Low Weight, Morbidity, and Mortality in Children With Cerebral Palsy: New Clinical Growth Charts

Page created by Sharon Walker
 
CONTINUE READING
Low Weight, Morbidity, and Mortality in Children With Cerebral Palsy: New Clinical Growth Charts
ARTICLES

Low Weight, Morbidity, and Mortality in Children With
Cerebral Palsy: New Clinical Growth Charts
AUTHORS: Jordan Brooks, MPH,a,b Steven Day, PhD,a                        WHAT’S KNOWN ON THIS SUBJECT: Weight-for-age percentiles of
Robert Shavelle, PhD,a and David Strauss, PhD,a                          children with cerebral palsy are lower than in the general
aLife Expectancy Project, San Francisco, California; and                 population. This is especially true in children with more severe
bDepartment of Biostatistics, University of California, Berkeley,        motor dysfunction. Poor growth, loosely defined, is associated
Berkeley, California
                                                                         with increased hospitalization and school absences.
KEY WORDS
growth charts, cerebral palsy, mortality, morbidity                      WHAT THIS STUDY ADDS: This article reports evidence-based
ABBREVIATIONS                                                            thresholds for low weight and provides estimates of associated
CDC—Centers for Disease Control and Prevention                           increases in mortality risk. These estimates are illustrated on
GMFCS—Gross Motor Function Classification System
CDER—Client Development Evaluation Report
                                                                         new clinical growth charts for children with cerebral palsy,
                                                                         stratified according to gender and Gross Motor Function
www.pediatrics.org/cgi/doi/10.1542/peds.2010-2801
                                                                         Classification System levels.
doi:10.1542/peds.2010-2801
Accepted for publication Apr 7, 2011
Address correspondence to Jordan Brooks, MPH, Life
Expectancy Project, 1439 17th Ave, San Francisco, CA 94122.
E-mail: brooks@lifeexpectancy.org
PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275).
                                                                    abstract
                                                                    OBJECTIVE: To determine the percentiles of weight for age in cerebral
Copyright © 2011 by the American Academy of Pediatrics
                                                                    palsy according to gender and Gross Motor Function Classification
FINANCIAL DISCLOSURE: The authors have indicated they have
no personal financial relationships relevant to this article to      System (GMFCS) level and to identify weights associated with negative
disclose.                                                           health outcomes.
COMPANION PAPER: A companion to this article can be found on        PATIENTS AND METHODS: This study consists of a total of 102 163
page e436 and online at www.pediatrics.org/cgi/doi/10.1542/
peds.2011-1472.                                                     measurements of weight from 25 545 children with cerebral palsy who
                                                                    were clients of the California Department of Developmental Services
                                                                    from 1988 through 2002. Percentiles were estimated using generalized
                                                                    additive models for location, scale, and shape. Numbers of comorbidi-
                                                                    ties were compared using t tests. The effect of low weight on mortality
                                                                    was estimated with proportional hazards regression.
                                                                    RESULTS: Weight-for-age percentiles in children with cerebral palsy
                                                                    varied with gender and GMFCS level. Comorbidities were more com-
                                                                    mon among those with weights below the 20th percentile in GMFCS
                                                                    levels I through IV and level V without feeding tubes (P ⬍ .01). For
                                                                    GMFCS levels I and II, weights below the 5th percentile were associated
                                                                    with a hazard ratio of 2.2 (95% confidence interval: 1.3–3.7). For chil-
                                                                    dren in GMFCS levels III through V, weights below the 20th percentile
                                                                    were associated with a mortality hazard ratio of 1.5 (95% confidence
                                                                    interval: 1.4 –1.7).
                                                                    CONCLUSIONS: Children with cerebral palsy who have very low
                                                                    weights have more major medical conditions and are at increased risk
                                                                    of death. The weight-for-age charts presented here may assist in the
                                                                    early detection of nutritional issues or other health risks in these
                                                                    children. Pediatrics 2011;128:e299–e307

