Aerobic exercise in severe mental illness: requirements from the perspective of sports medicine

Major depression, bipolar disorder, and schizophrenia are severe mental illnesses. Despite receiving psychopharmacological and psychosocial treatments, about half of patients develop a chronic course with residual cognitive and negative symptoms and have a high risk for cardiovascular disease and reduced life expectancy. Therefore, add-on innovative treatment approaches are needed to improve outcome. Aerobic exercise interventions have been shown to improve global functioning, cognition, and negative and depressive symptoms in these patients. The basic mechanism of these exercise-related changes has been reported to be improved brain plasticity, e.g., increased volume of disease-related brain regions such as the hippocampus. The optimal type, duration, and frequency of exercise have not yet been determined and need to be addressed in supervised physical exercise studies. Because of the low physical activity levels, lack of drive related to negative and depressive symptoms, and high prevalence of cardiovascular comorbidities in patients with severe mental illness, besides aiming to improve symptoms of mental illness, exercise interventions should also aim to increase cardiorespiratory fitness, which they should comprehensively assess by direct measurements of maximal oxygen uptake. Based on the recommendations for developing cardiorespiratory fitness by the American College of Sports Medicine, 150 min moderate-intensity training per week or vigorous-intensity exercise training for 75 min per week are appropriate. Most studies have had relatively short intervention periods, so future studies should focus on long-term adherence to exercise by implementing motivational strategies supported by telemedicine and by identifying and targeting typical barriers to exercise in this patient population.


Introduction
In 2010, more than 25% of the population in Europe was diagnosed with a mental illness [1]. Among mental illnesses, major depression (MDD), bipolar disorder (BD) and schizophrenia (SZ) belong to the 20 most burdensome disorders and result in annual costs of 207 billion euro in Europe. Thus, besides cardiovascular diseases [2,3], mental disorders-especially MDD-are one of the leading illnessrelated causes of years lived with disability worldwide [4]. Furthermore, more than 50% of patients in Europe with a mental illness, equivalent to about 25 million Europeans, develop a relapsing, chronic course of their illness with residual symptoms, which are associated with poor functional outcome [5].
Epidemiological research showed that in the last 100 years the long-term outcome of SZ has remained relatively stable [6], despite the introduction of psychotropic medications 70 years ago. Furthermore, psychotropic medications also have not improved outcome dimensions like cognition or negative symptoms. Overall, about 20% of patients with SZ have a good outcome and are able to participate in the primary job market, maintain a stable partnership and have longer phases with no need for psychopharmacological support [7,8]; about 30% of patients will have a good prognosis but will need to take long-term psychotropic medication, which can have burdensome side effects, especially related to metabolic syndrome [9]; and about 50% of patients will develop a chronic course (defined as continuous symptoms over a period of at least 2 years) with different degrees of residual symptoms and disability [5]. Residual symptoms include cognitive dysfunction, impaired mood, reduced drive, and reduced ability to cope with stress. These individuals have an unfavorable long-term social outcome, meaning that they have difficulties finding a long-term job on the primary job market or maintaining a stable partnership [8].
Besides having a direct influence on functional outcome, environmental risk factors also affect mortality: The mortality rate from physical causes, including suicide is 20-fold higher in patients with unipolar depression than in the general population, 15-fold higher in patients with BD, and 12-fold higher in patients with SZ [10][11][12][13]. Another important cause is the high incidence of medical comorbidities due to unhealthy lifestyle habits, such as high rates of cigarette smoking and low levels of physical activity. Estimates indicate that together suicides and lifestyle-related factors reduce life expectancy by nearly 10 to 20 years compared with the general population [14,15].
