Individuals affected by severe mental disorders (SMD), such as bipolar and psychotic disorders, have a significantly shorter lifespan when compared to unaffected peers, equal to approximately 15–20 years¹. In addition to mortality rate, individuals with SMD tend to be disproportionately affected by somatic diseases associated with advanced chronological age, such as metabolic syndromes, cardiovascular disease, and dementia^2,3. One proposed explanation for these differences is that individuals with SMD may possess a faster rate of biological ageing⁴, as assessed through measurements of telomere length (TL).

Telomeres are non-coding endpoints of human chromosomes, comprised of repeating TTAGGG nucleotides⁵. The main function of telomeres is to protect chromosomal ends during DNA replication⁶. Telomeres shorten through successive cell divisions and from cellular and environmental stressors. Telomeres length can be extended by telomerase; however, this enzyme is inactive in most somatic cells in the human body⁷. Once a critical length is reached, cells will enter a state of replicative senescence or undergo apoptosis⁸, slowing tissue and organ renewal⁹. This is thought to be one of the main processes underlying ageing¹⁰.

Telomere attrition affects all humans; however, epidemiological studies of TL have noted differences between a number of groups, focusing on gender¹¹, ethnicity¹², and more recently, individuals with SMD⁴. While variability exists, studies in psychiatric populations have noted that individuals with SMD tend to possess shorter telomeres compared to healthy controls^13,14. One mechanism which has been proposed is that individuals with SMD have less healthy lifestyle behaviours compared to unaffected controls¹⁵.

Reviews of SMD patient lifestyles have consistently noted an increased number of unhealthy behaviours such as poor diet, low physical activity, smoking and substance use¹⁵. This has been argued to be partly a result of symptom interference¹⁶, and medication side effects such as lethargy and increased appetite.^17,18 These are often compounded by adverse childhood experiences,¹⁹ which are quite prevalent in SMD. Analyses of these health domains have supported their influence on TL shortening. Particularly, substance use has been found to significantly reduce TL²⁰, across a wide range of substances. This may be due to the increase in oxidative stress and inflammation caused by such substances^21,22. In addition, evidence from the general population indicates reduced TL in individuals reporting poorer diet²³, sedentary behaviour²⁴, and smoking²⁵. These are all behaviours associated with oxidative stress and inflammation^26,27, to which TL is sensitive²⁸. Interestingly, coffee consumption seems to be associated with better telomere maintenance²⁹. This has been speculated to potentially be the result of coffee’s antioxidant properties³⁰.

However, despite the seeming association between individual health factors and TL, few studies examine the association between TL and ‘global’ lifestyle across several different facets (diet, exercise, substance use, etc.). One study³¹ found longer telomeres in middle-aged women who maintained healthier lives at one-year follow-up. Yet, despite their findings, a notable gap in the literature exists due to such studies being carried out in the general population. Thus, it is difficult to generalise these findings to the SMD population, where biological ageing might already be accelerated.

The main aim of the current study is to provide a comprehensive evaluation of how health behaviours may interact with TL in individuals with SMD. To the best of our knowledge, there are no previous studies investigating the cumulative effect of lifestyle/health behaviours on TL in this specific population. Based on previous literature, we anticipate that 1) TL will decrease with increasing unhealthy behaviours, 2) individuals with the healthiest lifestyles will possess the longest telomeres.

The health behaviours analysed in the study comprised diet, exercise, alcohol consumption, smoking, substance use, and coffee consumption. To determine alcohol consumption the Alcohol Use Disorder Identification Test (AUDIT)³³ was used. Following scoring guidelines, sum scores ranging from 1–7 indicate low-risk alcohol consumption (‘healthy’ group), and scores above 8 indicate harmful alcohol use (‘unhealthy’ group).

Self-report on illicit substance use was provided by participants and dichotomised into those with any lifetime substance use and those without any substance use. The substances included within the assessment were: cannabis, cocaine, heroin, hallucinogens, amphetamines, opioids, and PCP. Information on daily smoking was provided by participants and dichotomised into daily smoking and non-smoking groups.

Self-report information regarding exercise duration was provided by participants and split into ‘unhealthy’ (less than 150 min per week) and ‘healthy’ (more than 150 min per week) groups based on the World Health Organisation (WHO) recommendation of a minimum weekly exercise duration of at least 150 min³⁴.