PEDIATRICS Volume 128, Number 2, August 2011                                                                                           e299
                                   Downloaded from pediatrics.aappublications.org by guest on July 24, 2015
Growth charts are standard tools for        tion charts are truly prescriptive or        METHODS
monitoring pediatric growth, develop-       in any sense more useful than the
                                                                                         Inclusion and Exclusion Criteria
ment, and overall health. They contain      descriptive CDC reference curves is
estimated weight-for-age percentiles        an open question.17                          The study population included children
based on a reference population. If a                                                    with cerebral palsy who were clients
                                            Whether such a select sample for
child’s weight falls well outside age                                                    of the California Department of Devel-
                                            cerebral palsy growth curves would
norms, it may raise clinical concern.                                                    opmental Services between January
                                            be helpful is far from clear. Cerebral
The standard charts in pediatric prac-                                                   1988 and December 2002. Clients of the
                                            palsy growth patterns are depen-
tices are those of the Centers for Dis-                                                  Department of Developmental Ser-
                                            dent on the severity of disabilities,4
ease Control and Prevention (CDC) for                                                    vices are assessed annually with the
                                            and children with more severe dis-
boys and girls in the US general popu-                                                   Client Development Evaluation Report
                                            abilities are likely to have signifi-
lation.1 These charts may not be help-                                                   (CDER).18 This report contains over 200
                                            cant comorbidities. Thus, defining a
ful for children with cerebral palsy,                                                    medical, functional, behavioral, and
                                            “healthy” cerebral palsy population
whose growth patterns may be mark-                                                       cognitive items. For each client, a team
                                            becomes a difficult and somewhat
edly different from those of the gen-                                                    headed by a pediatric neurologist
                                            arbitrary task. Perhaps a more rea-
eral pediatric population.2–15                                                           makes medical diagnoses, including
                                            sonable approach to growth-chart
                                                                                         the assessment of cerebral palsy,
Krick et al2 produced the first cerebral     construction is to begin with a clini-
                                                                                         whereas functional items (crawling,
palsy–specific growth charts based on        cally appropriate reference popula-
                                                                                         walking, and feeding, etc) may be
the weight and stature of children with     tion to the construct charts then
                                                                                         assessed by other professionals
severe quadriplegia. The North Ameri-       analyze empirical data to determine
                                                                                         familiar with that aspect of the client’s
can Growth in Cerebral Palsy Research       growth thresholds that are associ-
                                                                                         development.
Collaboration has produced curves for       ated with good or bad health out-
several other growth parameters, in-        comes in that population. This ap-           Children who had a CDER with an Inter-
cluding weight, knee height, upper-         proach was taken by Stevenson                national Classification of Disease,
arm length, mid– upper arm muscle           et al3 and Samson-Fang et al,5 who           Ninth Revision19 code for any of several
area, triceps skinfold, and subscapu-       showed that poor growth, measured            degenerative conditions or condi-
lar skinfold.3 Recently, Day et al4 con-    by a combination of weight and               tions acquired after infancy were
structed a series of height, weight, and    other parameters, was associated             excluded from all analyses. The
BMI charts stratified by motor and           with increased health care use               inclusion-exclusion algorithm is
feeding skills.                             and decreased social-participation           shown in Fig 1.
Some researchers and practitioners          outcomes.
                                                                                         Gross Motor Classification
have raised concerns over the useful-       The following were the goals of the
ness of growth charts as diagnostic or      present study:                               Growth patterns in children with cere-
prognostic tools. One concern is that                                                    bral palsy vary with motor and feeding
                                            1. Estimate reference weight-for-age
existing charts are descriptive refer-                                                   abilities.4 The classification system for
                                               percentiles for children with cere-       motor disability in children with cere-
ences rather than prescriptive stan-           bral palsy at each Gross Motor
dards, showing how a particular group                                                    bral palsy used most commonly in clin-
                                               Function Classification System             ical and research settings is the 5-level
of children grew rather than how a             (GMFCS) level.
particular child should grow. Recently,                                                  GMFCS20:
                                            2. Test for associations between               I. Walks without limitations
the World Health Organization at-
                                               weight for age and morbidity and
tempted to address this concern by                                                        II. Walks with limitations
                                               mortality and quantify those that
constructing growth charts based on                                                      III. Walks using a hand-held mobility
                                               are significant.
a select sample of “healthy children                                                          device
living under conditions likely to favor     3. Construct cerebral palsy growth
                                               charts that clearly illustrate poten-     IV. Self-mobility with limitations, may
achievement of their full genetic
                                               tially unhealthy low weights.                 use powered mobility
growth potential [and whose moth-
ers] engaged in fundamental health-         4. Design the charts to mimic the CDC         V. Transported in a manual wheelchair
promoting practices, namely breast-            charts so that they may easily be         The specific criteria for each level are
feeding and not smoking.”16 Whether            integrated into existing clinical         age dependent and were developed
the resulting World Health Organiza-           practice.                                 with the intent that children would

e300   BROOKS et al
                               Downloaded from pediatrics.aappublications.org by guest on July 24, 2015
ARTICLES

