The electronic databases Embase, Lilacs, PEDro, PubMed and SCOPUS were searched using the terms: “Exercises” OR “Physical Activity” OR “Activities, Physical” OR “Activity, Physical” OR “Physical Activities” OR “Exercise, Physical” OR “Exercises, Physical” OR “Physical Exercise” OR “Physical Exercises” OR “Acute Exercise” OR “Acute Exercises” OR “Exercise, Acute” OR “Exercises, Acute” OR “Exercise, Isometric” OR “Exercises, Isometric” OR “Isometric Exercises” OR “Isometric Exercise” OR “Exercise, Aerobic” OR “Aerobic Exercise” OR “Aerobic Exercises” OR “Exercises, Aerobic” OR “Exercise Training” OR “Exercise Trainings” OR “Training, Exercise” OR “Trainings, Exercise” AND “Children” OR “Preschool Child” OR “Children, Preschool” OR “Preschool Children” AND “Grafting, Bone Marrow” OR “Bone Marrow Grafting” OR “Transplantation, Bone Marrow” OR “Bone Marrow Cell Transplantation” OR “Transplantation, Bone Marrow Cell” OR “Marrow, Bone” OR “Red Marrow” OR “Marrow, Red” OR “Yellow Marrow” OR “Marrow, Yellow” OR “Marrow” OR “Leukemia” OR “Cancer” AND “randomized controlled trial” OR “controlled clinical trial” OR “randomized controlled trials” OR “random allocation” OR “double-blind method” OR “single-blind method” OR “clinical trial” OR “clinical trials” OR NOT animal[mh] NOT human[mh].

The searches were performed from July to August 2019 by two independent authors. Title and abstract screening were applied to identify relevant articles and remove articles that were not eligible. Based on the satisfaction of the inclusion/exclusion criteria, the article was read in full.

Inclusion criteria included randomized controlled trials (RCTs) and non-randomized controlled trials (non-RCTs) including children and adolescents who had undergone bone marrow transplantation and had been part of exercise sessions, including aerobic or resistance training, in the inpatient or post-discharge phases, on whom studies had been written in the English language. The exercise intervention needed to be somewhat structured and described in enough detail to be repeated. A control group received usual care and no additional formal exercise intervention in their usual daily routine. Studies which compared one modality of exercise to another and still met the above criteria were included. No restrictions were applied as to the publication date of the articles, owing to the limited number of studies in a relatively new field.

Exclusion criteria included studies not written in English, providing no original data or that were not peer-reviewed. Studies that were written as abstracts only, rather than full papers, were also excluded.

The reviewers independently assessed the risk of bias in the inclusion of studies. The rating of the level of scientific evidence for each identified study was carried out using the Physiotherapy Evidence Database (PEDro) scale. Its criteria evaluate methodological qualities and validity characteristics internally, such as doing the study blindly, using statistics and analyzing with the intent to treat.

Measures of treatment effect were performed using the software program Review Manager 5 (RevMan 2014). Continuous data were analyzed using the mean difference (MD), 95% confidence intervals (CIs) being reported for all estimates.

The fixed-effect and random-effect models were applied, depending on the statistic heterogeneity. Moreover, the generic inverse variance method was used to combine RCTs into a single meta-analysis.

As a means to evaluate the heterogeneity of the data, an estimation of the treatment effects reported by individual studies, combined with the heterogeneity between trials, was performed. It included Floret plot analysis and quantification of the heterogeneity impact, using the I² statistic. The degree of heterogeneity was classified as less than 25% (without heterogeneity), 25%–49% (low heterogeneity), 50%–75% (moderate heterogeneity) or greater than 75% (substantial heterogeneity), as recommended by the Cochrane group. When there was evidence of apparent or moderate to substantial heterogeneity (I² greater than 50%), the possible causes were evaluated, searching for evidence of bias or methodological differences among the clinical trials.

