Using Obese-Specific Charts of Height and Height Velocity for Assessment

Home , Human height

Using obese-speci®c charts of height and height velocity for assessment of growth in obese children and adolescents during weight excess reduction

A Falorni1*, G Galmacci2, V Bini1, F Faraoni3, D Molinari1, G Cabiati1, M Sanasi1, F Celi1, G Di Stefano1, MG Berioli1, G Contessa1 and ML Bacosi1

1 Clinica Pediatrica, Servizio Regionale di Diabetologia Pediatrica, UniversitaÁ di Perugia, Azienda Ospedaliera di Terni; 2 Dipartimento di Scienze Statistiche, UniversitaÁ di Perugia; and 3Divisione di Pediatria, Ospedale Civile, Rieti, Italy

Objective: The in¯uence of weight excess reduction on height and height velocity of obese subjects should be evaluated on the basis of appropriate standards, since the pattern of growth of obese subjects is different from that of normal weight subjects. Design, subjects and measurements: Height, weight and triceps skinfold thickness were recorded from 17987 school subjects (9256 males and 8731 females), 3 ± 18 y of age, from three provinces of central Italy, and a growth reference curve of height was constructed. Using BMI (as computed using the tables of Rolland-Cachera et al) and triceps skinfold thickness, normal-weight subjects (NWS) and obese subjects (OS) were identi®ed and speci®c reference curves (mean Æ s.d. every sixth month of age) were developed for both groups. Centiles of height were also calculated for OS. Various (2 ± 4) measurements of height in school subjects were performed and a graph of height velocity (HV) was constructed in NWS and in OS using the JPPS method. The yearly mean Æ s.d. of HV was also calculated, based on square root transformed data (in order to realise a Gaussian distribution), deriving from successive measurements in total subjects, in NWS and in OS. The z-scores of height and of the square root of HV were calculated in 217 obese subjects (125 males and 92 females) before and during a weight excess reduction programme (WERP). Obese subjects in WERP who showed a reduction of z-score of BMI were considered as `responsive'; those who either maintained or showed an increase of z-score of BMI were considered as `non-responsive'. Obese subjects in WERP were followed for 1 ± 4 y, giving the following results: 0 ± 1 y, 142 responsives and 75 non-responsives; 0 ± 2 y, 76 responsives and 33 non-responsives; 0 ± 3 y, 35 responsives and 30 non-responsives; 0 ± 4 y, 24 responsives and 18 non-responsives. Results: Compared to NWS, OS showed a signi®cantly greater HV in 4 ± 9 y males and in 4 ± 8 y females, but in older children the pubertal spurt was reduced and more precocious. As a result, the height of OS, which was greater in 3 ± 13 year-old males and in 3 ± 11.5 year-old females, subsequently showed a reduction, as compared to that of NWS, in 16 ± 18 year-old males and in 13 ± 18 year-old females. In both responsive and non-responsive groups of obese subjects in WERP, the z-scores of height showed a reduction during WERP when evaluated using the reference curve of the total school population. In contrast, when their growth was evaluated according to the obese-speci®c reference curve, no signi®cant variation was observed comparing both z-scores before and during the WERP. Conclusions: More appropriate information on the growth of obese subjects may be obtained when evaluating the height and HV according to obese-speci®c reference standards from the same population of origin. Adopting this modality, no signi®cant variation of height resulted during WERP in obese children. Descriptors: growth; obese adolescents; obese children