                                                                                               recorded weights that were well
                Persons with CP, 1988–2002             Exclusion conditions: 8034
                          48 447                        Chromosomal anomalies: 818             above or below biologically plausible
                                                       Degenerative conditions: 4295
                                                            Traumatic brain injury,            limits. In addition, some assess-
                                                   motor vehicle accident, near drowning, or   ments suggested extreme rates of
                                                         other acquired injury: 2034
                  No exclusion conditions
                                                                 Autism: 532                   weight change; for example, a 5-year-
                                                                 Cancer: 355
                          40 413                                                               old child gaining 50 pounds during a
                                                             Age 25 y
                                                                                               1-year period. Together, all such
                                                                 10 950                        doubtful observations made up less
                        Age 1–25 y
                         29 463                                                                than 0.1% of our study sample and
                                                       Missing motor function or               were excluded from additional
                                                       not classified by GMFCS
                                                                   199                         consideration.
                    Classified by GMFCS
                            29 264                                                             Gender- and GMFCS-specific reference
                                                  Missing or implausible weight values
                                                                                               percentiles (growth curves) were esti-
                                                                  18                           mated for children with cerebral palsy
                 Weight percentile estimation                                                  who were aged 2 to 20 years (data on
                           29 246                                                              children aged 1 to 25 years were used
                                                          Age < 2 or Age > 20                  to improve the precision of weight per-
                                                                 3701                          centiles at ages 2 and 20). This age
              Morbidity and mortality analysis                                                 range was selected to match the stan-
                          25 545
                                                                                               dard CDC charts. Percentiles were es-
FIGURE 1                                                                                       timated with generalized additive mod-
Study population inclusion-exclusion algorithm. CP indicates cerebral palsy.                   els for location, scale, and shape
                                                                                               (GAMLSS), with a Box-Cox power expo-
                                                                                               nential distribution. This is a semipa-
maintain the same GMFCS level                       study. A relatively small number of        rametric statistical-modeling tech-
throughout childhood and adoles-                    children (⬍1%) were not assigned to        nique that allows estimation of
cence. Wood and Rosenbaum21 docu-                   any GMFCS level because they had           age-specific percentiles and z scores.26
mented the reliability of GMFCS from                missing functional assessments or be-      Models were fit in accordance with
the age of 2 to 12 years to be 0.79.                cause they had rare combinations of        World Health Organization methodol-
For the present study, the age-specific              abilities and disabilities. These chil-    ogy using cubic smoothing splines.
GMFCS criteria were approximated                    dren were excluded from additional         Model selection was based on penal-
with functional items from the CDER                 analysis.                                  ized maximum likelihood.27
based on the classification algorithm
used in Krach et al.25 Functional-item              Weight-for-Age Growth Curves               Morbidity
data have been independently validat-               Weight measures for the CDER were          Separately for each GMFCS level, the
ed22–24 and have interrater reliability             taken directly or, in some cases, re-      mean number of chronic major medi-
exceeding 0.85.24 Because the pres-                 ported by a parent or other caregiver.     cal conditions was calculated within
ence of a feeding tube may affect                   Discrepancies between weights re-          weight-for-age quintiles. According to
growth, GMFCS level V was subdivided                corded on the CDER and those in an         the Department of Developmental Ser-
into children who fed orally without a              individual’s actual medical records        vices, chronic major medical condi-
feeding tube (GMFCS V-NT) and those                 were found in 9% of a random sample,       tions are “major, chronic medical
who had a feeding tube (GMFCS V-TF).                but these were small enough to be ig-      problems that limit or impede the cli-
The vast majority of feeding tubes                  nored as immaterial.22                     ent or significantly impact the provi-
(well over 90%) are gastrostomy                     For approximately one-third of the         sion of service” and “include, but are
tubes. In the United States, nasogas-               assessments, weight values were            not limited to, diabetes mellitus, hyper-
tric feeding is rarely used for ex-                 carried over from a previous CDER.         tension, congenital or arteriosclerotic
tended periods.                                     Because such observations do not           heart disease, upper respiratory infec-
Some children gained or lost abilities              accurately represent age-specific           tions, etc.”18 Differences in the mean
and were represented in 1 or more                   weights, we excluded them from ad-         number of chronic major medical con-
GMFCS levels over the course of the                 ditional analysis. Few individuals had     ditions, for people in the extreme

PEDIATRICS Volume 128, Number 2, August 2011                                                                                       e301
                                Downloaded from pediatrics.aappublications.org by guest on July 24, 2015
TABLE 1 Study Population
                                                                                                                  GMFCS Level
                                                           I                     II                    III                       IV               V-NTa                 V-TFb
No. of assessments                                      14 030                31 808                 13 994                   24 744             11 919                  5668
Male, %                                                     61                    57                     55                       57                 54                    54
Age, median (interquartile range), y                 4.5 (1.7–6.8)         4.5 (2.7–6.4)          4.6 (3.2–6.5)            4.4 (2.6–6.4)      3.9 (2.2–6.3)          4.1 (2.4–6.4)
Has a feeding tube, %                                        1                     2                      5                        9                  0                   100
Orally fed by others, %                                      1                     2                     11                       33                 85                      0
Has severe (IQ 20–34) or profound                           11                    22                     34                       50                 68                    84
  (IQ ⬍ 20) mental retardation, %
Low birth weight (⬍2500 g) or preterm                          25                     28                     35                   31                  23                   22
  labor (⬍37 wk), %
Weight, median (interquartile range), kg            28 (16–50)            27 (19–44)            26 (18–39)                21 (15–32)         18 (13–26)            23 (16–31)
Observations are 102 163 CDERs from 25 545 subjects with cerebral palsy, who received services from the California Department of Developmental Services between 1988 and 2002. Some
children contributed observations to more than 1 group.
a Is fed orally.

b Dependent on a feeding tube.

