The Benefits of Physical Exercise on Mental Disorders and Quality of Life in Substance Use Disorders Patients. Systematic Review and Meta-Analysis

Physical exercise seems to have a promising effect on numerous variables related to the recovery of drug-dependent patients. However, some contradictions are found in the literature. The aim of this study was to perform a systematic review and meta-analysis in order to identify the effect of physical exercise on mental disorders, quality of life, abstinence, and craving, and make a comparison of the effect of exercise depending on the type of program. A search for articles was conducted using PubMed, Web of Science, and Scopus databases. Studies were selected that measured the acute effects or long-term effect (≥2 weeks) of exercise in patients who met criteria for alcohol use disorders or substance use disorders. A total of 59 studies were included. An effect of exercise on mental disorders (standardized mean differences (SMD) = 0.66 (confidence interval (CI): 0.46, 0.86); z = 6.50; p < 0.00001) and quality of life (SMD = 0.69 (95% CI: 0.53, 0.84); z = 8.65; p < 0.00001) was identified. Subgroup analysis revealed an effect of exercise in craving (SMD = 0.80 (CI: 0.07, 1.53); z = 2.15, p = 0.03), stress (SMD = 1.11 (CI: 0.31, 1.91); = 2.73; p = 0.006), anxiety (SMD = 0.50 (CI: 0.16, 0.84); z = 2.88; p = 0.004) and depression (SMD = 0.63 (CI: 0.34, 0.92); z = 4.31; p < 0.0001). Body–mind activities and programs based on improving physical conditions produced similar results in mental disorders and quality of life. Available evidence indicates that physical exercise, both body–mind and physical fitness programs, can be effective in improving mental disorders, craving, and quality of life in drug-dependent patients.