Introduction for ponderal excess in the child and adolescent seems justi®ed. However, the increasing prevalence of childhood Child and adolescent obesity tends to persist and to entail obesity can lead to an increase in the indiscriminate use of numerous risks in adulthood. A greater incidence of cardi- low-calorie diets. A careful evaluation of the effects of an ovascular pathologies and increased mortality have been absolute or relative ponderal reduction in the child and noted in subjects who were obese during adolescence and adolescent is therefore important. who were followed long-term (Must et al, 1992; Nieto et al, The effect of ponderal reduction on height growth has 1992; Must, 1996). As a consequence, corrective treatment been widely studied, but the results appear uncertain and contradictory. Various studies have reported a reduction in the speed of height when a ponderal reduction was obtained *Correspondence: Professor Adriano Falorni, Institute of Pediatrics, using low-calorie diets, so the use of dietary restrictions in University of Perugia, Hospital ``S. Maria'', 05100 Terni (Italy). Received 14 August 1998; revised 24 September 1998; accepted 12 the child has been advised against, particularly during the October 1998. pubertal spurt (Marshall, 1971; Brook et al, 1974; Archi- Growth in obese children and adolescents A Falorni et al 182 bald et al, 1983; Dietz & Hartung, 1985; Paige, 1986; the two speci®c reference curves, the NWS and OS selected Kneebone, 1990). The negative effects on growth appear re¯ected the described anthropometric criteria in all suc- analogous to those observed in states of undernutrition or cessive measurements. anorexia nervosa (Paige, 1986). However more recent Growth was examined in 217 obese subjects (125 males investigations have shown that moderate dietary restric- and 92 females) who attended the Paediatric Clinic of Terni tions associated with a correct control of patients have not and were included in a WERP. They came from the same had any negative in¯uence on the growth of obese children area of the school population studied for the construction of who have been followed for 10 years and whose heights the reference curves and, in part, they were recruited from a have been compared with those of their parents (Epstein et programme of diagnostic evaluation and therapy of a al, 1990, 1993). school population survey as reported in detail elsewhere When evaluating growth in obese subjects it is necessary (Falorni et al, 1997b; Falorni et al, 1998). Pathologies to bear in mind that their modality of stature growth differs affecting endocrine ± metabolic status, stature or ponderal from that of non-obese children. It has in fact been widely growth (with the exception of common obesity), or pubertal reported that obese children are taller, have an accelerated development were exclusion criteria. Height and HV of all growth rate and advanced bone age in pre-puberty, but these obese subjects were detected before WERP started, show a reduced height velocity and a less prominent spurt and the pubertal age was evaluated according to Tanner & during pubertal development (Bruch, 1941; Wolf, 1955; Whitehouse (1976). Modi®cation of eating behaviour was Garn & Haskell, 1959; Hammar et al, 1972; Garn & Clark, proposed, to reduce, hold or control the weight in order to 1975; Bernfeld & Dahan, 1976; Forbes, 1977; Laron et al, obtain a reduction of weight excess. Maintenance or con- 1978; Paige, 1986; Vignolo et al, 1988; Polito et al, 1995; trolled increase of weight were obtained by an educational De Simone et al, 1995; Falorni et al, 1997a). programme, by physical activity and by the correction of The aim of this research was to evaluate growth in a alimentary allowance. A balanced calorie-de®cient diet of large number of obese children and adolescents followed 5016 kJ=d (1200 kcal=d) or 20 ± 30% reduction of the for 1 ± 4 y at the Pediatric Clinic of Terni during a weight energy intake appropriate to age, and protein intake 1.5 ± excess reduction programme (WERP). To this end, a height 2.0 g=kg of ideal body weight per day was used for 2 ± 5 reference chart was constructed using data obtained from a months, when a reduction of the absolute value of weight school population-based survey in three provinces of cen- was advisable. Vitamin and mineral integration was tral Italy. An obese-speci®c reference curve was also adopted when appropriate. The obese subjects were seen constructed using the anthropometric data of school sub- at 2-week to 3-month intervals, adjustments of dietary jects, so it was possible to evaluate the growth of obese intake were made when necessary, and periodic laboratory children during the WERP by comparing it to the growth of examinations were performed. The follow-up period was obese subjects from the same population origin. 