Support Care Cancer
DOI 10.1007/s00520-016-3198-y
ORIGINAL ARTICLE
Physical activity and health-related quality of life
in pediatric cancer patients following a 4-week inpatient
rehabilitation program
Carsten Müller 1,2 & Konstantin A. Krauth 3 & Joachim Gerß 4 & Dieter Rosenbaum 2
Received: 1 October 2015 / Accepted: 28 March 2016
# Springer-Verlag Berlin Heidelberg 2016
Abstract
Purpose Chronic health conditions and impaired quality of
life are commonly experienced in childhood cancer survivors.
While rehabilitation clinics support patients in coping with the
disease, studies evaluating an inpatient rehabilitation program
on promoting physical activity (PA) and health-related quality
of life (HRQoL) are missing.
Methods A 4-week inpatient rehabilitation program was pro-
spectively evaluated. One hundred fifty patients with leukemia
or lymphoma (N = 86), brain tumors (N = 38), and sarcomas
(N = 26) were enrolled on average 17 months after cessation of
acute medical treatment. PA amount and cadence (indicating
the intensity of walking activity) using the StepWatch™ 3
* Carsten Müller
c.mueller@uni-muenster.de
Konstantin A. Krauth
Krauth@BadOexen.de
Joachim Gerß
joachim.gerss@ukmuenster.de
Dieter Rosenbaum
diro@uni-muenster.de
1
Institute of Sports Science, Work Unit Human Performance and
Training in Sports, University of Münster, Horstmarer Landweg 62b,
48149 Münster, Germany
2
Institute of Experimental Musculoskeletal Medicine, Movement
Analysis Laboratory, University Hospital Münster, Domagkstr. 3,
48149 Münster, Germany
3
Klinik Bad Oexen, Kinderhaus, Oexen 27, 32549 Bad
Oeynhausen, Germany
4
Institute of Biostatistics and Clinical Research, University of
Münster, Schmeddingstr. 56, 48149 Münster, Germany
Activity Monitor and HRQoL global and physical well-being
scores using the KINDL® questionnaire were assessed before,
immediately after, and 6 and 12 months following the program
and analyzed using multiple linear mixed models.
Results Significant effects on PA were only found at 12-month
follow-up for amount and cadence variables (all p < 0.05).
While leukemia and lymphoma patients revealed the highest
PA level throughout the study, rehabilitation effects were more
pronounced for cadence variables in brain tumor and sarcoma
patients. The rehabilitation program had immediate (t = 4.56,
p < 0.001) and sustainable effects on HRQoL global scores (6-
month follow-up, t = 4.08, p < 0.001; 12-month follow-up,
t = 3.13, p < 0.006).
Conclusions Immediate and sustainable increases in HRQoL
indicate that a 4-week rehabilitation program is beneficial for
improving psychosocial well-being, while the significant in-
crease in PA levels could be related to general recovery as
well. The lack of a control group hampers the evaluation of
the rehabilitation program on promoting PA levels in pediatric
cancer patients.
Keywords Physical activity . Quality of life . Rehabilitation .
Pediatric oncology . Childhood cancer . Physical therapy
Introduction
Progress in multimodal treatment and supportive care resulted
in substantially improved 5-year survival rates in pediatric he-
matology and oncology from below 20 % before 1950 to above
80 % since 1995 [1]. Consequently, the growing number of
childhood cancer survivors has raised awareness concerning
negative long-term side effects of treatment and quality of life.
Previous studies demonstrated that pediatric cancer survivors
are a high-risk population very likely to experience chronic

