点击显示 收起
1 From the Department of Gastroenterology (FEA, HS, and EVO), the James Fairfax Institute of Paediatric Nutrition (JRA, KJG, and MAG), the University of Sydney (FEA and SLC), and the Department of Nuclear Medicine (JNB and RBH-G), The Children’s Hospital at Westmead, Westmead, Australia
2 Supported by the National Health and Medical Research Council of Australia, the James Fairfax Institute of Paediatric Nutrition, and Nutricia Australia Pty Ltd. 3 Reprints not available. Address correspondence to EV O’Loughlin, Department of Gastroenterology, The Children’s Hospital at Westmead, Locked Bag 4001, Westmead NSW 2145, Australia. E-mail: tedo{at}chw.edu.au.
ABSTRACT
Background: No studies have directly measured body protein or validated skinfold-thickness anthropometry and dual-energy X-ray absorptiometry (DXA) to assess body protein in children with spastic quadriplegic cerebral palsy (SQCP).
Objective: We aimed to measure and evaluate body protein and to determine whether skinfold-thickness anthropometry and DXA can predict body protein in children with SQCP.
Design: This was a cross-sectional study of 59 children (22 girls, 37 boys) aged 3.9–19.5 y with SQCP. The children underwent measurements of anthropometric indexes, lean tissue mass by DXA (LTMDXA), and total body protein by neutron activation analysis (TBPNAA). In addition, TBP was estimated from both skinfold-thickness anthropometry (TBPSKIN) and DXA (TBPDXA). The agreement of TBPSKIN and TBPDXA was tested against TBPNAA by using Bland and Altman plot analysis.
Results: Height and weight SD scores (
Key Words: Cerebral palsy • children • anthropometry • total body nitrogen • body protein • neutron activation analysis • dual-energy X-ray absorptiometry • lean tissue mass
INTRODUCTION
Numerous studies have described the body composition of children with spastic quadriplegic cerebral palsy (SQCP) (1–8). Most of these studies focused on measurements of body fatness and the validation of various methods to predict body fat stores (6, 9–11). No studies have directly measured body protein in children with SQCP or validated indirect methods of assessing body protein. Body protein is important to measure because immune function, response and recovery to disease, and skeletal respiratory muscle function are severely impaired when body protein levels are low (12, 13).
Body protein content can be assessed by the measurement of total body nitrogen (TBN) with prompt gamma neutron activation analysis (NAA). The advantage of NAA over other techniques for estimating protein is that it is a direct measure of body nitrogen that is not affected by other factors such as hydration status (14). Indirect methods for estimation of protein involve measurement of fat-free mass (FFM) or lean tissue mass (LTM) from skinfold-thickness anthropometry and dual-energy X-ray absorptiometry (DXA).
Skinfold-thickness anthropometry divides body weight into 2 compartments, fat mass and FFM, whereas DXA provides a three-compartment model of body composition that consists of fat, bone-free LTM, and bone mineral content (BMC). Skinfold-thickness anthropometry is a relatively simple, inexpensive, and portable technique that can be easily performed by trained technicians in the clinic and community setting where sophisticated technology such as direct body protein measurement by NAA and DXA may not be readily available. DXA is becoming more widely available in hospital settings and is therefore an increasingly popular method of assessing body composition. DXA is ideal for children, because it is fast and involves minimal exposure to radiation. However, DXA and skinfold-thickness anthropometry do not directly measure body protein and rely on assumed hydration factors of LTM estimated from healthy children. If hydration is abnormal, this can lead to errors in the estimation of LTM (15). Furthermore, anthropometry and DXA have not been validated for estimating body protein in children with SQCP. Hence, the aims of the present study were 1) to directly measure and evaluate the body protein of children with SQCP by NAA, and 2) to determine whether skinfold-thickness anthropometry and DXA can predict body protein in these children.
SUBJECTS AND METHODS
Subjects
Fifty-nine children (22 girls, 37 boys) between the ages of 4 and 19 y with a diagnosis of SQCP were recruited through the Children’s Hospital at Westmead (CHW) between 2000 and 2004. Children were excluded if they had a known metabolic disorder, genetic syndrome, or chromosomal abnormality. All children were wheelchair-bound and were totally dependent on their caregivers for their everyday needs. They had all been seen in a multidisciplinary clinic that assesses developmentally disabled children with complex feeding and nutritional problems. Nineteen of the 59 children with SQCP (5 girls, 14 boys) were fed partially or completely by either a gastrostomy (n = 16) or a nasogastric tube (n = 3) at the time of the study.