Assessment of diet was based on interviews with patients of diet patterns during the last six months and grouped into healthy to poor diet following WHO guidelines of healthy diet³⁵. Due to previous findings associating coffee consumption with longer telomeres in the sample³⁶, we included information on coffee consumption. The data was provided as a self-report variable, which was dichotomised into coffee drinkers and non-drinkers.

Subsequently, participants were scored based on the number of unhealthy behaviours they exhibited (min score 0 and max score 6), to produce a lifestyle variable. The results were dichotomised into a ‘healthy' (participants reporting 0 unhealthy behaviours; n = 37) vs ‘unhealthy' (participants reporting 1 or more unhealthy behaviours; n = 373) lifestyle. In addition, a dose lifestyle variable ranging from 0 to 5 was also formed. The latter was done as only one participant (from the affective disorders group) possessed 6 unhealthy behaviours. In effect, groups 6 and 5 were collapsed into group 5. For an overview of N for each respective group, please see Supplementary Material Table S1.

In addition to health behaviours, the study also included data on childhood trauma exposure, evaluated using the childhood trauma questionnaire (CTQ)³⁷. For more information about trauma operationalisation, see Aas et al., 2019³⁸. Lastly, patient records were used to calculate the daily defined dose (DDD) of psychotropic medication.

Telomere length was assessed in all participants using peripheral blood leukocytes, through a quantitative real-time polymerase chain reaction (qPCR) method^38,39. The qPCR involved the extraction of 10 ng of DNA from leukocyte cells, which was combined with 5 µl of SYBR®Green JumpStart Taq Ready Mix and 0.25 µl of ROX reference dye. All assessment were carried out using a 384-well plate Applied Bio- systems 7900HT Fast Real Time qPCR. All inconclusive or extreme samples (top and bottom 5%) were re-evaluated. For further details on the methods, see Mlakar et al., 2024³⁹.

The qPCR provided researchers with a telomere to single copy ratio (T/S ratio), which was used to estimate mean TL. Smaller T/S ratios indicated shorter mean TL. Analyses of variation revealed that the sample possessed an intra-assay coefficient of 6.07% and an inter-assay coefficient of 6.08%. All blood samples were stored in The Biobank, located in Oslo, Norway.

In addition to the T/S ratio provided by the qPCR analysis, we estimated differences in base-pair attrition. This was carried out using a previously proposed quantitative estimate of years of accelerated aging, with an average of 70 base-pair reduction per year⁴⁰. In order to estimate TL attrition, we subtracted the base-pair differences between different groups (e.g.: mean TL in healthy lifestyle - mean TL in unhealthy lifestyle) and divided the difference by 70. The end result provides a rough estimate of TL attrition.

Demographic overview

Table 1 provides the descriptive statistics for the study sample (n = 410). In summary, the schizophrenia group was significantly younger and had a higher proportion of men than the affective disorder group. There was no significant difference in ethnicity between the diagnostic groups, with both comprising primarily of individuals of European ancestry.

With respect to medication, data were available for 213 participants. Of these, in the schizophrenia sample, 83.1% were taking antipsychotic medication, 2.3% lithium, 26.8% antidepressant medication, and 13.6% other mood stabilisers. In the affective disorder group, 59% were taking antipsychotic medication, 22.6% lithium, 28.9% antidepressant medication, and 52.4% other mood stabilisers. The affective disorder group had a trend for longer telomeres (F = 2.97, p = 0.09). There was no statistical difference concerning childhood trauma exposure between groups (p > 0.1).

In this study of individuals with SMDs, increasingly unhealthy behaviours were associated with shorter telomeres. Individuals who reported having an unhealthy lifestyle had significantly shorter telomeres compared to those with healthy lifestyles, equating to a biological age difference of roughly six years.

These results align with previous research proposing a protective effect of healthy lifestyle on TL in thee general population. In the study by Puterman and colleagues³¹, it was highlighted that a healthier lifestyle allowed for better stress management and telomere maintenance at a one-year follow-up, in a sample of women from the general population. Similarly, assessments of individual lifestyle domains (e.g.: diet, exercise, and substance use), have all noted their impact on telomere attrition^21,23,24. However, this is to our knowledge, the first study to investigate lifestyle and its association with TL in a psychiatric sample. In effect, our findings help extend existing knowledge by highlighting the importance of promoting healthy behaviours, in a population where accelerated biological ageing may already be occurring.