Introduction
Drug use is a worldwide problem that persists throughout the years. According to the World Drug Report (2016) about 250 million people between 15 and 64 years tried some type of drug in the year 2014. Among them, one in ten was a "problematic consumer", resulting in physical, psychological and/or social problems. Therefore, approximately 30 million people are suffering from problems derived from drug use (diseases, psychiatric disorders, physical deterioration, social exclusion, family problems, emotional disorders, etc.) [1].
The consequences of drug use include a serious deterioration in quality of life and mental health. This means that drug addiction rehabilitation treatments should not only be focused on achieving mere abstinence, but should also try to recover the quality of life of the patient and their physical and mental health [2].
Given this problem, performing physical exercise can be an effective tool to help in the prevention and treatment of drug dependencies at different levels. On one hand, physical exercise seems to be a preventative factor for the consumption of drugs. People who practice physical exercise on a regular basis have a lower rate of drug use problems and vice versa [3].
In addition to the effect on consumption and craving, exercise has also been proven effective as a method to improve mental disorders such as stress reduction [15], anxiety [16,17] and depression [16,18]. The DSM-5 defines a mental disorder as "a syndrome characterized by clinically significant disturbance in an individual's cognition, emotion regulation, or behavior that reflects a dysfunction in the psychological, biological, or developmental processes underlying mental functioning." As a consequence, there is usually a distress or impairment in personal, family, social, educational, occupational, or other important areas of functioning [19].
Likewise, some studies have observed that a program combining aerobic exercise with strength training can improve the quality of life in a drug-dependent population [6,20] and that yoga can be an equally valid tool for improving the quality of life [21]. In the same line, recent studies describe the influence that physical exercise can have at a cognitive level [22] and even in satisfaction of life [23]. Quality of life may be defined as "an individual's perception of their position in life in the context of the culture and value systems in which they live and in relation to their goals, expectations, standards and concerns. It is a broad ranging concept affected in a complex way by the person's physical health, psychological state, personal beliefs, social relationships and their relationship to salient features of their environment" [24].
However, benefits are not always obtained after participation in a physical exercise program, as recent studies show. For example, a varied physical exercise program of 12 weeks in heroin addicts undergoing rehabilitation was not proven to be effective to obtain significant benefits in stress, anxiety, and quality of life [25]. Therefore, it is likely that the type of exercise program used and the structural parameters of the exercise (type of activity, intensity, volume, or frequency) can play a key role in the effects obtained after participating in a physical exercise program.
Analyzing the literature, there are studies that have achieved positive effects using an aerobic exercise program [5,7,8,26], others that reach similar goals with a combined program of aerobic exercise or strength [6,20,27], some who have used a high-intensity interval training program [18] or others who have used activities related to meditation or relaxation, such as yoga, tai-chi or qigong with satisfactory results [9,21,[28][29][30][31].
In this way, generalizing about the effectiveness of the practice of physical exercise in the treatment of drug dependence can be complex, since the studies carried out employ different modalities of physical activity/physical exercise and different intensities, volumes and frequencies.
For this reason, the production of a review and meta-analysis of studies that investigate the influence of physical exercise in the population with substance use disorders (SUD) is of special importance, taking into account the different types of programs used, and identifying from a broad and global perspective the different benefits observed after participating in one exercise program or another.
In recent years, some review studies have been published trying to identify the benefits of physical exercise in participants who met the criteria for alcohol use disorders (AUD) or other substances use disorders [32][33][34][35][36][37][38] at different levels. Some meta-analysis have also been carried out with the same objective, focusing on SUD patients [39] and AUD patients [34]. These meta-analysis found, on the one hand, that physical exercise can increase the abstinence rate and reduce withdrawal symptoms, anxiety, and depression symptoms in SUD patients [39]. On the other hand, Hallgren and colleagues found that exercise can improve fitness and reduce depression symptoms in AUD patients [34].
However, to the best of our knowledge, there is no meta-analysis focusing on the benefits of exercise on quality of life, stress or craving in drug-dependent patients.
There are also very few studies that compare the effectiveness of one type of physical exercise/physical activity or another in terms of the benefits obtained. Aerobic exercise is the most commonly used type of exercise for drug addicts [34,39]. However, body-mind activities are increasingly used, such as yoga, tai chi, or qigong [33]. Both practices seem to be beneficial for the treatment of drug dependence, but it is unknown which of the two can generate greater benefits, and whether the use of aerobic exercise or body-mind practices may be more appropriate for aspects of mental disorders and quality of life. In this sense, Wang and colleagues [39] compared aerobic exercise with "mind-body exercise", concluding that both activities are effective for the improvement of abstinence, the symptoms of withdrawal syndrome, anxiety and depression, without appreciable differences between one type of activity and another. In the same way, they did not find differences regarding the type of intensity used. However, further research is required to identify the differences between types of exercise as well as the optimal intensity, frequency or duration of the programs, especially for the quality of life of this population.
Therefore, the objective of this review and meta-analysis was to perform an updated analysis of the benefits of physical exercise in AUD or SUD patients, quantify the effect of physical exercise on mental disorders, craving, and quality of life, and compare the effect of exercise programs with the most commonly used body-mind practices.

Method
This study follows the guidelines of the Preferred Reporting Items for Systematic reviews and Meta-Analysis (PRISMA) [40] and the Meta-Analysis and Systematic Reviews of Observational Studies (MOOSE) [41].

Research Strategies
For the selection of articles, a systematic search was made of articles published in PubMed, Web of Science and Scopus databases. Combinations of the key words "exercise, physical activity, qigong, taichi, yoga, substance use disorders, substance abuse, drug abuse, cocaine, cannabis, marijuana, alcohol, heroin, amphetamine, hypnotics, inhalants, hallucinogens, mdma, methamphetamine" were used.
The bibliographic references of the studies found were also examined manually to obtain new references. All of the identified articles were independently analyzed by two investigators of the present study, and the final selection of the articles to be included was taken by both, in a consensual manner.
We contacted some authors of the identified studies to ask for additional information or clarify some data for the meta-analysis.

Selection of Studies
Studies were selected that measured the acute effect of exercise (a single session) or experimental or quasi-experimental studies that measured the long-term effect of exercise in AUD or SUD patients. The studies that employed body-mind activities such as tai chi, yoga, or qigong, were also included.
Studies that fulfilled any of the following characteristics were discarded: studies conducted with animals, patients under 18 years of age, studies that exclusively analyzed tobacco addiction, review studies, observational studies or those that do not apply any program of physical exercise, and case studies. For the meta-analysis, studies without defined relevant statistical data, whose authors could not be contacted or did not provide the data when contacted, were also discarded.
All the studies addressed the analysis of any of the following variables: physical condition/ fitness/physiological markers, quality of life, stress, anxiety, depression, craving, abstinence/indications of consumption, other psychiatric or psychological variables. Figure 1 shows the process followed from the initial search to the final selection of the studies.

