e.g. AIDS, cancer, amyotrophic lateral sclerosis, Creutzfelt–Jakob disease), (f) not pregnant or breastfeeding (g) or if none of their family members or spouse had previously participated in BEL-AGE, or were scheduled to participate. Depression, anxiety, and endocrine disorders were not excluded. Skin cancer was not excluded because of its high prevalence and benign course with an early diagnosis. CAD was defined as having previously suffered at least one myocardial infarction, coronary artery bypass, coronary angioplasty, or stenosis of more than 50% on an angiography. An absence of CAD, angina, arrhythmia, congenital heart disease, heart failure, cardiomyopathy, and stroke was required for non-CVD status. Those who were considered healthy had no notable or chronic health issue. Medical history was self-reported and verified with the participants' medical file in the case of CAD. A total of seventy-eight participants were excluded from the current study analyses either because they did not meet the inclusion criteria (n = 22), had missing LTL or psychological data (n = 26), or were suffering from no health conditions (n = 30), resulting in a final sample of 1247 individuals.
Procedure
Potential participants were contacted by telephone and screened for eligibility. Appointments were scheduled on weekdays between 8:00 and 10:00 a.m. in order to control for circadian rhythms. Except for water and prescribed medication, participants were asked to refrain from eating and drinking 12 h before their appointment and using illicit substances or alcohol 24 h before their appointment. Participants who did not adhere to these requirements or presented with colds or other acute symptoms were rescheduled. After written consent was provided, anthropomorphic data (weight, height, and waist circumference) were measured, and a 35 mL sample of blood was obtained. Participants were questioned regarding their demographic and medical information, underwent a cognitive screen for other purposes, and completed several questionnaires. Only their travel/parking costs were reimbursed.
Measures
Sociodemographic and health variables
Data on sex, age, ethnicity, years of schooling, civil status, personal and family income, personal and family medical history were obtained, as was information on the participant's health behaviors (tobacco use, alcohol, caffeine consumption, diet, and physical activity). The participants provided a current list of medications.
The Childhood Trauma Questionnaire (CTQ; Bernstein et al., Reference Bernstein, Fink, Handelsman, Foote, Lovejoy, Wenzel and Ruggiero1994; Paquette, Laporte, Bigras, and Zoccolillo, Reference Paquette, Laporte, Bigras and Zoccolillo2004) was used to assess childhood and adolescent maltreatment. This questionnaire is comprised of 25 items with 5 subscales including: physical, sexual, and emotional abuse, as well as physical and emotional neglect. Each item uses a five-point scale ranging from ‘never true’ to ‘very often true’. The sum of scores is calculated and participants are also categorized as having experienced childhood maltreatment when they have a score within the moderate-to-severe range on at least one subscale: emotional abuse (≥13), physical abuse (≥10), sexual abuse (≥8), emotional neglect (≥15), and physical neglect (≥10). This questionnaire has high internal consistency (α = 0.89) and good test-retest reliability over a 4-month period in both English and French versions (Bernstein & Fink, Reference Bernstein and Fink1998; Paquette et al., Reference Paquette, Laporte, Bigras and Zoccolillo2004). In the current sample, overall internal consistency was excellent (α = 0.91) and ranged from 0.82 (emotional abuse, physical abuse, sexual abuse, and emotional neglect) to 0.58 for the physical neglect subscale, consistent with previous studies (Dekker et al., Reference Dekker, Girman, Rhodes, Nijpels, Stehouwer, Bouter and Heine2005; Ekelund et al., Reference Ekelund, Brage, Froberg, Harro, Anderssen, Sardinha and Andersen2006).
The Center of Epidemiological Studies Depression Scale Revised (CES-D-R; Van Dam and Earleywine, Reference Van Dam and Earleywine2011), measures the presence of depressive symptoms over the past 2 weeks on a likert scale from ‘not at all or less than one day’ to ‘nearly every day for 2 weeks’. Consisting of 20 items, total scores range from 0 to 60, with scores higher than or equal to 16 indicating greater risk of clinical depression. This questionnaire has excellent internal consistency (α > 0.85) for both English (Van Dam & Earleywine, Reference Van Dam and Earleywine2011) and French versions (Caron, Reference Caron2001). Test-retest reliability over 3–6 months is 0.54–0.59 for the English (Radloff, Reference Radloff1977; Roberts, Reference Roberts1980) and French (Caron, Reference Caron2001) versions, respectively.
