Effects of Acute Exercise on Affect, Anxiety, and Self-Esteem in Poly-Substance Dependent Inpatients

Introduction: Negative affect and anxiety frequently precede the onset of drug use in those with substance use disorder (SUD). Low self-esteem may increase the risk of relapse. We examined the short-term effects of exercise on affect, anxiety, and self-esteem in inpatients with poly-SUD. Methods: This is a multicenter randomized control trial (RCT) with a crossover design. Thirty-eight inpatients (37.3 ± 6.4 years; 84% male) from three clinics participated in 45 min of soccer, circuit training, and control condition (psychoeducation) in a random order. Positive and negative affect (PANAS), state anxiety (single item), and self-esteem (Rosenberg SE-scale) were measured immediately before, immediately after, 1-h, 2-h, and 4-h post-exercise. Heart rate and ratings of perceived exertion were taken. Effects were assessed using linear mixed effects models. Results: Compared to the control condition, there were significant post-exercise improvements in positive affect (β = 2.99, CI = 0.39: 5.58), self-esteem (β = 1.84, CI = 0.49: 3.20), and anxiety (β = −0.69, CI = −1.34: −0.04) after circuit training (shown) and soccer. Effects persisted 4-h post-exercise. Reductions in negative affect were observed 2-h (circuit training: β = −3.39, CI = −6.35: −1.51) and 4-h (soccer: β = −3.71, CI = −6.03: −1.39) post-exercise, respectively. Conclusion: Moderately strenuous exercise undertaken in naturalistic settings may improve mental health symptoms in poly-SUD inpatients for up to 4-h post-exercise.


Introduction
Globally, an estimated 36 million people have a substance use disorder (SUD) and receive no treatment [1]. Those with SUD experience a significant health burden, including a higher prevalence of diabetes, cardiovascular disease, cancer, infectious diseases, decreased quality of life, and high levels of co-occurring mental disorders [2][3][4]. Poly-SUDs (i.e., dependence on two or more substances) are highly prevalent and associated with greater psychopathology, poorer karger@karger.com www.karger.com/ear treatment outcomes, and higher rates of fatal overdoses than individual SUDs [5,6]. Treatments include psychological/behavioral therapy and medication typically administered during hospitalization. These treatments are effective [7], but relapse rates remain high (e.g., 80% during the first month of abstinence [8]), and they do not directly address the somatic health problems associated with poly-SUD [9].
Unlike mood, which is considered a more pervasive and sustained emotional "climate," affect refers to shorter more reactive emotional states [10]. Affective dysregulation is associated with the development and maintenance of SUDs [3,11]. Negative affect is an important predictor of relapse, and the perceived inability to cope with negative affective states is frequently reported as a reason for using drugs [12]. Relapse is also associated with low levels of positive affect and poor self-esteem [13]. Similarly, anxiety has been linked to drug use and relapse in SUD populations [14]. To increase the likelihood of sustained remission, treatments are needed that focus on improving physical, social, and mental health symptoms -especially positive/negative affect, anxiety, and self-esteem [15]. Examining the short-term affective responses to exercise in SUD patients could inform the development of long-term exercise regimes and potentially optimize their impact.
Physical activity and its structured subset "exercise" are shown to benefit mental health in the general population [16,17], and among those with chronic health conditions [18,19], including depression and anxiety [20,21]. Exercise is also shown to improve positive affect and lower anxiety in those with alcohol and nicotine dependence [22], but studies involving poly-SUD populations are scarce. Brellenthin et al. [23] (2021) examined the effects of aerobic exercise plus intensive outpatient treatment on psychological variables and endocannabinoids in individuals with SUDs (n = 21). Participants engaged in supervised, moderate-intensity exercise for 30 min, 3 times/week for 6 weeks. Compared to outpatient treatment, the exercise group experienced acute increases in vigor and circulating concentrations of the endocannabinoid, anandamide [23]. Recently, Wang et al. [6] (2020) reported that 20-min of aerobic exercise (70-80% of maximum heart rate) reduced heroin cravings and promoted inhibitory control among adults with heroin dependency (n = 60). In a feasibility study involving nine inpatients with poly-SUD, we found non-significant but positive trends in affect and self-esteem after acute aerobic exercise and circuit training [24]. No previous studies have examined the effects of exercise in "real-life" settings (most have been laboratory based), which limits external validity, and the effects of different types of exercise remain unclear. Available studies have focused on pre-to post-assessments; little is known about the maintenance of exercise effects over time (i.e., 2-4 h postexercise), when affective states could return to preexercise levels.
We recently demonstrated that acute exercise reduces cravings for drugs among poly-SUD inpatientseffects that lasted 4 h post-exercise [25]. The current study uniquely focuses on the effects of acute exercise on secondary mental health outcomes, including positive/ negative affect, anxiety, and self-esteem. We compared the effects of a single session of soccer and circuit training to psychoeducation (control). Soccer was chosen because it is currently offered to our inpatients as part of routine care and is popular in the Nordic countries. Resistance exercise (circuit) training is not routinely offered but is appealing to many of those who attend SUD clinics and is shown to have beneficial effects on mental health [26]. We hypothesized that exercise would be associated with larger magnitude improvements in all secondary outcomes and that the effects would persist for 4-h postexercise.