Data Extraction
The studies were thoroughly analyzed and the relevant data were collected in a table that contained the following categories: Population/substance of consumption, sample (experimental + control), type of exercise program, adherence to the program, measured variables and main results obtained.

Quality Analysis of the Studies
To define the quality of the studies, the widely used and validated "Physiotherapy Evidence Database (PEDro)" scale was taken into account [42][43][44]. This scale is based on the Delphi list [45], but it adds two criteria (items 8 and 10) that are not included in the Delphi list and are important when analyzing the quality of the studies. Of the 11 items on the PEDro scale, 5 and 6 were discarded for the analysis of the studies, as did other similar previous studies, among them, Wang and colleagues [39], since the blinding of patients and the instructors regarding the type of treatment used could not be applied in the studies that were contemplated.
In each study, it was analyzed if they met each of the nine criteria or not, adding one point for each criterion met and zero points for each criterion not met or not specified, with 9 points being the highest possible score and 0 points the minimum.

Meta-Analysis
For the meta-analysis, standardized mean differences (SMD) were calculated, with the respective 95% confidence interval (CI95%). The value of statistical significance was established at p

Data Extraction
The studies were thoroughly analyzed and the relevant data were collected in a table that contained the following categories: Population/substance of consumption, sample (experimental + control), type of exercise program, adherence to the program, measured variables and main results obtained.

Quality Analysis of the Studies
To define the quality of the studies, the widely used and validated "Physiotherapy Evidence Database (PEDro)" scale was taken into account [42][43][44]. This scale is based on the Delphi list [45], but it adds two criteria (items 8 and 10) that are not included in the Delphi list and are important when analyzing the quality of the studies. Of the 11 items on the PEDro scale, 5 and 6 were discarded for the analysis of the studies, as did other similar previous studies, among them, Wang and colleagues [39], since the blinding of patients and the instructors regarding the type of treatment used could not be applied in the studies that were contemplated.
In each study, it was analyzed if they met each of the nine criteria or not, adding one point for each criterion met and zero points for each criterion not met or not specified, with 9 points being the highest possible score and 0 points the minimum.

Meta-Analysis
For the meta-analysis, standardized mean differences (SMD) were calculated, with the respective 95% confidence interval (CI95%). The value of statistical significance was established at p < 0.05. On the other hand, the following criteria of statistical heterogeneity (I2) were established: 0-40% mild, 40-70% moderate and 70-100% high. A random effects model was selected when the heterogeneity values were adequate (I2 > 50%). Difference from post-intervention control group to post-intervention experimental group was calculated. In addition, the difference in the post-intervention control to post-intervention experimental change scores between participants in programs based on improving physical conditions and those based on oriental practices. A subgroup analysis was performed, according to the values of quality of life, mental health, and according to the physical activity modality investigated, for which effect estimates were obtained with the analysis model described above. The software proposed by Cochrane, Review Manager 5.3 (RevMan) was used to calculate the effect estimates.
To test for possible publication bias, we used the trim-and-fill funnel plot derived method, which identifies possible publication bias by estimating the number of unpublished studies and adjusts results for this bias. Meta-Essentials tool was used to build funnel plots and to statistically test for funnel plot asymmetry with the Trim and Fill test.
Aerobic exercise was the most used type of program, finding 27 studies that used this type of exercise [4,5,7,8,[10][11][12][13]16,18,22,26,46,49,50,56,57,[59][60][61]67,[74][75][76][77][78][79]. In addition, 14 studies were found that used exercise programs related to body-mind practices, such as yoga, tai chi or qigong [9,15,21,[28][29][30]54,55,[64][65][66]70,71,73], 10 who used combined strength and aerobic exercise [6,17,20,27,32,47,48,52,68,72], five that included varied physical activities [14,25,53,58,69], one that used a strength program [80], one that used softball [62] and finally, a study that compared the effects of three different exercise programs (aerobic, strength, and circuit training) [63]. Table 1 shows the analysis of the quality of the included studies, finding scores ranging from 1 point to 8 points. The average score of the studies was 4.97 (±1.85) out of 9. Several of the studies did not provide enough information to understand if some of the criteria were met or not. In these cases, the criterion was given as not fulfilled. Based on the Trim and Fill test, the lack of imputed data points in the funnel plots indicates the absence of asymmetry in the distribution of effect sizes when considering the effects of all programs together, the physical fitness programs and oriental practices on mental disorders and craving. Asymmetries were only found when considering the effect of all programs on quality of life, especially on Role Emotional (two missing studies) and Mental Health (one missing study) (see Figure 2; Figure 3, respectively).