State-Trait Anxiety Inventory-state version. (STAI; Spielberger, Gorsuch, Lushene, Vagg, and Jacobs, Reference Spielberger, Gorsuch, Lushene, Vagg and Jacobs1983) is a 20-item questionnaire assessing participant's current state of anxiety. Participants responded to each item with a 4-point Likert scale ranging from ‘not at all’ to ‘very much’. Total scores range from 20–80, while a score of 40 or greater is considered clinically elevated. This questionnaire has excellent internal consistency (average α = 0.89) and a test-retest reliability over a period of 3.5 months of 0.75 (Spielberger, Gonzalez-Reigosa, Martinez-Urrutia, Natalicio, & Natalicio, Reference Spielberger, Gonzalez-Reigosa, Martinez-Urrutia, Natalicio and Natalicio1971; Spielberger et al., Reference Spielberger, Gorsuch, Lushene, Vagg and Jacobs1983). Similar values were found for the French-Canadian version (Gauthier & Bouchard, Reference Gauthier and Bouchard1993).
The Perceived Stress Questionnaire (PSQ; Levenstein et al., Reference Levenstein, Prantera, Varvo, Scribano, Berto, Luzi and Andreoli1993). Consisting of 30 items, this questionnaire, specifically designed for psychosomatic research, covers seven dimensions including: harassment, overload, irritability, lack of joy, fatigue, worries, and tension. It is used to measure perceived stressful life events, circumstances, and reactions to stress. Participants rate how frequently they experienced each of the stress-related statements over the past 2 years on a four-point scale ranging from ‘almost never’ to ‘usually’. It is scored using the sum of the individual items, as well as the PSQ index (raw score-30)/90), which represents the severity of stress of each participant, varying from 0 (lowest level of stress) to 1 (highest level of stress). A score of 0.45–0.60 on the PSQ index indicates a moderate stress level, while scores >0.60 reflect high stress. This questionnaire has excellent internal consistency for the English version (α = 0.90), and a test-retest reliability over a period of 6 months of 0.82 (Lehman, Burns, Gagen, & Mohr, Reference Lehman, Burns, Gagen and Mohr2012; Levenstein et al., Reference Levenstein, Prantera, Varvo, Scribano, Berto, Luzi and Andreoli1993; Montero-Marin, Piva Demarzo, Pereira, Olea, & García-Campayo, Reference Montero-Marin, Piva Demarzo, Pereira, Olea and García-Campayo2014), with similar values for the French version (Consoli, Taine, Szabason, Lacour, & Metra, Reference Consoli, Taine, Szabason, Lacour and Metra1997).
We have shown in previous research that the psychometric properties for each of these psychological variables within our sample are consistent with the existing literature (Vaillancourt et al., Reference Vaillancourt, Busseuil and D'Antono2021).
Leukocyte telomere length
Quantitative polymerase chain reaction (qPCR) was used to measure TL, using the modified method (Epel et al., Reference Epel, Blackburn, Lin, Dhabhar, Adler, Morrow and Cawthon2004) of the protocol described by Cawthon (Reference Cawthon2002) as previously reported in Starnino et al. (Reference Starnino, Dupuis, Busque, Bourgoin, Dubé, Busseuil and D'Antono2021). This method is based on determining the number of telomere DNA sequences (T) on the number of copies of a control gene (RPLP0, 60S acidic ribosomal protein P0) (S). Briefly, DNA was extracted from blood leukocytes using BioRobot M48 system (Qiagen, Germany). Serial dilutions of a reference DNA from a single normal individual were made to establish a standard curve (mean efficiency for T = 120% and S = 102%). The T/S was normalized to 1 and used to calculate the T/S ratio for each sample of DNA tested. The relative ratio of a sample thus represents the number of copies of telomeres relative to the reference sample. All qPCR was performed on the ViiA™7 real-time PCR system (Applied Biosystems, Massachusetts, USA) and with the QuantStudio analysis software (Applied Biosystems, Massachusetts, USA). All samples were measured in triplicate, and their mean was used for analyses. If the standard deviation of cycle threshold (Ct s.d.) was over 0.3, the outlier was omitted, and duplicate was considered. Otherwise, the sample was repeated.
