Canadian Psychiatric Association Association des psychiatres du Canada Original Research Physical Activity as a Predictor of Clinical Trial Outcomes in Bipolar Depression: A Subanalysis of a Mitochondrial-Enhancing Nutraceutical Randomized Controlled Trial The Canadian Journal of Psychiatry / La Revue Canadienne de Psychiatrie 1-13 ª The Author(s) 2019 Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/0706743719889547 TheCJP.ca | LaRCP.ca L’activité physique comme prédicteur des résultats d’un essai clinique en dépression bipolaire : une sous-analyse d’un essai randomisé contrôlé d’un nutriceutique améliorant les mitochondries Melanie M. Ashton, BSc, GDip1,2,3, Mohammadreza Mohebbi, PhD4, Alyna Turner, PhD1,5, Wolfgang Marx, PhD1,6, Michael Berk, MD, PhD1,3,5,7,8 , Gin S. Malhi, MD9,10,11, Chee H. Ng, MBBS, MD2, Sue M. Cotton, PhD7,8, Seetal Dodd, PhD1,5,7, Jerome Sarris, PhD2,12, Malcolm Hopwood, MD13, Brendon Stubbs, PhD14,15, and Olivia M. Dean, PhD1,3 Abstract Objectives: Individuals with bipolar disorder (BD) generally engage in low levels of physical activity (PA), and yet few studies have investigated the relationship between PA and change in BD symptom severity. The aim of this subanalysis of an adjunctive nutraceutical randomized controlled trial for the treatment of bipolar depression was to explore the relationship between PA, the active adjunctive treatments (a nutraceutical “mitochondrial cocktail”), and clinical outcomes. Methods: Participants with bipolar depression were randomized to receive N-acetylcysteine alone, N-acetylcysteine with a combination of nutraceuticals (chosen for the potential to increase mitochondrial activity), or placebo for 16 weeks. Participants (n ¼ 145) who completed the International Physical Activity Questionnaire–Short Form (IPAQ-SF; measured at Week 4) were included in this exploratory subanalysis. Assessments of BD symptoms, functioning, and quality of life were 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 IMPACT Strategic Research Centre, School of Medicine, Barwon Health, Deakin University, Geelong, Victoria, Australia Professorial Unit, The Melbourne Clinic, Department of Psychiatry, University of Melbourne, Richmond, Victoria, Australia The Florey Institute of Neuroscience and Mental Health, Parkville, Victoria, Australia Biostatistics Unit, Faculty of Health, Deakin University, Geelong, Victoria, Australia School of Medicine and Public Health, Faculty of Health and Medicine, University of Newcastle, Callaghan, NSW 2308, Australia Department of Rehabilitation, Nutrition and Sport, School of Allied Health, College of Science, Health and Engineering, La Trobe University, Bundoora, Victoria, Australia Centre of Youth Mental Health, University of Melbourne, Parkville, Victoria, Australia Orygen, Parkville, Victoria, Australia Academic Department of Psychiatry, Northern Sydney Local Health District, St Leonards, New South Wales, Australia Faculty of Medicine and Health, Department of Psychiatry, Northern Clinical School, University of Sydney, New South Wales, Australia CADE Clinic, Royal North Shore Hospital, Northern Sydney Local Health District, St Leonards, New South Wales, Australia NICM Health Research Institute, Western Sydney University, Westmead, New South Wales, Australia Professorial Psychiatry Unit, Albert Road Clinic, Department of Psychiatry, University of Melbourne, Melbourne, Victoria, Australia Physiotherapy Department, South London and Maudsley NHS Foundation Trust, London, United Kingdom Department of Psychological Medicine, Institute of Psychiatry, Psychology and Neuroscience, King’s College London, United Kingdom Corresponding Author: Melanie M. Ashton, BSc, GDip, IMPACT Strategic Research Centre, Deakin University, P.O. Box 281, Geelong, Victoria 3220, Australia. Email: m.ashton@deakin.edu.au 2 The Canadian Journal of Psychiatry completed at monthly visits up until Week 20. Generalised Estimating Equations were used to explore whether IPAQ-SF scores were a moderator of treatment received on outcomes of the study. Results: Week-4 PA was not related to changes in Montgomery Åsberg Depression Rating Scale scores across the study until Week 20. However, participants who engaged in more PA and who received the combination treatment were more likely to have a reduction in scores on the Bipolar Depression Rating Scale (P ¼ 0.03). However, this was not consistent in all domains explored using the IPAQ-SF. Participants who engaged in higher levels of PA also experienced greater improvement in social and occupational functioning and less impairment in functioning due to their psychopathology and improvement in quality of life at Week 20, irrespective of treatment. Conclusions: This study provides novel evidence of the association between PA and reduction in BD symptoms in a nutraceutical clinical trial. However, further research assessing the potential synergistic effects of PA in BD is required. Abrégé Objectifs : Les personnes souffrant d’un trouble bipolaire (TB) ne s’adonnent généralement qu’à de faibles taux d’activité physique (AP), et pourtant, peu d’études ont recherché la relation entre l’AP et le changement de la gravité des symptômes du TB. L’objet de cette sous-analyse d’un essai randomisé contrôlé d’un adjuvant nutriceutique pour le traitement de la dépression bipolaire était d’explorer la relation entre l’AP, les traitements adjuvants actifs (un « cocktail mitochondrial » dans un nutriceutique), et les résultats cliniques. Méthodes : Les participants souffrant de dépression bipolaire ont reçu de façon aléatoire soit la N-acétylcystéine seulement, soit la N-acétylcystéine avec une combinaison de nutriceutiques (choisis pour leur potentiel d’accroı̂tre l’activité mitochondriale), soit un placebo pendant 16 semaines. Les participants (n ¼ 145) qui ont rempli la version abrégée du questionnaire d’activité physique international (IPAQ-SF; mesuré à la 4e semaine) ont été inclus dans cette sous-analyse exploratoire. Les évaluations des symptômes de TB, du fonctionnement et de la qualité de vie ont été effectuées lors de visites mensuelles, jusqu’à la 20e semaine. Des modèles linéaires mixtes ont servi à explorer si les scores à l’IPAQ-SF étaient un modérateur du traitement reçu dans les résultats de l’étude. Résultats : À la 4e semaine, l’AP n’était pas liée aux changements des scores à l’échelle de la dépression de Montgomery Åsberg dans toute l’étude jusqu’à la 20e semaine. Toutefois, les participants qui faisaient plus d’AP et qui recevaient un traitement combiné étaient plus susceptibles d’avoir une réduction de leurs scores à l’échelle de dépression bipolaire (P ¼ 0.03). Cependant, cela n’était pas constant dans tous les domaines explorés à l’aide de l’IPAQ-SF. Les participants qui se sont adonnés à des taux d’AP plus élevés ont aussi connu une plus grande amélioration du fonctionnement social et professionnel, et moins de déficience du fonctionnement en raison de leur psychopathologie et de la qualité de vie à la 20e semaine, sans égard au traitement. Conclusions : Cette étude apporte de nouvelles données probantes de l’association entre l’AP et la réduction des symptômes de TB dans un essai clinique nutriceutique. Il faut cependant plus de recherche pour évaluer les effets synergiques de l’AP dans le TB. Keywords physical activity, exercise, bipolar disorder, bipolar depression, mitochondrial agents, nutraceuticals, N-acetylcysteine Introduction Bipolar depression is often difficult to treat. One approach to optimize the effects of current therapeutics may be through lifestyle interventions such as engagement in physical activity (PA). Despite many known