The control children were from existing DXA (n = 172) and TBN (n = 72) databases and were of similar age (4-19 y) to the children with SQCP (16–18). The control children were not significantly different from reference data in height or weight as indicated by their SD scores and were therefore considered to be representative of the healthy population.
Protocol
During an outpatient visit to the hospital, the children had measurements of anthropometry, DXA, and TBN. The CHW Ethics Committee approved the study, and written informed consent was obtained from each parent or caregiver.
Anthropometry
Two trained observers performed all measurements by the use of standardized techniques (19). For the children with SQCP, knee height (±0.1 cm) was measured with an anthropometer (Holtain Ltd, Crosswell, United Kingdom) with the knee and ankle bent to a 90° angle. The distance from the heel to the anterior surface of the thigh over the femoral condyles was measured (20). The knee height measurement was used to estimate height on the basis of the equations of Stevenson 1995 (for children aged 12 y) and Chumlea et al 1994 (for children aged >12 y) (20, 21). For the control children, standing height without shoes was measured by using a Harpenden stadiometer (Holtain Ltd). Weight measurements (±0.1 kg) were made with electronic scales (A&D Mercury Pty Ltd, South Australia) with the children wearing minimal clothing. Because the children with SQCP were unable to stand on the scales, the caregiver stood on the scales while holding the child, and then the caregiver was weighed on his or her own, and the difference between the 2 weights was calculated to give the weight of the child. Weight and height measurements were converted to SD scores and were compared with the Centers for Disease Control and Prevention growth reference data (22). Percentage of ideal weight-for-height was calculated for the children with SQCP by first estimating the child’s height-age (the age at which the child’s estimated height would be on the 50th percentile) and then dividing the child’s actual weight by the weight corresponding to the 50th percentile for the child’s height-age. The result was then multiplied by 100 to give a percentage.
Skinfold thicknesses were measured in duplicate on the right side of the body at 4 sites (triceps, biceps, subscapular, and suprailiac) with a Harpenden caliper (British Indicators Ltd, St Albans, Hertfordshire, United Kingdom). An estimate of percentage body fat was calculated from the skinfold thicknesses by using the equations of Brook (23) for prepubertal children and those of Durnin and Rahaman (24) for pubertal children. Fat-free mass (FFMSKIN) was derived by subtracting FM from body mass. Midupper arm circumference was measured with a flexible steel tape at the midpoint between the acromion process and the head of the radius on the right arm. An estimate of upper arm muscle area was derived from the midupper arm circumference and triceps skinfold-thickness measurement. The upper arm muscle area was converted to a SD score derived from the Frisancho reference data (25).
Knee height was not measured in the first 17 subjects enrolled in the study. Therefore, data on height, height SD scores, and percentage of ideal weight-for-height are unavailable for 8 girls and 9 boys with SQCP.
Total body protein from neutron activation analysis
Total body protein was measured by the method of prompt gamma NAA as previously described by this unit (14, 16). The basis of this technique is that the subject is bilaterally irradiated with neutrons from 2 252Cf sources while lying supine. 14N is converted to 15N with the emission of a 10.8-MeV gamma ray that is specific for nitrogen. TBN is calculated by measuring the integral under the nitrogen peak centered on the 10.8 MeV of the gamma ray spectrum. By measuring TBN, total body protein (TBPNAA) can then be determined by using the following relation: mass of protein = 6.25 x mass of nitrogen (14, 17). The total exposure time is 15 min and the effective dose equivalent delivered during a scan is <0.15 mSv (quality factor = 20 for fast neutrons). The technique has a precision and accuracy of 1.4–5.4% and 97–101.5%, respectively, in child-sized phantoms (14). The TBPNAA of the children with SQCP was compared with the TBPNAA predicted from the control data for age, height, and weight.
Eight (1 girl, 7 boys) children with SQCP were excluded from the analysis of TBPNAA predicted for age, height, and weight because their ages fell outside the range for which the predictive equations were developed from control data (girls, 5.0–14.4 y; boys, 4.0–15.9 y). There were no significant differences between girls and boys in either the SQCP group or the control group, and therefore the data were analyzed with the sexes combined.
