点击显示 收起
1 From the Department of Medicine, Monash University; the Department of Medical Physics and the Body Composition Laboratory, Monash Medical Centre; the Department of Gastroenterology, Royal Children's Hospital; and the Department of Child Psychiatry, Austin Hospital, Melbourne.
2 DJ Borovnicar was funded by a Monash University Departmental Scholarship.
3 Address reprint requests to DB Stroud, Medical Physics, Monash Medical Centre, 246 Clayton Road, Clayton, VIC 3168, Melbourne, Australia. E-mail: d.stroud{at}shcn.com.au.
ABSTRACT
Background: Symptoms of cystic fibrosis (CF) may limit the utility of total body chlorine (TBCl) and total body potassium (TBK) measurements for assessing body fluid compartments of children.
Objective: This study assessed relations among independent measurements of TBCl, TBK, and total body water (TBW) in children with CF.
Design: We compared cross-sectional measurements of TBCl by in vivo neutron activation analysis, TBK by whole-body counting of 40K, TBW by D2O dilution [TBW(D2O)], and TBW from TBCl and TBK [TBW(Cl + K)] in 19 prepubertal children (13 boys) aged 7.6–12.5 y who had mild symptoms of CF. Body-composition measurements were compared with data from previous studies of healthy children.
Results: Subjects with CF had deficits in TBCl, TBK, TBW, and body weight compared with control reference data (P < 0.05). The ratios (TBCl + TBK)/TBW and TBCl/TBK were not significantly different from control reference values, and plasma chlorine and potassium concentrations were within control reference ranges. The sum of TBCl and TBK correlated with TBW(D2O) (r2 = 0.79, P < 0.001), and TBW(Cl + K) correlated with TBW(D2O) (r2 = 0.78, P < 0.001). TBW(Cl + K) was similar to TBW(D2O) ( ± SEM: 19.0 ± 0.5 compared with 19.4 ± 0.5 L; NS).
Conclusions: Prepubertal children with mild symptoms of CF can develop deficits in TBCl, TBK, and TBW that reflect chronic energy malnutrition. Mild symptoms of CF do not appear to affect normal relations among TBCl, TBK, and TBW. Measurements of TBCl and TBK may be used to assess body fluid compartments in these patients.
Key Words: Body composition • total body chlorine • total body potassium • total body water • extracellular water • intracellular water • cystic fibrosis • prepubertal children
INTRODUCTION
Protein-energy malnutrition is a common sequela in children with cystic fibrosis (CF) (1) and may result in abnormal whole-body distributions of electrolytes and fluids (2–4). The assessment of whole-body electrolyte and fluid compartments is therefore important in determining the nutritional status of children with CF.
Measurements of total body chlorine (TBCl) and total body potassium (TBK) have been used to derive estimates of extra- and intracellular fluid compartments in normal-weight and obese adults (5, 6). This approach is based on the assumptions that 88% of TBCl is found in extracellular water (ECW) (7) and that >96% of TBK is found in intracellular water (ICW) (8).
In children with CF, TBCl and TBK measurements may be influenced by factors that include protein-energy malnutrition and the transepithelial chlorine transport abnormality that is the intrinsic defect in CF (9). Although TBCl, TBK, and fluid compartments were studied previously in children with CF (2–4, 10–12, and IRJ Humphries and KJ Gaskin, unpublished observations, 1996), interpretation of these data is problematic because of the small numbers of subjects, wide variations in disease severity, incomplete reporting of disease symptoms and body components, and use of body-composition methods that may be invalid in CF populations.
The aim of this study was to define body composition in a group of clinically stable prepubertal children with mild symptoms of CF by 1) comparing independent measurements of TBCl, TBK, and total body water (TBW) and 2) investigating the respective relations of TBCl and TBK to ECW and ICW.
