Resistance training is an underused and promising tool in eating disorder recovery: a narrative review

Abstract

Background

Individuals with eating disorders may experience reduced bone mineral density and muscle mass, lower quality of life, and debilitating psychopathology. Abstinence from exercise has historically been recommended in the treatment of eating disorders. However, evidence shows that well-planned, supervised exercise can be safe and beneficial. Resistance training is an exercise modality of particular interest because of its ability to protect bone and muscle mass, in addition to its mental health and quality of life benefits. The purpose of this narrative review is to examine the use of resistance training in the treatment of eating disorders to determine whether its incorporation in treatment programs should be recommended.

Methods

A search of the literature was conducted via PubMed, Web of Science, ScienceDirect, and Sage Journal. Articles published before the final database search (March 22nd, 2025) were considered. Inclusion criteria: Studies must examine how the use of resistance training affects people with anorexia nervosa, bulimia nervosa, or binge eating disorder, and must be published in English. Both quantitative and qualitative literature were included.

Results

A total of 17 studies were examined. Quantitative research revealed significant improvements in body composition, strength, and quality of life when resistance training was included in treatment. The available evidence suggests that higher loading is well-tolerated and is more effective than low loading at improving body composition. Qualitative research demonstrated participants’ positive attitudes toward the use of resistance training during treatment. The participants reported improved relationships with exercise, citing the emphasis on building strength as an empowering motivator that deemphasizes weight loss.

Conclusion

In combination with a multidisciplinary team including a dietitian, psychologist, and medical provider, resistance training has the potential to play an important role in improving psychopathology, quality of life, and body composition in patients with eating disorders. More research is needed in this area before its inclusion in treatment can be confidently recommended.

Plain english summary

People with eating disorders experience reduced bone mineral density muscle mass, and report low quality of life. Although abstinence from exercise has historically been recommended in the treatment of eating disorders, its intentional application may confer meaningful benefits to both physical and mental health. Resistance training is a modality that shows promise for improving many eating disorder symptoms, including protecting and increasing bone and muscle health and improving quality of life.

1. Background

Eating disorders (EDs) are psychiatric illnesses that affect people of all ages and genders, most commonly originating in adolescence or young adulthood [1]. This review focuses on resistance training (RT) as a potential tool for improving the treatment of anorexia nervosa (AN), bulimia nervosa (BN), and binge eating disorder (BED).

Anorexia nervosa can be categorised as either restricting type (AN-R) or binge-eating/purging type (AN-BP) [2]. A 2017 review found lifetime prevalence rates of AN among females to range from 1.7–3.6% [3]. The lifetime prevalence rates of BN and BED among the entire population are 0.63% and 1.53%, respectively [4]. The mortality rate of eating disorders is among the highest of any psychiatric condition, ranging from 5–21%, with the number increasing with age and length of time spent with the illness [5–7].

Physical and exercise therapies are not widely utilised at any level of care, and are usually confined to yoga and similar low-intensity modalities when they are included in treatment [8]. Historically, patients with ED have been placed on bed rest to reduce metabolic demands [9]. Clinicians’ hesitancy to incorporate exercise in ED treatment is due to both the physical health profile with which many patients present and the prevalence of maladaptive exercise (ME) in this population, which is 22–80% [10–12]. ME is associated with worse ED psychopathology and greater risk of relapse [12, 13]. ME can be understood as exercise that negatively impacts health and/or quality of life [14]. Researchers have used several terms across the literature to describe this phenomenon and its various presentations, including “compulsive exercise,” “excessive exercise,” “driven exercise,” among others [14–16]. In the current review, ME refers to the broad category that encompasses these terms.

Quantitatively, ME encompasses “excessive exercise,” which describes an amount of exercise that is inappropriately high for the individual [15]. The amount of exercise that is inappropriate must be considered on an individual basis, as the amount of physical exertion a person can tolerate at a given time is not universal [15, 17]. ME also refers to the qualitative experience of exercise. Patients with ED often engage in exercise in a rigid or compulsive manner [15, 16, 18]. Changes in their exercise schedule may elicit extreme distress [15, 16, 18]. Patients may feel obligated to exercise in order to burn calories or earn the right to eat [15, 16, 18]. There is often a fixation on controlling body shape, accompanied by a fear of what will happen if they do not engage in exercise [15, 16, 18]. Bratland-Sanda et al. have emphasized the importance of recognising “exercise obsession” as a subtype of ME that does not always occur alongside excessive exercise behaviours [19]. The authors noted that patients with BN or BED have been assessed as exhibiting both compulsive exercise (per Compulsive Exercise Test scores) and sedentary behaviour. In these cases, patients may have obsessive thoughts about the exercise they should do or intend to do, yet may be unable to carry out much activity in practice [19].

While the the above factors have led to the avoidance of physical activity in treatment, a growing body of research shows supervised exercise to be both safe and beneficial for patients with EDs [20–23]. Including exercise in treatment may help prepare patients to incorporate exercise in their lives in an adaptive rather than maladaptive manner [23]. This is an important skill given the risk of relapse associated with ME, as well as the health consequences associated with inactivity in the general population [13, 24].

RT is an exercise modality that may be particularly useful for the ED population due to both its physical and psychological benefits [25, 26]. RT refers to exercise that involves muscular contraction against an external load, which can come from free weights, machine weights, bodyweight, or resistance bands [27, 28]. RT is more effective at improving bone mineral density (BMD), muscle mass, and muscle strength than aerobic training [29]. These are relevant benefits for the ED population, who often experience reduced bone mineral density and myopathy as a result of their disorder or its side effects [30–33]. RT has been shown to improve mental health conditions that are frequently comorbid to EDs, including depression and anxiety, as well as increase quality of life (QoL) [34–37]. As people with EDs report low QoL compared with their healthy counterparts, and low QoL is associated with worse treatment outcomes, the role of RT in this area is also worth exploring [6, 38–40]. RT may be a particularly useful physical therapy in this population due to its superior role in preserving lean mass, as well as its lower calorie expenditure compared with aerobic training, making it less likely to interfere with weight restoration [25, 41]. Although other authors have presented the evidence for physical therapy as a component of ED treatment in review articles, none have focused specifically on resistance training [8, 20, 23]. Therefore, there is a need for a review of the current evidence for RT in the care of patients with ED. The aim of this review is to examine the existing research on RT in populations with AN, BN, or BED and determine whether its incorporation in treatment programs should be recommended.

2. Methods

A thorough search of the literature was conducted via PubMed, Web of Science, ScienceDirect, and Sage Journal. Search terms included: “anorexia nervosa” OR “bulimia nervosa” OR “binge eating disorder” OR “eating disorder*” AND “resistance training” OR “strength training” OR “weight training” OR “pilates” OR “physical therapy”. Both qualitative and quantitative studies published through March 22nd, 2025 (date of final search) were considered. Criteria for inclusion were: Studies must be original, peer-reviewed research examining how the use of RT affects people with AN, BN, or BED; studies must be published in English. Other specified feeding or eating disorder (OSFED) was not included despite its prevalence because it encompasses a range of disorders of varying characteristics and etiologies that would need to be researched individually, which is beyond the scope of this review. However, four of the five OSFED subcategories (atypical AN, BN of low frequency and/or limited duration, BED of low frequency and/or limited duration, and purging disorder) share traits with AN, BN, and BED, and thus this research may be of potential relevance to these populations as well. See Figure 1 for a PRISMA flow chart of the article selection process.

