Effectiveness of Resistance Training on Fatigue in Patients Undergoing Cancer Treatment: A Meta-Analysis of Randomized Clinical Trials

Objective To assess the effectiveness of the resistance training to improve fatigue levels in people with cancer who are enrolled in adjuvant and/or neoadjuvant treatment. Methods MEDLINE, Web of Science, Embase, SPORTDiscus, LILACS, CENTRAL, and CINAHL databases were searched from May to December 7, 2021. Randomized clinical trials (RCT) that evaluate the effects of resistance training on fatigue levels in people undergoing cancer treatment were included. The PEDro scale was considered to assess methodological quality of studies, and the evidence was summarized through the GRADE system. The standardized average differences, effect size, and inverse variance model for meta-analysis were calculated. Results Fifteen RCT for qualitative synthesis and thirteen for meta-analysis were selected. A moderate to high level of evidence of resistance training was identified to improve fatigue in people undergoing cancer treatment. Meta-analysis showed a significant reduction in fatigue (SMD = −0.31, CI 95% = −0.58, -0.12, P = 0.001) after 10 to 35 sessions of resistance training. Conclusion The 10 to 35 sessions of resistance training are effective in reducing fatigue level in cancer patients who are undergoing cancer treatment and have a moderate level of quality evidence.


Introduction
According to the World Health Organization (WHO), cancer incidence and mortality increase each year, making it the second leading cause of death in the Americas after cardiovascular diseases. Cancer treatment can be neoadjuvant (presurgery) or adjuvant (postsurgery), which can consist of radiation therapy, chemotherapy, hormone therapy, androgen deprivation therapy, and hematopoietic call transplantation [1][2][3]. Cancer treatment is aimed at restoring the function of the organ affected by the malignant tumor [4] or controlling the progression of neoplasia through DNA damage and mitotic spindle inhibition of cancer cells [5,6]. The adjuvant and/or neoadjuvant treatment increases cancer patients' survival; however, it can generate negative side effects such as muscle atrophy, loss of functionality, reduced mobility, and fatigue related to cancer [7].
Cancer patients can present a high degree of fatigue perception that can last 5 to 6 weeks after cancer treatment and drastically affect quality of life [8]. The mechanisms involved in fatigue have been associated with the release of proinflammatory cytokines by the malignant neoplasia and cancer treatment that alters the hypothalamic-pituitary-adrenal axis [9][10][11] and generates a dysfunction of the autonomic nervous system [12]. Fatigue is defined as a "sensation of tiredness or distressing, persistent, subjective exhaustion related to cancer or cancer treatment that is not proportional to recent activity and interferes with normal functioning" [13].
Resistance training in cancer patients improves quality of life and increases adherence and success of cancer treatment, in addition to contributing to the increase in lean mass and body weight, preventing muscle mass loss, reducing its side effects, and increasing survival [14][15][16]. There are two systematic reviews that reported the effects of exercise and physical activity on fatigue levels in cancer survivors and patients with metastasis [17,18]; however, these studies did not include specific information regarding the dosage of resistance training and its effect during cancer treatment. The objective of this study is to evaluate the effectiveness of the resistance training (RT) to improve fatigue levels in people with cancer who are undergoing adjuvant and/or neoadjuvant treatment.

Material and Methods
2.1. Protocol and Registration. The present systematic review of randomized clinical trials and meta-analyses was performed according to the report of elements for systematic reviews and meta-analyses (PRISMA) [19] and Cochrane recommendations [20]. The reference number of the International Registry of Prospective Systematic Review (PROS-PERO) is CRD 42020196180.