Studies retrieved by the details of the results of the literature search are provided in Figure 1. The search criteria resulted in the retrieval of 732 articles, 723 of which were subsequently excluded after their titles and abstracts had been screened according to the eligibility criteria. Full-text screening of the remaining seven studies resulted in two RCTs and five non-RCTs that met all eligibility criteria and were included for review.

Figure 1.

Flow chart of search results.

(0.27MB).

Thus, seven RCTs (7 publications) met the eligibility criteria, reporting at least one of the outcomes established for this study. All were parallel studies. In summary, this is the general information on the included studies:

• •

  Sample size ranging from 6 to 70 children and adolescents;

• •

  Participant age, ranging from 3 to 19 years old, with an average between 8 to 9 years of age.

• •

  The studies included both hospitalized and discharged post-HCST patients. Three RCTs described inpatient interventions, three RCTs combined inpatient and home-based exercises and one RCT described only home-based exercises (Table 1).

  Table 1.

  Outcomes from exercise intervention.

  Study	Group	type	Frequency	Duration	Variables	Results
  Chamorro-Vina et al.11	Exercise (EG)	In-hospital	5×∼50min per week	3 weeks	• Immune cells (blood counts of leukocytes, monocytes, lymphocytes, and lymphocyte subpopulations	The immune cells measurement showed no significant responses pre- and post-transplantation in both groups. However, the post-transplant decrease was more abrupt in the CG than that in the EG.
  	n=7	Exercise program (EP)
  	Control (CG)	EG: aerobic and resistance exercises			• Anthropometric variables (body mass, body mass index and estimated fat-free mass)	Significant effect of the EG versus time for all anthropometric variables (weight, body mass index, body fat, and fat-free mass), indicating an increase over the hospitalization period in the EG compared to CG.
  	n=13	CG: non-exercise program
  Dubnov-Raz et al.13	EG	Home-based EP	3×60min per week,	6 months	• Physical fitness	No differences found between the groups in changes in physical fitness indexes, body composition or mental health. Significant correlations were found in bone density and lean body mass in all participants.
  	n=12	EG: endurance, strengthening and stretching training.			• Body composition
  	CG	GC: non-exercise program			• Bone density
  	n=12				• Mood
  					• Quality-of-life (QoL)
  Kabak et al.14	EG	Combined in-hospital and home-based EP	7×(20–40)min per week.	1 month	• Physical function	The study determined that the exercise program had positive effects in physical function and functional performance in the EG, in contrast with the worsened performance in the CG. Regarding ADL, no difference was found between the two groups.
  	n=11	EG: strengthening, endurance, stretching, and relaxation exercises.			• Functional performance
  	CG	CG: non-exercise program			• Activities of daily life (ADL)
  	n=11				• 6MWT
  Chamorro-Vina et al.11	EG	Combined in-hospital and home-based EP	5×(20–60)min per week	10 weeks	• Immunological recovery	Significant increase in NKCC concentration eight times greater than that found initially among EG.
  	n=3	EG: aerobic exercise strength training, stretching exercises
  	CG	GC: non-exercise program
  	n=3
  Wallek et al.16	EG	In-hospital EP	5×(40–60)min per week	Up to hospital discharge	• 6-min walking distance (6MWD)	Intergroup comparison revealed significant differences between the two groups. The EG increased their 6MWD by +8%, whereas both CGs presented a decline in the 6MWD. At discharge, the EG achieved 80% of healthy reference values.
  	n=27	EG: endurance training, muscle strengthening and flexibility training
  	CG	GC: non-exercise program
  	n=27
  Senn-Malashonak et al.15	EG	In-hospital EP	5×(30–60) min	Up to hospital discharge	• Maximal isometric knee extension strength (KES)	KES, 6MWD, and HGS significantly decreased in the CG, while there were no changes in the IG. Pre- to post-changes in 6MWD and HGS differed significantly between groups. QoL declined in both groups.
  	n=35	EG: resistance, endurance, and flexibility training			• Hand grip strength (HGS)
  	CG	GC: mental and relaxation training			• 6MWD
  	n=35				• QoL
  					• Medical parameters
  Yildiz Kabak et al.19	EG	Combined in-hospital and home-based EP	5×(30–60)min	Up to hospital discharge	• Depression	Pain intensity and days with pain were lower in IG (p<0.05).
  	n=15	EG: resistance, endurance, and flexibility training		And 1 month after discharge	• Pain	QoL declined in both groups.
  	CG	CG: non-exercise program			• Fatigue	Level of depression and Fatigue were lower in IG.
  	n=11				• QoL
  					• Parents QoL

