Prevalence and Factors Associated With Cancer‐Related Fatigue Among Children and Adolescents Undergoing Cancer Treatment: A Systematic Review and Meta‐Analysis Authors: Sapna Oberoi, Beili Huang, Rasheda Rabbani, Nicole Askin, George Okoli, Richa Jain, Lillian Sung, Maya M. Jeyaraman, Alyson Mahar, Roberta Woodgate, Ryan Zarychanski Published: 2024 ABSTRACT ABSTRACTBackgroundThe literature on fatigue in children and adolescents undergoing cancer treatment is highly variable, creating uncertainties about its prevalence and identifying those at higher risk.ObjectivesThe primary purpose was to describe the prevalence of fatigue among patients (< 21 years) undergoing cancer treatment across cancer types. Secondary outcomes included the prevalence of severe fatigue and factors associated with fatigue.MethodsSystematic searches of MEDLINE, Embase, Cochrane Central Register of Controlled Trials, CINAHL, and PsycINFO were conducted from inception to May 22, 2023. Two reviewers independently identified relevant citations and extracted data. Pooled prevalence estimates were derived using an inverse variance, random‐effects model. We used Joanna Briggs's critical appraisal checklist to assess study quality. (PROSPERO: CRD42020179307).ResultsWe included 47 studies: 26 for prevalence and 29 for factors associated with fatigue. The pooled prevalence of fatigue was 73% (95% [Confidence Interval, CI: 66%–79%; I 2 96%; 26 studies; 2699 patients], and severe fatigue was 30% [95% CI 14%–46%, I 2 98%; 8 studies; 1027 patients]). Subgroup analyses based on cancer type, study design, fatigue scale, fatigue reporting personnel, sample frame, and response rate did not reveal significant differences in fatigue prevalence. Fatigue prevalence significantly differed by treatment setting (inpatient [83%] vs. outpatient [55%] vs. inpatient and outpatient [69%]; p: 0.02). Due to considerable heterogeneity among studies, data on fatigue‐associated factors are presented descriptively.ConclusionsThe prevalence of fatigue among children and adolescents undergoing cancer treatment is variable but notably high. Systematic evaluation of factors associated with fatigue is essential to understanding which children are at high risk of developing fatigue.Trial RegistrationPROSPERO: CRD42020179307 Introduction Fatigue is a debilitating symptom commonly experienced by children and adolescents with cancer throughout their cancer journey [1]. Fatigue can result from a range of factors, including cancer itself, treatment‐related toxicities such as pain, infection, and anorexia, as well as psychological distress, anxiety, depression, and sleep disturbances. Fatigue significantly impacts children's ability to engage in daily activities, such as attending school, participating in sports, and maintaining social relationships. Beyond physical tiredness, it also affects their cognitive and psychological functioning, leading to psychological‐emotional distress, particularly in adolescents with cancer [2, 3]. Fatigue has been reported to be one of the most crucial factors impacting the quality of life of these children and adolescents [4]. The reported prevalence of fatigue among children and adolescents with cancer during treatment is heterogeneous, varying from 14% to 67%, and has been shown to depend on various demographic and treatment‐related factors [5, 6, 7, 8]. While a few systematic reviews have analyzed interventions to mitigate fatigue in adults and pediatric patients with cancer, none have evaluated the prevalence and correlates of fatigue in the pediatric population during treatment [9, 10, 11]. Such a review can help assess the overall prevalence of fatigue in children with various cancer diagnoses during treatment by synthesizing data from multiple studies to provide a more precise estimate of fatigue prevalence. Investigating potential factors associated with fatigue is also essential to identify children and adolescents at risk of experiencing more fatigue during cancer treatment. Overall, this information can provide evidence‐based insights to inform the development of strategies for screening, preventing, and treating fatigue in children and adolescents during treatment. Therefore, the primary objective of this systematic review and meta‐analysis was to determine the prevalence of fatigue, regardless of its severity and etiology, in children and adolescents receiving cancer treatment. The secondary objectives included assessing the prevalence of severe fatigue and identifying factors associated with the occurrence of fatigue of any severity in this patient population. Methods We conducted this systematic review following the methodological approaches outlined in the Cochrane Handbook for Systematic Reviews of Interventions and adhered to the Meta‐analysis of Observational Studies in Epidemiology (MOOSE) guidelines for reporting meta‐analyses of observational studies [12, 13]. This review was registered with PROSPERO (CRD42020179307) [14]. Search Strategy and Study Selection A systematic search was conducted using Medline (Ovid), Embase (Ovid), Cochrane Central (Wiley), CINAHL (EBSCO), and PsycINFO (Ovid) from inception to May 22, 2023 (Table ). Reference lists of narrative and systematic reviews and the included studies were screened for additional citations. Additionally, we searched the gray literature to capture relevant studies. Inclusion