From withdrawal to motivation: The feasibility, acceptability and early clinical benefits of adapted physical activity during inpatient alcohol withdrawal

Abstract

Introduction

Physical activity (PA) may improve physical and mental health in alcohol use disorder (AUD), yet poor adherence limits its clinical impact. Inpatient alcohol withdrawal represents a critical window for initiating health behavior change, although patients are often encouraged to rest.

Objectives

We evaluated whether supervised adapted physical activity (APA) can be feasibly implemented during inpatient withdrawal care and whether it influences short-term health and exercise engagement and motivation compared with free access to a PA room (FPA).

Methods

In this real-world study, 64 participants with severe AUD (26.5% women; mean age = 45.8 ± 11 years) completed pre- and post-assessments during a three-week inpatient withdrawal program. Participants were sequentially allocated to supervised APA or FPA with general advice. Feasibility (attendance and adverse events), session satisfaction, health outcomes (anxiety, depressive symptoms, self-esteem, craving, sleep quality, and physical functional capacity), and PA motivation (amotivation, extrinsic, and intrinsic motivation) were assessed after acute withdrawal at hospital discharge (post-test).

Results

APA participants engaged in significantly more PA (p < 0.05), with high levels of session satisfaction (median = 8.86 and 8.83/10), and an attendance rate of 83%. Health outcomes and amotivation improved over time in both groups (p < 0.001), without group-by-time interaction. Motivation showed a significant group and interaction effect (p < 0.05): APA participants scored higher post-test, while FPA participants decreased in extrinsic motivation (p = 0.013). Intention-to-treat analyses revealed consistent findings.

Conclusions

Implementing supervised APA during early withdrawal treatment is feasible and acceptable. While short-term clinical improvements were similar across groups, APA strengthened motivational processes relevant to PA engagement during recovery.

Graphical abstract

Highlights

• •Adapted Physical Activity is feasible and well-accepted early in detoxification.
• •More than half of inpatients do not exercise when it is not supervised.
• •Planned and supervised adapted physical activity increases exercise motivation.

1. Introduction

Alcohol Use Disorder (AUD) is a chronic, relapsing, and multifaceted mental disorder marked by a loss of control and persistent alcohol consumption despite negative consequences, as well as craving (American Psychiatric Association & Task, 2013; Koob and Le Moal, 2008; Koob and Volkow, 2016). Among patients seeking treatment for AUD, up to 79% present with at least one physical health problem and almost 60% have multiple physical alterations (Gossop et al., 2007). Physical fitness is markedly reduced in adults with AUD (Hallgren et al., 2021; Vancampfort et al., 2019). Altered physical fitness and physical disabilities in AUD are related to low levels of physical activity (PA) (Vancampfort et al., 2015a; Vancampfort et al., 2019) frequently observed in this population (Hallgren et al., 2021; Vancampfort et al., 2019). Insufficient PA is associated with greater AUD severity (Ternay et al., 2022), and higher levels of depressive, stress, and anxiety symptoms, nicotine dependence, and comorbid opioid use (Kelly et al., 2024; Ternay et al., 2022; Vancampfort et al., 2015a). Physically inactive individuals with AUD are more likely to experience poor sleep quality and insomnia (Bolstad et al., 2023; Duquet et al., 2025). In this context, interventions promoting PA among people with AUD are promising to counteract alcohol-related physical and psychological alterations exacerbated by inactivity (Cabé et al., 2021).

Most studies converge in showing the beneficial effect of PA on physical fitness among adults with AUD (Brown et al., 2014; Giménez-Meseguer et al., 2023; Piché et al., 2023; Sinyor et al., 1982). Sleep quality improvements have also been reported, although it remains unclear whether these changes are attributable to exercise or to its combined effects with alcohol drinking cessation (Duquet et al., 2025; Gunillasdotter et al., 2022). Three meta-analyses reported reductions in depressive symptoms (Hallgren et al., 2017a; Li et al., 2024; Piché et al., 2023) with greater effects observed in longer interventions involving more PA sessions (Hallgren et al., 2017a), but the impact of PA on mental health appears heterogeneous in AUD. Two meta-analyses found no effect on anxiety (Gür and Can Gür, 2020; Hallgren et al., 2017a) while two others reported improvements (Li et al., 2024; Piché et al., 2023).

The effect of PA on alcohol-related outcomes remains unclear. Some meta-analyses found no significant effects on risky or hazardous alcohol consumption (Gür and Can Gür, 2020; Li et al., 2024; Thompson et al., 2020). Other authors found small but significant reductions in weekly alcohol intake (Hallgren et al., 2017a; Lardier et al., 2021). A 3-week aerobic exercise intervention had a short-term effect with lower alcohol craving and fewer drinks consumed per week (Pennington et al., 2022).

PA is thus a promising and acceptable adjunctive treatment for AUD (Cabé et al., 2021; Stoutenberg et al., 2015), yet participation remains low (Giesen et al., 2016; Hallgren et al., 2017a) with a substantially higher dropout rate of 40,3% than that observed for other severe mental illnesses (Stubbs et al., 2016; Vancampfort et al., 2016) or chronic conditions such as cancer, cardiovascular diseases, and diabetes (Bullard et al., 2019). Adherence to PA interventions is commonly observed to be low among people with AUD and tends to decrease over time (Brown et al., 2009; Welford et al., 2023).

