Abstract

Introduction:

Supervised exercise sessions in pediatric oncology offer numerous benefits, including mitigated treatment-related symptoms and improved physical and psychological outcomes. However, systematic assessment of adverse events (AEs) remains limited which restricts implementation and optimization of exercise safety in this vulnerable population. We aimed to prospectively assess AEs during exercise sessions, describe their characteristics and provide recommendations for action to reduce AEs.

Methods:

A prospective observational study was conducted at six pediatric oncology centers over three years to systematically record AEs and related information occurring during usual care exercise sessions from acute cancer treatment to post-treatment phase. An AE was defined as any adverse event temporally associated with an exercise session, regardless of whether it was causally related to the exercise itself. Each AE was categorized and graded according to the Common Terminology Criteria for AEs, and additional contextual information was documented in a centralized database. Based on the findings, AE characteristics were analyzed and a multidisciplinary consensus process was used to develop recommendations.

Results:

In total, 178 (75% grade 1; 23% grade 2; 2% grade 3) AEs were documented across 74,083 supervised exercise sessions, corresponding to 1 AE per 416 sessions. No life-threatening AEs were observed. Of the 178 AEs, 85% (151/178) were judged as exercise-related, resulting in an incidence of 204 per 100,000 exercise sessions (0.2%). Most common AE types were pain (53%; n = 94), nausea or vomiting (20%; n = 35), and circulatory problems (17%; n = 30). Overall, AEs were primarily triggered by physical (over)exertion (63%; n = 112), medical treatments (44%; n = 78), and fall-related incidents (23%; n = 41). Based on these findings and existing guidelines, 11 recommendations for action to reduce AEs were developed, including a consensus-based risk assessment, multiprofessional collaboration, and continuous professional education of exercise experts providing the exercise sessions.

Conclusion:

We found an overall low incidence of AEs during supervised exercise sessions in pediatric oncology, of generally low grade. The study highlights the need for prospective studies to refine evidence-based prevention strategies, test their effectiveness and implement them.

1 Introduction

Supportive therapies in oncology have the potential to improve patients' physical and emotional well-being. Consequently, a variety of interventions are being implemented and evaluated in both adult and pediatric oncology settings (1, 2). Among these, supervised regular exercise sessions have gained particular prominence (3). Pediatric cancer encompasses a heterogeneous group of diseases, most commonly including leukemia, central nervous system tumors, lymphomas, and solid tumors such as neuroblastoma and sarcomas (4). This diagnostic heterogeneity is associated with diverse treatment regimens. Growing evidence supports the benefits of supervised regular exercise sessions in children and adolescents during active cancer treatment (5) and in the post-treatment phase (6), including improvements in muscle strength (7), endurance (8–10), functional mobility (11), gait function (12), immune function (13), cardio protection (14), and quality of life (15–17). Preliminary evidence also suggests reductions in cancer-related fatigue (8, 18, 19) and hospitalization time (20). Physical activity interventions during acute treatment seem to be of particularly important, as a decline in exercise capacity can be prevented through targeted exercise (5), thereby helping to reduce therapy- and inactivity-related side effects and support patients in coping with the physical burden of cancer treatment (21).

Against this background, the safety of supervised regular exercise sessions has become an increasing focus of attention. Current evidence in pediatric oncology suggests that targeted, supervised exercise during the acute treatment phase (1, 10, 22, 23) and in the post-treatment phase (24) is generally safe. However, no detailed registry or systematic, prospective assessment of adverse events (AEs) during supervised exercise sessions in pediatric oncology currently exists (25).

Only a small number of available supervised exercise trials document the occurrence of AEs in a systematic way, with many studies not providing details on how and when AEs were assessed or reported (26). Reviews usually indicate very few small of AEs in the included studies; however, it remains unclear whether AEs were absent or simply not recorded (25, 27).