health conditions and physical limitations in later life like
cardiovascular diseases, severe musculoskeletal conditions,
and impaired quality of life [2–5]. More specifically, the
Childhood Cancer Survivor Study revealed that the highest
prevalence of performance limitations was found in pa-
tients with bone sarcoma and brain tumors, being twice as
high compared to patients with leukemia or non-Hodgkin’s
lymphoma [6].
In the general population, regular physical activity (PA) is
associated with a variety of immediate health-promoting out-
comes, and motor skill development during childhood has
long-lasting effects into adulthood [7]. In particular, there is
strong evidence for the effectiveness of regular PA in children
and adolescents to improve cardiorespiratory and muscular
fitness, improve bone health, and improve cardiovascular
and metabolic health biomarkers and a favorable body com-
position [8]. However, survivors of childhood cancer rarely
reach adequate PA levels [9]. Approximately 50 % of survi-
vors do not meet current PA guidelines and every fourth sur-
vivor reports an inactive lifestyle [10]. Furthermore, physical
health has great potential to influence the three primary health-
related quality of life (HRQoL) domains comprising physical,
psychological, and social well-being in survivors [11, 12].
Since significant impairments of motor performance have
been demonstrated in the majority of children and adolescents
at the end of acute cancer treatment regarding strength or
muscular endurance [13], it is not surprising that patients tend
to report lower HRQoL following acute treatment compared
to healthy controls [14–16].
Medical rehabilitation clinics offer a wide range of thera-
peutic measures supporting patients in coping with the disease
and its physical and psychosocial late effects. Inpatient reha-
bilitation is usually offered for 4 weeks to juvenile patients
and their families or to adolescents not accompanied by their
parents. The results from intervention studies on motor per-
formance and quality of life in pediatric cancer patients during
and after treatment are promising [17, 18], but often limited in
significance due to small sample sizes (usually N ≤ 20).
Moreover, PA is commonly assessed using subjective methods
like the Godin Leisure Time Exercise Questionnaire [19, 20],
a brief four-item query suitable for children and adolescents
aged 10–16 years and allowing for the differentiation between
light, moderate, and strenuous daily PA. However, subjective
methods reveal validity and reliability issues when estimating
the amount and intensity of PA in children and adolescents
[21]. Furthermore, studies on the effectiveness of inpatient
rehabilitation programs for pediatric cancer patients aiming
at promoting daily PA and HRQoL are missing.
Therefore, our intention was to evaluate immediate and
long-term results of a 4-week inpatient rehabilitation program
for children and adolescents after cessation of acute cancer
treatment with respect to objectively assessed PA as the primary
outcome and HRQoL as the secondary outcome. Furthermore,
Support Care Cancer
the aim of our study was to evaluate results by the three main
diagnosis groups comprising leukemia and lymphoma pa-
tients, as well as brain tumor and sarcoma patients. Finally,
we evaluated whether changes in PA and HRQoL were
related.
Methods
Study population
Participants for this prospective, uncontrolled single-center
study with 12-month follow-up were recruited between
September 2010 and November 2012. The sample size was
determined by the size of the recruited cohort within the funded
period. No additional sample size calculation was performed.
Patients aged 4–18 years were eligible after completion of
acute treatment for hematopoietic malignancies (leukemia
and lymphoma), brain tumor, or sarcoma. Exclusion criteria
were unwillingness to participate, lack of capacity to consent
due to language problems or cognitive impairments, and the
need of a wheelchair. Two hundred eighty-four consecutive
patients scheduled for inpatient rehabilitation were suitable
for inclusion. One hundred fifty of these patients could finally