SUBJECTS AND METHODS
Subject selection
The study group consisted of 19 children (13 boys) with CF who were aged 7.6–12.5 y. All subjects were selected from the CF clinic of the Royal Children's Hospital, Melbourne, according to the following criteria: 1) age 7.0–13.0 y; 2) prepubertal (ie, Tanner stages 1 or 2); 3) positive diagnosis of CF on the basis of genotyping or pulmonary and gastrointestinal symptoms complemented by increased sweat chloride concentrations; 4) mild symptoms of lung disease as judged by a physical examination, chest X-rays, and spirometry tests; 5) no pulmonary exacerbations that required hospitalization within 2 mo of the study; 6) no hospitalization within 2 mo of the study; 7) weight change <1.5 kg over the 2 mo before the study; 8) no comorbidities likely to confound body-composition measurements (eg, clinical or biochemical evidence of liver disease or hepatosplenomegaly, pancreatitis, diabetes, or edema); and 9) no history of long-term (>1 y) use of medications likely to confound body-composition measurements (eg, corticosteroids, growth hormone, or diuretics). Reference data from previous studies of healthy children were used to interpret clinical, biochemical, and body-composition measures.
Study protocol
All subjects were studied as outpatients in 2 phases with the approval of the human ethics committees of both the Royal Children's Hospital and the Monash Medical Centre and the signed, witnessed, and informed consent of subjects and their parents or guardians. In phase 1, clinical status, diet, and pancreatic function were assessed. In phase 2, anthropometry, body composition, and plasma biochemistry were assessed within a 4-h morning session. Each subject participated in both phases within a period of 12 wk.
Clinical assessment
Body weight was measured with a digital scale and body height was determined by using a stadiometer according to the standards of Lohman et al (13). Anthropometric measures were interpreted by using World Health Organization (WHO) normal population percentiles for weight and height (14). Growth velocities were expressed as a percentage of the normal 50th-percentile growth velocities reported by Tanner et al (15), and pubertal status was determined according to the Tanner stages of development (16).
Medical records of spirometry test results were reviewed to determine each subject's measure of forced expiratory volume in 1 s (FEV1), expressed as a percentage of that expected of healthy children of the same age and sex. The pulmonary status of each subject was assessed by a thoracic physician and assigned a Holzer score (17) as follows: 0 (no lung disease) for subjects with no persistent cough, no sputum production, a normal chest radiograph, and an FEV1 >80%; 1 (minimal lung disease) for subjects with persistent cough, no sputum production, a chest radiograph showing minimal streaking, and an FEV1 >60% but <80%; and 2 (mild lung disease) for subjects with persistent cough, sputum production <10 mL/d, a chest radiograph showing bronchial-wall thickening and minimal hyperinflation, and an FEV1 >65% but <85%.
Subjects receiving pancreatic enzymes at the time of the study were considered to be pancreatic insufficient. Pancreatic status was also assessed by a 3-d fat-clearance study (18). Each subject completed a 3-d weighed-food diary and stool collection within 2 wk of body-composition assessment. The fat content of stools was assessed by using the method of Van der Kamer et al (19), and a coefficient of fat absorption was calculated as 100 x [fat intake (g) - fat excreted in stool (g)]/[fat intake (g)].
The nutrient contents of dietary food items were analyzed by using NUTRITIONIST III, version 4 (N-Squared Computing, Salem, OR; modified for use by Monash University, Melbourne, Australia) and food-composition tables (20). Energy and protein intakes were expressed as a percentage of the recommended dietary intakes (RDIs) published by the National Health and Medical Research Council of Australia (20). Medical histories were reviewed to determine each subject's use of corticosteroids.
Laboratory assessment
Subjects were designated as either homozygous or heterozygous for the F508 mutation on the basis of DNA testing by the Victorian Clinical Genetics Service, Melbourne. Fasting venous blood samples were collected for laboratory analysis immediately before each subject's body-composition assessment.