Motivation, in particular autonomous motivation, is a central aspect for sustaining PA engagement (Blevins et al., 2017; Weinstock, 2010), making it essential to embed motivational strategies into PA intervention design (Roessler et al., 2017; Vancampfort et al., 2015b; Vancampfort et al., 2025). The Self-Determination Theory (Deci and Ryan, 2000; Deci & Ryan, 2008) frames motivation as a continuum ranging from amotivation to intrinsic motivation (Fig. 1). Between these extremes, extrinsic motivation comprises four types of regulation that differ in their degree of internalization: external regulation, introjected regulation, identified regulation, and integrated regulation. Autonomous motivation appears to sustain PA by fostering intrinsic interest, personal endorsement, and satisfaction of the basic psychological needs for autonomy, competence, and relatedness (Deci & Ryan, 2008).

Fig. 1.

The self-determination continuum according to Deci et Ryan,2000.

Adapted PA (APA) is a research, theory, and service delivery directed toward persons of all ages underserved by the general sport sciences, disadvantaged in resources, or lacking power to access equal PA opportunities and rights (Hutzler and Sherrill, 2007; International Federation of Adapted Physical Activity, 2025). The tailored and supervised nature of APA intervention makes it suitable for much earlier integration into standard care than previously assumed, including during alcohol withdrawal and among inpatients with severe forms of AUD (sAUD). APA appears to be a relevant approach to enhance recovery and well-being in individuals with AUD, notably through enhanced exercise motivation (Cabé et al., 2021; Hallgren et al., 2017b; Horrell et al., 2020). Furthermore, alcohol withdrawal is a critical period for establishing new health-related behaviors (National Institute on Drug Abuse, 2012), but also a time when many barriers to PA can be overcome. Nevertheless, PA is still rarely offered in that context, with patients often advised to observe near-complete rest (Cabé et al., 2021; Horrell et al., 2020).

The present real-world interventional study explored early PA implementation during alcohol withdrawal hospitalization in adult inpatients with sAUD who require complete hospitalization. The first objective was to evaluate the feasibility and session satisfaction of an APA intervention for inpatients with sAUD undergoing alcohol withdrawal treatment. The secondary objective was to compare the short-term effect of supervised APA and free access to a PA room (FPA) on anxiety and depressive symptoms, self-esteem, alcohol craving, subjective sleep quality, physical functional capacity and exercise motivation. The hypothesis was that the APA condition would yield higher levels of satisfaction, greater adherence, and improvements in overall health and exercise motivation compared with the FPA condition.

2. Materials and methods

This study was designed as a real-world, embedded clinical study and conducted in line with the Declaration of Helsinki with an ethic approval from the Local Ethical Committee for Health Research, of the University Hospital of Caen (N°99). Sequential allocation was used to fit logistical constraints related to the presence of an APA instructor and minimize contamination between groups within a shared inpatient environment: APA (April–May 2018, 2019, and 2021) and FPA (June 2019–January 2020). Clinical outcomes are reported as exploratory, while feasibility outcomes are primary.

2.1. Participants

Eighty-seven inpatients aged 18 or over, weighing less than 130 kg, and diagnosed with sAUD according to the DSM-5 criteria (American Psychiatric Association & Task, 2013) were recruited in the Addiction Department at the University Hospital of Caen. The complete inclusion criteria and the flow chart for study recruitment are shown in Fig. 2. Psychiatrists specialized in addiction medicine and APA instructors jointly checked each patient to verify eligibility, provided information about the study protocol, and collected written informed consent.

Fig. 2.

Flow diagram of participant recruitment. Including reasons for non-inclusion and drop-out.

APA = Adapted Physical Activity; FPA = Free Physical Activity.

2.2. Experimental design

The usual inpatient treatment program consisted of a three-week hospitalization, beginning with a medically supervised acute alcohol detoxification. Alcohol withdrawal syndrome was managed pharmacologically when necessary (Laniepce et al., 2021). During the first week of usual care, which is dedicated to detoxification and assessment, patients remained in their rooms most of the time. This was followed by two weeks of individual and group therapy, with motivational enhancement therapy provided throughout the entire stay.

As for neuropsychological practice assessments, the pre-test was conducted once inpatients were considered detoxified by the medical team, as indicated by a low and stable Cushman score of two or less (Cushman Jr et al., 1985) and discontinuation of benzodiazepines. This was done to minimize the bias of acute alcohol withdrawal, but before initiating any moderate-intensity PA. Post-test was performed on the day of hospital discharge, replicating the initial evaluation battery (Fig. 3).

Fig. 3.

Experimental design oftheAPAddicto protocol.

APA = Adapted Physical Activity; PA = Physical Activity; FPA = Free Physical Activity.

