Growth failure in children with intractable epilepsy is not due to increased resting energy expenditure

Abstract

The aim of this study was to evaluate the resting energy expenditure (REE) of children with intractable epilepsy (IE) compared with healthy children, and to determine factors that contribute to the pattern of REE. REE, growth status, and body composition were assessed in 25 prepubertal children with IE (15 males, 10 females; mean age 5y 5mo [SD 2y 2mo] range 2–9y) with and without cerebral palsy (CP) and compared with those in 75 healthy children of similar age, sex, and fat free mass (FFM; 43 males, 32 females; mean age 6y 4mo [SD 1y 8mo], range 2–9y). Of the 25 children with IE, 12 had generalized and 13 partial seizures; 10 children had CP (four hemiplegia, one diplegia, and five tetraplegia); 18 were ambulators. REE (kcal/d), determined by indirect calorimetry, was expressed as a percentage of that predicted using Schofield equations. Energy intake from 3-day weighed food records was assessed for children with IE only and expressed as a percentage of estimated energy requirement. Compared with healthy children, children with IE had significantly lower percentage (Student’s t-test, p<0.05) of predicted REE (111 [SD 13] vs 104 [SD 4]), weight z-score, body mass index z-score, and FFM. Using multiple regression, REE adjusted for FFM, fat mass, and sex were significantly lower in children with IE and CP (–110 kcal/d, 95% confidence interval –199 to –21, p=0.016). In children with IE, energy intake was also a statistically significant predictor of REE. CP largely explained the suboptimal growth status and lower REE of children with IE compared with healthy children.

Epilepsy is a heterogeneous, common neurological disorder that affects 3 million people in the USA, half of whom are children. Although most patients with epilepsy have their seizures controlled with medication, 30% become refractory to medical therapy and require multiple treatments.¹ Children with intractable epilepsy (IE) are at a high risk of treatment side effects and continued seizure activity.^2,3 Suboptimal growth and suboptimal nutritional status are common⁴ in children with IE but little is known about the energy balance (i.e. energy intake and energy expenditure). The purpose of this study was to evaluate the resting energy expenditure (REE) of children with IE in comparison with an age-matched healthy children, and to determine factors that contribute to REE in both groups. Understanding energy expenditure and requirements in children with IE may lead to interventions to optimize growth as well as nutritional, and health status.

Method

PARTICIPANTS

The children with IE in this report were part of a prospective randomized trial of ketogenic diet efficacy.⁵ The study was performed at the Children’s Hospital of Philadelphia, PA, USA. The protocol was approved by the hospital’s institutional review board. Informed consent was obtained from a parent or guardian and assent was given by children who were cognitively able.

The cohort consisted of 48 children aged from 1 to 14 years with a history of one or more seizures per 28 days and for whom antiepileptic medications had failed to control seizures. For this cross-sectional study, REE was assessed at the baseline visit before treatment with the ketogenic diet. We restricted analysis to prepubertal children (age 2–9y) to avoid the complication of variable maturity rates through puberty. Only children who remained awake and who did not experience a seizure during the study were included. The final sample available for analysis was 25. A sample of historical and contemporary healthy children whose REE was assessed at the Nutrition and Growth Laboratory over a period that coincided with this study was selected as the comparison group. The comparison group included prepubertal children of similar sex, age range (2–9y), and fat-free mass (FFM) to those of the IE sample. Comparison children and children with IE were evaluated with the same research, growth, and metabolic equipment and protocols.

ANTHROPOMETRIC MEASURES

Weight and height were measured with standard techniques,⁶ using a scale accurate to 0.1kg (Scaletronix, White Plain, NY, USA) and a stadiometer accurate to 0.1cm (Holtain, Crymych, UK); body mass index (BMI) was calculated. Weight, height, and BMI values were compared with National Center for Health Statistics reference standards, and z-scores for weight, adjusted height, and BMI were computed.⁷ Upper arm circumference was measured with a flexible plastic measuring tape (Ross Laboratories, Columbus, OH, USA), and a skinfold calliper (Holtain, Crymych, UK) was used to measure triceps biceps, subscapular, and suprailiac skinfolds on the right side, or least affected side if asymmetric. Total upper arm muscle and areas of fat were calculated⁸ and z-scores for upper arm muscle area and upper arm fat area were computed.⁹ Using age-specific and sex-specific equations, FFM in kilograms and fat mass (FM) in kilograms were calculated from the four skinfolds.^10,11