be analyzed. The detailed recruitment process is presented in
Fig. 1. The study protocol was generally approved by the local
Ethics Committee but under the condition not to recruit a con-
trol group. Informed consent was obtained from the patients
and their parents or legal guardian.
Intervention and outcome measures
We evaluated a 4-week inpatient rehabilitation program indi-
vidually tailored to the requirements of the patients as well as
in consultation with them after a 1-h consultation with a phys-
iotherapist and according to the medical examination of the
chief pediatrician. The program consisted of individual mea-
sures like physiotherapy comprising land-based and aquatic
exercises and hippo therapy, as well as group measures like
exercise training and sports games (Table 1). PA and HRQoL
were assessed at the same time. The first assessment took part
the week before patients commenced their rehabilitation pro-
gram. For this purpose, patients and parents or legal guardians
were contacted by phone by a nurse from the rehabilitation
clinic and informed about methods and objectives of this
study. Upon consent, the PA monitor and HRQoL question-
naire were sent to the patients no later than 10 days prior to the
start of the rehabilitation program. The second measurement
was performed the day after leaving the rehabilitation clinic
and the third and fourth measurements 6 and 12 months fol-
lowing rehabilitation, respectively. All measurements were
performed in the home environment.
Support Care Cancer

Fig. 1 Flow diagram of patient

recruitment and data analysis

Sixty patients had to be excluded from this study (Fig. 1).
They received their approval for the rehabilitation program
too late to participate in the initial PA assessment before they
started their program. Although these patients took part in the
study, their data had to be excluded due to the various exercise
and social activity opportunities available during the rehabil-
itation program, thus adding substantial bias to the initial
physical activity assessment as the primary outcome.
PA was assessed using the StepWatch™ 3 Activity Monitor
(SAM, modus health llc, 1054 31st Street NW, Suite 318,
Washington, DC 20007, USA). The SAM is an ankle-mounted,
small (70 × 50 × 20 mm), and light-weight (38 g) uniaxial
accelerometer, recording the number of gait cycles for up to
2 months using 1-min recording intervals. It is programmed
according to the height and gait characteristics of each patient
and does not provide feedback to the patient, thus minimizing

Table 1 Patient characteristics

Study sample Leukemia/lymphoma Brain tumor Sarcoma Significance of

(N = 150) (N = 86) (N = 38) (N = 26) group differences
Mean (SD) Mean (SD) Mean (SD) p†

Gender female/male (female percentage) 73/77 (49 %) 35/51 (41 %) 21/17 (55 %) 17/9 (65 %) n.s.‡
Age (years) 10.7 (4.3) 10.1 (4.5) 10.1 (3.5) 13.4 (3.5) 0.001b,c
BMISDS 0.30 (1.22) 0.50 (1.20) 0.06 (1.16) 0 (1.31) n.s.
Time since diagnosis (months) 24.4 (22.1) 20.5 (16.3) 31.8 (25.3) 26.3 (30.4) 0.010a
Duration of acute treatment (months) 7.2 (4.2) 6.9 (3.9) 6.9 (5.5) 8.6 (2.1) 0.011b
Time since cessation of acute treatment (months) 17.2 (21.6) 13.6 (15.3) 25.0 (25.3) 17.7 (30.1) 0.010a,c
Chemotherapy (yes/no) 135/15 (90 %) 85 /1 (99 %) 26/12 (68 %) 24/2 (92 %) <0.001#
Surgery (yes/no) 59/91 (39 %) 0/86 (0 %) 34/4 (90 %) 25/1 (96 %) <0.001#
Radiation therapy (yes/no) 48/102 (32 %) 13/73 (15 %) 27/11 (71 %) 8/18 (31 %) <0.001‡
Previous rehabilitation (yes/no) 32/118 (21 %) 16/70 (19 %) 10/28 (26 %) 6/20 (23 %) n.s. ‡
Individual physiotherapy (number of sessions) 15.1 (14.4) 6.3 (7.9) 26.9 (12.0) 27.2 (13.6) <0.001a,b
Group exercises (sports therapy, number of sessions) 22.4 (10.5) 26.1 (9.7) 17.1 (9.9) 18.0 (8.8) <0.001a,b
Movement therapy sessions (sum) 37.6 (11.1) 32.4 (8.7) 44.0 (9.4) 45.2 (11.5) <0.001a,b
a
Significant difference (patients with leukemia/lymphoma vs. patients with brain tumors), b significant difference (patients with leukemia/lymphoma vs.
patients with sarcomas), c significant difference (patients with brain tumors vs. patients with sarcomas), † Kruskal-Wallis H test and Mann-Whitney U
test, ‡ χ2 test, # Fisher’s exact test

