Archives de Pédiatrie 27 (2020) 95–103

Available online at

ScienceDirect
www.sciencedirect.com

Research paper

Nutritional status of neurologically impaired children:

Impact on comorbidity
M. Leonard a,b,*, E. Dain a,b, K. Pelc a,c, B. Dan a,c, C. De Laet a,b
a
Université Libre de Bruxelles (ULB), 1050 Brussels, Belgium
b
Service de maladies métaboliques et nutrition, Hôpital des Enfants Reine Fabiola, 1020 Brussels, Belgium
c
Inkendaal Revalidatieziekenhuis, 1602 Sint-Pieters-Leeuw, Belgium

A R T I C L E I N F O A B S T R A C T

Article history: Background and aims: Malnutrition is common in neurologically impaired (NI) children. It is, however, ill-
Received 25 June 2019 defined and under-diagnosed. If not recognized and treated, it increases the burden of comorbidities and
Received in revised form 24 September 2019 affects the quality of life of these children. The aim of this study was to characterize the nutritional status
Accepted 11 November 2019
of a cohort of children followed up at a reference center for cerebral palsy (CP) in Brussels, Belgium, and
Available online 29 November 2019
to investigate possible links with the occurrence of comorbidities.
Material and methods: We conducted a single-center retrospective study including all the children
Keywords:
followed up at the Inter-university Reference Center for Cerebral Palsy ULB-VUB-ULg. The data were
Cerebral palsy
Malnutrition
obtained by reviewing medical files. Anthropometric measurements as well as the etiology of
Pressure ulcer neurological impairment, comorbidities, feeding patterns, and laboratory test results were collected. The
Pneumonia children were assigned a nutritional diagnosis according to the World Health Organization and
Mid-upper arm circumference Waterlow definitions.
Vitamin D deficiency Results: A total of 260 children with cerebral palsy were included, 148 males and 112 females. Their mean
age was 10.9  4.3 years. The gross motor function classification system (GMFCS) level was I for 79 children,
II for 63 children, III for 35 children, IV for 33 children, and V for 50 children. Of the children, 54% had a normal
nutritional status, 34% showed malnutrition, and 8% were obese; 38% had oropharyngeal dysphagia. The
sensitivity of mean upper arm circumference of < p10 to detect severe malnutrition was 95%. Specific growth
charts for CP were neither sensitive nor specific for predicting the risk of comorbidities. Malnutrition was
associated with an increased risk of comorbidities (relative risk of 2.4 [1.7; 3.4]). It was also associated with
the occurrence of pneumonia, pressure ulcers, and pathological bone fracture.
Discussion and conclusion: Children who are NI should be systematically and thoroughly screened for
malnutrition, in the hope of offering early nutritional support and reduce comorbidities.
C 2019 French Society of Pediatrics. Published by Elsevier Masson SAS. All rights reserved.

1. Introduction and morbidity of these children. Malnutrition leads not only to

stunting but also to muscle wasting and thus to a decreased ability
Neurological impairment comprise a heterogeneous group of to participate in everyday life or leisure activities. A decrease in
disorders that primarily relate to the central nervous system, axial muscle tone affects food intake and adds to the risk of choking
composed of the brain and spinal cord, affecting an individual’s and aspiration pneumonia. Moreover, the alteration of skin
speech, motor skills, vision, memory, muscle actions, and learning trophism along with sedentariness and macro- and micronutrient
abilities. Nutrition is a crucial component of health in neurologi- deficiencies promote poor healing and pressure ulcers [2–4].
cally impaired (NI) children. Nutritional evaluation and manage- Malnutrition in NI children is multifactorial. It is related in part
ment of NI children are a constant challenge for health-care to feeding difficulties associated with axial and face muscle
providers. However, malnutrition often remains under-diagnosed weakness; it can be worsened by oropharyngeal dysphagia, or
[1]. Malnutrition and obesity impact negatively the quality of life gastrointestinal disorders such as gastrointestinal reflux disease
(GERD) and constipation, which are more prevalent in this
population. Limited speech or intellectual disability can impair
* Corresponding author at: Service de maladies métaboliques et nutrition,
communication concerning hunger, thirst, or satiety, and thus lead
Hôpital des Enfants Reine Fabiola, 1020 Brussels, Belgium. to inadequate food or liquid intake. Moreover, various medications
E-mail address: marie.leonard.be@gmail.com (M. Leonard). such as antiepileptic drugs can modify appetite and induce

https://doi.org/10.1016/j.arcped.2019.11.003
0929-693X/ C 2019 French Society of Pediatrics. Published by Elsevier Masson SAS. All rights reserved.
96 M. Leonard et al. / Archives de Pédiatrie 27 (2020) 95–103

