Show summary
Introduction: Schizophrenia is associated with a broad spectrum of symptoms and impairments, some insufficiently responsive to or addressed by existing or available treatment. These include neurocognitive deficits, negative symptoms, depressive symptoms and low physical fitness. Given their negative effects on daily functioning and well-being, long-term functional outcome and even lifespan expectancy, there is a need to expand treatment options. Physical exercise has received growing attention for its beneficial effects on physical- and mental health. It represents a potential treatment targeting multiple aspects of neuropsychiatric disorders, such as schizophrenia. Investigation of physical exercise as a potential clinical treatment to enhance physical and mental health in different populations spans three core themes: i) The efficiency of physical exercise in improving core aspects of physical and mental health, ii) the mechanistic pathway(s) linking physical exercise to mental health and iii) moderating factors that may contribute to explain response variability. The current thesis aimed at addressing these overall research foci. The four included papers address the efficiency and efficacy of physical exercise in improving neurocognition (primary outcome), alleviating clinical symptoms and enhancing physical fitness (secondary outcomes) in individuals with schizophrenia. They proceed by exploring mechanistic pathways and moderating factors, focusing on the Brain Derived-Neurotropic Factor (BDNF; secondary outcome) and Volume of Oxygen consumed (VO2max).
Methods: The four papers included in this thesis build on the clinical trial “Effects of Physical Activity in Psychosis” (EPHAPS), organized at Vestfold Hospital Trust (ClinicalTrials.gov, number 02205684, 31.07.14). Designed as an observer-blinded randomized controlled trial (RCT), it compares the High-Intensity Interval Training (HIIT) to an active comparison intervention consisting of computerized sport-simulation (Active-Video-Gaming, AVG), serving to control for time spent and the possible effects of social interaction due to supervision. Both were provided as supervised sessions twice a week for twelve consecutive weeks. Eligible for participation were individuals fulfilling the diagnostic criteria for schizophrenia spectrum disorder, recruited from two outpatient clinics at Vestfold Hospital Trust (n=82) and randomly allocated to HIIT (n=43) or AVG (n=39). Assessments were performed at baseline, post-intervention and follow up (four months after the intervention ceased). The primary RCT analyses were performed by Mixed Effects Models, according to intention-to-treat (efficiency) and per protocol (efficacy) principles. We hypothesized that HIIT was more efficient than AVG in improving physical fitness (Paper I), neurocognition (Paper II) and overall clinical symptoms (Paper III). We proceeded by testing if statistical models of two core mechanistic hypotheses, the BDNF hypothesis (Paper IV) and the cardiovascular fitness hypothesis (Paper III), were supported by the data from the trial. Through this we also addressed variability in response. These secondary post-hoc analyses were performed by MEdiation and MOderation for REpeated measures (MEMORE). Attrition rate was 13% at post-intervention, cumulating to 31% at follow-up. At study end, 65% of the participants were defined as protocol-Compliant study-Completers (CC), excluding participants with protocol-violation (n=10) and/or lost to attrition. The mean attendance to sessions was 71% (82% in CC). The mean attendance to HIIT intervals was 73% (80% in CC).
Results: We estimated no significant changes in mean levels of VO2max or supplementary measures of physical fitness (body composition and lung function), from baseline to post-intervention. In post-hoc exploratory analyses we observed that 47% of the participants in HIIT, versus 27% in AVG, increased their VO2max by ≥5%. VO2max increments correlated positively with number of intervals performed at an intensity of ≥85% of maximum heartrate. Participants supervised by exercise-competent personnel, compared to other personnel, showed significantly larger increase in VO2max. Post hoc analyses indicated that HIIT improves VO2max significantly more than AVG if supervised by exercise-competent personnel. The subgroup with VO2max response (>1 metabolic equivalent of tasks) had higher age and lower physical fitness at baseline compared to the non-response group. Within the HIIT group, 61% showed significant increase in work capacity (measured as inclination and speed on the treadmill). We found a significant increase in the neurocognitive composite score from baseline to post-intervention and follow-up in both groups, reaching small to moderate pre-post effect sizes, albeit with no significant difference between HIIT and AVG. A similar pattern of change was estimated for most of the neurocognitive subdomains. We observed variability in response. Baseline levels of peripheral mature(m)BDNF significantly moderated the effect of physical exercise on neurocognition, regardless of group, and explained 7% of the variance in neurocognitive change. Higher baseline values of mBDNF, compared to lower, were associated with larger increase in the neurocognitive composite score. From baseline to post-intervention and follow-up, the mean level of mBDNF decreased: independently of group, differing by sex and associating negatively with change in VO2max. Change in BDNF was not significantly related to change in neurocognition. More specifically, our data did not support a mediating role of mBDNF on exercise effects on neurocognition. Beyond the decrease in mBDNF on group level, we estimated different pathways of change in BDNF moderated by its baseline value. For the clinical symptoms, we estimated a small reduction in overall levels (PANSS total), with no significant between-group difference. On a subfactor level (PANSS five factor), we found that HIIT reduced depressive symptoms (PANSS depressed factor) significantly more than AVG, reaching mild to moderate between-group effect size. Change in depressive symptoms associated negatively with change in VO2max. When testing a mediating role of VO2max on exercise effects on depressive symptoms, moderated by group, data tended to support this, but the moderated mediation index did not reach statistical significance. Change in overall symptoms was also negatively associated with change in VO2max. No significant association was found between change in neurocognition and change in VO2max.
