- Email: [email protected]
Anthropometric assessment of nutritional status in newborn infants. Discriminative value of mid arm circumference and of skinfold thickness
Early Human Development, Elsevier
11 (1985) 169-178
169
EHD 00638
Anthropometric assessment of nutritional status in newborn infants. Discriminative value of mid arm circumference and of skinfold thickness Jean Louis Excler a, LCon Sam a, Yves Lasne b and Jacques Picard ’ 0Service de Nkonatologie (Professeur M. Bethenod), Hspital Debrousse, 29 rue Soeur Bouvier, 69005 Lyon, ’ Laboratoire des Isotopes, Hijpital E. Herriot, 69003 Lyon, and’ CRESSA, 69003 Lyon, France Accepted
for publication
16 December
Hbpital Militaire Desgenettes,
1984
Summary 74 appropriate-for-gestational age (AGA) and 22 small-for-gestational age (SGA) Caucasian infants were studied for anthropometric parameters: mid arm circumference (MAC), triceps and subscapular skinfold thickness (TSKF and SSKF) recorded at 15 and 60 s, chest circumference (cc), head circumference, birth weight and length. MAC is highly correlated with birth weight either in AGA (r = 0.936; P < 0.001) or in SGA infants (r = 0.860; P < 0.001). MAC is also correlated with gestational age in AGA (r = 0.850; P < 0.001) and SGA infants (r = 0.76; P < 0.001). Similar correlations were found between TSKF, SSKF and birth weight or gestational age. Arm muscle and fat areas are also positively correlated with birth weight and gestational age, in AGA and SGA infants. A multiple regression analysis of our data allowed a classification of the best discriminant anthropometric parameters between AGA and SGA infants. MAC, SSKFlS, SSKF60 and chest circumference were selected. An equation was established in AGA infants with these four parameters giving a predictive gestational age (weeks) = 1.216 MAC (cm) - 3.588 SSKF15 (mm) + 0.263 CC age : gestational (cm) + 17.9. The ratio of predicted gestational age to the real gestational age was 1.0 + 0.044 in AGA versus 0.896 & 0.034 in SGA infants. Our data suggest that MAC and SSKF provide a simple measure of body composition of neonates and a useful tool for determining the degree of maturity of a newborn independent of birth weight. anthropometry;
0378-3782/85/$03.30
nutritional
status;
maturity;
0 1985 Elsevier Science Publishers
newborn
infants
B.V. (Biomedical
Division)
170
Anthropometric parameters, such as mid arm circumference (MAC), skinfold thickness (SKF) and indicators derived from these measurements, such as arm muscle and fat areas, have been advocated and widely used to determine nutritional status and body composition in adults and in children [13,18]. The measures are easy to perform and non-invasive. Arm muscle indicators (MAC, arm muscle area) have been advocated as indices of protein nutritional status [14], whereas SKF and arm fat area estimate the energy status of children [8]. However, MAC, fat and muscle areas have been little studied in newborn infants, since only two papers report these parameters in newborn infants who were more than 34 weeks of gestational age 14,101.In addition, no distinction is clearly established for these parameters between appropriate-for-gestational age (AGA) and small-for-gestational age (SGA) infants. Therefore in the present study we analyze the relationship between MAC, SKF, upper arm fat, muscle areas and birth weight or gestational age, both in AGA and SGA infants.
