APPLIED NUTRITIONAL INVESTIGATION INTRODUCTION MATERIALS AND METHODS

Transcription

1 APPLIED NUTRITIONAL INVESTIGATION Bioelectrical Impedance Vector in Pre- and Postmenarcheal Females Roberto Buffa, PhD, Giovanni Floris, MD, and Elisabetta Marini, PhD From the Department of Experimental Biology, Anthropological Science Section, University of Cagliari, Cagliari, Italy OBJECTIVE: To present what is to our knowledge the first study of the effects of puberty on the distribution of bioimpedance vectors. METHODS: Resistance and reactance (standard, tetrapolar analysis at 50-kHz frequency) were measured in 143 pre- and postmenarcheal Sardinian girls aged 10 to 15 y. RESULTS: The data agreed well with the reference values proposed for the Italian population. Bioelectrical impedance vector analysis showed a clear differentiation between pre- and postmenarcheal girls, whereas no significant age-related differences were observed within groups. Postmenarcheal girls had a shorter impedance vector (reduced R/H and Xc/H) and a greater phase angle than did premenarcheal girls. CONCLUSIONS: Sexual maturity is the major cause of differentiation in the body composition of adolescent females. We recommend that this source of variation be considered in clinical and anthropologic analyses of bioelectrical variables. Nutrition 2002;18: Elsevier Science Inc KEY WORDS: body composition, sexual maturity, Sardinia INTRODUCTION Impedance analysis of body composition constitutes a very promising application by allowing, among other things, the clinical examination and follow-up of patients with critical illness, renal and liver diseases, and cardiac failure. 1 A methodological variant of the classic procedure involves the direct use of bioelectrical parameters. 2 4 In the application proposed by Piccoli et al. 2, the values of resistance and reactance, standardized by stature, are plotted in a Cartesian plane (RXc graph). The bioelectrical parameters provide an evaluation of body composition because they are sensitive to variations of the state of hydration. The physiologic meaning of these measures is still the object of experimental research, and the observations are interpreted in terms of comparison with bioelectrical values deriving from healthy and ill individuals. 2,5,6 In this way, one can qualitatively assess the state of hydration combined with the structure and mass of soft tissues. Recent studies conducted on a large Italian sample have provided the first definition of normal bioimpedance values of adults 3 and children. 4 The proposed reference values were divided into specific age groups, with a subdivision between sexes from the age of 14 y. The bioelectrical vector showed a shortening trend with growth. A similar pattern was observed in a previous study. 7 Growth processes certainly involve quantitative and qualitative changes in body composition and thus bioimpedance parameters. Moreover, puberty, whose onset is extremely variable among individuals but generally occurs before age 14 y, is characterized by more radical and rapid changes. Therefore, it is important to isolate and study the behavior of bioelectrical parameters in relation to sexual development. We investigated the puberty effect on the bioelectrical vector. The problem was analyzed in a female sample, in which we were This research was supported by M.U.R.S.T. 40% and 60% funds. Correspondence to: Giovanni Floris, MD, Department of Experimental Biology, Anthropological Science Section, Cittadella Universitaria Monserrato, Monserrato (Cagliari), Italy. floris@unica.it able to identify a definite moment representative of puberty, i.e., menarche. We also evaluated the suitability of the normal Italian values in the population of Sardinia, a region that is geographically and culturally distinct from populations on the Italian peninsula. 8,9 MATERIALS AND METHODS The Sample We examined 143 girls, between 10 and 15 y of age, from the city of Cagliari and its environs, all daughters of Sardinian parents. All subjects were apparently healthy. The sampling was conducted in five schools believed to be similar in terms of the socioeconomic level of the territory in which they are located. The mean age at menarche determined by the retrospective method was y. Measurements All the measurements were taken by the same experienced operator (R.B.). Anthropometric variables were measured according to the International Biological Programme. 