Fig. 1.

Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. https://doi.org/10.1136/bmj.n71

The 17 articles included for review encompass randomised controlled trials (RCTs), qualitative interviews, case studies, a case series, a case report, a non-randomised controlled trial, and an uncontrolled pilot study. The results of these studies were analyzed and described, and implications for effects of RT in ED treatment were outlined. Given that the studies vary widely in type, design, and outcomes measured, a systematic review and meta-analysis of the data were not conducted. Author V.B. worked independently to search the specified databases, screen each record, and determine eligibility for inclusion. A narrative review was conducted using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines as a framework due to the varied methodologies and study designs relevant to this topic. The Scale for the Assessment of Narrative Review Articles (SANRA) was used as a guideline for conducting this review in order to ensure high quality standards were met [42]. The narrative review format allowed the included studies to be interpreted together to provide context for clinical practice and future research.

Due to the inclusion of studies that are subjective, uncontrolled, non-randomized, or, in some cases, have a sample size of one, there is a risk of bias in the reported outcomes. Even in RCTs, double blinding is impossible for the question being examined. There is also risk of bias due to the reasonable caution researchers take when selecting participants, screening for maladaptive exercise and requiring certain percentages of weight restoration in inclusion criteria. These inherent biases are understood and expected. The intention of this review is to take a critical look at the range of study designs that have been conducted on this under-researched topic in order to present a full picture of the current knowledge base.

3. Results

Seventeen studies were included in this review (see Figure 1). Of the included studies, nine were randomised controlled trials, two were qualitative interviews, two were case studies, one was a case series, one was a case report, one was a non-randomised controlled trial, and one was an uncontrolled pilot study. The results are presented here organised by study design, with key findings highlighted. Details about the intervention, participant demographics, and results are described in the text and summarized in Table 1.

Table 1.

Summary of Findings

Author (Year), study design	Number of participants	Age (Years)	Sex	BMI	Intervention	Muscular loading	Duration	Key findings
Touyz et al. (1993) NRCT	INT = 19 (4 weeks), 17 (6 weeks) CTRL: n = 20	INT: 15.942±2.45 CTRL: 20.00±5.28	Not reported	14.82±1.01	3 hrs/week of structured exercise including stretching, posture enhancement, weight training, social sport, and occasional low-impact aerobic activities	Not reported	4-6 weeks	No significant between-group difference in rate of weight gain at 4 or 6 weeks
Brooks et al. (1999) Case study	n = 1	41	Female	16.9	- Heavy RT 3x/week - Daily jumping on a hard surface for high-impact exercise - 800 mg calcium supplementation	high loading (80% 1RM)	12 months	BMD and muscular strength increased.
Szabo et al. (2002) RCT	All groups: n = 7	All groups: 20-22	Not reported	15.1±1.1	Two 60-min RT sessions/week using BW, resistance bands, and 2.5 kg dumbbells	Low loading (exact measures unspecified)	8 weeks	- Body mass, BMI, BF%, FM, LBM significantly increased in both INT and CTRL. - ED psychopathology significantly improved in both groups.
Chantler, et al. (2006) RCT	All groups: n = 7	INT: 20±5 CTRL1: 22±6 CTRL2: 23±3	Female	15.1 ± 1.1	Two 60-min RT sessions/week using BW, resistance bands, and 2.5 kg dumbbells	Low loading (exact measures unspecified)	8 weeks	INT significantly increased peak torque of knee extensor, knee flexor, and elbow flexor.
Fernández-del-Valle et al (2010) RCT	Both groups: n = 11	INT: 14.7±0.6 CTRL: 14.2±1.2	20 females, 2 males	18.7±1.7	- Two RT sessions/week for 60-70 mins each - BW, free weights, and machines used for resistance	Low loading (20-60% 6RM, gradually progressed over 12 weeks)	12 weeks	No significant differences between intervention and control groups.
Fisher et al. (2012) Case report	n = 1	48	Female	8.9	30-min sessions 4-6x/week training flexibility, strength training, functionality, postural stability, and endurance	Low loading (BW, resistance bands)	9 weeks	- BMI increased. - Pt improved from needing total assistance for ADLs to modified independence. - Fall risk decreased. - Gait speed increased.
Fernández-del-Valle, et al (2014) RCT	Both groups: n = 18	INT: 12.61±0.59 CTRL: 13.00±0.60	Female	17.28±2.55	- Three 50- to 60-min RT sessions/week - Exercises: lateral row, bench press, leg press, leg extension, lateral pull-down, abdominal crunch, low back extension, push-ups	High loading (3 sets of 8-10 repetitions at 70-100% 6RM. gradually progressed)	8 weeks	INT increased leg press, lateral row by, and bench press.
Fernández-del-Valle, et al (2016) RCT	Both groups: n = 18	INT: 13.0±0.6 CTRL: 12.6±0.6	Female	>14.0 kg/m2	- Three 50- to 60-min RT sessions/week - Exercises: lateral row, bench press, leg press, leg extension, lateral pull-down, abdominal crunch, low back extension, push-ups	High loading (3 sets of 8-10 repetitions at 70-100% 6RM. gradually progressed)	8 weeks	- INT increased skeletal muscle mass and ratios of strength to BW and strength to skeletal muscle mass - RT group had slow, sustained fat gain; control group had rapid fat gain
Bratland-Sanda, et al. (2018) Case study	n = 1	25	Female	17.6	- Three 60-min RT sessions/week - Exercises: deadlift, squat, bench press, pull-down	Maximal loading (3 sets of 5 repetitions at 5RM)	16 weeks	- Strength increased in upper body and legs. - BMD increased in lumbar spine and right and left femoral neck. - No change in ED psychopathology, though pt reported enjoying the focus on strength over cardio.
Bakland et. al (2019) Interviews	n = 15	19-42	Female	17.5-35 at time of PED-t intervention	Participants were interviewed after taking part in PED-t protocol.	Medium (10RM) to heavy (up to 2RM, progressed gradually) loading	16 weeks	Participants reported positive experiences regarding “obtaining tools” and “developing new perspectives”, “planning and framing”, and “therapists’ competence.”
Mathisen, et al. (2020) RCT	INT: n = 76 CTRL: n = 73	INT: 28.2±6.2 CTRL: 27.7±5.3	Female	17.5-35	PED-t protocol: - 2 RT sessions and 1 interval running session per week, 40-60 mins each session - 20 dietary therapy sessions	Medium (10RM) to high (up to 2RM, progressed gradually) loading	16 weeks	Similar improvements were seen in both groups in ED psychopathology and satisfaction with life. Depression scores significantly reduced only in INT. Between-group difference was not significant.
Galasso et al. (2020) RCT	INT: n = 10 CTRL: n = 9	INT: 54±11 CTRL: 53±13	Female	38±6	Four 90-min sessions per week: 60 mins aerobic activity, 20 mins strength training, 10 mins cool-down	Not reported (likely low due to emphasis on aerobic training)	6 months	- ED psychopathology decreased in both groups. - Both groups improved endurance, with greater improvements in INT.
Martínez-Sánchez et al. (2020) Uncontrolled pilot study	n = 12	14.6±1.7	Female	19.6±2.2	Three 60-min Pilates classes/week	Low loading (BW)	10 weeks	- Body dissatisfaction scores significantly decreased. - QoL domain “physical well-being” increased and “autonomy and parent relation” decreased.
Hockin-Boyers et al. (2021) Interviews	n = 19	17-38	Female	Not reported	Qualitative interviews with women who cited weightlifting as a factor in their ED recovery	Not reported (likely medium to high loading based on training styles described)	N/A	Participants described improvements related to “structure and managing routines,” “what the body can do,” and “regulating affect”.
Agne, et al (2022) RCT	INT: n = 19 CTRL: n = 22	12.78±0.88	Female	17.0±2.1	- Three 50-min RT sessions/week - Exercises: bench press, leg press, lateral row, leg extension, lateral pull down, abdominal crunch, low back extension and push-ups	High loading (3 sets of 8-10 repetitions starting at 70% 6RM and progressed)	8 weeks	- INT had greater improvements than CTRL in QoL. - INT improved mental health scores to no longer be significantly different from normative values. - INT significantly increased arm circumference.
Healy et al. (2024) Case series Inpatient	Case 1: n = 1 Case 2: n = 1 Case 3: n = 1	Pt 1: 18 Pt 2: 29 Pt 3: 20	Female	14.5-16.3	- 3 RT session/week including leg press, bench press, pulldown, and plyometric dropdowns	Maximal loading (85% 1RM for 3 sets of 5 repetitions)	6 weeks	- Pts 1 and 3 increased strength in all exercises, BW, and BMI - Pt 2 increased strength in 2 out of 3 exercises, decreased BW and BMI - All pts decreased anxiety around meals
Minano-Garrido et al. (2025) RCT	Both groups: n = 12	INT: 22.00±9.02 CTRL: 22.49±9.04	Female	15.01±1.79	- 3 sessions/week for 20-30 mins - Manual counter-resistance and body awareness/relaxation exercises	Low loading	4 weeks	- INT had greater QoL improvements. - INT had greater increases in strength in 3 of the 13 muscles tested.