test bias. The monitor has demonstrated high validity and reli-
ability [22]. Patients were instructed to wear the monitor con-
tinuously for 7 days during waking hours. Valid measurements
required a wear time of at least 8 h per day. During waking
hours, non-wear time was defined as at least 60 min of physical
inactivity and subtracted from total daily wear time. Four PA
measurements were evaluated. All analyzed measurements
were performed in the patient’s home environment.
PA was assessed by means of activity amount and cadence.
PA amount was taken as the average number of gait cycles per
day and per hour. Gait cycles per hour were calculated by
dividing daily gait cycles by monitor wear time in order to
minimize effects of different monitor wear times on PA
amount. Activities with cadences above 40 gait cycles per min-
ute were classified as moderate to vigorous PA (MVPA). Peak
cadence was determined as an indicator of the intensity of
walking activity [23]. For this purpose, gait cycles accumulated
during the 10 most active, but not necessarily consecutive,
minutes of a day were analyzed and averaged across all mea-
surement days, indicating the highest cadence. Furthermore,
the most active 30 consecutive minutes were identified and
averaged across all measurement days as an indicator of endur-
ance capacity [23].
HRQoL was assessed using the KINDL® questionnaire, a
generic instrument suitable for healthy as well as (chronically)
ill children and adolescents aged 4 years and older through
parent reports (children aged 4 to 7 years) and self-reports
(≥8 years). Acceptance, psychometric quality, and feasibility
have been previously demonstrated for the KINDL® ques-
tionnaire [24]. Six dimensions of HRQoL are assessed on a
five-point Likert scale, comprising physical and emotional
well-being, self-esteem, family, friends, and school, each
consisting of four items. Based on the six dimensions, a
global HRQoL score can be calculated after transforming
the raw data on a scale ranging from 0 to 100, with higher
scores indicating better HRQoL.
Statistical analyses
Data analyses were performed using IBM SPSS Statistics ver-
sion 22 and SAS version 9.4 for Windows (SAS Institute Inc.,
Cary, NC, USA). Descriptive statistics included means, stan-
dard deviations, and 95 % confidence intervals. Differences in
patient characteristics were analyzed using non-parametric
Kruskal-Wallis H test and Mann-Whitney U test, as well as
χ2 test and Fisher’s exact test. Testing data for normal distri-
bution was performed using Shapiro-Wilk tests. Multiple PA
and HRQoL data were found to be skewed and could not be
converted by using transformation techniques. Multiple linear
mixed models with subject-specific random intercept (vari-
ance component models, corresponding to multilevel models)
were established, including the independent variables diagno-
sis group, tumor surgery, radiation and chemotherapy as acute
Support Care Cancer
treatment modalities, gender, age at baseline, BMISDS, and
duration and time since cessation of acute treatment, as well
as time of assessment and diagnosis group, in order to predict
the outcomes PA and HRQoL levels at baseline and follow-up
assessment. Models were fitted using a restricted maximum
likelihood approach (REML). Model-based mean estimates of
PA and HRQoL with associated standard errors and 95 %
confidence intervals were derived. Mean differences of post-
rehabilitation assessments versus baseline and between the
three diagnosis groups were calculated and evaluated for sig-
nificance using Bonferroni adjustment. Spearman’s rank cor-
relations were calculated to evaluate relationships between
changes in PA and HRQoL. The local significance level was
set at alpha = 0.05. p values were regarded noticeable
(Bsignificant^) in case p ≤ 0.05.
Results
Data sets of 150 patients were evaluated. Detailed patient char-
acteristics are listed in Table 1. Reasons for missing data were
lack of interest (PA N = 8, HRQoL N = 5 measurements),
incomplete data (PA N = 3, HRQoL N = 2), accelerometers
and questionnaires lost in the mail (PA N = 1 and HRQoL
N = 1), and patients being unable to be traced for the last
measurement (N = 4). Finally, relapse occurred in one lympho-
ma patient resulting in missing the 6-month post-rehabilitation
PA and HRQoL assessment, as well as in one brain tumor and
one sarcoma patient resulting in missing 6- and 12-month
post-rehabilitation measurements for PA and HRQoL.
Physical activity
Mean accelerometer wear time was 6.2 to 6.5 days with a
duration of 11.6 to 12.1 h per day. The last PA assessment
revealed a significantly longer daily monitor wear time com-
pared to previous measurements (p < 0.05). Overall, levels of
PA amount were high. Estimated means for gait cycles were
between 6507 and 7216, corresponding to 13,014 to 14,432
steps per day (Fig. 2a), while the average peak 10-min ca-
dence ranged from 54.0 gait cycles per minute at baseline to
56.9 gait cycles per minute at 12-month follow-up (Fig. 2b).
The rehabilitation program elicited no immediate effects on
PA amount or cadence. Significantly higher PA levels were
found at 12-month follow-up for PA volume, including gait
cycles per day (t = 3.76, p < 0.001), gait cycles per hour
(t = 2.56, p = 0.032), and MVPA (t = 3.03, p = 0.008), as well
as for cadence including peak 10-min cadence (t = 4.96,
p < 0.001) and the most active 30 consecutive minutes
(t = 3.83, p < 0.001).
PA levels differed substantially between diagnosis groups.
At baseline, leukemia and lymphoma patients accumulated
significantly more gait cycles per day compared to patients
Support Care Cancer
Fig. 2 PA following the
intervention period. a Average
gait cycles per day. b Average 10-
min peak cadence. Bars represent
mean and standard error. Stars
indicate a statistically significant
change compared with the
baseline assessment after
Bonferroni adjustment
with brain tumors (t = 3.49, p = 0.002) and sarcomas (t = 3.13,
p = 0.003) (Fig. 2a). Similarly, peak 10-min cadence was
significantly higher in leukemia and lymphoma patients com-
pared to patients with brain tumors (t = 4.56, p < 0.001) and
sarcomas (t = 4.93, p < 0.001) (Fig. 2b). Group differences
were apparent throughout the observation period and also at
12-month follow-up.
Estimated changes in PA levels between baseline and
follow-up measurements are presented in Table 2. Compared
with the initial PA assessment, leukemia and lymphoma
patients significantly increased the number of gait cycles
per day at 12-month follow-up. The remaining comparisons
revealed no significant changes. Beneficial effects on PA
levels in patients with brain tumors and sarcomas were sig-
nificant for peak 10-min cadence immediately following
rehabilitation and at 12-month follow-up as well as for the
most active 30 consecutive minutes at 12-month follow-up
(Table 2).
Health-related quality of life
The rehabilitation program resulted in immediate and long-
term beneficial effects on global HRQoL scores. Effects on
physical well-being scores were lower but still significant ex-
cept at 12-month follow-up (Fig. 3). Large differences were
found between the diagnosis groups: sarcoma patients report-
ed the highest HRQoL global scores and physical well-being
throughout the observation period, while the lowest scores
were reported by leukemia and lymphoma patients. Group
differences did not reach significance, except for physical
well-being at baseline between sarcoma and brain tumor
patients (difference 16.6 [95 % CI 6.6–26.6], t = 3.27,
p = 0.004).
Changes in HRQoL global and well-being scores follow-
ing the rehabilitation program were slightly higher for brain
tumor patients compared to leukemia and lymphoma patients.
However, the effect did not reach significance in brain tumor
Table 2 Mixed models estimated changes in physical activity levels between baseline and follow-up measurements