DIETARY INTAKE ASSESSMENT

For children with IE, dietary intake was assessed using 3-day (including one weekend day) weighed food records. Families were provided with a calibrated food scale (CS-200; Ohaus Corporation, Pinebrook, NJ, USA), accurate to 0.1g, and instructed in detail by a Clinical and Translational Research Center Registered Dietitian to weigh and report everything that the child consumed. Dietary intake data were collected and analyzed with the Nutrition Data System for Research software (version 2005; Nutrition Coordinating Center, University of Minnesota, Minneapolis, MN, USA). Energy intake (kcal/d) was assessed with adjustments for height, body weight, and physical activity levels and expressed as a percentage of the estimated energy requirement (EER) for children with a low active physical activity level.¹²

RESTING ENERGY EXPENDITURE MEASURES

REE (kcal/d) was measured by open-circuit, indirect calorimetry during an admission. Participants fasted for 12 hours before REE measurement and were brought in a wheelchair from the Clinical and Translational Research Center to the Nutrition and Growth Laboratory. A computerized metabolic cart (Sensor medic 2900 Z; Sensor Medics, Yorba Linda, CA, USA) was used to measure REE for at least 60 minutes in a quiet, thermally neutral room. Participants rested in a supine position while under a large, clear, ventilated hood. Expiratory gases were sampled and analyzed every second, and 1-minute averages were recorded. The first 10 minutes of the measurement period were devoted to environmental acclimation of the child and were not used in calculations. Additionally, periods of significant movement or seizure were documented, and if associated with changes in REE they were not used in calculations. The remaining data points were averaged and REE was calculated from oxygen consumption and carbon dioxide production by the equations of Weir.¹³ For this analysis, the participants were restricted to those who were resting and awake during the REE measurement, because of known alteration of REE during sleep (a decrease of about 10%)^14,15 and to match the comparison sample. Only the awake segment of the REE measurement was used. Children with IE who were asleep for the complete record were excluded as were children who had a seizure during the assessment. REE values (kcal/d) are also expressed as percentages of predicted values using Schofield equations¹⁶ for basal metabolic rate, which adjust for stature, weight, age, and sex.

STATISTICAL ANALYSIS

Data analysis occurred in two phases. In the first phase, descriptive statistics were performed, and group differences between children with IE and healthy comparison children, and also between children with IE with and without cerebral palsy (CP), were assessed. All variables were tested for normality and nonparametric tests were applied as appropriate. Differences in growth status, body composition, and REE between children with IE (with and without CP) and healthy comparison children, and differences in disease status and history between children with IE with and without CP were determined using either a Student’s t-test or a Wilcoxon rank sum test, as appropriate, for continuous variables. χ² tests were used to test for differences of categorical variables. In the second phase of analysis, differences in REE (kcal/d) between children with IE with CP, children with IE without CP, and healthy comparison children were assessed by multiple regression analysis after adjusting for body composition (FFM and FM) and sex. Multiple regression models were also used to explore significant predictors of REE in healthy children and in children with IE separately. The error rate was held constant at α=0.05. Pearson’s or Spearman’s rank correlation coefficients, as appropriate, were used to explore relationships between growth status, body composition, energy intake, and REE in the children with IE. A total of 54 correlations were tested for statistical significance. The adjusted hypothesis-wise error rate was calculated to be α=0.007 across all 54 related correlations, using the adjustment of Tukey et al. for multiple, highly related endpoints.^17,18 All analyses were conducted with STATA (version 7.0 or 8.0).

Results

Clinical characteristics of the 25 children with IE (15 males, 10 females; mean age 5y 5mo [SD 2y 2mo], range 2–9y) and of the 75 comparison children (43 males, 32 females; mean age 6y 4mo [SD 1y 8mo], range 2–9y) are presented in Table I. In comparison with children with IE without CP (n=15), those with CP (n=10) did not differ in age at onset of their seizure disorder (1–3y on average), duration of seizures (about 2y), or seizuretype, (evenly split between generalized partial). Each group had a similar history of status epilepticus, abnormal magnetic resonance imaging results, and microcephaly. However, moderate and severe mental retardation^* was more common in children with CP (χ² test, p=0.005) and they were significantly less likely to be ambulators (χ² test, p<0.005). Children with IE had suboptimal energy intake, with 65% of the group below the percentage EER for children with a low active physical activity level.