benefits of PA in the general population1 and increasing evidence that individuals with other serious mental disorders such as schizophrenia2 and major depression3,4 can also benefit, limited research has investigated PA and symptom severity in bipolar disorder (BD; for reviews5,6). To date, the literature is largely based on cross-sectional, prospective cohort, or small pilot studies, all of which suggest that engagement in PA improves mood and quality of life, but the evidence base is limited.5,7-9 Individuals with BD engage in lower levels of PA, are less likely to meet recommended international guidelines for exercise (World Health Organization [WHO]10), and are more likely to be sedentary versus age- and sex-matched controls.11 Therefore, not surprisingly, people with BD demonstrate lower levels of cardiorespiratory fitness compared to healthy controls.12,13 Previous research has suggested that increased PA is associated with better cognition in euthymic females with a diagnosis of BD.14 Achieving an adequate level of PA has been included in the current National Institute of Health guidelines for treating BD, but only in the broad sense of improving general health.15 In the general population, it is recommended that individuals achieve 150 min of moderate or 75 min of vigorous PA per week.10 The literature to date in the general population has found that both continuous and interval aerobic PA at a moderate to high intensity can improve mitochondrial function.16- 3 La Revue Canadienne de Psychiatrie 19 An emerging evidence base also advocates that resistance training, specifically targeting the loading and strengthening of skeletal muscles, can also improve mitochondrial function.19,20 While people with BD may have mitochondrial dysfunction,21 it is unclear whether PA at moderate to high intensity at recommended guidelines such as those recommended by WHO10 can influence mitochondrial function. PA is low-cost, safe, and tolerable and therefore could be an effective adjunct to improve response to treatment in BD; however, this has been largely unexplored. Therefore, we aimed to investigate whether PA was associated with changes in symptoms, functioning, and quality of life in BD. This study was embedded in a double-blind randomized controlled trial (RCT) evaluating the efficacy of adjunctive nutraceuticals for the treatment of bipolar depression. The adjunctive nutraceuticals were specifically selected due to their potential mitochondrial-enhancing properties,22 and there may be a relationship between PA and the nutraceuticals via mitochondrial biogenesis.23 There were three arms of the RCT: Nacetylcysteine (NAC) alone, a combination treatment (CT) of nutraceuticals including NAC and placebo. We hypothesized that reported PA would be an effect modifier for the relationship between those receiving NAC alone or the CT, and an improvement on depression, functioning, and quality of life outcomes. We also hypothesized that PA in categorical terms, according to the scoring guide of the PA scale (low, moderate, and high), would be an effect modifier for the relationship between those receiving NAC alone or CT, and outcomes (detailed above). Finally, when utilizing data categorized by WHO recommendations, we hypothesized that PA (according to WHO recommendations) would be an effect modifier for the relationship between treatment with NAC alone or CT, and outcomes (detailed above). Methods Ethics The study was run in accordance with International Council for Harmonisation Good Clinical Practices Guidelines.24 Ethical approval was granted from Barwon Health Human Research and Ethics Committee (HREC), Northern Sydney Local Health District HREC, The Melbourne Clinic Research Ethics Committee and Deakin University HREC. The study is registered on the Australian and New Zealand Clinical Trial Registry (ACTRN12612000830897). Trial Study Design Participants (n ¼ 181) who were randomized received the study medication for 16 weeks and visited study sites (Melbourne, Geelong, and Sydney) every 4 weeks for clinical interviews with a research assistant up until Week 20. Inclusion criteria were a diagnosis of BD, determined by the Mini-International Neuropsychiatric Interview 5.025 and a current moderate to severe depressive episode measured by a score 20 on the Montgomery Åsberg Depression Rating Scale (MADRS). 26 Full study protocol 16 and primary results20 have been published previously. The primary aim of the trial was to assess the efficacy of the two active arms of the study (NAC alone and CT) compared to placebo for treating depressive symptoms (measured by the MADRS) at Week 16. Primary results of the study at the primary endpoint were not significant at Week 16.27 However, at Week 20 (4 weeks post–study medication discontinuation), CT was superior to placebo at improving the following outcome measures; changes in depression symptoms measured by the MADRS which was the primary outcome measure in the study; bipolar depression symptom severity measured by the Bipolar Depression Rating Scale (BDRS)28; Social and Occupational Functioning Assessment Scale (SOFAS), 29 a clinician-rated measure of functioning; The Longitudinal Interval Follow-Up Evaluation–Range of Impaired Functioning (LIFE-RIFT),30 a clinician-rated measure of impairment in functioning from psychopathology and the Clinical Global Impressions Scales Bipolar Version–Improvement (CGI-I), 31 a 1-item clinician-rated scale measuring improvement. Participants also completed The Quality of life Enjoyment and Satisfaction Questionnaire–Short Form (Q-LES-Q-SF),32 a self-report measure of quality of life. There was no significant relationship between CT versus placebo in regard to Q-LES-Q-SF scores, but this outcome was included in the subanalysis because of the association between PA and quality of life in BD.13 Total possible scores for each outcome measure and indication of direction for improvement can be found in Supplemental Table 1. PA The International Physical Activity Questionnaire–Short Form (IPAQ-SF)33 was administered at Week 4 to measure each participant’s general level of PA. The IPAQ-SF is a 10-item self-report questionnaire where participants recall the number of days and minutes of vigorous activity, moderate activity, walking and sitting time, over the past 7 days. The IPAQ-SF has been used extensively in other mental health disorder populations and has acceptable validity and reliability.34 The IPAQ-SF was administered at Week 4 to reduce participant burden at the baseline visit and to coincide with collection of dietary intake data. The IPAQ-SF was administered as secondary outcomes’ data and has been included in the protocol16; however, this measure was inadvertently omitted from the trial registry. Data were cleaned using IPAQ-SF recommendations33 that include removing cases with missing values and removing cases with values too low (less than 10 min of activity per day). There was no missing data for vigorous, moderate activity, or walking. Two participants had missing values for the “time spent sitting” item. Both these participants remained in the analysis as this item is not used to calculate total scores or categorical scores. Minimum and maximum values were implemented to remove outliers. As a result, one 4 participant was removed for too few minutes (6 min) of activity. To normalize the data, the protocol suggests truncating each daily activity time to no more than 180 min. This rule was employed for five participants reporting vigorous activity, seven for moderate activity, and six for walking. Of note, one participant filled in the IPAQ-SF questionnaire at Week 8, not Week 4 but remained in the analysis. Weekly metabolic equivalent of task (MET)-minute scores for each activity type were first calculated as follows: Note. Abbreviations: BMI ¼ body mass index; CT ¼ combination treatment; IPAQ ¼ International Physical Activity Questionnaire; NAC: N-acetylcysteine. *P < 0.05. **P < 0.01. 