%FM: percent fat mass;

AN: anorexia nervosa; AN-R: anorexia nervosa, restrictive type; BDI: beck depression inventory; BED: binge eating disorder; BF: body fat; BMD: bone mineral density; BMI: body mass index; BN: bulimia nervosa; BW: body weight; CIA 3.0: clinical impairment assessment; EDE-Q: eating disorder examination questionnaire; EDI: eating disorder inventory; EDNOS: eating disorder not otherwise specified; FM: fat mass; LBM: lean body mass; QoL: quality of life; RT: resistance training; SF-36: short form 36; SMM: skeletal muscle mass; SWLS: satisfaction with life scale.

3.1 Randomised controlled trials

3.1.1 Szabo and green (2002)

Szabo and Green (2002) studied the effects of an 8-week RT program on hospitalised patients with AN [43]. The program consisted of two 60-min sessions per week using low loading (bodyweight, resistance bands, 2.5 kg dumbbells).

3.1.1.1 Psychopathology findings

ED psychopathology as assessed using the Eating Disorder Inventory significantly improved in both the RT (p = 0.002) and control (p = 0.0112) groups. Beck Depression Inventory scores showed significant improvement in the control group (p = 0.0128) and not the RT group (p = 0.2188).

3.1.1.2 Body composition findings

Body weight, body mass index (BMI), body fat percentage, fat mass, and lean body mass significantly increased in both the RT group and the control group, with no significant difference between groups [43].

3.1.2 Chantler et al. (2006)

Chantler et al. (2005) studied the effects of the same low-loading RT protocol used by Szabo and Green (2002) on muscular strength and body composition in hospitalised patients with AN [44].

3.1.2.1 Muscular strength findings

Strength was assessed by measuring peak torque of elbow and knee extensors and flexors. The RT group significantly increased peak torque of the knee extensors (p < 0.001), knee flexors (p < 0.0001), and elbow flexors (p < 0.01). These measures did not significantly increase in the control groups. There was no significant change in peak torque of elbow extensor in any group [44].

3.1.2.2 Body composition findings

Body weight, BMI, body fat percentage, fat mass, and lean body mass significantly increased in both the RT group and the control group, with no significant difference between groups [44].

3.1.3 Fernández-del-Valle et al. (2010)

Fernández-del-Valle et al. (2010) measured the effects of RT with low loading (20–60% 6RM) on anthropometrics, body composition, muscular strength, functional mobility, and quality of life in outpatients with AN-R. The program involved two supervised RT sessions per week for 12 weeks. The only signficant improvement that could be solely attributed to the RT program was the lateral row strength test (p = 0.009 group x time effect). There were no significant differences between the groups in any other measure [45].

3.1.4 Fernández-del-Valle et al. (2014)

Fernández-del-Valle et al. (2014) looked at the effects of RT with higher loads on patients with AN-R attending a structured day program [46]. The experimental group participated in three supervised RT sessions per week for eight weeks and consumed an additional 150-kcal high-protein milkshake to account for added energy expenditure during exercise. RT sessions included exercises such as bench press, leg press, and lateral row, which were performed for 3 sets of 8–10 repetitions. Higher loading was utilised, beginning at 70% 6-repetition maximum (6RM) and increasing as participants progressed.

3.1.4.1 Muscular strength and agility findings

Strength was measured by testing 6RM for leg press, bench press, and lateral row. The intervention group increased leg press by 52%, and bench press by 41%, and lateral row by 37%. The group x time interaction was significant (p = 0.000 for each). Agility was measured using the Timed Up and Go (TUG) 3 m and 10 m tests and the Timed Up and Down Stairs (TUDS) test. The RT group significantly improved TUG-10m pre- to post-intervention (p = 0.003). No significant main or interaction effects were found. However, a magnitude-based analysis showed high probability of a beneficial effect across time in the RT group for all agility tests, while low probability of any effect across time was seen for the control group [46].

3.1.4.2 Body composition findings

There were no significant differences in body weight or BMI between groups [46].

3.1.5 Fernández-del-Valle et al. (2016)

Fernández-del-Valle et al. (2016) investigated the effects of RT on body composition and strength in adolescent females with AN-R participating in a structured day program [47]. The program was the same 8-week protocol described in Fernández-del-Valle et al. (2014), including high loads beginning at 70% 6RM and gradually increasing to 100% 6RM. A 150-kcal milkshake was also added to the RT group’s diet to account for energy expenditure during exercise.