weight quintiles versus those in the 3                         RESULTS                                                          boys and girls were similar up to about
middle quintiles were assessed with t                                                                                           the age of 15 years. Girls plateaued
                                                               Descriptive Statistics
tests.                                                                                                                          earlier than boys, and between the
                                                               The study population included 25 545                             ages of 15 and 20 years boys tended to
Mortality                                                      children (56% male, aged 2–20 years)                             weigh more than girls. Gender differ-
                                                               who contributed 102 163 weight mea-                              ences were smaller in the more se-
Electronic death records were ob-
                                                               surements (Table 1). Age, gender, pre-                           verely affected groups. For example, at
tained from the California Department
                                                               maturity or low birth weight, and cal-                           age 20 years the difference in median
of Health Services. Individuals surviv-
                                                               endar year of CDER did not vary                                  weights for boys and girls in GMFCS
ing 3 or more years after their last
                                                               significantly by GMFCS level. The most                            level I was 7.3 kg; the difference was
weight measure were censored at 3
                                                               frequent level in our study population                           only 1.8 kg in the GMFCS V-TF group.
years. All individuals surviving to De-
                                                               was GMFCS level II (31%). This was fol-                          Figure 2A shows a scatter plot of weight-
cember 31, 2002, were administra-
                                                               lowed by levels IV (24%), V (17%), I                             for-age data in boys from GMFCS level I,
tively censored at that date.                                  (14%), and III (14%). The proportion                             along with estimated weight-for-age per-
We used Cox proportional hazards re-                           with severe feeding and cognitive dis-                           centiles and the CDC percentiles for boys
gression analysis with time-varying co-                        abilities increased with increasing                              in the general population. The 90th per-
variates28 to relate survival time to                          GMFCS level. For example, 2% of chil-                            centile in GMFCS level I closely tracked
weight percentiles. This enabled us to                         dren in GMFCS level I were either tube                           that of the general population. The me-
control for other variables, such as                           fed or orally fed by others compared                             dian was lower, and the difference in me-
feeding skills, that might confound or                         with 42% of children in GMFCS level IV                           dians increased with age. The 10th per-
modify the effect of low weight on mor-                        and 90% in GFMCS level V. Eleven percent                         centile was markedly lower at all ages.
tality. Separate models were fit for                            of children in GMFCS level I had severe or                       Children in GMFCS level V exhibited more
GMFCS levels I and II and GMFCS levels                         profound mental retardation compared                             linear growth patterns (ie, no growth
III through V because children in these                        with 50% of children in GMFCS level IV                           spurt), with a plateau in late adoles-
groups tend to be different with re-                           and 73% in GFMCS level V.                                        cence (Fig 2B).
spect to functional skills beyond gross
motor function, feeding and cognition,                         Weight-for-Age                                                   Morbidity
age-specific weight values, age-                                In all but the most severe group                                 The mean number of chronic major
specific mortality patterns, and secu-                          (GMFCS level V), weight-for-age data                             medical conditions increased mod-
lar trends. Low-weight cutoffs were se-                        exhibited nonlinear dependence on                                estly with GMFCS level. The most strik-
lected on the basis of maximum                                 age, with a visible growth spurt be-                             ing marker for chronic medical condi-
likelihood. Data were managed in SAS                           tween ages 9 and 13 years and plateau                            tions was the presence of a feeding
version 9.12,29 and analyzed by using R                        in late adolescence. For each GMFCS                              tube. For example, children in GMFCS
version 2.9.30                                                 level, weight-for-age percentiles for                            V-TF had, on average, twice as many

e302       BROOKS et al
                                           Downloaded from pediatrics.aappublications.org by guest on July 24, 2015
ARTICLES

                                                                                                     time. There were 1496 deaths, for an
                                                                                                     overall mortality rate of 9 deaths per
                                                                                                     1000 person-years. For GMFCS levels III
                                                                                                     through V, children with weight for age
                                                                                                     below the 20th percentile had signifi-
                                                                                                     cantly higher mortality rates com-
                                                                                                     pared with children with weight for
                                                                                                     age in the 20th to 80th percentile range
                                                                                                     (P ⬍ .01) (Fig 4). The excess death rate
                                                                                                     in this lowest quintile increased steadily
                                                                                                     with GMFCS level (0.3 per 1000 person-
                                                                                                     years [GMFCS level I] up to 26 per 1000
                                                                                                     person-years [GMFCS V-TF]). Weight
                                                                                                     above the 80th percentile was not asso-
                                                                                                     ciated with differential mortality.
                                                                                                     Because mortality rates in children
FIGURE 2                                                                                             with cerebral palsy vary strongly with
Weight-for-age data and fitted percentiles.                                                           the severity of disabilities, for model-
                                                                                                     ing purposes the data were divided
                                                                                                     into 2 groups: mild to moderate
major medical conditions as those in               conditions than the middle 60% (P ⬍
                                                                                                     (GMFCS levels I and II) and severe
GMFCS V-NT (Fig 3). Among children in              .0001). The mean number of major med-
                                                                                                     (GFMCS levels III through V). Within
the GMFCS levels I through IV and the              ical conditions for children with weights
                                                                                                     each group, we fit unadjusted Cox pro-
level V-NT groups, those with weights              above the 80th percentile was not signif-
                                                                                                     portional hazard regression models and
below the 20th percentile had more                 icantly different from that of children
                                                                                                     also more complex models with baseline
major medical conditions than chil-                with weights in the middle 60%.
                                                                                                     hazard functions stratified by GMFCS
dren whose weights fell in the middle
                                                   Mortality                                         level and adjusted for time-varying cova-
60% (P ⬍ .01). In contrast, children in
                                                                                                     riates, including age, gender, mobility,
GMFCS V-TF who had weights below the               Study participants contributed a total
                                                                                                     feeding, mental retardation, low birth
20th percentile had fewer major medical            of 166 327 person-years of follow-up
                                                                                                     weight or prematurity, and calendar
                                                                                                     year. Unadjusted and adjusted hazard
                                                                                                     ratios from the models are given in Ta-
                                                                                                     bles 2 and 3. For GMFCS levels I and II,
                                                                                                     weight below the 5th percentile was
                                                                                                     associated with an adjusted hazard ra-
                                                                                                     tio of 2.2 (95% confidence interval: 1.3–
                                                                                                     3.7). For GMFCS levels III through V,
                                                                                                     weight below the 20th percentile was
                                                                                                     associated with increased mortality
                                                                                                     (adjusted hazard ratio: 1.5 [95% confi-
                                                                                                     dence interval: 1.4 –1.7]). The relative
                                                                                                     mortality risk associated with low
                                                                                                     weight did not vary with gender, age,
                                                                                                     or calendar year. Sensitivity analyses
                                                                                                     confirmed that the pattern of missing
                                                                                                     age-specific weights were noninforma-
                                                                                                     tive with respect to survival and there-
                                                                                                     fore did not influence these results.
FIGURE 3
Mean number of chronic major medical conditions according to weight quintile. a Significant differ-   These mortality risk research findings
ence from the middle 3 quintiles (P ⬍ .01).                                                          are illustrated on newly developed