Setting and Design
Inpatients under treatment for SUDs were recruited from three treatment centers in Oslo, Norway, that offer long-term inpatient SUD treatment for individuals over 18 years. Most patients are male, poly-substance users, with comorbid mental (e.g., depression/anxiety) and somatic health problems (e.g., hepatitis C and lifestyle-related disorders). The study was a short-term, multicenter randomized controlled trial (RCT) with a crossover design and random order condition allocation. Affect, self-esteem, and state anxiety were measured immediately before (baseline) and then immediately after, 1-h, 2-h, and 4-h post-exercise. Effects were compared to a control condition (psychoeducation). The study was retrospectively registered on November 11, 2019 (on German Clinical Trials [DRKS 00018869]). The protocol has been published [27].

Participants
Participants (n = 38) were over 18 years and diagnosed with two or more SUD diagnoses using the Statistical Classification of Diseases and Related Health Problems (ICD-10) (WHO, 2004). Participants were in different treatment stages, but all had completed detoxification. The only exclusion criteria were medical conditions contraindicated to physical exercise. No participants were excluded on this basis. Prior to analysis, one participant was dropped due to withdrawal of consent. The trial was approved by the Regional Committee for Medical and Health Research Ethics in South-East Norway [2018/1275]. Participation was voluntary, and all participants provided written informed consent.

Procedure
Recruitment took place between February and August 2019. Patients joined a meeting at the treatment centers where they were given information about the study from a research group member and were invited to join. To minimize dropouts, participants were recruited no more than 2 weeks before the trial. The order of the study conditions was randomized using a random number generator.

Measures
During a clinical interview, the participants' therapist filled out a survey using clinical data (SUD and comorbid diagnosis, primary drug use, drug use history, and medication) and sociodemographic characteristics. Height and weight were also recorded. Participants were asked about exercise prior to admission and their experience with exercise (see below). Cortisol samples were also taken as per the study protocol.

Exercise Variables
Participants completed Borg's Rating of Perceived Exertion Scale (RPE) [28], a single-item scale ranging from 6 (no exertion at all) to 20 (maximal exertion), immediately after each intervention. To collect an objective measure of exercise intensity, average and maximum heart rates were measured using a Polar Heart Rate Monitor.

Outcome Variables
The following validated questionnaires were completed immediately before, immediately after, 1-h, 2-h and 4-h after each intervention. Participants were asked to rate how they felt "at this moment." Positive and Negative Affect The Positive and Negative Affect Schedule (PANAS) consists of twenty items rated on a 5-point Likert scale ranging from 1 (very slightly/not at all) to 5 (extremely) [29]. Examples of items include (feeling) irritated, optimistic, proud, frustrated, scared. Half of the items represent positive affect; the other half negative affect. The PANAS has been widely used to assess affective responses to exercise, including in SUD populations, where reliability estimates are reported to be excellent [30].
State Anxiety. State anxiety was assessed with a single-item Visual Analogue Scale (VAS), ranging from 0 (no anxiety) to 10 (full panic). Studies have shown that single-item anxiety measures with either a Likert Scale or VAS response are an adequate replacement for the State Trait Anxiety Inventory (STAI) in repeated measures designs and reduce participant burden [31].
Self-Esteem. Participants completed the 10-item Rosenberg Self-Esteem Scale [32], measuring global (both positive and negative) feelings about oneself using a 4-point Likert scale format ranging from "strongly agree" to "strongly disagree." Scores range from 0 to 30. Scores between 15 and 25 are within the normal range; scores below 15 suggest low self-esteem. The scale is widely used to assess changes in SUD populations and has excellent psychometric properties [33].