Methodological Quality of the Selected Studies
Based on the Trim and Fill test, the lack of imputed data points in the funnel plots indicates the absence of asymmetry in the distribution of effect sizes when considering the effects of all programs together, the physical fitness programs and oriental practices on mental disorders and craving. Asymmetries were only found when considering the effect of all programs on quality of life, especially on Role Emotional (two missing studies) and Mental Health (one missing study) (see Figure 2; Figure 3, respectively).

Acute Effect of the Exercise
Seven studies were found that analyzed the acute effect of exercise, five of them in AUD patients, in which inconsistent results were obtained on the effect of exercise in this population [5,[77][78][79][80] and two in methamphetamine dependents, where an effect of acute exercise on craving is shown [10,11]. In all of them, an aerobic exercise session was performed, except for one [80], which performed a strength session. The results are summarized in Table 2. Based on the Trim and Fill test, the lack of imputed data points in the funnel plots indicates the absence of asymmetry in the distribution of effect sizes when considering the effects of all programs together, the physical fitness programs and oriental practices on mental disorders and craving. Asymmetries were only found when considering the effect of all programs on quality of life, especially on Role Emotional (two missing studies) and Mental Health (one missing study) (see Figure 2; Figure 3, respectively).

Acute Effect of the Exercise
Seven studies were found that analyzed the acute effect of exercise, five of them in AUD patients, in which inconsistent results were obtained on the effect of exercise in this population [5,[77][78][79][80] and two in methamphetamine dependents, where an effect of acute exercise on craving is shown [10,11]. In all of them, an aerobic exercise session was performed, except for one [80], which performed a strength session. The results are summarized in Table 2.

Effect Size
Studies Combined effect size Imputed data points CES Adjusted

Acute Effect of the Exercise
Seven studies were found that analyzed the acute effect of exercise, five of them in AUD patients, in which inconsistent results were obtained on the effect of exercise in this population [5,[77][78][79][80] and two in methamphetamine dependents, where an effect of acute exercise on craving is shown [10,11]. In all of them, an aerobic exercise session was performed, except for one [80], which performed a strength session. The results are summarized in Table 2.

Benefits for the quality of life
This review includes 10 studies that specifically analyze the influence of exercise on quality of life: five of them use yoga as an experimental treatment [15,21,55,70,71] a study uses tai chi [73], two studies apply a program of mixed exercise and aerobic exercise [6,72] and another study employs a varied exercise program [25].

Benefits for mental disorders
A total of 23 studies are included that analyze the influence of exercise on stress, anxiety and/or depression. Of these 23 studies, six studies assess the effect of exercise on stress. Two studies showed improvements in the levels of this variable after participation in a yoga program [15,55], while there are four studies where, despite showing better levels after the end of the program, these improvements were not significant in comparison with the control group, both in yoga programs [29,66], and in varied exercise programs [25,69].
In anxiety, six studies were found that showed statistically significant benefits after participation in programs of aerobic exercise [16,51], strength [63], mix of aerobic exercise and strength [17], yoga [21] or qigong [65], while in seven other studies, anxiety levels were improved, but this improvement was not statistically significant or was only intra-group, both in yoga programs [28,29] and in aerobic exercise programs [4,18,57] or mixed [20,51].
Finally, we found eight studies that show a statistically significant improvement in depression after participation in aerobic exercise programs [16,51], strength [63], mixed [17], yoga [21,54,70] or qigong [9]. On the other hand, another eight studies were found that, in spite of improving the levels, did not show statistically significant improvements compared to the control group after participation in aerobic exercise programs [4,18,57], mixed [21], yoga [28,29], tai chi [30] or combined programs of yoga with aerobic work [53].
The following tables show the results and characteristics of the studies conducted with patients addicted to alcohol (Table 3) or other drugs (Table 4).    There was no effect of the exercise program on alcohol consumption A greater protective effect of excessive alcohol consumption was observed in patients with a higher level of physical activity. There were no significant changes between the control and experimental groups in any of the 4 measured subscales.     The exercise group increased its daily exercise levels significantly, but no significant improvement was observed in any of the other variables measured.
Trivedi et al. [74] Stimulants 302 (152+150)  The experimental group improved the speed of information processing and fat oxidation.