Statistical analyses
Preliminary analyses
Data analysis was performed using IBM SPSS, Version 26. Moderation and mediation analyses were performed using the computational tool PROCESS version 3.5 (Hayes, Reference Hayes2020). Log-transformations were applied to depressive symptoms and anxiety data, while square root transformations were applied to childhood maltreatment data to normalize their distributions.
Covariates, including age, hours of exercise per week, civil status, years of education, smoker status, metabolic triad (diagnosis of hypercholesterolemia, hypertriglyceridemia, and /or essential hypertension), body mass index (BMI), and family income, were chosen based on the existing literature on childhood maltreatment and LTL (Bürgin et al., Reference Bürgin, O'Donovan, d'Huart, di Gallo, Eckert, Fegert and Boonmann2019; Ridout et al., Reference Ridout, Levandowski, Ridout, Gantz, Goonan, Palermo and Tyrka2018).
Main analyses
Bivariate correlations were conducted to examine the relationship between childhood maltreatment and LTL. To examine the association between childhood maltreatment and LTL controlling for covariates, a hierarchical linear regression was conducted with the covariates as well as sex and CAD-status variables in block 1, and childhood maltreatment in block 2.
Moderation analyses were performed to verify whether sex and/or CAD-status moderated the association between childhood maltreatment and LTL, controlling for other covariates. Similar analyses were conducted to verify whether depression, anxiety, and stress independently moderated the relation between childhood maltreatment and LTL. The 2- and 3-way interactions of childhood maltreatment with sex, CAD status, and/or psychological variables in moderation analyses were constructed from centered variables. If the interactions terms were not significant, they were removed from the equation and the more parsimonious model(s) were retained. Significant interactions were explored via simple slope analyses.
A parallel mediation analysis was conducted to investigate whether depression, anxiety, and stress (entered simultaneously) mediated the relationship between childhood maltreatment and LTL, controlling for covariates, including sex and CAD status. The indirect effects were tested using non-parametric bootstrapping, with effects considered statistically significant when 0 falls outside the 95% confidence intervals. Bootstrapping was performed using 5000 samples.
A two-sided p value < 0.05 was considered statistically significant for main effects, and interactions meeting a p value < 0.10 were examined to avoid type II errors (Winer, Reference Winer1962). Main analyses were conducted using continuous measures of CTQ, depression, anxiety, and stress.
Results
Demographic characteristics of men and women, and individuals with and without CAD are presented in Table 1. While childhood maltreatment did not differ as a function of sex nor CAD status, women had significantly longer LTL compared to men, and non-CVD individuals had significantly longer LTL than those with CAD. Furthermore, women reported significantly more depression, anxiety or stress compared to men.
Table 1. Participant characteristics (Mean ± s.d.)
Note: CVD, cardiovascular disease.
Childhood maltreatment and LTL are associated
Childhood maltreatment correlated with significantly shorter LTL (r = −0.059, p = 0.038). Table 2 presents the detailed results of the hierarchical linear regression controlling for covariates, in which the relationship between childhood maltreatment and LTL remained significant.
Table 2. Hierarchical regression analysis details for TL
b, estimated value of unstandardized regression coefficient; t, student's t distribution; LL, lower limit; UL, upper limit.
Note: Metabolic triad refers to the presence of hypercholesterolemia, hypertriglyceridemia, and/or essential hypertension.
The relation between childhood maltreatment and LTL is not moderated by examined variables
The relation between childhood maltreatment and LTL was not moderated by sex (p = 0.896), CAD status (p = 0.971), nor by depression (p = 0.470), anxiety (p = 0.448), and stress (p = 0.673).
Chronic stress mediates the relation between childhood maltreatment and LTL
Results of the parallel mediation analysis revealed that while childhood maltreatment predicted significantly higher levels of depression, anxiety, and stress (all ps < 0.001), only stress partially and significantly mediated the relationship between childhood maltreatment and LTL [Indirect effect – IEPSQ = −0.0041, s.e. = 0.002, 95% CI (−0.0085 to −0.0002)], controlling for the specified covariates. The total (b = −0.016, p = 0.006) and direct (b = −0.012, p = 0.041) effects of childhood maltreatment on LTL were both negative and statistically significant. Results are depicted in Fig. 1.