and Exclusion Criteria Studies were included if they met the following criteria: (1) Participants were under 21 years of age with cancer or hematopoietic stem cell transplant (HCT) recipients; (2) the intervention included chemotherapy, radiotherapy, surgery, HCT, targeted therapy, or immunotherapy for cancer treatment; (3) reported prevalence, risk factors, or other variables associated with fatigue irrespective of the cause of fatigue; and (4) the study design was randomized controlled trials (RCTs), cohort studies, controlled before‐and‐after studies, or cross‐sectional studies. RCTs and controlled before‐and‐after studies had to report the prevalence of fatigue at baseline before the initiation of the intervention, provided they did not have the presence or severity of fatigue as an eligibility criterion for enrolling patients. Studies were excluded if they (1) included < 75% of children and adolescents with cancer; (2) exclusively included children and adolescents with relapsed/refractory cancer or those receiving palliative care or had > 25% of children and adolescents with relapsed/refractory cancer or those receiving palliative care; (3) solely included childhood cancer survivors or had > 25% of the population off cancer treatment; (4) did not report any outcomes of interest to this review; (5) reported fatigue only at cancer diagnosis; or (6) were non‐English language studies. Outcomes The primary outcome of interest was the prevalence of overall self‐ or proxy‐report fatigue measured by any fatigue assessment scale, except when fatigue was reported as toxicity using Common Terminology Criteria for Adverse Events (CTCAE) grading. Secondary outcomes included the prevalence of self‐ or proxy‐report study‐defined severe fatigue measured by any fatigue assessment scale, except for CTCAE, and study‐reported patient‐, disease‐, or treatment‐related factors associated with the presence of fatigue. Data Extraction Two reviewers (SO, RJ, BH, GO, or OL) independently evaluated the titles and abstracts identified by the search strategy. Any citation deemed potentially relevant by either reviewer was retrieved in full‐text format and assessed for eligibility. Discrepancies between the two reviewers were resolved through consensus, with adjudication by a third reviewer (GO or SO) if necessary. We extracted the following data: (a) Study‐level variables, including study design and setting (inpatient vs. outpatient), year of enrollment and publication, country of study, inclusion and exclusion criteria, and quality assessment items; (b) Population‐, disease‐, and treatment‐related variables: such as age at diagnosis, sex and ethnicity of participants, cancer type, cancer stage (nonmetastatic vs. metastatic), percentage of participants with relapsed/refractory disease, type of cancer treatment, and phase of cancer treatment; (c) Outcome‐related variables including the type of fatigue assessment scale used, outcome definition (i.e., specific diagnostic criteria or fatigue assessment scale cut‐off), time point of fatigue assessment, the person reporting fatigue, and the number of patients assessed for or presenting with fatigue and severe fatigue and factors associated with fatigue. In case of missing or unclear data, authors of the individual studies were contacted to provide the required information. Studies using the same patients or datasets were grouped as companion studies and treated as a single study in this systematic review. When studies reported fatigue prevalence from both patient and parent perspectives, the patient‐reported data were used in our meta‐analysis. We used the average prevalence for longitudinal studies that provided prevalence estimates at multiple time points during cancer treatment. Study Quality Assessment Two reviewers (SO and RJ, BH, OL, or GO) assessed the quality of observational studies using the Joanna Briggs Institute's (JBI) critical appraisal checklist for studies reporting prevalence data [15]. Nine domains of each study were considered for quality assessment, including sample frame, sampling method, sample size, study setting, sample coverage, validation of fatigue scale, fatigue measurement, statistical analysis, and response rate (Table S2). Data Synthesis We performed descriptive statistics using Microsoft Excel 2019 [Excel v15, Microsoft Corp., Redmond, WA, United States of America (USA)] and pooled aggregate data at the study level. If study‐level proportions and standard errors were not reported, they were calculated before meta‐analyses. We averaged the prevalence estimates for longitudinal studies reporting prevalence estimates at different time points during cancer treatment. The weighted summary proportions of fatigue and corresponding 95% confidence intervals (CI) were calculated by pooling the study‐specific estimates using generic inverse variance random‐effects models. Statistical heterogeneity between studies was assessed using the I 2 statistic, which describes the percentage of total variation across studies due to heterogeneity rather than chance; values around 30%–60% represent moderate heterogeneity, 50%–90% substantial heterogeneity, and 75%–100% considerable heterogeneity [16]. Potential publication bias was evaluated using visual inspection of funnel plot analysis and Egger's regression test for the primary outcome of fatigue prevalence [17, 18]. To explore