360.0 990.0 Male gender, n (%) 17 (34.7) 18 (36.0) 16 (34.8) 51 (35.2) Age, M (SD) 45.88 (11.9) 45.0 (12.1) 47.7 (13.3) 46.1 (12.4) 21.3 to 72.0 BMI, M (SD) 30.3 (7.9), n ¼ 47 27.9 (6.3) 28.2 (6.8) 28.8 (7.0), n ¼ 143 16.82 to 52.8 Total weekly physical activity (MET-min), M (SD) 2,024.4 (2,477.8) 1,766.3 (2,433.2) 1,603.1 (1,718.9) 1,801.8 (2,239.3) 0 to 11,118.0 IPAQ categorical Low, n (%) 22 (44.9) 27 (54.0) 23 (50.0) 72 (49.7) Moderate, n (%) 13 (26.5) 11 (22.0) 14 (30.4) 38 (26.2) High, n (%) 14 (28.6) 12 (24.0) 9 (19.6) 35 (24.1) WHO recommendations No physical activity, n (%) 7 (14.3) 4 (8.0) 5 (10.9) 16 (11.0) Below WHO recommendations, n (%) 9 (18.4) 17 (34.0) 14 (30.4) 40 (27.6) Within WHO recommendations, n (%) 9 (18.4) 10 (20.0) 4 (8.7) 23 (15.9) Above WHO recommendations, M (SD) 24 (49.0) 19 (38.0) 23 (50.0) 66 (45.5) Minutes spent sitting per weekday, M (SD) 421.9 (236.6) 427.6 (251.6) 441.8 (239.9), n ¼ 44 430.0 (241.4), n ¼ 143 21 to 1,260 Characteristics Table 1. Study Participants’ Characteristics. Placebo (n ¼ 49) NAC (n ¼ 50) CT (n ¼ 46) Total (n ¼ 145) Range Median Cutpoint The Canadian Journal of Psychiatry  Vigorous activity, minutes/week ¼ total minutes per week of vigorous activity  8.0 METs  Moderate activity, minutes/week ¼ total minutes per week of moderate activity  4.0 METs  Walking, minutes/week ¼ total minutes per week of walking  3.3 METs Each total activity-MET score was then summed to create a continuous total PA score. In addition to total PA scores, a categorical value was produced for each participant. The categories were low, moderate, or high PA and were calculated for each participant in accordance with IPAQ-SF scoring protocol.33 Within this protocol, participants’ activity levels were deemed high if they engaged in at least 3 days of vigorous activity and achieving a total activity of at least 1,500 MET min/week, or a combination of all intensity levels for 7 days or more and achieving a total activity of at least 3,000 MET min/week. Moderate activity category was achieved if participants engaged in at least 20 min of vigorous activity for 3 days, or at least 30 min of walking and/or moderate activity for 5 days, or a combination of any activity level for 5 or more days and achieving a total activity of at least 600 MET-min/week. Lastly, participants’ activities were categorized as low if they did not fit into either of the above categories. A summary of categorical scores for the sample can be found in Table 1. The last item of the IPAQ-SF is “time spent sitting” and is used to assess participants’ rates of sedentary behaviors. Sitting has been presented as a separate variable, measured in average minutes per typical weekday. In addition to the validated exploration of the IPAQ scale, further analysis was conducted using WHO recommendations. This was completed to provide preliminary data for guidelines and clinical practice and to provide real-world advice to patients. To explore these data in relation to WHO recommendations, total PA data in MET-min/week were categorically scored. These additional categories were utilized to aid direct interpretation of the results to participant adherence to WHO recommendations as outlined below. This quick interpretation allows results from this study to be easily translated into policy and clinical care. 1. No PA—All activity < 10 min duration (equivalent of 0 MET-min/week). 2. Below WHO recommendations—Less than 150 min of moderate activity or 75 min of vigorous activity per week. 5 La Revue Canadienne de Psychiatrie Equivalent of energy expenditure between 0 and 600 METmin/week (not inclusive). 3. Within WHO recommendations—At least 150 min of moderate activity or 75 min of vigorous activity per week. Equivalent of energy expenditure between 600 and 1,200 MET-min/week (inclusive). 4. Exceeding WHO recommendations—WHO recommends for greater health benefits, at least 300 min of moderate activity and 150 min of vigorous activity. Equivalent of energy expenditure greater than 1,200 MET-min/week. Statistical Analysis Generalized estimating equations (GEE) were used to assess whether PA (as a total score, categorical value, and according to WHO recommendations) were predictors of outcomes from the nutraceutical RCT (MADRS, BDRS, SOFAS, LIFE-RIFT, Q-LES-Q-SF, and CGI-I scores). Each predictor was assessed individually including an exploration of each of the treatment arms (NAC alone or CT) compared to placebo across the study up until Week 20. By using GEE, the analyses are able to take into account the longitudinal nature of data (i.e., measurement autocorrelation in follow-ups). The primary outcome of the study followed a modified intention-to-treat analysis whereby participants with post– baseline data were included in the analysis.27 First, the original RCT analyses were replicated by including treatment arms as a nominal factor, log of follow-up time as a covariate, and the two-way interaction between log(time) and treatment arms was replicated, followed by including each predictor (each PA score) in a separate model to evaluate whether it is a predictor of outcomes. The latter model contained treatment arms as a nominal factor, log of follow-up time as a covariate, predictor of interest, all possible two-way interactions and the three-way interactions between treatment arms, log of follow-up time, and the predictor of interest. Three-way models evaluated the effect of each predictor on the outcome measure, across time in the study, for each treatment arm. Treatment by PA two-way interactions explored the role of the predictor for each of the study, independent of time. Each model utilized Baron and Kenny 35 criteria guidelines as first described by Kraemer et al.36 Each model for each of the predictors is described below. Categorical PA We took into account the ordinal nature of PA categories when modeling the IPAQ-SF as low, moderate, and high. The model included a fixed-effect treatment group and categorical (ordinal) PA, and logarithm of time as covariates, all two-way interactions and the three-way interactions. As above, three-way interactions were then removed to explore two-way interactions. Total PA was also assessed as a continuous score, details of which are outlaid in Supplemental Material. PA According to WHO Recommendations PA according to WHO recommendations was assessed as nominal data and included in the model as a factor. The initial model included a fixed-effect treatment group and PA according to WHO recommendations, and logarithm of time as covariate, all two-way interactions and all three-way interactions. After this model was run for each outcome, three-way interactions were then removed to explore all two-way interactions for each outcome. The P value for all overall three-way interactions were reported alongside Wald w2 statistic (used to measure parameter effects). In addition, for each treatment group (NAC alone or CT), three-way interactions were reported with P value and Wald w2 statistic, alongside their corresponding b coefficients and 95% confidence intervals (CIs) to measure association. After examining three-way and two-way interactions of interest for each predictor, the data were then further explored for nonspecified predictors. Nonspecified predictors demonstrated a relationship with change