2.3. Interventions

2.3.1. Adapted physical activity (APA)

According to recommendations for PA adherence in people suffering from an AUD and French health recommendations in APA (HAS, 2022; Horrell et al., 2020; Roessler et al., 2017; Thompson et al., 2020), APA instructors tailored each session according to treatment constraints, as well as individual needs and abilities. During the alcohol detoxification phase, two 45–60-min familiarization sessions of low-intensity PA occurred in the PA room, combining breath exercises, stretching, and self-massage techniques. Over the following two weeks, APA instructors supervised a total of five 60-min group sessions of moderate-intensity PA. Each session began with a 10-min warm-up and ended with a 5–10-min cool-down period involving stretching and discussion. Sessions included aerobic exercise, as well as resistance and balance exercises with gradual intensity progression. In addition, during the second weekend, APA instructors encouraged the inpatients to exercise independently. Inpatients could choose aerobic or resistance exercise sessions, which were listed on a sheet for the weekend. The APA teacher recorded inpatient attendance for each PA session.

2.3.2. Free physical activity (FPA)

Upon their inclusion, participants assigned to the FPA group received advice on the benefits of regular PA. The information sheet explicitly encouraged them to engage in PA after completing the detoxification phase, provided they had the necessary medical authorization giving them unrestricted access to the PA room. This room was equipped with four stationary bikes, a rowing machine, and some strength-training equipment such as weights and elastic bands, to allow aerobic or resistance exercise. Participants who used the PA room were asked to report on the duration of each exercise session.

2.4. Assessments

We collected sociodemographic and clinical characteristics by reviewing medical records (Table 2). Pre-test was conducted after completion of the acute detoxification phase and before any moderate-intensity PA sessions. The two low-intensity PA sessions were designed for safety and engagement and were not intended as the therapeutic exposure to be assessed on clinical outcomes. Pre-test evaluated alcohol misuse and PA level using the Alcohol Use Disorders Identification Test (AUDIT) (Gache et al., 2005) and Ricci & Gagnon's (Laureyns and Séné, 2011) questionnaires. These scientifically validated assessment tools for adults, along with the health and motivation outcome assessment tools used during the pre- and post-tests, are described in Table 1. The pre-post-test tools included the Hospital Anxiety and Depression Scale (HADS) (Zigmond and Snaith, 1983), the Rosenberg scale (Rosenberg, 1965; Vallieres and Vallerand, 1990), the Obsessive-Compulsive Drinking Scale (OCDS) (Ansseau et al., 2000; Anton et al., 1995), the 6-Minute walking test (6MWT) (ATS Statement, 2002), and Pittsburgh Sleep Quality Index (PSQI) (Buysse et al., 1989). The Sports Motivation Scale (SMS) (Brière et al., 1995) assessed amotivation, extrinsic motivation and intrinsic motivation.

Table 2.

Sociodemographic and clinical characteristics at baseline.

	APA		FPA		Statistic W	p-value
	N = 40		N = 24
	Median [Q1-Q3]	(Min - Max)	Median [Q1-Q3]	(Min - Max)
Age	47 [38.75–52.25]	(32–70)	40.5 [35–51.25]	(24–71)	573.5	0.197
Female. N (%)	10 (25.0%)		7 (29.2%)		χ2(1) =0.01	0.942
Years of education	12 [11–14]	(4–17)	11 [11−11]	(8–15)	595.5	0.096
Weight (kg)	71.5 [66–83]	(44–95)	74.5 [64.75–84]	(53–105)	457	0.755
BMIa (kg / m2)	24.49 [21.37–26.72]	(16.28–32.01)	24.41 [22.12–27.8]	(18.08–37.2)	411	0.427
Level of PAb (RG score)	19 [14.75–25.25]	(8–41)	21 [16–26.5]	(4–34)	412	0.497
Status of PAb. N (%)					χ2(2) =2.82	0.245
inactive	16 (40.0%)		7 (30.4%)
active	21 (52.5%)		16 (69.6%)
very active	3 (7.5%)		0 (0.0%)
Anxiolytic treatment. N (%)	20 (50.0%)		16 (66.7%)		χ2(1) =1.08	0.298
Antidepressant treatment. N (%)	14 (35.0%)		9 (37.5%)		χ2(1) =0.0	1.000
AUDIT (score)	31 [27.75–34]	(16–40)	30 [27.5–34.25]	(18–38)	484	0.961
Years of OH misuse	20.5 [11–28]	(4–38)	15 [11–21.25]	(3–39)	561.5	0.261
Daily total standard drink in previous month	18 [13–22.75]	(6–41)	22 [13.75–30]	(5–50)	405.5	0.304
Number of withdrawals	2 [1–4]	(0–6)	2 [1–3]	(0–7)	546.5	0.348
Number of substances used in previous month	3 [2–3]	(1–5)	2.5 [2–3]	(1–4)	525	0.513
Substance used. N (%)
Alcohol	40 (100%)		24 (100%)
Tobacco	32 (80.0%)		20 (83.3%)		χ2(1) =0.0	1.00
Cannabis (THC)	10 (25.0%)		5 (20.8%)		χ2(1) =0.01	0.939
CBD	2 (5.0%)		0 (0.0%)		χ2(1) =0.14	0.71
Benzodiazepine	20 (50.0%)		10 (41.7%)		χ2(1) =0.15	0.698
Heroine	1 (2.5%)		0 (0.0%)		χ2(1) =0.0	1.00
Cocaine	2 (5.0%)		0 (0.0%)		χ2(1) =0.14	0.71

Notes. Continuous variables were not normally distributed and are summarized as Median[Q1-Q3] (Min-Max) unless otherwise specified. APA = Adapted Physical Activity; FPA = Free Physical Activity; BMI = Body Mass Index; PA = Physical Activity; RG = Ricci & Gagnon; AUDIT = Alcohol Use Disorder Identification Test; THC = tetrahydrocannabinol; CBD = cannabidiol

^aBMI data were missing for 1 APA inpatient

^bLevel of PA and Status of PA data were missing for 1 FPA inpatient

Table 1.