Quality of life
The results of the analysis showed a significant effect of physical exercise (k = 6) in the eight variables included in the SF36 test (Figure 4)

Quality of life
The results of the analysis showed a significant effect of physical exercise (k = 6) in the eight variables included in the SF36 test (Figure 4)

Mental disorders and craving
The

Mental disorders and craving
The results show a significant effect of exercise (including all exercise programs here) on depression

Mental disorders and craving
The

Quality of life
When analyzing the subcategories of the SF-36, a greater effect of OP (k = 3) is observed in the variables physical function and pain, a greater effect of PF (k = 3) in the variables "physical role", "vitality" and "social function" and an similar effect in the variables "general health", "emotional role" and "mental health" ( Figure 6; Figure 7).

Quality of life
When analyzing the subcategories of the SF-36, a greater effect of OP (k = 3) is observed in the variables physical function and pain, a greater effect of PF (k = 3) in the variables "physical role", "vitality" and "social function" and an similar effect in the variables "general health", "emotional role" and "mental health" (Figure 6; Figure 7).

Quality of life
When analyzing the subcategories of the SF-36, a greater effect of OP (k = 3) is observed in the variables physical function and pain, a greater effect of PF (k = 3) in the variables "physical role", "vitality" and "social function" and an similar effect in the variables "general health", "emotional role" and "mental health" ( Figure 6; Figure 7).

Mental disorders and craving
When analyzing the subcategories of mental disorders, a greater effect of OP is observed in the variables "anxiety" and "depression" and a greater effect of PF in the variable "stress".

Mental disorders and craving
When analyzing the subcategories of mental disorders, a greater effect of OP is observed in the variables "anxiety" and "depression" and a greater effect of PF in the variable "stress".

Discussion
The present study shows important results about the benefits of physical exercise in numerous variables related to health and quality of life in drug-dependent persons. The results indicated that physical exercise can be a good tool to improve physical conditions, mental health, quality of life and craving in drug-dependent persons and that different physical exercise practices produce similar results.

Physical fitness
First, the results show that the practice of aerobic exercise or combinations of aerobic exercise and strength allows patients to improve their physical conditions. Other review studies conducted with AUD patients came to the same conclusion [34,81].
Drug-dependent patients usually show deterioration in their physical health, either due to direct drug consumption, loss of healthy habits or both circumstances [26]. This makes drug-dependent patients more likely to develop diseases, such as metabolic disorders [82] or cardiac pathologies [83]. These types of pathologies are highly related to the level of physical condition [34]. Thus, an improvement in the physical condition of drug-dependent patients may potentially result in an improvement in their health with a decrease in the risk of developing comorbidities associated with drug use, highlighting the importance of improving physical conditions in this population.

Discussion
The present study shows important results about the benefits of physical exercise in numerous variables related to health and quality of life in drug-dependent persons. The results indicated that physical exercise can be a good tool to improve physical conditions, mental health, quality of life and craving in drug-dependent persons and that different physical exercise practices produce similar results.

Physical fitness
First, the results show that the practice of aerobic exercise or combinations of aerobic exercise and strength allows patients to improve their physical conditions. Other review studies conducted with AUD patients came to the same conclusion [34,81].
Drug-dependent patients usually show deterioration in their physical health, either due to direct drug consumption, loss of healthy habits or both circumstances [26]. This makes drug-dependent patients more likely to develop diseases, such as metabolic disorders [82] or cardiac pathologies [83]. These types of pathologies are highly related to the level of physical condition [34]. Thus, an improvement in the physical condition of drug-dependent patients may potentially result in an improvement in their health with a decrease in the risk of developing comorbidities associated with drug use, highlighting the importance of improving physical conditions in this population.
However, the "improvement of physical condition" is something very generic, and to this day it is not clear which aspects of physical condition are more effective. Practically, all of the studies included in this review employ aerobic exercise programs, expressing the improvements in physical conditions expressed in terms of VO2, heart rate or performance in indirect tests that measure aerobic capacity. Some studies employ combinations of aerobic and strength exercise, while studies employing only strength or flexibility training are practically nil. In this sense, it would be convenient to carry out studies that also investigate the benefits that exercise programs may have for other physical abilities, such as strength and/or flexibility.