1 ± 4 y. The anthropometric data of these obese subjects were included in the construction of reference growth curves of height and HV only when collected before the Subjects and methods WERP started. Therefore, the data obtained during WERP To construct a growth curve, the height of 17987 school were not used for this purpose. Two separate groups of subjects (9256 males and 8731 females), 3 ± 18 year-old, obese subjects in WERP were formed as follows: subjects who participated in an extensive research project on the who showed a reduction of z-score of BMI > 0.25 during 1 growth of school children in three provinces of central Italy year of follow up, or a reduction of z-score of BMI > 0.50 (Perugia, Terni, Rieti), were collected from January 1993 to during 2 y of follow up, or a reduction of z-score of BMI January 1998. Height, weight and triceps skinfold thickness > 0.75 during 3 y of follow up, or a reduction of z-score of were taken during school hours by expert anthropometrists BMI > 1.0 during 4 y of follow up, were designated as of our group using the appropriate equipment and methods. `responsive'; subjects who showed a lower reduction or an Data were recorded on two occasions in all the 17987 increase of z-score of BMI during the various follow-up subjects, on three occasions in 6120 subjects, and on four periods were designated as `non responsive'. occasions in 2326 subjects. The interval between succes- The study was carried out with the collaboration of the sive measurements was 8 ± 12 months in subjects below three county Directors of Education and approved by the 10 ± 11 years of age, and 6 ± 10 months in older subjects. Ethics Committee of the Region of Umbria. Informed Height velocity (HV) (cm=y) was grouped according to the consent was obtained from the parents of the children mean age at the time of two successive measurements. included in this study. Speci®c curves for normal-weight subjects (NWS) and for obese subjects (OS) were constructed, separating the two different groups of subjects by the use of the following Statistical methods anthropometric criteria: Quetelet's index or Body Mass The seven-parameter (JPPS) asymptotic growth curve of Index (BMI) (weight in kilograms divided by the square Jolicoeur et al (1988) was ®tted to height data and the ®rst- of the height in meters), as computed using the table of derivative curve was used to represent the height velocity in Rolland-Cachera et al (1991), which are preferred as a both sexes in each group. To report the mean Æ s.d. of HV standard for European children by the members of Eur- data, their distribution was normalized by a square-root opean Childhood Obesity Group (ECOG) (Poskitt, 1995); transformation. ANOVA, Kolgomorov ± Smirnov test to triceps skinfold thickness, evaluated according to the tables assess the statistical requirement for meeting normality of of Must et al (1991). NWS were selected when their BMI distribution, paired and unpaired t-test, and partial correla- was between the 25th and 75th centile, and OS were tion coef®cient analysis were performed using Statgraphics selected when their BMI was > 95th centile. Triceps software (Graphic Software System, Rockville, MD, skinfold results were between 25th and 75th centile in USA). The level of statistical signi®cance was set at NWS and > 85th centile in OS. For the construction of P < 0:05. Growth in obese children and adolescents A Falorni et al 183 The distribution of BMI in a population is generally age In OS, HV was higher in younger children, but their dependent and tends to be positively skewed so that the pubertal spurt was reduced and precocious. Differences z-score cannot be calculated from the mean and s.d. Values existed between the two sexes, the pubertal height spurt of z-score of BMI were obtained for each age and sex group being earlier in females than in males in the two groups. using the LMS method of Cole (1988, 1990). The smoothed Mean Æ s.d. of HV1=2 of NWS and of OS are reported in L, M, and S values at six-month intervals were kindly Table 2. HV1=2 of OS was signi®cantly (P < 0.05) greater furnished by Dr M.F. Rolland-Cachera for the BMI in 5 ± 9 year-old males, and in 5 ± 8 year-old females. An reported in her paper (Rolland-Cachera et al, 1991). inversion with greater HV1=2 values (P < 0.01) in NWS was detected in 13 ± 15 year-old males and in 12 ± 13 year-old females. Results The anthropometric characteristics of the obese subject All the measurements of height of school subjects exam- submitted to WERP, subdivided into responsive and non- ined in this study were used to construct the height responsive groups, are reported in Table 3 and Table 4, reference charts. Height values had a satisfactory Gaussian respectively. Four subgroups were subdivided according to distribution in all the groups, subdivided according to sex the follow-up duration of WERP. The z-scores of height and age. Values (mean Æ s.d.) of height at 6-month inter- and of HV1=2 were computed either using the reference vals of the total school population, of NWS and of OS, are curves of total school subjects (z-score(1)) or using the reported in Table 1. Comparing NWS and OS, the height of obese-speci®c reference curves (z-score(2)). ANOVA did obese subjects was signi®cantly (P < 0:01) greater in 3 ± 13 not reveal differences among the four subgroups at the start year-old males and in 3 ± 11.5 year old females. In older of WERP. A signi®cant reduction of the z-score of height in subjects, the height of OS shifted progressively towards both responsive and non-responsive obese subjects was values lower than those of NWS, the differences being observed, during year 0 to 1 of WERP, when the total signi®cant (P < 0:05) in 16 ± 18 year-old males and in subjects reference curves were used (z-scores(1)). How- 13.5 ± 18 year-old females. Centiles of height of OS are ever, when the z-scores were estimated using the obese- reported in Figure 1. speci®c growth curves (z-scores(2)), no signi®cant varia- Height velocity (50th centile) of NWS and of OS, tions were observed during WERP in the various subgroups graphically reproduced by the JPPS method (Jolicoeur et of both responsive and non-responsive obese subjects al, 1988) are reported in Figure 2. This method allows the subdivided according to the duration of follow-up. The calculation of the age of minimum (pre-pubertal) and variations of the z-scores of height (differences between maximum (pubertal spurt) values of HV and ®nal height. before and after WERP) (mean Æ s.d.), in obese subjects