Search Strategy and Article Selection.
Studies were selected if they met the following inclusion criteria based on publication types based on the participants, interventions, comparisons, outcomes, and study design (PICOS): (1) population: subjects over 18 years old with cancer who are undergoing cancer treatment; (2) intervention type: RT; (3) comparisons: other training type controls without intervention; (4) results: fatigue; and (5) studies: randomized clinical trials (RCT), published in English or Spanish. The articles were excluded if the sample consisted of people who were not undergoing cancer treatment.
An electronic search was performed until December 7, 2021, in the following electronic databases: MEDLINE (via PubMed), Embase, Web of Science, CINAHL, SPORTDiscus, LILACS, PEDro, Cochrane, and EBSCO. The following Medical Subject Headings (MeSH) terms were included and combined for sensitive searches: "Cancer"; "radiotherapy"; "chemotherapy"; "fatigue" and "resistance exercise". The search strategy used in the MEDLINE was as follows: ((((fatigue) AND (cancer)) AND (radiotherapy)) OR (chemotherapy)) AND (resistance exercise). The strategy was adapted according to each database and is specified in a supplementary file (available here).
The articles were initially selected for screening of the title and the abstract, ending with a full text evaluation. Two evaluators (LB, RC) evaluated each scientific article independently. In case of disagreement or discrepancy in the inclusion or exclusion of any article, it was sent to an independent reviewer (IL), who decided in this respect to agree on its selection for the complete analysis of the article. The same process was performed to evaluate both methodological quality and level of evidence of included articles.

Data Extraction and Analyses.
For data extraction, standardized forms adapted from the Cochrane Collaboration model [20] were used. The following aspects were considered to describe the characteristics of the studies: population (sex, age, number of participants per group, cancer type, and cancer treatment), intervention (type, duration, and frequency), and dosage of resistance exercise (type of external load, maximum repetition percentage (%1RM), increase load, number of sets, and repetitions). To describe the results of the selected studies, the following were considered: questionnaire type used to assess fatigue, effect size, and fatigue levels preand post-RT (variation percentage). The results of the perception of fatigue were evaluated according to total training time. Additionally, the effect size (ES) was calculated to determine the clinical relevance of the interventions with a 95% confidence interval, considering pre-and postintervention values. A small effect of the intervention was considered <20; a moderate effect, between 21 and 0.79; and a great effect, >0.80 [21]. The negative effect size represents a "worsening of symptoms" while the positive one represents an improvement [22]. All primary studies that presented descriptive values as mean and standard deviation were considered to ES calculation.

Quality Assessment.
To evaluate the methodology quality and the risk of bias of the included RCT, we used the Physiotherapy Evidence Database (PEDro) scale [23]. The PEDro scale allows the bias rating of articles according to a cut-off score. The studies with a 10-9 score correspond to an excellent quality methodology, those with an 8-6 score correspond to a good methodology quality, those with a 4-5 score correspond to an acceptable methodology quality, and those that had a score lower than 4 correspond to a bad methodology quality [24,25].
The Grades of Recommendation, Assessment, Development, and Evaluation (GRADE) scale was used to evaluate the level of evidence categorizing it into 4 levels: high, moderate, low, and very low [26]. GRADE considers 5 criteria: limitations in study design, indirectness of evidence, heterogeneity, imprecision of results, and publication bias. Each criterion is evaluated as follows: not serious, serious, and very serious considering a consensus between two raters [21]. The high characterization reflects that there is a high confidence in the coincidence between the real and the estimated effect; the moderate quality level corresponds to a moderate confidence in the estimate effect; that is, there is a possibility that the real effect is far from the estimated effect. On the contrary, a low and very low quality of evidence reflects that there is uncertainty regarding the intervention effects [26,27].

Statistical Methods.
A descriptive synthesis was performed considering the mean, standard deviation, distribution by frequency, and percentage of the results of the primary studies. In addition, the meta-analysis was calculated to determine the effects of resistance training on fatigue levels. The analysis was performed using means and standard deviations from each selected clinical trial. Results of some questionnaires indicated an increase in fatigue levels while others indicated a decrease, according to the direction of questions. The mean values of a group of studies were multiplied by -1 to ensure that all instruments point in the same direction [20]. The difference in standardized means and the 95% confidence interval (CI) were calculated using an inverse variance model of random effects for the meta-analysis, considering postintervention data.

2
International Journal of Breast Cancer Heterogeneity data between studies was evaluated using the I 2 statistic, and P value was set at 0.01. The statistical analysis was performed with five revisions of RevMan management software (version 5.3).

Study Selection.
The electronic search identified a total of 523 studies across 8 scientific databases, and 15 randomized clinical trials were included. Details on the selection of studies are in Figure 1.