Interventions
Participants completed two 45-min supervised group exercise sessions (soccer and circuit training) and one 45-min control condition (psychoeducation) in random order. Between interventions, there was a 1-day non-exercise "washout" period to control for possible carry-over effects. A 1-day interruption was feasible and enabled the study to be completed within 5 days, thereby avoiding weekends when drug use was a heightened risk. The interventions took place at the clinics between 9: 30 a.m. and 11 a.m. on Monday, Wednesday, and Friday for 1 week. Between the sessions, participants carried on with their treatment as usual. A nurse with formal education in physical activity for people with mental health and addiction disorders supervised the exercise sessions.
The interventions are described in detail in the published protocol [27]. Briefly, in circuit training, participants completed four circuits of bodyweight exercises with a 2-min rest between each circuit. Exercises included air squats, inchworm, dips, frog jumps, sit-ups, push-ups, walking lunges, and back hyperextensions (see online suppl. material; for all online suppl. material, see https://doi.org/10.1159/000531042 for illustrated examples). Each exercise was performed for 40 s followed by 20 s of rest before transitioning to the next station. The first circuit functioned as a warm-up with a lower intensity. Participants were asked to "give their best," to perform as many repetitions as possible, using proper form. Soccer was played indoors in a 20 × 40 m court. Each team had five participants. The match lasted for 45-min with a 2-min break after 15 and 30 min, respectively. Participants were asked to play a "friendly game" rather than a highly competitive match; goals were counted.
The control condition was a 45-min lecture on the health benefits of physical activity using a PowerPoint presentation and was similar to the psychoeducation that patients usually receive at the centers. All conditions were group sessions with the same participants in the three conditions. Participants could converse during all conditions.

Statistical Analyses
Descriptive statistics (means and SDs for continuous variables, counts and percentages for categorical data) were calculated for all participant characteristics. Between intervention differences in RPE, maximum and average HR were tested with paired samples t test. To analyze between intervention differences over time, we used linear mixed models (LMM) with a random intercept for each participant. Each outcome variable (positive affect, negative affect, state anxiety, and self-esteem) was analyzed separately. The models included a main effect of time (dummy-coding with baseline as reference) and the time by intervention state interaction (dummycoding of intervention state with the control state as reference). All models were adjusted for baseline variation by excluding the main effect of intervention state from the model. Estimated regression coefficients (representing estimated mean differences in the outcome variables between the given conditions) with 95% CIs and p values are reported. Period and sequence effects were checked by assessing main effects of the order of the trial days and the conditions for all outcome variables. For state anxiety, there was a significant main effect of period (trial day) (p = 0.004); state anxiety was significantly higher on the first day of the trial (Monday). Changes in state anxiety are reported with adjustment for the period effect. We also checked for main effects of the trial site (none were found). We used LMMa change from the study protocoldue to its ability to handle missing data. Participants with missing data were entered into the analysis. Data were analyzed using SPSS version 25 our feasibility study [24]. Sample sizes of ≥21 in each group would provide >80% statistical power (two-tailed α = 0.05, five repeated measures) to detect a small-to-moderate effect of exercise on negative affect. Power was estimated using G*Power (version 3.1.9.7).
Retention and Adherence Twenty-five (65.8%) participants completed all interventions, eight (21.1%) completed two, and five (13.2%) only one intervention. Factors unrelated to the trial and beyond participant`s control (e.g., illness, medical advice, commencement of new medication) contributed to most of the non-adherence (see online suppl. material for detailed descriptions and CONSORT Figure). No adverse events were reported.

Changes in Outcome Variables
To illustrate the changes within the three study conditions, estimated marginal mean scores from the model, stratified by study condition, are presented in Figure 1. Changes over time in the exercise interventions compared to the control condition, after baseline adjustment, are shown in Table 1. No period or sequence effects were found.
Positive and Negative Affect Results from linear mixed models are shown in Table 1. Compared to baseline, there was a significant increase in positive affect at all assessment times within circuit training and soccer. Compared to control, the increases in positive affect were significantly larger for both exercise conditions at all assessment points. The mean pre-to-post-intervention increase in positive affect was 2.68 (10.9%) in control, 7.63 (31.9%) in circuit training, and 6.27 (23.6%) in soccer. Compared to control, there was an estimated mean difference of 4.88 for circuit training and 5.5 for soccer. There were significant reductions in negative affect all time points within the circuit and soccer conditions (p < 0.001 all points). The mean pre-to-post-intervention decrease was 2.48 (12.3) in the control condition, 3.9 (19.2%) in circuit training, and 3.73 (18.3%) in soccer. Compared to control, there was an estimated mean difference of −2.53 for circuit training and −1.45 for soccer. The decrease in negative affect was significantly larger for circuit training compared to control up to 2 h post-intervention. For soccer the decrease was significantly larger than control at all time points except immediately after the intervention. There was no significant difference between soccer and circuit training.

State Anxiety
Within the exercise interventions there was a significant reduction in anxiety at all measurement points (p < 0.001). Compared to control, anxiety reduced significantly in the exercise interventions at all time points. The mean pre-to-post-exercise decrease was 1.5 points (i.e., 46.9%) in circuit training, and 1.16 points (i.e., 44.6%) in soccer. Compared to control, there was an estimated mean difference of −1.33 for both circuit training and soccer. There were no significant differences between soccer and circuit training.