Baseline Post-rehabilitation vs. baseline 6-month follow-up vs. baseline 12-month follow-up vs. baseline

Estimate Mean change (95 % CI) Mean change (95 % CI) Mean change (95 % CI)

PA amount (GC/day)
Total sample 6564 −57 (−383 to 268) t = −0.35, p = 1.0 −17 (−345 to 313) t = −0.10, p = 1.0 652 (311 to 992) t = 3.76, p < 0.01
Leukemia/lymphoma 6784 −308 (−735 to 118) t = −1.42, p = 0.47 −329 (−755 to 97) t = −1.52, p = 0.39 586 (145 to 1026) t = 2.62, p = 0.03
Brain tumor 4025 226 (−410 to 863) t = 0.70, p = 1.0 56 (−587 to 699) t = 0.17, p = 1.0 653 (−10 to 1315) t = 1.94, p = 0.16
Sarcoma 4135 370 (−423 to 1163) t = 0.92, p = 1.0 1089 (246 to 1932) t = 2.54, p = 0.03 891 (71 to 1711) t = 2.14, p = 0.10
PA amount (GC/h)
Total sample 579 −4.3 (−31.8 to 23.2) t = −0.31, p = 1.0 1.8 (−26.1 to 29.7) t = 0.13, p = 1.0 37.6 (8.8 to 66.5) t = 2.56, p = 0.03
Leukemia/lymphoma 596 −24.3 (−60.5 to 11.9) t = −1.32, p = 0.56 −20.7 (−56.9 to 15.5) t = −1.12, p = 0.79 29.9 (−7.6 to 67.3) t = 1.57, p = 0.35
Brain tumor 431 11.0 (−43.0 to 65.1) t = 0.40, p = 1.0 11.3 (−43.4 to 65.9) t = 0.41, p = 1.0 38.4 (−17.9 to 94.8) t = 1.34, p = 0.54
Sarcoma 430 41.5 (−25.9 to 108.9) t = 1.21, p = 0.68 72.8 (1.0 to 144.5) t = 1.99, p = 0.14 64.6 (−5.2 to 134.3) t = 1.82, p = 0.21
MVPA (min)
Total sample 40.5 −0.8 (−4.3 to 2.8) t = −0.43, p = 1.0 −0.6 (−4.2 to 2.9) t = −0.34, p = 1.0 5.6 (2.0 to 9.3) t = 3.03, p < 0.01
Leukemia/lymphoma 42.3 −3.4 (−8.0 to 1.2) t = −1.46, p = 0.44 −3.5 (−8.1 to 1.1) t = −1.50, p = 0.40 5.4 (0.6 to 10.2) t = 2.23, p = 0.08
Brain tumor 20.7 2.1 (−4.8 to 9.0) t = 0.59, p = 1.0 −0.3 (−7.2 to 6.7) t = −0.07, p = 1.0 5.0 (−2.1 to 12.1 t = 1.39, p = 0.50
Sarcoma 23.1 4.1 (−4.5 to 12.7) t = 0.94, p = 1.0 10.5 (1.4 to 19.7) t = 2.27, p = 0.07 8.1 (−0.8 to 17.0) t = 1.79, p = 0.22
Highest cadence (GC/min)
Total sample 54.0 1.0 (−0.1 to 2.0) t = 1.74, p = 0.25 1.2 (0.1 to 2.3) t = 2.20, p = 0.09 2.9 (1.7 to 4.0) t = 4.96, p < 0.01
Leukemia/lymphoma 55.6 −0.8 (−2.2 to 0.6) t = −1.15, p = 0.76 −0.5 (−1.9 to 0.9) t = −0.65, p = 1.0 1.4 (−0.1 to 2.8) t = 1.88, p = 0.18
Brain tumor 43.4 3.4 (1.3 to 5.5) t = 3.20, p < 0.01 2.0 (−0.1 to 4.1) t = 1.85, p = 0.20 3.8 (1.7 to 6.0) t = 3.47, p < 0.01
Sarcoma 41.5 3.3 (0.7 to 5.9) t = 2.49, p = 0.04 6.2 (3.4 to 8.9) t = 4.40, p < 0.01 6.3 (3.6 to 9.0) t = 4.60, p < 0.01
Endurance (GC/min)
Total sample 31.4 0.9 (−0.3 to 2.1) t = 1.54, p = 0.37 1.2 (0 to 2.4) t = 1.96, p = 0.15 2.5 (1.2 to 3.7) t = 3.83, p < 0.01
Leukemia/lymphoma 32.3 0.2 (−1.4 to 1.8) t = 0.26, p = 1.0 0.3 (−1.3 to 1.9) t = 0.40, p = 1.0 1.4 (−0.2 to 3.1) t = 1.73, p = 0.26
Brain tumor 21.6 2.0 (−0.3 to 4.4) t = 1.68, p = 0.28 1.7 (−0.7 to 4.1) t = 1.42, p = 0.47 3.6 (1.1 to 6.1) t = 2.86, p = 0.01
Sarcoma 21.3 1.8 (−1.2 to 4.7) t = 1.18, p = 0.72 3.6 (0.5 to 6.8) t = 2.26, p = 0.07 4.1 (1.0 to 7.2) t = 2.64, p = 0.03

Bonferroni significance level p ≤ 0.05

Support Care Cancer
Support Care Cancer

Fig. 3 HRQoL following the

intervention period. a KINDL®
global score (norm value 76.9
[24]). b KINDL® physical well-
being score (norm value 77.2
[24]). Bars represent mean and
standard error. Stars indicate a
statistically significant change
compared with the baseline
assessment after Bonferroni
adjustment

patients at 12-month follow-up for the global score and at 6-
and 12-month follow-up for the well-being score, while all
comparisons were significant in patients with leukemia and
lymphoma. Generally, sarcoma patients revealed small differ-
ences in the global score between baseline and follow-up mea-
surements. Contrary, an increasingly negative effect was
found for physical well-being that reached significance at
12-month follow-up (t = −2.41, p = 0.05).
Relationship between changes in PA and HRQoL
following rehabilitation
in the remaining KINDL® subscales were also small (ρ = −0.16
to 0.11) and not significant.
The largest correlation coefficients were found in patients
with brain tumors for changes in PA and physical well-being
(ρ = 0.15 to 0.32). These were also positively correlated in
patients with leukemia and lymphoma (ρ = 0.15 to 0.22). In
contrast, changes in PA were negatively associated with
changes in the global score (ρ = −0.20 to −0.35) and physical
well-being (ρ = −0.22 to −0.33) in patients with sarcoma.

In general, changes in PA and HRQoL from baseline to post- Discussion

rehabilitation were not related. Changes in PA amount and
cadence showed negative correlations with changes in the
global score (ρ = −0.09 to −0.01, n.s.) and positive correla-
tions with changes in physical well-being (ρ = 0.06 to 0.16,
n.s.). Correlation coefficients for changes in PA and HRQoL
The aim of the present study was to evaluate a 4-week inpa-
tient rehabilitation program for a mixed group of pediatric
cancer patients with respect to changes in PA and HRQoL.
While immediate and persistent effects on HRQoL were