Table 1 Mean Æ s.d. of height in total subjects, normal-weight subjects and obese subjects from a school population subdivided according to sex and age

Total subjects Normal-weight subjects Obese subjects

Males Females Males Females Males Females Age (y) n Mean s.d. n Mean s.d. n Mean s.d. n Mean s.d. n Mean s.d. n Mean s.d.

2.751 ± 3.250 353 97.22 4.14 344 96.68 4.29 100 96.61 4.38 86 95.40 4.11 20 99.63 4.09 16 100.53 4.18 3.251 ± 3.750 360 101.29 4.39 359 100.58 4.49 149 100.76 4.51 129 99.36 4.29 21 103.77 4.33 23 104.50 4.36 3.751 ± 4.250 359 104.99 4.52 359 104.14 4.50 104 104.55 4.55 100 103.01 4.45 37 107.51 4.39 14 108.05 4.47 4.251 ± 4.750 309 108.42 4.63 310 107.47 4.51 99 108.07 4.57 120 106.43 4.57 36 110.96 4.47 21 111.30 4.68 4.751 ± 5.250 311 111.66 4.76 326 110.62 4.57 93 111.37 4.58 117 109.70 4.65 36 114.21 4.63 23 114.35 4.85 5.251 ± 5.750 388 114.76 4.93 339 113.67 4.67 124 114.49 4.59 107 112.85 4.69 55 117.31 4.87 44 117.27 4.97 5.751 ± 6.250 399 117.75 5.12 349 116.65 4.82 122 117.48 4.61 121 115.91 4.70 61 120.31 5.13 40 120.09 5.05 6.251 ± 6.750 852 120.68 5.32 870 119.59 4.99 154 120.34 4.71 146 118.90 4.72 67 123.25 5.37 53 122.87 5.13 6.751 ± 7.250 936 123.55 5.49 860 122.50 5.18 157 123.11 4.86 139 121.83 4.76 88 126.14 5.54 65 125.61 5.24 7.251 ± 7.750 1058 126.39 5.62 976 125.38 5.39 158 125.79 5.04 148 124.70 4.88 89 129.00 5.66 65 128.33 5.41 7.751 ± 8.250 1246 129.19 5.71 1131 128.25 5.63 201 128.40 5.20 183 127.52 5.07 132 131.82 5.71 96 131.34 5.61 8.251 ± 8.750 1353 131.97 5.77 1210 131.10 5.89 316 130.94 5.32 240 130.31 5.34 242 134.60 5.75 189 133.72 5.83 8.751 ± 9.250 1294 134.72 5.83 1221 133.96 6.17 261 133.45 5.42 233 133.08 5.67 239 137.34 5.78 169 136.34 6.03 9.251 ± 9.750 994 137.44 5.92 861 136.85 6.44 222 135.93 5.51 180 135.90 6.02 215 140.02 5.85 144 139.06 6.33 9.751 ± 10.250 1133 140.14 6.08 1107 139.80 6.88 266 138.44 5.62 242 138.79 6.55 251 142.65 5.98 190 141.72 6.53 10.251 ± 10.750 1514 142.84 6.32 1533 142.85 6.99 312 141.01 5.79 378 141.82 6.71 363 145.22 6.18 314 144.38 6.73 10.751 ± 11.250 1456 145.58 6.64 1334 145.97 7.18 276 143.72 6.05 256 145.00 6.88 335 147.76 6.42 220 147.01 6.24 11.251 ± 11.750 1046 148.40 7.02 957 149.12 6.96 319 146.63 6.41 351 148.25 7.03 302 150.32 6.69 241 149.56 6.11 11.751 ± 12.250 966 151.37 7.43 884 152.13 6.86 330 149.79 6.85 352 151.43 6.77 301 152.95 6.95 219 151.93 5.95 12.251 ± 12.750 1010 154.54 7.79 981 154.82 6.56 342 153.24 7.32 359 154.32 6.62 305 155.75 7.13 236 154.00 5.77 12.751 ± 13.250 985 157.91 8.05 928 157.04 6.34 396 156.89 7.76 385 156.72 6.40 277 158.73 7.21 175 155.70 5.61 13.251 ± 13.750 933 161.38 8.16 816 158.74 6.14 379 160.61 8.08 346 158.58 6.17 245 161.82 7.36 179 157.00 5.51 13.751 ± 14.250 797 164.77 8.06 694 159.96 6.00 324 164.16 8.19 323 159.91 5.99 188 164.80 6.98 116 157.94 5.49 14.251 ± 14.750 463 167.82 7.76 338 160.79 5.93 177 167.33 8.40 156 160.82 5.91 97 167.35 6.69 41 158.59 5.48 14.751 ± 15.250 414 170.34 7.30 316 161.33 5.90 168 169.96 7.66 165 161.43 5.89 84 169.30 6.35 34 159.02 5.45 15.251 ± 15.750 376 172.25 6.79 326 161.69 5.88 144 172.02 7.10 168 161.82 5.88 67 170.63 6.04 36 159.30 5.40 15.751 ± 16.250 363 173.62 6.46 360 161.91 5.85 135 173.56 6.72 206 162.07 5.86 77 171.48 5.90 31 159.48 5.35 16.251 ± 16.750 335 174.55 6.37 308 162.06 5.70 127 174.67 6.42 162 162.23 5.83 60 171.99 5.88 26 159.60 5.31 16.751 ± 17.250 330 175.16 6.29 374 162.15 5.47 130 175.45 6.31 187 162.33 5.68 46 172.28 5.85 26 159.68 5.30 17.251 ± 17.750 359 175.56 6.23 352 162.21 5.38 129 175.99 6.19 176 162.40 5.56 64 172.45 5.82 37 159.72 5.28 17.751 ± 18.250 313 175.81 6.17 292 162.25 5.37 122 176.36 6.05 162 162.44 5.41 48 172.55 5.76 18 159.76 5.25 Growth in obese children and adolescents A Falorni et al 184

Figure 1 Centiles (3rd, 10th, 25th, 50th, 75th, 90th, 97th) of height of school population obese subjects (OS) subdivided according to sex.

Table 2 Mean Æ s.d. of square root of height velocity (cm=year) in total subjects, normal-weight subjects, and obese subjects from a school population subdivided according to sex and age

Total subjects Normal-weight subjects Obese subjects

Males Females Males Females Males Females Age (y) n Mean s.d. n Mean s.d. n Mean s.d. n Mean s.d. n Mean s.d. n Mean s.d.

4.501 ± 5.500 306 2.54 0.23 312 2.55 0.22 40 2.60 0.16 35 2.57 0.16 40 2.72 0.17 30 2.69 0.20 5.501 ± 6.500 539 2.53 0.23 535 2.54 0.25 86 2.57 0.16 65 2.54 0.16 64 2.65 0.22 53 2.63 0.22 6.501 ± 7.500 455 2.42 0.34 505 2.41 0.32 102 2.40 0.29 91 2.37 0.29 69 2.52 0.31 40 2.50 0.29 7.501 ± 8.500 656 2.41 0.39 654 2.43 0.38 111 2.39 0.35 133 2.40 0.31 96 2.50 0.39 70 2.49 0.30 8.501 ± 9.500 1481 2.37 0.42 1495 2.44 0.42 306 2.36 0.38 265 2.46 0.43 281 2.45 0.44 236 2.43 0.39 9.501 ± 10.500 1404 2.34 0.43 1336 2.49 0.46 235 2.30 0.39 212 2.51 0.46 346 2.36 0.47 246 2.44 0.44 10.501 ± 11.500 1426 2.31 0.45 1411 2.52 0.48 264 2.26 0.42 293 2.53 0.49 366 2.32 0.49 305 2.46 0.52 11.501 ± 12.500 1898 2.4 0.48 1748 2.33 0.5 373 2.45 0.46 375 2.42 0.50 530 2.40 0.50 426 2.22 0.55 12.501 ± 13.500 2093 2.59 0.52 1832 1.97 0.52 537 2.67 0.52 435 2.11 0.52 572 2.48 0.50 393 1.86 0.60 13.501 ± 14.500 1027 2.5 0.57 657 1.68 0.54 288 2.57 0.58 177 1.67 0.55 240 2.36 0.54 118 1.77 0.64 14.501 ± 15.500 510 2.03 0.63 342 1.34 0.57 171 2.11 0.63 142 1.36 0.58 99 1.90 0.61 36 1.49 0.65 15.501 ± 16.500 324 1.7 0.68 237 1.26 0.61 86 1.77 0.64 98 1.26 0.61 63 1.71 0.77 21 1.35 0.62