Fig. 1. Mediation of the relation between childhood maltreatment and TL. A parallel mediation analysis revealed a significant indirect effect of childhood maltreatment on TL through stress perception [b = −0.0041, 95% CI (−0.0085 to −0.0002)].
Post-hoc analyses
A three-way (Sex by CAD-Status by Childhood Maltreatment) ANCOVA was conducted with LTL as dependent variable and childhood maltreatment as a dichotomous variable (based on CTQ cut-off) to examine whether comparing more severe to no/less severe maltreatment altered results. Significant main effects of childhood maltreatment [F(1,1221) = 10.10, p = 0.002] and sex [F(1,1221) = 13.09, p < 0.001], as well as a 3-way CAD-status by Sex by Childhood maltreatment interaction [F(1,1221) = 4.425, p = 0.036] were observed. More specifically, childhood maltreatment was associated with significantly shorter LTL in women with CAD (p = 0.041) and in non-CVD men (p = 0.016), with a trend towards the same in non-CVD women (p = 0.272). In men with CAD, there was no relation (p = 0.954) (see Fig. 2).
Fig. 2. Group differences in relative telomere length as a function of sex, CAD status, and childhood maltreatment. Childhood maltreatment was associated with significantly shorter TL in women with CAD (p = 0.041) and non-CVD men (p = 0.016), and a trend towards the same in non-CVD women (p = 0.272). In men with CAD, there was no relation (p = 0.954). Significant differences between those exposed to childhood maltreatment and those not exposed are identified with asterisks within the figure.
Bivariate correlations between LTL and the number of maltreatment subtypes (number of CTQ subscales for which each person surpassed the cut-off) were performed to examine whether exposure to more types of childhood maltreatment was associated with shorter LTL. Number of maltreatment subtypes was associated with significantly shorter LTL (r = −0.059, p = 0.038).
Discussion
The current project evaluated the association between childhood maltreatment and LTL among older men and women with CAD or other non-CVD illness, and whether individual characteristics (sex, CAD status, and correlates of emotional dysregulation) of participants moderated and/or mediated this relation. Findings confirmed that greater maltreatment in childhood was associated with significantly shorter LTL. None of the individual characteristics examined moderated this relation when childhood maltreatment was examined as a continuous variable. However, when childhood maltreatment was dichotomized to capture more severe maltreatment, men with CAD did not show the expected association. Perceived chronic stress partially and significantly mediated the relation between maltreatment experiences and later LTL. Findings were independent of sociodemographic, medical, and lifestyle variables.
That individuals with childhood maltreatment had shorter LTLs in the current study is consistent with the results of previous meta-analyses (Bürgin et al., Reference Bürgin, O'Donovan, d'Huart, di Gallo, Eckert, Fegert and Boonmann2019; Li et al., Reference Li, He, Wang, Tang and Chen2017; Ridout et al., Reference Ridout, Levandowski, Ridout, Gantz, Goonan, Palermo and Tyrka2018). In contrast to Li et al. (Reference Li, He, Wang, Tang and Chen2017), however, our finding remained significant after adjusting for age (and other covariates with the potential to impact on LTL). While robust, the effect accounted for only 0.6% of the variance in LTL. This may seem small, but the association was actually larger than the other variables included, with the exception of age and sex. Furthermore, childhood maltreatment was associated with up to 5 years greater biological age compared to those who were not exposed to childhood maltreatment (see online Supplementary Table S1 for the calculation of biological years difference from T/S ratio). Nonetheless, overall, only 12.6% of variance in LTL was explained by the model, indicating need for continued research.
The few studies that examined sex differences in the relation between childhood maltreatment and TL reported similar effects of maltreatment in men and women (Bürgin et al., Reference Bürgin, O'Donovan, d'Huart, di Gallo, Eckert, Fegert and Boonmann2019). This was consistent with our own results using a continuous measure of childhood maltreatment. Thus, while boys and girls differ in the types of maltreatment they are more likely to be exposed to (Burczycka & Conroy, Reference Burczycka2017), it may be the experience of any severe maltreatment that contributes to premature aging, though exposure to a greater number of maltreatment types may be more damageable as suggested by this and other research (Mayer et al., Reference Mayer, Prather, Puterman, Lin, Arenander, Coccia and Epel2019; Ridout et al., Reference Ridout, Ridout, Kao, Carpenter, Philip, Tyrka and Price2015; Surtees et al., Reference Surtees, Wainwright, Pooley, Luben, Khaw, Easton and Dunning2012; Theall, Brett, Shirtcliff, Dunn, & Drury, Reference Theall, Brett, Shirtcliff, Dunn and Drury2013).