10 Stature was measured with a movable anthropometer to a precision of 0.1 cm, and weight was measured with a portable spring scale to a precision of 0.1 kg. Body mass index (BMI) was also calculated. Bioelectrical parameters of resistance (R, ) and reactance (Xc, ) were measured with a monofrequency impedance analyzer (BIA-101, RJL/Akern, Firenze, Italy) that emitted 800- A and 50-kHz alternating sinusoidal currents. The positioning of outer and inner electrodes was the standard one, and the entire procedure was performed according to the indications of the National Institutes of Health Technology Assessment Conference Statement. 1 The phase angle ( ) was obtained from the values of resistance and reactance according to the equation: arctan(xc/r). 11 The impedance measurements were standardized by stature (R/H, /m, and Xc/H, /m). The parameters R/H and Xc/H and the phase angle defined the bioelectrical impedance vector (Z, ). Nutrition 18: , /02/$22.00 Elsevier Science Inc., Printed in the United States. All rights reserved. PII S (02)

2 Nutrition Volume 18, Number 6, 2002 Bioelectrical Impedance and Menarche 475 TABLE I. DESCRIPTIVE STATISTICS AND ANALYSIS OF VARIANCE RESULTS FOR THE COMPARISON BETWEEN AGE GROUPS 10 11y(n 43) 12 y (n 20) 13 y (n 28) 14 15y(n 52) Mean SD Mean SD Mean SD Mean SD F Stature (cm) Weight (kg) BMI (kg/m 2 ) * R( ) Xc ( ) R/H ( /m) Xc/H ( /m) Phase angle ( ) r (R, Xc) Statistical Analyses Statistical analyses were performed with Statistica 4.0 (Statsoft Inc.). Software provided by A. Piccoli (BIVA Programs, release 1998) was used to draw confidence ellipses containing the Z mean vector. Graphs were made with AutoCAD (Autodesk). COMPARISON WITH THE NORMAL ITALIAN VALUES. Survey conditions were almost identical to those reported by De Palo et al., 4 and the sample was subdivided into the same age classes: 10 11, 12, 13, and y. We calculated the descriptive statistics for all the variables in each group and the coefficient of correlation between the parameters R/H and Xc/H. Agreement of the Sardinian values with the Italian standards was evaluated qualitatively by analysis of their distribution in the tolerance ellipses proposed by De Palo et al. 4 EFFECT OF MENARCHE. To find variations related to growth, analysis of variance (ANOVA) was applied to the sample subdivided by age groups. To evaluate the effect of sexual maturation on the bioelectrical parameters, we divided the overall sample into two subsamples based on the presence or absence of menarche at the time of the investigation. In each subsample, divided into 1-y age groups, we applied ANOVA to assess the possible persistence of age-dependent variations. Further, to compare the impedance variables in pre- and postmenarcheal girls, we applied Hotelling s T 2 test and we drew the 95% probability confidence ellipses. The bioelectrical menarche effect was thoroughly examined and its role was isolated from that of other concomitant causal factors, in particular body size (as expressed by BMI). Using the analysis of covariance, 12 we adjusted the bioelectrical values for the effect of BMI and performed an extra comparison between pre- and postmenarcheal girls. TABLE II. DESCRIPTIVE STATISTICS AND ANALYSIS OF VARIANCE RESULTS FOR THE COMPARISON BETWEEN AGE GROUPS IN THE PREMENARCHEAL GIRLS 10y(n 20) 11 y (n 15) 12 y (n 10) 13 y (n 7) 14 y (n 3) Mean SD Mean SD Mean SD Mean SD Mean SD F Stature (cm) Weight (kg) BMI (kg/m 2 ) R( ) XC ( ) R/H ( /m) Xc/H ( /m) Phase angle ( ) r(r, Xc) *

3 476 Buffa et al. Nutrition Volume 18, Number 6, 2002 TABLE III. DESCRIPTIVE STATISTICS FOR POSTMENARCHEAL GIRLS 11y(n 8) 12 y (n 10) 13 y (n 21) 14 y (n 29) 15 y (n 20) Mean SD Mean SD Mean SD Mean SD Mean SD F Stature (cm) Weight (kg) BMI (kg/m 2 ) R( ) Xc ( ) R/H ( /m) Xc/H ( /m) Phase angle ( ) r(r, Xc) RESULTS Comparison With the Normal Italian Values Table I shows the descriptive statistics for the anthropometric and bioelectrical variables in the sample subdivided by age group. The same table shows the correlation values between the parameters R and Xc. In the comparison with the results of De Palo et al., 4 we observed that the mean values of the Sardinian girls were always within the 50% tolerance intervals of the values of the Italian population. Moreover, the distribution of the individual values was more concentrated: in the general sample, 55.9% of cases fell within the 