during WERP, were evaluated according to general school Discussion population reference curve (z-score(1)) and according to the obese-speci®c reference curve (z-score(2)); the compar- The study was carried out on a large number of school isons are reported in Table 5. In all subgroups of subjects, subjects in a population located in a geographical area either responsive or non-responsive, the Á mean values corresponding to three provinces of central Italy. The data were signi®cantly greater when evaluated according to the relating to the heights of subjects who may be considered as total school subjects reference curve. obese from their anthropometric features were used to The correlation between the reduction of z-score of BMI construct an obese-speci®c growth curve and were com- during WERP (Áz-score of BMI), and either variation of pared to those of subjects having normal ponderal char- the z-scores(2) of height and of HV1=2, or absolute value of acteristics. With respect to the NWS, the OS during the z-score(2) of HV1=2, adjusted for sex, age, pubertal childhood and early adolescence showed greater height, stage and WERP duration, was computed using the partial associated with a greater HV, until the age of 9 in boys and correlation analysis. A signi®cant positive correlation until the age of 8 in girls. However, the pubertal growth (P < 0:05) was found only between the Áz-score of BMI spurt of OS was reduced. This latest evidence emerges from 1=2 1= and the absolute value of z-score(2) of the HV . a comparison of the averages of HV 2 (OS vs NWS) as Growth in obese children and adolescents A Falorni et al 185

Figure 2 JPPS height velocity curve (50th centile) in normal weight subjects (solid line) (NWS), and obese subjects (dotted line) (OS) from a school population, subdivided according to sex. In the panels, age of minimum (pre-pubertal) and maximum (pubertal peak value) values of height velocity, and the prediction of ®nal height in each of the two groups are reported. well as from curves constructed by the JPPS method earlier pubertal development in obese subjects (Wolf, (Jolicoeur et al, 1988). The peak height velocity in OS 1955; Garn & Haskell, 1959; Hammar et al, 1972; Garn was located earlier than in NWS: 12.5 year-old male OS vs & Clark, 1975; Bernfeld & Dahan, 1976; Forbes, 1977; 12.9 year-old male NWS, and 10.9 year-old female OS vs Vignolo et al, 1988; Frisch, 1985; De Simone et al, 1995; 11.2 year-old female NWS. This pattern agrees with those Falorni et al, 1997a). The height of OS was signi®cantly from previous observations of more rapid growth and lower than that of NWS > 16.0 y of age in males and over

Table 3 Pubertal stage, age and anthropometrics in obese children and adolescents (responsive) subdivided according to weight excess reduction programme (WERP) duration

WERP duration (y) 1 2 3 4 n(<=,) 142(87=55) 76(45=31) 35(19=16) 24(11=13) Tanner stage(<=,) I 59=30 27=18 13=10 8=7 II 17=7 10=5 5=3 2=4 III 7=8 4=6 0=2 1=2 IV ± V 4=10 4=2 1=1 0=0 Age (y)(<=,)a 10.8 Æ 2=9.9 Æ 2 11.1 Æ 2=9.7 Æ 1.8 11.1 Æ 1.7=9.2 Æ 2 9.8 Æ 1.8=9 Æ 2.3

Before After Before After Before After Before After

Height (cm)a 144.9 150.3 144.2 155.3 142 158.1 135.8 155.2 Æ 12.4 Æ 12.3 Æ 12.1 Æ 12 Æ 11.2 Æ 10.1 Æ 12.6 Æ 8.2 z-score(1) of heighta,b 0.37* 0.28 0.16 0.1 0.13 0 0.02 70.15 Æ 1.09 Æ 1.09 Æ 1.01 Æ 0.97 Æ 0.87 Æ 0.7 Æ 0.78 Æ 0.51 z-score(2) of heighta,b 0.09 0.1 7 0.12 0.02 7 0.2 0.04 7 0.35 7 0.03 Æ 1.1 Æ 1.12 Æ 1.03 Æ 0.96 Æ 0.85 Æ 0.74 Æ 0.82 Æ 0.48 z-score(1) of square root of height 7 0.10 7 0.17 7 0.1 7 0.08 7 0.16 7 0.13 7 0.26 7 0.24 velocitya,b Æ 0.81 Æ 0.75 Æ 0.6 Æ 0.55 Æ 0.54 Æ 0.57 Æ 0.47 Æ 0.32 z-score(2) of square root of height 7 0.17 7 0.14 7 0.05 0 0.01 7 0.04 7 0.23 7 0.21 velocitya,b Æ 0.78 Æ 0.76 Æ 0.66 Æ 0.59 Æ 0.53 Æ 0.58 Æ 0.5 Æ 0.41 Weight (kg)c 55.7 55.7 55.1 59.5 55.4 60.4 46 59.2 45.4 ± 64.4 45.3 ± 65.4 45.2 ± 64.3 47.8 ± 67.9 40.8 ± 58 53 ± 67.8 36.6 ± 59.1 49.9 ± 67.8 BMI (kg=m2)c 25.8 23.8 25.5 23.7 25.3 23 25.7 23.9 24 ± 28.7 22.1 ± 26.4 23.7 ± 28.6 21.8 ± 26.4 22.5 ± 27.4 21.8 ± 25.5 23.2 ± 27.2 21.6 ± 25.9 z-score of BMIa 3.91 2.93 3.77 2.46 3.79 2.18 4.07 2.14 Æ 1.04 Æ 0.97 Æ 1.07 Æ 0.85 Æ 1.01 Æ 0.71 Æ 0.98 Æ 0.79 a Mean Æ s.d. c Median and interquartile range. b z-score(1) referred to total population reference curve; z-score(2) referred to obese speci®c curve. * P < 0:01 before vs after WERP. Growth in obese children and adolescents A Falorni et al 186 Table 4 Pubertal stage, age and anthropometrics in obese children and adolescents (non-responsive) subdivided according to weight excess reduction programme (WERP) duration