Study
Characteristics. The characteristics of the fifteen studies selected are shown in Table 1. The sample studied in general included 1439 adults with an age range from 18 years to 78 years old. Regarding cancer type, 6 articles were found on subjects with stage 0-III breast cancer [28][29][30][31][32][33], two on subjects with head-neck cancer [34,35], four on subjects of prostate cancer [36][37][38][39], two articles on subjects with hematological neoplasms (leukemia, lymphoma, multiple myeloma, and myelodysplastic syndrome) [40,41], and one article on subjects of germ cell cancer [42]. According to the type of cancer treatment, in five articles, subjects were   International Journal of Breast Cancer  International Journal of Breast Cancer   Hacker et al., [40] EG: traditional muscle strength training. Seated leg press, seated row machine, trunk flexion, knee flexion machine, bench press, trunk extension machine, push press, standing planted flexion, and frontal pulldown    Segal et al., [37] EG: progressive resistance exercises. Leg press, chest press, leg extension, leg curl, shoulder press, seated side pull row, calf raise,  [28][29][30][31]41]; in seven clinical trials, the sample received radiotherapy [28,32,33,37,38]; in two articles, the subjects were undergoing hematopoietic stem cell transplantation [40,41]; and in two studies, the subjects underwent androgen deprivation therapy [36,39].
The control group in 12 studies did not receive any type of intervention, and the subjects continued their daily activities and medical care as usual [28-30, 33-38, 40-42]. In two articles, the control group subjects received an intervention corresponding to progressive muscular relaxation according to Jacobson [31]. Only in one study did the control group perform aerobic training [39] (Table 1).

Quality Assessment.
Regarding the methodological quality of the primary studies, 86.6% presented a good methodological quality and 13.3% acceptable methodological quality. According to critical criteria, the sample was randomly assigned in all the selected studies, the allocation was concealed in 66.6%, and in 33.3% of the articles, it was not concealed; the groups were similar at the baseline regarding the most important prognostic indicators in all studies. Regarding the criteria for blinding the therapist and subjects, these were not met in all the selected articles; likewise and in relation to the blinded evaluators, only two study met this criterion [36,42]. Details of the methodological quality of the studies are presented in a supplementary file.
The effect of the intervention on fatigue levels is described in detail in Table 2. Three articles reported lower fatigue levels in the group that underwent progressive resistance training compared to a control group [36,37,42] and compared to the group that underwent muscle relaxation [33]. In addition, only tree studies that used a resistance training with resistance bands showed to be effective to reduce the fatigue compared to the control group [28,38,41]. On the other hand, nine clinical trials that used RT as an intervention did not report significant changes in fatigue levels of the intervened subjects [29-32, 34, 35, 39, 40].
3.6. Publication Bias. Publication bias was estimated for eight studies that performed an intervention of 10 to 35 sessions. According to the funnel plot, there is no evident asymmetry, therefore indicating an absence of publication bias in this meta-analysis ( Figure 4). On the other hand, five studies that carried out an intervention of more than 35 sessions have a publication bias, since all the studies are the outside area in the funnel plot ( Figure 5).

Sensitivity Analysis.
A sensitivity analysis was performed to test the effect of studies with fair methodological quality and studies with smaller sample sizes. As for the influence of studies with regular methodological quality [41], the results did not change in favor of resistance training compared to the control groups and the SMD was reduced from -0.31 to -0.28, without significant heterogeneity (I 2 = 0%; P = 0:83). Likewise, regarding studies with a small sample size [34,35,41], these do not influence the results reported in the meta-analysis; the SMD changes from -0.31 to -0.32 without significant heterogeneity (I 2 = 0%; P = 0:95).