Description of assessment tools.

Outcomes		Assessment tools	Assessment time
Alcohol misuse		AUDIT (Gache et al., 2005)	Self-report questionnaire 10 items; 5-point Likert scale (0–4) Total score range: 0–40 Scores ≥8 indicate hazardous or harmful alcohol use Internal consistency: α = 0.87 for the French version	Pre-test
Physical Activity Level		Ricci & Gagnon (Gache et al., 2005, Laureyns and Séné, 2011)	Self-report questionnaire 9-items, 5-point Likert scale (1–5) 3 profiles depending on the total score: inactive (<18); active (18 < score ≤ 35) or very active (>35)	Pre-test
Health outcomes	Anxiety and depression	HADS (Zigmond and Snaith, 1983, Laureyns and Séné, 2011)	Self-report questionnaire 14-items, 7 items for anxiety and 7 items for depression symptoms 4-point Likert scale (0–3) Total score range: 0–42 Higher the total score, higher the anxiety and depression symptoms Internal consistency: α = 0.78–0.93	Pre- and post-test
	Self-esteem	Rosenberg scale (Rosenberg, 1965, Vallieres and Vallerand, 1990)	Self-reported questionnaire, 10 items, 4-points Likert-scale (1–4) Score range: 0–40 Higher the score, the higher the self-esteem Internal consistency: α = 0.90 for the French version	Pre- and post-test
	Alcohol craving	OCDS (Ansseau et al., 2000, Anton et al., 1995)	Self-reported questionnaire, 14 items, 5-point Likert-scale from 0 to 4 Obsessive thought and compulsive craving Total score range: 0–40 Higher the total, higher the alcohol craving Internal consistency: α = 0.88 for the French version	Pre- and post-test
	Functional capacity	6MWT (ATS Statement, 2002)	Timed submaximal intervention Flat corridor with two cones spaced 30 m apart; marks every 3 m on the sides Walk as fast as possible for 6 min, back and forth between the cones Distance covered measured and recorded to the nearest meter at the end Greater distance covered, the higher functional capacity	Pre- and post-test
	Subjective sleep quality	PSQI (Buysse et al., 1989)	Self-reported questionnaire Assessing sleep quality over the past month: sleep latency, duration, efficiency, disturbances, use of sleep medication, daytime dysfunction due to sleepiness Total score range: 0–21 Higher total score, the greater sleep disturbance Internal consistency: α = 0.83	Pre- and post-test
Motivation for exercise	Amotivation	SMS (Brière et al., 1995)	Self-reported questionnaire Question: “Why do you practice your sport?” 28 items with 4 items per subscale 7-point Likert scale (1 = does not apply at all, 7 = applies exactly) Scores: •Amotivation score (range 0–28) •Extrinsic motivation composite score (range: 0–84): external, introjected, identified regulation subscales •Intrinsic motivation composite score (range: 0–84): knowledge, accomplishment, stimulation subscales Internal consistency: •Amotivation score, α = 0.71 •Extrinsic motivation, α = 0.73–0.86 •Intrinsic motivation, α = 0.86–0.92	Pre- and post-test
	Extrinsic motivation
	Intrinsic motivation

Notes. APA = Adapted Physical Activity, FPA = Free Physical Activity, AUDIT = Alcohol Use Disorders Identification Test, HADS = Hospital Anxiety and Depression Scale, OCDS = Obsessive Compulsive Drinking Score, 6MWT = 6 Minutes Walking Test, PSQI = Pittsburgh Sleep Quality Index, SMS = Sport Motivation Scale.

To evaluate the feasibility of the APA intervention, PA adherence was quantified based on the number of sessions, mean session duration and total PA duration performed weeks two and three (when both groups practiced). The safety of PA was assessed by the occurrence of PA-related adverse events during and after PA sessions.

Session satisfaction of the APA intervention was evaluated using two self-reported Likert scales, each ranging 0 to 10, which measured perceived enjoyment and well-being at the end of each session. Inpatients answered the following questions: “How much did you enjoy exercising?” and “ How would you rate your well-being?”

2.5. Data analysis and statistics

Descriptive statistics describe participant characteristics, with the sample median imputed for missing education data. Sociodemographic and clinical characteristics were compared between groups at pre-test using Wilcoxon or chi-square tests, when appropriate.

Between-group comparisons of PA adherence were conducted using Wilcoxon tests. Session satisfaction of APA sessions was presented using descriptive statistics. Health and motivation outcomes were analyzed using general linear mixed models (GLMMs), including group, time, and their interaction as fixed effects, with participants as random effects and age, sex, and education as covariates. For the 6MWT, height and weight were additionally controlled. We finally conducted within- and between-group planned comparisons (pre- versus post-test for each group, FPA versus APA at each assessment time).

All statistical analyses were conducted using RStudio (Posit Software, PBC., 2023) with an alpha level of 0.05.