found following the rehabilitation program, significant in-
creases in PA levels were only found at 12-month follow-up.
Subgroup analyses indicated that brain tumor and sarcoma
patients increased their PA levels particularly with regard to
cadence following rehabilitation. However, it has to be noted
that the lack of a control group is a limiting factor in our
evaluation because we cannot differentiate between effects
that can be attributed to the rehabilitation program or are due
to general recovery.
Several pilot studies indicated the feasibility and benefits of
exercise interventions on outcome in different stages during
and after acute treatment [19, 25]. Given the paucity of research
with respect to objectively assessed PA as a main outcome
measure, comparisons with previous intervention studies re-
main difficult. However, PA amount and intensity were dem-
onstrated to increase following exercise interventions. For ex-
ample, patients with lower extremity sarcoma accumulated a
median of only 517 gait cycles per day 6 weeks after tumor
surgery and improved to 5023 gait cycles 18 months post-
surgery [26]. Similarly, six patients with ALL who accumu-
lated 2770 gait cycles (5539 pedometer steps) on average per
day at baseline also improved over the course of their mainte-
nance therapy following a home-based exercise and nutrition
program [27], supporting previous assumptions suggesting
that supervised hospital trainings are more effective compared
with home-based or community-based programs during after-
care [28].
It should be noted that baseline values for PA amount in our
study were considerably higher due to longer time intervals
since primary diagnosis, particularly in patients with leukemia
and lymphoma who accumulated 6784 gait cycles or 13,568
steps on average per day. Thirteen thousand steps per day
correspond to norm values of healthy children and adolescents
and are associated with 60 min of MVPA [29, 30], currently
being recommended by the World Health Organization for
leading a healthy lifestyle. Compared with previously provid-
ed age- and sex-specific normative values of pedometer-
determined step activity in healthy children and adolescents
[31], 65 % of patients with leukemia and lymphoma in our
study exceeded the reported average values. Furthermore, nor-
mative data for accelerometer-determined cadence patterns
from NHANES (N = 2610) indicate that healthy children
and adolescents spend 24 min on average at cadences above
40 gait cycles [32], an intensity we defined as MVPA. We
found that leukemia and lymphoma patients accumulated
42 min of MVPA at baseline, thus exceeding normative values
by more than 50 %. Obviously, with high initial activity levels,
the detection of intervention effects is virtually impossible.
In contrast, though brain tumor and sarcoma patients
achieved 21 and 23 min of MVPA on average at baseline,
which is comparable to normative values [32], both groups
were able to increase daily MVPA to 26- and 31-min 12-
month post-rehabilitation. Similarly, mean cadence values
Support Care Cancer
increased from baseline to post-rehabilitation measurements,
indicating opportunities for PA interventions in both patient
groups. However, whether increased PA indicates the sustain-
ability of rehabilitation effects or simply results from disease
recovery remains a matter of debate due to the lack of a control
group.
In our sample of pediatric cancer patients, initial HRQoL
(mean 74.3) was lower than norm data of healthy children and
adolescents (mean 76.9) [24]. HRQoL was also reduced in
comparison with a mixed childhood cancer group of 23 pa-
tients aged 4–17 years after cessation of acute treatment and
within 5 years after primary diagnosis as assessed with the
KINDL® questionnaire (mean 79.5) [33], thus being compara-
ble to our study sample. The significant increase immediately
following the rehabilitation program as well as the sustainabil-
ity of effects during follow-up indicates that short-term im-
provements can be attributed to the rehabilitation program
rather than to general recovery from the disease. In this regard,
our results confirm the majority of previous studies evaluating
the effects of clinical exercise interventions on HRQoL in
pediatric oncology [17, 18].
Given their poorer physical functioning in patients with
sarcomas and brain tumors compared to leukemia and lym-
phoma patients, we expected compromised KINDL® global
scores and particularly lower scores in the subscale physical
well-being. This was not confirmed in sarcoma patients, who
scored above norm values and substantially higher for physi-
cal well-being compared to patients with leukemia, lympho-
ma, and brain tumors. This result reflects general issues in
HRQoL assessment in sarcoma patients, who are diagnosed
during adolescence, whereas patients with brain tumors, leu-
kemia, and lymphoma are typically diagnosed during child-
hood. The difference in age and maturity may evoke bias due
to the fact that sarcoma patients are more likely to be aware of
their life-threatening disease, which in turn may induce denial
in order to maintain continuity with the past, generating higher
than expected HRQoL values [34].
Based on previously reported associations between PA and
HRQoL [35, 36], we expected changes in both variables fol-
lowing the rehabilitation program to be related. We found
positive correlations in patients with leukemia, lymphoma,
and brain tumors. However, PA changes in sarcoma patients
were negatively associated with changes in HRQoL global
and physical well-being scores. This contradictory outcome
is likely to result from continuously increased PA levels fol-
lowing the rehabilitation program and concurrent over-
reporting of HRQoL particularly for physical functioning, as
previously described in sarcoma patients [37]. Moreover, ad-
olescents taking part in longitudinal studies may come to a
point where a strategy of denial can no longer be maintained.
Confrontations with late effects and resulting consequences
for the future (e.g., family, career) will certainly yield psycho-
logical adaptations resulting in adapted HRQoL levels.
Support Care Cancer
Limitations comprise the lack of a control group impeding the
differentiation between rehabilitation effects and disease recovery
in longitudinal studies. Furthermore, we used 1-min epochs of
PA recordings, which is in accordance with the methods used in a
study that raised the most comprehensive accelerometer-derived
