Supervised Exercise Interventions in Childhood Cancer Survivors: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

Background: Childhood cancer survivors (CCSs) may suffer from a multitude of health impairments, resulting in a compromised quality of life (QoL). This review’s objective was to examine CCSs’ adherence to supervised exercise training interventions and the impact of these interventions on health outcomes. Methods: The following databases were searched in May 2022: PubMed, Embase, Cochrane Library, and Web of Science. The included studies were limited to randomized controlled trials (RCTs), published in English involving CCSs aged 18 years and below. Results: Nine RCTs (n = 642 participants) were included in the systematic review, and seven of them (n = 551 participants) were included in the meta-analysis. Both the mean retention rate and adherence to the supervised exercise interventions were 87%. Supervised exercise interventions significantly improved muscle strength (standardized mean difference (SMD) = 1.42, p = 0.03), level of daily physical activity (SMD = 1.05, p < 0.001), body mass index (BMI) (mean difference (MD) = 1.06, p = 0.03), and fatigue (SMD = −0.44, p < 0.001), while there was no statistical significance in the quality of life (QoL) (SMD = 0.21, p = 0.20). Conclusions: The adherence of CCSs to supervised exercise interventions is high, and supervised exercise interventions are safe and effective.


Introduction
Cancer and anticancer therapies are associated with many adverse effects on childhood cancer survivors (CCSs). Research has shown that, during treatment, CCSs may experience fatigue and a decline in cardiopulmonary function, muscle strength, functional performance, and quality of life (QoL) [1][2][3]. Although, the five-year survival rate of children diagnosed with cancer has reached nearly 85% due to the significant advancements in cancer treatment [4]. CCSs face a high risk of severe and even fatal late health consequences of cancer or treatment [5]. It is estimated that approximately two-thirds of survivors may experience at least one complication, such as obesity, diabetes, osteoporosis, cardiovascular disease, and secondary malignant tumors [5][6][7][8]. These adverse effects can have a negative impact on CCSs' health outcomes.
As an effective nonpharmacological therapy, exercise plays an essential role in the treatment of CCSs, which can improve motor function ability and exercise tolerance [9,10]. However, up to 60~75% of CCSs' level of daily physical activity does not meet the World Health Organization's (WHO) recommended average of 60 min per day of at least moderateintensity physical activity [11][12][13]. Physical inactivity has been identified as an important reason for diminished physical function in CCSs [14]. Previous meta-analyses of CCSs found that exercise interventions can reduce the side effects of cancer treatment in CCSs and bring health benefits [14,15]. In particular, supervised exercise is one of the most effective exercise modalities representing modifiable health behavior. The implementation of a supervised exercise intervention can reduce cancer-related sequelae (e.g., fatigue) and improve

Data Extraction
Two authors (Q.S. and J.Z.) independently screened the titles and abstracts of the articles to determine eligibility. Then, full texts of potentially relevant studies were retrieved for further assessment of their eligibility. Data from the included studies were independently extracted and summarized by two authors (Q.S. and J.Z.) using a standardized data extraction form. If there was any disagreement, the third author was involved in the discussion until a consensus was reached. We extracted and summarized the following information for all of the included studies: Study design, basic information of the studies, characteristics of the participants, intervention characteristics (i.e., frequency, intensity, time, type, and setting), outcome measures (i.e., retention, adherence rates and safety, cardiorespiratory fitness, muscle strength, functional performance, flexibility, balance, level of daily physical activity, body mass index (BMI), total lean and fat mass, fatigue, QoL, and self-efficacy).

Quality Assessment
The quality and risk of bias of the studies were assessed according to the Cochrane Handbook for Systematic Reviews of Interventions [25]. Two authors (Q.S. and J.Z.) evaluated the following criteria: random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessor (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and other potential sources of bias. Each domain was classified into three levels: "Low risk", "high risk", or "unclear". Quality assessment was evaluated by two independent authors (Q.S. and J.Z.), and disagreements were resolved through discussion with the third author (K.L.).

Statistical Analysis
Review Manager 5.3 and STATA software were used to perform a meta-analysis of the included studies. The means and standard deviations (SDs) from baseline to postintervention were recorded. For continuous outcomes, if the measuring tool was the same, we adopted the mean difference (MD) with 95% confidence intervals (95% CIs); if the measuring instruments were inconsistent, we used the standard mean difference (SMD) with 95% CIs. Statistical heterogeneity was calculated with I 2 statistic, which was explained using the following cut-off parameters: non-important heterogeneity, 0% to 40%; moderate heterogeneity, 30% to 60%; substantial heterogeneity, 50% to 90%; and considerable heterogeneity, 75% to 100% [25]. If the heterogeneity was substantial (I 2 > 50%), a random-effects model was applied; otherwise, a fixed-effect model was used [27]. If there were several articles based on the same study, only one study was included in the meta-analysis. When measurements were conducted at different time points, the data closest to the end of the intervention were included. If the heterogeneity was substantial (p < 0.1, I 2 > 50%), sensitivity analysis was conducted by excluding studies one by one to explore the possible source of heterogeneity. Potential publication bias was evaluated by Egger's test. A two-sided p < 0.05 was considered to be statistically significantly different.