Discussion: Findings from EPHAPS show that physical exercise may enhance both physical and mental aspects of schizophrenia. They also underscore the need to address variability in response to refine knowledge about how to prescribe exercise if used as adjunctive treatment in a clinical setting. The strict answer to the aims and hypotheses is: HIIT is more efficient than AVH in alleviating depressive symptoms, but not in improving neurocognition, alleviating overall clinical symptoms or increasing physical fitness. These treatment effects are based on between-group comparisons of average responses. Beyond this, we found improved neurocognition in both groups, supported by concurrent research, reaching a small to moderate pre-post effect size. Improvements spanned across multiple neurocognitive subdomains frequently affected in schizophrenia. This holds large clinical value. Baseline levels of mBDNF had a moderating role of exercise effects on neurocognition and explained a small part of the observed variability in response. The mBDNF hypothesis was not supported by the current data, albeit cannot be disclaimed. Neurocognitive response was neither related to change in BDNF nor VO2max. The findings support that neurocognition can be improved by various exercise modes (HIIT and AVG) also in absence of improved VO2max. Overall clinical symptoms were alleviated in the total sample, but HIIT was more efficient than AVG in alleviating depressive symptoms, corresponding to a mild to moderate effect size. Reduction in depressive symptoms, and overall clinical symptoms, were associated with improved VO2max, tending to support the cardiorespiratory hypothesis but not reaching significance in a mediation model. Relationships with mBDNF were not explored. HIIT may potentially be used to alleviate mild to moderate depressive symptoms, also before they reach clinical depression. Effects on both neurocognition and clinical symptoms sustained four months after the intervention ceased. While the mean level of VO2max did not improve in the HIIT group, change in VO2max correlated positively with number of intervals at sufficiently high intensity and was larger if supervised by exercise-competent personnel compared with other mental health staff. Variability in response was observed for both physical and mental outcomes. Going beyond efficiency to build knowledge about mediating and moderating factors may contribute to refining interventions, understanding variability in responses, explaining discrepancies across studies and guiding clinical prescriptions. Potentially, this may facilitate more consistent responses and/or advance individualized treatment principles within the field of physical exercise and health. More detailed recommendations for prescription (mode, dose, intensity – potentially depending on the outcome) are needed to guide clinical choices. Apart from being under-powered, the design and methodology in EPHAPS is evaluated as robust, with low risk of biases, and the main results are regarded as trustworthy. Explorative results must be interpreted with caution and preferably used to generate new research hypotheses. If provided as e directed clinical treatment, physical exercise should be supervised, preferentially by exercise-professionals, and responses need to be monitored. Integrating exercise clinics into mental health services may facilitate this and promote integrated treatment of mental and physical aspects associated with schizophrenia. Conducted in a realistic clinical outpatient setting, both HIIT and AVG appeared safe, feasible and applicable. Few exclusion criteria facilitate the generalisability of results. Overall, EPHAPS supports the use of physical exercise as an adjunctive treatment for individuals with schizophrenia. The current findings comply with the emerging guidelines, recommending supervised physical exercise as part of multimodal treatment for individuals with schizophrenia to target both physical and mental health.