WERP duration (y) 1 2 3 4 n(<=,) 75(38=37) 33(21=12) 30(17=13) 18(10=8) Tanner stage(<=,) I 34=23 19=8 12=8 8=7 II 4=8 1=1 4=2 2=1 III 0=2 1=1 1=2 0=0 IV ± V 0=4 0=2 0=1 0=0 Age (y)(<=,) a 10.1 Æ 2=9.8 Æ 2.3 10.5 Æ 2.1=9.6 Æ 2.4 10.3 Æ 1.7=8.8 Æ 2.2 10.4 Æ 1.6=8.2 Æ 0.3

Before After Before After Before After Before After

Height (cm)a 139.9 145 139.7 149.2 138.2 153.8 136.7 159.5 Æ 12.4 Æ 12.2 Æ 11.9 Æ 12 Æ 12.8 Æ 12.1 Æ 8.6 Æ 8 z-score(1) of heighta,b 7 0.2* 7 0.54 0.05 7 0.01 7 0.04 7 0.22 7 0.26 7 0.49 Æ 1.21 Æ 1.15 Æ 1.1 Æ 1.09 Æ 1.36 Æ 1.2 Æ 1.28 Æ 1.31 z-score(2) of heighta,b 7 0.27 7 0.24 7 0.53 7 0.71 7 0.41 7 0.27 7 0.67 7 0.42 Æ 1.13 Æ 1.13 Æ 1.18 Æ 1.18 Æ 1.39 Æ 1.34 Æ 1.27 Æ 1.46 z-score(1) of square root of height 7 0.32 7 0.36 7 0.46 7 0.43 7 0.3 7 0.32 7 0.19 7 0.17 velocitya,b Æ 0.75 Æ 0.68 Æ 0.59 Æ 0.56 Æ 0.44 Æ 0.36 Æ 0.45 Æ 0.38 z-score(2) of square root of height 7 0.4 7 0.37 7 0.41 0.4 0.25 7 0.29 7 0 7 0.08 velocitya,b Æ 0.82 Æ 0.72 Æ 0.66 Æ 0.58 Æ 0.48 Æ 0.36 Æ 0.51 Æ 0.39 Weight (kg)c 47.3 53.1 50 57 45.4 63.3 41.8 67.7 41 ± 58 45 ± 65 38 ± 66 45.4 ± 79.6 39.7 ± 58 56.2 ± 74.2 37.2 ± 54.2 60.5 ± 80.8 BMI (kg=m2)c 24.7 25.5 24.7 26.3 25.3 27.5 23.4 27.5 22.4 ± 26.8 23.3 ± 28.1 22.3 ± 28.1 23.4 ± 30.9 22.9 ± 27.7 25.5 ± 29.3 22.1 ± 26.3 24.8 ± 29.7 z-score of BMIa 3.7 3.5 3.7 3.4 4 3.5 3.7 3.2 Æ 0.9 Æ 0.8 Æ 1 Æ 0.8 Æ 1.3 Æ 0.9 Æ 0.9 Æ 1.1

a Mean Æ s.d. c Median and interquartile range. b z-score(1) referred to total population reference curve; z-score(2) referred to obese speci®c curve. * P < 0:01 before vs after WERP.