Quality of life
One of the novelties of the present study is the analysis performed in regards the effect of exercise on the quality of life of drug-dependent patients. Only one previous systematic review included three studies [32] analyzing the quality of life in patients addicted to methamphetamine. To the best of our knowledge, no meta-analysis has been published that analyzes the quality of life in a drug-dependent population.
The results of our review, like those obtained by Morris et al. [32] in patients who met criteria for methamphetamine dependence, indicated an effect of exercise on quality of life, finding that eight articles out of 10 show significant improvements after participating in yoga programs [22,55,70,71], tai chi [73] or in aerobic and strength exercise programs [6,20,72]. In the same way, the results of the meta-analysis show a significant effect of the exercise on quality of life. This result is particularly important, both due to the lack of meta-analysis that address the study of the quality of life in the drug-dependent population, and the importance of improving the quality of life in these types of patients, since drug-dependent patients suffer from important deterioration of the quality of life, presenting significantly lower values than the healthy population [84].
In this sense, the data presented in this meta-analysis provides evidence for the use of physical exercise, especially programs that use combinations of aerobic exercise and strength and oriental practices, such as yoga or tai chi, for improving quality of life during the recovery process from drug dependency.

Mental disorders
The results obtained in mental disorders seem to show an effect of exercise on stress, anxiety and depression, although the results should be viewed in perspective. When analyzing the studies included in this review, it is observed that there are both studies that show significant improvements in stress, anxiety or depression, as studies that did not find significant improvements in these variables, or that these improvements were only intra-group, practically identifying the same number of studies in both senses. This fact could prevent us from drawing conclusions about the effectiveness of exercise in improving these variables. However, when analyzing in depth the studies that do not show significant improvements in stress, anxiety or depression, some methodological or structural aspects can be identified that could explain the non-achievement of statistically significant results, such as the short duration of the programs [53] and sparse samples [18,29,57,66]. Likewise, when observing the results of the meta-analysis, a significant effect of the exercise on mental disorders is identified, obtaining an average effect in anxiety and depression and a high effect in stress, for which, analyzing together the obtained data, it seems that positive effects of exercise on stress, anxiety and depression can be deduced.
The results obtained in anxiety and depression are consistent with those obtained in other reviews, where the beneficial role of exercise in these variables has been shown in patients who met criteria for methamphetamine dependence [32], AUD [34,81] or SUD [39]. It should be noted that the meta-analysis conducted by Hallgren et al. [34] found a significant effect of exercise on depression, but not on anxiety in AUD patients. Additionally, the review conducted by Giesen et al. [81] found a slight tendency to improve anxiety, but found no effect in depression in the same population. This contradiction in the benefits of exercise on anxiety and depression in AUD patients may be due to methodological irregularities and the disparity of measures used in some of the studies included in Giesen et al. or to the fact that only three studies were included in the meta-analysis conducted by Hallgren et al. Therefore, more randomized studies that assess the effect of exercise on anxiety and depression in this specific population and that allow meta-analysis with higher samples are required.
Assessing the results obtained in mental disorders in this meta-analysis highlights the effect of exercise on stress, finding a high effect of exercise on this variable. There are very few studies that have shown the effects of exercise on stress and no previous meta-analysis analyzing this variable in drug-dependent patients. In the present review, six experimental or quasi-experimental studies were identified that analyzed the effect of exercise on stress. Of these six, only two [15,55] showed statistically significant improvements compared to the control group after participating in two yoga programs. However, when observing the meta-analysis, more conclusive results are shown, identifying a great effect of the exercise on this variable. Once again, there is a need to carry out randomized studies with sufficiently representative samples and with structured exercise programs that rigorously allow the assessment of the effect of exercise on stress.
The management of stress in drug-dependent patients is crucial. This population has very high levels of stress, anxiety and depression [85]. In addition, stress is a variable directly related to the likelihood of relapse, increasing the patient's vulnerability to relapse [86]. Therefore, improving stress through physical exercise could potentially exert a protective effect on patients and a boost to maintain abstinence.