3. Results

3.1. Sociodemographic and clinical characteristics

No significant differences were observed between the APA and FPA groups in terms of sociodemographic or clinical characteristics (all p > 0.05; Table 2).

3.2. Intervention adherence

Thirty-five sAUD inpatients initiated the FPA intervention, of whom 24 completed the program. Similarly, among the 52 sAUD inpatients included in the APA group, 40 completed the program. There was no significant difference in group dropout (Χ² = 0.382, p = 0.536), without any dropout reason specifically related to PA intervention. Subsequent statistical analyses were therefore conducted on the 64 subjects who completed both assessments (Fig. 2).

3.3. Session satisfaction and feasibility

We reported no undesirable effects or adverse events related to PA. The APA group completed more PA sessions (p = 0.027, r = 0.277 (small)) and spent more time exercising per session (p < 0.001, r = 0.854 (large)) than the FPA group, with a significantly higher total PA duration during hospitalization (p < 0.001, r = 0.484 (moderate)). Within the FPA group, two distinct subgroups emerged: 13 inpatients reported no PA during hospitalization (FPA-) and 11 engaged in at least one exercise session (FPA+).

In the APA group, the attendance rate, enjoyment and well-being scores were consistently high. The median attendance rate was 83%, and mean ratings ranged from 7.67 to 10. The median values were 8.86 for enjoyment and 8.83 for well-being (Table 3).

Table 3.

Physical activity engagement between groups as well as enjoyment and well-being in the APA group.

Measures		APA	FPA	Statistic W	p-value
		N = 40	N = 24
Number of PA session	Median [Q1-Q3] Min-Max	5 [4–6] 2–6	0 [0–9] 0–13	637	0.027*
Mean duration of PA session (min)	Median [Q1-Q3] Min-Max	54 [54–55] 48.3–62.5	0 [0–28.2] 0–45	960	<0.001***
Total duration of PA (min)	Median [Q1-Q3] Min-Max	270 [200–270] 110–270	0 [0–215] 0–585	752	<0.001***
Attendance rate (%)	Median [Q1-Q3] Min-Max	83 [67–100] 33–100	NA	NA	NA
Enjoyment ratea (0−10)	Median [Q1-Q3] Min-Max	8.86 [8.37–9.71] 7.67–10	NA	NA	NA
Well-being ratea (0–10)	Median [Q1-Q3] Min-Max	8.83 [8.33–9.73] 7.67–10	NA	NA	NA

Notes. Values are presented as Median[Q1-Q3] Min-Max. APA = Adapted Physical Activity; FPA = Free Physical Activity; PA = physical Activity

^aEnjoyment and well-being rates data were missing 1 APA inpatient

3.4. Health outcomes

The GLMMs revealed no significant group effect on health outcomes (HADS total score, Rosenberg score, OCDS total score, 6MWT distance and PSQI total score). However, a significant time effect indicated improvements in both groups by the end of the intervention. No significant group-by-time interaction was found. Planned comparisons confirmed the absence of differences between or within groups at the pre-test (all p > 0.05, Fig. 4).

Fig. 4.

Changes in health outcomes between pre- and post-assessment as a function of APA and FPA groups. Estimated scores are adjusted by age. Sex and education. Distance in 6MWT adjusted by age, sex, education, weight and height. HADS and OCDS data were missing for 2 FPA inpatients; 6MWT data were missing for 1 FPA inpatient; PSQI data were missing for 1 APA inpatient and 4 FPA inpatients.

*** p-value < 0.001

APA = Adapted Physical Activity. FPA = Free Physical Activity. HADS = Hospital Anxiety Depression Scale. OCDS = Obsessive Compulsive Drinking Score. 6MWT = 6-min walking test. PSQI = Pittsburgh Quality Sleep Index.

3.5. Motivation outcomes

All motivation results are summarized and illustrated in Fig. 5.

Fig. 5.

Changes in amotivation and motivation scores between pre- and post-assessment as a function of APA and FPA groups. Estimated scores were adjusted by age., sex and education. Amotivation and extrinsic motivation data were missing for 3 APA inpatients and 7 FPA inpatients; Intrinsic motivation data were missing for 2 APA inpatients and 8 FPA inpatients.

*p-value < 0.05; **p < 0.01; *** p-value < 0.001.

APA = Adapted Physical Activity. FPA = Free Physical Activity.

3.5.1. Amotivation

There was no group effect, but a significant effect of time (p = 0.008), reflecting a decrease in amotivation from pre- to post-test. No significant group-by-time interaction was found (p > 0.05, marginal R² = 0.068).

3.5.2. Extrinsic motivation

We found a significant group effect (APA > FPA, 0.001), no significant effect of time, and a significant group-by-time interaction (marginal R² = 0.154, p = 0.004). Planned comparisons revealed a decline in extrinsic motivation during FPA intervention (estimate = −0.204, 95% CI [−0.365; −0.044], p = 0.013) and higher post-test scores in APA compared with FPA (estimate = 0.316, 95% CI [0.121; 0.511], p = 0.001).

3.5.3. Intrinsic motivation

The GLMM revealed a significant group effect (APA > FPA, p < 0.001), but no significant effect of time. However, a significant group-by-time interaction was observed (marginal R² = 0.197, p = 0.037). Planned comparisons indicated no significant difference between the groups at pre-test, but higher intrinsic motivation at post-test in APA compared with FPA (estimate = 0.376, 95% CI [0.185; 0.566], p < 0.001).