sources of heterogeneity and to provide estimates of fatigue prevalence among relevant subgroups, stratified/subgroup analyses were planned based on a priori‐defined subgroups based on age, sex/gender, presence of mood or anxiety disorders and sleep disorders, cancer type and stage, treatment type, treatment setting, the person reporting fatigue, type of fatigue scale used, study type, and adequate sample frame, sampling and response rate [14]. These subgroup comparisons provided detailed insights into how fatigue varied within distinct patient populations and treatment contexts. We used the p value for subgroup difference to assess whether the prevalence of fatigue varied across subgroups, with a p value < 0.05 indicating a statistically significant effect. The meta‐analysis was performed using the metafor R package in R software (version 4.3.0) [19]. Results We identified 12,589 citations and screened 9016 after removing duplicates. Following a full‐text review of 236 articles, 47 studies met the eligibility criteria and were included in the review (Figure 1). PRISMA flowchart of the selection of the included articles. Study Characteristics Table 1 lists the baseline characteristics of all included studies. Of the 47 studies [5, 6, 8, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64] 26 contributed to the primary outcome of overall fatigue prevalence [5, 6, 8, 20, 21, 22, 23, 25, 26, 27, 30, 33, 34, 35, 36, 37, 40, 42, 44, 45, 49, 50, 53, 54, 56, 62] 29 to factors associated with fatigue [5, 20, 21, 22, 24, 26, 28, 29, 32, 35, 36, 38, 39, 41, 42, 43, 44, 46, 47, 51, 52, 55, 56, 57, 58, 60, 61, 65, 66] and 8 to the prevalence of severe fatigue [8, 20, 40, 49, 53, 54, 62, 67]. Geographically, 47% of studies were conducted in the USA, 13% in the USA and Canada, 6% in China, 4% in Canada, and 4% in Turkey. Prospective cohort studies comprised 60%, while cross‐sectional and phase 3 RCTs comprised 30% and 9%, respectively. Among the 30 studies from the USA and Canada, 25 reported enrolments of non‐white participants with a median representation of 28% (range: 8%–75%). Sex/gender distribution was reported in all studies, with a median of 56% males (range: 41%–80%). Cancer diagnoses varied widely: 64% of studies included patients with different cancer types, 21% focused solely on acute lymphoblastic leukemia (ALL), and 4% included both acute myeloid leukemia (AML) and ALL. Two studies exclusively enrolled patients with Central Nervous System (CNS) tumors. Regarding the study setting, 47% included patients receiving inpatient and outpatient treatments, while 19% and 17% focused exclusively on inpatient or outpatient settings. Modalities of cancer treatment included chemotherapy (51%), chemotherapy/radiation/surgery (13%), chemotherapy/radiation (4%), radiation (2%), and HCT (2%). TABLE 1Summary of studies included in the systematic review (N = 47).Study ID (author and year)Country of studyStudy typeStudy settingInclusion of HSCT patientsAge of participants (years), mean (range)Ethnicity (non‐White%)Sex (% male)Type of cancerTotal no. of participants in studyPerson reporting fatigueFatigue prevalence reportedGandy et al. 2022 [20]USAProspective cohortNRNo8.3 (3–16)54.557.6CNS tumor37ParentYesIrestorm et al. 2022 [21]NetherlandsProspective cohortNRNR6.3 (NR)NR56.5ALL127ParentYesJacobs et al. 2022 [22]USA and CanadaProspective cohortBothNR13.0 (7–18)22.251.4LL257PatientYesWeaver et al. 2022 [23]USAProspective cohortBothNo13.0 (7–17)43.554.7Mixed a 492PatientYesWu et al. 2022 [24]ChinaCross‐sectionalBothNR8.9 (NR)NR60.0Mixed a 40PatientNoBradford et al. 2021 [25]AustraliaCross‐sectionalOutpatientNo12 (8–18)NR69.0Mixed a 48PatientYesCheng et al. 2021 [26]ChinaCross‐sectionalBothNRNR100.066.8Mixed a 187PatientYesCheng KK et al. 2021 [27]SingaporeProspective cohortOutpatientYes13.7 (10–18)NR62.0Mixed a 50PatientYesBrown et al. 2021 [28]USACross‐sectionalNRNR8.5 (2.6–17.3)75.056.0ALL171Parent and patientNoDaniel et al. 2020 [29]USACross‐sectionalOutpatientNR10.1 (5–17)24.544.4Mixed a 59PatientNoLi et al. 2020 [30]ChinaCross‐sectionalBothNo8.9 (NR)NR61.0ALL and AML159Parent and patientYesRostagno et al. 2020 [31]ItalyProspective cohortBothNR11.7 (5–17)NR54.5Mixed a 134Parent and patientYesSteur et al. 2020 [32]NetherlandsProspective cohortOutpatientNR(1–19)NR57.0ALL151ParentNoCadamuro et al. 2020 [33]BrazilCross‐sectionalBothYesNR52.249.7Mixed a 157Parent and patientYesKudubes et al. 2019 [34]TurkeyPhase 3 RCTInpatientNR9.4/9.1 (control and experimental group, NR)NR57.5Mixed a 80Parent and patientYesNagarajan et al. 2019 [35]USA and CanadaProspective cohortBothNo(2–18)17.052.6AML560Parent and patientNoRogers et al. 2019 [36]USAPhase 3 RCTInpatientYes9.5 (4–19)18.260.6Medulloblastoma43Parent and patientYesTomlinson et al. 2019 [8]USA and CanadaCross‐sectionalBothYesNRNR61.3Mixed a 366 b PatientNoHockenberry et al. 2018 [65]USAProspective cohortBothNRNR59.155.9ALL191Parent and patientNoMacpherson et al. 2018 [37]USAProspective cohortBothNo(8–18)56.354.1Mixed a 96PatientYesDobrozsi et al. 2017 [38]USAProspective cohortBothNR11.7 (5–21)28.058.0Mixed a 41PatientNoRodgers et al. 2016 [39]USAProspective cohortNRNR(3–12)47.045.0ALL38Parent and patientNoBastani et al. 2015 [40]IranPhase 3 RCTInpatientNR10.0 (8–12)NR68.3ALL120PatientYesCrabtree et al. 2015 [41]USAProspective cohortNRNR7.7 (2–18)19.051.0Mixed a 170Parent and