in the outcome measure independent of what treatment was received and time. Each model for nonspecified predictors included the main effects of treatment group, the predictor, and logarithm of time. This model assesses for nonspecified predictors as it explores the predictors’ response in the sample as a whole (combining all treatment groups). The GEE technique was implemented for model estimation using an unstructured working correlation matrix and a robust variance estimator.37 Statistical analyses were completed using IBM SPSS Statistics for Windows, Version 25.38 Results Participants Of the 181 participants in the clinical trial, 33 participants were excluded from the analysis for not having any post– baseline data, 2 participants excluded for missing IPAQ-SF data, and 1 participant excluded due to insufficient activity (less than 10 min activity). Therefore, 145 participants were included in the current analysis. The average age of the sample was 46.14 years (SD ¼ 12.38), ranging from 21 to 72 years of age, and 51% were male. Participants were randomized to receive NAC (n ¼ 50), CT (n ¼ 46), and placebo (n ¼ 49). A full list of study sample characteristics can be found in Table 1. Analysis of Predictors Change scores were calculated for each outcome measure (except CGI-I that self-evidently had no baseline data available). Mean change (Week 20 minus baseline scores) for each outcome variable per treatment group is shown in Supplemental Table 2. On average, participants in all treatment arms improved across all outcome measures. As CGI-I 6 Table 2. IPAQ Scores Categorized into Low, Moderate, and High According to IPAQ-SF Guidelines as a Predictor of Mean Change Scores for Each Treatment Arm. Placebo METCategorical Low Placebo CT NAC MADRS change Mean (SD) 12 (10.1) 14.1 (9.5) 15.1 (10.8) n 20 18 19 BDRS change Mean (SD) 9.5 (9.8) 10.2 (10.3) 13.4 (9.7) n 19 16 19 SOFAS change Mean (SD) 12.7 (12.9) 17.3 (12.7) 13.3 (10.6) n 19 18 18 LIFE-RIFT change Mean (SD) 2.3 (4.4) 3.7 (5.0) 3.7 (3.7) n 19 16 19 Q-LES-Q change Mean (SD) 12.7 (23.4) 16.7 (18.0) 18.1 (18.1) n 19 18 19 CGI-I Week 20b n 20 18 19 NAC CT Interactiona b Coefficient (95% CI) Moderate CT 4.9 (11.2) 10 18.7 (7.6) 11 15.4 (10.8) 14.8 (12.2) 22.5 (3.5) 7 10 2 2.1 (12.4) 8 17.1 (7.9) 11 13.8 (9.3) 6 NAC High 11.7 (9.3) 10 CT 10.6 (8.1) 9 21.0 (2.8) 2 Interaction Test Interaction Test P Value NAC Interactiona b Coefficient (95% CI) 0.02 (2.9 to 2.8) w2(1) < 0.01 P ¼ 0.987 2.8 (0.01 to 5.6) w2 (1) ¼ 3.8 P ¼ 0.051 9.2 (7.2) 9 0.2 (2.7 to 3.0) w2(1) ¼ 0.01 P ¼ 0.917 2.9 (0.03 to 5.7) w2(1) ¼ 3.9 P ¼ 0.047 Placebo P Value 17.5 (14.5) 11 14.5 (14.5) 6 14.7 (12.0) 10 12.5 (7.8) 2 13.0 (11.2) 9 0.7 (4.8 to 3.4) w2(1) ¼ 0.1 P ¼ 0.736 0.2 (3.8 to 3.5) w2(1) ¼ 0.01 P ¼ 0.932 0.8 (3.6) 8 5.6 (3.8) 11 4.7 (2.5) 6 3.7 (4.2) 10 4.0 (1.4) 2 3.6 (2.0) 9 0.1 (1.5 to 1.3) w2(1) ¼ 0.01 P ¼ 0.909 0.5 (0.7 to 1.8) w2(1) ¼ 0.7 P ¼ 0.417 <0.01 (15.0) 9 23.1 (19.8) 11 17.1 (10.0) 7 21.8 (18.0) 9 26.8 (12.6) 2 15.9 (23.7) 9 1.8 (6.7 to 3.2) 0.2 (0.4 to 0.8) w2(1) P w2(1) P 2.6 (12.7) 9 9 11 7 10 2 9 ¼ 0.5 ¼ 0.486 ¼ 0.4 ¼ 0.546 5.2 (11.0 to 0.6) 0.05 (0.5 to 0.6) w2(1) P w2(1) P ¼ 3.1 ¼ 0.077 ¼ 0.03 ¼ 0.864 Note. Abbreviations: BDRS ¼ Bipolar Depression Rating Scale; CGI-I ¼ Clinical Global Impression Improvement; CT ¼ combination treatment; LIFE-RIFT ¼ Longitudinal Interval Follow-Up Evaluation–Range of Impaired Functioning Tool; MADRS ¼ Montgomery Åsberg Depression Rating Scale; NAC ¼ N-acetylcysteine; SOFAS: Social and Occupational Functioning Scale. a Three-way interaction between potential predictor, time and treatment group, reference group was placebo. b As CGI-I is not administered at baseline, mean score change has not been measured. High and low levels of each predictor were determined by median split. 7 La Revue Canadienne de Psychiatrie represents a single score of change from baseline, mean Week-20 CGI-I scores per treatment group are summarized in Supplemental Table 3. On average, research clinicians rated participants as improving across the study. For all models with a significant interaction, age, sex and body mass index were explored as potential confounders, and no factors had a statistically significant impact on the relationships. PA as a Categorical Variable PA scores on the IPAQ-SF were categorized as low, moderate, or high using scale recommendations. From the whole sample, 49.7% of participants were categorized as engaging in low weekly PA, 26.2% engaging in moderate weekly PA, and 24.1% engaging in high weekly PA. A visual representation of data has been included in Supplemental Figure 1b. Categorical PA was not significantly associated with scores for MADRS, SOFAS, LIFE-RIFT, Q-LES-Q-SF, or CGI-I (see Table 2). There was a three-way interaction between taking NAC and engaging in high exercise and participant’s BDRS outcomes. Compared to placebo, participants receiving NAC and engaging in a high amount of exercise showed an increase in BDRS scores, indicating a worsening of symptoms across the trial. For every one-level increase in level of PA (i.e., level of PA according to IPAQSF categorical scores), mean BDRS on NAC further increased by 2.85 (95% CI, 0.03 to 5.7) units when compared with placebo group with similar PA level. There were no significant two-way interactions between treatment received and categorical PA. WHO Recommendations PA scores were represented in terms of WHO recommendations. From the whole sample, 11% engaged in no weekly PA, 27.6% engaged in weekly PA under the WHO recommendations, 15.9% engaged in weekly PA within the WHO recommendations, and 45.5% engaged in weekly PA greater than, or, exceeding the WHO recommendations. A visual representation of data has been included in Supplemental Figure 1c. Results of the effect modification analysis are shown in Table 3. PA according to WHO recommendations was not significantly associated with scores for MADRS, SOFAS, LIFE-RIFT, Q-LES-Q-SF, or CGI-I. There was a significant three-way interaction between treatment received, PA according to WHO recommendations, and time. Participants who were randomized to receive CT and engaged in more PA had a greater reduction in BDRS scores, indicating an improvement in symptoms. For every one-level increase in PA (i.e., level of PA categorized according to WHO recommendations), mean BDRS in the combination therapy group further decreased by 2.15 (95% CI, 4.07 to 0.23) units when compared with the placebo group with similar PA levels. There were no significant two-way interactions between treatment received and PA in terms of WHO recommendations. Total PA Scores Total PA, as a continuous score, was not significantly associated with MADRS, SOFAS, LIFE-RIFT, Q-LES-Q-SF, or CGI-I scores (see Supplemental Table 4). There was, however, a significant three-way interaction between participants taking CT and engaging in more PA and participant’s BDRS outcomes. Compared to placebo, participants receiving CT and engaging in a high amount of exercise showed a decrease in BDRS scores at Week 20 indicating an improvement in symptoms across the trial. For every 10% increase in participants’ total MET score, BDRS scores decreased by 0.09 (95% CI, 1.8 to 0.1) units. There were no significant two-way interactions between treatment received and log-transformed total PA. Nonspecified Predictors’ Analysis Results of the nonspecified predictors of outcomes analysis can be found in Table 4. Total PA was not significantly related to MADRS or CGI-I outcomes. Total PA was a significant nonspecified predictor of SOFAS, LIFE-RIFT, and Q-LES-Q-SF scores at Week 20, irrespective of treatment received. For