Estimating Body Composition in Cystic Fibrosis Patients (2015) PDF

Journal of Cystic Fibrosis 14 (2015) 784 – 791 www.elsevier.com/locate/jcf Original Article...

Journal of Cystic Fibrosis 14 (2015) 784 – 791 www.elsevier.com/locate/jcf Original Article Estimating body composition from skinfold thicknesses and bioelectrical impedance analysis in cystic fibrosis patients Gianfranco Alicandro a , Alberto Battezzati b , Maria Luisa Bianchi c , Silvana Loi a , Chiara Speziali a , Arianna Bisogno a , Carla Colombo a,⁎ a Department of Paediatrics, Cystic Fibrosis Center, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, University of Milan, Milan, Italy b International Center for the Assessment of Nutritional Status, University of Milan, Milan, Italy c Bone Metabolism Unit, Istituto Auxologico Italiano IRCCS, Milan, Italy Received 1 October 2014; revised 29 July 2015; accepted 29 July 2015 Available online 15 August 2015 Abstract Background: The accuracy of body composition estimates based on skinfold thickness measurements and bioelectrical impedance analysis (BIA) is not yet adequately explored in cystic fibrosis (CF). Using DXA as reference method we verified the accuracy of these techniques and identified predictors of body composition specific for CF. Methods: One hundred forty-two CF patients (age range: 8–31 years) underwent a DXA scan. Body fat percentage (BF%) was estimated from skinfolds, while fat free mass (FFM) from single-frequency 50 kHz BIA. Results: Bland–Altman analysis showed poor intra-individual agreement between body composition data provided by DXA and BF% estimated from skinfolds or FFM estimated from BIA. The skinfolds of the upper arm were better predictors of BF% than BMI, while compared to other BIA measurements the best predictor of FFM was the R-index (Height2/Resistance). Conclusions: Due to poor accuracy at individual level, the estimates of body composition obtained from these techniques cannot be part of the standard nutritional assessment of CF patients until reliable CF-specific equations will become available. BMI has limited value in predicting body fatness in CF patients and should be used in combination with other predictors. Skinfolds of the upper arm and R-index are strongly related to BF% and FFM and should be tested in a large CF population to develop specific predictive equations. © 2015 European Cystic Fibrosis Society. Published by Elsevier B.V. All rights reserved. Keywords: Body composition; DXA; Skinfold thicknesses; Bioelectrical impedance analysis; Cystic fibrosis 1. Background (CFTR) gene. The major clinical consequences of impaired CFTR function are pancreatic insufficiency, recurrent infections Cystic fibrosis (CF) is a common recessive genetic disease, and progressive lung disease. CF patients are at high risk for caused by mutations of CF trans-membrane conductance regulator malnutrition because of anorexia with reduced food intake, increased fecal energy loss and high resting energy expenditure during pulmonary exacerbations. Furthermore chronic inflam- Abbreviations: BIA, bioelectrical impedance analysis; BMD, bone mineral density; CF, cystic fibrosis; FEV1, forced expiratory volume in one second; mation and reduced physical activity contribute to alterations in FM, fat mass; FFM, fat free mass; BF%, body fat percentage; ICC, intraclass body composition. In CF patients the chronic inflammatory coefficient of correlation; LA, limits of agreement; R, resistance; RMSE, root process has been shown to reduce fat free mass (FFM), that in mean square error; SF, skinfold; SF-BIA, single-frequency bioelectrical turn leads to a reduction in skeletal muscle mass, inspiratory impedance analysis; MF-BIA, multi-frequency bioelectrical impedance analysis. muscle wasting and impairment of inspiratory muscle function ⁎ Corresponding author at: Cystic Fibrosis Center, Fondazione IRCCS Ca' Granda, Ospedale Maggiore Policlinico, University of Milan, Via Commenda [1–3]. In addition, recent studies have documented an association 9, 20122 Milan, Italy. Tel.: + 39 02 55032456; fax: +39 02 55032814. between FFM depletion and poor respiratory muscle strength or E-mail address: [email protected] (C. Colombo). impaired pulmonary function [4,5]. http://dx.doi.org/10.1016/j.jcf.2015.07.011 1569-1993/© 2015 European Cystic Fibrosis Society. Published by Elsevier B.V. All rights reserved.  