50% tolerance interval, 85.3% within the 75% interval, and 98.6% within the 95% interval. In the overall sample subdivided by age, ANOVA showed significant differences in the distribution of the anthropometric and bioelectrical values in most cases (Table I). As expected, the values of stature, weight, and BMI followed an increasing trend. The values of R/H and Xc/H showed a corresponding decrease. The descriptive and F statistics for the pre- and postmenarcheal subsamples subdivided into 1-y age classes (Tables II and III, respectively) showed a rather homogeneous within-group distribution. With the exception of stature, which increased significantly with age in both groups but remained at lower levels in the premenarcheal girls, the other anthropometric variables did not show a statistically significant age-related trend. In the postmenarcheal girls the values of weight and BMI were similar in all age groups (BMI range: ). BMI and weight were lower in the premenarcheal girls, in whom BMI decreased slightly with age, from 17.6 at age 11 y to 15.2 at age 14 y. Thus, there was a higher percentage of thin girls among those with late sexual maturation. The bioelectrical variables standardized by stature (R/H, Xc/H) also showed no significant differences in the distribution of the values in relation to age in the premenarcheal girls or the postmenarcheal girls. A graphical representation of this situation is shown in Figure 1. As can be seen, the location of the mean vectors in the RXc graph is very concentrated among postmenarcheal girls and they are also quite near to each other in the premenarcheal ones. The homogeneity within the two subsamples (pre- and postmenarcheal girls) permitted the pooling of the different age classes. Table IV shows the statistics relative to the entire pre- and postmenarcheal groups. Figure 2 shows the 95% confidence ellipses of the two groups and no overlap between the two curves. Hotelling s T 2 test for comparison of the pre- and postmenarcheal groups showed a significant difference in the impedance vectors (Table IV). The postmenarcheal girls (ellipse at lower left) presented lower values of R/H and Xc/H, whereas the phase angle was higher. A similar result was obtained after adjustment of the bioelectrical values for the associated effect of BMI. The new 95% confidence ellipses (Fig. 3), drawn with adjusted values of R/H and Xc/H, are once more significantly separated (Hotelling s T ; P 0.001). Effect of Menarche FIG. 1. Location of the mean vectors relative to pre- (solid circles) and postmenarcheal (open circles) groups subdivided age group.

4 Nutrition Volume 18, Number 6, 2002 Bioelectrical Impedance and Menarche 477 TABLE IV. STATISTICAL COMPARISON BETWEEN PRE- AND POSTMENARCHEAL FEMALES R( ) Xc ( ) R/H ( /m) Xc/H ( /m) Phase angle ( ) n Mean SD Mean SD Mean SD Mean SD Mean SD Premenarche Postmenarche * Hotelling s test (Z/H): T , F(2,140) 25.53, P SD, standard deviation; R, resistance; R/H, resistance standardized by stature; Xc, reactance; DISCUSSION AND CONCLUSIONS Our data represent the first information about the bioelectrical characteristics of the population of Sardinia. The results show that the distribution of the values is similar to that of the Italian population, 4 even though sampling from Sardinia was not considered for its definition. Like De Palo et al., 4 we observed a significant trend toward shortening and steepening (as indicated by the higher phase angle) of the Z vector with age. A similar trend was observed in adults as a consequence of variations in tissue hydration. In our case the vector migration can be related to the overall maturation process of the subjects, which implies modifications in body electrical properties, tissue changes, and increasing body size. Because the bioelectrical parameters were standardized by stature, a residual size influence could be exerted by the different cross-sectional areas of the body. The observed vector migration also can be interpreted as a consequence of sexual maturation. Hotelling s T 2 test showed a clear differentiation between pre- and postmenarcheal girls. Otherwise, no significant age-related difference in the distribution of the bioelectrical parameters was observed within groups, as indicated by the ANOVA results and graphical representation in Figure 1. These results suggest that in the considered growth period the principal cause of modification of body composition is related to sexual development. Accordingly, the mean impedance vector of postmenarcheal girls occupied an RXc graph position similar to that of the older girls. This result suggests a variation in the electrical properties of tissues and the body size of sexually mature girls. Further, data in the literature showed that obese patients present a similar, shorter and steeper vector than do healthy adults. 