osteopenia through enzymatic induction of cytochrome CYP450, and the expected weight for a defined height, was expressed as a
which enhances vitamin D metabolism [2,5,6]. percentage.
The prevalence of malnutrition in NI children is high; it is The weight, height, and BMI were also plotted on the specific
estimated to be 46–90% [2]. Caloric intake in this population is charts for children with cerebral palsy (CP), according to their
often low, which is justified in some cases by decreased energy GMFCS level [10]. They were expressed in percentile ranges.
expenditure due to lack of mobility. Yet low caloric intake can lead Weight was classified as out of or in the red zone, which is below p5
to malnutrition [4,7]. for GMFCS I and II and below p20 for GMFCS III to V. This red zone
Adapted strategies are required to evaluate the nutritional was associated with an increased risk of comorbidities in the
status of NI children. Measuring their weight and height using cohort from which the charts were built [11].
common measuring tools is often challenging. A weighting chair The MUACs were plotted on the Centers for Disease Control and
and a mobile measuring rod to measure the child in a supine Prevention (CDC) charts [12], because although WHO data were
position are often necessary. The patient may also be weighed collected for children up to 19 years [13], charts were not available
while being carried in the arms of an adult. Height measurement for children over 5[14]. The values are expressed in percentile
may be skewed by musculoskeletal deformities, in which case it ranges, and were considered too low if they were < p10, as
can be estimated from segmental measurements (tibial length, recommended by the guidelines of the European Society for
knee height) using appropriate equations [6]. However, this Pediatric Gastroenterology Hepatology and Nutrition (ESPGHAN)
evaluation is somewhat imprecise, particularly for body mass [15].
index (BMI) calculation. Mean upper arm circumference (MUAC)
and skinfold measurements allow for additional evaluation with 2.3.2. Definitions of nutritional status
data concerning body composition since they quantify the muscle On the basis of their anthropometric data, a nutritional
and subcutaneous tissues [8,9]. Furthermore, laboratory analyses diagnosis was attributed to the children. Malnutrition was defined
complement the evaluation assessing micronutrient status [2,8]. in our cohort using a combination of BMI z-score and Waterlow
In order to evaluate the approach of nutritional assessment in index [16].
NI children and the comorbidities associated with malnutrition, we Acute malnutrition was defined as a BMI z-score of < 2
performed a retrospective study of the cohort of children followed standard deviation (SD), and/or a WFH of < 90%. It was considered
up at the Inter-university Reference Center for Cerebral Palsy ULB- moderate if BMI was between 2 and 3 SD and/or WFH was
VUB-ULg (Centre Interuniversitaire de Référence pour l’Infirmité between 80 and 89%, and severe if BMI was < 3 SD and/or
motrice Cérébrale ULB-VUB-ULg [CIRICU]) in Brussels, Belgium. WFH < 80%. In the case of weight loss or stagnation during the
Our aim was to characterize the nutritional status in order to previous 6 months, the subject was also considered acutely
improve the follow-up scheme of these children. malnourished.
Chronic malnutrition was defined as a height z-score
of < 2 SD for ambulant children (GMFCS I–III). However, we
2. Material and methods considered that it was not correct to apply this definition to
wheelchair-bound children (GMFCS IV and V) as they have a lower
2.1. Participants growth [17]. We therefore used a lower threshold of < 3 SD for
children in GMFCS levels IV and V. The malnutrition was defined as
We included all the NI children followed up at the CIRICU of the moderate between 2 and 3 SD for GMFCS I–III and between 3
Queen Fabiola Children’s University Hospital (HUDERF) in Brussels, and 4 SD for GMFCS IV and V, and as severe below 3 SD for
Belgium on June 30, 2017. GMFCS I–III and below 4 SD for GMFCS IV and V.
Obesity was defined using WFH for children under 5 years of
2.2. Data collection age according to the WHO guidelines, with a threshold of 120%, and
using the BMI z-score for children over 5 years of age with a
We reviewed the medical files of all patients actively enrolled in threshold of > 2 SD.
CIRICU on June 30, 2017 and collected data regarding age, sex, last If the anthropometric data were incomplete for a child but the
anthropometric assessment (weight, height, growth curves, MUAC, available data were within the normal range, we considered that
tibial length, knee height), anamnestic data (etiology of neurologi- the child was neither malnourished nor obese, and had a normal
cal impairment, comorbidities, gross motor function classification nutritional status.
system [GMFCS] level, food inquiry including type of diet and
difficulties with meals), laboratory test data (complete blood 2.4. Endpoints
count, iron status, serum protein level, serum albumin level,
vitamin D level, and zinc level). We also noted the outcome of The primary endpoints were the description of the nutritional
eventual orthopedic surgical interventions in terms of complica- status and feeding pattern of this single-center cohort of NI
tions (wound infection, respiratory infections, urinary infections, children, and to establish a correlation between the nutritional
and pressure ulcers). status and the occurrence of comorbidities.
The secondary endpoints were to evaluate the utility of GMFCS-
2.3. Data processing specific growth charts and of the MUAC measurement in
nutritional evaluations. We also looked for a correlation between
Data were anonymized. macro- and micronutrient deficiencies and the occurrence of
If height measurement was unavailable for a child but comorbidities.
segmental measurements were recorded, we estimated the height Finally, we evaluated the role of nutritional status in post-
using the equations published by Scarpato et al. [6]. orthopedic surgery outcome.