PEDIATRICS Volume 128, Number 2, August 2011                                                                                              e303
                               Downloaded from pediatrics.aappublications.org by guest on July 24, 2015
medical conditions for children in the
                                                                                                                               GMFCS V-TF group is unclear. It may be
                                                                                                                               that some very-low-weight children
                                                                                                                               have feeding tubes placed strictly to
                                                                                                                               address weight issues even in the
                                                                                                                               absence of comorbidities, whereas
                                                                                                                               heavier children have feeding tubes to
                                                                                                                               reduce risks from aspiration pneumo-
                                                                                                                               nias or to address other medical is-
                                                                                                                               sues. Additional research is necessary
                                                                                                                               to fully understand this.
                                                                                                                               The concept of failure to thrive is used
                                                                                                                               frequently in general pediatric prac-
                                                                                                                               tice without much evidence regarding
                                                                                                                               its associations with health out-
                                                                                                                               comes.31 It is interesting to note that
FIGURE 4                                                                                                                       our evidence-based GMFCS levels I and
Crude mortality rates according to weight quintile. a Significant difference from the middle 3 quintiles                        II low-weight threshold (ie, the 5th per-
(P ⬍ .05).                                                                                                                     centile) is broadly consistent with an-
TABLE 2 Cox Regression Results for Children in GMFCS Levels I and II                                                           thropometric failure-to-thrive crite-
                                                                               Hazard Ratio for Death and
                                                                                                                               ria.32 This threshold also is consistent
                                                                                95% Confidence Interval                         with studies of the general population
                                                                   Unadjusted                                  Adjusteda       that have found the 10th percentile of
Weight below the 5th percentileb                                   3.2 (1.9–5.3)                             2.2 (1.3–3.7)     adult BMI to be associated with mod-
Based on 45 838 evaluations of 13 118 individuals in GMFCS levels I or II. The cohort experienced 125 deaths over 76 733       estly increased mortality.33,34 That the
person-years of follow-up.
a Adjusted for gender, age, stair climbing ability, mental retardation, feeding, and low birth weight or prematurity.
                                                                                                                               low-weight percentile threshold for
b GMFCS- and age-specific 5th percentile.                                                                                       GMFCS levels I and II is lower than that
                                                                                                                               for GMFCS levels III through V (5th ver-
                                                                                                                               sus 20th percentile) reflects the fact
TABLE 3 Cox Regression Results for Children in GMFCS Levels III Through V
                                                                                                                               that children in GMFCS levels I and II
                                                                               Hazard Ratio for Death and
                                                                                95% Confidence Interval
                                                                                                                               weigh more than those in GMFCS levels
                                                                                                                               III through V.
                                                                   Unadjusted                                  Adjusteda
Weight below the 20th percentileb                                  1.6 (1.4–3.8)                             1.5 (1.4–1.7)     It may seem counterintuitive that
Both models account for functional skills that vary over time (ie, time-varying covariates). The baseline hazard functions     high weights were not associated
were stratified by GMFCS level. Based on 56 325 evaluations of 14 688 individuals in GMFCS levels III through V. The cohort     with increased mortality or morbid-
experienced 1371 deaths over 89 594 person-years of follow-up.
a Adjusted for gender, age, head-lifting ability, feeding, mental retardation, low birth weight or prematurity, and calendar   ity, particularly because obese chil-
year.                                                                                                                          dren may be subject to additional
b GMFCS- and age-specific 20th percentile.
                                                                                                                               comorbidities and may require mod-
                                                                                                                               ified care regimes. The most likely
growth charts with shaded weight-for-                            available at www.lifeexpectancy.org/                          explanation may be that the effects
age values where mortality risk is                               articles/newgrowthcharts.shtml.                               of overweight or obesity do not no-
significantly increased. Fig 5 shows                                                                                            ticeably increase mortality risk until
weight-for-age charts for girls in                               DISCUSSION                                                    adulthood. The impact of childhood
GMFCS level IV and boys in GMFCS level                           Among children in GMFCS levels I                              obesity on adult outcomes in people
V who are tube fed. The new charts are                           through IV, and level V who are not                           with developmental disabilities re-
styled after the standard CDC charts                             tube fed, low weight was, as expected,                        mains an open question.
and include designated areas to re-                              associated with an increase in the                            The proper clinical interpretation of
cord patient name, dates, parental                               number of concurrent chronic major                            the risks discussed here deserves ad-
height and weight, and general notes.                            medical conditions. Why very low                              ditional comment. A practicing clini-
The full set of growth charts is                                 weight is associated with fewer major                         cian may ask, “Do these risks apply to