every 10% increase in participants’ total MET score, SOFAS scores increased by 0.06 (CI, 0.01 to 1.31) units, LIFE-RIFT scores decreased by 0.02 (95% CI, 0.41 to 0.08) units, and Q-LES-Q-SF scores would increase by 0.09 (95% CI, 0.13 to 1.80) units. Categorical PA did not significantly predict Week-20 MADRS scores. Higher PA categories, according to the IPAQ-SF scoring protocol, was a nonspecified predictor of SOFAS, LIFE-RIFT, CGI-I, and Q-LES-Q-SF. Higher activity levels were more likely to be associated with slightly improved scores for these measures, regardless of treatment received. For every one-level increase in level of PA according to IPAQ-SF categorical scores (i.e., moderate to high), mean SOFAS scores at Week 20 increased by 2.27 (95% CI, 0.24 to 4.30) units, mean LIFE-RIFT scores decreased by 0.67 (95% CI, 1.23 to 0.11) units, mean Q-LES-Q-SF scores increased by 2.39 (95% CI, 0.03 to 4.75) units, and mean CGI-I scores decreased by 0.16 (95% CI, 0.31 to 0.01) units. PA according to WHO recommendations was not associated with Week-20 MADRS or CGI-I scores. Higher PA categories, according to WHO recommendations, was a nonspecified predictor of SOFAS, LIFE-RIFT, and Q-LESQ-SF. Higher activity levels were more likely to be associated with slightly improved scores for these measures, regardless of treatment received. For every one-level increase in level of PA according to WHO recommendations (i.e., from below to within recommendations), mean SOFAS scores at Week 20 increased by 1.80 (95% CI, 0.35 to 3.25) units, mean LIFE-RIFT scores decreased by 0.71 (95% CI, 8 Table 3. Physical Activity According to WHO Recommendations as a Predictor of Mean Change Scores for Each Treatment Arm. Placebo MET–WHO CT No Physical Activity NAC Placebo CT NAC Placebo Below WHO Recommendations MADRS change Mean (SD) 11.3 (14.5) 9.0 (13.9) 15.0 (7.2) 14.1 (5.0) 14.8 (9.1) 14.9 (10.1) n 7 3 4 8 12 12 BDRS change Mean (SD) 14.1 (11.0) <0.01 (14.1) 14.5 (11.1) 12.1 (7.5) 9.9 (9.8) 12.9 (9.3) n 7 2 4 7 11 12 SOFAS change Mean (SD) 15.3 (16.0) 7.0 (15.9) 19.8 (10.1) 13.9 (11.3) 18.0 (11.3) 12.9 (8.9) n 7 3 4 7 12 11 LIFE-RIFT change Mean (SD) 3.0 (4.5) <0.01 (7.1) 3.5 (5.1) 3.7 (3.5) 3.7 (4.2) 3.4 (4.0) n 7 2 4 7 11 12 Q-LES-Q change Mean (SD) 13.8 (22.2) 10.1 (14.5) 24.6 (11.4) 19.4 (23.7) 18.8 (16.2) 16.4 (18.9) n 7 3 4 7 12 12 b CGI-I Week 20 n 7 3 4 8 12 12 CT NAC Within WHO Recommendations Placebo CT Exceeding WHO Recommendations 3.0 (11.2) 15.3 (7.1) 8 4 18.0 (13.7) 13.0 (11.1) 19.9 (7.0) 7 17 12 0.6 (11.1) 16.8 (5.1) 7 4 15.0 (11.1) 6 1.6 (11.3) 8 16.0 (15.4) 4 9.7 (16.8) 6 NAC 10.4 (8.5) 12 Overall Three-Way Interaction Test P Value Interactiona b Coefficient (95% CI) w2(2) ¼ 3.3 P ¼ 0.19 0.8 (2.8 to 1.2) Interaction Test P Value w2(1) ¼ 0.6 P ¼ 0.440 Interactiona b Coefficient (95% CI) Interaction Test P value 1.1 w2(1) ¼ 1.0 (1.1 to 3.2) P ¼ 0.319 8.5 (10.2) 18.1 (7.5) 16 12 9.8 (10.5) 12 w2(2) ¼ 6.2 P ¼ 0.046 2.2 w2(1) ¼ 4.8 0.1 (4.1 to 0.2) P ¼ 0.028 (2.2 to 2.0) w2(1) <0.01 P ¼ 0.951 12.1 (12.4) 16 18.9 (13.4) 12 13.7 (10.5) 12 w2(2) ¼ 1.4 P ¼ 0.504 1.7 (1.1 to 4.5) w2(1) ¼ 1.4 P ¼ 0.243 0.8 w2(1) ¼ 0.4 (1.8 to 3.4) P ¼ 0.546 0.4 (4.7) 7 5.0 (4.1) 4 4.7 (2.0) 6 2.3 (4.2) 16 5.7 (4.2) 12 3.9 (4.0) 12 w2(2) ¼ 0.8 P ¼ 0.666 0.4 (1.3 to 0.5) w2(1) ¼ 0.8 P ¼ 0.382 0.1 w2(1) ¼ 0.03 (1.0 to 0.8) P ¼ 0.866 3.1 (16.0) 7 17.0 (11.9) 4 13.8 (15.8) 7 14.2 (20.9) 16 23.7 (22.8) 12 17.9 (21.1) 12 16 12 12 w2(2) ¼ 1.3 P ¼ 0.512 w2(2) ¼ 0.8 P ¼ 0.67 1.3 (2.5 to 5.1) 0.2 (0.6 to 0.2) w2(1) ¼ 0.5 P ¼ 0.496 w2(1) ¼ 0.7 P ¼ 0.402 1.0 w2(1) ¼ 0.2 (5.1 to 3.1) P ¼ 0.640 <0.01 w2(1) <0.01 (0.4 to 0.4) P ¼ 0.999 8 4 7 Note. Abbreviations: BDRS ¼ Bipolar Depression Rating Scale; CGI-I ¼ Clinical Global Impression Improvement; CT ¼ combination treatment; LIFE-RIFT ¼ Longitudinal Interval Follow-Up Evaluation–Range of Impaired Functioning Tool; MADRS ¼ Montgomery Åsberg Depression Rating Scale; NAC: N-acetylcysteine; SOFAS ¼ Social and Occupational Functioning Scale. a Three-way interaction between potential predictor, time and treatment group, reference group was placebo. b As CGI-I is not administered at baseline, mean score change has not been measured. High and low levels of each predictor were determined by median split. 9 La Revue Canadienne de Psychiatrie Table 4. Total Weekly Physical Activity, IPAQ Categorical Scores, and Physical Activity Categorized by WHO Recommendations as Nonspecified Predictors of Outcomes. Predictor b Coefficient (95% CI) Main Effect Total weekly physical activity MADRS 0.2 (0.6 to 0.3) w2(1) ¼ 0.6, P ¼ 0.458 BDRS 0.1 (0.5 to 0.2) w2(1) ¼ 0.5, P ¼ 0.498 SOFAS 0.7 (0.01 to 1.3) w2(1) ¼ 4.0, P ¼ 0.046 LIFE-RIFT 0.2 (0.4 to 0.1) w2(1) ¼ 8.5, P ¼ 0.004 Q-LES-Q 1.0 (0.1 to 1.8) w2(1) ¼ 5.2, P ¼ 0.023 CGI-I 0.01 (0.1 to 0.04) w2(1) ¼ 0.1, P ¼ 0.709 IPAQ scores in categorical MADRS 1.1 (2.2 to 0.03) w2(1) ¼ 3.7, P ¼ 0.056 BDRS 0.7 (1.9 to 0.4) w2(1) ¼ 1.5, P ¼ 0.218 SOFAS 2.3 (0.2 to 4.3) w2(1) ¼ 4.8, P ¼ 0.028 LIFE-RIFT 0.7 (1.2 to 0.1) w2(1) ¼ 5.6, P ¼ 0.018 Q-LES-Q 2.4 (0.03 to 4.8) w2(1) ¼ 3.9, P ¼ 0.047 CGI-I 0.2 (0.3 to 0.01) w2(1) ¼ 4.4, P ¼ 0.036 Physical activity categorized by WHO recommendations MADRS 0.6 (1.5 to 0.3) w2(1) ¼ 1.5, P ¼ 0.219 BDRS 0.4 (1.3 to 0.4) w2(1) ¼ 1.0, P ¼ 0.311 SOFAS 1.8 (0.3 to 3.3) w2(1) ¼ 5.9, P ¼ 0.015 LIFE-RIFT 0.7 (1.1 to 0.3) w2(1) ¼ 13.7, P < 0.001 Q-LES-Q 2.2 (0.5 to 4.0) w2(1) ¼ 6.0, P ¼ 0.014 CGI-I 0.1 (0.2 to 0.1) w2(1) ¼ 1.1, P ¼ 0.286 Note. Abbreviations: BDRS ¼ Bipolar Depression Rating Scale; CGI-I ¼ Clinical Global Impression Improvement; CT ¼ Combination Treatment; LIFE-RIFT ¼ Longitudinal Interval Follow-Up Evaluation–Range of Impaired Functioning Tool; MADRS ¼ Montgomery Åsberg Depression Rating Scale; NAC ¼ N-acetylcysteine; SOFAS: Social and Occupational Functioning Assessment Scale. Bolded p-values highlight significant values. 1.09 to 0.34) units, and mean Q-LES-Q-SF scores increased by 2.25 (95% CI, 0.45 to 4.04) units. Discussion The aim of this subanalysis of a nutraceutical RCT was to assess the relationships between PA, treatment received, and changes from baseline to Week 20 in outcomes measures for individuals with BD. Results suggest that there may be an association between PA and some of the depression and functioning outcomes of the study, but this was not consistent for all outcome measures. In regard to depression symptoms, PA was unrelated to change across the study from baseline to Week 20 on the primary outcome measure, the MADRS. However, for participants receiving CT, total PA significantly predicted changes in bipolar depression symptoms (measured by the BDRS). There was a robust relationship between participants receiving CT who exceeded WHO recommendations for PA. These participants showed a greater reduction in the BDRS depression symptoms, compared to participants receiving placebo at a similar level of PA, in a dosedependent manner; however, the differences between the groups were minimal. In contrast, participants who received NAC and engaged in higher levels of PA demonstrated a worsening of their BD symptoms, but this was not consistent across all measures. After some types of strenuous, high intensity, or endurance PA, there is evidence of a shortterm acute inflammatory response in some people39-42 that adapts over time. Inflammation is a necessary part of muscular recovery from exercise, and anti-inflammatory medication such as NAC may be inhibiting this process.43,44 There may be a delicate balance between anti-inflammatory use and benefits of exercise, potentially leading to the need for targeted and timed anti-inflammatory medication.44 As use