Study Selection
The initial search retrieved a total of 8786 records, which was reduced to 8034 studies after removing duplicates. After screening the titles and abstracts of the identified articles, 7907 articles were excluded as they did not meet the inclusion criteria. The full texts of the remaining 127 articles were reviewed; finally, nine articles were eligible for inclusion and were included in this review [28][29][30][31][32][33][34][35][36]. Four articles were published on the same two studies, with the remaining seven RCT studies for the quantitative analysis ( Figure 1) [29,31,32,[34][35][36].

3.
Functional performance: Only one study analyzed the effect of the exercise intervention on functional performance, using the 3 m Timed Up and Go (TUG) test and Timed Up and Down Stairs (TUDS) test [29]. This study did not report a beneficial training effect of the exercise intervention on functional performance [29].

4.
Flexibility and balance: One study assessed the effect of the exercise intervention on flexibility and balance [35]. Flexibility and balance were measured by the sit and reach test and flamingo balance test, respectively. The results showed that the program can effectively improve the flexibility and balance of CCSs.

5.
Level of daily physical activity: Six studies analyzed the effect of exercise interventions on the level of daily physical activity [29,30,[32][33][34]36]. These studies adopted the Chinese University of Hong Kong Physical Activity Rating for Children and Youth scales [32][33][34], the German Momo questionnaire [36], or acceleration for objective measurement [29,30]. Three studies found that the level of daily physical activity increased after exercise interventions [32][33][34]. Based on the meta-analysis, compared to the control group, supervised exercise can significantly increase the level of daily activity of CCSs in the experimental group (n = 4 studies, n = 374 participants, SMD = 1.05, 95% CI = 0.60~1.50, p < 0.001), with substantial heterogeneity between studies (I 2 = 66%, p = 0.03) (Figure 3b) [29,32,34,36]. The results of the sensitivity analysis demonstrated that the removal of any studies had no significant effect on the overall results, indicating that this meta-analysis is robust (Figure 4b).

Publication Bias
Overall, there was no evidence of publication bias in muscle strength (Egger's tes
Self-efficacy: Three studies adopted the Physical Activity Self-Efficacy scale (PA-SE) to evaluate the effect of the exercise interventions on self-efficacy, the results of which showed that exercise interventions had a significant effect on self-efficacy [32][33][34].