A meta-analysis of data from 29 countries on 6 continents confirmed that people with mental disorders have a significantly higher mortality rate and that, in 65 studies, the highest mortality rate (relative risk 2.54; 95% CI 2.35-2.75) was among patients with psychosis [15][16][17][18]. People with severe mental illness have a higher risk of developing coronary heart disease than controls (adjusted hazard ratio 1.54; 95% CI 1.30-1.82) and a higher rate of autonomic nervous system dysfunction, including diminished heart rate variability, hypertension, alterations of the QT interval, and lipid pattern abnormalities [19]. Unfortunately, the currently available psychopharmacological and psychosocial treatments do not ameliorate or improve these symptoms significantly and therefore do not help to improve functional outcome or increase life expectancy.
Surprisingly few studies have examined whether aerobic exercise combined with diet and psychosocial interventions can reduce the mortality gap between patients with mental disorders and the general population, and the studies included only small samples of patients [20]. The seminal paper by the Lancet Commission [21] identified an unhealthy diet and a sedentary lifestyle as major risk factors for the physical illness burden of patients with MDD, BD, and SZ, diseases that are also collectively referred to as affective and nonaffective psychoses. The paper gives recommendations to influence these modifiable risk factors, e.g., including regularly exercise in treatment programs.
This qualitative review will outline the effects of aerobic exercise on clinical outcome in patients with MDD, BD, and SZ and will evaluate the quality of intervention studies on the basis of the requirements of modern sports medicine.

Physical activity, physical fitness, and medical health outcome in patients with MDD, BD, and SZ
Low physical activity levels and poor cardiorespiratory fitness (CRF) are associated with a high risk of cardiovascular disease and all-cause mortality [22]. Exercise and good CRF play an important role in mitigating cardiovascular disease risk factors, such as metabolic syndrome, which is defined as a combination of increased waist circumference; elevated fasting glucose, triglycerides, and low high-density lipoprotein cholesterol; and high blood pressure [23,24]. A metaanalysis determined that the risk of metabolic syndrome was elevated in all patients with affective and nonaffective psychoses (32.6%, 95% CI 30.8-34.3%) and that the prevalence did not differ between patients with MDD, BD, or SZ [25].
Another meta-analysis found that patients with MDD had lower levels of physical activity (standardized mean differenc, − 30; 95% CI − 0.40 to 0.21) and higher levels of sedentary behavior (standardized mean difference 0.99; 95% CI 0.01-0.18) than healthy controls [26]. Therefore, researchers concluded that less physically active patients might benefit from specific aerobic exercise interventions aimed at increasing physical fitness [27]. An important aspect to consider in this context is that low physical activity is related to negative symptoms such as amotivation [28], so aerobic exercise interventions must be supervised by experienced sports scientists to ensure that patients adhere to the intervention [29].
Physical fitness and physical activity are low not only in patients with MDD, but also in those with BD [30]. However, this diagnostic group is highly underrepresented in physical activity studies.
A study in patients with SZ showed that low physical fitness was associated with a higher prevalence of metabolic syndrome and more severe cognitive, negative, and positive symptoms [27]. The exercise capacity (measured by the distance covered in the 6-min walking test) of patients with SZ and prediabetes was reduced and the body mass index was increased; in addition, patients with SZ and manifest type 2 diabetes were less physically active [31].
Previous aerobic exercise studies showed the feasibility of endurance training in patients with SZ, and adaptations to aerobic endurance training in patients were comparable to those in healthy controls, as assessed by physical working capacity and maximal achieved power. However, differences were detected in changes of performance at a lactate concentration of 3 mmol/L, i.e., patients with SZ showed an impaired increase in lactate [29].