Systematic documentation of AEs is crucial, as cancer treatment, inactivity, and late effects of multimodal therapy may increase the risk of AEs during exercise due to pain (28, 29), balance disturbances (30, 31), functional impairments (32), bone density loss (33), muscle weakness, and neurological impairments (29, 30, 34). Despite these concerns, comprehensive data on the frequency, consequences, outcomes, and contextual conditions of AEs during supervised exercise in children and adolescents with cancer remain limited, particularly with regard to therapy phase, diagnosis, and age. This gap restricts reliable conclusions on exercise safety in this population. A retrospective analysis conducted in Germany in 2020 addressed this issue by documenting AEs during 35,110 supervised regular exercise sessions from 24 centers (22). The study reported 983 grade 1 AEs, with muscle soreness being the most frequent (43%), and a very low incidence of grade 2 and 3 AEs (0.017%).

Some recommendations for action to reduce AEs in pediatric exercise oncology are already addressed in the German S2k guidelines on supervised regular exercise sessions in pediatric oncology developed by the Association of the Medical Societies in Germany (AWMF) (35, 36). These guidelines provide general specifications regarding blood count thresholds, exercise expert-to-patient ratios, and the overall framework for supervised exercise. However, specific instructions for preventing AEs have not yet been incorporated, primarily due to limited empirical data. Similarly, the international Pediatric Oncology Exercise Guidelines (iPOEG) refer to AEs, but do not include concrete, evidence-or consensus-based recommendations (3).

Accordingly, there is a clear need for systematic and prospective investigation of AEs during supervised exercise sessions with children and adolescents during and after cancer treatment. The aims of the present study are therefore:

(1)

to prospectively describe incidence and characteristics of AEs during exercise sessions in pediatric oncology,

(2)

to examine explore potential trends in AE grade and AE triggers, and

(3)

to derive practical recommendations to reduce the risk for AEs during supervised exercise sessions in this population, informed by the empirical findings.

2 Materials and methods

This prospective, multicenter observational study (January 2021–December 2024) used a structured, guideline-based assessment framework to prospectively and comprehensively collect AEs during supervised regular exercise sessions, helping to reduce the risk of reporting bias and enabling characterization of the AEs and their settings. As illustrated in Figure 1, a three-phase mixed-methods design was applied, combining a descriptive analysis of the AEs with a systematic method for developing practical, empirically informed recommendations to reduce AEs.

Three-phase mixed-methods design for characterizing adverse events during exercise sessions, and developing practical, evidence-based recommendations for action to reduce adverse events. AE, adverse event; n, number of participants; NAOK, Network ActiveOncoKids.

The responsible ethics committee (Ethics Committee of the Medical Faculty, University of Duisburg-Essen (Germany) confirmed in advance that no formal application was required because of the study's anonymized and non-interventional design.

2.1 Study population

All participating acute-treatment hospitals were affiliated with the German Society for Pediatric Oncology and Hematology (GPOH) and were part of the nationwide Network ActiveOncoKids (NAOK) initiative in Germany. The study involved five acute-care pediatric oncology centers and one rehabilitation facility, each contributing data for at least 12 months from usual care exercise programs. Eligible participants were children aged 3 years and older, adolescents, and young adults up to 21 years of age who had been diagnosed with cancer and who participated in supervised exercise programs during the acute, maintenance, or post-treatment phases at the respective centers. Exercise programs were generally offered to all patients and tailored to the patient's current health status, treatment phase, and clinical condition in consultation with the treating medical team (35, 36). Due to data protection regulations, detailed information on individual diagnoses and treatment regimens was not available for the present analysis. However, participating centers covered a broad spectrum of pediatric oncology diagnoses, including leukemia, central nervous system tumors, lymphomas, and solid tumors.

No minimum fitness level was required for participation in the exercise sessions or the present study (see Inclusion and Exclusion Criteria), as the exercise program was tailored to each participant's fitness level. Tailoring was based on participants' age, individual goals, prior exercise experience, personal preferences, overall health status, and day-to-day condition. Accordingly, key training parameters such as exercise type, intensity, duration, and progression were adjusted. When necessary, programs were further modified in consultation with the interdisciplinary treatment team, including specialists from cardiology, nursing, physiotherapy, pulmonology, and psychology.