G. Alicandro et al. / Journal of Cystic Fibrosis 14 (2015) 784–791 785 However standard nutritional indicators such as height, 2.1. Skinfold thickness measurements weight and BMI do not allow identifying the body composition alteration frequently found in CF patients [5–7]. SF thickness was determined to the nearest 0.2 mm at the Given the above mentioned reasons, the possibility to evaluate left biceps, triceps, sub-scapular and supra-iliac sites using a body composition in these patients is clinically relevant and Holtain skinfold caliper calibrated to exert a constant pressure in-depth assessment of nutritional status should include fat mass of 10 g/mm2 (Holtain Ltd, Crymych, UK). Triplicate readings (FM) and FFM analysis. were made as recommended by Lohman et al.. In subjects The European CF bone mineralization guidelines recom- aged less than 17 years, BF% was predicted by the equation of mend routine bone density DXA scans in the age range Slaughter et al., while in older patients BF% was predicted between 8 and 10 years to assess bone mineral density (BMD) using the equation by Durnin and Womersley.. DXA is a widely recognized method of body composition analysis that beyond BMD provides information of the 2.2. Bioelectrical impedance analysis nutritional status of the patients, including fat reserve and lean soft tissue. DXA can also be used for nutritional monitoring BIA was performed using a single-frequency BIA analyzer over an extended period of time because the radiation exposure (BIA 101 RJL, Akern Bioresearch, Firenze, Italy) which is low. However, access to DXA may be limited in many applies a 50 kHz oscillating current of 800 μA. The subject situations and the cost of the measurement may be relevant, was set in supine position and the electrodes placed on the therefore DXA may not be used extensively for nutritional dorsal surface of the right foot and ankle, as well as on the right follow-up. wrist and hand. Resistance (R) was recorded and normalized by For many years researchers have been looking for simple, easy height (R-index: height2/R). FFM (kg) was predicted by the to perform and sufficiently accurate methods to estimate and software BodyGramPro 3.0®. monitor body composition in routine clinical practice. Skinfold thickness measurement and bioelectrical impedance analysis 2.3. Dual X-ray absorptiometry (BIA) have proved to be useful in population studies, but perform worse on individual basis. Participants underwent whole-body DXA scanning (DXA Several predictive equations based on skinfolds and BIA Hologic Discovery Hologic, Waltham, USA), a technique that measurements in healthy subjects and in patients affected by is able to estimate FM and FFM by measuring the X-ray different diseases have been published. These equations have attenuation of different tissue components, after the instrument shown low accuracy when applied in a population different calibration with a phantom (including six fields of lucite and from that where they were derived. aluminum of varying thickness and known absorptive proper- Therefore the objectives of this study were: 1) to verify the ties) scanned as an external standard. The whole-body DXA agreement between SF and BIA with DXA in CF patients, scans were performed at the Istituto Auxologico Italiano and using DXA as reference method; and 2) to identify predictors of analyzed with the same protocol throughout the entire duration body composition in a CF population. of the study. At our laboratory, the coefficients of variation (CVs) for FM and FFM were 2.9% and 1.7%, respectively. Body composition was assessed by the three compartment 2. Methods model (version 12.3); including FM, bone mineral content, and FFM. FFM was calculated using the sum of the estimates of One hundred forty-two Caucasian CF patients attending the CF lean tissue mass and bone mineral content. BMD was Center of Milan underwent DXA scan, BIA and skinfold determined on total body in adults and in total body excluding measurements on the same day. the head in pediatric patients, because the skull does not Height was measured to the nearest 0.5 cm on a standardized develop at the same rate as the other parts of the skeleton. stadiometer. Body weight was determined to the nearest 0.1 kg Age and gender-specific Z-scores of BMD were calculated. The by a standard physician scale with the subjects dressed only in Z-scores were calculated with reference to the BMD of healthy light underwear and without shoes. BMI was calculated as weight age- and sex-matched Italian population obtained in our (kg) divided by height squared (m2). Age- and gender-specific laboratory. Our reference database includes the anthropometric, Z-scores of height, weight and BMI were calculated using the gender, age, puberty, and DXA data of 410 healthy Italian CDC references. Forced expiratory volume in one second subjects, 2–25 years of age. (FEV1) values were considered to define the severity of lung disease according to the Cystic Fibrosis Foundation (CFF) 