Different types of interventions have been evaluated that aim to improve physical health in patients with mental illness. A meta-analysis of 47,231 patients with SZ summarized and compared the effects of pharmacological and nonpharmacological interventions [32]. The authors showed that the most effective interventions for weight reduction were individual lifestyle counseling and exercise interventions, followed by psychoeducation, augmentation with the atypical antipsychotic aripiprazole, topiramate add-on therapy, and dietary interventions. The best efficacy in reducing glucose levels was found for a switch from olanzapine (the atypical antipsychotic with the highest risk for metabolic syndrome) to aripiprazole and add-on medication with metformin. Efficacy was also shown for treatment with glucagon-like peptide-1-receptor agonists, dietary interventions, and aripiprazole augmentation. Insulin resistance improved best followed by metformin treatment. Metformin also had the greatest effects on total cholesterol and high-density lipoprotein cholesterol. The best effect on triglycerides and low-density lipoprotein cholesterol was achieved with topiramate. Importantly, only exercise interventions increase exercise capacity [32]. Recent efforts to increase the efficacy of exercise include the use of high-intensity interval training (HIIT). In a randomized controlled HIIT study, compliant patients with overweight and SZ showed improvement in waist circumference, negative symptoms, and psychosocial functioning [33]. HIIT may be a feasible and effective way to improve CRF and metabolic parameters and has been established as such in physical disorders. It may also have more beneficial effects on the metabolic state than more moderate and continuous endurance training methods [20].

Effects of aerobic exercise on symptoms of MDD
Aerobic exercise, often revered to as "endurance exercise", is defined as physical activity with a predominant metabolic pathway that uses oxygen to meet energy demands (oxidative phosphorylation) and leads to only low blood lactate levels [34]. In practice, aerobic exercise is usually characterized by repeated sequences of physical activity in a light to moderate intensity for extended periods of time. Aerobic exercise improves especially CRF and includes typically activities such as walking, swimming or cycling [35]. Contrastingly, anaerobic exercise refers to short-term high-intensity efforts with a preponderance of metabolic pathways not using oxygen (phosphagens metabolic pathway and glycolytic pathway) [36]. In most studies, the term "anaerobic training" is used to describe high-intensity exercise intervals with a duration of up to several minutes that result in increased lactate levels.
Aerobic exercise studies addressing MDD are summarized in Table 1. An aerobic exercise training study of 12 × 75-min sessions over a period of 4 weeks in patients with SZ and MDD revealed improvements in cognition, which were more pronounced in the patients with SZ; however, the patients with MDD showed a greater reduction in depressive symptoms and anxiety [37]. Aerobic exercise studies in MDD showed that exercise improves working memory and psychosocial functioning and reduces depressive symptoms [38,39]. In particular, one study provided evidence for an effect of aerobic exercise on remission in MDD by showing that 29.5% of patients with unremitted MDD remitted after 3 months of aerobic exercise treatment [40].
In a randomized, controlled trial, 50 min' add-on supervised aerobic exercise training 3 times a week for 4 months decreased symptoms of depression, anxiety, and stress compared with pharmacotherapy with antidepressants [41]. After an 8-week walking or running aerobic exercise program in local sports clubs, patients with MDD showed a large reduction in depressive symptoms compared with patients on a waiting list [42]. Moreover, an 8-week study found that highfrequency exercise was superior to low-frequency exercise with respect to depressive symptoms [43]. In an unsupervised study of physical activity in patients with MDD given access to fitness center resources, an increase in moderateto-vigorous activity was associated with improvements in depressive symptoms [44]. In an 8-week study both aerobic and non-aerobic training methods had favorable effects on depression scores [45].
Different types of exercise have been studied in the last decade. A meta-analysis revealed small effects of aerobic exercise and yoga in outpatients with MDD, whereas the effects of Tai Chi were insufficient to enable conclusions to be drawn [46]. Additionally, aerobic exercise was superior to basic body awareness therapy with respect to depressive symptoms and cardiovascular fitness [47]. In patients with MDD randomized to 4 weeks' sprint interval training or continuous aerobic exercise training, improvements in CRF were observed in both groups and were associated with improved depressive symptoms, emotional wellbeing, and sleep [48]. In contrast, another study found no improvements in depression score in the Hamilton Rating Scale for Depression after a 4-month strength and aerobic exercise training in patients with MDD [49]. Using mendelian randomization methods on genomic and phenotypic data from the UK biobank, beneficial effects of exercise were detected in depression but not in SZ [50,51] (Table 1).