3.1.5.1 Muscular strength findings

Ratios of strength to body weight (BW) and strength to skeletal muscle mass (SMM) were measured by calculating the kilograms of weight loaded per kilogram of BW or SMM. The RT group significantly increased both these measures (p < 0.001), while the control group did not.

3.1.5.2 Body composition findings

Height, BW, and BMI significantly increased in both groups. The RT group showed significant improvement in SMM after the intervention (p = 0.045, d = 0.60), whereas the control group showed no changes in this measure. The control group had a rapid increase in fat mass, while the intervention group had slow but sustained increases [47].

3.1.6 Mathisen et al. (2020)

Mathisen et al. conducted a trial in which outpatients diagnosed with BN or BED participated in 16 weeks of either enhanced cognitive behavioural therapy (CBT-E), which is a form of CBT that targets ED psychopathology, or the authors’ novel treatment approach involving physical exercise and dietary therapy (PED-t) [48–50]. The intervention protocol included three exercise sessions per week, two of which were RT sessions supervised by qualified trainers [48, 49]. Loading was medium to high, beginnig at 10RM and increasing to 2RM. The third weekly session was unsupervised and consisted of pyramid interval running (intensive work periods followed by active rest, increasing the duration of work periods from 1 to 4 min) [49]. The PED-t protocol also included 20 dietary therapy sessions over the course of the 16-week trial [48, 49].

3.1.6.1 Psychopathology findings

Both the PED-t group and the CBT group significantly improved ED psychopathology as measured by the Eating Disorder Examination Questionnaire, Clinical Impairment Assessment, and Satisfaction with Life Scale. Depression, as measured by the Beck Depression Inventory, significantly improved only in the PED-t group (p < 0.001; g = 0.93). Both groups signifcantly reduced driven exercise, binge eating, and self-induced vomiting. The number of patients in full remission increased from baseline to post-trial in both the PED-t (p < 0.001) and CBT (p = 0.001) groups [48].

3.1.7 Galasso et al. (2020)

Galasso et al. (2020) examined the effects of 6 months of combined aerobic and anaerobic exercise training (CAAET) on patients with BED attending a structured day program [51]. The intervention group and control group both received CBT and dietary support, while only the intervention group participated in CAAET. The CAAET program consisted of four 90-min workouts per week for 6 months. The workouts included 60 mins of aerobic activity (i.e. walking), 20 mins of strength training, and a 10-min cool-down. The loading of the strength exercises was not reported. However, given that aerobic exercise made up the bulk of the workout, loads were likely low. Additionally, the authors assessed strength using a squat test that scored patients based on the number of bodyweight squats they could do in 30 seconds [51]. As this test focuses on muscular endurance rather than maximum strength output, it is unlikely heavy loads were used in training.

3.1.7.1 Psychopathology findings

Psychopathology was measured using the Binge Eating Scale (BES) and the Bulimic Investigatory Test Edinburgh (BITE). BITE measures both binge eating and purgative behaviors and provides scores for symptoms and severity. Both the CAAET and control groups significantly improved in all of these measures (p < 0.05). The effect size between groups for BITE severity was large (d = − 0.8) [51].

3.1.7.2 Body composition, strength, and aerobic fitness findings

Both groups signficantly reduced BW and BMI, with no significant difference between groups. Strength, as measured using the squat test described above, significantly improved in both groups. The difference between the groups was not significant, although the effect size was large (d = 0.8). Aerobic fitness was measured using the Six-Min Walk Test, in which participants walked as quickly as they could for six minutes without running or jogging. Both groups significantly improved, with greater improvements seen in the CAAET group compared with the control (p < 0.05) [51].

3.1.8 Agne et al. (2022)

Agne et al. (2022) compared the effects of progressive RT on QoL and anthropometrics in adolescents diagnosed with AN-R who were part of a structured day program with those in a control group [52]. Both groups received the same cognitive behavioral therapy (CBT) and diet monitoring. The experimental group participated in three supervised 50-min RT sessions per week for eight weeks and consumed an additional 150-kcal high-protein milkshake to account for added energy expenditure during exercise. RT sessions included exercises such as bench press, leg press, and lateral row, which were performed for 3 sets of 8–10 repetitions. Loading began at 70% 6-repetition maximum (6RM) and increased as participants progressed.

3.1.8.1 Quality of life findings

QoL was assessed pre- and post-intervention via the health-related quality of life questionnaire short-form 36 (SF-36). The questionnaire utilises eight domains (physical functioning, role limitations due to physical health, bodily pain, general health, vitality, social functioning, role limitations due to emotional problems, mental health) to develop a score from each. These scores contribute to a physical component scale (PCS) and a mental component scale (MCS). At baseline, both the experimental and control groups had significantly lower SF-36-derived scores than normative values in all categories except the PCS. “Normative values” refers to healthy standards established by the SF-36. Significant group x time effects were seen on SF-36 domains “role limitations due to physical health” (p = 0.018; d = 0.8) and “physical functioning” (p = 0.031; d = 0.72). Mental health scores in the RT group improved to no longer be significantly different from normative values after the intervention (p = 0.080), an improvement that was not observed in the control group [52].

3.1.8.2 Body composition findings

Body weight significantly increased in both the RT (p = 0.022) and control groups (p = 0.012). Only the RT group significantly increased upper arm circumference both contracted (p = 0.013) and relaxed (p = 0.021).

3.1.9 Minano-Garrido et al. (2025)

Minano-Garrido et al. (2025) examined the effects of four weeks of low-loading RT combined with body awareness exercises on inpatient females with AN-R or AN-BP [53]. The intervation, called Anorexia Nervosa Inpatient Physiotherapy Adapted Progream (ANIPAP), involved three sessions per week for 20–30 mins per session. The authors specify that loading was low, using manual counter-resistance of upper and lower limbs, as well as trunk rotation. The program also included body awareness, breathing, and relaxation exercises [53].

3.1.9.1 Quality of life findings

Health-related QoL was measured using the Eating Disorger Quality of Life Test, which includes four subscales: Psychological, Physical/Cognitive, Financial, and Work/School. Significant group x time interactions were reported for the total score, as well as for the psychological subscale (p < 0.05).

3.1.9.2 Anthropometric, strength, and aerobic fitness findings

BW and BMI increased in both groups, with no significant difference between groups. The Six-Min Walk Test showed no significant difference within or between groups. The authors measured pulmonary function by testing peak respiratory flow. The intervention group improved pulmonary function (measured by testing peak respiratory flow), with a significant group x time effect (p = 0.033). Strength was measured byassessing hand grip strength and performing manual muscle testing (MMT). MMT is a low-impact method to assess muscle strength [54, 55]. Hand grip strength significantly increased in both groups, with no difference between them. MMT showed significant group x time interactions for triceps brachii (p = 0.00), thumb opposition (p = 0.015), and quadriceps (p = 0.013).

