Effects of combined exercise training for adults with resistant major depression: a pilot study from the TRACE-RMD project

Introduction

Major depressive disorder (MDD)—a complex, multifactorial, and multigenic condition—is considered one of the most common types of depressive disorders worldwide (Cichon et al., 2009). Even among patients with MDD receiving adequate treatment (i.e., the use of psychotherapy, a variety of antidepressants, and combined strategies of therapeutic enhancement), only 30% experience full recovery or remission (Trivedi et al., 2006). Patients whose depressive disorder does not respond satisfactorily to adequate treatment have harder-to-treat depression, generally referred to as resistant major depression (RMD) (Thase & Rush, 1997).

People with RMD generally practice an unhealthy lifestyle characterised by a lack of physical activity (PA) (Vancampfort et al., 2017), low cardiorespiratory fitness (CRF), tobacco use, high blood pressure, and overweight/obesity. In addition, the side effects of pharmacological treatments contribute to the appearance of different diseases, a high prevalence of metabolic and inflammatory dysregulation, as well as decreased life expectancy and quality of life (QoL) (ten Doesschate et al., 2010). Thus, there seems to be a bidirectional relationship, with RMD acting as both a cause and a consequence of physical inactivity (Howland, 2010; Mura et al., 2014). Savitz et al. (2025) recently reported that adults with depression who have high systemic levels of inflammation (defined by serum C-reactive protein (CRP) > 3 mg/L) may be at an increased level of anhedonia and worsening of clinical symptoms. Consistently, there is an evolutionary framework that links depression and chronic low-grade inflammation. It proposes that modern sedentary lifestyles diverge significantly from the PA patterns of our ancestors and contribute to increased vulnerability to mental health disorders (Carrera-Bastos et al., 2025). Insufficient PA plays a central role in the dysregulation of immune function, promoting systemic inflammation that is closely associated with the pathophysiology of MDD (Schuch et al., 2016a). Drawing on evolutionary biology, regular PA, particularly exercise aligned with basic movement patterns, might act as a powerful anti-inflammatory intervention (Carrera-Bastos et al., 2025).

Given this scenario, various non-pharmacological strategies have been explored as adjuncts to pharmacological treatment to enhance the prognosis and remission rates of MDD, including exercise and electroconvulsive therapy interventions (Howland, 2010; Mura et al., 2014). According to the latest World Health Organization guidelines on PA and sedentary behaviour, recommendations for people living with a disability (including those with RMD) do not differ from those of the general elderly population (i.e., a minimum of 150 min per week of moderate-to-high intensity activity or 75 min of high-to-severe intensity activity across the week, as well as at least three days a week of varied multicomponent PA that prioritise functional balance and moderate-to-higher resistance training) (Bull et al., 2020).

Based on a recent systematic review, in previous studies on populations with RMD, the exercise programmes consisted of 1-h sessions twice a week—typically involving aerobic exercises on a cycle ergometer or treadmill—over a period exceeding 10 weeks. These programmes have been shown to improve several parameters of depression and functioning based on the Hamilton Depression Rating Scale, the Beck Depression Inventory, the Clinical Global Impressions Scale (CGI-S), and the Global Assessment of Functioning (Etxaniz-Oses et al., 2024). In one study, patients in an exercise group had a 26% remission rate, compared with 0% in the control group (Mota-Pereira et al., 2011). Only one study incorporated a combined regimen of aerobic and resistance training (Mather et al., 2002). Of note, none of the cited studies controlled individualised exercise intensity.

Although general PA recommendations encourage the implementation of a moderate-to-high intensity exercise to obtain more significant health benefits, previous research has demonstrated that low-to-moderate intensity training effectively enhances various outcomes, such as CRF and body mass index (BMI), particularly in adults with chronic illnesses and a low starting level of aerobic fitness (Horváth et al., 2022). Moreover, in one study, it improved overall well-being and helped fill the 3-week gap that antidepressants require before showing therapeutic effects (Schuch et al., 2016b). Thus, low-intensity aerobic interval training (LIIT) exercise (i.e., alternating loads of light-to-moderate intensity exercise) could be a useful and effective tool for people with RMD, especially those with low CRF (De Oliveira et al., 2016). However, to our knowledge, there is no scientific evidence regarding the effects of an exercise programme combining LIIT and resistance training in people with RMD. Therefore, we aimed to determine changes in body composition, CRF, clinical symptoms, the biochemical profile, and QoL following a 12-week combined exercise intervention for adults with RMD. We hypothesised that clinical, physical, and physiological variables concomitantly improve after the intervention.