2.4. Statistical analysis criteria (severe FEV1 b 40%, moderate 40–69%, normal/mild ≥ 70%). Intraclass coefficients of correlation were calculated to Pubertal status was determined using breast and pubic hair analyze correlation between body composition data estimated stages in girls, testicular and pubic hair stages in boys, according from skinfolds or BIA and those obtained by DXA. Mean values to the Tanner criteria. Written and informed consent was of body composition data were compared among methods by obtained from all of the subjects or their parents. The study paired t-test to verify the agreement on the overall population. protocol was approved from the local ethical committee. The percent relative difference between skinfolds or BIA data and 786 G. Alicandro et al. / Journal of Cystic Fibrosis 14 (2015) 784–791 DXA (bias) and the limits of agreement (Bland–Altman 95% healthy population. As expected, males had lower FM compared limits of agreement) were calculated to verify the intra-individual to females in both age categories. Males aged less than 17 years agreement. had higher reduction of BMD compared to females (P b 0.0001). To identify predictors of body composition in CF patients FEV1 was not different between genders. Due to the longer we considered DXA as the reference method and generated duration of sexual maturation in males, fewer males had reached linear regression models, setting as dependent variables BF% sexual maturation compared to females. or FFM obtained by DXA and as independent variables the anthropometric and BIA indices. The R-square and the root 3.2. Comparison of methods mean square error (RMSE) were considered as measures of ability of the models to estimate the body component (BF% Intraclass coefficients of correlation indicated good correla- or FFM). For all the analysis the alpha level of statistical tions between skinfolds or BIA and DXA in both genders and significance adopted was P b 0.05. Data were analyzed by age categories. using SPSS, version 15.0 software (SPSS, Inc, Chicago, IL). In patients aged less than 17 years mean BF% obtained by the equation of Slaughter et al. was lower compared to DXA in both 3. Results genders. In older males mean BF% predicted by the equation of Durnin and Womersley was lower compared to DXA, while it 3.1. Patients' population was not significantly different in females. Mean FFM obtained by BIA was not significantly different The enrolled population included 142 CF patients [63 (44%) from DXA, indicating a good agreement at population level males, mean age 16 ± 6 years, range 8–31 years, of whom 52 (Table 2). (37%) were adults]. Pancreatic insufficiency was present in 102 Fig. 1 shows the Bland–Altman analysis of differences between (72%) patients. Six (4%) patients had severe lung disease, 22 skinfolds or BIA and DXA plotted against the values obtained (16%) moderate lung disease, 112 (80%) mild lung disease from DXA. As indicated by the wide limits of agreement, the and 2 were not able to perform pulmonary function tests. The intra-individual agreement of SF and BIA with DXA was poor. characteristics of the pediatric and adult population with The differences between skinfolds or BIA and DXA tended to be particular regard to anthropometric and body composition consistent across the values obtained from DXA, but in patients characteristics are shown in Table 1. Nutritional indicators and over 17 years there was some over-prediction in the lower and BMD adjusted for age and gender were lower compared to the under-prediction in the upper end of FFM distribution. Table 1 Anthropometric and body composition characteristics of the study population. Patients b17 years Patients ≥ 17 years Male Female Male Female (n = 33) (n = 46) (n = 30) (n = 33) Age Mean ± SD 12 ± 3 12 ± 3 22 ± 4 23 ± 4 Range (years) 8–16 8–16 17–31 17–31 Height cm 144 ± 14 145 ± 13 169 ± 6*** 161 ± 7 Z-score − 0.54 ± 0.94 − 0.21 ± 1.05 − 1.08 ± 0.86 − 0.29 ± 1.03 Weight kg 36.3 ± 10.7 38.2 ± 10.1 61.5 ± 9.6*** 51.9 ± 7.8 Z-score − 0.56 ± 0.99 − 0.26 ± 0.94 − 0.93 ± 1.31 − 0.88 ± 1.18 BMI kg/m2 17.2 ± 2.1 17.8 ± 2.5 21.6 ± 3* 19.9 ± 2.1 Z-score − 0.36 ± 0.89 − 0.17 ± 0.9 − 0.57 ± 1.3 − 0.70 ± 0.80 FFMDXA (kg) 29.7 ± 9.7 29.1 ± 7.5 51.4 ± 6.7*** 39.0 ± 4.9 FMDXA (kg) 7.0 ± 2.6** 9.4 ± 3.6 10.4 ± 5.4* 13.2 ± 3.9 BFDXA (%) 19.4 ± 5.4*** 24.2 ± 5.8 16.3 ± 6.1*** 24.9 ± 4.5 BMC (g) 1227 ± 377 1306 ± 357 2212 ± 400*** 1868 ± 332 BMD (Z-score) − 1.87 ± 1.04*** − 0.72 ± 1.11 − 1.21 ± 1.23 − 0.72 ± 1.26 FEV1 (% of predicted) 93 ± 19 96 ± 15 76 ± 25 76 ± 25 Pre-pubertal, n (%) 10 (29) 14 (30) – – Pubertal, n (%) 21 (64)** 14 (30) – – Post-pubertal, n (%) 2 (6)** 18 (39) – – Menarche, n (%) – 18 (39%) – – Data are presented as mean ± SD or counts (%). Abbreviations: BF = body fat; FM = fat mass; FFM = fat free mass; BMD = bone mineral density; FEV1 = forced expiratory volume in one second. Comparison between genders: ***P b 0.0001, **P b 0.01, *P b 0.05.  