norm values for peak cadence in healthy children and adolescents
so far. Thus, we were able to compare cadence of the patients
with norm values [32]. However, it has to be noted that using 1-
min recording intervals is associated with accuracy issues in
determining PA intensity in children whose PA is characterized
by intermittent, high-intensity activities in short bouts that may
be misclassified as lower intensity when averaged over longer
epochs of 1 min. Hence, according to more recent guidelines,
recording epochs of 15 s are recommended [38, 39]. With
regard to HRQoL, we relied on single information of children
and adolescents aged 8 years and older, while parents com-
pleted the questionnaires for children aged 4 to 7 years.
However, current opinion is that information is derived from
both patients and parents in order to get a more comprehensive
picture [34]. In contrast, the strengths of the study are related
to the longitudinal design, a large sample size allowing for
comparisons of diagnoses groups, and the implementation of
objective PA assessments. Activity monitoring allows for the
objective assessment of the amount and intensity of physical
activities in daily life. Hence, PA has been well documented
with respect to functional aspects and also proved to be helpful
in the evaluation of the rehabilitation program.
Acknowledgments We thank the participating patients for their
sustained commitment to the study. The help of Andrea Kelter-Klöpping,
Pascal Mailand and Kai Lindkamp in recruiting the participants and data
acquisition is gratefully acknowledged. This study was funded by
Arbeitsgemeinschaft für Krebsbekämpfung (ARGE Krebs NW).
Compliance with ethical standards The study protocol was generally
approved by the local Ethics Committee but under the condition not to
recruit a control group. Informed consent was obtained from the patients
and their parents or legal guardian.
Conflict of interest The authors declare that they have no conflict of
interest.
References
1. Rossig C, Juergens H, Schrappe M, Moericke A, Henze G, von
Stackelberg A, Reinhardt D, Burkhardt B, Woessmann W,
Zimmermann M, Gadner H, Mann G, Schellong G, Mauz-
Koerholz C, Dirksen U, Bielack S, Berthold F, Graf N, Rutkowski
S, Calaminus G, Kaatsch P, Creutzig U (2013) Effective childhood
cancer treatment: the impact of large scale clinical trials in Germany
and Austria. Pediatr Blood Cancer 60:1574–1581
2. Blaauwbroek R, Stant AD, Groenier KH, Kamps WA, Meyboom
B, Postma A (2007) Health-related quality of life and adverse late
effects in adult (very) long-term childhood cancer survivors. Eur J
Cancer 43:122–130
3. Speechley KN, Barrera M, Shaw AK, Morrison HI, Maunsell E
(2006) Health-related quality of life among child and adolescent
survivors of childhood cancer. J Clin Oncol 24:2536–2543
4. KC O, AC M, CA S, Kawashima T, MM H, AT M, DL F, Marina N,
Hobbie W, NS K-L, CL S, Leisenring W, LL R, Childhood Cancer
Survivor Study (2006) Chronic health conditions in adult survivors
of childhood cancer. N Engl J Med 355:1572–1582
5. Hudson MM, Ness KK, Gurney JG, Mulrooney DA, Chemaitilly
W, Krull KR, Green DM, Armstrong GT, Nottage KA, Jones KE,
Sklar CA, Srivastava DK, Robison LL (2013) Clinical ascertain-
ment of health outcomes among adults treated for childhood cancer.
JAMA 309:2371–2381
6. Ness KK, Hudson MM, Ginsberg JP, Nagarajan R, Kaste SC,
Marina N, Whitton J, Robison LL, Gurney JG (2009) Physical
performance limitations in the childhood Cancer survivor study
cohort. J Clin Oncol 27:2382–2389
7. Loprinzi PD, Cardinal BJ, Loprinzi KL, Lee H (2012) Benefits and
environmental determinants of physical activity in children and
adolescents. Obes Facts 5:597–610
8. Evenson KR, Mota J (2011) Progress and future directions on phys-
ical activity research among youth. J Phys Act Health 8:149–151
9. Winter C, Muller C, Hoffmann C, Boos J, Rosenbaum D (2010)
Physical activity and childhood cancer. Pediatr Blood Cancer 54:
501–510
10. Ness KK, Leisenring WM, Huang S, Hudson MM, Gurney JG,
Whelan K, Hobbie WL, Armstrong GT, Robison LL, Oeffinger
KC (2009) Predictors of inactive lifestyle among adult survivors
of childhood cancer: a report from the childhood Cancer survivor
study. Cancer 115:1984–1994
11. McDougall J, Tsonis M (2009) Quality of life in survivors of child-
hood cancer: a systematic review of the literature (2001-2008).
Support Care Cancer 17:1231–1246
12. Zebrack BJ, Chesler MA (2002) Quality of life in childhood cancer
survivors. Psychooncology 11:132–141
13. Gotte M, Kesting SV, Winter CC, Rosenbaum D, Boos J (2015)
Motor performance in children and adolescents with cancer at the
end of acute treatment phase. Eur J Pediatr 174:791–799
14. Mitchell HR, Lu X, Myers RM, Sung L, Balsamo LM, Carroll WL,
Raetz E, Loh ML, Mattano LA, Winick NJ, Devidas M, Hunger SP,
Maloney K, Kadan-Lottick NS (2015) Prospective, longitudinal
assessment of quality of life in children from diagnosis to three
months off treatment for standard risk (SR) acute lymphoblastic
leukemia (ALL): results of children’s oncology group study
AALL0331. Int J Cancer 138(2):332–339
15. Penn A, Lowis SP, Hunt LP, Shortman RI, Stevens MC, McCarter
RL, Curran AL, Sharples PM (2008) Health related quality of life in
the first year after diagnosis in children with brain tumours com-
pared with matched healthy controls; a prospective longitudinal
study. Eur J Cancer 44:1243–1252
16. Hinds PS, Gattuso JS, Billups CA, West NK, Wu J, Rivera C,
Quintana J, Villarroel M, Daw NC (2009) Aggressive treatment
of non-metastatic osteosarcoma improves health-related quality of
life in children and adolescents. Eur J Cancer 45:2007–2014
17. Braam KI, van der Torre P, Takken T, Veening MA, van Dulmen-den
Broeder E, Kaspers GJ (2013) Physical exercise training interven-
tions for children and young adults during and after treatment for
childhood cancer. Cochrane Database Syst Rev 4:CD008796
18. Baumann FT, Bloch W, Beulertz J (2013) Clinical exercise inter-
ventions in pediatric oncology: a systematic review. Pediatr Res
74:366–374
19. Beulertz J, Prokop A, Rustler V, Bloch W, Felsch M, Baumann FT
(2015) Effects of a 6-month, group-based, therapeutic exercise pro-
gram for childhood cancer outpatients on motor performance, level
of activity, and quality of life. Pediatr Blood Cancer 63(1):127–132
20. Wurz A, Chamorro-Vina C, Guilcher GM, Schulte F, Culos-Reed
SN (2014) The feasibility and benefits of a 12-week yoga