2.3.1. Units 2.5. Statistical analysis

Weight, height, and BMI were expressed as z-scores according
to the World Health Organization (WHO) charts. The weight-for- Statistical analysis was performed using the Statistica1
height (WFH), i.e., the ratio between the current weight of a child software (version 13.2, Tulsa, Oklahoma). We used
 M. Leonard et al. / Archives de Pédiatrie 27 (2020) 95–103 97

2  2 contingency tables and performed Chi2 tests when the were obese; one of these also had chronic malnutrition. A normal
expected values were > 5, or unilateral Fisher or Chi2 tests with nutritional status was found for 140 children. A nutritional
Yates correction if one of the expected values was < 5. We also diagnosis could not be attributed to 10 children: seven with no
used linear regressions and mean comparisons via unequal anthropometric data recorded; two with a very low height
variances t test. The normality of distribution was tested using a ( 4.1 SD in a child with Bohring–Opitz syndrome associated with
Shapiro–Wilk test. The significance level of P was set at 0.05 for epileptic encephalopathy but high BMI z-score and high WFH; and
each of these analyses. Subgroup analyses were performed in order 7.0 SD for a child with holoprosencephaly but normal BMI z-score
to address potential biases of positive findings as needed. and normal WFH) that was considered part of their syndrome; one
with a weight of 3.4 SD and no record of height.
The prevalence of malnutrition was significantly lower in the
3. Results
GMFCS I group compared with the GMFCS II–V groups and
significantly higher in GMFCS V compared with GMFCS I–IV. Severe
3.1. Data obtained malnutrition was also significantly more prevalent in the GMFCS V
group (Fig. 2).
A total of 260 children were included; 148 were males and
112 females, aged 18 months to 18 years. Their mean age was 3.2.2. Prevalence of oropharyngeal dysphagia
10.9  4.3 years. Weight was recorded for 251 children, height for Oropharyngeal dysphagia, defined as difficulty with chewing or
209. Height was estimated based on measurements of segments for swallowing, was present in 98 patients (38%). In the GMFCS I
an additional 20 patients. BMI was calculated for 218 children and group, 76% reported having a normal diet, while 22% reported
WFH for 206. The last anthropometric data dated from the previous having some difficulties with swallowing or prolonged meal
12 months for 165 children (63%), between 1 and 5 years for duration. Only two of them (3%) needed modification of the diet
72 children (28%), between 5 and 10 years for 14 children (5%), and texture (mixed, soft, finely cut). The prevalence of oropharyngeal
more than 10 years in two children (1%). No anthropometric data dysphagia increased according to the GMFCS level. Only children
were found for seven subjects (3%). with GMFCS IV–V levels required enteral nutrition: 6% and 40% of
For 53 children (20%), there was a specific nutrition follow-up in them, respectively (Fig. 3).
addition to CIRICU follow-up. Choking episodes were reported by parents for 28% of the
children. They were more prevalent in the GMFCS V group (60%),
3.2. Cohort characteristics and were more often described as ‘‘frequent episodes’’ than in the
other GMFCS levels (Fig. 4). It should be noted that among the
The etiology of the neurological impairment is detailed in 24 children with frequent episodes of choking, five still reported
Table 1. having a normal diet without modification of texture.
The severity of motor disability was categorized as GMFCS I for In total, 73% of the children with obesity were free from
79 children, GMFCS II for 63 children, GMFCS III for 35 children, symptoms of oropharyngeal dysphagia, 18% had some difficulty
GMFCS IV for 33 children, and GMFCS V for 50 children. swallowing or prolonged meal duration, and 9% required a
We confirmed that non-ambulant children (GMFCS IV and V) modified-texture diet; 13% also reported occasional choking.
were significantly shorter with a mean height of 2.3  1.8 SD These proportions are not different from the rest of the cohort,
compared with walking children (GMFCS I–III), whose height was taking into account the degree of motor disability.
0.7  1.5 SD (normal distribution, unequal variances t test,
P < 10 4). 3.3. Comorbidities

3.2.1. Prevalence of malnutrition There was a documented history of pneumonia in 52 children

In all, 55 children had acute malnutrition (28 moderate, (20%), among whom nine (3.5%) required (invasive or noninvasive)
25 severe, two unclassified), 47 children had chronic malnutrition respiratory support in a pediatric intensive care unit. Furthermore,
(29 moderate, 18 severe) (Fig. 1); 13 were malnourished both 39 children (15%) had GERD and 40 (15%) showed constipation.
acutely and chronically (mixed malnutrition). A total of 22 children Three patients (1.2%) presented with pressure ulcers, and four
(1.5%) with pathological bone fractures.