of NAC appears to demonstrate a worsening of BD symptoms for those in a high category of the IPAQ compared to placebo, this may be a demonstration of a disruption to this delicate balance and warrants further investigation. As the CT group demonstrates improvement on this same depression scale, there is potentially an element within the CT, which is protective and counteracting the negative effects of NAC. However, due to the exploratory nature of this subanalysis and the low number of participants, cautious interpretation is required. It is possible that the combination of mitochondrialenhancing PA and the mitochondrial-enhancing CT may be an important interaction for improving bipolar depression symptoms. This is in keeping with the hypothesis that BD is at its heart a mitochondrial disorder manifested by decreased biogenesis in depression and excess energy generation in mania.21 Previous research has also found a reduction of depression (unipolar and bipolar) with PA at higher levels.45 The potential for PA in BD is profound, given its positive effects on neuroplasticity,46 hippocampal volume,47 increasing brain-derived neurotrophic factor,48 mitochondrial activity, and neurogenesis23 potentially mediated by peroxisome proliferator-activated receptor-gamma coactivator (PGC)1a.49 These are all processes that are disturbed in BD, giving rise to the possibility that PA may improve symptoms of BD via improving mitochondrial dysfunction and neuroplasticity. The additional benefits of receiving CT and engaging in higher levels of PA may be achieved via synergistic effects on the pathway regulating mitochondrial energy generation, such as PGC-1a.49,50 There were no significant relationships between participants’ PA, the treatment they received on the study and functional outcomes (LIFE-RIFT and SOFAS), quality of life (Q-LES-Q-SF), or clinician-rated improvement (CGI-I). However, there were relationships between the PA predictors and outcome measures, irrespective of what treatment they received. PA (including all variations on the scale) was a nonspecified predictor of improvement in social and occupational functioning (SOFAS), psychopathologyinduced impairment of functioning (LIFE-RIFT), and quality of life (Q-LES-Q-SF) at Week 20. These results are in keeping with previous research suggesting improved outcomes for those who engage in more PA.51 One interpretation could be a bidirectional relationship between functioning and PA. For instance, if a participant has adequate physical functioning levels, then they may have a 10 greater motivation or ability to engage in PA. However, as PA is only measured once, we cannot determine causality. Strengths of this study include the design of the doubleblind adjunctive RCT adjunctive, allowing for robust clinical trial data. PA has been measured according to a validated scale with two possible outcome measures for interpretation (continuous weekly score and categorical weekly score).33 This scale takes a conservative approach in truncating and removing data for less skew. In addition, PA has been categorized according to WHO recommendations allowing for real-world, practical interpretations and has implications for public health messages. Results of this study should be cautiously interpreted due to its limitations. In particular, the phasic nature of BD may interact with PA levels of participants. Given the scale was administered at Week 4, we cannot guarantee the phase of BD that participants were in is consistent across the sample. In addition, there is no measure of activity later in the study to assess change in participants’ level of PA. The disparity of energy expended in different states in BD highlights the potential for a bipolar-specific PA scale with populationspecific standards. In terms of the PA Scale used (IPAQ-SF), limitations exist due to the nature of self-report and can be prone to error and recall bias.52 In addition, the IPAQ considers the intensity of PA but does not record the types of exercise participants have engaged in. To reduce recall bias and to be able to review types of exercise, actigraphy could be used in addition to PA questionnaires.52 The nature of exploratory subanalyses in general poses further limitations. The RCT was powered for the primary outcome, that is, change in depression for the active treatment groups, which means the subanalysis is likely underpowered. Due to the small sample size of the data, there was insufficient power for a robust response to assess the categorical data measured from the IPAQ-SF scoring guide as nominal and as a factor within the model. PA is measured only once and as a covariate that is not directly being intervened, which limits interpretability of results. Lastly, the results presented in this subanalysis are statistically significant, but they represent small changes in outcome and thus small clinical significance. Future studies directly assessing the impact of PA programs should be powered to see greater changes in outcomes. Post hoc analyses always need to be interpreted with caution, as is the case for multiple comparisons. Conclusion and Future Directions Engaging participants to increase their activity may be a cost-effective way of improving treatment outcomes with additional health benefits for comorbid physical disorders This subanalysis of an adjunctive nutraceutical RCT adds some further support to the association between PA and mental health, and in particular, BD. PA measured at the beginning of this study was associated with functioning and quality of life at the end of the study. This subanalysis The Canadian Journal of Psychiatry suggests that measures of PA may be useful when analyzing outcomes of a new treatment. Future research may clarify the potential adjunctive effects of higher PA and mitochondrialenhancing therapies in treating bipolar depression symptoms, possibly through mitochondrial biogenesis. Authors’ Note The sponsors and funding bodies have played no role in collection, analysis, interpretation of results, or writing of the manuscript. Acknowledgments The authors would like to thank all participants of the study. The authors would also like to acknowledge the following health services involved in this study: Barwon Health, The Geelong Clinic, The Melbourne Clinic, and the University of Sydney CADE Clinic based at Royal North Shore Hospital. The authors are grateful to the Stanley Medical Research Institute, The CRC for Mental Health and the National Health and Medical Research Council for funding the study. M.M.A would like to thank Australasian Sociesty for Bipolar Depressive Disorders, Lundbeck, Australian Rotary Health, Ian Parker Bipolar Research Fund and Deakin University for scholarship support. Declaration of Conflicting Interests M.M.A. has received grant/research support from Deakin University, Australasian Society for Bipolar Depressive Disorders, Lundbeck, Australian Rotary Health, Ian Parker Bipolar Research Fund, and Cooperative Research Centre for Mental Health. A.T. has received travel or grant support from the NHMRC, Deakin University, AMP Foundation, National Stroke Foundation, Hunter Medical Research Institute, Helen Macpherson Smith Trust, Schizophrenia Fellowship NSW, SMHR, ISAD, and the University of Newcastle. M.B. has received grant support from NIH, Simons Autism Foundation, Cancer Council of Victoria, CRC for Mental Health, Stanley Medical Research Foundation, MBF, NHMRC, Beyond Blue, Geelong Medical Research Foundation, Bristol Myers Squibb, Eli Lilly, GlaxoSmithKline, Organon, Novartis, Mayne Pharma, and Servier. M.B. has been a speaker for Astra Zeneca, Bristol