3.6. Exploratory analyses

3.6.1. Intention-to-treat analyses

Exploratory intention-to-treat (ITT) analyses, using the full dataset (n = 87), with last observation carried forward imputation for missing post-test data, confirmed the primary findings, with minor variations. A significant group effect on self-esteem persisted (p = 0.015), with APA exhibiting higher Rosenberg scores than FPA at both assessments. Amotivation significantly decreased only in the APA group (p = 0.010). The group-by-time interaction remained significant for intrinsic and extrinsic motivation (p = 0.024 and p = 0.010, respectively). The APA group demonstrated higher levels of intrinsic motivation at both assessment times (pre: p = 0.011 and post: p < 0.001), while its decline in FPA was non-significant in ITT (p = 0.173). These nuances did not alter the overall interpretation of the results.

3.6.2. Subscale analyses

Exploratory analyses of the SMS subscales confirmed the overall pattern for the extrinsic and intrinsic composite scores (Supplementary Materials, S1). The significant interaction effects were primarily driven by an increased tendency for motivation in APA and/or decrease tendency in the FPA group over time. Planned comparisons showed that both groups had a similar level of motivation at pre-test, except for intrinsic motivation linked to knowledge. However, post-test, APA inpatients demonstrated higher levels of motivation across all subscales, particularly concerning the introjected, identified, and intrinsic motivation.

3.6.3. Subgroup analyses

There were no significant differences between APA, FPA(+) and FPA(−) groups for any sociodemographic or clinical characteristics, as determined by non-parametric Kruskal–Wallis ANOVAs (all p > 0.05; Table 4).

Table 4.

Demographics and baseline characteristics by PA adherent subgroups.

Measures	APA (N = 40)	FPA+ (N = 11)	FPA- (N = 13)	Statistic W	p-value
	Median [Q1-Q3]	Median [Q1-Q3]	Median [Q1-Q3]
Age	47 [38.8–52.2]	40 [36.5–56.5]	43 [32–47]	2.073	0.355
Female. N (%)	10 (25.0%)	2 (18.1%)	5 (38.5%)	Χ2 (2) =1.3897	0.499
Years of education	12 [11–14]	11 [11−13]	11 [11–11]	4.104	0.129
Height (cm)	175 [169–179]	175 [167.5–176.5]	171 [165–175]	1.442	0.486
Weight (kg)	71.5 [66–83]	68 [62–75.2]	80 [65–84]	1.107	0.575
BMI (kg/m2)a	24.5 [21.4–26.7]	23.3 [20.5–26.2]	27.4 [23.5–29.1]	3.038	0.219
Level of PAb (RG score)	19 [14.8–25.2]	22 [19–26.5]	18.5 [14.2–24.2]	2.504	0.286
Status of PAb . N (%)				Χ2 (4) =7.081	0.132
inactive	16 (40.0%)	1 (9.1%)	6 (46.2%)
active	21 (52.5%)	10 (90.9%)	6 (46.2%)
very active	3 (7.5%)	0 (0.0%)	0 (0.0%)
Anxiolytic treatment. N (%)	20 (50.0%)	7 (63.6%)	9 (69.23%)	Χ2 (2) =1.769	0.413
Antidepressant treatment. N (%)	14 (35.0%)	6 (54.5%)	3 (23.08%)	Χ2 (2) =2.604	0.272
AUDIT	31 [27.8–34]	30 [27–34]	30 [28–34]	0.053	0.974
Years of OH misuse	20.5 [11–28]	16 [11.5–23.5]	13 [8–20]	2.171	0.338
Daily total standard drink in previous month	18 [13–22.8]	17 [10–22.5]	25 [19–35]	4.759	0.093
Nb of substances used in previous month	3 [2–3]	2 [2–2.5]	3 [2–3]	2.384	0.304
Substance used. N (%)
Alcohol	40 (100%)	11 (100%)	13 (100%)		1.000
Tobacco	32 (80.0%)	9 (81.8%)	11 (84.6%)	Χ2 (2) =0.140	0.932
Cannabis (THC)	10 (25.0%)	1 (9%)	4 (30.8%)	Χ2 (2) =1.706	0.426
CBD	2 (5.0%)	0 (0.0%)	0 (0.0%)	Χ2 (2) =1.239	0.538
Benzodiazepine	20 (50.0%)	4 (36.6%)	6 (46.2%)	Χ2 (2) =0.648	0.723
Heroine	1 (2.5%)	0 (0.0%)	0 (0.0%)	Χ2 (2) =0.610	0.737
Cocaine	2 (5.0%)	0 (0.0%)	0 (0.0%)	Χ2 (2) =1.239	0.538

Notes. Median [IQR]. IQR = interquartile range. PA = Physical Activity. APA = Adapted Physical Activity. FPA + = Free Physical Activity adherent. FPA- = Free Physical Activity non-adherent. BMI = Body Mass Index. RG = Ricci & Gagnon. AUDIT = Alcohol Use Disorder Identification Test. THC = tetrahydrocannabinol. CBD = cannabidiol