Effects of physical activity on symptoms of BD
No interventional studies have examined the effects of exercise in patients with BD. However, in a study examining the effects of N-acetylcysteine treatment, physical activity was not related to improvements in depressive symptoms, although those participants who engaged in higher levels of physical activity had greater improvements in social and occupational functioning [52].

Aerobic exercise training improves cognition and symptoms in patients with SZ
Several studies have demonstrated beneficial effects of physical exercise on symptoms of SZ (Table 2). For example, a well-cited meta-analysis showed that in patients with SZ aerobic exercise improves negative, positive, and depressive symptoms and global functioning, as measured by the Global Assessment of Functioning (GAF) score [53].
In addition, another meta-analysis focusing on cognition demonstrated improved global cognition, working memory, social cognition, and attention after aerobic exercise in patients with SZ [54]. A recent meta-analysis of randomized controlled trials found that aerobic exercise had small beneficial effects on negative symptoms in patients with SZ [55]. Across aerobic exercise studies, symptom improvement was seen in interventions consisting of 90 min of moderate exercise per week [56]. This finding is in line with our own work, which showed that 3 × 30 min of aerobic exercise per week alleviated negative symptoms and significantly improved global functioning and short-term memory in patients with SZ [57,58]. Moreover, we found preliminary evidence that the improvements in level of functioning might be sustained even after exercise cessation [59]. Effects of aerobic exercise on cognition have been observed also in patients with first-episode SZ. After a 12-week supervised circuit-training program, improvement was seen in processing speed, visual learning, and visual attention domains [60]. In 75 patients with SZ randomized to 12 weeks of either moderate-intensity treadmill exercise or stretching and toning exercise, aerobic exercise improved processing speed and attention [61]. However, in a pilot randomized controlled trial in a small sample, group aerobic exercise over 12 weeks showed similar improvements in cognition and symptoms as treatment as usual [62]. After patients with SZ performed 12 weeks of treadmill training, their general and psychopathology and total score on the Positive and Negative Symptom Scale (PANSS) and aerobic capacity improved [63]. In a 6-month randomized study comparing aerobic exercise with occupational therapy in patients with SZ, exercise reduced symptoms, depression, and need of care and increased cardiovascular fitness [64]. In contrast, after a single session of aerobic exercise and yoga, patients from both groups showed only decreased anxiety and psychological stress and increased subjective wellbeing [65]. In meta-analyses of meditation-based mind-body interventions, small effect sizes have been observed for yoga in SZ [66]. Besides yoga, other exercise interventions such as Tai Chi have been applied in SZ patients and led to improvements in PANSS score, negative symptoms, and aggressive behavior [67]. A study that compared a 12-week Tai Chi program with aerobic exercise showed improved negative and depression symptoms [68]. In a randomized 8-month study of a Greek traditional dancing program, the dancing group showed improved positive and negative symptoms, GAF score, and quality of life compared with a sedentary group [69]. Finally, resistance training was studied in patients with SZ and improved negative symptoms [70] and level of functioning assessed with the GAF [71] (Table 2).

Neuroplasticity effects of aerobic exercise
Animal models and basic research in humans clearly show that aerobic exercise has favorable neurobiological effects. These effects may involve epigenetic alterations, synaptic plasticity, differentiation of glial cells and neurogenesis, the hypothalamus-pituitary-adrenal axis, growth factors, immune-related mechanisms, neurotransmitters, and the endocannabinoid system [72]. In 2103 adults from the general population, CRF, measured as peak oxygen uptake (VO2 peak ), was related to higher gray matter volume and showed a strong association with gray matter volume of the left middle temporal gyrus, right hippocampus, left orbitofrontal cortex, and bilateral cingulate cortex [73]. A metaanalysis of hippocampal volume in 737 voluntary participants revealed significant positive effects of aerobic exercise on left hippocampal volume but not on total hippocampus volume [74]. These results may be relevant for MDD, BD, and SZ because these brain disorders have been repeatedly shown to involve structural and functional alterations in the hippocampal formation [75,76]. Moreover, a 7-Tesla magnetic resonance imaging study in older adults found a prominent volume increase in the left cornu ammonis (CA) subregions of the hippocampus and a trend for a volume increase in the left CA4/dentate gyrus after physical activity [77].