G. Alicandro et al. / Journal of Cystic Fibrosis 14 (2015) 784–791 787 Table 2 The addition of R-index in a model containing age and weight Comparison between body composition estimates obtained from skinfold markedly improved the prediction of FFM (without R-index: thickness measurements or bioelectrical impedance analysis and DXA. adjusted R2 = 0.93, RMSE = 2.99 kg; adding R-index: adjusted BF% estimated by SF thickness measurements vs. DXA R2 =0.97, RMSE = 1.88 kg) (Table 4). ICC SF DXA Mean percent Mean (SD) Mean (SD) Relative difference (LA) 4. Discussion Patients b 17 years Males (n = 33) 0.80 15.6 (5.6)*** 19.4 (5.4) − 19% (− 57; 19%) The aims of the study were to evaluate the agreement between Females (n = 46) 0.81 20.4 (5.9)*** 24.2 (5.8) − 15% (− 47; 17%) body composition data obtained from skinfolds or BIA with DXA Patients ≥ 17 years and to identify predictors of body composition specific for CF Males (n = 30) 0.87 14.7 (5.8)* 16.3 (6.1) − 9% (− 44; 26%) patients. This work adds a novel contribution to existing knowledge Females (n = 33) 0.80 25.7 (4.8) 24.9 (4.5) 3% (− 20; 26%) for its large sample size, which enabled a thorough evaluation according to the stages of sexual maturation. In addition, a detailed FFM (kg) estimated by BIA vs. DXA analysis of the individual contribution of the anthropometric and ICC BIA DXA Mean percent impedentiometric measures to the prediction of body composition Mean (SD) Mean (SD) Relative difference (LA) was performed for the first time in CF and new multiple regression Patients b 17 years models were developed. Males (n = 33) 0.99 29.7 (9.4) 29.7 (9.7) 0% (− 10; 10%) As regards the first aim, using predictive equations Females (n = 46) 0.98 29.4 (7.9) 29.1 (7.5) 1% (− 12; 13%) developed in healthy subjects we obtained estimates of body Patients ≥ 17 years composition based on skinfold or BIA measurements that were Males (n = 30) 0.95 51.4 (6.0) 51.4 (6.7) 0% (− 8; 9%) compared with DXA and showed that agreement between these Females (n = 33) 0.89 39.9 (4.4) 39.0 (4.9) 2% (− 9; 14%) techniques and DXA is poor at individual level. Abbreviations: BF% = body fat percentage; BIA = bioelectrical impedance Previous studies that have addressed this relevant issue did analysis; FFM = fat free mass; ICC = intraclass coefficient of correlation; not allow drawing firm conclusion, because the CF populations LA = limits of agreement; SF = skinfold. analyzed were very small [17–23]. Only three studies enrolled Comparisons between techniques: *P b 0.05, ***P b 0.001. more than 50 CF patients and achieved conflicting results: King et al. found that skinfolds and BIA incorrectly estimated FFM in 76 adults with CF ; Wells et al. found that the equation of The analysis by pubertal status indicated that the difference Slaughter et al. based on skinfold measurements had adequate compared to DXA is related to sexual maturation. Mean percent accuracy to estimate body composition in 55 pediatric patients relative bias for BF% was greater in pre-pubertal and pubertal with CF ; Beaumesnil et al. found that, despite good patients compared to post-pubertal (− 26%, − 18% and − 4% correlations between skinfold measurements or BIA and DXA, respectively, P b 0.0001). Similarly mean percent relative skinfolds and BIA overestimated FFM. difference for FFM obtained from BIA was different among In our study we found that the agreement with DXA was levels of sexual maturation (− 3% in pre-pubertal patients, + 3% affected by age and sexual maturation with greater bias in in pubertal and 0% in post-pubertal patients, P b 0.0001). pre-pubertal and pubertal compared to post-pubertal patients. The prediction of body composition in puberty is challenging, due to the large changes induced by puberty, and predictive 3.3. Predictors of body composition in cystic fibrosis equations based on skinfolds and BIA measurements may not be sufficiently accurate in patients who have not completed Age and gender significantly affected the ability of sexual maturation. Sexual development affects body composi- anthropometric and BIA indices to predict body composition tion and during puberty males gain greater amounts of fat free (Table 3). However, skinfolds of the upper arm were better mass and skeletal mass, whereas females acquire more fat mass. predictors than those of the trunk, explaining from 65% to 71% Both genders reach peak bone accretion during the pubertal of the variance in