patientNoMDR Nunes et al. 2015 [42]USADescriptive with repeated measuresOutpatientNo12.8 (8–17)65.748.6Mixed a 42PatientYesRogers et al. 2014 [43]USA and CanadaProspective cohortOutpatientNR8.8 (5–18)18.365.5ALL100Parent and patientNoAmeringer et al. 2013 [44]USAProspective cohortBothNR15.3 (13–18)66.755.6Mixed a 9PatientYesHinds et al. 2013 [45]USACross‐sectionalBothNR12.9 (8–17)NR55.5Mixed a 93PatientYesWesley et al. 2013 [46]USACross‐sectionalBothNR15.6 (13–19)38.054.9NR123PatientNoHooke et al. 2011 [47]USAProspective cohortNRNR(6–17)NR66.7Mixed a 30PatientNoMiller et al. 2011 [48]USAProspective cohortInpatientNR13.5 (10–17)59.043.6Mixed39PatientYesBaggott et al. 2010 [5]USAProspective cohortBothNR14.8 (10–18)NR51.5Mixed a 66PatientYesDupuis et al. 2010 [49]CanadaCross‐sectionalBothNR9.4 (4–18)NR55.0Mixed a 200ParentYesErickson et al. 2010 [50]USAProspective cohortNRNR16.1 (12–19)15.050.0Mixed20PatientYesHockenberry et al. 2010 [51]USAProspective cohortBothNR(7–18)52.057.0Mixed a 67Parent and patientNoWalker et al. 2010 [6]USAProspective cohortBothNR14.2 (10–19)25.056.9Mixed a 51PatientYesZupanec et al. 2010 [52]CanadaCross‐sectionalOutpatientNR(4–18)48.479.7ALL77Parent and patientNoSitaresmi et al. 2009 [53]IndonesiaProspective cohortBothNR7.1 (2–16)NR63.0ALL51ParentYesYeh et al. 2009 [54]TaiwanProspective cohortBothNR14.2 (10–18.9)NR58.0Mixed a 144PatientYesEkti Genc et al. 2008 [55]TurkeyPhase 3 RCTInpatientNR9 (7–12)NR61.7ALL and AML60Parent and patientNoEnskar et al. 2008 [56]SwedenCross‐sectionalBothNo9.6 (NR)NR59.0Mixed17PatientYesPerdikaris et al. 2008 [57]GreeceProspective cohortNRNR8.9 (7–12)NR65.0Mixed a 40NRNoWhitsett et al. 2008 [58]USA and CanadaProspective cohortInpatientNR13.7 (9–17)8.366.7Mixed12Parent and patientNoYeh et al. 2008 [59]TaiwanProspective cohortInpatientNR11.5 (7–17)NR54Mixed a 48Parent and patientNoHinds et al. 2007 [60]USAProspective cohortInpatientNR12.5 (7–18)28.041.0Mixed a 29Parent, patient and healthcare providerNoHinds et al. 2007 [61] (cancer journal paper)USA and CanadaProspective cohortOutpatientNR(5–18)21.062.0ALL100Parent and patientNoWilliams et al. 2006 [62]USACross‐sectionalInpatientNo10.4 (2–18)9.045.0Mixed a 11PatientYesAbbreviations: ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; CNS, central nervous system; HCT, hematopoietic stem cell transplant; LL, leukemia and lymphoma; NR, not‐reported; USA, United States of America. a Mixed: mixed cancer types defined as a category where more than two types of patients with cancer were included. b Study enrolled 502 patients, of which, 366 were on treatment. Summary of studies included in the systematic review (N = 47). Abbreviations: ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; CNS, central nervous system; HCT, hematopoietic stem cell transplant; LL, leukemia and lymphoma; NR, not‐reported; USA, United States of America. Mixed: mixed cancer types defined as a category where more than two types of patients with cancer were included. Study enrolled 502 patients, of which, 366 were on treatment. Regarding fatigue assessment, 49% of studies used patient‐reported measures, 11% utilized parent‐reported measures, and 38% employed both measures. Fatigue assessment across the 47 included studies [5, 6, 8, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64] utilized 13 scales (Table S3). The most commonly used scales were the Fatigue Scale‐Child (FS‐C), Fatigue Scale‐Adolescent (FS‐A), or Fatigue Scale‐Parent (FS‐P) in 35% of studies, followed by the Pediatric Quality of Life‐Multidimensional Fatigue Scale (PedsQL‐MFS) in 17% and the Patient‐Reported Outcomes Measurement Information System (PROMIS) scale in 14%. Study Quality Assessment Among all studies, 38% had an appropriate sampling frame, 17% had adequate sample size and 36% had suitable study settings (Figure S1). A valid fatigue measurement scale was utilized in 90% of studies, and standard and reliable outcome measurement was employed in 98%. However, blinding was inadequate due to non‐blinded outcome measures. Appropriate statistical methods were used for the overall prevalence analysis in all 26 studies reporting prevalence (Figure S2). Among the 29 studies exploring factors associated with fatigue as a secondary outcome, only 41% adjusted for potential confounders. The response or fatigue assessment rate was deemed adequate in 68% (N = 32) of studies, indicating minimal attrition bias. Primary Outcome In total, 26 studies, comprising 2699 patients (aged 2–18 years), reported the prevalence of overall fatigue, with 2522 patients assessed (Table S4). The reported fatigue prevalence varied from 45.8% to 100.0%. The pooled prevalence estimate across these studies was 73% (95% CI, 66%–79%; I 2 = 96%) (Figure 2). Forest plot of the meta‐analysis of overall fatigue prevalence in included studies (N = 26). Subgroup analysis of overall fatigue prevalence based on predefined factors is detailed in Table 2. Significant differences in fatigue prevalence were observed only in subgroups categorized by treatment setting (p < 0.05). Subgroup analyses by age, sex, mood, or anxiety disorder, sleep disorder, and cancer stage were not feasible due to insufficient reporting in the included studies. TABLE 2Subgroup analysis of the primary outcome of overall prevalence of fatigue.SubgroupNo. of studiesNo. of participants enrolledPrevalence (%)95% CI (%) I 2 (%) p Value