13.5 y of age in females. This study therefore shows that evaluation on the basis of this element. We nevertheless the OS of the studied population had lower values of height feel con®dent that given the large number of subjects in late adolescence and of height prediction in adulthood, examined, the in¯uence of such factors was considerably than those of NWS and of the general population. Similar reduced. results were reported by De Simone et al (1995) and by The question of which reference standard must be used Vignolo et al (1988) even though the anthropometric data for the assessment of growth of obese children clearly of their obese children did not derive from a population emerges from the results of this research. Statistical analy- study and were estimated according to the growth reference sis demonstrated that the two reference curves (general curves of Tanner & Whitehouse (1976). population and obese-speci®c) constructed on the popula- We have evaluated growth in a large number of obese tion studied in this research were not compatible when the subjects submitted to dietary treatment to reduce pond- z-scores for height were calculated. Recent studies on the eral excess. In contrast to other similar studies, the obese growth of obese subjects under dietary treatment had subjects in WERP of this research were examined on the already pointed out that the use of reference curves of basis of an obese-speci®c growth chart constructed on the growth of the general population can induce erroneous same population of origin. The interpretation of the deductions (Epstein et al, 1990, 1993). In this study, we results of this study needs to bear in mind certain have demonstrated that when the growth of the responsive limits imposed by the way in which the anthropometric obese subjects in the WERP was evaluated according to the data were collected. The growth curves of the school reference curve of the general school population, a reduc- subjects do not respond precisely either to the requisites tion was observed in the z-score of the height, even though of a cross-sectional or a longitudinal curve. Furthermore, it was signi®cant only during the ®rst year of follow-up. A the values of the pubertal growth spurt were not cor- signi®cant reduction was also found in the non-responsive rected for early and late maturers and therefore are obese subjects in WERP, on whom the dietary treatment ¯attened. On the other hand, the curves were used to may have had little or no in¯uence on growth. However, compare different categories of the same population when the growth of these obese subjects was evaluated on studied and served as a point of reference in order to the basis of an obese-speci®c reference curve (z-score(2) of evaluate growth in obese subjects in treatment, whose Tables 3 and 4) no signi®cant variation in the z-scores of anthropometric data were also collected during the course height or of HV during the 1 ± 4 y of the WERP was found of a follow-up varying from 1 ± 4 y. We must also in either groups (responsive and non-responsive). Also the remember that growth differences in obese children Áz-score of height during WERP (the difference between have been reported in relation to the age when obesity before and after) signi®cantly increased when comparing z- arose (Bernfeld & Dahan, 1976). Our study lacks ana- score(1) (on the basis of the general population reference mnestic data relating to the age of onset of obesity in OS curve) vs z-score(2) (on the basis of obese-speci®c refer- and so it was not possible to make a differentiated ence curve). Growth in obese children and adolescents A Falorni et al 187 In conclusion, our data con®rm that, in general, a WERP 8) nce = correctly constructed and conducted with clinical and 0.637 0.245 0.570 4 0.226* Æ Æ laboratory controls, does not have a signi®cant in¯uence refere 18(10 7 on the growth of obese children. This result appears to be ects con®rmed also by the analysis of the partial correlation subj between the reduction of BMI and z-scores (evaluated according obese-speci®c growth reference curves) of hool 1 2 4 8 1* sc

13) height and of HV /2. Our data agree with those reported = ects 0.51 0.13 0.48 3 0.18 by Epstein et al (1990, 1993), which emerged from a total subj Æ Æ 7

30(17 lengthy, longitudinal study comparing the parental heights to g with those of obese subjects undergoing a dietary treat-

obese ment. Our study provides the possibility of the assessment of the growth of obese subjects during the WERP using a accordin nsive model constructed from the obese subjects of the popula- 0 5 7 6* 12)

= tion of origin. But we must keep in mind that an absolute or n-respo aluated 0.57 0.17 2 0.55 0.34 relative reduction in weight induces a reduction of fat-free ev No Æ Æ 33(21 mass, even when dietary protein allowance is high. For this reason, a note of caution is necessary, especially when one subjects confronts the critical phase of growth and pubertal development during which body mass tends to double in size obese (Paige, 1986). Moreover, one cannot exclude negative 5 4 6 0* 37) in = in¯uences of weight reduction on the growth of a single P) 0.26 0.03 1 0.26 0.06 Æ Æ subject. Therefore, it remains the job of health care profes- 7 75(38 sionals to weigh the advantages of a reduction in weight (WER excess against a possible negative outcome resulting from me theconstructionofweightreductionprogrammesforchildren. program 13) = 0.884 0.315 4 0.776 0.176* Acknowledgements ÐWe acknowledge the Directors of Education of Æ Æ

7 Perugia, Terni and Rieti counties for collaboration and Dr M.F. Rolland- 24(11

reduction Cachera, INSERM, Section Nutrition, 78110 Le VeÂsinet, France, who kindly furnished the tables of L, M and S values to calculate the SDS of BMI. We also thank Miss Myrna Goldstein for her valuable help in excess t reviewing this manuscript. 7 9 4 6* 16) = weigh g 0.66 0.23 3 0.64 0.12 Æ Æ 7 35(19 subjects durin References

s.d.) Archibald EH, Harrison JE & Pencharz PB (1983): Effects of a weight- obese Æ reducing high-protein diet on the body composition of obese adoles- ve cents. Am. J. Dis. Child. 137, 658 ± 662. nsi 8 5 4 1* 31) (mean Bernfeld J & Dahan O (1976): Croissance et deÂveloppement de l'enfant = obeÁse. Etude longitudinale. Rev. Fr. Endocrinol. Clin. 17, 141 ± 164. 0.14 0.42 2 0.39 0.10 Respo Brook CAD, Lloyd JK & Wolff OH (1974): Rapid weight loss in children. Æ Æ e(2)) 7 76(45 -score)

z Br. Med. J. 2, 44 ± 45. Á

-scor Bruch H (1941): Obesity in relation to puberty. J. Pediatr. 19, 365 ± 375. ( z (

ht Cole TJ (1988): Fitting smoothed centile curves to reference data (with discussion). J. R. Sratist. Soc. A. 151, 385 ± 418. heig

curve Cole TJ (1990): The LMS method for constructing normalized growth of 8 2 0 8* standards. Eur. J. Clin. Nutr. 44, 45 ± 60. 55) nce =

7 De Simone M, Farello G, Palumbo M, Gentile T, Ciuffreda M, Olioso P, 0.01 0.28 1 0.27 0.08 cores

height. Cinque M & De Matteis F (1995): Growth charts, growth velocity and -s refere Æ Æ 7 z of