Discussion
The objective of the present systematic review and metaanalysis was to evaluate the effectiveness of resistance training to better the fatigue levels in cancer patients that are undergoing adjuvant and/or neoadjuvant treatment.
In relation to the quantitative analysis, the results of the meta-analysis show that resistance training with 10 to 35 sessions using elastic bands and/or an external load between 60% and 80% of the 1RM generates significant changes to reduce fatigue levels compared to a control group. On the other hand, the group intervened with resistance training with a number equal to or greater than 35 sessions did not generate significant changes in fatigue levels compared to the control group.
The findings reported in this systematic review are like those obtained in two published meta-analyses [17,18]. The study by Hilfiker et al. [17] reported that any type of exercise significantly reduces fatigue levels in subjects who are receiving adjuvant/neoadjuvant therapy or after cancer treatment. Likewise, the study by Kessels et al. [18] reported the positives effects of aerobic exercises and combined exercises in reducing fatigue. Even though both meta-analyses [17,18] support exercises to reduce fatigue, they lack an analysis regarding the specificity of resistance exercise, dosage, and the level of scientific evidence. However, despite the lack of specificity in terms of dosage, exercise and physical activity have been shown to significantly reduce fatigue related to any type of cancer [43].   The etiology of fatigue related to cancer and cancer treatment is still not completely clear; however, there are multiple causal mechanisms that have been postulated, such as emotional and cognitive factors, genetic variants, proinflammatory mechanisms, immune response, and molecular mechanisms involved in cachexia that generates neuromuscular complications and that could further exacerbate fatigue [44]. Regarding proinflammatory mechanisms, it has been hypothesized that the beneficial effects of exercise are related to the release of IL-6 myokine from the muscle, which generates negative feedback on the production of proinflammatory cytokines IL1-β and TNF-α [45]. Only one study [32] included in this systematic review reported that resistance exercise reduces proinflammatory makers associated with a reduction in physical fatigue in subjects with breast cancer. However, no specific data were reported on fatigue levels. Likewise, a pervious study [46] reported an inverse correlation between improved strength/ muscle mass with reduced inflammation and decreased fatigue levels in survivors of breast cancer subjects trained with resistance exercise. Therefore, resistance exercise through its benefits in reducing proinflammatory markers could reduce fatigue levels in cancer patients. However, there is a lack of more ran-domized clinical trials that provide more specific and categorical estimates of mechanisms involved in reducing fatigue.
Resistance exercise can play an important role in cancer patients, since through the reduction in inflammation previously mentioned, the decrease in catabolism, and the increase in satellite cells in type II muscle fibers, it allows reducing muscle atrophy related to cachexia in cancer [47,48]. According to the above, adequate function and muscle mass generate significant improvements in fatigue symptoms, reduce the side effects of cancer treatment, reduce depression, improves the quality of life, and increases survival in patients undergoing resistance training [49,50].
The results of the systematic review and meta-analysis reported an update of the kinesiology clinical practice, with a synthesis at moderate to high level of evidence according to GRADE on the effectiveness of resistance training of 10 to 35 sessions to reduce fatigue levels in cancer with a moderate to strong magnitude of the effect. Likewise, the synthesis of the evidence according to GRADE is high on the effectiveness of resistance training greater than or equal to 36 sessions to reduce fatigue levels in cancer with minimal to moderate effect magnitude.  14 International Journal of Breast Cancer The clinical guidelines have established that resistance exercise is safe for cancer patients either during or after cancer treatment. It is recommended at a moderate intensity and a frequency of 2-3 times a week, involving large muscle groups; however, the lack of personalization of the exercise regimes is a limitation [51]. Future clinical trials should consider for targeted planning of each patient and a more specific evaluation that considers critical variables such as physical fitness, lifestyle habits such as a sedentary lifestyle and smoking, comorbidities present in the patient, cancer type, and treatment regimens as well as possible side effects of cancer treatment. In addition, they should consider the chronic effects of resistance exercise on fatigue levels, with long-term follow-up.
The limitations of this systematic review and metaanalysis are mainly related to the following: (1) different volumes of training and exercise protocols were used in the primary studies; (2) few studies greater than or equal to 36 sessions were incorporated into the meta-analysis; (3) moderate heterogeneity and publication bias of studies greater than or equal to 36 sessions may indicate variability in the estimation of the effect; (4) three included studies had a small sample size which could influence a possible overestimation; (5) two studies did not show descriptive data for the main variable; (6) different types of questionnaires were used for evaluating fatigue; (7) different cancer treatments were employed in the sample; and (8) the sample presents different types of cancer.

Conclusions
Based on the limitations of the current evidence mentioned above, we recommend for future research to conduct randomized clinical trials with larger sample size and duration of resistance training greater than 36 sessions in people with cancer during oncology treatment.
Despite the above limitations, it can conclude that resistance training of 10 to 35 sessions is effective in reducing fatigue levels in cancer subjects who are undergoing cancer treatment with a moderate quality level of evidence.

Data Availability
All data are presented in Results and supported by the supplemental file.