2.2 Inclusion and exclusion criteria

In accordance with current guidelines on exercise and physical activity in pediatric oncology (35, 36), no minimum fitness level was required for participation in the exercise sessions or for inclusion in the present study, as the program was individually tailored to each participant's physical condition. These guidelines specify that no absolute exclusion criteria exist, and that participation is determined by relative contraindications depending on the patient's current clinical status. Even very low-intensity interventions, such as progressive muscle relaxation in patients with poor general condition, were considered part of exercise therapy according to this definition of movement-based programs.

2.3 Supervised exercise

In addition to physiotherapy, all sites offered supervised regular exercise sessions. The goal of these interventions was to reduce therapy-associated and inactivity-related side effects, prevent late and long-term effects of medical treatment, and leverage the positive effects of physical activity on participants' physical and mental health. Exercise parameters were adjusted to each patient's current condition. They targeted core motor components such as strength, endurance, flexibility, coordination, speed, and mental training, which were combined and adjusted to the participant's exercise capacity. Activities ranged from light (e.g., reaction games for coordination training) to moderate (e.g., muscle strengthening with small hand weights while sitting or standing, or games involving throwing or returning a ball) to more intensive exercises (e.g., interval-based endurance training on an ergometer). All supervised exercise sessions were conducted by qualified exercise experts with specific expertise in pediatric oncology. While supervision by trained professionals was ensured across all centers, exercise professional-to-patient ratios during individual sessions were not systematically documented. Sessions were held 3 to 5 times a week, lasting at least 15 min and typically 30 to 60 min. Supervision was defined as continuous instruction and monitoring by an exercise expert, either in person or via real-time video communication. In the online setting, this enabled immediate feedback, correction of exercise technique, and individualized adjustment of exercises to ensure safety and proper execution. For online training, exercises were selected conservatively to minimize fall risk, and only participants deemed capable of safely performing the exercises at home were included. In addition, an adult was required to be present in the household during all online sessions. The number of exercise sessions conducted at each participating hospital was documented, reported every 6 months, and submitted to the study office.

2.4 Definition and recording of AEs

Consistent with the definition used by Gauss et al. (22), an AE was defined as any new physical or psychological symptom occurring for the first time during a supervised regular exercise session or worsening over the course of that session. This definition implies, for example, that pain following a fall onto the knee is judged as a new-onset AE, whereas symptoms already present before the session, such as nausea, are considered AEs when they worsen during the exercise session, for instance progressing to vomiting.

2.5 Documentation

AEs were reported by the responsible exercise expert to the study coordinator via telephone using a guideline-based interview that included both free-text fields and predefined answer categories. For some variables, only a single response option was available (e.g., location), whereas for others multiple responses could be selected (e.g., affected body parts).

The data collection process was iteratively refined over the course of the study within an agile framework. In particular, contextual variables such as consequences of AEs and their duration were introduced at later stages and were therefore not available for earlier cases. In some instances, missing information could be retrospectively derived from free-text entries; however, this was not consistently possible. Thus, missingness reflects the stepwise expansion of the documentation system rather than random data loss.

All information was entered in the REDCap digital database (Research Electronic Data Capture, developed by Vanderbilt University, Nashville, TN, USA; Version 15.9.1), hosted on a secure university hospital server (Essen University Hospital), and based on predefined questions covering four categories (Table 1) (entry template: https://www.activeoncokids.org/wp-content/uploads/2025/11/AERegistry_AERegistry-2.pdf or Supplementary Material File S1). Participants' fitness level was also assessed by the exercise experts and categorized using a three-level rating system (moderate, average, good), based on comparison with age- and sex-specific expected fitness levels. Definitions of AE triggers and AE types were specified a priori and are provided in Supplementary Material Table S1.

Category*Recorded variablesBasic informationDate; AE type**; AE trigger**; affected body parts**.JudgmentCTCAE grading, exercise related (yes vs. no), new onset and exercise-related.Consequences and outcomes (duration & intensity)Hospitalization; increased care needs; medication administration; fear and uncertainty** (among the treatment team, the affected individuals, the parents, and the exercise experts); refusal of further exercise offers; application of the RICE rule; temporary suspension of session; termination of session; programmatic adaptation* (communication strategy, equipment, exercise selection, intensity, location); life-threatening consequences or death.Background and settingTherapy phase; group size; patient age; time point; location; primary motor form*; fitness level.