  EG: exercise group; CG: control group; EP: exercise program; QoL: quality of life; IG: intervention group; KES: knee extension strength; HGS: hand grip strength; 6MWT: six minute walk test; ADL: activity daily life; NKCC: NK cell cytotoxity.

After full-text reading of the studies alleged for eligibility, three were excluded. San Juan et al.8 included healthy children who were not subjected to HCST in the control group. In addition, the Bogg et al.9 study did not have a control group, even though it had met several outcomes of interest outlined in this systematic review (SR) and meta-analysis. Lastly, the Chamorro-Viña et al.10 protocol was excluded.

A search was carried out on the clinicaltrials.org. Two ongoing studies were identified which involve pediatric and adolescent patients undergoing HCST who will take part in a physical exercise program. An Italian study (NCT04090268) started in 2017 aims to enroll 380 patients in a combined exercise protocol for 12 weeks. A French study in its initial phase (NCT04331483) was designed to treat 30 pediatric patients through a supervised EP four times per week for 3 months. The latter is expected to be completed in 2023.

After establishing similarities between the outcomes assessed in the RCTs, three significant results related to physical exercise in children and adolescents after HCST were selected. These are QoL; physical performance by the distance walked in meters in the 6MWT, as well as the percentage of the predicted distance.

The Dubnov-Raz et al.13 and Yildiz Kabak et al.19 studies, using the PedsQL to assess QoL, and data from 25 patients in the EG and 22 patients in the CG were included in the statistical analysis. The result represented in the forest plot shows statistical significance, with lower values for QoL in the CG. Physical exercise prevents a reduction in the QoL of these patients (Figure 2).

Figure 2.

Forest plot — quality of Life.

(0.17MB).

Three studies employed the 6MWT to assess the physical performance of post-HCST children and adolescents. Kabak et al.14 used the distance walked in meters as an outcome measure, whereas Senn-Malashonak et al.15 used the percentage from the 6MWT distance walked, as predicted by formulas. In contrast, Wallek et al.16 assayed the distance walked in meters and the percentage. The forest plots (Figures 3 and 4) present the statistical difference in the shorter distance walked in the CG, both in meters and percentage of the predicted value.

Figure 3.

Forest plot 6MWT (distance covered in meters).

(0.16MB).

Figure 4.

Forest plot 6MWT (distance walked in percentage of the predicted value).

(0.15MB).

The risk of bias was assessed using the PEDro scale. The methodological quality of the studies lost somewhat in relation to randomization, blinding and the analysis of the intention to treat. However, all of them presented baseline data, an intergroup comparison and a follow-up of at least 85% of the initial sample, indicating little significant sample loss (Table 2).

The assessment of evidence quality was carried out using the GRADE approach (Grading Recommendations Assessment, Development, and Evaluation), used to rate the body of evidence at the outcome level, rather than the study level. To determine the strength of the recommendation, GRADE takes into account key factors, such as the risk of bias, inconsistency, indirectness, publication bias and other biases. The quality rating has four levels in certainty in evidence, being very low, low, moderate and high, in which the high rate intervention should be strongly recommended.

Regarding the evidence quality, this SR and MA observed a low certainty in evidence in two outcomes (QoL and 6MWT). Although the studies individually show a moderate methodological quality, due to the small number of patients included in the general sample, the quality of the evidence is low, requiring further studies, as the research question has not yet been answered (Table 3).