3.2 Case studies

3.2.1 Brooks et al. (1999)

Brooks et al. (1999) measured the effects of 12 months of RT with high loading (80% 1RM) on body composition and muscular strength of a 41-year-old female with AN [56]. The trial took place in an outpatient setting. The intervention included heavy RT three days per week, as well as high-impact exercise in the form of daily jumping on a hard surface. Daily jumping was progressed by beginning with 10 jumps per day and increasing to 75 jumps per day over the 12-month period. The patient was also supplemented with calcium (800 mg) to promote bone remodeling. Diet counseling was provided intermittently, and the patient was instructed to increase energy intake by a minimum of 15%. However, assessments of her diet done via 24-hour recalls at the beginning, middle, and end of the program showed she may not have been meeting her energy needs.

3.2.1.1 Body composition findings

BMD, assessed using Dual-energy X-ray absorptiometry (DEXA), increased significantly (p < 0.05) in the trochanter, intertrochanteric region, and proximal femur. There were no significant BMD changes in the femoral neck or lumbar spine. Percent body fat decreased by 1.7%, and lean mass increased by 3%. Body weight remained about the same (+0.4 kg) [56].

3.2.1.2 Muscular strength findings

Muscular strength increased by 18% in the hamstrings, 116% in the biceps, 99% in the triceps, and 118% in the back extensors [56].

3.2.2 Bratland-Sanda et al. (2018)

Bratland-Sanda et al. (2018) examined the effects of RT using maximal loads on BMD, strength, and psychopathology of a 25-year-old woman who had a nine-year history of AN-R and met the criteria for osteopenia. The woman was not undergoing any other treatment at the time of the study. The program involved three 60-min RT sessions per week for 16 weeks. The workouts included deadlifts, squats, bench press, and pull-down, each performed for three sets of five repetitions at maximal loading (5RM). Prior to the intervention, the patient reported performing moderate to vigorous endurance exercise approximately six days per week. She was allowed to continue exercising outside of the program, and these sessions were logged and reported. Although she was counseled to increase energy intake, she did not adhere to these guidelines.

3.2.2.1 Psychopathology findings

ED psychopathology as measured via the Eating Disorders Examination 16.0 did not change. Maladaptive exercise was measured using the Compulsive Exercise Test and also showed no change. However, the patient reported enjoying the feeling of getting stronger and being able to move her body in different ways. Her exercise logs showed that she had continued to do vigorous endurance exercise outside of the strength sessions for the first three weeks of the intervention. After three weeks, she reduced the intensity and volume of these additional workouts, transitioning to moderate-intensity walking. She reported that this change was due to noticing how vigorous running affected her ability to lift heavy weights in her strength sessions.

3.2.2.2 Strength and body composition findings

Strength, measured by testing 1RM, increased by 36% in the upper body and 30% in the legs. BMD, assessed using DEXA, increased by 4% in the lumbar spine, 4% in the right femoral neck, and 3% in the left femoral neck. There was no significant change in body weight or BMI.

3.3. Case report

3.3.1 Fisher et al. (2012)

Fisher et al. (2012) reported on a unique physical therapy protocol used in the care of a 48-year-old female patient with AN in an acute care hospital setting. The patient participated in 30-min physical therapy sessions four to six times per week for nine weeks. The sessions included strength training in addition to flexibility training, functional training, postural stability training, and endurance training. Strength training involved low loads using body weight and resistance bands, progressed as tolerated [57].

3.3.1.1 Body composition and functionality findings

At admission, the patient weighed 42% ideal body weight and had a BMI of 8.9. At discharge, she was at 64% ideal body weight with a BMI of 13.2. Her functionality was assessed using the Functional Independence Measure, which scores functional independence in different categories from one (total assistance) to seven (complete independence). The patient’s score increased from one to six (modified independence) in each of the categories measured (bed, transfer, ambulation). The TUG test was used to assess functionality in tasks such as sit-to-stand transfers and turns. Her score improved from 19.27 seconds at admission to 11 seconds at discharge. Her fall risk and gait speed also improved from admission to discharge [57].

3.4 Case series

3.4.1 Healy et al. (2024)

Healy et al. (2024) report on three cases in which a six-week RT program with maximal loading was included in the treatment of patients with AN. The patients were women who were in either inpatient or day care treatment at the same facility. The patients participated in three RT sessions per week. The exercises performed were leg press, bench press, pull-down, and plyometric drop-downs. The exercises were performed for three sets of five repetitions set at maximal loading (85% 1RM). Two of the patients (Cases 1 and 3) were diagnosed with AN-R. The patient in Case 2 was diagnosed with AN-BP. This patient moved up to inpatient treatment at week 4 of the intervention, while the other two remained in the day care program.

3.4.1.1 Psychopathology findings

The first patient had improved ED psychopathology as measured by the Eating Disorder Inventory (EDI). Maladaptive exercise, measured via the Exercise Addiction Inventory (EAI), also improved. Her anxiety around meals decreased, as measured by the Visual Analogue Scale (VAS), whicih scores meal-based anxiety. The second patient’s EDI score worsened, and her EAI score had no change. Her anxiety around meals decreased. The third patient’s EDI score improved, as did her anxiety around meals. Her EAI score had no change.

3.4.1.2 Muscular strength and body composition findings

The first and third patients increased their strength (measured by testing 1RM) for all exercises. They also both increased body weight and BMI. The second patient increased strength for two out of the three exercises, but decreased for the pull-down. Her body weight and BMI also decreased.

3.5 Non-randomised controlled trial

3.5.1 Touyz et al. (1993)

Touyz et al. (1993) examined the effects of a structured exercise program that included RT on the rate of weight gain in a cohort of patients with AN in an inpatient unit. The cohort following this group acted as the control. The intervention included three hours per week of structured exercise for four to six weeks (depending the patients’ length of stay). In addition to weight training, the program included stretching, posture support, social support, and occasional low-impact aerobic activities. The loading was not specified, though the authors noted that weight training was the cornerstone of the intervention program and was individualised for each patient. There were no significant between-group differences in rate of weight gain at either 4 or 6 weeks.

3.6 Uncontrolled pilot study

3.6.1 Martínez-Sánchez et al. (2020)

Martínez-Sánchez et al. (2020) examined the effects of a pilates program on psychopathology, body image disturbance, and quality of life in patients with AN [58]. The participants were female patients at the same hospital, and were diagnosed with either AN-R or atypical AN. The program involved three 60-min pilates classes per week for 10 weeks. Pilates is a modality that primarily uses body weight, so loading was low. Each class focused on a pilates principle, including: core control, axial elongation, spinal articulation, dissociation of movement, limb alignment and coordination, and scapular alignment. Movements were each repeated 10 times.