BF%, with the exception of older males in years, though males develop a greater skeletal mass. whom skinfolds of the trunk performed better. The ability of To address the second aim we identified predictors of body Z-score of BMI to predict BF% was very low, with the composition specific for CF patients. exception of females aged less than 17 years (56% of the We found that Z-score of BMI alone was not able to reliably variance in BF% explained). estimate BF% in CF patients, especially in those with low BMI. R-index was a good predictor of FFM in all age and gender This finding is consistent with previous studies carried out categories, explaining from 67% to 95% of the variance in either in healthy subjects and CF patients. Indeed, BMI has FFM. limited predictive value in healthy children and adolescents, A multiple linear regression model containing age, gender especially in subjects with low BMI , and BF% estimation and skinfolds of the upper arm performed better (adjusted R2 = significantly improves when BMI is combined with skinfold 0.77, RMSE = 3.1%) than the model containing Z-score of BMI thickness. Similar results were found in CF children where (adjusted R2 = 0.49, RMSE = 4.6%) or skinfolds of the trunk weight-based indicators (weight-for-age and weight-for-height (adjusted R2 = 0.56, RMSE = 4.2%) (Table 4). Z-scores) showed poor sensitivity in detecting body cell mass 788 G. Alicandro et al. / Journal of Cystic Fibrosis 14 (2015) 784–791 Fig. 1. Differences in body composition estimates between skinfold thicknesses, bioelectrical impedance analysis and DXA. Panels A and B show the Bland–Altman analysis of the differences between body fat percentage (BF%) predicted by equations based on skinfold thickness (Slaughter et al. and Durnin and Womersley) and DXA. Panels C and D show the Bland–Altman analysis of differences between fat free mass (FFM) predicted by equations based on BIA measurements and DXA. Continues lines are mean differences and dotted lines are upper and lower limits of agreements (mean ± 1.96 SD). Table 3 Predictors of body composition in cystic fibrosis. Predictor Body component Males Females b17 years ≥17 years b17 years ≥ 17 years BMI BF% Z-score 0.27* 0.31* 0.56** 0.18* SF BF% Biceps + Triceps 0.68*** 0.71*** 0.66*** 0.65*** Sub-scapular + Supra-iliac 0.29** 0.80*** 0.51*** 0.39*** 4-SF 0.53*** 0.86*** 0.63*** 0.54*** BIA FFM (kg) Resistance 0.53*** 0.52*** 0.44*** 0.24* Reactance 0.03 0.04 0.12* 0.01 Phase angle 0.19* 0.32** 0.12* 0.06 R-index 0.95*** 0.83*** 0.91*** 0.67*** Results are presented as R2 obtained by linear regression models. Abbreviations: BIA = bioelectrical impedance analysis; BF%: percentage of body fat; FFM = fat free mass; R-index = resistance index; SF = skinfold. *P b 0.05, **P b 0.01, ***P b 0.001.  G. Alicandro et al. / Journal of Cystic Fibrosis 14 (2015) 784–791 789 Table 4 Predictive equations of body composition in cystic fibrosis. Body component Equation adjusted R2 RMSE BF% 0.060 * age – 6.385 * gender + 3.254 * BMI + 24.824 0.49 4.6% − 0.025 * age – 2.168 * gender + 10.660 * ln (BSF + TSF) – 7.574 0.77 3.1% − 0.278 * age – 5.015 * gender + 7.936 * ln (SSF + SISF) + 4.684 0.56 4.2% − 0.200 * age – 3.561 * gender + 11.028 * ln (4 SF) – 13.689 0.73 3.3% FFM (kg) 0.769 * weight + 3.377 * gender – 0.492 0.93 2.99 kg 0.730 * weight + 2.899 * gender + 1.035 * phase angle – 4.740 0.94 2.84 kg 0.379 * weight + 0.917 * gender + 0.532 * R-index – 1.689 0.97 1.88 kg Results are presented as β, adjusted R2 and RMSE obtained by multiple linear regression models. Predictors: age (years), BMI (Z-score), gender (M = 1, F = 0), phase angle (°), R-index (cm2/Ω), skinfolds (mm), and weight (kg). Before including in the model SF measurements were natural log transformed. Abbreviations: BMI = body mass index; BSF = biceps skinfold thickness; R-index = resistance index (height2/resistance); SISF = supra-iliac skinfold thickness; SSF = sub-scapular skinfold thickness; TSF = triceps skinfold thickness; 4SF = sum of 4 skinfold thickness. reduction estimated by total body potassium count. Moreover in our study we tested the conventional single- Moreover, in healthy adults BMI alone explains only 25% of frequency BIA (SF-BIA) but we cannot exclude that more the variance in BF% and in CF adults King et al. found a sophisticated approaches, such as multi-frequency devices low ability of BMI in identifying FFM depletion evaluated by (MF-BIA) could perform better. Indeed such devices have DXA. Therefore, BMI provides little information about been tested in a small group of children (n = 28) with CF BF% in thin individuals and is a bad predictor of BF% in CF showing little advantage over SF-BIA , whereas