Table I:

Clinical characteristics of children with intractable epilepsy (IE) with and without cerebral palsy (CP)

Characteristic	All IE	IE with CPa	IE without CP
Number, n	25	10	15
Age, mean (SD) y:mo	5:5 (2:2)	4:11 (2:1)	5:10 (2:4)
Male/Female, n	15/10	4/6	11/4
Age at first seizure, median (range) y	2.0 (0–6.4)	0.8 (0–6.4)	3.0 (0.4–6.0)
Duration of seizures, median (range) y	2.1 (0.5–7.6)	2.4 (1.5–6.8)	1.9 (0.5–7.6)
Nr of seizures per week, median (range)	3.5 (0.2–4116)	8.8 (0.2–196)	3.2 (0.5–4116)
Seizure type, n
Generalized seizures	12	6	6
Partial seizures	13	4	9
History of status epilepticus	11	6	5
Abnormal MRI, n	16	8	8
Mental retardation* (moderate or severe), n	17	10	7b
Microcephaly, n	4	1	3
CP, n	10	10	0c
Ambulating, n	18	4	14b
Energy intake, mean (SD) kcal/d	1321 (477)	1133 (410)	1446 (490)
EER low active, mean (SD) %	94 (28)	90 (28)	97 (30)

^aOf the 10 children with CP, eight had spastic and two hypotonic CP; there were four with hemiplegia, one with diplegia, and five with tetraplegia.

^bSignificant difference between IE with CP and IE without CP (χ² test): p<0.005

^cSignificant difference between IE with CP and IE without CP (χ² test): p<0.001. MRI, magnetic resonance imaging; EER (estimated energy requirement) low active, energy intake (kcal) as percentage of EER for children with a low active physical activity level.

^*UK usage: learning disability.

Growth status, body composition, and REE for all children with IE, children with IE with and without CP, and healthy children are presented in Table II. In comparison with the healthy children, children with IE as a group were younger, had significantly poorer weight and BMI status, lower FFM, and lower REE, expressed as a percentage of values predicted from the Schofield equations¹⁶ (–7% from predicted, Student’s t-test, p<0.05). Both healthy comparison children and children with IE had an REE close to predicted values. Children with IE and CP had particularly poor growth status, low FFM and FM, and lower REE compared with those of healthy children (–13% from predicted values, Student’s t-test, p<0.005). Children with IE without CP had somewhat poorer growth status, lower FFM, and lower REE (–3% from predicted values) than healthy comparison children; however, none of these differences was statistically significant. REE (kcal/d) as a function of FFM (kg) is presented in Figure 1 for the three groups: healthy comparison children, children with IE with CP, and children with IE without CP. Only children with both IE and CP had a significantly lower REE than the healthy children.

Table II:

Growth, nutritional status, and resting energy expenditure (REE) in children with intractable epilepsy (IE) with and without cerebral palsy (CP), and in healthy comparison children

Parameter	All IE	IE with CP	IE without CP	Healthy
Number, n	25	10	15	75
Age, y:mo	5:5 (2:2)a	4:11 (2:1)c	5:10 (2:4)	6:4 (1:8)
Male/Female, n	15/10	4/6	9/6	43/32
WAZ	−0.65 (1.51)a	−1.27 (1.61)e	−0.23 (1.33)	−0.05 (0.88)
HAZ	−0.28 (1.03)	−0.56 (1.22)c	−0.10 (0.88)	0.06 (0.76)
BMIZ	−0.79 (1.75)a	−1.50 (2.02)e	−0.33 (1.43)	−0.09 (1.01)
UAMAZ	−0.04 (1.07)	−0.24 (0.95)	0.10 (1.16)	0.32 (0.88)
UAFAZ	−0.24 (1.04)	−0.55 (0.93)	−0.04 (1.10)	0.15 (1.03)
FFM, kg	15.3 (4.2)b	13.4 (3.2)d	16.5 (4.5)	18.0 (3.7)
FM, kg	2.9 (1.4)	2.2 (1.2)c	3.4 (1.4)f	3.6 (1.9)
REE, Schofield % predicted	104 (14)a	98 (12)d	108 (14)	111 (13)

Results with parentheses are mean (SD). Differences were determined using either Student’s t-test or Wilcoxon rank sum test, as appropriate. There was no statistically significant difference between IE without CP and healthy comparison children.

^ap<0.05

^bp<0.005, significantly lower in all IE compared with healthy children

^cp<0.05

^dp<0.005

^ep<0.005, significant difference between IE with CP and healthy children

^fp<0.05, significant difference between IE with CP and IE without CP. WAZ, weight for age z-score; HAZ, height for age z-score; BMIZ, body mass index, z-score; UAMAZ, upper arm muscle area z-score; UAFAZ, upper arm area z-score; FFM, fat-free mass; FM, fat mass.