Recorded variables for each adverse event during supervised exercise sessions in childhood cancer.

AE, adverse event; CTCAE, common terminology criteria for AE.

**

The questions underlying these variables were designed as multiple-response items.

2.6 Grading of adverse events

To ensure systematic grading of the reported AEs, a multilevel evaluation procedure based on the CTCAE, version 5.0 (

37

,

38

) was conducted every 6 months by a multidisciplinary expert group comprising experts in medicine (CS), epidemiology (CSR), and pediatric exercise oncology (TS, HS, MG, GG), together with project partners. In the first step, each AE was independently and confidentially graded using a five-point scale (grades 1 to 5) (

38

), which had been preadapted to the context of supervised exercise sessions. The decisive criterion for grading was the objectively medical or subjectively perceived consequence for the affected individual (

Supplementary Table S2

):

Grade 1 Interruption of the exercise with content modification,

Grade 2: Medical intervention [e.g., first aid following the “Rest, Ice, Compression, and Elevation (RICE)” rule, provided it was administered to treat an actual injury and not solely for teaching purposes or to provide comfort],

Grade 3: Medically relevant consequences (e.g., surgical procedures or functional impairments affecting daily activities),

Grade 4: Life-threatening measures.

Grade 5: Death.

In cases of disagreement in the initial assessments, a moderated consensus process was used until consensus was reached. To this end, two online meetings were held annually, during which unclear cases were discussed, and final grades were agreed on.

Exercise experts judged each AE as exercise-related or not, and whether it represented a new-onset symptom or a worsening of a pre-existing condition This assessment was performed in close consultation with the treating medical team at each center. AEs were considered exercise-related if they were judged unlikely to have occurred or worsened in the absence of physical activity (e.g., at rest). This pragmatic approach was applied across centers but relied on clinical judgment rather than predefined decision rules or formal standardization. In cases of ambiguity, particularly for symptoms such as nausea, dizziness, or pain, classification was based on the temporal relationship to exercise, symptom progression during the session, and overall clinical context.

2.7 Phase 1: development of recommendations by primary authors

Phase 1 consisted of two steps resulting in establishing data-driven recommendations (Figure 1). First, a quantitative analysis was performed (see Statistical Analysis section). Second, the primary authors (GG, MG) qualitatively and contextually screened all free-text responses using qualitative content analysis (39) to systematically identify specific irregularities, cases of precedence, and particular learning cases that could inform the development of evidence-based recommendations for action to reduce AEs.

2.8 Phase 2: finalization of recommendations by a multidisciplinary expert group

Phase 2 comprised four steps (Figure 1). The aim was to revise the evidence-based and practice-oriented draft of the recommendations from phase 1 in detail. For this purpose, a multidisciplinary working group consisting of eight experts from the fields of medicine, sports science, physical therapy, epidemiology, and nursing, as well as patient representatives, was convened (MG, CSR, CS, TS, HS, GG, MD, MM). The group integrated professional perspectives from the involved disciplines. All participants were provided with comprehensive and transparent information on the development process of the recommendations, including their rationale and wording, in order to ensure traceability and plausibility. The complete documentation is available from the authors upon request. Subsequently, the recommendations for action to reduce AEs were discussed in a 120-minute consensus meeting (40). Following an initial revision, a written round of comments was conducted via email before the final version was approved by all members.

2.9 Phase 3: external validation using voting by practice-based experts

Phase 3 comprised two steps to externally validate the eleven recommendations for action to reduce AEs (Figure 1) (41). For this purpose, all exercise experts working in German-speaking pediatric oncology centers who had completed GPOH-certified BOP training [“Bewegungstherapeutische InterventiOnen der Pädiatrischen Onkologie” (Exercise Interventions in Pediatric Oncology)] were invited to provide their assessment. External validation was conducted using a standardized online survey via the LimeSurvey application (LimeSurvey GmbH, Hamburg, Germany; Version 6.10.3 + 250203). Participants were provided with background information on the development process of the recommendations. They were then asked to approve or reject the recommendations and were also given the opportunity to provide comments.