there is patients, while skinfold measurements are highly correlated some evidence suggesting better accuracy of MF- and with BF% and should be included in models predicting body segmental BIA compared to SF-BIA in healthy subjects , composition in this population. in patients with stable heart failure and in malnourished Some characteristics of CF may affect estimates of body patients. So far there is no general consensus on the composition, including hydration status (frequently impaired in advantage of MF-BIA over SF-BIA, but further researches are these patients) and alteration of tissue distribution. A preferen- needed to verify the usefulness of MF-BIA analyzers in tial loss of FFM and BMC compared with FM was documented nutritional assessment of CF patients. in CF patients, and the FFM loss was found in arm and leg Although we have shown that skinfolds and SF-BIA do not rather than in the trunk, suggesting a depletion of skeletal provide reliable estimate of body composition in CF, an open muscle mass [31, 32]. question of clinical relevance still remains about their usefulness The specific pattern of body composition in CF has in detecting changes in body composition. Actually the lack of important implications in the selection of predictors of body sufficient data in CF and the inconsistent results in other composition. Our results suggest that skinfolds of the upper conditions do not allow a definitive answer. Studies in obese arm are better predictors of %BF compared to those of the children and adults showed conflicting results [40, 41], but trunk. Indeed, anthropometric measurements of the limbs could there is little evidence of the utility of BIA during development better reflect the specific CF pattern of body composition and growth , losses with illness or gain after compared to those of the trunk. Similarly to what has been nutritional intervention in adolescents with anorexia nervosa shown in previous studies on healthy subjects, R-index is a. In a group of 31 CF children Quirk et al. suggested that good predictor of FFM and when combined with body weight SF-BIA is only marginally better than anthropometric mea- its ability to predict FFM improves. sures in predicting total body potassium in individual patients When interpreting our findings it should be considered that. These contrasting results may have been caused by different any in vivo technique for body composition analysis has some time periods between measurements and the use of predictive limitations related to their underlying assumptions. Age, body equations not specific for the population under study. Indeed, small size, fatness, level of sexual maturation and disease state may short term changes cannot be detected due to the minimum significantly affect DXA assumptions and may have impacted detectable change specific of each technique , which for BIA our findings. In fact, DXA underestimates %BF in leanest and skinfolds depends mostly on the precision of the predictive subjects as compared to the reference 4-component model of equation used. body composition and significant differences were observed In conclusion, the poor agreement at individual level between among DXA scanning systems [34, 35]. In this context, a skinfolds or BIA and DXA prevents their use in clinical setting 4-component model of body composition combining measure- while the better agreement at population level suggests their ments of total body water, body density, and total body bone usefulness in population based studies. The more relevant mineral to estimate %BF could have provided more accurate implication of this finding is that body composition estimates data. However DXA is considered clinically valuable because provided by these techniques cannot be part of the standard of its strong relationship with other criterion methods, relative nutritional assessment of CF patients until reliable CF-specific low cost and ease of use. predictive equations become available. 790 G. Alicandro et al. / Journal of Cystic Fibrosis 14 (2015) 784–791 Our results support the additional value of body composition Tanner JM. Physical growth and development. In: Forfar JO, Arnell CC, to weight-based anthropometric measurements for the assessment editors. Textbook of pediatrics. 