Myers Squibb, Eli Lilly, GlaxoSmithKline, Janssen Cilag, Lundbeck, Merck, Pfizer, Sanofi Synthelabo, Servier, Solvay and Wyeth, and served as a consultant to Astra Zeneca, Bristol Myers Squibb, Eli Lilly, GlaxoSmithKline, Janssen Cilag, Lundbeck, and Servier. C.H.N. had served as a consultant for Lundbeck, Grunbiotics, Servier, Janssen-Cilag, Wyeth and Eli Lilly, received research grant support from Wyeth and Lundbeck, and speaker honoraria from Servier, Lundbeck, Bristol-Myers Squibb, Organon, Eli Lilly, GlaxoSmithKline, Janssen-Cilag, Astra-Zenaca, Wyeth, and Pfizer. M.H. has received grant support from ISSCR, Servier, US DOD, and Bionomics, has been a speaker for Janssen-Cilag, Lundbeck, and Servier, and has been a consultant for AstraZeneca, Eli Lilly, Janssen-Cilag, Lundbeck, and Servier. J.S. has received either presentation honoraria, travel support, clinical trial grants, book royalties or independent consultancy payments from Integria Healthcare & MediHerb, Pfizer, Scius Health, Key Pharmaceuticals, Taki Mai, Bioceuticals & Blackmores, Soho-Flordis, Healthworld, HealthEd, HealthMasters, Elsevier, Chaminade University, International Society for Affective Disorders, Complementary Medicines Australia, Terry White Chemists, ANS, Society for Medicinal Plant and Natural Product Research, UBiome, 11 La Revue Canadienne de Psychiatrie Omega-3 Centre, the National Health and Medical Research Council, CR Roper Fellowship. S.D. has received grant support from Stanley Medical Research Institute, NHMRC, Beyond Blue, ARHRF, Simons Foundation, Geelong Medical Research Foundation, Fondation FondaMental, Eli Lilly, Glaxo SmithKline, Organon, Mayne Pharma, and Servier. He has received speaker’s fees from Eli Lilly, advisory board fees from Eli Lilly and Novartis, and conference travel support from Servier. J.S. has received either presentation honoraria, travel support, clinical trial grants, book royalties or independent consultancy payments from Integria Healthcare & MediHerb, Pfizer, Scius Health, Key Pharmaceuticals, Taki Mai, Bioceuticals & Blackmores, Soho-Flordis, Healthworld, HealthEd, HealthMasters, Elsevier, Chaminade University, International Society for Affective Disorders, Complementary Medicines Australia, Terry White Chemists, ANS, Society for Medicinal Plant and Natural Product Research, UBiome, Omega-3 Centre, the National Health and Medical Research Council, CR Roper Fellowship. G.S.M. has received grant or research support from National Health and Medical Research Council, Australian Rotary Health, NSW Health, Ramsay Health, American Foundation for Suicide Prevention, Ramsay Research and Teaching Fund, Elsevier, AstraZeneca, and Servier; has been a speaker for AstraZeneca, Janssen-Cilag, Lundbeck, and Servier; and has been a consultant for AstraZeneca, Janssen Cilag, Lundbeck, and Servier. S.M.C. has received grant support from the NHMRC, the Stanley Medical Research Institute, BeyondBlue, Movember, The University of Melbourne, Australian Catholic University, ARHRF, and Mental Illness Research Fund (Victoria Department of Human Services). O.M.D. is a R.D. Wright Biomedical Research Fellow and has received grant support from the Brain and Behavior Foundation, Simons Autism Foundation, Stanley Medical Research Institute, Deakin University, Lilly, NHMRC and Australasian Society for Bipolar and Depressive Disorders (ASBDD)/Servier. Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study has been funded by CRC for Mental Health, the Stanley Medical Research Institute, and an NHMRC Project Grant (APP1026307). M.M.A. is supported by Deakin University, Australasian Society for Bipolar and Depressive Disorders (ASBDD)/ Lundbeck, and Australian Rotary Health/Ian Parker Bipolar Research Fund. M.B. is supported by a National Health and Medical Research Council (NHMRC) Senior Principal Research Fellowship (APP1059660 and APP1156072). S.C. is supported by a NHMRC Senior Research Fellowship (APP1136344). J.S. is funded by an NHMRC Clinical Research Fellowship (APP1125000). W.M. is supported by Deakin postdoctoral fellowship. B.S. is supported by a Clinical Lectureship (ICA-CL-201703-001) jointly funded by Health Education England (HEE) and the National Institute for Health Research (NIHR). B.S. is part funded by the NIHR Biomedical Research Centre at South London and Maudsley NHS Foundation Trust. This paper presents independent research supported by the National Institute for Health Research (NIHR) and the views expressed are those of the author(s) and not necessarily those of the (partner organisation), the NHS, the NIHR or the Department of Health and Social Care. O.M.D. is supported by a NHMRC R.D. Wright Biomedical Research Fellowship (APP1145634). ORCID iD Michael Berk, MD, PhD https://orcid.org/0000-0002-5554-6946 Supplemental Material Supplemental material for this article is available online. References 1. Piercy KL, Troiano RP, Ballard RM, et al. The physical activity guidelines for Americans. JAMA. 2018;320(19): 2020-2028. 2. Firth J, Cotter J, Elliott R, French P, Yung AR. A systematic review and meta-analysis of exercise interventions in schizophrenia patients. Psychol Med. 2015;45(7):1343-1361. 3. Schuch FB, Vancampfort D, Richards J, Rosenbaum S, Ward PB, Stubbs B. Exercise as a treatment for depression: a metaanalysis adjusting for publication bias. J Psychiatr Res. 2016; 77:42-51. 4. Stubbs B, Vancampfort D, Hallgren M, et al. EPA guidance on physical activity as a treatment for severe mental illness: a meta-review of the evidence and Position Statement from the European Psychiatric Association (EPA), supported by the International Organization of Physical Therapists in Mental Health (IOPTMH). Eur Psychiatry. 2018;54:124-144. 5. Bauer IE, Galvez JF, Hamilton JE, et al. Lifestyle interventions targeting dietary habits and exercise in bipolar disorder: a systematic review. J Psychiatr Res. 2016;74:1-7. 6. Thomson D, Turner A, Lauder S, et al. A brief review of exercise, bipolar disorder, and mechanistic pathways. Front Psychol. 2015;6:147. 7. Ng F, Dodd S, Berk M. The effects of physical activity in the acute treatment of bipolar disorder: a pilot study. J Affect Disord. 2007;101(1-3):259-262. 8. Sylvia LG, Nierenberg AA, Stange JP, Peckham AD, Deckersbach T. Development of an integrated psychosocial treatment to address the medical burden associated with bipolar disorder. J Psychiatr Pract. 2011;17(3):224-232. 9. Sylvia LG, Salcedo S, Bernstein EE, Baek JH, Nierenberg AA, Deckersbach T. Nutrition, exercise, and wellness treatment in bipolar disorder: proof of concept for a consolidated intervention. Int J Bipolar Disord. 2013;1(1):24. 10. World Health Organization. Global recommendations on physical activity for health. Geneva, Switzerland: World Health Organization; 2010. 11. Vancampfort D, Firth J, Schuch FB, et al. Sedentary behavior and physical activity levels in people with schizophrenia, bipolar disorder and major depressive disorder: a global systematic review and meta-analysis. World Psychiatry. 2017;16(3): 308-315. 12. Vancampfort D, Stubbs B, Sienaert P, et al. A comparison of physical fitness in patients with bipolar disorder, schizophrenia and healthy controls. Disabil Rehabil. 2016;38(20): 2047-2051. 13. Vancampfort D, Hagemann N, Wyckaert S, et al. Higher cardio-respiratory fitness is associated with increased mental 12 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. The Canadian Journal of Psychiatry and physical quality of life in people with bipolar disorder: a controlled pilot study. Psychiatry Res. 2017;256:219-224. Fellendorf F, Kainzbauer N, Platzer M, et al. Gender differences in the association between physical activity and cognitive function in individuals with bipolar disorder. J Affec Disord. 