Body-composition assessment
TBCl was measured by the method of prompt gamma-ray in vivo neutron activation analysis (IVNAA) as described elsewhere (21, 22). Subjects were irradiated unilaterally by a 0.2-GBq 252Cf neutron source from shoulder to midthigh in both supine and prone positions. Prompt gamma-ray spectra were measured with 2 pairs of NaI(Tl) crystals (each crystal: 10 x 10 x 15 cm) positioned on both sides of the subject. TBCl was estimated from the ratio of chlorine-to-hydrogen counts as determined from the measurement of 8.57 and 2.22 MeV prompt gamma-rays from the respective reactions 35Cl(n,)36Cl and 1H(n,)D. The chlorine-to-hydrogen count ratio was corrected for the effect of body width and thickness on background and gamma-ray attenuation. Total body hydrogen was used as an internal standard, which was determined by using a 4-compartment model of body weight (22). The total exposure time per measurement scan was 630 s and the effective dose equivalent for a small child was 0.25 mSv (Q = 20, indicating exposure to neutrons). During this study, the assay error and precision were determined to be 3% and 10% (CV), respectively, for TBCl measurements.
TBK was measured by counting 1.46 MeV gamma-ray emissions from 40K by using a shadow shield counter (23). Subjects rested supine on a moveable bed that passed through a symmetric array of four 40 x 10 x 10-cm NaI(Tl) detectors shielded within an open-ended, box-like steel chamber; the total scan time was 30 min. Measurements were calibrated against a potassium-rich, child-size box phantom, and a correction factor was applied to account for the effect of body thickness and width on gamma-ray attenuation (21). During this study, the error and precision of TBK measurements were determined to be 2.5% and 4% (CV), respectively.
TBW was measured by using the D2O isotope dilution method (24). A baseline venous blood sample was first collected from the fasting subject and then an oral dose of 0.4 g 99.9% pure D2O/kg body wt was administered with a mouthwash of deionized water (100 mL). The subject fasted for a further 2–3-h equilibration period until a second venous blood sample was collected. The D2O plasma concentration of each plasma sample was determined in triplicate by using the Fourier transform infrared analysis technique and the average measure was then used to determine TBW by using the equation
RESULTS
Clinical assessment
Given that there were no significant sex differences in any of the clinical measures investigated, the clinical characteristics of boys and girls are combined in Table 1. The CF group had subnormal z scores for weight-for-age, weight-for-height, and BMI and a mean height z score indicative of a nonsignificant degree of stunting. Negative linear correlations were observed between age and z scores for weight-for-age (r2 = 32.2, P = 0.01) and height-for-age (r2 = 37.0, P < 0.01). Individual z scores for weight-for-age, weight-for-height, and height-for-age all fell within 2 SDs of reference medians. BMI-for-age z scores of all but 2 subjects fell within 2 SDs of the reference median.
View this table:
TABLE 1.. Clinical and laboratory assessment of children with cystic fibrosis1
Subjects showed a subnormal rate of growth in weight and a normal rate of growth in height. All subjects were prepubertal with Tanner scores of 1 or 2. Subjects had minimal symptoms of lung disease, as judged by a mean (±SEM) Holzer score of 0.9 ± 0.2, and subnormal lung function (FEV1 < 100%; P < 0.001). At the individual level, 6 subjects (all boys) showed no clinically overt symptoms of lung disease, 9 (5 boys) showed minimal symptoms of lung disease, and 4 (2 boys) showed mild symptoms of lung disease.
All subjects were pancreatic insufficient and used pancreatic-enzyme supplements. Dietary fat absorption tests were successfully completed for 15 of 19 subjects. These 15 subjects could absorb only 86.4% of ingested dietary fat with the use of enzyme supplements.
At the time of study, all subjects were consuming an ad libitum diet. The 3-d weighed-food diary was completed by 18 of 19 subjects. A comparison with RDIs showed a significant deficit in total energy intake and a significant surfeit of protein intake. Only 2 subjects (1 boy) reported taking salt-tablet supplements in quantities of 1000 mg/d over the 3-d period of dietary assessment.
Only 7 subjects (3 boys) had used steroids, ranging in frequency from once or twice in their lifetime through to several times/y. There were 17 d on which body-composition measurements were performed. The mean (±SD) maximum temperature for these 17 d was 22.3 ± 6.2°C (range: 15.0–36.5).