Table 1 3.3.1. Factors influencing the overall rate of comorbidities

Etiology of neurological impairment (n = 260). The overall rate of comorbidities increased according to GMFCS
level (Fig. 5).
Etiology Number of children
Malnutrition (acute and chronic) was also associated with an
Preterm birth, n (%) 88 (34)
increased risk of comorbidities with and relative risk (RR) of
Prenatal stroke, n (%) 47 (18)
Genetic or malformative disease, n (%) 48 (18) 2.4 [95% confidence interval (95% CI): 1.7; 3.4] (Fig. 6). It was even
Postnatal stroke, n (%) 10 (4) more increased in cases of severe malnutrition (RR = 3 [95 CI: 2.1;
Congenital heart disease 6 4.3]). The analysis of the GMFCS subgroups yielded similar results
Sickle cell disease 1 for the GMFCS I and III levels (Fisher test, P < 0.05). The analysis of
Myocarditis 1
Cardiac arrest 1
subgroups according to the type of malnutrition (acute or chronic)
Systemic lupus erythematosus 1 yielded similar results (x2 test, P < 10 4). We did not observe a
Perinatal asphyxia, n (%) 30 (12) significant difference in the occurrence of comorbidities between
Other, n (%) 37 (14) children with obesity and those with normal nutritional status.
Shaken baby syndrome 4
Epileptic encephalopathy 4
Congenital infection 5 3.3.2. Factors influencing the rate of pneumonia
Kernicterus 2 The occurrence of pneumonia was strongly correlated with
Meningitis/encephalitis 6 frequent choking episodes, independently of GMFCS level, with an
Sequelae of brain cancer 1 RR of 5.1 [95% CI: 3.1; 8.3] (Fig. 7). Occasional choking did not
Unknown 15
significantly increase the risk of pneumonia.
98 M. Leonard et al. / Archives de Pédiatrie 27 (2020) 95–103

Fig. 1. Distribution of nutritional status (n = 260).

Fig. 2. Distribution of nutritional status according to GMFCS level (n = 260). Prevalence of malnutrition is significantly lower in GMFCS I compared with the four other GMFCS
groups and is significantly higher in GMFCS V compared with the four other GMFCS groups (x2 tests, P < 0.05). Prevalence of severe malnutrition is significantly higher in
GMFCS V compared with GMFCS I–IV groups (x2 tests, P < 0.05). GMFCS: Gross motor function classification system.

Malnutrition was also a risk factor for pneumonia (RR = 3 [95% to the type of malnutrition (significant difference in severe
CI: 1.6; 5.4] for moderate malnutrition; RR = 3.8 [95% CI: 2.1; 7.0] malnutrition, acute malnutrition, and mixed malnutrition) with
for severe malnutrition with x2 tests, P < 0.05), even after Fisher tests, P < 0.05.
neutralizing the confounding factor ‘‘choking’’ via a subgroup
analysis (Fig. 7). In this subgroup analysis, malnutrition was a 3.3.4. Factors influencing the rate of pathologic bone fractures
risk factor for pneumonia in children having no choking episodes Pathological bone fractures occurred significantly more often in
or occasional choking, but not those with frequent choking the case of mixed malnutrition (Fisher test, P = 0.015).
episodes.
In total, 14 patients with missing data regarding choking were 3.4. Comorbidities occurring after orthopedic surgery
excluded from these analyses. They had significantly more
episodes of pneumonia than the rest of the cohort (x2 test, Of the children, 74 had had orthopedic surgery; 11 of them
P = 0.003). underwent nutrition follow-up and four had enteral nutritional
support. Eight patients (10.8%) developed postoperative compli-
3.3.3. Factors influencing the rate of pressure ulcers cations: pressure ulcers (n = 2), pneumonia (n = 4), urinary tract
The occurrence of pressure ulcers was significantly correlated infection (n = 1), and complex regional pain syndrome (n = 1). The
with malnutrition in the full cohort, and in subgroups according rate of complications was significantly higher after a bone surgery
 M. Leonard et al. / Archives de Pédiatrie 27 (2020) 95–103 99

Fig. 3. Distribution of oropharyngeal dysphagia according to GMFCS level (n = 260). Proportion of absence of oropharyngeal dysphagia (normal diet) is significantly lower in
GMFCS V group compared with the four other GMFCS levels (x2 tests, P < 10 4). Proportion of need for a modified-texture diet is significantly higher in GMFCS III group
compared with GMFCS I (Fisher test, P = 0.001) and GMFCS II groups (Fisher test, P = 0.040). GMFCS: Gross motor function classification system.

Fig. 4. Prevalence of choking episodes according to GMFCS level (n = 260). Choking and frequent choking are significantly more prevalent in GMFCS V group compared with
GMFCS I–IV groups (x2 tests, P < 10 4). GMFCS: Gross motor function classification system.