Self-Esteem
Within intervention changes were significant at all time points for circuit training and soccer (p < 0.001). Compared to control, the increase in self-esteem was significant for both exercise interventions at all measurement times. The mean pre-to-post-intervention increase was 2.47 (9.9%) in circuit training and 2.69 (10.9%) in soccer. Compared to control, there was an estimated mean difference of 2.55 for circuit training and 2.34 for soccer. There were no significant differences between soccer and circuit training.

Discussion
This is the first trial to examine the effects of acute exercise, performed in non-laboratory settings, on mental health symptoms in poly-SUD inpatients. Participation in a single session of circuit training or a soccer match was associated with increased positive affect and self-esteem, and decreased state anxiety and negative affect. These effects lasted up to 4-h post-exercise. Previous studies have demonstrated similar acute exercise effects in populations suffering from anxiety, depression, and individual SUDs [34,35]. Findings extend the results from our pilot study, where we observed a non-significant trend of increased self-esteem and positive affect following soccer and circuit training [24]. Unlike previous trials limited to pre-post assessments, we observed effects that persisted 4 h after exercise cessation, indicating that the benefits were maintained for several hours after exercise cessation. Soccer and circuit training did not differ in their effects on the outcomes. This suggests that exercise effects are not dependent on exercise type when intensity is similar.
Several models have been proposed to explain the relationship between affect and drug use. Negative affective states are understood to act as conditioned stimuli that are capable of eliciting conditioned drug responses [36,37]. These conditioned responses are assumed to provide the motivation to engage in substance use. Acute bouts of exercise can regulate affective states, lower cravings for alcohol and other drugs, and enhance selfesteem [24,38]. We speculate that, over time, repeated pairings of acute exercise with inate biological and psychological rewards may reduce the use of drugs in those with SUD. Few studies have examined the effects of acute exercise on self-esteema potential mediator of the exercise-drug use relationshipbut exercise-related improvements in self-esteem have been reported in mental health populations [38], and an immediate increase has been reported in a youth SUD population [39].
The effects of acute exercise on mood (an emotional state with relative persistence over time) have varied in previous SUD studies. Ussher et al. [40] (2004) did not observe a pre-to-post-exercise effect on mood in alcohol-  [42]. Several trials have examined the effects of exercise as treatment for SUDs, including alcohol [41,[43][44][45], smoking [46,47], and methamphetamine dependence, where poly-substance dependence is common [48,49]. Recent reviews have concluded that exercise is an effective adjunct treatment for SUDs, with positive effects on drug craving and use and mental health reported [19,22,50]. Available evidence remains limited, and there have been calls for adequately powered trials [22].
Our findings have potential clinical implications. SUD inpatients could be supported to introduce exercise as a strategy to regulate affective states, which in turn may reduce the risk of relapse [51]. Increased self-esteem could improve patients' belief in their ability to cope with life stressors typically encountered in a treatment and rehabilitation process without the use of drugs. Enhanced self-esteem might also improve subjective wellbeing and potentially encourage patients to engage in other treatment interventions [52].
Strengths of our study include the randomized crossover design, where participants acted as their own controls. The exercises were performed in naturalistic settings encountered in everyday life and used in inpatient treatment, which may increase the external validity of our findings. Although the inpatient setting protected the participants from environmental factors they might encounter in an outpatient setting Ellingsen/Clausen/Johannesen/Martinsen/ Hallgren or upon finishing treatment (e.g., exposure to previous risk situations or availability of drugs), the generalizability of our finding is expected to be large for these populations. While previous studies have mainly assessed pre-post changes, we assessed effects up to 4 h after exercise and demonstrated that the benefits persisted beyond the initial exercise stimulus. Some potential limitations are acknowledged. As participation in the trial was voluntary, we cannot exclude the possibility that patients were biased toward expecting favorable effects of exercise. Assuming this is true, our findings may not be generalizable to those who do not have positive attitudes toward exercise. Most participants were male, which limits generalizability to females with poly-SUDs. Dropout and non-adherence rates were higher than expected, although within the rates of previous exercise studies for SUDs (7-54%) [53]. We used a singleitem measure of anxiety which cannot assess or detect clinical levels of anxiety. Although the reasons for noncompliance were mostly beyond the participants' control, future studies should explore factors that might contribute to increase participation in exercise interventions.
In conclusion, compared to a control condition, participation in circuit training or soccer was associated with improvements in positive affect, self-esteem, and anxiety in poly-SUD inpatients, and the effects remained 4-h postexercise. Reductions in negative affect were also observed in both exercise groups. Moderate-to-vigorous exercise performed in non-laboratory settings may improve mental health symptoms in poly-SUD inpatients which, in turn, could alleviate cravings for drugs and ultimately reduce the risk of relapse. Future studies should examine the mediating role of affect in the association between exercise and drug use in this population. Long-term trials of exercise for poly-SUD are also warranted.