3.6.1.1 Psychopathology, body image disturbance, and quality of life findings

ED psychopathology, measured using the Eating Disorder Inventory, did not significantly decrease. Body image disturbance was assessed using the Contour Drawing Rating Scale, which measures the discrepency between an individual’s perception of their own body shape and the body shape they would like to have. These scores significantly improved (p = 0.046). Health-related quality of life was evaluated using KIDSCREEN-27, which is a validated tool to assess QoL in adolescents. The results showed a significant increase in the category “physical well-being” (p = 0.008), and a significant decrease in the category “autonomy and parent relation” (p = 0.021). The authors speculate that this decrease may indicate a beneficial shift toward a more autonomous and self-aware mindset. Patients reported high satisfaction with the program via a questionnaire (12.8±1.9 out of a possible 15 points) [58].

3.7 Qualitative studies (interviews)

3.7.1 Bakland et. al (2019)

Bakland et. al (2019) conducted a follow-up study to the PED-t protocol mentioned above [59]. This was a qualitative investigation in which patients who particpated in the aforementioned PED-t protocol were interviewed in order to understand their perceptions of the experience. The participants reported positive feelings about the use of the PED-t program in “obtaining tools” and “developing new perspectives.” Many participants reported that they previously had been doing cardio as their form of exercise, which they found exhausting, and they were constantly thinking about ways to burn more calories. They stated that the resistance exercises they learned during the program were more energising, and their goals shifted from burning calories to achieving a strong and healthy body. Some participants noted they enjoyed feeling stronger, and that this also made them feel stronger mentally. Opinions were mixed on the group environment. Most participants reported positive experiences regarding “learning and comfort in meeting others” and “motivating and supporting each other”. However, others found it frustrating when time was spent discussing topics that did not relate to them. They also noted some people required more support than others, which resulted in the therapists’ attention being unevenly divided [59].

3.7.2 Hockin-Boyers et al. (2021)

Hockin-Boyers et al. (2021) conducted a qualitative study of 19 women who cited weightlifting as a key part of their ED recovery. The participants were amateur lifters of different styles of weightlifting, including strength training, powerlifting, bodybuilding, CrossFit, and strongwoman. All the participants reported having suffered from various forms of EDs, including AN, BN, BED, eating disorder not otherwise specified (EDNOS), or a combination of multiple EDs. EDNOS is a diagnosis that was included in the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) and was replaced by OSFED and unspecified feeding or eating disorder in the DSM-V [60]. Slightly more than half of the participants reported having undergone formal ED treatment, whereas the others had never received professional support. Many of the women who had accessed formal treatment reported difficulty engaging with their treatment protocols, and reported weightlifting to be the major impetus in their recovery. The participants cited various aspects of weightlifting as beneficial for them. Themes identified from participant responses included “Structure and Managing Routines,” “What the Body Can Do,” and “Regulating Affect”. The responses related to the first theme (Structure and Managing Routines) mentioned finding the structure reassuring. The participants stated that having a clear plan to follow with a trajectory for progress, as is typical of progressive strength programs, allowed them to feel calmer about taking rest days because they understood the importance of recovery and its role in their training goals. Responses categorised under “What the Body Can Do” mentioned that the culture around this type of training was particularly useful for reframing ideas around exercise and body ideals with its emphasis on “gains,” including building muscle and improving strength and performance. Along these lines, food is framed as working with the body and necessary for the desired “gains”. Responses under the “Regulating Affect” theme discussed how weightlifting helped them clear the mind and feel more present in their bodies. The participants noted that lifting made them feel empowered and accomplished. They also stated it was an outlet for emotions and provided stress relief. They reported increased self-confidence and a more positive attitude as a result of RT [61].

4. Discussion

The available research shows RT can be an effective tool to improve body composition, psychopathology, and quality of life in patients with AN, BN, or BED, though clinical considerations for the individual must be taken into account.

4.1 Body composition

Loading appears to be an important variable in ensuring the effectiveness of an RT program on body composition. Of the studies using low loading, only the case report presented by Fisher et al (2012) showed significant improvements in key measures of functionality. Although this was considered to be low loading due to the use of body weight and bands as resistance, it was sufficiently challenging, progressive, and individualised to elicit improvements in a patient who was severely malnourished [57]. Brooks et al. (1999) and Bratland-Sanda et al. (2018) both showed significant improvements in BMD using high to maximal loads [56, 62]. This is promising given the prevalence of reduced BMD in these populations. Among adolescents with AN, 50% have osteopenia and 25% have osteoporosis [31]. Ninety percent of adult females with AN have reduced bone density, and 38% have osteoporosis [31]. RT plays an important role in building and maintaining bone density by providing a mechanical stimulus for bone building, an effect that is stronger in RT than in aerobic training [25, 63].

In addition to increasing BMD, high to maximal loads may be ideal for increasing muscle mass and improving fat distribution, as shown by Fernández-del-Valle et al. (2016) [47]. Fat regain after a period of starvation, as is the case in AN, favours visceral fat accumulation [64]. While fat eventually redistributes after approximately one year of recovery, fat accumulation at the belly may be psychologically distressing to patients and contribute to relapse [64, 65]. RT has been shown to reduce visceral fat while preserving lean mass [66]. The results seen in Fernández-del-Valle et al. (2016) suggest this benefit may occur when RT is included in treatment [47].

4.2 Psychopathology and quality of life

RT seems to either improve or not effect ED psychopathology, with most results showing no difference between the intervention and control groups in the various scoring tools used. Still, the lack of negative impact is promising. The PED-t protocol described by Mathisen, et al. (2020) was shown to be as effective as CBT-E at improving ED psychopathology, depression, and satisfaction with life. CBT is one of the most researched and utilised methods used in ED treatment, with CBT-E in particular standing out for its efficacy [50, 67, 68]. However, only 40% of patients with BN and 50% of patients with BED abstain from behaviors after receiving CBT or other psychological treatments [69, 70]. PED-t may be a viable alternative for those who do not respond to CBT. Patients who participated in the PED-t protocol emphasized the value of learning from qualified professionals how to appropriately incorporate exercise and fuel their bodies [59]. This allowed the patients to let go of misinformed beliefs about nutrition and fitness and form new habits in these areas. Incorporating intentional, well-planned exercise may help reduce depression and increase feelings of self-efficacy [34, 71]. PED-t offers a way to put these skills into practice.

The three studies that assessed QoL all showed RT to be beneficial [52, 53, 58]. Effects on QoL may be greater when higher loads are used in training. This may be related to greater improvements in strength, as high loading has been shown to result in greater strength increases than low loading [72]. Higher muscular strength has been associated with higher QoL in non-ED populations, such as older adults and breast cancer survivors [73, 74].

Qualitative studies show that RT has potential to meaningfully impact the lives of people with EDs. Hockin-Boyers et al. (2021) and Bakland et al. (2019) both described a shift in mindset toward building strength and fueling for health over burning calories and getting smaller [59, 61]. In this way, food and exercise may be reframed in the context of what they do for the body. This practical application of training skills is a feature that sets RT programs such as PED-t apart from CBT-E.

