The following electronic databases were searched: PubMed (MEDLINE), PsycArticles, PsycInfo, CINAHL, and SPORTDiscus (EBSCO). In brief, the search blocks included keywords related to: (1) VO2peak; (2) Children; and (3) Cerebral Palsy. The full literature search is provided in the Supplementary material Table S1.

For this review, studies using an incremental exercise test aimed at testing the maximal exercise fitness were included. Studies using sub-maximal exercise tests were excluded. Original studies that reported findings on VO2peak measured with a gas analysis system, in children with CP, aged 18 years or younger, measured with a maximal exercise test were included. Exclusion criteria were (i) mixed diagnosis groups, unless data of the CP group were separately described, (ii) articles written in other languages than English, Dutch, German, or French, and (iii) systematic reviews, letters to the editor, and other studies without original VO2peak data. If TD peers were tested in the included studies, their VO2peak was also extracted. No exclusion criteria were applied for this control group.

Two reviewers (E.J.W. and H.B.) screened the search results by title and abstract, and subsequent full-text, using the web-tool Rayyan.17 Additionally, the reference lists of included studies and systematic reviews were screened for any potentially relevant studies. In case of a disagreement between reviewers, a discussion to include or exclude the article took place. The searches are up to date until November 15, 2023.

A pre-designed extraction form was used to collect relevant information from each included study. Data extraction was done by one reviewer (E.J.W.) and cross-verified by a second reviewer (H.B.). Extracted study and child characteristics included: last name first author; year of publication; number of participants, mean age; sex; GMFCS level; distribution and motor type of CP; type of maximal exercise test; VO2peak; maximal heart rate (HRmax); and respiratory exchange ratio (RER). When data were available, VO2peak was reported per subgroup according to the type of maximal exercise test, GMFCS level, distribution type of CP, and sex. The main outcome of interest was VO2peak as expressed in mL/kg/min or L/min.

The GMFCS describes the functional mobility level of children with CP.18,19 Children functioning at GMFCS level I are able to walk without limitations, those classified as GMFCS level II experience difficulty walking on uneven terrain, inclines, and in crowds or confined spaces. GMFCS level III reflects children who walk with a walking aid and use a wheelchair when covering longer distance. Children at GMFCS level IV are dependent on physical assistance or powered mobility in most settings, while children who are classified at GMFCS V are not able to walk or use a wheelchair by themselves.18,19

The motor type of CP can be subdivided into 3 types based on the dominant motor disorder: spastic, dyskinetic, or ataxic. In spastic CP, spasticity is the predominant disorder, with spasticity characterized by hypertonia and pathological reflexes, in particular increased stretch reflexes. Hypertonia can be elicited at the start of a movement, in which fast passive stretch results in a velocity dependent increase in muscle resistance. Children with spastic CP can be further characterized by the distribution of involved limbs (unilateral/bilateral). For the purpose of this study, hemiplegia was categorized as unilateral CP, whereas diplegia, tetraplegia, and quadriplegia were classified as bilateral CP. Children with dyskinetic CP predominantly have involuntary sustained or intermittent muscle contractions causing stereotyped movements and abnormal postures, which can be subdivided in dystonia and choreo-athetosis. Often, primitive reflexes persist. In ataxic CP, damage to the cerebellum causes lack of muscle coordination. Common features are balance and coordination problems.2

The main aim of our systematic review and meta-analysis was to summarize reported VO2peak values in youth with CP in all included literature, regardless of study design or main objective of the original studies. Therefore, the most important methodological quality question for our systematic review concerned the valid, unbiased measurement of VO2peak. The risk of biased VO2peak measurement was judged by means of three criteria: did the study report (1) how the VO2peak value was defined, (2) prior stated additional peak criteria related to maximal heart rate, respiratory exchange ratio, or signs of perceived exhaustion, and (3) reporting the number of children who successfully completed the maximal exercise test.12,15 The results of the quality assessment are reported in Supplementary material Table S2.

The study and child characteristics from the included studies were reported using descriptive statistics. The VO2peak outcome of the same children was re-used in a large number of included articles. Therefore, for data pooling, unique or so-called ‘independent’ VO2peak outcomes were used. Hereto, the outcomes of the article with the most complete (subgroup) data were used. In studies in which a maximal exercise test was repeatedly performed (e.g. test-retest study, intervention studies), data of the first test was used to exclude any learning and/or intervention effects. Verschuren and Takken20 used external reference data of 336 healthy controls for comparison. In this review, these external TD reference data were excluded. Two articles21,22 which fulfilled our inclusion criteria showed VO2peak data only in graphs. Authors of these studies were contacted and asked to share the VO2peak values needed for meta-analysis.