intervention for pediatric cancer out-patients. Pediatr Blood
Cancer 61:1828–1834
21. Welk GJ, Corbin CB, Dale D (2000) Measurement issues in the
assessment of physical activity in children. Res Q Exerc Sport
71:S59–S73
22. McDonald CM, Widman L, Abresch RT, Walsh SA, Walsh DD
(2005) Utility of a step activity monitor for the measurement of
daily ambulatory activity in children. Arch Phys Med Rehabil
86:793–801
23. Barreira TV, Katzmarzyk PT, Johnson WD, Tudor-Locke C (2013)
Walking cadence and cardiovascular risk in children and adoles-
cents: NHANES, 2005-2006. Am J Prev Med 45:e27–e34
24. Ravens-Sieberer U, Ellert U, Erhart M (2007) Health-related quality
of life of children and adolescents in Germany. Norm data from the
German Health Interview and Examination Survey (KiGGS).
Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz
50:810–818
25. Speyer E, Herbinet A, Vuillemin A, Briancon S, Chastagner P
(2010) Effect of adapted physical activity sessions in the hospital
on health-related quality of life for children with cancer: a cross-
over randomized trial. Pediatr Blood Cancer 55:1160–1166
26. Winter CC, Muller C, Hardes J, Gosheger G, Boos J, Rosenbaum D
(2013) The effect of individualized exercise interventions during
treatment in pediatric patients with a malignant bone tumor.
Support Care Cancer 21:1629–1636
27. Moyer-Mileur LJ, Ransdell L, Bruggers CS (2009) Fitness of chil-
dren with standard-risk acute lymphoblastic leukemia during main-
tenance therapy: response to a home-based exercise and nutrition
program. J Pediatr Hematol Oncol 31:259–266
28. Huang TT, Ness KK (2011) Exercise interventions in children with
cancer: a review. Int J Pediatr 2011:461512
29. Tudor-Locke C, Craig CL, Beets MW, Belton S, Cardon GM,
Duncan S, Hatano Y, Lubans DR, Olds TS, Raustorp A, Rowe
DA, Spence JC, Tanaka S, Blair SN (2011) How many steps/day
are enough? for children and adolescents. Int J Behav Nutr Phys
Act 8:78–5868–8-78
Support Care Cancer
30. Adams MA, Johnson WD, Tudor-Locke C (2013) Steps/day
translation of the moderate-to-vigorous physical activity guide-
line for children and adolescents. Int J Behav Nutr Phys Act 10:
49–5868–10-49
31. Craig CL, Cameron C, Tudor-Locke C (2013) CANPLAY pedom-
eter normative reference data for 21,271 children and 12,956 ado-
lescents. Med Sci Sports Exerc 45:123–129
32. Barreira TV, Katzmarzyk PT, Johnson WD, Tudor-Locke C (2012)
Cadence patterns and peak cadence in US children and adolescents:
NHANES, 2005-2006. Med Sci Sports Exerc 44:1721–1727
33. Beulertz J, Bloch W, Prokop A, Baumann FT (2013) Specific def-
icit analyses in motor performance and quality of life of pediatric
cancer patients—a cross-sectional pilot study. Pediatr Hematol
Oncol 30:336–347
34. Eiser C (2009) Assessment of health-related quality of life after
bone cancer in young people: easier said than done. Eur J Cancer
45:1744–1747
35. Keats MR, Culos-Reed SN (2008) A community-based physical
activity program for adolescents with cancer (project TREK): pro-
gram feasibility and preliminary findings. J Pediatr Hematol Oncol
30:272–280
36. Paxton RJ, Jones LW, Rosoff PM, Bonner M, Ater JL, Demark-
Wahnefried W (2010) Associations between leisure-time physical
activity and health-related quality of life among adolescent and adult
survivors of childhood cancers. Psychooncology 19:997–1003
37. Badr H, Chandra J, Paxton RJ, Ater JL, Urbauer D, Cruz CS,
Demark-Wahnefried W (2013) Health-related quality of life, life-
style behaviors, and intervention preferences of survivors of child-
hood cancer. J Cancer Surviv 7:523–534
38. Reilly JJ, Penpraze V, Hislop J, Davies G, Grant S, Paton JY (2008)
Objective measurement of physical activity and sedentary behav-
iour: review with new data. Arch Dis Child 93:614–619
39. Gabrys L, Thiel C, Tallner A, Wilms B, Muller C, Kahlert D, Jekauc
D, Frick F, Schulz H, Sprengler O, Hey S, Kobel S, Vogt L (2015)
Accelerometry for measuring physical activity—recommendations
on methods 45:1–9