Lean tissue mass from DXA
Total-body LTM was assessed by DXA (LTMDXA). All DXA measurements were performed and analyzed by trained staff in the Department of Nuclear Medicine at CHW with a commercial X-ray bone densitometer (DPX Lunar Radiation Corp, Madison, WI). Whole-body scans were performed by using the recommended scan modes for weight. Software version 4.7 was used to determine body-composition parameters (Lunar Radiation Corp). The total scan time was 10 min with a total radiation dose of 0.2 µSv. Most of the children with SQCP were lightly sedated with 0.35 mg midazolam/kg given orally 30 min before the scan to reduce movement during the test. The precision of the technique to measure LTM in children as assessed at CHW is 0.82% (18).
The LTMDXA results were significantly different between the sexes for both the children with SQCP and the control children; therefore, the data for girls and boys were analyzed separately. Three children with SQCP (1 girl, 2 boys) had metallic implants and were removed from the analysis.
Total body protein from DXA and anthropometry
Total body protein was calculated from the DXA measurement (TBPDXA) and also from skinfold-thickness anthropometry (TBPSKIN) according to the method of Fuller et al (12) as follows:
RESULTS
Subject characteristics
The anthropometric characteristics of the children with SQCP and the control children are shown in Table 1. Except for percentage body fat in the control group, there were no significant differences between boys and girls. Therefore, the sexes were analyzed together, excluding percentage body fat, which is presented in the text below.
View this table:
TABLE 1. Anthropometric characteristics of the children with spastic quadriplegic cerebral palsy (SQCP) and of the control children1
The mean percentage of ideal weight-for-height of the children with SQCP was 96.3 ± 14.9% (n = 42), which was not significantly different from 100%. In contrast, percentage body fat was significantly lower in the female children with SQCP than in the female controls (12.5 ± 7.3% and 22.5 ± 7.3%, respectively; P < 0.001) and in the male children with SQCP than in the male controls (14.2 ± 7.0% and 17.6 ± 4.5%, respectively; P < 0.01).
Height, weight, and FFMSKIN were significantly positively correlated with age in both the children with SQCP and the control group (data not shown). However, height SD scores, weight SD scores, upper arm muscle area SD scores, and subscapular skinfold-thickness SD scores in the children with SQCP were all significantly negatively correlated with age (data not shown).
Total body protein from neutron activation analysis
Fifty-three (21 girls, 32 boys) children with SQCP underwent measurements of TBPNAA. The mean TBPNAA of the children with SQCP was 3.0 ± 1.2 kg, whereas that of the controls was 6.4 ± 2.9 kg (43 girls, 29 boys). The TBPNAA results for the children with SQCP, expressed as a percentage of predicted values from control data for age, height, and weight, respectively, were significantly low: 56.1 ± 17.3% (P < 0.001; n = 45), 81.5 ± 15.7% (P < 0.001; n = 34), and 83.9 ± 17.2% (P < 0.001; n = 45).
TBPNAA was significantly positively correlated with age in both the children with SQCP (r = 0.74, P < 0.001) and the control group (r = 0.89, P < 0.001). However, the percentage TBPNAA predicted for age from the control data was significantly negatively correlated with age (r = –0.58, P < 0.001; n = 45) in the children with SQCP (Figure 1). In the children with SQCP, TBPNAA was highly correlated with FFMSKIN (r = 0.86, P < 0.001), weight (r = 0.80, P < 0.001), height (r = 0.75, P < 0.001), upper arm muscle area (r = 0.69, P < 0.001), and midupper arm circumference (r = 0.56, P < 0.001). There were no significant correlations between TBPNAA and measures of body fatness (data not shown).
FIGURE 1.. Percentage of total body protein by neutron activation analysis (TBPNAA) predicted for age versus age in children with spastic quadriplegic cerebral palsy. The solid line indicates the mean of the control group (100%); the dashed line is the regression line (n = 45, r = –0.58, P < 0.001).