4.3 Clinical considerations

Supervised RT does not appear to detract from recovery goals and may offer significant benefits. RT did not impair weight restoration in any controlled study. However, Healy et al. (2024) described three patients in a case series, one of whom decreased in both BW and BMI over the course of the 6-week trial. Of note, this patient had a longer duration of illness (seven years) than the other two patients included in the case series (three years and one year). This patient was also regressing in her condition, as evidenced by her admission to inpatient care four weeks into the study. This suggests that patients whose conditions worsen over the course of an RT program may not be safe to continue with RT. Indeed, some of the RCTs included in this review stipulated as part of continuation criteria that patients not experience significant losses in weight or BMI due to voluntarily reducing their caloric intake or develop maladaptive exercise behaviors [46, 47, 52]. These may be resonable markers to determine if an individual is safe to continue participating in RT.

The available research also demonstrated the importance of having a skilled, ED-informed exercise therapist design and supervise training [49, 59, 62, 75]. Participants may be more likely to buy into the program if they trust their support team, including their trainer [59, 61, 62]. This allows for the proper intensity and volume to be prescribed for positive adaptations to occur without overtraining, thus emphasising recovery and diminishing maladaptive exercise [48, 59, 61, 62, 76]. Currently, exercise professionals do not receive education about eating disorders in their curriculum [75]. As physical therapy is not considered part of the standard of care for eating disorders, few opportunities exist for these professionals to further their learning in this area [75]. Thus, if an exercise therapist is to be included in the care team, it is important that they receive training on working with ED patients.

An important factor to consider is the role of nutrition in RT and ED recovery. The included studies show that patients have better outcomes when they have adequate nutritional intake and receive dietary support [46–48, 52]. Studies in which participants did not follow dietary recommendations resulted in insufficient weight restoration [56, 62, 77]. Without adequate nutrition, exercise (including RT) results in further breakdown of body tissues and a heightened state of stress [78].

Clinical judgement should be used to determine when in the disease sequelae to introduce RT, the appropriate volume and intensity, and how much supervision is needed. The SEES (Safe Exercise at Every Stage) framework can be used as a guideline for these markers, although these recommendations do not replace clinical judgement [17]. This framework uses a psychological and physical assessment to identify a patient’s risk level, ranging from level A (high risk) to level D (low risk). Under this framework, patients at high risk have severely compromised laboratory work, which may include bradycardia or tachycardia, hypotension, and electrolyte imbalances, among other markers. These patients must be under medical supervision and should limit exercise to static stretching [17].

4.4 Strengths and limitations

To the authors’ knowledge, this review is the first to specifically assess the evidence regarding the use of RT in the treatment of eating disorders. This adds to the available knowledge base to inform future research and practice. The inclusion of both quantitative and qualitative literature allows for interpretation of both objective outcomes and subjective experiences.

A limitation of the current review is that only one author, V.B., conducted the literature search, screened and selected records for inclusion, and assessed the available research. This increases the risk of bias due to potential blind spots and limited perspectives. Future reviews would benefit from multiple authors handling these tasks.

The risk of bias inherent in the studies examined is another limitation. As research was included that that was non-randomised, uncontrolled, subjective, or had small sample sizes, results cannot be generalised. Additionally, many of the included studies were conducted by similar groups of authors. More variety in research groups would provide more information about the reproducibility of the findings in different settings and reduce the effects of a single group’s particular biases. There was also little diversity among the participants. Almost all the participants were female, and most were under 30 years old. The findings therefore cannot be extrapolated to broader populations.

4.5 Recommendations for future research

Future research should include more diverse populations and longer intervention periods. Studying the effects of RT in males with AN, BN, or BED would be particularly valuable, as males are more prone to muscle dysmorphia and thus may be more likely to develop maladaptve exercise habits with RT [79]. Studies involving longer intervention periods would allow more time to demonstrate how both physical and psychological responses to training may change over time. The available research points to greater loads being more effective at producing results than lighter loads, but this must be shown in more studies conducted by different groups before it can be confidently recommended. Examining the effects of greater loads vs. lighter loads specifically would be a valuable addition to the literature. Additionally, future research should use gold-standard tools such as DEXA to measure body composition. This would improve the accuracy of measurements and create a standardisation so that results can be more precisely compared between studies.

5. Conclusions

The available research indicates that, in combination with support from a multidisciplinary team including a dietitian, psychologist, and medical provider, resistance training led by a qualified exercise therapist has the potential to play an important role in improving psychopathology, quality of life, and body composition in patients with AN, BN, or BED. Resistance training is a promising tool that may fill important gaps in current treatment methods, and is worthy of further study and consideration.