2.10 Statistical analysis

To standardize exercise units in relation to the occurrence of AEs, each child's and adolescent's participation in an individual or group exercise session lasting at least 15 min was counted as one exercise unit. Accordingly, ten children participating in individual sessions corresponded to ten exercise units, while a group session involving 15 adolescents corresponded to 15 exercise units. Longer interventions (e.g., activity days or multi-day programs) were likewise counted per participating child/adolescent and per session, such that a six-day skiing program with eight children completing two sessions per day resulted in a total of 96 exercise units (8 × 6 × 2). The number of sessions stratified by format (e.g., individual vs. group) was not systematically recorded across centers.

Descriptive analysis of the 178 recorded AEs was conducted using Python (Python scripts were developed by SM), and the aggregated dataset was made available in tabular form on Github (https://github.com/smeisegeier/sport-adverse-events/tree/v1.0) (Supplementary Dataset 1). Multiple responses were documented sequentially and were not assigned to mutually exclusive categories.

Complete information was available for 105 recorded AEs. Missing data most frequently affected contextual variables, particularly consequences of AEs, their duration, and session-specific characteristics. Within this subgroup, we investigated differences in AE grade and trigger in relation to AE type, therapy phase, group size, exercise relatedness, impact on session continuation, age, location, and motor performance using univariate group comparisons. CTCAE grades 2 and 3 were merged into a single moderate-to-severe category. Group differences were assessed using chi-square tests in IBM SPSS Statistics (IBM Corp., Armonk, NY, USA; Version 31.0.1.0), with a significance level of α = 0.05. As assumptions for chi-square testing were frequently violated, p values were calculated using Fisher's exact test with Monte Carlo simulation. Multiple responses were treated as independent categories; for example, AEs attributed to both medical therapy and physical (over)exertion were counted as an additional combined category. Given the exploratory nature of the analyses, no formal correction for multiple testing was applied. Accordingly, reported p-values should be interpreted as descriptive and hypothesis-generating rather than confirmatory, and the risk of type I error should be considered. Proportions are reported with 95% confidence intervals (CI). All analyses were exploratory and descriptive in nature.

3 Results3.1 Incidence of adverse events

During the 3-year observation period, 178 AEs were documented across a total of 74,083 exercise units of supervised exercise sessions. Of these AEs, 133 (75%) were judged as grade 1, 42 (23%) as grade 2, and three (2%) as grade 3. The overall incidence rate was 240 AEs per 100,000 exercise units (0.2%), corresponding to one AE per 416 exercise units. On average, one grade 1 AE approximately every 561 exercise units, one grade 2 AE every 1,723 exercise units, and one grade 3 AE every 24,694 exercise units. Overall, a grade 2 or 3 AE occurred every 1,611 exercise units. Of the 178 AEs documented during the exercise sessions, 146 (82%) were judged as new-onset AEs, indicating that no symptoms had been present prior to the respective exercise unit. In 32 AEs (18%), pre-existing symptoms were present and worsened during the session. Among the 178 AEs, 151 (85%) were judged as exercise-related and were considered unlikely to have occurred in the absence of the exercise session. In contrast, 27 AEs were judged as non–exercise-related AEs and were considered likely to have occurred independently of physical activity, for example while watching a movie. Overall, 132 of the 178 documented AEs (74%) met both criteria simultaneously, namely new-onset and exercise-relatedness, indicating that the AE occurred during the exercise session and did not represent a worsening of any pre-existing symptoms (Table 2).

ParameterGrade 1Grade 2Grade 3TotalN (%)N (%)N (%)N (%)CTCAE grading133 (75%)42 (23%)3 (2%)178 (100%)Onset status132 (74%)43 (24%)3 (2%)178 (100%)New onset109 (61%)35 (20%)2 (1%)146 (82%)Pre-existing and aggravated23 (13%)8 (5%)1 (1%)32 (18%)Exercise-related113 (63%)36 (20%)2 (1%)151 (85%)New onset + exercise- related97 (74%)33 (25%)2 (2%)132 (71%)

Distribution of adverse events by CTCAE grading, onset status, and exercise relatedness during supervised exercise sessions.