Patients and Methods 96 Caucasian newborn infants (57 males and 39 females) were studied within 24 h of birth for anthropometric parameters. They were selected randomly from the population admitted into the neonatal unit. Infants with malformations were excluded from the study. Informed parental consent was obtained for all the infants and the protocol was approved by the human subjects committee of the hospital. Gestational age was determined from the menstrual history of the mother or from electroencephalographic criteria [6] when menstrual history was uncertain. A gestation of 37 weeks or less defined prematurity. According to Lubchenco’s classification [15] a birth weight below the 10th percentile for gestational age defined SGA. Birth weight was measured to the nearest 10 g. 74 infants were AGA (mean + S.D.): gestational age = 35.4 weeks $- 34 (range 29-42), birth weight = 2269 + 673 g (range: 880-3950) and 22 were SGA: gestational age = 37.4 * 2.4 weeks (range: 33-42) birth weight = 1852 + 375 g (range: 1040-2360). All measures were performed by the same person (J.L.E.). Length was taken to the nearest cm on a standard measuring table. Head (fronto-occipital), chest (at the level of nipples in expiration) and mid arm (mid-distance between acromion and tip of olecranon) circumferences were measured to the nearest mm with plastic tape. Subscapular skinfold thickness (SSKF) and triceps skinfold thickness (TSKF) were measured with the Harpenden caliper using the technique described by Brans et al. [3]. Measurements were obtained to the nearest 0.05 mm, exactly 15 and 60 s after application of the caliper. Dynamic skinfold (A% SKF) was calculated as follows [3]: A% SKF = (SSKF15 - SSKF60)/SSKFlS X 100. Upper arm area (A), upper arm muscle area (M) and upper arm fat area (F) were derived from measures of MAC, converted to mm (C) and SSKF60 (T) using the following procedure previously described [8,9]: A(mm2)=$xd2,whered=$
171
M
(mm*) =
(’---T)2 andF(mm*)=A=A-M
Ponderal index was calculated as follows [19]: ponderal index = birth weight (g)/length
(cm3)
x
100
All these estimates were calculated for each subject. As pointed out in previous publications [8,9], the calculations A and F are only approximations. First, the formulae assume that the upper arm is cylindrical in form, an assumption subject to some inaccuracy. Second, the estimates of muscle area do not take into account the humeral diameter and any variation in humeral diameter is therefore not taken into account. However, we measured the humeral diameter on X-ray films of the upper arm in 26 infants included in the study: the mean humeral diameter was 0.44 + 0.06 cm. There was no change of this diameter with gestational age (34.2 f 6.9 weeks, range: 29-41). The results were analyzed with a computer using linear correlations and discriminative and multiple regression analysis with significance at the 95% level.
Results
Inter- and intraindividual coefficients of variation were 8.9% for SKF15 and for SKF60 (standard deviation for either TSKF or SSKF = 0.20 mm, n = 18), and 2.46% for MAC (S.D. = 0.16 cm, n = 9). Sex did not contribute to the discrimination for the different anthropometric parameters. Boys and girls were therefore considered together in our study. The individual values of MAC with gestational age or birth weight are shown in Figs. 1 and 2, respectively. Table I gives the different correlations between anthropometric parameters and gestational age or birth weight. MAC was highly correlated with birth weight either in AGA (r = 0.936, P < 0.001) or in SGA infants (r = 0.86, P < 0.01). MAC was also well correlated with gestational age in AGA (r = 0.85, P < 0.001) and in SGA infants (Y = 0.76, P < 0.01). Although the regression lines of MAC in AGA and SGA had a tendency to converge with increasing birth weight (Fig. 1) and to diverge with increasing gestational age (Fig. 2), their slopes were not significantly different. The discrimination between AGA and SGA was poor when MAC and birth weight were considered together. However, the discrimination appeared better with gestational age, since all but five SGA infants were situated below the interval of prediction of AGA (Fig. 2). A close correlation was found between MAC and ponderal index in AGA (r = 0.612; P < 0.001) and in SGA infants (Y = 0.678; P < 0.001). In AGA infants, SSKFlS and SSKF60 correlated with birth weight, respectively r = 0.67 (P < 0.001) and r = 0.73 (P < 0.001). The same correlations were found between TSKFlS and TSKF60 with birth weight, respectively r = 0.70 (P < 0.001) and r = 0.73 (P < 0.001). In SGA infants, similar strong correlations were obtained between SSKF60 and birth weight (r = 0.67; P < 0.001). No correlation was obtained between birth weight and the difference between SKF15 and SKF60. However, in AGA infants, A% SKF
172
??AGA
n=74 y=o,OO169 x + 4,403 r = 0,936 P< 0,001
n=22 oSGAy=0,00209x r=0,865 P< 0,001
+ 3,571
4
L
I
1000
4000
3000
2000 6lRlHWElGHl
(9)
Fig. 1. Correlation between mid arm circumference and birth weight in appropriate-for-gestational (AGA) and small-for-gestational age (SGA) infants within the first 24 h of life. 0. ,.I-
ll-
age
SD
9.
7-
Ii-
5-
n=22 c SGA y= 0,2912x -3,065 r= 0.759 PC 0,001
<,/-
,’
d-
27
,’
/’
29
29
30
31
32
33
GESTATIONAL
34
35 AGE
36
37
39
39
40
41
, 12
(weeks)
Fig. 2. Correlation between mid arm circumference and gestational age in AGA and SGA infants, within 24 h of birth.