Figure 1:

Regression analysis of resting energy expenditure and fat-free mass in children with intractable epilepsy (IE) with and without cerebral palsy (CP).

Differences between the three groups in REE (kcal/d) were further explored using multiple regression models adjusting for FFM, FM, and sex (Table III). The healthy comparison children and female sex were the reference groups in the model. FFM and male sex were positively and significantly associated with REE. After adjustment for these factors, children with IE with CP had a significantly lower REE than healthy children (–110kcal/d, 95% confidence interval −199 to −21, df=99, p=0.016). Together these variables explained 58% of the total variance in REE.

Table III:

Multiple regression models predicting resting energy expenditure

Parameter	Coefficient (kcal/d)	SEM	t	Wald statistic p	R2	p
All (n=100)
Fat-free mass, kg	31.7	4.1	7.80	<0.001	0.58	<0.001
Fat mass, kg	−0.9	6.9	−0.10	0.921
Sexa	73.2	25.8	2.84	0.005
IE with CPb	−110.1	44.9	−2.45	0.016
IE without CPb	−41.7	35.9	−1.09	0.248
Constant	−463.1	64.3	7.20	<0.001
Healthy comparison children (n=75)
Fat-free mass, kg	36.1	4.7	7.63	<0.001	0.57	<0.001
Fat mass, kg	−2.7	9.5	−0.29	0.773
Sexa	89.4	28.6	3.12	0.003
Constant	381.7	72.9	5.24	<0.001
Children with intractable epilepsy (n=25)
Fat-free mass, kg	20.5	7.2	2.83	0.011	0.56	0.001
Fat mass, kg	–9.1	22.0	−0.41	0.685
Sexa	−15.7	51.4	−0317	0.763
CPc	−128.5	52.4	−2.45	0.024
EER low active,d %	2.0	0.8	2.40	0.027
Constant	843.6	166.2	5.02	<0.001

^aFemale is reference group

^bhealthy comparison children are reference group

^cIntractable epilepsy (IE) without cerebral palsy (CP) is reference group

^denergy intake ( ) as percentage of estimated energy requirement (EER) for children with a low active physical activity level. SEM, standard error of the mean; EER low active, energy intake (kcal) as percentage of EER for children with a low active physical activity level.

Regression models were then run separately for the healthy comparison group and the children with IE, to determine significant predictors of REE within each group (Table III). FFM was a significant positive predictor of REE in both groups, whereas FM was not related to REE. In healthy comparison children, males had a significantly higher REE than females; however, sex differences in REE were not evident in children with IE. FFM and male sex explained 57% of the variance in REE in healthy children. For children with IE, those with CP as a comorbidity had a significantly lower REE (−129kcal/d), and energy intake expressed as a percentage of EER was positively and significantly associated with REE (2kcal/d increase with every 1% increase in EER). Together, FFM, energy intake, and having CP as a comorbidity accounted for 56% of the variance in REE in children with IE.

Associations of REE expressed as kcal/d and percentage of predicted values¹⁶ with age, seizure history, growth status, body composition, and energy intake were explored for children with IE. As expected, the absolute calories expended were significantly and positively associated with age, height, weight, and FFM (Pearson’s correlation coefficient r=0.55–0.64, p<0.01). They were also significantly associated with energy intake (kcal/d; Pearson’s correlation coefficient r=0.63, p<0.001). However, REE expressed as percentage predicted (Schofield equations), adjusted for age, sex, height, and weight, was not significantly associated with seizure history, growth status, body composition, or energy intake (data not shown).

Discussion

The cohort of children with IE in this study presented with significantly lower weight, BMI, and FFM than the healthy comparison children. Their poorer growth status suggests chronic insufficient calorie status or negative energy balance, rather than the positive energy balance required for a growth pattern reflecting the child’s genetic potential. This study evaluated REE, nutrition, and growth status of children with IE with and without CP and showed that REE adjusted for FFM was significantly lower in children with IE and CP than in healthy comparison children. Thus, excess REE (seizures, CP, movement disorder) was not the cause of negative energy balance and growth failure in IE.