for subgroup difference Study type 0.40Cross‐sectional88726957–8194Prospective cohort1515847263–8093Phase 3 RCT32438764–10089 Setting 0.02Inpatient52938366–9990Outpatient31405544–6641Both inpatient and outpatient1520776961–7793 Type of cancer 0.58Mixed cancer types1919057062–7995Hematological malignancies57147965–9397CNS tumors2807747–10089 Type of cancer treatment 0.38Chemotherapy1110967868–8995Chemotherapy/Surgery/Radiation44756847–8998 Type of fatigue scale 0.48Childhood Fatigue Scale41857550–9995Memorial Symptom Assessment Scale53657260–8488Symptom Screening in Pediatrics Tool22055539–7176PROMIS Fatigue Subscale410296455–7387PedsQL Multidimensional Fatigue Scale44116952–8794 Person reporting fatigue 0.33Parent/Caregiver44157764–9081Patient1717116960–7896Patient and parent/Caregiver55738166–9595 Adequate sample frame 0.95Yes1117517164–7991Unclear128297361–8697No31197345–10094 Adequate sampling 0.78Yes1110337364–8392Unclear1112067463–8597No44606645–8797 Adequate response rate 0.90Yes1713887465–8396Unclear34257142–10098No68867161–8088Abbreviations: CI, confidence interval; CNS, central nervous system; PedsQL, Pediatric Quality of Life; PROMIS, Patient‐Reported Outcomes Measurement Information System; RCT, randomized controlled trial. Subgroup analysis of the primary outcome of overall prevalence of fatigue. Abbreviations: CI, confidence interval; CNS, central nervous system; PedsQL, Pediatric Quality of Life; PROMIS, Patient‐Reported Outcomes Measurement Information System; RCT, randomized controlled trial. Secondary Outcomes Severe Fatigue Severe fatigue prevalence was reported in 8 (17%) studies, encompassing 1027 patients (Figure 3 and Table S3) [20, 31, 40, 49, 53, 54, 62, 63]. These studies utilized six fatigue scales, with PedsQL‐MFS being the most frequently employed (25%, N = 2). Among the eight studies, 5 (62%) included patients with different cancer diagnoses. The pooled prevalence of severe fatigue across these studies was 30% (95% CI 14%–46%, I 2 = 98%). Forest plot of pooled prevalence of severe fatigue (N = 8). Factors Associated With Fatigue Twenty‐nine studies reported potential factors associated with fatigue among children and adolescents with cancer during active cancer treatment (Table 3). Meta‐analysis for this outcome was not conducted due to the considerable heterogeneity in the methods used to analyze and report (e.g., correlation, univariable, or multivariable analysis) and the study populations; therefore, data on the factors associated with fatigue are presented descriptively: (a) Demographic factors: Among six studies, three found older age associated with higher fatigue older age and higher fatigue [8, 22, 35, 36], two found no association [28, 60], and one found younger age associated with higher fatigue [38]. Seven studies examined the association of sex/gender with fatigue, with five finding no association and two identifying higher fatigue in females [22, 57]. Four studies explored race/ethnicity, with one finding Hispanic patients reporting higher fatigue, while the other three found no such association [22, 28, 35, 46]; (b) disease‐related factors: One study found that patients with CNS/solid tumors experienced less fatigue than those with leukemia/lymphoma [38], while another found no association with the type of leukemia/lymphoma [22]. Two studies in acute lymphoblastic leukemia (ALL) risk grouping showed mixed results, with one finding no association and the other reporting less fatigue in high‐risk and very high‐risk ALL patients compared to low and average‐risk patients [21, 28]; (c) treatment‐related factors: Among brain tumor patients undergoing radiotherapy, cranial spinal radiation dose was a significant predictor of fatigue in one study [20]. Dexamethasone pulse phases and steroid use during chemotherapy were associated with higher fatigue in ALL patients [32, 59]. In acute myeloid leukemia patients, the number of CTCAE toxicities was significantly associated with general fatigue [35]. Fatigue varied during treatment, decreasing over time in ALL patients and during the initial weeks of leukemia/lymphoma treatment [5, 41, 65]; (d) Sleep‐related factors: Seven studies found that poor sleep measures, such as inconsistent sleep habits, fragmented sleep, high caregiver‐reported sleep problems, frequent night awakenings and high nighttime activity, were associated with higher fatigue [32, 36, 44, 52, 60, 68, 69]; (e) Other clinical and psychosocial symptoms: Two studies found a positive association between fatigue and pain [26, 46] and two studies found a positive correlation between depressive symptoms and fatigue [44, 51, 58]. TABLE 3Factors associated with fatigue in the individual study included (N = 29).Study IDType of cancerFactors analyzedType of analysisFactors Associated/Correlated with greater fatigueFactors Associated/Correlated with lower fatigueGandy et al. 2022 [20]CNS tumorRadiotherapy (proton vs. photon), hydrocephalus (None vs. Any), craniospinal radiation, age at radiation therapyMultivariable regression analysisCraniospinal radiation during first week of therapyNoneIrestorm et al. 2022 [21]ALLCouse of fatigue, sex, risk group at diagnosis, serious adverse eventMultivariable regression analysisCourse of fatigue during treatment predicted fatigue reported at follow‐upNoneJacobs et al. 2022 [22]Leukemia and lymphomaGender, age, ethnicity, time since