Craving
The results of this study indicate a positive influence of exercise on craving, finding five studies that show significant improvements in craving after participation in exercise programs and only one [8] where improvements were not obtained in craving after participation in a four-week aerobic exercise program. In this study, perhaps the short duration of the program could have influenced the lack of improvements in craving. However, studies analyzing the acute effects of exercise on craving showed an improvement of craving after performing aerobic exercise, both in AUD patients [77] and in methamphetamine dependence patients [10,11]. Likewise, the results of the meta-analysis reinforce the conclusion that exercise can be a good way to reduce craving levels, finding a high effect of exercise on this variable. Previous review studies with AUD patients showed similar results [35,81], but to date, no known meta-analysis addressed this issue. Nevertheless, future studies should investigate the appropriate type and duration of exercise programs in order to maximize benefits in this regard.
The impulse and need to consume despite the serious consequences it entails, is one of the effects of drug-addiction [85]. These impulses are those that lead the drug-dependent patient to consume again and again before any other need. In periods of abstinence, the level of craving is directly related to the probability of having a relapse [87]. Therefore, to identify that physical exercise may reduce the levels of craving in this population is a very relevant result and gives physical exercise an important role in the rehabilitation of drug addictions, proving itself as an ideal means to increase the chances of success of the treatments.

Abstinence/consumption
The results of this review seem to indicate a certain benefit of the exercise on abstinence and drug use, finding numerous studies that show satisfactory results in abstinence or consumption after participating in exercise programs of different kinds. However, there are also several studies that did not show improvements in these variables in a significant way or that the improvements were only intra-group [8,16,29,52,59,60,69]. When analyzing the methodological quality and the possible biases in the studies, no substantial difference was found between the studies that showed statistically significant benefits on abstinence and those that did not. Therefore, no firm conclusions about the effect of exercise in the maintenance of abstinence can be drawn based on the systematic review. In similar lines, there were two reviews [35,81] and a meta-analysis performed in AUD patients [34], where they did not find a significant effect of exercise on alcohol consumption.
On the other hand, a meta-analysis conducted with SUD patients did draw firm conclusions in this field [39], stating that physical exercise can have an important effect for the maintenance of abstinence in tobacco, alcohol, and illicit drugs. The reason for this disparity in the results could be due to small sample sizes or unstructured exercise programs with a short duration or frequency.
Another reason that could influence the achievement or failure of significant results could be the adherence to the exercise programs. For example, in a randomized controlled study with a large sample [74], it was found that after the end of the exercise program there were no significant differences in abstinence compared to the control group. However, when performing the analysis controlling adherence to the program, they did find statistically significant improvements, concluding that mere participation in the program was not enough to achieve improvements in abstinence and that adherence to the exercise program was a determining factor in the achievement of the desired benefits. Therefore, future studies should investigate possible methodological strategies to increase adherence to exercise programs in this population, with the aim of increasing the effectiveness of the programs.

Aerobic-strength exercise vs. Body-mind activities
When analyzing the subcategories of mental disorders and quality of life, some differences can be observed. In quality of life, a greater effect of oriental practices was observed in the subcategories "physical function" and "pain", while a greater effect of aerobic-strength exercise was shown in the variables "physical role", "vitality" and "social function". A similar effect on the variables "general health", "emotional role" and "mental health" was observed.
In regards to mental disorders, oriental practices showed a greater effect on "anxiety" and "depression", while aerobic-strength exercise shoed a greater effect on the variable "stress". These results are not in agreement with those obtained by Wang et al. [39], where no differences were identified between both exercise practices in anxiety and depression. Notwithstanding, the results of the present study in this regard should be interpreted with caution, since when performing the analysis classifying by type of program, the number of included studies was lower. Furthermore, the number of studies included for each type of program was different, which may condition the results obtained.
Taking everything together, it can be deduced that both exercise practices are effective for the improvement of mental disorders and quality of life. Future meta-analysis studies that include more randomized studies of both practices and that control the bias that may exist at the time of performing the analysis, as well as randomized studies that compare the effects of each type of practice, should determine the suitability of applying one type of practice or another depending on the objectives of the treatment.