Laboratory assessment
Seven subjects (5 boys) were homozygous for the F508 allele, 9 (6 boys) were heterozygous for the F508 allele, and 3 (2 boys) had no F508 allele. The CF group had plasma chlorine, sodium, and potassium concentrations within the control reference ranges. Plasma chlorine and potassium concentrations showed no significant association with age, sex, genotype expressed in terms of the F508 allele, steroid use, or any of the variables of anthropometry, growth, pulmonary status, gastrointestinal status, or diet.
Body-composition assessment
General effects of cystic fibrosis on body composition
The body composition of the 19 subjects studied is summarized in Table 2. Deficits were observed for both age- and height-standardized values of TBCl, TBK, and TBW. Negative linear correlations were observed between age and age-standardized values of TBCl (r2 = 29.2, P = 0.02), TBK (r2 = 30.6, P = 0.01), and TBW (r2 = 41.2, P < 0.01) considered separately. By contrast, mean values for the ratios (TBCl + TBK)/TBW and TBCl/TBK did not differ significantly from the cited reference values. Neither of these ratios was significantly associated with age, genotype expressed in terms of the F508 allele, steroid use, or any of the variables of anthropometry, growth, pulmonary status, gastrointestinal status, or diet. Sex did not significantly affect any body-composition measure.
View this table:
TABLE 2.. Body-composition assessment of children with cystic fibrosis1
Comparison of TBCl and TBK with TBW(D2O)
There was a relatively strong linear correlation between the sum of TBCl and TBK and TBW(D2O), as shown in Figure 1. The sum of TBCl and TBK was a significant predictor of TBW(D2O) (P < 0.001), accounting for 79% of the variation in TBW(D2O).
FIGURE 1. . Total body water as measured by D2O dilution [TBW(D2O)] compared with the sum of total body chlorine plus total body potassium (TBCl + TBK) in 19 children (13 boys) with cystic fibrosis. Regression equation: TBW(D2O) = 0.00594 x + 2.67 (SEE = 1.09 L, r2 = 0.79, P < 0.001).
Comparison of TBW(Cl + K) with TBW(D2O)
Absolute mean values of TBW(D2O) and TBW(Cl + K) in boys and girls and in the combined group are shown in Table 3. The paired t test showed no significant difference between the 2 estimates of TBW. In addition, no significant sex difference was found in estimates of TBW by either method when the unpaired t test was used.
View this table:
TABLE 3.. Total body water (TBW) as measured by D2O dilution [TBW(D2O)] and from the sum of total body chlorine plus total body potassium [TBW(Cl + K)] in children with cystic fibrosis1
A linear regression analysis of TBW(D2O) compared with TBW(Cl + K) in the CF group showed TBW(Cl + K) to be a significant predictor of TBW(D2O) (P < 0.001). The slope of the regression line was not significantly different from 1.0 (one-sample t test), and the intercept was not significantly different from 0 (one-sample t test).
Results from the Bland and Altman analysis of agreement between the D2O dilution method and the method of summing TBCl and TBK are summarized in Table 4 and Figure 2. The 95% CI for the average measurement bias between the 2 methods included zero for both boys and girls and for the combined group. For individual subjects, the variation of differences between the 2 methods are indicated by the limits of agreement, which spanned 4.95 L in boys, 2.87 L in girls, and 4.69 L in the combined group. Also, measurement biases between the 2 methods were not significantly associated with age, sex, genotype expressed in terms of the F508 allele, steroid use, or any of the variables of anthropometry, growth, pulmonary status, gastrointestinal status, or diet.
View this table:
TABLE 4.. Comparison of total body water (TBW) measured by D2O dilution [TBW(D2O)] with TBW estimated from the sum of total body chlorine plus total body potassium [TBW(Cl + K)] in children with cystic fibrosis1
FIGURE 2. . Differences between total body water measured by D2O dilution [TBW(D2O)] and TBW estimated from the sum of total body chlorine plus total body potassium [TBW(Cl + K)] in 19 children (13 boys) with cystic fibrosis. The mean difference (bias), the 95% CI of the mean difference, and the limits of agreement are also shown.