Meta-analyses in CP and TD subgroups were conducted on VO2peak data in mL/kg/min. A limited number of studies reported VO2peak in L/min or VO2peak adjusted for lean body mass, resulting in insufficient data for conducting meta-analyses. We estimated the pooled VO2peak with corresponding 95% confidence intervals (CI) according to the generic inverse variance method, using a random effect model. The difference between groups was tested with an unpaired t-test. To assess heterogeneity between study outcomes, I2 statistic was used: an I2 value > 75% was considered high heterogeneity. Meta-analysis were performed using SPSS version 28 (IBM Corp, Armonk, NY). A p-value < 0.05 was considered statistically significant.

The literature search yielded a total of 480 records, of which 36 studies20-55 met the inclusion criteria (Fig. 1). Children with CP were included from 13 different countries. Papers were published between 1978 and 2023. Data from 16 intervention studies, 9 cross-sectional case-control studies comparing children with CP and TD, 4 clinimetric studies (one test-retest study,27 and three studies47,53.54 comparing two different test modalities) were included. Seven studies had another study design (Table 1). Further details, including the risk of bias score of the 36 studies, ordered by the maximal exercise test used, are presented in Supplementary material Table S2.

Fig. 1.

Flow diagram of study selection.

(0.42MB).

Among the 36 included studies, there was overlap between study populations in 16 studies (44.4%).20-55 The study populations of Hoofwijk et al.32 and Unnithan et al.50 were identical, and therefore only the paper of Hoofwijk et al.32 was used in this review. There was overlap in study populations in the papers of Verschuren and Takken,20 Verschuren et al.,53 and Zwinkels et al.55 For pooling of the child characteristics the 2004 cohort data from Verschuren et al.53 and the 2014 cohort described by Zwinkels et al.55 were used. The papers of Verschuren and Takken20 and Verschuren et al.53,54 were used for subgroup meta-analysis. Nsenga Leunkeu et al.44,45 also reported training results of the same children in two articles. Data from the 2013 article45 were used for pooling. The trial population of van Wely et al.52 was used for secondary analyses by Balemans et al.24,25 In the largest study with 70 children (Balemans et al.23), trial participants, but also children from the study by Brehm et al.,27 were included. In this systematic review, the largest study by Balemans et al.23 was the starting point for pooling unique VO2peak data. The papers of Kim et al.,34 Lee et al.39 Park et al.,46 and Suk and Kwon49 also had an overlap in study populations. In this case, the largest data set of Kim et al.34 was used.

Taking the overlap between studies into account, 510 unique children with CP and 173 unique TD peers were included. Table 1 shows the extracted child characteristics (age, sex, GMFCS, and distribution and motor type of CP) of each included paper. Boys, and children with bilateral spastic CP at GMFCS levels I and II were in the majority.

VO2peak was measured using bicycle ergometer (n = 16), treadmill walking (n = 13), arm crank ergometer (n = 6), shuttle run test (n = 3), and shuttle ride test (n = 1). VO2peak values and physiological characteristics per study are presented in Supplementary material Table S3.

Table 2(A-D) presents the pooled VO2peak (in mL/kg/min) in subgroups of children with CP and TD peers. The overall estimated VO2peak in children with CP was 32.84 mL/kg/min (SE 1.28) and 45.02 mL/kg/min (SE 1.32) in TD peers, with a mean difference between CP and TD of −12.17 mL/kg/min (95% CI diff: −16.70, −7.64). (Table 2A) Figs. 2 and 3 show the forest plots of study outcomes in children with CP and TD peers.

Fig. 2.

Forest plot of VO2peak outcomes (in mL/kg/min) in children with CP. Studies are ordered by type of exercise test.

(0.39MB).

Fig. 3.

Forest plot of VO2peak outcomes (in mL/kg/min) in TD children. Studies are ordered by type of exercise test.

(0.19MB).

On all types of exercise tests, TD peers scored higher VO2peak values than children with CP (Table 2A). In children with CP, the highest pooled VO2peak was found on the shuttle run test: 37.23 mL/kg/min (SE 0.79).