Fig. 5. Prevalence of comorbidities according to GMFCS level (n = 260). GERD: gastrointestinal reflux disease; PICU: pediatric intensive care unit; GMFCS: Gross motor
function classification system. Rate of comorbidities is significantly higher in GMFCS IV and V groups compared with GMFCS I–III groups (x2 tests, P < 0.5).

(arthrodesis or osteotomy) than after a soft tissue surgery (tendon 3.5. Utilization of the red zone of the specific charts for CP
transposition or lengthening): 25% vs. 2.9% (Fisher test, P = 0.027).
There was, however, no significant difference in the rate The data of 246 children were plotted on the GMFCS-
of complications according to nutritional status or nutrition specific growth charts. Among them, 21 were in the red zone
follow-up. for weight.
100 M. Leonard et al. / Archives de Pédiatrie 27 (2020) 95–103

Fig. 6. Prevalence of comorbidities according to nutritional status (n = 250). Normal: encompasses normal nutritional status and obesity; MM: moderate malnutrition; SM:
severe malnutrition. Rate of comorbidities is significantly higher in moderate malnutrition compared with normal nutritional status, and significantly higher in severe
malnutrition compared with moderate malnutrition (x2 tests, P < 0.05). NB: children with unknown nutritional status were excluded from this analysis.

Fig. 7. Prevalence of pneumonia according to nutritional status and choking (n = 246). Rate of pneumonia is significantly higher in cases of frequent choking episodes (x2 tests,
P < 0.05 in the full cohort and in the GMFCS V subgroup). Rate of pneumonia is significantly higher in malnourished patients, if they had no choking episode or occasional
choking (x2 test, P = 0.030 and Fisher test, P = 0.006), but not if they had frequent choking episodes. NB: children with missing data regarding choking were excluded from this
analysis; GMFCS: Gross motor function classification system.

In comparison with our definitions of malnutrition using WHO associated with malnutrition, including isolated acute and isolated
and Waterlow criteria, 17 of these 21 patients in the red zone were chronic malnutrition (n = 82, x2 tests, P < 0.05). The sensitivity of
considered severely malnourished, three had moderate malnutri- an MUAC of < p10 was particularly good in detecting severe
tion, and one had a normal nutritional status (the aforementioned malnutrition. For overall severe malnutrition, the sensitivity was
child with syndromic short stature). However, 68 malnourished 95% and specificity 75%; it was 100% and 72% concerning severe
children, including 21 who were severely malnourished, were acute malnutrition, respectively; and 90% and 64% for severe
above this red zone of the GMFCS growth charts. chronic malnutrition, respectively.
Of the 21 children in the red zone, 16 (76%) had comorbidities,
including the three children with pressure ulcers. Of the 3.7. Macro- and micronutrient deficiencies
68 malnourished children not classified in the red zone, 33
(48%) also had comorbidities, including four children admitted to a Iron deficiency was found in 43 of the 114 children who were
pediatric intensive care unit because of pneumonia and one child tested (38%). Anemia was found in 24 of 171 patients for whom a total
with pathological bone fracture. blood count was performed (14%), 83% of whom had iron-deficiency
anemia. We could not find a significant correlation in this cohort
3.6. Utilization of the MUAC between iron status and type of diet, nutritional status or pressure
ulcers. Anemia was not correlated with pressure ulcers either.
In total, 35 of 82 patients whose MUAC was measured had an Total serum protein and serum albumin were identified in
MUAC of < p10. Having an MUAC of < p10 was significantly 129 children. Nine of them had hypoproteinemia (low protein
 M. Leonard et al. / Archives de Pédiatrie 27 (2020) 95–103 101

Fig. 8. Vitamin D status according to nutritional status (n = 260). MM: moderate malnutrition; SM: severe malnutrition. Rate of vitamin D level  30 ng/mL is significantly
higher in the severe malnutrition group compared with the rest of the cohort (x2 test, P = 0.006). Rate of vitamin D deficiency (< 20 ng/mL) significantly higher in the obese
group compared to the rest of the cohort (Fisher test, P = 0.0006). Rate of vitamin D monitoring significantly higher in the severely malnourished children compared to the rest
of the cohort (Chi2 test, P < 10 4).