2nd ed. Edinburgh, Scotland: Churchill Livingstone; 1978. p. 249–303. of nutritional status in CF. Indeed, BMI has limited value in Lohman TG, Roche AF, Martorell R. Anthropometric standardization predicting body fatness in CF patients and should be used in reference manual. Champaign, IL: Human Kinetics; 1988. combination with other predictors of body composition, such as Slaughter MH, Lohman TG, Boileau RA, Horswill CA, Stillman RJ, Van skinfolds of the upper arm and R-index. These predictors are Loan MD, et al. Skinfold equations for estimation of body fatness in children and youth. Hum Biol 1988;60:709–23. strongly related to body composition data obtained from DXA Durnin JV, Womersley J. Body fat assessed from total body density and and should be tested in a large CF population to develop specific its estimation from skinfold thickness: measurements on 481 men and predictive equations. women aged from 16 to 72 years. Br J Nutr 1974;32:77–97. Bianchi ML, Romano G, Saraifoger S, Costantini D, Limonta C, Colombo C. BMD and body composition in children and young patients affected by Funding statement cystic fibrosis. J Bone Miner Res 2006;21:388–96. Azcue M, Fried M, Pencharz PB. Use of bioelectrical impedance analysis The study was supported by a grant from the Italian Cystic to measure total body water in patients with cystic fibrosis. J Pediatr Fibrosis Research Foundation – ONLUS (Prog. FFC #21/2013). Gastroenterol Nutr 1993;16:440–5. Borowitz D, Conboy K. Are bioelectric impedance measurements valid in patients with cystic fibrosis? J Pediatr Gastroenterol Nutr 1994;18:453–6. Author contributions de Meer K, Gulmans VA, Westerterp KR, Houwen RH, Berger R. Skinfold measurements in children with cystic fibrosis: monitoring fat-free mass and exercise effects. Eur J Pediatr 1999;158:800–6. G. Alicandro designed the study; G. Alicandro, S. Loi, C. Hollander FM, De Roos NM, De Vries JH, Van Berkhout FT. Assessment Speziali, and A. Bisogno conducted research; G. Alicandro of nutritional status in adult patients with cystic fibrosis: whole-body analyzed data; G. Alicandro and C. Colombo wrote the paper; bioimpedance vs body mass index, skinfolds, and leg-to-leg bioimpedance. J M.L. Bianchi supervised all DXA scans and analysis, and Am Diet Assoc 2005;105:549–55. performed evaluation of BMD Z-scores; G. Alicandro, C. Lands LC, Gordon C, Bar-Or O, Blimkie CJ, Hanning RM, Jones NL, et al. Comparison of three techniques for body composition analysis in Colombo, A. Battezzati and M.L. Bianchi contributed to the cystic fibrosis. J Appl Physiol 1993;75:162–6. interpretation of the data and reviewed the manuscript for Spicher V, Roulet M, Schaffner C, Schutz Y. Bio-electrical impedance intellectual content. All authors approved the final manuscript. analysis for estimation of fat-free mass and muscle mass in cystic fibrosis patients. Eur J Pediatr 1993;152:222–5. Ziai S, Coriati A, Chabot K, Mailhot M, Richter MV, Rabasa-Lhoret R. References Agreement of bioelectric impedance analysis and dual-energy X-ray absorptiometry for body composition evaluation in adults with cystic Ionescu AA, Nixon LE, Shale DJ. Cellular proteolysis and systemic fibrosis. J Cyst Fibros 2014;13:585–8. inflammation during exacerbation in cystic fibrosis. J Cyst Fibros 2004;3: King S, Wilson J, Kotsimbos T, Bailey M, Nyulasi I. Body composition 253–8. assessment in adults with cystic fibrosis: comparison of dual-energy X-ray Ionescu AA, Nixon LS, Luzio S, Lewis-Jenkins V, Evans WD, Stone MD, absorptiometry with skinfolds and bioelectrical impedance analysis. et al. Pulmonary function, body composition, and protein catabolism in Nutrition 2005;21:1087–94. adults with cystic fibrosis. Am J Respir Crit Care Med 2002;165:495–500. Wells GD, Heale L, Schneiderman JE, Wilkes DL, Atenafu E, Coates AL, King SJ, Nyulasi IB, Bailey M, Kotsimbos T, Wilson JW. Loss of fat-free et al. Assessment of body composition in pediatric patients with cystic mass over four years in adult cystic fibrosis is associated with high serum fibrosis. Pediatr Pulmonol 2008;43:1025–32. interleukin-6 levels but not tumour necrosis factor-alpha. Clin Nutr 2014; Beaumesnil M, Chaillou E, Wagner AC, Rouquette A, Audran M, Giniès JL. 33:150–5. Body composition analysis in patients with cystic fibrosis. Comparison of 3 Enright S, Chatham K, Ionescu AA, Unnithan VB, Shale DJ. The methods: dual energy x-ray absorptiometry, bioelectrical impedance influence of body composition on respiratory muscle, lung function and analysis, and skinfold measurements. Arch Pediatr 2011;18:370–5. diaphragm thickness in adults with cystic fibrosis. J Cyst Fibros 2007; Loomba-Albrecht LA, Styne DM. Effect of puberty on body composition. 