Subgroup meta-analysis of six studies20,32,36,37,41,47 revealed that boys with CP scored higher VO2peak values than girls with CP: VO2peak of 39.43 mL/kg/min (SE 3.50) and 34.64 mL/kg/min (SE 1.26), respectively, with a mean difference of 4.79 mL/kg/min (95% CI: −8.10, 17.68). The sex-specific VO2peak data of TD peers were non-significantly higher (Table 2B). The difference between boys with CP and TD boys equalled −9.41 mL/kg/min (95% CI: −22.47, 3.65); between girls with CP and TD girls −3.10 mL/kg/min (95% CI: −17.68, 11.47).

Overall, seven studies reported on the VO2peak per GMFCS level, showing gradual differences between children at GMFCS level I, II, III, and III/IV with respectively a pooled VO2peak of 35.41 (SE 2.94), 32.05 (SE 2.57), 30.23 (SE 1.90), and 25.70 (SE 0.97) mL/kg/min (Table 2C).23,26,29,30,34,54,55

Seven studies reported on the VO2peak for CP subgroups with bilateral involvement or unilateral involvement (Table 2D).25,31,36,37,41,47,51 Children with bilateral involvement had a pooled VO2peak 32.77 mL/kg/min (SE 3.15); one study25 presented VO2peak data of children with unilateral involvement, i.e. 33.50 mL/kg/min (SE 1.21).

With regard to clinical practice and public health, it is highly relevant to understand the impact of low aerobic fitness of children with CP. A low aerobic fitness in childhood disability has clearly proven to have negative health consequences in the short term as well as the long term. Maximum oxygen uptake values below the threshold of 42 mL/kg/min for boys and 35 mL/kg/min for girls indicate potential cardiovascular risk.9 Our meta-analysis showed that both boys and girls with CP scored below these minimal recommended thresholds associated with positive health. Taken together, these findings highlight the importance to identify and monitor children with increased cardiovascular disease risk, and to utilize all opportunities to improve their aerobic fitness at the start of childhood.6,7,9 Even small increases in cardiorespiratory fitness are associated with considerably lower adverse cardiovascular event rates.6

Physical activity is essential for the growth, development, well-being, and socialization of every child, especially children with disabilities.4,10,74 It is not obvious for every child to pursue a healthy lifestyle and avoid sedentary behavior. Several barriers to being physically active for children with a disability are identified.74–77 Better sporting facilities for children with a disability, and more awareness is needed to keep children and parents informed about available possibilities. To reach this goal, co-creation, teamwork, and intersectoral collaboration remains required between the child, parents, health care professionals (e.g. pediatric physicians and pediatric physical therapists), schools, sport coaches, and municipality.4,10,78

To the best of our knowledge, no previous systematic review and meta-analysis has been published summarizing the VO2peak in children with CP compared with TD peers. VO2peak values were measured directly by maximal exercise tests instead of estimated from submaximal tests. However, the results should be viewed in the light of the following limitations. In this systematic review we were not able to use an existing, valid risk of bias tool to evaluate the quality of included studies and weighing the level of evidence according to the GRADE (Grading of Recommendations, Assessment, Development and Evaluation) method.79 Considering the main aim of this review, the risk of bias of each study was assessed by means of well-established exercise physiology criteria regarding the unbiased measurement of VO2peak.12,15,67 Based on our quality assessment, some studies reported only data from children who met the VO2peak criteria, while other studies also included children who did not meet the VO2peak criteria in their analyses. In clinical practice, it is quite often difficult for children to comply with the instructions and to reach their maximum exercise level during testing. This may imply that the pooled VO2peak values of children with CP as well as TD children in this review are an underestimation of their maximal aerobic capacity. A difficulty in pooling the data was the large overlap in study populations in 16 of the 36 studies. We carefully analyzed the studies to avoid using duplicate samples. Furthermore, there was a large heterogeneity of included studies, which may be a potential source of bias. For example, the population characteristics of children with CP differed, varying protocols for exercise testing were used, and tests were performed under different conditions (laboratory and field tests).14 The current review is based on aggregated data, i.e. combining the grouped (average) data of primary studies published between 1978 and 2023, and reflects the state of pediatric exercise physiology research over the past 50 years. With univariate subgroup analyses we were able to reduce some of the heterogeneity. However, the available aggregated data did not allow further refinement, i.e. multivariable meta-analysis combining different child characteristics.