Appendix 1

PRISMA checklist

Section and Topic	Item #	Checklist item	Location where item is reported
TITLE
Title	1	Identify the report as a systematic review.	Line 2: narrative review
ABSTRACT
Abstract	2	See the PRISMA 2020 for Abstracts checklist.
INTRODUCTION
Rationale	3	Describe the rationale for the review in the context of existing knowledge.	Lines 56-117
Objectives	4	Provide an explicit statement of the objective(s) or question(s) the review addresses.	Lines 115-117
METHODS
Eligibility criteria	5	Specify the inclusion and exclusion criteria for the review and how studies were grouped for the syntheses.	Lines 126-128
Information sources	6	Specify all databases, registers, websites, organisations, reference lists and other sources searched or consulted to identify studies. Specify the date when each source was last searched or consulted.	Lines 121-126
Search strategy	7	Present the full search strategies for all databases, registers and websites, including any filters and limits used.	Lines 121-126
Selection process	8	Specify the methods used to decide whether a study met the inclusion criteria of the review, including how many reviewers screened each record and each report retrieved, whether they worked independently, and if applicable, details of automation tools used in the process.	Lines 141-142
Data collection process	9	Specify the methods used to collect data from reports, including how many reviewers collected data from each report, whether they worked independently, any processes for obtaining or confirming data from study investigators, and if applicable, details of automation tools used in the process.	Lines 141-142
Data items	10a	List and define all outcomes for which data were sought. Specify whether all results that were compatible with each outcome domain in each study were sought (e.g. for all measures, time points, analyses), and if not, the methods used to decide which results to collect.	N/A
	10b	List and define all other variables for which data were sought (e.g. participant and intervention characteristics, funding sources). Describe any assumptions made about any missing or unclear information.	N/A
Study risk of bias assessment	11	Specify the methods used to assess risk of bias in the included studies, including details of the tool(s) used, how many reviewers assessed each study and whether they worked independently, and if applicable, details of automation tools used in the process.	Lines 149-156
Effect measures	12	Specify for each outcome the effect measure(s) (e.g. risk ratio, mean difference) used in the synthesis or presentation of results.	N/A
Synthesis methods	13a	Describe the processes used to decide which studies were eligible for each synthesis (e.g. tabulating the study intervention characteristics and comparing against the planned groups for each synthesis (item #5)).	N/A
	13b	Describe any methods required to prepare the data for presentation or synthesis, such as handling of missing summary statistics, or data conversions.	N/A
	13c	Describe any methods used to tabulate or visually display results of individual studies and syntheses.	N/A
	13d	Describe any methods used to synthesize results and provide a rationale for the choice(s). If meta-analysis was performed, describe the model(s), method(s) to identify the presence and extent of statistical heterogeneity, and software package(s) used.	N/A
	13e	Describe any methods used to explore possible causes of heterogeneity among study results (e.g. subgroup analysis, meta-regression).	N/A
	13f	Describe any sensitivity analyses conducted to assess robustness of the synthesized results.	N/A
Reporting bias assessment	14	Describe any methods used to assess risk of bias due to missing results in a synthesis (arising from reporting biases).	N/A
Certainty assessment	15	Describe any methods used to assess certainty (or confidence) in the body of evidence for an outcome.	N/A
RESULTS
Study selection	16a	Describe the results of the search and selection process, from the number of records identified in the search to the number of studies included in the review, ideally using a flow diagram.	Figure 1
	16b	Cite studies that might appear to meet the inclusion criteria, but which were excluded, and explain why they were excluded.	N/A
Study characteristics	17	Cite each included study and present its characteristics.	Table 1
Risk of bias in studies	18	Present assessments of risk of bias for each included study.	N/A
Results of individual studies	19	For all outcomes, present, for each study: (a) summary statistics for each group (where appropriate) and (b) an effect estimate and its precision (e.g. confidence/credible interval), ideally using structured tables or plots.	Table 1
Results of syntheses	20a	For each synthesis, briefly summarise the characteristics and risk of bias among contributing studies.	N/A
	20b	Present results of all statistical syntheses conducted. If meta-analysis was done, present for each the summary estimate and its precision (e.g. confidence/credible interval) and measures of statistical heterogeneity. If comparing groups, describe the direction of the effect.	N/A
	20c	Present results of all investigations of possible causes of heterogeneity among study results.	N/A
	20d	Present results of all sensitivity analyses conducted to assess the robustness of the synthesized results.	N/A
Reporting biases	21	Present assessments of risk of bias due to missing results (arising from reporting biases) for each synthesis assessed.	N/A
Certainty of evidence	22	Present assessments of certainty (or confidence) in the body of evidence for each outcome assessed.	N/A
DISCUSSION
Discussion	23a	Provide a general interpretation of the results in the context of other evidence.	Lines 555-648
	23b	Discuss any limitations of the evidence included in the review.	Lines 660-667
	23c	Discuss any limitations of the review processes used.	Lines 656-659
	23d	Discuss implications of the results for practice, policy, and future research.	Lines 669-681
OTHER INFORMATION
Registration and protocol	24a	Provide registration information for the review, including register name and registration number, or state that the review was not registered.	Line 736
	24b	Indicate where the review protocol can be accessed, or state that a protocol was not prepared.	Lines 739-740
	24c	Describe and explain any amendments to information provided at registration or in the protocol.	N/A
Support	25	Describe sources of financial or non-financial support for the review, and the role of the funders or sponsors in the review.	Lines 746-747
Competing interests	26	Declare any competing interests of review authors.	Line 743
Availability of data, code and other materials	27	Report which of the following are publicly available and where they can be found: template data collection forms; data extracted from included studies; data used for all analyses; analytic code; any other materials used in the review.	N/A

From: Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. https://doi.org/10.1136/bmj.n71. This work is licensed under CC BY 4.0. To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/

Section and Topic	Item #	Checklist item	Reported (Yes/No)
TITLE
Title	1	Identify the report as a systematic review.	Yes (narrative review)
BACKGROUND
Objectives	2	Provide an explicit statement of the main objective(s) or question(s) the review addresses.	Yes
METHODS
Eligibility criteria	3	Specify the inclusion and exclusion criteria for the review.	Yes
Information sources	4	Specify the information sources (e.g. databases, registers) used to identify studies and the date when each was last searched.	Yes
Risk of bias	5	Specify the methods used to assess risk of bias in the included studies.	N/A
Synthesis of results	6	Specify the methods used to present and synthesise results.	N/A
RESULTS
Included studies	7	Give the total number of included studies and participants and summarise relevant characteristics of studies.	Yes
Synthesis of results	8	Present results for main outcomes, preferably indicating the number of included studies and participants for each. If meta-analysis was done, report the summary estimate and confidence/credible interval. If comparing groups, indicate the direction of the effect (i.e. which group is favoured).	Yes
DISCUSSION
Limitations of evidence	9	Provide a brief summary of the limitations of the evidence included in the review (e.g. study risk of bias, inconsistency and imprecision).	N/A
Interpretation	10	Provide a general interpretation of the results and important implications.	Yes
OTHER
Funding	11	Specify the primary source of funding for the review.	N/A
Registration	12	Provide the register name and registration number.	N/A

From: Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. https://doi.org/10.1136/bmj.n71. This work is licensed under CC BY 4.0. To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/

Abbreviations

AN-BP

Binge-eating/purging type anorexia nervosa

AN-R

Restricting type anorexia nervosa

AN

Anorexia nervosa

ANIPAP

Anorexia nervosa inpatient physiotherapy adapted progream

BED

Binge eating disorder

BES

Binge eating scale

BITE

Bulimic investigatory test edinburgh

BMD

Bone mineral density

BMI

Body mass index

BN

Bulimia nervosa

BW

Body weight

CAAET

Combined aerobic and anaerobic exercise training

CBT

Cognitive behavioral therapy

DEXA

Dual-energy X-ray absorptiometry

DSM

Diagnostic and statistical manual of mental disorders

EAI

Exercise addiction inventory

ED

Eating disorder

EDI

Eating disorder inventory

MCS

Mental component scale

ME

Maladaptive exercise

MMT

Manual muscle testing

OSFED

Other specified feeding or eating disorder

PCS

Physical component scale

PED-t

Physical exercise and dietary therapy

PRISMA

Preferred reporting items for systematic reviews and meta-analyses

QoL

Quality of life

RCT

Randomised controlled trial

RM

Repetition maximum

RT

Resistance training

SEES

Safe exercise at every stage

SMM

Skeletal muscle mass

TUDS

Timed up and down stairs

TUG

Timed up and go

VAS

Visual analogue scale

Biographies

Victoria Bongiorno

is a registered dietitian, certified personal trainer, and writer working in outpatient, intensive outpatient, partial hospitalisation, and residential levels of eating disorder care.

Martica Heaner

is a doctor of behavioural nutrition & physical therapy, health writer, and adjunct professor at Hunter College.

Author contributions

V.B. conceptualised the research, performed the literature review, interpreted the findings, and wrote the manuscript. M.H. provided guidance during the research process and made edits to the manuscript. All authors read and approved the final manuscript.

Funding

This research was funded by the authors. No outside funding was provided for this research.

Availability of data and materials

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable

Consent for publication

Not applicable

Competing interests

The authors declare no competing interests.