Lean tissue mass from DXA
Total LTMDXA was significantly lower in the children with SQCP (21 girls, 34 boys) than in the controls (84 girls, 88 boys): girls, 14.3 ± 4.1 compared with 29.4 ± 9.2 kg (P < 0.001); boys, 17.6 ± 6.0 compared with 33.6 ± 15.1 kg (P < 0.001). There was no significant group-by-sex interaction. LTMDXA was significantly positively correlated with age in both the control group (r = 0.86, P < 0.001) and the children with SQCP (r = 0.73, P < 0.001). There was a significant correlation between LTMDXA and TBPNAA in both the children with SQCP and in a subset of similarly aged control children who had both DXA and NAA measurements: r = 0.87, P < 0.001 (n = 55), and r = 0.995, P < 0.001 (n = 28), respectively.
Total body protein from DXA and skinfold-thickness anthropometry
Complete data sets for measures of NAA, DXA, and skinfold-thickness anthropometry were available for 52 (21 girls, 31 boys) children with SQCP and 28 (17 girls, 11 boys) controls. There were no significant group-by-sex interactions. Mean TBPDXA and TBPNAA did not differ significantly in the children with SQCP (3.1 ± 1.2 kg and 3.0 ± 1.2 kg, respectively; P = 0.12), but did differ significantly in the control group (6.2 ± 3.1 and 7.2 ± 3.6 kg, respectively; P < 0.001). Mean TBPSKIN was significantly different from TBPNAA in the children with SQCP (3.4 ± 1.1 and 3.0 ± 1.2 kg, respectively; P < 0.001) and in the controls (6.1 ± 2.8 and 7.2 ± 3.6 kg, respectively; P < 0.001). There was a strong positive correlation between TBPNAA and TBPDXA, and also between TBPNAA and TBPSKIN, in both the children with SQCP (r = 0.91, P < 0.001, and r = 0.90, P < 0.001, respectively) and the control group (r = 0.99, P < 0.001, and r = 0.99, P < 0.001, respectively).
Bland and Altman plots were constructed to test agreement between TBPNAA and TBPDXA and TBPNAA and TBPSKIN (Figure 2 and Figure 3). For the children with SQCP, there was a mean difference between TBPDXA and TBPNAA of 0.1 ± 0.5 kg. As shown in Figure 2A, the 95% range of agreement was –0.9 to 1.1 kg (–2 SDs to +2 SDs). For the average child in this study, this equated to a variation of 1.0 kg, or 33.3%. Conversely, for the control children, TBPDXA underestimated body protein by a mean of 1.0 ± 0.6 kg with a variation of ±15.4%. Statistical analysis (univariate ANCOVA) showed that the pattern of the Bland and Altman plots was significantly different between the children with SQCP and the controls (P = 0.02).
FIGURE 2.. Bland and Altman plots for total body protein measured by neutron activation analysis (TBPNAA) versus that measured by dual-energy X-ray absorptiometry (TBPDXA) in children with spastic quadriplegic cerebral palsy (A; n = 52) and in control children (B; n = 28). The dashed line indicates the regression line. In A: r = 0.06, P = 0.69;
FIGURE 3.. Bland and Altman plots for total body protein measured by neutron activation analysis (TBPNAA) versus that measured by skinfold-thickness anthropometry (TBPSKIN) in children with spastic quadriplegic cerebral palsy (A; n = 52) and in control children (B; n = 28). The dashed line indicates the regression line. In A: r = –0.17, P = 0.24;
TBPSKIN overestimated body protein by a mean of 0.3 ± 0.5 kg in the children with SQCP. As shown in Figure 3A, the 95% range of agreement was –0.7 to 1.3 kg (–2 SDs to +2 SDs). For the average child in this study, this equates to a variation of 1.0 kg, or 33.3%. For the control children, TBPSKIN underestimated body protein by a mean of 0.7 ± 0.6 kg with a variation of ±16.5%. Statistical analysis (univariate ANCOVA) showed that the pattern of the Bland and Altman plots was not significantly different between the children with SQCP and the controls (P = 0.46). Further analysis showed that the intercepts of the regression lines on the y axes were significantly different between the 2 groups (P < 0.001).
DISCUSSION
In the present study, we found that children with the most severe form of cerebral palsy were stunted and had significantly lower body fat and protein than did similarly aged controls. We also found that the body protein of children with SQCP was markedly lower than that of healthy controls of similar height. Furthermore, the difference in body protein between the children with SQCP and the control group appeared to be greater as the children aged; in particular, there was an increasing divergence in body protein from 10 y of age (Figure 1). This finding is supported by Stallings et al (6), who showed a similar deviation in FFM with age by using an indirect method to measure FFM.