173 TABLE
I
Equation of regression between the different following anthropometric parameters in AGA and SGA infants: MAC (cm), SSKF (mm), TSKF (mm), chest circumference (cm), birth weight (g), gestational age (weeks), ponderal index. n
Equations
r
AGA = 74
MAC = 0.00189 BW + 4.403 BW = 466.71 MAC- 1179.57 MAC = 0.3527 GA - 3.798 SSKF15 = 0.0012 BW + 0.8146 SSKF60 = 0.0013 BW + 0.8146 TSKF15 = 0.0009 BW + 1.631 TSKF60 = 0.0009 BW + 1.176 SSKFl5 = 0.1763 GA- 2.172 SSKF60 = 0.2088 GA - 3.660 TSKFIS = 0.1536 GA - 1.7548 TSKF60 = 0.1599 GA - 2.4367 CC = 0.0041 BW + 19.744 CC = 0.906 GA - 3.223
0.936 0.936 0.853 0.674 0.726 0.697 0.733 0.516 0.644 0.600 0.657 0.874 0.878
* * * * * * * * * * * * *
SGA = 22
MAC = 0.00209 BW + 3.571 MAC = 0.2812 GA- 3.065 MAC = 3.00 PI + 0.696 SSKFl5 = 0.00157 BW -0.0152 SSKF60 = 0.00135 BW + 0.0664 TSKF15 = 0.00121 BW + 0.4418 TSKF60 = 0.00114 BW + 0.2323 SSKF15 = 0.1897 GA - 4.2444 SSKF60 = 0.1579 GA- 3.3323 TSKFl5 = 0.1407 GA - 2.5733 TSKF60 = 0.1329 GA - 2.6236 CC = 0.0066 BW + 14.903 CC = 0.814 GA- 3.127
0.865 0.758 0.678 0.688 0.678 0.651 0.680 0.541 0.516 0.492 0.516 0.902 0.816
* * * * * * * ** ** *** **
GA, gestational age; BW, birth weight; PI, ponderal * P < 0.001; ** P < 0.01; *** P < 0.02. TABLE
index; CC, chest circumfence.
II
Equations of regression AGA and SGA infants.
between
upper arm areas (mm’),
birth weight (g) and gestational
n
Equations
r
AGA = 74
Muscle area = 0.1816 BW + 69.514 Muscle area = 38.4610 GA - 890.64 Fat area = 0.0729 BW - 26.980 Fat area = 13.5296 GA - 341.670
0.862 0.765 0.831 0.646
* * * *
SGA = 22
Muscle area = 0.1770 BW + 34.467 Muscle area = 24.6228 GA - 557.69 Fat area = 0.060 BW - 26.752 Fat area = 7.4687 GA - 194.568
0.806 0.730 0.173 0.625
* * * *
BW, birth weight; * P < 0.001.
GA, gestational
age.
age (weeks) in
174
SGA
500-
“E &
n-22
400.
a w : 5 a
300.
zoo-
100. I
I 1000
I
2000
BIRTHWEIGHT
3000
4000
(9)
Fig. 3. Correlations of arm muscle (AMA) and fat areas (AFA) with birth weight (BW) in AGA and SGA infants, within 24 h of birth. AGA: AMA = 0.1816 BW +69.514; r = 0.862 *. AFA = 0.0729 BW - 26.980; r = 0.831 *. SGA: AMA = 0.177 BW + 34.467; r = 0.806 *. AFA = 0.060 BW - 26.752; r = 0.773 *. * P < 0.001.