2017;221:232-237. National Collaborating Centre for Mental Health. Bipolar disorder: The NICE guideline on the assessment and management of bipolar disorder in adults, children and young people in primary and secondary care. London, United Kingdom: British Psychological Society; 2018. Daussin FN, Zoll J, Dufour SP, et al. Effect of interval versus continuous training on cardiorespiratory and mitochondrial functions: relationship to aerobic performance improvements in sedentary subjects. Am J Physiol Regul Integr Comp Physiol. 2008;295(1):R264-R272. Fritzen AM, Thøgersen FB, Thybo K, et al. Adaptations in mitochondrial enzymatic activity occurs independent of genomic dosage in response to aerobic exercise training and deconditioning in human skeletal muscle. Cells. 2019;8(3):237. Taivassalo T, Shoubridge EA, Chen J, et al. Aerobic conditioning in patients with mitochondrial myopathies: physiological, biochemical, and genetic effects. Ann Neurol. 2001;50(2): 133-141. Pesta D, Hoppel F, Macek C, et al. Similar qualitative and quantitative changes of mitochondrial respiration following strength and endurance training in normoxia and hypoxia in sedentary humans. Am J Physiol Regul Integr Comp Physiol. 2011;301(4):R1078-R1087. Porter C, Reidy PT, Bhattarai N, Sidossis LS, Rasmussen BB. Resistance exercise training alters mitochondrial function in human skeletal muscle. Med Sci Sports Exerc. 2015;47(9): 1922. Morris G, Walder K, McGee SL, et al. A model of the mitochondrial basis of bipolar disorder. Neurosci Biobehav Rev. 2017;74(Pt A):1-20. Dean OM, Turner A, Malhi GS, et al. Design and rationale of a 16-week adjunctive randomized placebo-controlled trial of mitochondrial agents for the treatment of bipolar depression. Braz J Psychiatry. 2015;37(1):3-12. Sylvia LG, Ametrano RM, Nierenberg AA. Exercise treatment for bipolar disorder: potential mechanisms of action mediated through increased neurogenesis and decreased allostatic load. Psychother Psychosom. 2010;79(2):87-96. Food Drug Administration. International conference on harmonisation: guidance on statistical principles for clinical trials (ich-e9). Fed Regist. 1997;62:25692-25709. Sheehan DV, Lecrubier Y, Sheehan KH, et al. The miniinternational neuropsychiatric interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry. 1998; 59(Suppl 20):22-33;quiz 34–57. Montgomery S, Asberg M. A new depression scale designed to be sensitive to change. Br J Psychiatry. 1979;134: 382-389. 27. Berk M, Turner A, Malhi GS, et al. A randomised controlled trial of a mitochondrial therapeutic target for bipolar depression: mitochondrial agents, N-acetylcysteine, and placebo. BMC Med. 2019;17(1):18. 28. Berk M, Dodd S, Dean OM, Kohlmann K, Berk L, Malhi GS. The validity and internal structure of the Bipolar Depression Rating Scale: data from a clinical trial of N-acetylcysteine as adjunctive therapy in bipolar disorder. Acta Neuropsychiatr. 2010;22(5):237-242. 29. Morosini PL, Magliano L, Brambilla L, Ugolini S, Pioli R. Development, reliability and acceptability of a new version of the DSM-IV Social and Occupational Functioning Assessment Scale (SOFAS) to assess routine social functioning. Acta Psychiatr Scand. 2010;101(4):323-329. 30. Leon AC, Solomon DA, Mueller TI, Turvey CL, Endicott J, Keller MB. The range of impaired functioning tool (LIFERIFT): a brief measure of functional impairment. Psychol Med. 1999;29(4):869-878. 31. Spearing MK PR, Leverich GS, Brandt D, Nolen W. Modification of the clinical global impressions (CGI) scale for use in bipolar illness (BP): the CGI-BP. Psychiatry Res. 1997;73: 159-171. 32. Endicott J, Nee J, Harrison W, Blumenthal R. Quality of Life Enjoyment and Satisfaction Questionnaire: a new measure. Psychopharmacol Bull. 1993;29(2):321-326. 33. Craig CL, Marshall AL, Sjostrom M, et al. International physical activity questionnaire: 12-country reliability and validity. Med Sci Sport Exerc. 2003;35(8):1381-1395. 34. Soundy A, Roskell C, Stubbs B, Vancampfort D. Selection, use and psychometric properties of physical activity measures to assess individuals with severe mental illness: a narrative synthesis. Arch Psychiatr Nurs. 2014;28(2): 135-151. 35. Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Personal Social Psychol. 1986; 51(6):1173-1182. 36. Kraemer HC, Stice E, Kazdin A, Offord D, Kupfer D. How do risk factors work together? Mediators, moderators, and independent, overlapping, and proxy risk factors. Am J Psychiatry. 2001;158(6):848-856. 37. White H. A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica. 1980;48(4):817-838. 38. IBM Corp. IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY: Author; 2017. 39. Lira FS, Dos Santos T, Caldeira RS, et al. Short-term high- and moderate-intensity training modifies inflammatory and metabolic factors in response to acute exercise. Front Physiol. 2017; 8:856-856. 40. Monteiro PA, Campos EZ, De Oliveira FP, et al. Modulation of inflammatory response arising from high-intensity intermittent and concurrent strength training in physically active males. Cytokine. 2017;91:104-109. 41. Kaspar F, Jelinek HF, Perkins S, Al-Aubaidy HA, De Jong B, Butkowski E. Acute-phase inflammatory response to single- La Revue Canadienne de Psychiatrie 42. 43. 44. 45. 46. 47. bout HIIT and endurance training: a comparative study. Mediators Inflamm. 2016;2016:5474837-5474837. Zwetsloot KA, John CS, Lawrence MM, Battista RA, Shanely RA. High-intensity interval training induces a modest systemic inflammatory response in active, young men. J Inflamm Res. 2014;7:9-17. Peake JM, Neubauer O, Della Gatta PA, Nosaka K. Muscle damage and inflammation during recovery from exercise. J Appl Physiol (Bethesda, MD: 1985). 2017;122(3):559-570. Chazaud B. 2016. Inflammation during skeletal muscle regeneration and tissue remodeling: application to exercise-induced muscle damage management. Immunol Cell Biol. 94(2):140-145. Schuch FB, Vancampfort D, Firth J, et al. Physical activity and incident depression: a meta-analysis of prospective cohort studies. Am J Psych. 2018;175(7):631-648. Phillips C. Physical activity modulates common neuroplasticity substrates in major depressive and bipolar disorder. Neural Plasticity. 2017;2017:7014146-7014146. Firth J, Stubbs B, Vancampfort D, Schuch F, Lagopoulos J, Rosenbaum S, Ward PB. Effect of aerobic exercise on hippocampal volume in humans: a systematic review and meta-analysis. Neuroimage. 2018;166:230-238. 13 48. Kerling A, Kuck M, Tegtbur U, et al. Exercise increases serum brain-derived neurotrophic factor in patients with major depressive disorder. J Affect Disord. 2017;215: 152-155. 49. Granata C, Jamnick NA, Bishop DJ. Principles of exercise prescription, and how they influence exercise-induced changes of transcription factors and other regulators of mitochondrial biogenesis. Sport Med. 2018;48(7):1541-1559. 50. Nierenberg AA, Ghaznavi SA, Sande Mathias I, Ellard KK, Janos JA, Sylvia LG. Peroxisome proliferator-activated receptor gamma coactivator-1 alpha as a novel target for bipolar disorder and other neuropsychiatric disorders. Biol Psychiatry. 2018;83(9):761-769. 51. Sylvia LG, Friedman ES, Kocsis JH, et al. Association of exercise with quality of life and mood symptoms in a comparative effectiveness study of bipolar disorder. J Affect Disord. 2013; 151(2):722-727. 52. Paul DR, McGrath R, Vella CA, Kramer M, Baer DJ, Moshfegh AJ. Understanding the nature of measurement error when estimating energy expenditure and physical activity via physical activity recall. J Phys Activ Health. 2018;15(7):543-549.