DISCUSSION
Effect of mild symptoms of cystic fibrosis on TBCl, TBK, and TBW
The CF group in this study had a body-composition pattern consistent with compromised growth of new tissue and loss of existing tissue. This was evidenced by 1) significant deficits in age- and height-standardized measures of TBCl, TBK, TBW, and body weight; 2) age-standardized deficits in TBCl, TBK, TBW, and body weight that increased significantly with age; and 3) a subnormal weight velocity. By contrast, individual subjects appeared capable of attaining normal TBCl, TBK, TBW, or body weight for age or height.
There is a paucity of TBCl data from subjects with CF. Measurements of TBCl in apparently well-nourished children with CF were recorded (IRJ Humphries and KJ Gaskin, unpublished observations, 1996). TBCl was determined by IVNAA in 55 fibrocystic children with an average age of 10.8 y and 28 healthy children with an average age of 11.1 y. Overall, the CF group weighed 8 kg less and had 8% less TBCl than did the control group. However, when TBCl was adjusted for body weight, there was no significant difference in TBCl between the 2 groups. Humphries and Gaskin concluded that, on a body-weight basis, well-nourished children with CF have normal amounts of TBCl. The only other report of TBCl in subjects with CF appears to be that of Cheek (4), who used the NaBr dilution technique to measure TBCl in 5 malnourished, fibrocystic children aged between 0.5 and 4 y. Four of the 5 children had an increased ratio of TBCl to body weight; this finding was interpreted to signify a loss of total body fat. The TBCl observations in the present study extend the findings of Cheek (4) and are consistent with the data of Humphries and Gaskin. The TBK and TBW observations in this study support results of past studies showing that children with CF can attain normal amounts of TBK and TBW and that age- and height-related deficits in TBK and TBW develop in relation to the severity of CF symptoms (10–12, 36–38).
Multiple factors may affect TBCl, TBK, and TBW in children with CF. These include the normal influences of age, sex, growth, and physical activity (8, 26, 29–32, 39, 40) and the abnormal influences of malnutrition secondary to pulmonary and gastrointestinal symptoms of CF (12, 41, 42) and inadequate dietary intake (1, 43, 44). In this study, CF-related lung disease, pancreatic insufficiency, and an inadequate dietary energy intake were the factors most likely to influence body-composition abnormalities. This statement is supported by the observation that 18 of 19 subjects had minimal or mild symptoms of lung disease. Also, all subjects were pancreatic insufficient, and the CF group reported a reduced dietary energy intake equivalent to 89% of the RDI (see Table 1).
Comparisons between TBCl, TBK, and TBW
The results of this study suggest that children with mild symptoms of CF can maintain normal relations among TBCl, TBK, and TBW. This is supported by the following findings. First, a positive linear correlation was observed between TBCl + TBK and TBW(D2O) (see Figure 1). This result was independent of sex. Because TBCl is largely confined to the extracellular compartment (88% by weight) (3, 7), TBCl should be related to the volume of this compartment. Similarly, TBK is largely confined to the intracellular compartment (>96% by weight) (7, 33) and should therefore be related to the volume of this compartment. The observed correlation is not unexpected and is consistent with similar findings reported in healthy children of similar age (26) and in healthy adults (5).
Second, the ratio (TBCl + TBK)/TBW(D2O) in the CF group attained a mean (±SEM) value of 145.3 ± 1.9 mmol/L, which was similar to the control value of 146 mmol/L calculated from the reference data of Cheek (26). Although this ratio showed wide variability in individual subjects (126.2–155.7 mmol/L), the ratio was not significantly associated with age, sex, or the measures of anthropometry, growth, genotype, pulmonary and gastrointestinal status, diet, or steroid use.
Third, the ratio TBCl/TBK was not significantly different from the control value derived from the reference data of Cheek (see Table 2). Our results are consistent with those of past studies of healthy humans that showed the ratio TBCl/TBK falling within a relatively constant range from 0.6 to 0.8 mmol/mmol during childhood, adolescence, and adulthood (26, 28).