Materials & Methods

Study design

The OSI Araba Research Ethics Committee (PI2021045) approved this pilot study, and all participants provided written informed consent before any data were collected. This study was registered with the International Standard Randomised Controlled Trial (Code NCT 05136027). Ethical approval for this study was granted by a local ethics committee on 7 May 2021 (Certificate No. 2021-045), in accordance with the Declaration of Helsinki. All measurements were taken before and after the 12-week intervention period, during which time the participants were involved in a supervised exercise group (Fig. 1).

Results

At baseline, the participants showed some cardiovascular risk factors (Table 1): a BMI > 25 kg/m², indicative of overweight; 34.8% fat mass, indicative of obesity; deficient CRF (V̇O_2peak = 17.3 ± 6.1 mL kg⁻¹ min⁻¹ for women and 25.7 ± 10.2 mL kg⁻¹ min⁻¹ for men); an upper optimal LDL-C concentration > 100 mg/dL; an atherogenic index (TC/HDL-C) > 3.5; fasting plasma glucose > 100 mg/dL; and a proinflammatory status with CRP > 3 mg/L and IL-1ß < 0.5 pg/mL. Conversely, the participants had normal blood pressure (123/79 mmHg), HDL-C (>40 mg/dL), TC (<200 mg/dL), TGs (<200 mg/dL), insulin (<16.7 μU/mL), HOMA-IR (<3.8), liver enzymes (AST <30 U/L, ALT <30 U/L, and GGT <50 U/L), testosterone (0.2 (0.02) ng/dL for women and 3.8 (1.3) ng/dL for men), IL-10 (<7 pg/mL), and IL 6 (<9 pg/mL). Overall, at baseline the participants were metabolically unhealthy with overweight/obesity according to the Wildman-modified criteria, considering glucose and CRP levels (Martínez-Larrad et al., 2014), and had low CRF.

Table 2 presents the psychiatric and QoL characteristics at baseline. The participants had moderate depression based on the MADRS (29.1) and were moderately ill based on the CGI-S (4.2). The SDS score (25.0 ± 6.0) indicated high disability (83.3%), considering that a value of 30 represents a 100% impact on functionality. Furthermore, the participants had low QoL values for the physical (<50) and mental (<50) component summary analyses of the SF-36. Most participants were taking one (41.2%) or two (41.2%) antidepressants (Table 2) combined with antipsychotics (29.4%) and stabilisers (58.8%).

After the intervention, four participants dropped out for logistical (not clinical) reasons. Therefore, the total sample for statistical analysis after the intervention was 14. There were no significant changes (P > 0.05) in the body composition variables, submaximal and peak physiological variables, or the biochemical profile, except for haemoglobin (mean difference = 0.64 mg/dL, 95% CI [0.18–1.10] mg/dL, P = 0.010,), haematocrit (mean difference = 2.21 mg/dL, 95% CI [0.92–3.51] mg/dL, P = 0.003), and AST (mean difference = 3.07 mg/dL, 95% CI [0.62–5.52] mg/dL, P = 0.018). After the intervention, there was a significant reduction in the clinical symptoms of depression based on the MADRS (mean difference = −8.31, 95% CI [−15.1 to −1.5], P = 0.021), the CGI-S (mean difference = −1.17, 95% CI [−1.92 to −0.41], P = 0.006), and the SDS (mean difference = −5.46, 95% CI [−10.8 to −0.12], P = 0.046) (Fig. 2). Regarding changes in the domains of health-related QoL (Fig. 3), after the intervention there was an increase in general health (mean difference = 13.8, 95% CI [2.9–24.8], P = 0.017), vitality (mean difference = 10.4, 95% CI [0.26–20.5], P = 0.045), social functioning (mean difference = 25.9, 95% CI [8.1–43.7], P = 0.008), and the mental component summary (mean difference = 6.9, 95% CI [1.1–12.8], P = 0.024).

Discussion

The present pilot study is unique in that it investigated the effects of a supervised, combined exercise training programme (LIIT + resistance exercise training) in adults diagnosed with RMD. Based on the preliminary findings, the combined exercise programme induced positive and beneficial changes in functionality, clinical symptoms (based on the MADRS, CGI-S, and SDS), and QoL. However, there were no associated changes in body composition, submaximal and peak physiological exercise-related variables, or some of the analysed biochemical variables.