was poorly but significantly correlated with birth weight (r = - 0.315; P < 0.01) and with gestational age (I = - 0.306; P < 0.01). As shown in Table II, muscle areas and fat areas showed strong correlations with birth weight, respectively in AGA (r = 0.86; P < 0.001 & r = 0.83; P < 0.001) and in SGA infants (r = 0.80; P < 0.001 & r = 0.77; P < 0.001). However, both muscle and fat areas were less correlated with gestational age either in AGA or in SGA infants. Muscle area regression lines in AGA and in SGA infants were parallel, whereas fat area lines had a tendency to diverge with increasing birth weight (Fig. 3) and with increasing gestational age (Fig. 4). However, the slopes of these different regression lines were not significantly different between AGA and SGA infants. Chest circumference also showed a good correlation with birth weight in AGA (r = 0.87; P < 0.001) and SGA infants (r = 0.90; P < 0.001). Multiple regression allowed a classification of the best discriminant anthropometric parameters between AGA and SGA infants. It was found that parameters such as birth weight, length and head circumference were poorly discriminative compared with MAC, SSKF and chest circumference between AGA and SGA infants. An equation was established in AGA infants between gestational age and the best discriminant anthropometric parameters: gestational age (weeks) = 1.216 MAC (cm) - 3.588 SSKFlS (mm) + 3.963 SSKF60 (mm) + 0.263 chest circumference (cm) + 17.9
175
600-
+ AGA
n=74
SGA
n=22
500-
400-
300-
zoo-
100.
L,,.,..,....,,., 21 28
29
30
31
32
33
GESTATIONAL
34
35 AGE
36
31
38
39
40
41
(weeks)
Fig. 4. Correlations of arm muscle (AMA) and fat areas (AFA) with gestational age (GA) in AGA and SGA infants, within 24 h of birth. AGA: AMA = 38.461 GA- 890.64; r = 0.765 *. AFA = 13.523 GA - 341.67; r = 0.646 *. SGA: AMA = 24.623 GA - 551.69; i- = 0.730 *. AFA = 7.469 GA - 194.569: r = 0.625 **. * P < 0.001; ** P < 0.1.
The ratio between this calculated gestational age and the known gestational age was (mean f 1 SD.): 1.0 &-0.044 in AGA infants versus 0.896 * 0.034 in SGA infants. Discussion
Measurement of MAC is non-invasive, painless and extremely easy to perform in neonates. The method is reproducible, since the intra- and interindividual coefficients of variation were low, similar to those found in the literature for children [18,21] and neonates [10,16]. MAC is highly correlated with gestational age and birth weight. Since humeral diameter showed no variance with gestational age in our study, three other factors could contribute to the variations of MAC: muscle, fat and extracellular interstitial water. All but five SGA infants had a MAC situated below the interval of prediction of AGA infants. The ponderal index of these five infants did not differ from the other SGA’s ponderal index. Although correlated with MAC, ponderal index seems to be of poor discriminative value compared to MAC. In our study, arm muscle area is well correlated with birth weight (Fig. 3). Arm muscle area is closely related to creatinine excretion in children and offers a useful index of body muscle mass [21]. In newborn infants, the muscle mass accounts for 25% of body weight, irrespective of gestational age [24] and represents a major
176
protein reservoir. The regression lines of arm muscle areas in our study increase linearly with birth weight and are parallel to the total body protein content in neonates given by Brans et al. [l]. Moreover, the ratio of protein to total body weight is also approximately constant (12%) during the third trimester of gestation [1,27]. These data suggest that arm muscle area could be a valuable index of protein body composition in newborn infants and to a lesser extent of muscle nitrogen since 90% of muscle nitrogen belongs to proteins [24]. The separation between the regression lines of arm muscle areas of AGA and SGA infants either with birth weight or gestational age suggests that SGA infants have a decreased muscle mass compared to AGA infants [26]. On the other hand, the parallelism of these regression lines is in agreement with the findings of protein accretion in SGA similar to their normal growth controls [2]. However, we must be aware of the limits of arm muscle area. As suggested by Heymsfield et al. [ll], connective tissues represented by neurovascular bundles can lead to a significant over-estimation (5-10%) for muscle area. This factor was very difficult to evaluate and has not been studied in our work. According to Heymsfield [ll], it could be evaluated by ultrasound and might deserve further study in newborn infants. As previously described by Brans et al. [3], we found a positive correlation between TSKF60, birth weight and gestational age. Similar results were found at the subscapular site at 60 s. Correlations were better with birth weight than with gestational age, and much weaker in SGA than in AGA infants. These results suggest that subcutaneous fat deposition estimated by SKF60 measurements is affected both by impaired intrauterine growth and by dysmaturity. Usher [22] and Usher and McLean [23] have shown the same effect of intrauterine growth retardation on the abdominal skinfold. Arm fat area is better correlated with birth weight than SKF in AGA and SGA infants (Fig. 3). The present results in preterm and full-term neonates confirm the situation observed in children and in adults that fat areas are systematically better estimations than the corresponding SKF for estimating fat body weight [9,12]. Arm fat area is also correlated to a lesser extent with gestational area. These findings are similar to those described in the study of carcasses of infants [25] and the study of Enzi et al. [7]: in AGA newborns, SKF and body fat mass progressively increase from the 30th to the 41st week of gestation showing a significant correlation with gestational age. In SGA infants, the fat accretion is reduced with increasing birth weight [7]. Thus, SKF and particularly arm fat area constitute a valuable approach to body fat mass in neonates. Brans et al. [3] suggested that the dynamic skinfold thickness (A% SKF) reflects the extracellular water content and this was recently confirmed by Thornton et al. 1201.In the present study, a weak but significant negative correlation was observed in AGA infants between A% SKF and gestational age and birth weight. This finding agrees with the previous results of Brans et al. [3], suggesting that the proportion of body water decreases with maturity. However, the correlation was much weaker than in their study but similar to the results of Bustamente et al. [5]. The absence of correlation in SGA infants was not unexpected since the body water content is much larger in these infants regardless of maturity [l].