and/or low albumin). Malnourished children had a higher risk of index somewhat inappropriate in some children; the alternative
hypoproteinemia (RR 9.1 [95% CI: 1.2; 70.5], Fisher test, P = 0.010). index is the WFH [15,19].
Hypoalbuminemia was a major risk factor for the occurrence of It was difficult to establish cut-offs in our cohort, resulting in a
pressure ulcers with an RR of 41 [95% CI: 4.3; 391.5] (Fisher test, composite definition of acute malnutrition using the WHO and
P = 0.005). Waterlow definitions (with BMI z-score and WFH percentages)
Vitamin D was monitored in 94 patients. The level was and a dynamic component that was the absence of weight gain
< 20 ng/mL (< 50 mmol/L) in 32 children (34%), between 20 and during the previous 6 months. We arbitrarily decided to use a
30 ng/mL (50–75 mmol/L) in 21 children (22%) and  30 ng/dL different threshold of height z-score to define chronic malnutri-
( 75 mmol/L) in 41 children (44%). Vitamin D level was well tion in non-ambulant children (GMFCS IV and V) since their
correlated (but inversely) with the nutritional status (Fig. 8). In 64% growth potential is intrinsically altered by their pathology. We
of the severely malnourished patients had an optimal vitamin D were reassured with this choice because they were overall
level ( 30 ng/mL), while none of the obese children had an significantly shorter than the rest of the children, with a mean
optimal level. On the contrary, 89% of the obese children had height under 2 SD. This is also supported by the specific growth
vitamin D deficiency (< 20 ng/mL), which is an RR of 3.1 [95% CI: charts for children with CP, in which median height drops
2.1; 4.7] to be vitamin D-deficient if obese in our cohort. Vitamin D according to the GMFCS level [11].
was monitored in 76% of the severely malnourished children Because of the challenges in measuring anthropometric
compared with 30% for the rest of the cohort, which is a statistical parameters in NI children, the nutritional evaluation must be
difference. Vitamin D deficiency was not associated with the multimodal. Commonly used anthropometric data, such as BMI,
occurrence of pathological bone fracture. Out of the four patients WFH, and height-for-age, are the basis of the evaluation.
who had a pathological bone fracture, two had a vitamin D level According to our results, the red zone for weight depicted in the
of > 30 ng/mL, one of < 20 ng/mL, and one was unknown. Three of specific charts for children with CP is a criterion of the severity of
these children were on antiepileptic drugs. malnutrition, but is not sufficient to predict the risk of
Zinc levels were monitored in 40 children and 52% of them had comorbidities.
a zinc deficiency. No correlation was found with the nutritional The MUAC measurement provides additional information on
status. No increase in comorbidities, in particular pressure ulcers, body composition. It quantifies bone, muscle, and the adipose
was noted in the children with zinc deficiency in this cohort. panicle (representative of the energetic balance). Its interpretation
with standards calculated on a reference population is biased in NI
children, whose body composition is altered owing to their
4. Discussion pathology (muscle wasting and osteopenia, partly as a conse-
quence of lack of mobility). The MUAC was measured in a third of
4.1. How to define malnutrition in NI children our cohort, and analyzed with CDC standards making our results
subject to caution. However, it seems to be a good screening tool
Malnutrition is difficult to define in all groups of people, since it for malnutrition, with a very high sensitivity regarding severe
is a dynamic condition, implying intake that is inadequate to allow malnutrition, and it is easily measured with a simple tape, even on
for homeostasis and physical activity, and in children, growth children with reduced mobility. WHO charts for MUAC are now
[18]. Anthropometric data provide the most practical way to available for children until the age of 19 and should also be
approach the concept of malnutrition in clinical practice. evaluated in this setting [13].
Given the specific phenotype of NI children, there is no clear Skinfold measurements were not recorded in our study because
definition of malnutrition in this population, as highlighted in the of the scarcity of available data. They are, however, regarded as a
guidelines from ESPGHAN [15]. ESPGHAN and its North American better measure of the adipose panicle and provide the best
equivalent NASPGHAN suggest using BMI to track acute malnutri- estimation of body composition [20]. They require access to an
tion, although imprecision in height measurements makes this adipometer, and appropriate training for accurate measurements.
102 M. Leonard et al. / Archives de Pédiatrie 27 (2020) 95–103