^aBMI data were missing for 1 APA inpatient

^bLevel of PA and Status of PA data were missing for 1 FPA(−) inpatient

Exploratory subgroup analyses confirmed the pattern observed between APA and FPA regarding the amotivation and extrinsic composite scores. For the intrinsic motivation composite score, the subgroup effect was significant, but not the interaction (p > 0.05) (Fig. 6). At pre-test, no difference was observed in terms of amotivation or motivation (all p > 0.05). Following the intervention, amotivation decreased significantly only in the APA group (p = 0.007), whereas extrinsic motivation decreased in the FPA(−) subgroup (p = 0.028). Furthermore, both extrinsic and intrinsic motivation composite scores were significantly higher in APA than in FPA(−) at post-test (p = 0.006 and < 0.001, respectively). The extrinsic and intrinsic motivation scores of the FPA(+) subgroup did not decrease significantly and were not different from those of the other groups at both assessment times.

Fig. 6.

Changes in amotivation and motivation scores from pre- to post-assessment across APA. FPA(+) and FPA(−) subgroups. Estimated scores were adjusted by age. Sex and education. Amotivation and extrinsic motivation data were missing for 3 APA inpatients. 4 FPA(+) inpatients and 3 FPA(−) inpatients; Intrinsic motivation data were missing for 2 APA inpatients. 4 FPA(+) inpatients and 4 FPA(−) inpatients.

*p-value<0.050. **p-value<0.010. *** p-value<0.001.

APA = Adapted Physical Activity. FPA + = Free Physical Activity with at least one PA session. FPA -: Free Physical Activity reporting no PA session.

4. Discussion

This study is the first to implement an APA intervention from the earliest stage of alcohol withdrawal hospitalization. Our findings suggest that detoxification treatment could be a key factor in improving the short-term health outcomes in both groups, while supervised APA can be safely integrated into withdrawal care and has a specific added value in strengthening inpatients' motivation and adherence to PA compared with unsupervised access to PA facilities. Thus, these findings imply that early supervised and structured APA is a promising intervention to change behavior by increasing motivation for exercise rather than a clinical treatment with free access to a PA room during withdrawal hospitalization.

4.1. Feasibility and satisfaction of APA during inpatient alcohol withdrawal

The APA program was safe and readily implemented, even when introduced early after alcohol cessation, as reflected by the high enjoyment and well-being scores reported by APA inpatients and the absence of dropout and adverse events related to PA. The delivery of the program within the ward was smooth and caused no operational difficulties. While most inpatients report believing that PA can be initiated during alcohol treatment (Stoutenberg et al., 2015), only half of FPA inpatients initially intending to exercise at the start of the study finally engaged in PA during hospitalization. In contrast, APA inpatients exercised at least 2 sessions, indicating that APA intervention integrated into usual care is associated with higher engagement for sAUD inpatients compared to spontaneous exercise (Stoutenberg et al., 2015). Such programs and sessions include goal setting, positive feedback, enjoyment and social support through supervision and/or group-based activities, all of which can enhance acceptability, self-efficacy, and motivation (Hallgren et al., 2017b; Weinstock, 2010). By its personalized and motivational approach, APA appears as a promising approach.

4.2. Health-related outcomes

Although PA adherence was significantly higher in the APA group, physical and mental health outcomes improved similarly in both groups. This suggests that these changes could reflect common factors in hospital care, such as stopping drinking alcohol and following the entire therapeutic program. The early weeks of abstinence are typically associated with significant physical and psychological recovery, including reduction in anxiety, depressive symptoms and craving (Manning et al., 2008; Petit et al., 2017).

The overall therapeutic effects of alcohol withdrawal and inpatient stabilization may have overshadowed the specific contribution of PA intervention. Moreover, the PA intervention was designed to fit the actual duration of alcohol withdrawal hospitalization. However, three weeks were likely too short to induce measurable health changes. Previous studies and guidelines and clinical guidelines suggest that exercise intervention should last at least 12 weeks to observe a significant effect on clinical outcomes (Gür and Can Gür, 2020; Piché et al., 2023). Therefore, the absence of significant effects in our study may be partly explained by the short intervention duration. While previous studies have reported greater improvements in psychoaffective outcomes and craving with longer interventions, suggesting a possible dose-response relationship (Abrantes et al., 2017; Roessler et al., 2013), our study was not designed to evaluate such an effect. Further studies with longer intervention periods are needed to determine whether dose-response relation exists in this population. However, follow-up studies reported that PA practice diminished after the end of intervention (Fagan et al., 2021; Patten et al., 2003; Weinstock et al., 2020), highlighting the importance of promoting long-term PA adherence after hospital discharge.

4.3. Motivation and self-determination processes

Even though a 3-week program was likely too short to elicit specific health benefits, significant between-group differences emerged regarding changes in exercise motivation.