Limitations and future studies
This study presents a series of limitations, mainly due to the type of studies included. First, both randomized and nonrandomized studies were included. In addition, many of them presented methodological limitations, such as the non-performance of inter-group analysis, excessive number of dropouts, lack of control group or unrepresentative samples. These aspects may bias the results. However, it would also under represent the entire field of study.
In the same way, the high heterogeneity of the included studies can represent a bias when synthesizing the results. The selected studies include populations who met criteria of dependency for different substances, and different exercise programs with different structures in terms of duration, volume, intensity or frequency. Likewise, there are very different variables and measurement instruments and in some cases lack of relevant data. For example, many of the studies included did not report the severity of addiction prior to treatment.
All of this caused the loss of some studies at the time of performing the meta-analysis and may imply a bias in the research that must be taken into account.
On the other hand, of the 59 studies included in this research, 35 did not control adherence to the intervention program. However, adherence to exercise programs may determine the achievement of the intended objectives [74]. Drug-dependent patients have an average of dropouts in exercise programs much higher than dropouts registered in other special populations [34]. Not taking this issue into account can result in contradictory or misleading results, which can create confusion when drawing conclusions. In this way, control of adherence to exercise programs is especially important when performing research in this population.
Another methodological issue is the lack of information about the period of time that the benefits are maintained after the exercise intervention. Very few studies have included a follow up. In this sense, Sinyor et al. [50] showed that the benefits found in the abstinence of a sample of patients with AUD after an exercise program were maintained for 18 months after the end of the program. However, there is a clear need for more studies including follow-ups in order to draw firm conclusions about this issue.
There is currently no scientific evidence in regards to which specific physical capacities (endurance, strength, flexibility, balance, coordination) are more important from the point of view of the health and recovery of the drug-dependent patient. There is also no clear evidence about what volumes, intensities and frequencies are ideal for working with this population. Wang et al. [39] found that there was no difference between the application of exercise programs in terms of type of exercise and intensity, showing that mild and moderate intensities of the exercise achieved similar effects and that different types of exercise programs achieved similar effects. On the contrary, Wang et al. [11] concluded that a moderate or vigorous intensity of exercise supposed greater acute benefits in craving compared to a mild intensity after a single session of aerobic exercise. Therefore, there is still no clear evidence on the parameters of the exercise in terms of type of exercise, intensity, volume or frequency in this population. Thus, there is a need for randomized controlled studies with larger samples including structured exercise programs (detailing type, intensity, duration and frequency) and comparing different types, while controlling for the adherence to the program and including follow-ups, which will allow us to obtain reliable and comparable results, with the final aim of finding the optimal exercise prescription for this population.

Conclusions
This study shows relevant results about the benefits of exercise for health in drugdependent patients.
First, it can be observed that aerobic physical exercise and combinations of aerobic exercise and strength are effective for the improvement of physical conditions in drug-dependent patients. Aerobic exercise programs are the most widely used in this population, although some studies combine aerobic and strength exercise, with satisfactory results.
One of the relevant results of this study is the significant effect of exercise on the quality of life in drug-dependent patients, with no previous meta-analysis addressing this issue. Similarly, it highlights the significant effect of exercise on stress, anxiety, depression and craving.
Despite the promising effects of physical exercise on abstinence, based on the results obtained in this study it cannot be concluded that there is solid evidence on this issue.
Finally, when comparing the effects of aerobic-strength exercise with oriental practices, no relevant differences were found about the overall value of quality of life and mental health. On the other hand, differences were observed in some of the subcategories of quality of life and mental health, although future studies must delve deeper into these issues to draw firm conclusions in this regard.
Thus, it can be concluded that aerobic physical exercise or combinations of aerobic exercise and strength are viable and effective practices for the improvement of physical conditions in this population. In addition, physical exercise, whether through pure aerobic exercise, combinations of aerobic exercise and strength or oriental practices such as yoga, tai chi or qigong, can be an effective way to improve stress, anxiety, depression, quality of life and craving in drug-dependent patients.
Therefore, it is recommended that physical exercise programs in rehabilitation centers are included in order to optimize the patient's recovery process.