For patients with RMD, exercise interventions have been considered a non-pharmacological treatment to help improve prognosis and remission rates of MDD (Howland, 2010; Mura et al., 2014). In the present pilot study, the participants showed improvement in all assessed depression and functioning parameters (Figs. 2 and 3). These results are consistent with previous reports concerning the positive effects of exercise on patients with RMD (Mota-Pereira et al., 2011) and MDD (Trivedi et al., 2006). Treatment with antidepressants has been shown to increase the volume of the hippocampus, anterior cingulate, and orbitofrontal cortex; to improve white matter integrity; and to alter the functional dynamics of frontal-limbic neural networks in adults with depression (Singh & Gotlib, 2014). However, growing evidence suggests that both exercise and antidepressant use may relieve depression through similar neuromolecular mechanisms (Garza et al., 2004; Hird et al., 2024). Increased expression of neurotrophic factors, such as brain-derived neurotrophic factor (Garza et al., 2004), greater availability of serotonin and norepinephrine, regulation of hypothalamic–pituitary–adrenal axis activity, reduced systemic inflammatory signalling (Mathur & Pedersen, 2008), and enhanced dopamine transmission can be achieved with PA and exercise interventions (Hird et al., 2024). These processes help develop new neurons, enhance synaptic connections, and increase cerebral vasculature and hippocampal plasticity (Sun et al., 2022). Hence, once can assume that if clinical symptoms improve, there is a reciprocal interaction with improvements in psychosocial functioning and QoL (Papakostas et al., 2004). After the 12-week intervention, the participants showed an increase in general health, vitality, social functioning, and mental component summary scores (Fig. 2). Although a previous study implementing a 12-week walking programme (30–45 min/day) suggested it may not be sufficient to significantly improve several domains of QoL (except for the physical functioning domain) in patients with RMD (Mota-Pereira et al., 2011), another more recent study found improvements in vitality and rumination with only 1-month of mindful walking in people with depressive disorder (El-Sayed et al., 2024).

After analysing the baseline variables, we categorised the participants as metabolically unhealthy, with overweight/obesity, according to some inflammatory biochemical parameters, overweight values, and low CRF (Table 1). The related clinical guidelines on cardiovascular disease prevention strongly recommend intensified attention and support to improve adherence to lifestyle changes in patients with mental disorders, including a reduction in sedentary time and a recommendation to follow general PA and exercise recommendations (Visseren et al., 2021). Based on our hypothesis, we expected that the exercise-mediated mitochondrial functions (i.e., mitophagy) would increase CRF and improve physical and biochemical inflammation-related parameters and, consequently, the antidepressant effect (Hird et al., 2024; Sun et al., 2022). However, the lack of changes in the participants’ metabolic profile and CRF (no improvement or worsening) requires further consideration of our results. First, the effect of low-to-moderate-intensity PA may have been amplified by reducing clinically related depression symptoms, although there were no changes in other parameters, consistent with previous studies (Morga et al., 2021; Mota-Pereira et al., 2011). Second, the intrinsic correlation between the pathophysiology of depression itself and inflammatory processes, with elevated CRP and IL-1β levels in the present study, mediated by the tryptophan-kynurenine pathway, makes the lack of favourable changes in the parameters related to body composition, CRF, and the biochemical profile very interesting. The maintenance of these parameters could be beneficial, considering the functional and clinical improvements (Ren & Xiao, 2023). Finally, high-intensity interval training seems to be more effective at inducing favourable changes to body composition, CRF, and inflammatory markers compared with low-to-moderate-intensity exercise, with exerkines, mitophagy, and lactate as the major brain-mediated effectors (Heo, Noble & Call, 2023; Mishra & Thakur, 2023).

Our study has several strengths, including the fact that there have been very few studies about the RMD population and none on an exercise intervention combining LIIT and resistance training. However, we must also note the limitations. First, we did not include a control group, which restricts our ability to control for confounding variables that might influence the observed outcomes. Moreover, we did not collect comprehensive data on potential covariates such as diet, sleep, or other lifestyle factors, which could impact physiological or psychological responses. These omissions were intentional, given the exploratory nature of this pilot study and the need to test feasibility with a limited sample size. Future studies should include randomised controlled designs and collect relevant covariate data to allow for adjusted analyses and stronger causal interpretations. Second, we did not have an equal number of men and women, which raises statistical problems. Future studies, particularly those using interventions and questionnaires, should seek to recruit an equal number of men and women. Third, four participants (22%) did not complete the post-intervention assessments. While missing data can bias the results, we note that in this case, dropout was due to logistical factors, such as early hospital discharge and geographic distance, rather than lack of clinical benefit. Finally, the sample represented the RMD population from community mental health centres in a single city, where mental health and social care are highly regarded and may not be representative of other towns or communities worldwide.

Conclusions

We demonstrated that a supervised, combined exercise programme (LIIT and resistance training) for people with RMD can induce positive and beneficial changes in functionality, clinical symptoms, and QoL, and maintain body composition, CRF, and biochemical variables. These results highlight the critical role of supervised exercise, regardless of intensity, in improving clinical symptoms in people with RMD. Thus, exercise might tentatively be considered in co-adjuvant programmes for the treatment of certain RMD populations.

Supplemental Information