Deficits in both episodic and working memory are related to hippocampal abnormalities and are hallmarks of an unfavorable outcome in MDD [78] and SZ [79]. Our first study to investigate the effects of aerobic endurance training in a small sample of patients with multi-episode SZ showed a significant 10% increase in hippocampal volume after 3 months [57]. In our subsequent study in 20 patients, which combined 3 months of aerobic endurance training with cognitive remediation, we found no changes in hippocampal volume in the exercise group [58], but we did find a significant             correlation between exercise-related volume increases in the CA4/dentate gyrus subregion of the hippocampus and the SZ polygenic risk score (SZ-PRS, [80]). Using cell-specific PRS, we found that this volume effect in CA4 was also caused by oligodendrocyte precursor cell-related pathways [81], which is also in line with our post-mortem finding of reduced oligodendrocyte number in the CA4 subregion [82].
In SZ and MDD it has been hypothesized that metabolic coupling may link oligodendrocyte to interneuron pathology [83]. Other studies found no changes in total hippocampal volume after aerobic exercise in MDD [84] or SZ [85]. However, after a 12-week aerobic exercise training, hippocampal volume in the CA1 subregion increased in SZ patients, whereas hippocampal vascular volume was unchanged, indicating no effect of aerobic exercise on blood vessels [86].
Additionally, a study that compared aerobic exercise training with table soccer in patients with SZ and healthy controls showed an increased volume of the right entorhinal cortex compared with baseline after 6 weeks' training [87] and of the left superior, middle, and inferior anterior temporal gyri after 3 months' training; but patients with SZ who played table soccer showed increased volumes in the motor and anterior cingulate cortices [88]. After 6 weeks' aerobic exercise training, a magnetic resonance spectroscopy study in patients with SZ found increased N-acetyl-aspartate/total creatine levels in the left dorsolateral prefrontal cortex in both the aerobic exercise and table soccer groups [89], indicating improved neuronal viability. Additionally, a 6-month aerobic exercise program improved the integrity of motor function-related white matter fiber tracts compared with a life-as-usual condition [90]. Taken together, these results indicate that in SZ exercise has neuroplastic effects in brain regions that are affected by the disease itself. The effects of aerobic exercise on brain volume changes and underlying mechanisms warrant further study, not only in patients with SZ but also in those with MDD and BD.

Improvements of CRF in patients with severe mental illness
CRF is an important marker of cardiovascular health and should be comprehensively assessed in both clinical studies and clinical practice [22]. Especially in patients with severe mental illness and negative symptoms such as reduced drive and motivation, CRF serves as a control for the efficacy of an exercise intervention. Because of the above-mentioned low activity levels and high prevalence of cardiovascular comorbidities in patients with severe mental illness, besides aiming to improve symptoms of mental illness exercise interventions should also aim to increase CRF [91].