The relation of TBCl to ECW
In healthy children and adults, 88% of TBCl is located in disparate pools of extracellular fluid (3, 7). Although chlorine concentrations of extracellular fluids may vary considerably, the average chlorine concentration of the total ECW pool is closely approximated by the chlorine concentration of blood plasma (100 mmol/L).
The intrinsic defect identifying CF is abnormal transepithelial chlorine transportation. Whether this defect is associated with TBCl abnormalities in CF patients was questioned previously. The results of the present study suggest that children with mild symptoms of CF can maintain a normal relation between TBCl and ECW. Subjects showed a normal plasma chlorine concentration consistent with that reported in previous studies (45–47). This study also found a strong linear correlation and a good agreement between TBW(Cl + K) and TBW(D2O). In calculating ECW from TBCl, we assumed that 88% of TBCl is found in ECW and that the plasma chlorine concentration is representative of the average chlorine concentration of the total ECW pool. Our results suggest that these assumptions are suitable for clinically stable children with mild symptoms of CF.
The relation of TBK to ICW
This study identified the following supportive evidence for a normal relation between TBK and ICW in clinically stable children with mild symptoms of CF. First, plasma potassium concentrations were within the normal reference range for 17 of 19 subjects, slightly above the upper reference value for one subject (5.2 compared with 5.0 mmol/L), and slightly below the lower reference value for another (3.4 compared with 3.5 mmol/L).
Second, we observed a strong linear correlation and a good agreement between TBW(Cl + K) and TBW(D2O). ICW was derived from TBK by using a cellular model that assumes that 100% of TBK is found in the body cell mass, that the ratio of body cell mass to TBK is 0.00833 kg/mmol, and that 75% of the body cell mass is ICW by weight (33, 34). Our data suggest that this cellular model of TBK is also suitable in children with mild symptoms of CF.
The validity of TBW measurements by D2O dilution
This study used the D2O dilution method as a reference standard against which TBW estimates derived from TBCl + TBK were compared. In an ancillary experiment performed in the CF group, no significant difference was observed between TBW measured by D2O dilution and TBW derived from the 4-compartment model of body composition defined as body weight (measured by using digital scales) minus the sum of total body fat measured by dual-energy X-ray absorptiometry, total body protein determined by IVNAA, and total body bone mineral measured by dual-energy X-ray absorptiometry [ ± SEM: 19.4 ± 0.5 compared with 19.6 ± 0.5 L for TBW(D2O) and TBW(4-compartment model), respectively; P = 0.36 by paired t test] (21, 22). The significance of this comparison is that the 4-compartment model is based on methods that are independent of the primary CF defect.
Summary
This study showed independent, cross-sectional measurements of TBCl, TBK, and TBW in a group of clinically stable prepubertal children with mild symptoms of CF. Modest deficits in TBCl, TBK, and TBW were consistent with the effect of chronic energy malnutrition secondary to multiple factors that included CF lung disease, pancreatic insufficiency, and an energy-deficient diet. By contrast, relations among TBCl, TBK, and TBW appeared normal and unaffected by symptoms of CF. The clinical significance and implications of these findings are as follows. First, the abnormal CF gene that affects electrolyte and fluid transport across the epithelial surfaces of specific tissues and organs does not appear to affect to any significant extent normal relations between electrolytes and water compartments at a whole-body level. Second, noninvasive measurements of TBCl and TBK may be used as surrogate measures of ECW, ICW, and TBW in clinically stable children with mild symptoms of CF.
ACKNOWLEDGMENTS
We extend our appreciation to the children and parents who willingly participated in this study. We also gratefully acknowledge the support of Colin Robertson from the Department of Thoracic Medicine, Royal Children's Hospital, and Mark Wahlqvist from the Department of Medicine, Monash Medical Centre, and the technical assistance of Ding Wei Xiong from the Department of Medicine, Monash University, and Susan Forrest from the Murdoch Institute, Melbourne.
REFERENCES