An unhealthy lifestyle may influence TL through several biological pathways. The most documented of these are oxidative stress and inflammation. As mentioned previously, several behaviours such as smoking⁴¹, sedentary lifestyle⁴², diets low in nutrients and high in sugars⁴², substance and alcohol use²¹, have been associated with increases in reactive oxygen species (ROS) generating oxidative stress, as well as inflammation^43,44. It has been noted that ROS have the ability to directly damage telomeres⁴⁵, which increased inflammation may exacerbate, through the up regulation of pro-inflammatory cytokines such as IL-6 and TNF-α, which themselves release further ROS⁴⁶. Although we have found limited evidence of an association between immune activation and TL in SMD in a recent cross-sectional study⁴⁷, the relationship between immune activation and TL needs to be further investigated in longitudinal studies.

As a result of the sustained telomeric strain generated by both oxidative stress and inflammation, cells may undergo replication and proliferation, in order to counteract cell and tissue damage, which in turn leads to further telomeric shortening and senescence¹⁰. Moreover, both processes are also associated with reductions in telomerase activity⁴⁸. This has been argued to be the result of direct oxidative damage to the Telomerase Reverse Transcriptase (TERT) protein by ROS⁴⁹, as well as inflammatory downregulation of TERT expression⁵⁰. Lastly, certain health behaviours have been related to methylation changes of TL relevant DNA regions. A study by Gao and colleagues⁵¹ found seven CpG sites, where smoking altered methylation profiles which were strongly associated with TL shortening in their sample. In effect, lifestyle may influence TL not only through direct molecular damage, but also through epigenetic downregulation of TL-relevant genes and biological mechanisms.

Furthermore, when investigating the impact of childhood trauma, our results indicated that only sexual abuse was significantly associated with unhealthy lifestyle behaviours, suggesting a distinct role of trauma subtype in shaping health behaviour. Prior studies examining the impact of sexual abuse have noted a poorer lifestyle compared to non-exposed peers⁵². The psychological impacts of sexual abuse are widespread, however, common patterns noted by survivors have been poor self-image, emotional dysregulation, feelings of lack of control, hypervigilance, etc.⁵³ These experiences can be contextualised in terms of later health behaviours and lifestyle. For instance, overeating may serve as a self-soothing mechanism⁵⁴, with restrictive eating providing survivors with a sense of control over their life⁵⁵. In addition, substance use has been linked to self-medication and poor self-esteem^56.57, dissociation⁵⁸ or hypervigilance⁵⁹. Overall, our findings highlight the importance of trauma-informed lifestyle interventions for survivors of childhood sexual abuse.

Study limitations

Although not uncommon in lifestyle assessments, most of the lifestyle factors utilised in the study relied on self-reports by the participants. Due to the nature of the information being provided (i.e. health), the participants could potentially have misrepresented or underreported their own habits. Moreover, whilst we speculate that TL was increased due to potential antioxidant properties of certain health behaviours, we did not have data on peripheral antioxidant levels.

In addition, the study is only comprised of a psychiatric sample (schizophrenia & affective disorders) without a healthy control comparison group, as well as carried out cross-sectionally. This limits our ability to understand the causality between health behaviours and changes in TL over time.

Another notable limitation was that TL was measured using qPCR, which provided us with a mean TL measure, rather than the number of short telomeres within a sample³⁸. The qPCR is a well validated measure for assessing TL, however we cannot exclude the possibility that focusing on critically short TL, could have given alternative information about the relationship between health behaviours and TL. Lastly, our study only included one marker of biological ageing (TL), which could not capture the entire ageing process. Ideally several markers should have been included, such as epigenetic clocks and brain age measures.

In conclusion, our study suggests that a healthier lifestyle is associated with cross-sectionally longer TL in individuals with SMD. In addition, our study highlights the distinct influence of childhood sexual abuse on adult health behaviours. This highlights the importance of health behaviours as well as addressing trauma-specific needs, as potential clinical targets for ensuring healthier cellular ageing in psychiatric populations.