According to one meta-analysis (Bürgin et al., Reference Bürgin, O'Donovan, d'Huart, di Gallo, Eckert, Fegert and Boonmann2019), childhood trauma was more weakly associated with TL in investigations that included individuals suffering from medical conditions suggesting that health status may moderate the relation. To our knowledge, only one prior study specifically examined this issue (Malan-Müller et al., Reference Malan-Müller, Hemmings, Spies, Kidd, Fennema-Notestine and Seedat2013). Although they found that individuals with HIV had significantly shorter LTL than those without, HIV status did not moderate the association between maltreatment and LTL. This is consistent with our own results concerning CAD status. It is noteworthy that the post hoc analyses using a categorical approach to measure childhood maltreatment did find health status to influence the association as a function of participants' sex. Indeed, it was particularly among the women with CAD and men (and to a lesser degree, women) with non-cardiovascular chronic disease that childhood maltreatment was found to be associated with the greatest telomere attrition. Why men with CAD did not show the same pattern is unclear. One possible explanation may relate to survivor bias. Results from a meta-analysis suggest that psychological distress (that is greater in survivors of childhood maltreatment) may be associated with a slightly but significantly poorer prognosis in men with CAD as compared to women (Smaardijk, Maas, Lodder, Kop, & Mommersteeg, Reference Smaardijk, Maas, Lodder, Kop and Mommersteeg2020). If this is the case, it may be that those men with CAD who were able to participate in the study were more resilient for some reason to the effects of adversity, in contrast to others who may have already passed away or been too ill to participate, in effect concealing the true impact of maltreatment among these men. In addition, shorter telomeres have been observed in individuals with CAD compared to individuals without CAD in this and other research (Xu et al., Reference Xu, Hu, Lin, Huang, Ji, Li and Duan2019), and though limited, some data suggest that this difference may be particularly evident among men (Xu et al., Reference Xu, Hu, Lin, Huang, Ji, Li and Duan2019). Though additional research examining sex differences in the association between CAD and LTL is needed, it is plausible that the shorter (and less variable) LTL in men with CAD in the current study compared to the other groups may have precluded additional effects of childhood maltreatment, in essence creating a floor effect on LTL in men with CAD. Given the paucity of data on the potential moderating role of health status and sex, further research is warranted to replicate our findings.
The relation between childhood maltreatment and LTL was not found to be moderated by the various markers of emotional dysregulation (anxiety, depression, stress), but perceived stress over the past two years was found to partially mediate this relation. This is consistent with various theories that have proposed that the experience of childhood maltreatment may lead to a selective search for threat, difficulties perceiving social and environmental cues of safety, as well as overactive fight-flight-or freeze physiological and behavioral responses to these threats (Porges, Reference Porges2009; Thayer & Lane, Reference Thayer and Lane2000). In essence, a psychobiological context is created in which victims of childhood maltreatment are chronically perceiving, experiencing, and perhaps contributing to stress. According to the Hidden Talent Theory (Ellis et al., Reference Ellis, Bianchi, Griskevicius and Frankenhuis2017), these alterations were likely adaptive for a child exposed to constant threat because they helped promote immediate survival (e.g. social withdrawal in the context of abuse may help prevent unwanted attention) and temporarily reduce acute distress, but may hold long-term health costs when maintained. In the current sample, significantly greater symptoms of anxiety, depression, and chronic stress were reported by participants who experienced maltreatment in childhood indicating that such emotional dysregulation may persist well into older age. The precise mechanisms by which chronic stress or other markers of emotional dysregulation mediates the impact of childhood maltreatment on LTL have not been delineated, though poor health behaviors (in attempts to self-regulate negative mood; Khantzian, Reference Khantzian1997) and repeated solicitation of fight-flight-freeze responses and associated autonomic, metabolic, endocrine, and immune changes have been proposed to be involved (Porges, Reference Porges2009; Thayer & Lane, Reference Thayer and Lane2000). More specifically, it has been suggested that these recurrent physiological changes in response to perceived threat contribute to greater allostatic load (McEwen & Stellar, Reference McEwen and Stellar1993) and dysregulation of mitochondrial structure and function, and as a result altered TL regulation (Juster, Russell, Almeida, & Picard, Reference Juster, Russell, Almeida and Picard2016). Further research is planned to test whether telomere attrition in survivors of childhood maltreatment is explained by increased allostatic load.