e304       BROOKS et al
                                              Downloaded from pediatrics.aappublications.org by guest on July 24, 2015
ARTICLES

FIGURE 5
Clinical growth charts for children with cerebral palsy.

my patient? And, if so, for how long?”               The primary limitation of the study is     ment of Developmental Services be-
The evidence presented here is                       the lack of information regarding the      cause of a perceived lack of need.
generally applicable to all children                 etiology of low weight. Low weight is a    Thus, our GMFCS level I findings may be
with cerebral palsy, but additional                  known marker for nutritional deficits       valid only for children with ongoing
patient-specific features should al-                  and general frailty, which is a reason-    needs for services.
ways be considered. One potentially                  able mechanism for increased mor-
                                                                                                The study has a number of strengths. The
benign reason for low weight may                     bidity and mortality. On the other hand,
simply be small parental stature. In                 children may lose weight or have trou-     findings represent the first evidence-
cases where benign etiology has                      ble gaining it as a result of chronic or   based link between low weight and mor-
been ruled out, the excess risks as-                 acute illness. A secondary limitation is   tality risk in children with cerebral palsy.
sociated with low weight should be                   the apparent underrepresentation of        The large sample size allowed percentile
interpreted as persistent for as long                GMFCS level I (14%) in our study popu-     estimates that are robust to modeling
as the child remains in the low-                     lation. In other population-based cere-    assumptions. For example, the charts
weight category. On the other hand,                  bral palsy registries, the proportion of   presented here have estimated weight-
for the reasons stated above clini-                  those in level I ranged from 30% to        for-age percentiles that are consistent
cians should not take the findings of                 40%.35–37 It may be that the most mildly   with those in Day et al.4 The GAMLSS
this study to infer that overweight is               affected individuals in California dis-    growth chart methodology used here is
not a significant health risk in chil-                proportionately choose not to seek         consistent with that of both the CDC
dren with cerebral palsy.                            long-term services from the Depart-        and World Health Organization. It al-