diagnosis, diagnostic group (ALL/AML/NHL/HL), and caregiver's education levelMixed effects modelFemale, Hispanic patientsMale, older age, non‐HispanicWu et al. 2022 [24]MixedQuality‐of‐life distressMultivariable regression analysisQuality‐of‐ life distressNoneCheng et al. 2021 [26]MixedPain interference, depression and lower mobilityMultivariable regression analysisGreater pain interference, depressive symptoms and less mobilityNoneBrown et al. 2021 [28]ALLAge at diagnosis, BMI, gender, race/ethnicity, type of leukemia, CNS involvement at diagnosis, high‐or very high‐risk ALL, and asparagine and gamma‐glutamyl glutamine in CSFMultivariable regression analysisLow‐risk and average‐risk ALL, asparagine and gamma‐glutamyl glutamine in CSFHigh‐risk and very‐high risk ALLDaniel et al. 2021 [29]MixedSleep timing, sleep consistency, technology use, presence of someone else, sleep disturbance, sleep‐related impairment, pain interference, nausea and sleep consistencyCorrelation and Simple mediation modelsSleep disturbances, Sleep‐related impairment, pain interference and nauseaConsistent caregiver‐reported sleep routinesSteur et al. 2020 [32]ALLDexamethasone pulses during maintenance chemotherapy, sleep–wake rhythm, stable sleep–wake rhythm, robust sleep–wake rhythm, more physical activity during the day and fragmented sleep–wake rhythmMultivariable regression analysisDexamethasone pulses, fragmented sleep–wake rhythmRobust sleep–wake rhythm, less fragmented sleep and higher physical activity during dexamethasone‐free periodsNagarajan et al. 2019 [35]AMLAge, sex, White race, Hispanic ethnicity, insurance status, high‐risk AML, bortezomib arm assignment, days of neutropenia, and number of submitted CTCAE toxicitiesMultivariable regression analysisOlder age and number of submitted CTCAE toxicitiesNoneRogers et al. 2019 [36]MedulloblastomaAge, percent sleep, longest sleep episode and nighttime activity score on actigraphy. Amplitude, 24 h auto‐correlation, intra‐daily variability, inter‐daily stability, dichotomy index on actigraphyLinear mixed models and correlationHigher age, lower percent nighttime sleep and higher nighttime activity scores (adolescent‐reported fatigue), longest nighttime sleep episode (child‐reported fatigue), dysregulated amplitude, 24 h auto‐correlation, and intra‐daily variabilityLower ageHockenberry et al. 2018 [65]ALL3NT (protein 3‐nitrotyrosine) in CSF and time since induction treatmentLatent class growth analysis and mixed modelsHigher 3NT in CSFTime from one treatment phase to another when measured from post‐induction to 12 months post‐induction chemotherapyDobrozsi et al. 2017 [38]MixedAge, gender, type of cancer, time since diagnosis, and intensity of therapyLinear mixed modelsLeukemia/lymphomaOlder age and diagnosis of solid/CNS tumorsRodgers et al. 2016 [39]ALLReduced glutathione (GSH) and reduced/oxidized glutathione (GSH/GSSG) ratio in the CSFCorrelationLow mean GSH/GSSG ratios in CSFNoneCrabtree et al. 2015 [41]MixedType of cancer, gender, age, and socioeconomic status, steroid use, radiation and chemotherapy, insomnia, sleep hygiene, bedtime, wake time, total sleep time, or restless sleep within 30 days of diagnosis and 8 weeks laterUnivariable analysis and multivariable regression analysisLonger sleep duration (6–12‐year‐old)Younger children with leukemia/lymphoma had a significant decline in parent‐reported fatigue within 30 days of diagnosis and 8 weeks later than those with solid tumor/CNS tumorsMDR Nunes et al. 2015 [42]MixedAge, gender, cancer type, sleep durationUnivariable analysis and correlationAdolescents, females, sarcoma, less sleep durationYounger children, malesRogers et al. 2014 [43]ALLCircadian activity rhythm parameters—peak, midline estimating statistic of rhythm (MESOR), amplitude, acrophase and circadian quotientLinear mixed modelsNonePeak activity, MESOR and amplitudeAmeringer et al. 2013 [44]MixedAnxiety and sleep disturbancesCorrelationDisturbed sleepHigher trait anxietyWesley et al. 2013 [46]NRAge, gender, minority status, pain, nausea, positive and negative affect, stressful life events, family support, friend support and family functioningCorrelationPain, nausea, and positive affectNegative affectHooke et al. 2011 [47]MixedGender, type of cancer and time since first three cycles of chemotherapyUnivariable analysisNoneFor young children, fatigue significantly decreased during the first three cycles of chemotherapy, and the ALL group had a greater decrease in fatigue than the lymphoma or the solid tumor group from cycle 1 to cycle 3 of chemotherapyBaggott et al. 2010 [5]MixedTime since administration of a chemotherapy cycleMultilevel logistic regression analysisNoneFor each week following a cycle of chemotherapy, odds of reporting fatigue were lower than the previous week for 2 weeks after the initiation of chemotherapy cycleHockenberry et al. 2010 [51]MixedDepression on day 7 of chemotherapyCorrelationDepressive symptomsNoneZupanec et al. 2010 [52]ALLPerceived problems in sleep, different sleep since diagnosis, different sleep place, sleep in 20 min, moving to different bed in the night and duration of weeknight sleepCorrelationSleep problems (4–12 years), Different sleep since diagnosis (4–7 years), moving to another bed (4–12 years)NoneEkti Genc et al. 2008 [55]ALL & AMLSex, diagnosis, age, hemoglobin, mucositis, nausea and vomitingCorrelationNoneNoneEnskar