The decrease in amotivation observed at the end of the intervention in both groups may reflect the beneficial effects of alcohol withdrawal treatment on patients' general clinical condition. Improvements in withdrawal symptoms and overall functioning during inpatient care may contribute to greater perceived capacity and readiness to consider health-promoting behaviors, including PA. The absence of a specific effect of PA intervention on amotivation is consistent with previous findings in individuals with AUD, where amotivation was not associated with any stage of change for PA according to the Transtheoretical Model of Prochaska and DiClemente (Vancampfort et al., 2018). Similarly, the relationship between amotivation and PA appears to be complex and may depend on the level of engagement in exercise. Indeed, previous research reported no association between amotivation and exercise outcomes in samples including both exercisers and non-exercisers, whereas negative associations were observed among regular exercisers only (Teixeira et al., 2012). Although no significant difference in amotivation was observed between FPA(+) and FPA(−) groups, future studies should investigate whether sustained participation and adherence to PA programs may influence motivational regulation over time. This question may be particularly relevant in clinical populations, as previous research in individuals with severe mental illness has reported a negative association between amotivation and PA engagement (Farholm et al., 2017; Vancampfort et al., 2015b).

Furthermore, among FPA inpatients, particularly non-adherent ones (FPA-), extrinsic motivation decreased significantly during the intervention, suggesting that providing access to the PA room without program and supervision may be insufficient for patients lacking the necessary resources to exercise independently. Moreover, the APA group showed higher extrinsic and intrinsic motivation than the FPA group at post-test, primarily due to the low motivation of the FPA(−) group. PA practice seems to favor both types of motivation for exercise, regardless of the modality of PA (Farholm et al., 2017; Vancampfort et al., 2015b). These observations reinforce the relevance of supervised and structured APA programs and suggest that simply providing free access to a PA room may even be insufficient for the least active inpatients, who showed increased amotivation and decreased extrinsic motivation.

Concerning extrinsic motivation, APA inpatients showed significantly higher external and introjected regulation at post-test. Although such controlled motivation leads to the initiation of PA behavior, it is not sustainable in the long-term (Teixeira et al., 2012). Introjected regulation may help to initiate PA but APA interventions must ideally promote autonomous motivation (Vancampfort et al., 2015b; Vancampfort et al., 2025), which was associated with 31% higher moderate-to-vigorous PA in people with mental illness (Chapman et al., 2024). A higher level of identified regulation in the APA group could be linked to the exchange of knowledge about PA with a qualified APA supervisor and the acute effect of exercise, which may help participants to recognize and identify the benefits of PA within just three weeks. Previous research suggests that identified regulation is a strong predictor of exercise adoption, while intrinsic motivation plays a critical role in sustaining exercise over time (Teixeira et al., 2012). The APA supervisor likely contributes to these gains by providing information during practice, fostering enjoyment and support psychological basic needs of autonomy, competence and relatedness (Deci & Ryan, 2008; Farholm et al., 2017). As far as we know, only one clinical trial has examined exercise motivation based on the Self-Determination Theory in people with AUD. This study showed that a 12-week lifestyle PA intervention increased autonomous motivation (Blevins et al., 2017), which is consistent with our results. However, autonomous motivation was not related with PA duration but rather with a greater use of PA as a coping strategy for alcohol craving and negative affect (Blevins et al., 2017).

5. Limitations and future directions

Even though there was no between-group difference in sociodemographic and clinical characteristics, the small proportion of women remains a limitation that may prevent the generalizability of our finding for inpatient populations with a higher representation of women. Additionally, sequential design remains a limitation, and the use of self-reported measures could result in desirability and recall bias. Furthermore, the relatively small and unequal sample sizes may have reduced statistical power. Nonetheless, randomization could have been considered as a limitation given the inevitable interaction between APA and FPA participants in the Addiction department. Thus, the naturalistic and ecological setting of this study ensures external validity. In addition, we were unable to distinguish the respective effect of supervision and greater exposure to PA, as supervision was considered an integral component of the APA intervention. To better examine how an APA approach that supports basic psychological needs fosters self-determined motivation, future researches should use a more comprehensive measure of exercise motivation across the continuum of self-determination. Additionally, the dynamic relationships with PA adherence, craving and risk of relapse should be examined over a longer follow-up period.

6. Conclusion

This real-world study shows that APA can be safely and effectively integrated into inpatient alcohol withdrawal care, with excellent session satisfaction and greater adherence than unsupervised exercise. While short-term clinical improvements were comparable between groups, APA was associated with more favorable changes in exercise motivation. These findings highlight the early alcohol withdrawal period as a pivotal moment for re-engagement in health-promoting activities and a window for PA. By providing structured, individualized guidance and meaningful social support, supervised APA may address psychological needs, relevant for sustained engagement. Integrating tailored and supervised exercise programs from standard withdrawal care and for several weeks could therefore serve as a promising clinical lever to boost motivation, enhance recovery trajectories.

Credit authorship contribution statement

Aude Barraud: Writing – review & editing, Writing – original draft, Project administration, Methodology, Investigation, Formal analysis, Data curation. Alice Laniepce: Writing – review & editing, Methodology. Julien Cabe: Writing – review & editing, Methodology. Shailendra Segobin: Writing – review & editing, Methodology. Anne-Lise Pitel: Writing – review & editing, Writing – original draft, Supervision, Methodology, Formal analysis. Nicolas Cabe: Writing – review & editing, Writing – original draft, Supervision, Methodology, Investigation, Formal analysis, Data curation, Conceptualization.

Funding

This research was supported by the Institut Universitaire de France (IUF). The funders and sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary material: Results of Generalized Linear Mixed Models (GLMM) and planned comparisons on motivation subscale of the Sport Motivation Scale (SMS)

Data availability

Data will be made available on request.