PEDIATRICS Volume 128, Number 2, August 2011                                                                                           e305
                                Downloaded from pediatrics.aappublications.org by guest on July 24, 2015
lows for the calculation of both per-               dence to rely on, clinicians may be                   which to monitor a particular child’s
centiles and z scores, which have be-               forced to make important treat-                       growth. To facilitate integration into
come popular in both the research                   ment decisions on the basis of subjec-                current clinical practice, our growth
and clinical communities. Finally, the              tive impressions. The extent to which                 charts are styled in accordance with
use of a simple and reliable mea-                   today’s clinicians can practice                       those of the CDC and include desig-
sure, weight, may have practical ben-               evidence-based medicine depends                       nated areas to record patient char-
efits over using a more detailed but                 largely on the availability of tools                  acteristics and clinical notes. Ulti-
possibly unreliable combination of                  designed with these principles in                     mately, the utility of the charts will
measures, for example stature or                    mind.                                                 become more apparent as they are
skinfold thickness, in children with                The new cerebral palsy growth                         used in clinical practice.
cerebral palsy.                                     charts presented here are the first to
                                                    give a visual indication of potentially               ACKNOWLEDGMENTS
CONCLUSIONS                                         unhealthy weights. GMFCS is rela-                     Provision of data from the California
Evidence-based decision-making is                   tively stable throughout childhood                    Departments of Developmental Dis-
crucial in clinical and care-planning               and adolescence and thus provides a                   abilities and Health Services is grate-
settings. Without sound empirical evi-              useful stratification scheme from                      fully acknowledged.
REFERENCES
 1. Kuczmarski RJ, Ogden CL, Grummer-Strawn               Feeding dysfunction is associated with poor           growth charts. J Pediatr. 2008;153(5):
    LM, et al. CDC growth charts: United States.          growth and health status in children with             592–594
    Adv Data. 2000;314(314):1–27                          cerebral palsy. J Am Diet Assoc. 2002;          18.   California Department of Developmental
 2. Krick J, Murphy-Miller P, Zeger S, Wright E.          102(3):361–373                                        Services. Client Development Evaluation
    Pattern of growth in children with cerebral     11.   Shapiro BK, Green P, Krick J, Allen D, Capute         Report (CDER). Sacramento, CA: Califor-
    palsy. J Am Diet Assoc. 1996;96(7):680 – 685          AJ. Growth of severely impaired children:             nia Department of Developmental
 3. Stevenson RD, Conaway M, Chumlea WC, et               neurological versus nutritional factors. Dev          Services; 1986
    al. Growth and health in children with                Med Child Neurol. 1986;28(6):729 –733           19.   World Health Organization. International
    moderate-to-severe cerebral palsy. Pediat-      12.   Stallings VA, Charney EB, Davies JC, Cronk            Statistical Classification of Diseases, Inju-
    rics. 2006;118(3):1010 –1018                          CE. Nutritional status and growth of chil-            ries and Causes of Death, Ninth Revision.
 4. Day SM, Strauss DJ, Vachon PJ, Rosenbloom             dren with diplegic or hemiplegic cerebral             Geneva, Switzerland: World Health
    L, Shavelle RM, Wu YW. Growth patterns in a           palsy. Dev Med Child Neurol. 1993;35(11):             Organization; 1977
    population of children and adolescents                997–1006                                        20.   Palisano R, Rosenbaum P, Walter S, Rus-
    with cerebral palsy. Dev Med Child Neurol.      13.   Stallings VA, Charney EB, Davies JC, Cronk            sell D, Wood E, Galuppi B. Development
    2007;49(3):167–171                                    CE. Nutrition-related growth failure of               and reliability of a system to classify
                                                          children with quadriplegic cerebral                   gross motor function in children with ce-
 5. Samson-Fang L, Fung E, Stallings VA, et al.
                                                          palsy. Dev Med Child Neurol. 1993;35(2):              rebral palsy. Dev Med Child Neurol. 1997;
    Relationship of nutritional status to health
                                                          126 –138                                              39(4):214 –223
    and societal participation in children with
    cerebral palsy. J Pediatr. 2002;141(5):         14.   Zainah SH, Ong LC, Sofiah A, Poh BK, Hussain     21.   Wood EP, Rosenbaum PL. The Gross Motor
    637– 643                                              IH. Determinants of linear growth in Malay-           Function Classification System for cerebral
                                                          sian children with cerebral palsy. J Paedi-           palsy. Dev Med Child Neurol. 2000;42(5):
 6. Tobis JS, Saturen P, Larios G, Posniak AO.
                                                          atr Child Health. 2001;37(4):376 –381                 292–296
    Study of growth patterns in cerebral palsy.
    Arch Phys Med Rehabil. 1961;42:475– 481         15.   Shim ML, Moshang T Jr, Oppenheim WL,            22.   Citygate Associates, LLC. Independent Eval-
                                                          Cohen P. Is treatment with growth hor-                uation of the Department of Developmental
 7. Krick J, Van Duyn MA. The relationship be-
                                                          mone effective in children with cerebral              Services’ Community Placement Practices:
    tween oral-motor involvement and growth:
                                                          palsy? Dev Med Child Neurol. 2004;46:                 Final Technical Report. Sacramento, CA:
    a pilot study in a pediatric population with          569 –571                                              Citygate Associates, LLC; 1998
    cerebral palsy. J Am Diet Assoc. 1984;84(5):
                                                    16.   de Onis M, Onyango A, Borghi E, et al, eds.     23.   Arias M, Ito E, Takagi N. Concurrent valid-
    555–559
                                                          WHO Child Growth Standards: Length/                   ity of the client development and evalua-
 8. Stevenson RD, Roberts CD, Vogtle L. The ef-           Height-for-Age, Weight-for-Age, Weight-for-           tion report. In: Silverstein AB, ed. Pacific
    fects of non-nutritional factors on growth in         Length, Weight-for-Height and Body Mass               State Archives VIII. Pomona, CA: University
    cerebral palsy. Dev Med Child Neurol. 1995;           Index-for-Age: Methods and Development.               of California at Los Angeles, Developmen-
    37(2):124 –130                                        Geneva, Switzerland: World Health Organi-             tal Disabilities Immersion Program; 1983:
 9. Motion S, Northstone K, Emond A, Stucke S,            zation Press; 2006                                    28 –33
    Golding J. Early feeding problems in chil-      17.    Mei Z, Ogden CL, Flegal KM, Grummer-           24.   Harris CW, Eyman RK, Mayeda T. An Interra-
    dren with cerebral palsy: weight and neuro-           Strawn L. Comparison of the prevalence of             ter Reliability Study of the Client Develop-
    developmental outcomes. Dev Med Child                 shortness, underweight, and overweight                ment Evaluation Report: Final Report to the
    Neurol. 2002;44(1):40 – 43                            among US children aged 0 to 59 months by              California Department of Developmental
10. Fung EB, Samson-Fang L, Stallings VA, et al.          using the CDC 2000 and the WHO 2006                   Services. UCLA Mental Retardation Re-

e306     BROOKS et al
                                      Downloaded from pediatrics.aappublications.org by guest on July 24, 2015
ARTICLES