et al. 2008 [56]MixedLife satisfactionUnivariable analysisLess life satisfactionNonePerdikaris et al. 2008 [57]MixedGenderMultivariable regression analysisFemalesNoneWhitsett et al. 2008 [58]MixedDepressionCorrelationDepressionNoneYeh et al. 2008 [66]MixedSteroid use prior to start of chemotherapy cycle, steroids use per day of cycle, hemoglobin value, prior chemotherapy, cumulative doses of chemotherapy drugs in the cycleMultivariable analysisSteroids used before chemotherapy cycle, hemoglobin value, steroid use for each day of chemotherapy cycle and prior chemotherapyNoneHinds et al. 2007 [60]MixedNumber of nocturnal awakenings during inpatient hospital stay, age, diagnosis, gender, baseline fatigue, or length of hospitalization, hematocrit and hemoglobin levelMixed effect modelMore night awakenings (20 or more)NoneHinds et al. 2007 [61]MixedDexamethasone treatment, age, sex and ALL risk groupCorrelationDexamethasone treatmentNoneAbbreviations: ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; CNS; central nervous system; CSF, cerebrospinal fluid; CTCAE, common terminology criteria for adverse events; HL, Hodgkin lymphoma; LSS‐C, Life Situation Scale for Children; NHL, non‐Hodgkin lymphoma. Factors associated with fatigue in the individual study included (N = 29). Abbreviations: ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; CNS; central nervous system; CSF, cerebrospinal fluid; CTCAE, common terminology criteria for adverse events; HL, Hodgkin lymphoma; LSS‐C, Life Situation Scale for Children; NHL, non‐Hodgkin lymphoma. Publication Bias Visual inspection of the funnel plot exploring the publication bias of the studies contributing to the over fatigue prevalence showed that smaller studies reporting higher fatigue prevalence were underrepresented (Figure S3). Egger's regression test, assessing asymmetry in the funnel plot, demonstrated a statistically significant result (t = −3.77 and p—0.0009), indicating the presence of publication bias. Discussion In this review, we found high heterogeneity in summary effect measures, with the pooled prevalence of overall fatigue at 73% and severe fatigue at 30% in children and adolescents undergoing cancer treatment. These findings are comparable to a recent review demonstrating a prevalence of 62% among adults with cancer during treatment [70]. The high fatigue prevalence in our population also aligns with reports of increased fatigue among children and adolescents with chronic diseases such as cystic fibrosis and autoimmune diseases [71, 72]. Notably, this prevalence is seven times higher than the 10.1% prevalence of general fatigue among adolescents in the general population [73]. Stratified analyses by treatment setting revealed a lower prevalence among patients receiving outpatient treatment than those getting treatment either as an inpatient or both inpatient and outpatient. Typically, inpatient cancer treatments are more intensive than outpatient chemotherapy; plausibly, fatigue is more prevalent during inpatient chemotherapy cycles due to the increased intensity of therapy, increased distress associated with inpatient admissions impacting the psychological aspects of fatigue, fewer interactions with family and less time spent being physically active and more sleep disruptions due to night awakenings, and the noisy environment from the monitors [60, 74, 75, 76]. Additional studies are warranted to ascertain if outpatient delivery of a similar chemotherapy regimen leads to lesser fatigue than inpatient delivery. Studies reporting on fatigue‐related factors examined various demographic, disease‐related, treatment‐related, and psychological factors but revealed inconsistent findings due to substantial heterogeneity. However, most studies found consistent associations between sleep patterns and fatigue, with stable sleep habits linked to lower fatigue levels. The literature indicates a reciprocal and strong relationship between cancer‐related fatigue and sleep disturbances through shared physiological pathways [77, 78]. Future research should target pediatric‐specific interventions to improve fatigue and sleep quality in children and adolescents with cancer. Most studies demonstrated a positive association between pain, depression, and fatigue, in addition to sleep. This supports the concept of symptom clustering in oncology, an evolving concept that underscores the importance of clinicians recognizing the co‐occurrence of symptoms and the need to address multiple symptoms within a cluster to achieve better overall symptom control [51, 66]. Despite its heterogeneity, this systematic review's primary strength is in systematically presenting data on the prevalence of fatigue and associated factors among children and adolescents with cancer during treatment based on multiple studies conducted to date. A comprehensive search of several databases from their inception, utilizing broad search terms and robust methodology, was performed to avoid bias. Subgroup analyses were conducted to explore the causes of heterogeneity systematically. Finally, a multidisciplinary team provided expertise in conducting the review and interpreting the findings. This review has several notable limitations. The included studies exhibited substantial heterogeneity in patient demographics, cancer diagnoses, treatment modalities, fatigue assessment tools, and follow‐up