142(8 bone development in childhood obesity. Int. J. Obes. Relat. Metab. c Disord. 19, 851 ± 856. the eci® e(2) Dietz WH Jr & Hartung R (1985): Changes in height velocity of obese of preadolescents during weight reduction. Am. J. Dis. Child. 139, 705 ± -scor ions z 707. Á obese-sp Epstein LH, McCurley J, Valoski A & Wing RR (1990): Growth in obese vs to variat children treated for obesity. Am. J. Dis. Child. 144, 1360 ± 1364. after after ht n Epstein LH, Valoski A & McCurley J (1993): Effect of weight loss by obese ding and and children on long-term growth. Am. J. Dis. Child. 147, 1076 ± 1080. heig Falorni A, Bini V, Cabiati V, Papi F, Arzano F, Celi F & Sanasi M betwee of accor s (1997a): Serum levels of type I procollagen C-terminal propeptide, before before insulin-like growth factor-I (IGF-I), and IGF-binding protein-3 in obese and ight ight ore(1) (y) children and adolescents: relationship to gender, pubertal development, he he tween tween -sc z mparison growth, insulin and nutritional status. Metabolism 46, 862 ± 871. e(1)) of of be be tion Á

Co Falorni A, Bini V, Molinari D, Papi F, Celi F, Di Stefano G, Berioli G, 2) 1) 01 -scor Bacosi ML & Contessa G (1997b): Leptin serum levels in normal : dura z 5 0 ( ) rences P) rences P) weight and obese children and adolescents: relationship with age, sex, , < = -score( -score( ble pubertal development, body mass index, and insulin. Int. J. Obes. Relat. < z z P Ta * (diffe WER Á curve n( WERP (diffe WER Á Metab. Disord. 21, 881 ± 890. Growth in obese children and adolescents A Falorni et al 188 Falorni A, et al (1998): Fasting serum leptin levels in the analysis of body Must A, Dallal GE & Dietz WH (1991): Reference data for obesity: 85th mass index cut-off values: are they useful for overweight screening in and 95th percentiles of body mass index (wt=ht2) and triceps skinfold children and adolescents? A school population-based survey in three thickness. Am. J. Clin. Nutr. 53, 839 ± 846. provinces of central Italy. Int. J. Obes. Relat. Metab. Disord. (in press). Must A, Jacques PI, Dallal G, Bajema C & Dietz WH (1992): Long-term Forbes GB (1977): Nutrition and growth. J. Pediatr. 91, 40 ± 42. morbidity and mortality of overweight adolescents: a follow-up of the Frisch RE (1985): Fatness, menarche, and female fertility. Perspect. Biol. Harvard Growth Study of 1922 to 1935. N. Engl. J. Med. 327, 1360 ± Med. 28, 611 ± 633. 1365. Garn SM & Haskell JA (1959): Fat and growth during childhood. Science Nieto FJ, Szaklo M & Comstock GW (1992): Childhood weight and 130, 1711 ± 1712. growth rate as predictors of adult mortality. Am. J. Epidemiol. 136, Garn S & Clark DC (1975): Nutrition, growth, development and matura- 201 ± 213. tion: ®ndings from the ten-state nutrition survey of 1968 ± 1970. Paige DM (1986): Obesity in childhood and adolescence. Postgrad. Med. Pediatrics 56, 306 ± 319. 79, 233 ± 245. Hammar SL, Campbell MM, Campbell VA, Moores NL, Sareen C, Gareis Polito C, Di Toro A, Collini R, Cimmaruta E, D'Alfonso C & Del Giudice FJ & Lucas B (1972): An interdisciplinary study of adolescent obesity. G (1995): Advanced RUS and normal carpal bone age in childhood J. Pediatr. 80, 373 ± 383. obesity. Int. J. Obes. Relat. Metab. Disord. 19, 506 ± 507. Jolicoeur P, Pontier J, Pernin MO & SempeÂ M (1988): A life Poskitt EME (1995): De®ning childhood obesity: the relative body mass time asymptotic growth curve of human height. Biometrics 44, 995 ± index (BMI). Acta Pñdiatr. 84, 961 ± 963. 1003. Rolland-Cachera MF, Cole TJ, SempeÂ M, Tichet J, Rossignol C & Kneebone GM (1990): Childhood obesityÐthe diagnosis and manage- Charraud A (1991): Body Mass Index variations: centiles from birth ment. Aust. Fam. Physician 144, 1360 ± 1364. to 87 years. Eur. J. Clin. Nutr. 45, 13 ± 21. Laron Z, Ben-Dan, Shrem M, Dickerman Z & Lilos P (1978): Puberty in Tanner JM & Whitehouse RH (1976): Clinical longitudinal standards for simple obese boys and girls. In Obesity in Childhood ed. E Cacciari, Z height, weight, height velocity, weight velocity and stages of puberty. Laron & S Raiti, pp 29 ± 40. London: Academic Press. Arch. Dis. Child. 51, 170 ± 176. Marshall WA (1971): Evaluation of growth rate in height over periods of Vignolo M, Naselli A, Di Battista E, Mostert M & Aicardi G (1988): less than one year. Arch. Dis. Child. 46, 414 ± 420. Growth and development in simple obesity. Eur. J. Pediatr. 147, 242 ± Must A (1996): Morbidity and mortality associated with elevated body 244. weight in children and adolescents. Am. J. Clin. Nutr. 63 (Suppl.), Wolf OH (1955): Obesity in childhood: a study of birth weight, the height S445 ± S447. and onset of puberty. Q. J. Med. 24, 109 ± 118.