Total energy expenditure is composed of REE, physical activity energy expenditure, and, in children, energy expenditure for growth. REE is the largest component of total energy expenditure. REE may be expressed as an absolute value in kcal/day, and/or as a percentage of predicted values, such as those provided by the Schofield prediction equations,¹⁶ which adjust for age, sex, height, and weight. Our data suggest that REE adjusted for FFM was significantly lower in children with IE with CP (–110kcal/d) than in healthy children matched for age and sex. A recent report using a physical activity survey found that adolescents with epilepsy are less active than their siblings without epilepsy.¹⁹ The pattern of negative energy balance and growth failure found in some children with IE was not associated with a higher REE as with other chronic childhood diseases, such as cystic fibrosis and sickle cell disease.^20–22 Rather, in IE it is associated with a lower REE accompanied by lower energy intake and possibly also lower physical activity levels. Insufficient energy (food) intake is suspected to be the cause of growth faltering in IE²³ and requires intervention trials to determine whether increased food intake will prevent or improve growth failure. The lower REE in children with IE who have CP as a comorbidity in the present study may be due, at least in part, to their lower FFM (muscle, organ, and brain) and physical activity.

Although children with IE in general had a lower REE than healthy comparison children, those with CP had a particularly reduced REE accompanied by poorer growth, lower FFM, and lower FM. Seizure history, seizure type, and presence of microcephaly did not differ between children with and without CP. However, children with CP were much more likely to be non-ambulatory and to have mental retardation. Adjusted for FFM and sex, the REE of children with IE and CP was 129kcal/day lower than in children with IE without CP. Stallings et al.²³ also observed a lower REE (adjusted for FFM) in children with CP than in healthy comparison children. In their study of 32 children with the most severe form of CP, spastic quadriparesis, but without intractable epilepsy, they found a lower REE than expected and the lowest REE in the group with low fat stores, a recognized metabolic change in response to low energy intake. The severe motor deficits and abnormal oral motor skills resulting from the CP were felt to contribute significantly to poor caloric intake.

In a recent study, Bertoli et al.⁴ investigated 17 children with IE, aged 1 to 16 years, including four with CP, and only 30% ambulating. The authors found that 40% were malnourished (determined as a percentage of ideal body weight) and 24% were wasted (determined as a percentage of ideal body weight for height). Arm circumference and subscapular skinfold were, respectively, 6% and 4% lower than the reference value and the triceps skinfold was 20% higher. Children with IE had lower FFM and FM on the basis of European reference data for growth and adiposity. Using European dietary intake standards, Bertoli et al. concluded that REE was on average 5% lower in the IE group.

It is difficult to compare our findings for body composition directly with those of Bertoli et al.⁴ because of the data available and different reference data used. In our study, CP, not IE, was the major predictor of nutritional status, and only FM was significantly different in the IE without CP group. It is possible that our group was less neurologically impaired than the group studied by Bertoli et al. which may contribute to the differences in findings. In addition, there were differences in how REE was adjusted for FFM; our data were adjusted by regression analysis.²⁴

Our 25 children with IE were a subset of the larger group for which we have recently reported the dietary intake and compared it with the National Health and Nutrition Examination Survey and Dietary Reference Intake.²⁵ We found that caloric intake was significantly reduced; 49% of the children fell below the EER for sedentary children and 70% fell below the EER for children with a low active physical activity level. These differences were not related to ambulatory or CP status.

Frequent seizures in a child can result in poor nutritional and growth status because the child spends more of his or her day post-ictal or in an encephalopathic stupor with a resulting poor appetite or be unable to eat sufficient quantities. In the present study, the children with IE without CP had somewhat poorer growth and nutritional status than healthy comparison group; however, this did not reach statistical significance, possibly as a result of the small sample size.

There are limitations to our study. The small sample size is a limitation and, therefore, our results need to be confirmed on a larger sample of children with IE. The comparison children were healthy volunteers from studies conducted over several years in our laboratory, either preceding or contemporary with this study. Unplanned methodological drift in REE measurement over time or combining samples from different studies could result in differences between groups or could introduce bias. However, REE measurements for the comparison children were conducted in the same environment, using the same standard procedures, and with an ongoing quality assurance system to maintain methodological consistency.

Conclusion

In comparison with healthy children of similar age and sex, children with IE have suboptimal growth status and a significantly lower REE. The lower REE was largely explained by CP comorbidity (lower FFM) and, to a lesser degree, by poor energy intake and presumed low physical activity. These data suggest that growth faltering in children with IE was not due to elevated energy expenditure; rather, it was due to decreased energy intake and may respond to increased caloric intake. Future nutrition and behavioral interventional studies are needed to determine the impact of increased caloric intake on growth status and other health outcomes in children with IE.

List of abbreviations

EER

Estimated energy requirement

FFM

Fat-free mass

FM

Fat mass

IE

Intractable epilepsy

REE

Resting energy expenditure