durations. Consequently, despite pooling the prevalence data, our confidence in the estimated pooled prevalence remains low. The secondary objective of identifying factors associated with fatigue also suffered from inconsistent adjustment for potential confounders across studies. Only 3 out of 27 studies scored well on all the domains of study quality assessment for external validity, with 60% failing to specify sampling methods, thus raising concerns about selection bias. The reliance on self‐reported fatigue questionnaires precluded participant blinding, resulting in high detection bias. Additionally, variations in fatigue prevalence estimates may result from inconsistent reporting sources, such as caregiver versus child self‐reports. Caregiver/proxy reports may overestimate child‐reported fatigue, leading to discrepancies in prevalence estimates among studies [79]. Fatigue, a multifaceted construct, was assessed in this review as a general or total measure, limiting our ability to comment on specific dimensions of fatigue. The restriction to English‐language studies may have introduced publication bias, and the exclusion of studies focusing on children and adolescents with relapsed/refractory cancer or those receiving palliative care limits the applicability of the findings to these populations. Conclusions In conclusion, our systematic review and meta‐analysis reveal that 73% of children and adolescents undergoing cancer treatment experience fatigue, with nearly 30% enduring severe fatigue. Despite the significant heterogeneity among studies, indicating a low certainty in these estimates, the high prevalence and variable nature of fatigue during treatment are evident. The high prevalence and significant impact of fatigue on the quality of life and other critical aspects of life for children and adolescents with cancer underscore the need for routine assessment using valid, reliable, and psychometrically robust fatigue scales in clinical settings [3, 80]. Since effective interventions such as physical activity, mindfulness, and relaxation interventions to mitigate fatigue even in younger children are available, it is paramount that healthcare professionals should not consider fatigue an inevitable toxicity of treatment and utilize effective approaches to address it with their patients [81, 82, 83]. Future research should focus on identifying patterns and predictors of fatigue persistence beyond treatment, employing consistent assessment methods and multivariable analyses while ensuring the inclusion of younger children to address current gaps. Author Contributions Sapna Oberoi: conceptualization (lead), data curation (lead), formal analysis (lead), investigation (lead), methodology (lead), project administration (lead), resources (lead), visualization (lead), writing – original draft (lead), writing – review and editing (lead). Beili Huang: conceptualization (supporting), data curation (supporting), formal analysis (supporting), project administration (supporting), writing – review and editing (supporting). Rasheda Rabbani: conceptualization (supporting), formal analysis (equal), investigation (supporting), methodology (equal), software (lead), visualization (equal), writing – review and editing (supporting). Nicole Askin: data curation (supporting), investigation (supporting), methodology (supporting), resources (supporting), writing – review and editing (supporting). George Okoli: data curation (supporting), formal analysis (supporting), investigation (supporting), methodology (supporting), project administration (supporting), resources (supporting), writing – review and editing (supporting). Richa Jain: data curation (supporting), formal analysis (supporting), investigation (supporting), project administration (supporting), writing – review and editing (supporting). Lillian Sung: conceptualization (supporting), investigation (supporting), methodology (supporting), supervision (supporting), writing – review and editing (supporting). Maya M. Jeyaraman: conceptualization (supporting), formal analysis (supporting), investigation (supporting), methodology (supporting), supervision (supporting), writing – review and editing (supporting). Alyson Mahar: conceptualization (supporting), formal analysis (supporting), investigation (supporting), investigation (supporting), methodology (supporting), methodology (supporting), project administration (supporting), project administration (supporting), supervision (equal), supervision (equal), writing – original draft (supporting), writing – original draft (supporting), writing – review and editing (supporting), writing – review and editing (supporting). Roberta Woodgate: conceptualization (supporting), formal analysis (supporting), investigation (supporting), methodology (supporting), project administration (supporting), supervision (equal), writing – original draft (supporting), writing – review and editing (supporting). Ryan Zarychanski: conceptualization (equal), formal analysis (supporting), investigation (supporting), methodology (equal), project administration (equal), supervision (lead), writing – original draft (supporting), writing – review and editing (supporting). Ethics Statement This study is exempt from ethics approval as it is a systematic review that utilized data from previously published data from trials. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Data S1.