30(6):384–90. Curr Opin Endocrinol Diabetes Obes 2009;16:10–5. King SJ, Nyulasi IB, Strauss BJ, Kotsimbos T, Bailey M, Wilson JW. Fat- Demerath EW, Schubert CM, Maynard LM, Sun SS, Chumlea WC, free mass depletion in cystic fibrosis: associated with lung disease severity Pickoff A, et al. Do changes in body mass index percentile reflect changes but poorly detected by body mass index. Nutrition 2010;26:753–9. in body composition in children? Data from the Fels Longitudinal Study. McNaughton SA, Shepherd RW, Greer RG, Cleghorn GJ, Thomas BJ. Pediatrics 2006;117:e487–95. Nutritional status of children with cystic fibrosis measured by total body Freedman DS, Wang J, Ogden CL, Thornton JC, Mei Z, Pierson RN, et al. potassium as a marker of body cell mass: lack of sensitivity of The prediction of body fatness by BMI and skinfold thicknesses among anthropometric measures. J Pediatr 2000;136:188–94. children and adolescents. Ann Hum Biol 2007;34:183–94. Stapleton D, Kerr D, Gurrin L, Sherriff J, Sly P. Height and weight fail to Gallagher D, Visser M, Sepúlveda D, Pierson RN, Harris T, Heymsfield detect early signs of malnutrition in children with cystic fibrosis. J Pediatr SB. How useful is body mass index for comparison of body fatness across Gastroenterol Nutr 2001;33:319–25. age, sex, and ethnic groups? Am J Epidemiol 1996;143:228–39. Sermet-Gaudelus I, Bianchi ML, Garabédian M, Aris RM, Morton A, Bolton CE, Ionescu AA, Evans WD, Pettit RJ, Shale DJ. Altered tissue Hardin DS, et al. European cystic fibrosis bone mineralisation guidelines. distribution in adults with cystic fibrosis. Thorax 2003;58:885–9. J Cyst Fibros 2011;10(Suppl. 2):S16–23. Alicandro G, Bisogno A, Battezzati A, Bianchi ML, Corti F, Colombo C. Heymsfield S, Lohman T, Wang Z, Going S. Human body composition 2nd Recurrent pulmonary exacerbations are associated with low fat free mass ed.; 2005. and low bone mineral density in young adults with cystic fibrosis. J Cyst Kuczmarski RJ, Ogden CL, Guo SS, Grummer-Strawn LM, Flegal KM, Fibros 2014;13:328–34. Mei Z, et al. 2000 CDC growth charts for the United States: methods and Kyle UG, Bosaeus I, De Lorenzo AD, Deurenberg P, Elia M, Manuel development. Vital Health Stat 11 2002;246:1–190. Gómez J, et al. Bioelectrical impedance analysis part II: utilization in Patient Registry Annual Data Report. Cystic Fibrosis Foundation; 2013. clinical practice. Clin Nutr 2004;23:1430–53.  G. Alicandro et al. / Journal of Cystic Fibrosis 14 (2015) 784–791 791 Sopher AB, Thornton JC, Wang J, Pierson Jr RN, Heymsfield SB, Horlick overweight and obesity in children and adolescents; a descriptive review. M. Measurement of percentage of body fat in 411 children and Nutr Hosp 2013;28:52–62. adolescents: a comparison of dual-energy X-ray absorptiometry with a Minderico CS, Silva AM, Keller K, Branco TL, Martins SS, Palmeira AL, four-compartment model. Pediatrics 2004;113:1285–90. et al. Usefulness of different techniques for measuring body composition Malouf J, Di Gregorio S, Del Rio L, Torres F, Marin AM, Farrerons J, changes during weight loss in overweight and obese women. Br J Nutr et al. Fat tissue measurements by dual-energy x-ray absorptiometry: cross- 2008;99:432–41. calibration of 3 different fan-beam instruments. J Clin Densitom 2013;16: Phillips SM, Bandini LG, Compton DV, Naumova EN, Must A. A 212–22. longitudinal comparison of body composition by total body water and Quirk P, Ward LC, Thomas BJ, Holt TL, Shepherd RW, Cornish BH. bioelectrical impedance in adolescent girls. J Nutr 2003;133:1419–25. Multiple frequency bioelectrical impedance for the prediction of total body Klauke S, Fischer H, Rieger A, Frühauf L, Staszewski S, Althoff PH, et al. potassium in cystic fibrosis. Clin Nutr 1995;14:348–53. Use of bioelectrical impedance analysis to determine body composition Demura S, Sato S, Kitabayashi T. Percentage of total body fat as estimated changes in HIV-associated wasting. Int J STD AIDS 2005;16:307–13. by three automatic bioelectrical impedance analyzers. J Physiol Anthropol Mika C, Herpertz-Dahlmann B, Heer M, Holtkamp K. Improvement of Appl Human Sci 2004;23:93–9. nutritional status as assessed by multifrequency BIA during 15 weeks of Alves FD, Souza GC, Biolo A, Clausell N. Comparison of two refeeding in adolescent girls with anorexia nervosa. J Nutr 2004;134: bioelectrical impedance devices and dual-energy X-ray absorptiometry 3026–30. to evaluate body composition in heart failure. J Hum Nutr Diet 2014;27: Quirk PC, Ward LC, Thomas BJ, Holt TL, Shepherd RW, Cornish BH. 632–8. Evaluation of bioelectrical impedance for prospective nutritional assess- Pirlich M, Schütz T, Ockenga J, Biering H, Gerl H, Schmidt B, et al. ment in cystic fibrosis. Nutrition 1997;13:412–6. Improved assessment of body cell mass by segmental bioimpedance Ellis KJ. Selected body composition methods can be used in field studies. analysis in malnourished subjects and acromegaly. Clin Nutr 2003;22: J Nutr 2001;131:1589S–95S. 167–74. de Miguel-Etayo P, Moreno LA, Iglesia I, Bel-Serrat S, Mouratidou T, Garagorri JM. Body composition changes during interventions to treat