5 This indicates the possible influence of BMI variations in the different bioimpedance vector patterns. Therefore, the literature reports and the results of our study demonstrate that sexual maturation is accompanied by an increase in BMI. Thus the differences in impedance between pre- and postmenarcheal girls could be due to the corresponding differences in BMI. However, after correction of the bioelectrical values for the BMI effect, the difference between pre- and postmenarcheal girls, albeit reduced, was still significant (Fig. 3). This points to tissue maturation as an important source of bioelectrical variation during puberty. In conclusion, the results of this investigation suggest that the stage of sexual maturation should be considered in the application of bioelectrical parameters in growing individuals, particularly for clinical purposes. In fact, between the ages of 10 and 15 y, a difference in pubertal status implies a greater bioelectrical differentiation than a difference in age. Moreover, a significant component of the observed differentiation seems to be the variation in electrical properties of tissues that occurs with sexual maturation. FIG. 2. Mean impedance vectors with the 95% confidence ellipses relative to the pre- and postmenarcheal groups. FIG. 3. Mean impedance vectors (values adjusted for the effect of body mass index) with the 95% confidence ellipses relative to the pre- and postmenarcheal groups.

5 478 Buffa et al. Nutrition Volume 18, Number 6, 2002 A better definition of the vector migration with respect to the pubertal stage would be useful for applications involving the detection and follow-up of sexual maturation in individual girls. SUMMARY This article reports the bioelectrical impedance values of a sample of 143 Sardinian girls aged 10 to 15 y. The values agreed well with the normal values evaluated in the Italian population. Postmenarcheal females have a shorter impedance vector (reduced R/H and Xc/H) and a greater phase angle than premenarcheal females. ACKNOWLEDGMENTS The authors thank A. Piccoli for providing the BIVA program and F. Di Todaro for computer assistance. REFERENCES 1. NIH Technology Assessment Conference Statement. Bioelectrical impedance analysis in body composition measurements, 1994 Dec Am J Clin Nutr 1996;64:524S 2. Piccoli A, Rossi B, Pillon L, Bucciante G. A new method for monitoring body fluid variation by bioimpedance analysis: the RXc graph. Kidney Int 1994;46: Piccoli A, Nigrelli S, Caberlotto A, et al. Bivariate normal values of the bioelectrical impedance vector in adult and elderly populations. Am J Clin Nutr 1995;61: De Palo T, Messina G, Edefonti A, et al. Normal values of the bioelectrical impedance vector in childhood and puberty. Nutrition 2000;16: Piccoli A, Brunani A, Savia G, et al. Discriminating between body fat and fluid changes in the obese adult using bioimpedance vector analysis. Int J Obes Relat Metab Disord 1998;22:97 6. Toso S, Piccoli A, Gusella M, et al. Altered tissue electric properties in lung cancer patients as detected by bioelectric impedance vector analysis. Nutrition 2000;16: Piccoli A, Pillon L, Pisanello L, Zacchello G. Electrical maturation trajectory of human tissues identified by bioelectrical impedance vector analysis. Nutrition 1999;15:77 8. Cappello N, Rendine F, Griffo R, et al. Genetic analysis of Sardinia: I. Data on 12 polymorphisms in 21 linguistic domains. Am J Hum Genet 1996;60: Floris G. Sull evoluzione dei sardi dalla preistoria ad oggi. In: Floris G, Sanna E, eds. L Uomo in Sardegna. Aspetti di Antropobiologia ed Ecologia umana. Sestu (Cagliari): Zonza Editori, 1998: Weiner JS, Lourie JA. Practical human biology. London: Academic Press, Lukaski HC. Biological indexes considered in the derivation of the bioelectrical impedance analysis. Am J Clin Nutr 1996;64:397S 12. Marascuilo LA, Serlin RC. Statistical methods for the social and behavioral sciences. New York: Freeman, 1988 (For an additional perspective, see Editorial Opinions)