4.2. Prevalence of malnutrition feeding and sometimes unbalanced macro- and micronutrient
contents. They are not harmless. Iron deficiency is of course a risk
The prevalence of malnutrition in our cohort is comparable to factor for anemia.
that described in previous reports, both in terms of undernutrition We found an increased rate of hypoproteinemia in malnour-
and obesity [2,21]. Malnutrition is highly prevalent, with more ished children and a large increase in the risk of pressure ulcers in
than a third of children featuring some kind of malnutrition, those presenting with hypoalbuminemia (which actually reflects
enhancing the need for systematic nutritional screening in NI severe malnutrition).
children. We did not observe a significant relationship between vitamin D
status and pathological bone fracture, but we did not evaluate the
4.3. Prevalence and consequences of oropharyngeal dysphagia calcium intake of our subjects. Interestingly, vitamin D deficiency
was monitored much more often in malnourished children, but was
The rate of dysphagia in NI children has been reported to be significantly more prevalent in obese children. One explanation is
85%, if objectively evaluated during a meal with a score attributed that malnourished children are more spontaneously given supple-
by a speech therapist trained in swallowing/chewing evaluation mentation with vitamin D. As for obese children, several studies
[22]. If evaluated based on parents’ report alone, as is the case in have shown a correlation between obesity and vitamin D deficiency
our study, the rate of dysphagia drops to about 26%, according to in children without disabilities [27,28]. The pathophysiology
Lopes et al. [23] or by 50% compared with a speech therapist behind this link has not been very well explained so far.
observation, according to Benfer et al. [24]. Thus, when evaluating NI children for malnutrition, macro- and
The texture of the food consumed by the children in our cohort micronutrients have to be monitored, especially in patients with
is also similar to that described in the cohort of Benfer et al., in feeding difficulties and in malnourished children. Vitamin D
which 81% of the GMFCS I and II children had a normal diet, versus should probably be systematically supplemented in most NI
22% for the GMFCS IV and V [24]. children, regardless of their nutritional status, since they are a
Given the strong correlation between swallowing difficulties, group at risk for vitamin D deficiency owing to their lack of
especially choking, and the occurrence of pneumonia, it seems mobility and sun exposure [29]. Appropriate calcium intake should
important to systematically investigate the quality of swallowing be achieved by dairy consumption or supplementation.
in NI children through their history. Those who report difficulties
with food or pneumonia should also be objectively evaluated in 5. Conclusion
order to implement strategies that could make the meal times safer
(modified textures of food, good installation for eating, start of NI children are at high risk of malnutrition. This leads to
enteral feeding). comorbidities and an altered quality of life, which can be
prevented by efficient screening for the risk factors of malnutrition,
4.4. Link between malnutrition and comorbidities for malnutrition itself, and by targeted and multidisciplinary care.
More studies are needed to define biometric thresholds tailored to
We cannot draw conclusions on the rate of comorbidities in our NI children.
cohort, since the study design is based on a file review. However, In the setting of a multidisciplinary follow-up of NI children, a
malnutrition in NI children appears to be per se a risk factor of nutritionist physician should be involved from the start, before
comorbidities, in particular pneumonia and pressure ulcers, even malnutrition occurs and becomes a problem for the child. NI
when taking into account confounding factors such as severity children should be regularly measured and weighed, despite the
of handicap or choking. It is a strong argument in support of practical difficulties, in order to screen the development toward
systematic screening for malnutrition in all NI children. malnutrition or obesity. MUAC should also be monitored as an
Comorbidities of the gastrointestinal tract (GERD and consti- additional screening tool, since it is easy and quick to measure, or
pation) are highly prevalent. Previous studies report a rate twice or skinfold thickness if adequate tools are available. If needed, the
three times superior to ours, for the reasons explained above evaluation of swallowing should be performed by a trained speech
[2]. Even if these comorbidities do not seem to be influenced by therapist. A dietary evaluation should be obtained, in the form of a
nutritional status, they can impede nutritional care and cause standardized questionnaire, in order to avoid unbalanced macro-
significant discomfort. They should therefore be sought for and and micronutrient intake, which occurs much more frequently if
treated. the child has difficulties with meal times.
We could not establish a relation between nutritional status When a major surgery such as orthopedic surgery has to be
and postsurgical comorbidities. This is probably due to a relatively scheduled, the child should be reassessed as part of a nutrition
small cohort of children having had orthopedic surgery and to consultation in order to provide nutritional support when needed,
putative missing data. Such complications of surgery have and hopefully reduce the rate of postoperative complications.
previously been reported in several observational studies in adults
[25,26]. The scheduling of orthopedic surgery in a NI child, Ethics
especially if it is for a major surgery, should be combined with a
nutritional evaluation if not done previously, and the nutritional Approval from the Ethics Committee of HUDERF was obtained
status of the child should be restored if necessary prior to surgery. for this study. Since it was a retrospective survey, no informed
More prospective studies are needed in children to confirm the consent was required according to Belgian legislation.
benefits of such therapeutic options.
Disclosure of interest

4.5. Macro- and micronutrient deficiencies The authors declare that they have no competing interest.

Macro- and micronutrient deficiencies are abundant in our Acknowledgments

cohort when searched for, as previously reported [2], even though
prevalence cannot be calculated owing to numerous missing data. We want to thank Caroline De Longueville for her precious
These deficiencies occur in the setting of feeding difficulties and computer support and other advice, and the CIRICU team for their
reduced energy needs in some children, with reduced volume of collaboration.
 M. Leonard et al. / Archives de Pédiatrie 27 (2020) 95–103 103