The low TBP for height and age in our study could be a result of reduced muscle mass due to malnutrition coupled with muscle atrophy from the underlying neurologic impairment. The relative contribution of either malnutrition or neurologic impairment could not be determined in this study. A longitudinal, prospective nutritional rehabilitation study is required to resolve this issue. A few studies, mostly retrospective studies with small numbers, have shown weight gain and an increase in fat stores after gastrostomy insertion in children with cerebral palsy (29–34). However, none of these studies measured body protein.
The children with SQCP in the present study had major deficits in height and weight. However, their percentage of ideal weight-for-height was not significantly different from 100%, which implies that although they were stunted they were not wasted. Yet, the more detailed nutritional assessment including skinfold thicknesses and TBP found that they had depletions of both body fat and protein compartments for age and height that were consistent with wasting. These findings suggest that the commonly used Waterlow classification system of weight-for-height as an indicator of malnutrition with wasting is inappropriate for this group of children (35). Why has this apparent incongruity occurred? One possible explanation is that the use of knee height to estimate height could be inaccurate; this method has not been cross-validated in children with SQCP (20). It is also possible that malnourished children with SQCP have abnormally high TBW, which has been reported in malnutrition (36). Increased TBW could result in falsely elevated body weight and lean body mass (which would therefore result in increased body protein when calculated from DXA and skinfold-thickness anthropometry) but would not affect the measurement of TBP by NAA.
Others have measured TBW in children with cerebral palsy by using isotope-dilution methods and have reported various results (9, 11, 37). One study found a higher TBW (37), another found a lower TBW (11), and another found that TBW did not differ significantly from that in the control population (9). However, each of these studies had small numbers of subjects with varying severities of cerebral palsy and who were relatively well-nourished with equivalent or higher percentages of body fat compared with the healthy controls.
The second aim of the present study was to assess the validity of anthropometry and DXA to predict body protein in children with SQCP. Significant correlations were found between anthropometric measures and TBPNAA and between DXA and TBPNAA. Despite these significant correlations, however, agreement between the methods showed that both skinfold-thickness anthropometry and DXA produced wide variation in the estimation of TBP in children with SQCP.
The method used in the present study to estimate body protein from anthropometry and DXA was based on 2 assumptions. First, that the hydration of the LTM was estimated; and second, that the ratio of nonosseous mineral to bone mineral was 0.18:0.82, such that the mass of nonosseous mineral was 0.23 x BMC. The applicability of these assumptions to our population is questionable because it is unknown whether malnourished children with SQCP have normal body water, and, furthermore, most children with SQCP are osteopenic (1). Studies including measures of both TBW and TBP are required in children with SQCP to determine specific hydration factors to use with indirect measures of body protein.
In conclusion, the body protein of children with SQCP is significantly reduced for both age and height. A longitudinal prospective nutritional rehabilitation study is required to determine whether body protein can be improved with weight gain. In addition, the significant correlations between TBPNAA and anthropometric measures and DXA found in this study suggest that these measures can be used as indicators of body protein in children with SQCP. However, agreement analysis showed wide variation in some individuals, which indicates that the use of skinfold-thickness anthropometry and DXA to derive body protein should be treated with caution.
ACKNOWLEDGMENTS
We gratefully acknowledge Madeleine Thompson for her assistance with DXA scans and the parents, caregivers, and children who generously volunteered their time to participate in the study.
FEA coordinated the study as part of her PhD and was involved in the preparation of the study, ethics submissions, organization of subject testing, collection and analysis of data, and writing of the manuscript. SLC recruited, tested, and analyzed the data for the first 17 subjects enrolled in the study as part of her Master of Nutrition and Dietetics degree. JRA, KJG, MAG, HS, and EVO were involved in the conception of the study, ethics submissions, subject recruitment and testing, analysis of the data, and writing of the manuscript. KJG was involved in the conception and development of the NAA technique at CHW and holds the current radiation license for this instrument. JRA and MAG performed the TBN measurements. JNB and RBH-G were responsible for the collection and development of the DXA control database, the DXA subject testing and data analysis, and review of the manuscript. None of the authors had a financial conflict of interest.
REFERENCES