Remarks: The total number of events varies due to missing information in some of the variables.

AE, adverse event; CTCAE, common terminology criteria for AE; N, total number; n, number of participants.

3.2 Characteristics of adverse events

Overall, AEs were predominantly mild and observed across all age groups, therapy phases, and fitness levels. Higher-grade AEs were rare and were observed more frequently in specific age groups, therapy phases, and fitness levels. Grade 3 AEs were observed only in patients aged ≥15 years. In the >18-year age group, grade 2–3 AEs were descriptively more frequent than grade 1 AEs (5 vs. 3 cases), whereas grade 1 AEs predominated in all other age groups. During acute cancer treatment, the ratio of grade 2–3 to grade 1 AEs was 27:120, whereas during exercise in the post-treatment phase it was 16:9. All grade 3 AEs occurred in participants judged as having a “moderate” fitness level. Across all fitness levels, grade 1 AEs were the most frequently documented category (Table 3).

ParameterGrade 1Grade 2Grade 3TotalN (%)N (%)N (%)N (%)Age (n = 133)2–5 years13 (10%)9 (7%)022 (17%)6–9 years34 (26%)6 (5%)040 (30%)10–14 years30 (23%)9 (7%)039 (29%)15–18 years17 (13%)6 (5%)1 (1%)24 (18%)>18 years3 (2%)3 (2%)2 (2%)8 (6%)Therapy phase (n = 178)Acute treatment119 (67%)27 (15%)1 (1%)147 (83%)Maintenance4 (2%)2 (1%)06 (3%)Post-treatment9 (5%)14 (8%)2 (1%)25 (14%)Fitness level (n = 124)Moderate53 (43%)17 (14%)3 (2%)73 (59%)Average26 (21%)10 (8%)036 (29%)Good10 (8%)2 (2%)012 (10%)Unknown3 (2%)003 (2%)

Distribution of adverse events by patient age, therapy phase, and fitness level during supervised exercise sessions.

Remarks: The total number of events varies due to missing information in some of the variables.

CTCAE, common terminology criteria for AE; N, total number; n, number of participants.

As summarized in Table 4, AEs occurred across a broad range of AE types, AE triggers, and affected body parts. The most frequent AE type combinations were soft-tissue or tissue injuries accompanied by pain (n = 13; 7%) and superficial injuries with pain (n = 9; 5%), with pain being reported across all severity grades. Regarding AE triggers, the most frequently documented causes were physical (over)exertion in combination with the consequences of medical treatment (n = 64; 36%), followed by fall-related incidents due to coordination problems (n = 10; 6%). For fall-related incidents, the ratio of grade 2–3 to grade 1 AEs was 20:21. With respect to affected body parts, grade 2–3 AEs were as frequent as grade 1 AEs in cases involving head injuries, whereas grade 1 AEs predominated in all other body parts.

ParameterGrade 1Grade 2Grade 3TotalN (%)N (%)N (%)N (%)AE type (n = 178)*Pain62 (35%)30 (17%)2 (1%)94 (53%)Nausea/vomiting31 (17%)4 (2%)035 (20%)Circulatory problems28 (16%)2 (1%)030 (17%)Soft-tissue injury7 (4%)14 (8%)021 (12%)Superficial injuries7 (4%)8 (4%)015 (8%)Psychological stress reaction6 (3%)1 (1%)07 (4%)Muscle soreness4 (2%)2 (1%)06 (3%)Physical (over)exertion4 (2%)004 (2%)Coughing fit4 (2%)004 (2%)Itching3 (2%)1 (1%)04 (2%)Bone injuries002 (1%)2 (1%)Enuresis2 (1%)002 (1%)Nosebleed1 (1%)1 (1%)02 (1%)Spontaneous painful bowel movement2 (1%)002 (1%)Muscle cramps1 (1%)001 (1%)AE trigger (n = 178)*Physical (over)exertion93 (52%)17 (10%)2 (1%)112 (63%)Medical treatment67 (38%)11 (6%)078 (44%)Fall-related incident21 (12%)19 (11%)1 (1%)41 (23%)Coordination problems14 (8%)