References [15] Romano C, van Wynckel M, Hulst J, et al. European Society for Paediatric
Gastroenterology, Hepatology and Nutrition guidelines for the evaluation and
[1] Dan B. Nutrition, brain function, and plasticity in cerebral palsy. Dev Med Child treatment of gastrointestinal and nutritional complications in children with
Neurol 2016;58:890. neurological impairment. J Pediatr Gastroenterol Nutr 2017;65:242–64.
[2] Penagini F, Mameli C, Fabiano V, et al. Dietary intakes and nutritional issues in [16] Mehta N, Corkins M, Lyman B, et al. Defining pediatric malnutrition: a
neurologically impaired children. Nutrients 2015;7:9400–15. paradigm shift toward etiology-related definitions. J Parenter Enteral Nutr
[3] Samson-Fang L, Stevenson R. Identification of malnutrition in children with 2013;37:460–81.
cerebral palsy: poor performance of weight-for-height centiles. Dev Med Child [17] Grossberg R, Wolff J. Defining optimal growth in cerebral palsy: comparing
Neurol 2000;42:162–8. apples to apples? Dev Med Child Neurol 2016;58:892–3.
[4] Kilpinen-Loisa P, Pihko H, Vesander U, et al. Insufficient energy and nutrient [18] Cederholm T, Barazzoni R, Austin P, et al. ESPEN guidelines on definitions and
intake in children with motor disability. Acta Paediatr 2009;98:1329–33. terminology of clinical nutrition. Clin Nutr 2017;36:49–64.
[5] Quitadamo P, Thapar N, Staiano A, et al. Gastrointestinal and nutritional [19] Marchand V, Motil K. NASPGHAN committee on nutrition. Nutrition support
problems in neurologically impaired children. Eur J Paediatr Neurol for neurologically impaired children: a clinical report of the North American
2016;20:810–5. Society for Pediatric Gastroenterology, Hepatology, and Nutrition. J Pediatr
[6] Scarpato E, Staiano A, Molteni M, et al. Nutritional assessment and interven- Gastroenterol Nutr 2006;43:123–35.
tion in children with cerebral palsy: a practical approach. Int J Food Sci Nutr [20] Gurka M, Kuperminc M, Busby M, et al. Assessment and correction of skinfold
2017;68:763–70. thickness equations in estimating body fat in children with cerebral palsy. Dev
[7] Azcue M, Zello G, Levy L, et al. Energy expenditure and body composition in Med Child Neurol 2010;52:e35–41.
children with spastic quadriplegic cerebral palsy. J Pediatr 1996;129:870–6. [21] Pascoe J, Thomason P, Graham H, et al. Body mass index in ambulatory children
[8] Samson-Fang L, Bell K. Assessment of growth and nutrition in children with with cerebral palsy: a cohort study. J Paediatr Child Health 2016;52:417–21.
cerebral palsy. Eur J Clin Nutr 2013;67:S5–8. [22] Benfer K, Weir K, Bell K, et al. Oropharyngeal dysphagia and gross motor skills
[9] Kuperminc M, Gurka M, Bennis J, et al. Anthropometric measures: poor in children with cerebral palsy. Pediatrics 2013;131:e1553–62.
predictors of body fat in children with moderate to severe cerebral palsy. [23] Lopes P, Amancio O, Araújo R, et al. Food pattern and nutritional status of
Dev Med Child Neurol 2010;52:824–30. children with cerebral palsy. Rev Paul Pediatr 2013;31:344–9.
[10] Growth charts for children with cerebral palsy. Life expectancy; 2011. Avail- [24] Benfer K, Weir K, Bell K, et al. Food and fluid texture consumption in a
able from: http://www.lifeexpectancy.org/articles/NewGrowthCharts.shtml population-based cohort of preschool children with cerebral palsy: relation-
[Accessed 9th August 2017]. ship to dietary intake. Dev Med Child Neurol 2015;57:1056–63.
[11] Brooks J, Day S, Shavelle R, et al. Low weight, morbidity, and mortality in [25] Legg J, Davies E, Raich A, et al. Surgical correction of scoliosis in children with
children with cerebral palsy: new clinical growth charts. Pediatrics spastic quadriplegia: benefits, adverse effects, and patient selection. Evid-
2011;128:e299–307. Based Spine-Care J 2014;5:38–51.
[12] Fryar C, Gu Q, Ogden C. Anthropometric reference data for children and adults: [26] Jevsevar D, Karlin L. The relationship between preoperative nutritional status
United States, 2007–2010. Vital Health Stat 11 2012;252:1–48. and complications after an operation for scoliosis in patients who have
[13] Mramba L, Ngari M, Mwangome M, et al. A growth reference for mid upper arm cerebral palsy. J Bone Joint Surg Am 1993;75:880–4.
circumference for age among school age children and adolescents, and vali- [27] Peterson C, Belenchia A. Vitamin D deficiency & childhood obesity: a tale of
dation for mortality: growth curve construction and longitudinal cohort study. two epidemics. Mo Med 2014;111:49–53.
BMJ 2017;358:j3423. [28] Smotkin-Tangorra M, Purushothaman R, Gupta A, et al. Prevalence of vitamin
[14] The WHO Child Growth Standards. World Health Organization; 2006/2007. D insufficiency in obese children and adolescents. J Pediatr Endocrinol Metab
Available from: http://www.who.int/childgrowth/standards/en/ for children 2007;20:817–23.
up to 5 years, and https://www.who.int/growthref/en/ for children 5–19 years [29] Braegger C, Campoy C, Colomb V, et al. Vitamin D in the healthy European
[Accessed 9th August 2017]. paediatric population. J Pediatr Gastroenterol Nutr 2013;56:692–701.