      search Center. Pomona, CA: Lanterman                   London, United Kingdom: Chapman and Hall;             in a prospective cohort of U.S. adults. N Engl
      State Hospital; 1983                                   1984                                                  J Med. 1999;341(15):1097–1105
25.   Krach LE, Kriel RL, Day SM, Strauss DJ.          29.   SAS/STAT Software [computer program].           34.   Fontaine KR, Redden DT, Wang C, Westfall AO,
      Survival of individuals with cerebral palsy            Release 9.1. Cary, NC: SAS Institute; 2001            Allison DB. Years of life lost due to obesity.
      receiving continuous intrathecal                 30.   R Development Core Team. R: A Language                JAMA. 2003;289(2):187–193
      baclofen treatment: a matched-cohort                   and Environment for Statistical Computing.      35.   Nordmark E, Hagglund G, Lagergren J. Cere-
      study. Dev Med Child Neurol. 2010;52(7):               Vienna, Austria: R Foundation for Statistical         bral palsy in southern Sweden II: gross mo-
      672– 676                                               Computing; 2008                                       tor function and disabilities. Acta Paediatr.
26.   Stasinopoulos DM, Rigby RA. Generalized          31.   Wilcox WD, Nieburg P, Miller DS. Failure to           2001;90(11):1277–1282
      additive models for location scale and                 thrive: a continuing problem of definition.      36.   Howard J, Soo B, Graham HK, et al. Cerebral
      shape (GAMLSS) in R. Stat Softw 2007; 23(7):           Clin Pediatr (Phila). 1989;28(9):391–394              palsy in Victoria: motor types, topography
      1– 46                                            32.   Olsen EM, Petersen J, Skovgaard AM, et al.            and gross motor function. J Paediatr Child
27.   Borghi E, de Onis M, Garza C, et al. Construc-         Failure to thrive: the prevalence and con-            Health. 2005;41(9 –10):479 – 483
      tion of the World Health Organization child            currence of anthropometric criteria in a        37.   Gorter JW, Rosenbaum PL, Hanna SE, et al.
      growth standards: selection of methods for             general infant population. Arch Dis Child.            Limb distribution, motor impairment, and
      attained growth curves. Stat Med. 2006;                2007;92(2):109 –114                                   functional classification of cerebral
      25(2):247–265                                    33.   Calle EE, Thun MJ, Petrelli JM, Rodriguez C,          palsy. Dev Med Child Neurol. 2004;46(7):
28.   Cox DR, Oakes D. Analysis of Survival Data.            Heath CW Jr. Body-mass index and mortality            461– 467

PEDIATRICS Volume 128, Number 2, August 2011                                                                                                               e307
                                   Downloaded from pediatrics.aappublications.org by guest on July 24, 2015
Low Weight, Morbidity, and Mortality in Children With Cerebral Palsy: New
                         Clinical Growth Charts
       Jordan Brooks, Steven Day, Robert Shavelle and David Strauss
     Pediatrics 2011;128;e299; originally published online July 18, 2011;
                       DOI: 10.1542/peds.2010-2801
Updated Information &                including high resolution figures, can be found at:
Services                             http://pediatrics.aappublications.org/content/128/2/e299.full.h
                                     tml
References                           This article cites 28 articles, 3 of which can be accessed free
                                     at:
                                     http://pediatrics.aappublications.org/content/128/2/e299.full.h
                                     tml#ref-list-1
Citations                            This article has been cited by 4 HighWire-hosted articles:
                                     http://pediatrics.aappublications.org/content/128/2/e299.full.h
                                     tml#related-urls
Subspecialty Collections             This article, along with others on similar topics, appears in
                                     the following collection(s):
                                     Neurology
                                     http://pediatrics.aappublications.org/cgi/collection/neurology
                                     _sub
                                     Neurologic Disorders
                                     http://pediatrics.aappublications.org/cgi/collection/neurologic
                                     _disorders_sub
Permissions & Licensing              Information about reproducing this article in parts (figures,
                                     tables) or in its entirety can be found online at:
                                     http://pediatrics.aappublications.org/site/misc/Permissions.xht
                                     ml
Reprints                             Information about ordering reprints can be found online:
                                     http://pediatrics.aappublications.org/site/misc/reprints.xhtml

PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly
publication, it has been published continuously since 1948. PEDIATRICS is owned, published,
and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk
Grove Village, Illinois, 60007. Copyright © 2011 by the American Academy of Pediatrics. All
rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.

               Downloaded from pediatrics.aappublications.org by guest on July 24, 2015
Low Weight, Morbidity, and Mortality in Children With Cerebral Palsy: New
                         Clinical Growth Charts
       Jordan Brooks, Steven Day, Robert Shavelle and David Strauss
     Pediatrics 2011;128;e299; originally published online July 18, 2011;
                       DOI: 10.1542/peds.2010-2801

The online version of this article, along with updated information and services, is
                       located on the World Wide Web at:
          http://pediatrics.aappublications.org/content/128/2/e299.full.html

  PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly
  publication, it has been published continuously since 1948. PEDIATRICS is owned,
  published, and trademarked by the American Academy of Pediatrics, 141 Northwest Point
  Boulevard, Elk Grove Village, Illinois, 60007. Copyright © 2011 by the American Academy
  of Pediatrics. All rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.

              Downloaded from pediatrics.aappublications.org by guest on July 24, 2015
You can also read