Discussion
To the best of our knowledge, this is the first meta-analysis to quantitatively summarize the effects of supervised exercise interventions on CCSs. This systematic review and metaanalysis provides new evidence for the effect of supervised exercise interventions on CCSs. The results demonstrated that supervised exercise interventions had high retention and adherence rates, which could significantly improve muscle strength, level of daily physical activity and BMI, and reduce the fatigue of CCSs during and after treatment. In addition, no major adverse events or health-related problems related to exercise training were found. Therefore, supervised exercise interventions performed during and after treatment are safe and effective.
In this meta-analysis, substantial or considerable heterogeneity was observed in the included studies. Firstly, there were some differences in the study population. Although all of the participants were CCSs, they were in different stages of treatment when the supervised exercise interventions were implemented. Two studies were conducted when the CCSs had completed cancer treatment [33,34]; one study was in treatment or within the first year after cancer treatment [28]; and six studies were undergoing treatment, but the specific stage was not clear [29][30][31][32]35,36]. Secondly, notable differences were present in the supervised exercise interventions. The interventions in nine studies differed widely with regard to the types of exercise interventions (aerobic [31], aerobic and resistance [28][29][30]33,34,36], or multitype exercise [32,35]), duration (two days [31] to 24 weeks [32][33][34][35]), and number of sessions (three [33,34] to 57 sessions [29]). Finally, most of the outcome measures were different. For example, six measurement methods were used to measure muscle strength in this meta-analysis. However, due to the limited number of included studies, subgroup analysis could not be performed to analyze the definite source of heterogeneity in some of the outcomes.
Through supervised exercise interventions, health professionals can make exercise plans according to the current physical condition of CCSs and can provide timely feedback. Such feedback includes suggestions on the type, intensity, frequency, and duration of exercise and encouraging CSSs to exercise in their daily life [34]. This feedback can improve motivation to perform the exercise, resulting in increased adherence [41]. Previous studies have shown that the adherence to supervised exercise interventions was higher than that of home-based exercise interventions [42,43]. Supervised training programs had greater adherence among adolescents and increased training-induced adaptations than those unsupervised ones [44]. In this systematic review, the retention rate of supervised exercise interventions (87%) was slightly higher than that of non-supervised ones (85%), and adherence to supervised exercise interventions (87%) was similar to that with no supervision (88%) [22]. Unlike previous evidence [45], it was found that supervised exercise interventions in survivors do not lead to higher adherence than unsupervised exercise interventions. This may be because most of the participants in this systematic review were in their treatment period, and they were absent due to poor physical condition [46], while all of the participants in Mizrahi's [22] systematic review had completed their intensive cancer treatment regimens. Nevertheless, this review demonstrated a positive impact of supervised exercise interventions on CCSs, and the impact still existed after the interventions [32][33][34]. Therefore, CCSs benefited more from supervised exercise interventions than unsupervised or home-based interventions.
Cancer and its treatment result in impaired physical function in children [47]. Concretely, the muscle strength of CCSs is impaired to different degrees during and after treatment [48]. Muscle strength is indispensable in many daily-life activities for individuals to dress, walk, stand, climb stairs, etc. [49]. Lanfranconi [50] found that exercise can increase the arm and leg muscle strength of CCSs. However, there is no previous evidence specifically reporting the effects of a supervised exercise intervention on muscle strength in CCSs. Our meta-analysis demonstrated that supervised exercise interventions could significantly improve muscle strength in CCSs, both during and after treatment, which is consistent with the effect of exercise interventions in adult cancer patients [51]. Importantly, this systematic review and meta-analysis adds some details for exercise interventions modality and evidence of supervised exercise interventions on the muscle strength of CCSs.
The treatment of cancer and its sequelae can significantly reduce physical activity and increase the fatigue of CCSs [52,53]. Fortunately, supervised exercise interventions can significantly increase the level of daily physical activity and improve fatigue for CCSs, as demonstrated in this review. In contrast, distance-delivered interventions had no significant effect on the level of daily physical activity among CCSs [22]. There has been no previous meta-analysis that has evidenced that exercise interventions can improve the level of daily physical activity and fatigue in CCSs [14]. This means that greater exercise practice is driven by supervised exercise interventions. Moreover, this review also confirmed that supervised exercise interventions are one of the most effective nonpharmacological strategies for improving fatigue. This meta-analysis included Li's study; however, the intervention frequency of Li's study was low and the sample size was large, which may have a great impact on the meta-analysis results. Therefore, further RCTs of high-quality and reasonable intervention programs are needed to strengthen this evidence and encourage supervised exercise interventions for CCSs.
Although the results of these three studies showed no significant effect on BMI [29,35,36], our pooled meta-analysis demonstrated that supervised exercise interventions increase BMI. In three studies [29,35,36], CCSs had a relatively low BMI during treatment. Malnutrition is considered an important predictor of decreased overall survival [54]. Ouyang [55] found that 55.8% of children with cancer were malnourished, and 74.2% had a moderate-to-high risk of malnutrition. The increase in BMI suggested that supervised exercise might indirectly improve the body weight and nutritional status of CCSs. Duan [56] also found that exercise can promote BMI in adult cancer patients. Therefore, supervised interventions can be considered in daily practice to improve BMI.
The results of this meta-analysis showed no significant improvement in QoL. However, significant improvements in physical outcomes from individual RCTs, including cardiopulmonary fitness [35,36], muscle strength [29,32,35,36], flexibility and balance [35], level of daily physical activity [32][33][34], and fatigue [32,34,36], may be considered important factors in improving the QoL. Consistent with our results, the studies of the current meta-analysis did not find a significant effect of exercise on the QoL of CCSs [14,15,22]. Another study showed significant beneficial effects on QoL in adult cancer survivors through supervised exercise interventions, but not through unsupervised interventions [43]. Given the limited number of studies included in the analysis, more data from high-quality RCTs are required to derive stronger evidence on the effect of supervised exercise on QoL in CCSs.
There were several limitations in our review. This meta-analysis lacked evidence of some important outcome measures. The original meta-analysis program intended to include anxiety and depression as outcome variables. However, the included studies did not assess anxiety or depression. As a result, this review lacked evidence of the psychological impact of supervised exercise interventions. Moreover, the methodological differences between RCTs introduced a moderate risk of bias, including that some trials did not blind subjects, interveners, and/or outcome evaluators. Furthermore, the statistical heterogeneity of our results might impact the ability to draw strong conclusions from the effects of the supervised exercise interventions. Although each study used a supervised exercise program, the great variability in the outcome measures and the intervention dose may be the cause of the heterogeneity. Due to the high variability and limited studies included in the meta-analysis, subgroup analysis could not be conducted to obtain optimum results of the type, intensity, frequency, and duration of supervised exercise interventions.

Conclusions
In summary, the adherence of CCSs to supervised exercise interventions was high. Supervised exercise interventions were able to improve muscle strength, the level of daily physical activity, BMI, and fatigue in CCSs. However, supervised exercise interventions did not achieve a statistically significant level to improve QoL. More high-quality RCTs are needed to further explore the optimal type, intensity, frequency, and duration of supervised exercise interventions for CCSs and to determine their impact on psychological outcomes. Nevertheless, this evidence indicates that supervised exercise interventions are safe and effective intervention strategies for CCSs. Therefore, we recommend that supervised exercise programs be implemented to improve the physical condition of CCSs during and after treatment.

Acknowledgments:
We are very grateful to the experts for their guidance on this systematic review and meta-analysis.