Several limitations of this study bear mention. The measure of childhood maltreatment was retrospective and self-reported and could have been influenced by memory bias (Baldwin & Danese, Reference Baldwin and Danese2019). The cross-sectional nature of the study does not allow for causal inferences to be made. While prospective research is obviously required for mediation analyses, the current results nonetheless contribute to our understanding of the factors that may contribute to shorter LTL following childhood maltreatment. Results may not be generalizable to other ethnicities considering the majority of participants were French-speaking Caucasians and race/ethnicity differences in TL have previously been reported in the literature (Brown, Needham, & Ailshire, Reference Brown, Needham and Ailshire2017; Ly et al., Reference Ly, Walker, Berry, Snell, Marks, Thayer and Morton2019). Although not examined within the current study, it is possible that differences in the chronicity and age at onset of maltreatment, or genetic vulnerability may have confounded the results (White & Kaffman, Reference White and Kaffman2019).
The strengths include the use of validated questionnaires, a large sample size with a representative sample of men and women by which sex differences could be examined and validated qPCR LTL measurement methods that have been validated by our group in several studies (e.g. Buscarlet et al., Reference Buscarlet, Provost, Zada, Barhdadi, Bourgoin, Lépine and Busque2017; Mollica et al., Reference Mollica, Fleury, Belisle, Provost, Roy and Busque2009; Starnino, Busque, Tardif, & D'Antono, Reference Starnino, Busque, Tardif and D'Antono2016; Starnino et al., Reference Starnino, Dupuis, Busque, Bourgoin, Dubé, Busseuil and D'Antono2021). To our knowledge, this is the first study to examine the potential moderating and mediating role of psychological distress in the relation between childhood maltreatment and LTL relation among older individuals with chronic illnesses. The examination of three indices of emotional dysregulation is clinically useful, as each index has been shown to reflect different types of maladaptive emotional regulation strategies. For example, suppression may be more frequently used in the case of anxiety (Amstadter, Reference Amstadter2008) as compared to rumination in the case of depression (Joormann & Gotlib, Reference Joormann and Gotlib2010).
In conclusion, childhood maltreatment predicted accelerated biological aging, as measured by LTL independently of patient characteristics among older individuals, though this effect was less obvious in men with CAD. Chronic stress resulting from childhood maltreatment may contribute to this process. Whether stress management interventions in survivors of childhood maltreatment could slow or lower their risk for premature ageing requires investigation. Such treatments have been shown to alter telomere regulation processes in a variety of samples (Schutte & Malouff, Reference Schutte and Malouff2014), and to reduce risk for clinical events in individuals with established cardiovascular disease (Blumenthal et al., Reference Blumenthal, Sherwood, Smith, Watkins, Mabe, Kraus and Hinderliter2016).
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0033291722003543.
Acknowledgements
Special thanks to students and staff of the Heart and Mind Laboratory, participants and staff of the André and France Desmarais Hospital Cohort of the Montreal Heart Institute, as well as the Canadian Institutes of Health Research (CIHR) and the Montreal Heart Institute Foundation.
Author contributions
A. C. wrote the original drafts of the manuscript. A. C. and B. D'A. analyzed and interpreted the data. L. S. contributed to data collection. B. D'A. designed, obtained funding for, and led the BEL-AGE study. J.-C. T., chair of the MHI cohort's management committee, contributed to the BEL-AGE study design, as well as funding and conduct of the MHI cohort. M.-P. D. and L. B. were involved in the BEL-AGE study design particularly as relates to telomeres. D. B. was responsible for selection of participants, management of the collection of samples and part of the data for the MHI cohort. V. B. and L. B. were responsible for TL extraction from blood samples. All authors revised and approved the final manuscript.
Financial support
This research was supported by grants awarded to Bianca D'Antono by the Canadian Institutes of Health Research (CIHR; MOP # 111015) and the Montreal Heart Institute Foundation.
Conflict of interest
None.
Ethical standards
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.
Open access