Besides using different measurement methods, studies differ regarding the training modalities. To date, the effects of anaerobic exercise interventions have been investigated only scarcely. However, there are studies that combined aerobic and anaerobic training elements [90]. In addition, there is a growing body of studies investigating HIIT [33,47,[98][99][100]104], which is typically characterized by highintensity exercise at 4 × 4 min intervals (85-95% of maximum heart rate [HR max ]), with active breaks consisting of 3 min of moderate-intensity exercise (approximately 70% of HR max ). HIIT was shown to be effective in improving CRF in patients with SZ. A small, 8-week study showed that VO2 peak increased by 12% in the HIIT group (n = 12) but did not increase in the PC gaming group (n = 7) [104]. These results were confirmed by a recent randomized controlled trial on the effects of 12 weeks' HIIT on VO2 max in 21 patients with SZ. Like the study by Heggelund et al. [104], in this study the control group (n = 26) practiced their PC gaming skills. Although more than half of the patients in the HIIT group showed a significant increase in workload, a significant within-group difference in VO2 max was only observed when the physical activity competence of the health care providers was added into the statistical model. This result underlines the importance of professional and experienced supervision when aiming to successfully improve CRF in patients with SZ [99]. The study findings are supported by a similar study in which the training group (n = 25) performed aerobic interval training and received professional adherence support twice a week over the 12-week intervention period. The patients' VO2peak improved by 10%, while no change was observed in the control group (n = 23), who performed two supervised exercise sessions at the beginning of the study and were subsequently encouraged to continue exercising on their own [100].
To the best of our knowledge, only two studies have evaluated a form of interval training in patients with MDD.
Gerber et al. [98] found associations between an increase in VO2 max and improvements of symptoms in patients who performed a sprint interval training consisting of 25 repetitions of 30-s high-intensity bursts at 80% of maximal power output, followed by 30 s of total rest. Danielsson et al. [47] reported a significant increase of CRF in the intervention group, which performed intervals of exercise at higher perceived intensity during the aerobic exercise program, although training intensity was not strictly defined.
Exercise training of moderate intensity can also be effective in improving maximal oxygen consumption in patients with MDD [45,92,97] and SZ [29,57,63,85,105]. For example, in the most recent study in SZ, VO2 max improved by 18% in patients after a 12-week aerobic exercise program with intensities ranging from 60 to 75% of HR max (n = 16) but decreased by − 0.5% in the usual care group (n = 17) [105].
Mixed programs consisting of aerobic training combined with resistance training may also have the potential to improve CRF in patients with MDD [49,94] and SZ [85,102]. Although three studies measured CRF directly by cardiopulmonary exercise testing [49,85,94,106], Korman et al. [102] used a submaximal test (they assessed functional exercise capacity, a submaximal proxy measure of CRF, as the distance walked during the 6-min walking test). Moreover, two studies that evaluated the effect of dancing programs in SZ showed improvements in performance during the 6-min walking test [69] or the 3-min step test [107]. Overall, little evidence is available on the effects of mixed programs, so further studies are needed that use clearly defined exercise programs and high-quality CRF measurements.
In summary, in patients with severe mental illness aerobic exercise, especially endurance training, has shown beneficial effects on global functioning, cognition, and negative and depressive symptoms. It stimulates synaptic and brain plasticity and affects the volume of specific brain regions, with genetic risk (SZ-PRS) influencing the results. However, despite the growing body of literature, the type, duration, and frequency of exercise needed for beneficial effects in the long term have yet to be determined before aerobic exercise will be used widely in general practice [109]. Some recommendations for further studies can be given from the perspective of sports medicine: • The American College of Sports Medicine recommends that adults engage in moderate-intensity cardiorespiratory exercise training for ≥ 30 min/day on ≥ 5 days/ week for a total of ≥ 150 min/week, or vigorous-intensity cardiorespiratory exercise training for ≥ 20 min/day on ≥ 3 days/week for a total of ≥ 75 min/week [110]. • Further studies are needed to identify the most effective exercise interventions (type, duration, frequency).
• Because most studies were conducted over a relatively short intervention period of 3 to 4 months, were supervised and did not include a follow-up, future studies should focus on long-term adherence to exercise (e.g., by implementing motivational strategies supported by telemedicine and apps and by identifying and targeting typical barriers to exercise in this patient population). • This patient group has a high prevalence of cardiovascular disease, so researchers should consider measuring the associated risk factors when performing exercise intervention studies. • CRF should be comprehensively assessed in both clinical studies and clinical practice by direct measurements of maximal oxygen uptake. • Exercise interventional studies in patients with BD need to be conducted because this patient group is underrepresented.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.