171
When a multiple regression analysis was carried out, it appeared that four parameters showed the greatest value for discriminating AGA and SGA infants: MAC, SSKFlS. SSKF60 and chest circumference. This finding was not surprising for the skinfold thickness and chest circumference since they agree with previous investigations [3,17]. It also agrees with the marked differences between AGA and SGA infants observed with MAC determinations in the present study. An equation could be established in AGA infants between these parameters and gestational age, showing the influence of maturity on these parameters. However, in SGA infants the application of the same equation resulted in a lower calculated gestational age, suggesting the influence of dysmaturity and poor fetal nutrition on the maturation of these parameters of body composition. In conclusion, the present data suggest that MAC and SKF thickness provide a simple approach to the body composition of neonates. The multiple regression analysis offers a better utilization of the parameters for their relation to maturity. Acknowledgements
We are indebted to Y. Ingenbleek, M.D., Ph.D. (Nestle Research Department), and to A.G.S. Philip, M.D. (Division of Neonatology, Maine Medical Center, Portland, Maine, U.S.A.) for their helpful comments. References 1 Brans, Y.W. and Cassady, G. (1975): Fetal nutrition and body composition. In: Total parenteral nutrition, pp. 301-333. Editors: H. Ghadimi, John Wiley, New York. 2 Brans, Y.W. and Shannon, D.L. (1981): Chemical changes in human skeletal muscle during fetal development. Biol. Neonate, 40, 21-28. 3 Brans, Y.W., Sumners, J.E., Dweck, H.S. and Cassady, G. (1974): A non invasive approach of body composition in the neonate: dynamic skinfold measurements. Pediatr. Res., 8, 215-222. 4 Brook, O.G., Butters, F., Wood, C., Bailey, P. and T&ma&i, F. (1981): Size at birth from 37-41 weeks gestation: ethnic standards for British infants of both sexes. J. Hum. Nutr., 35, 415-430. 5 Bustamente, S., Jacobs, P. and Gaines, J. (1982): Body weight, static and dynamic skinfold thickness in small premature infants during first month of life. Pediatr. Res., 16, 109. 6 Dreyfus-Brissac, C. (1979): Ontogenesis of brain bioelectrical activity and sleep. Organisation in neonates and infants. In: Human Growth, pp. 157-178. Editors: F. Falkner and J.M. Tanner. Plenum Publishing Corp., New York. 7 Enzi, G., Zanardo, V., Caretta, F., Inelmen, E.M. and RubaltelIi, F. (1981): Intrauterine growth and adipose tissue development. Am. J. Clin. Nutr., 34, 1785-1790. 8 Frisancho, A.R. (1974): Triceps skinfold and upper arm muscle size norms for assessment of nutritional status. Am. J. Clin. Nutr., 27, 1052-1057. 9 Frisancho, A.R. (1981): New norms of upperlimb fat and muscle areas for assessment of nutritional status. Am. J. CIin. Nutr., 34, 2540-2545. 10 Frisancho, A.R., KIayman, J.E. and Matos, J. (1977): Newborn body composition and its relation to linear growth. Am. J. Clin. Nutr., 30, 704-711. 11 Heymsfield, S.B., McManus, C., Smith, J., Stevens, V. and Nixon, D. (1982): Anthropometric measurement of muscle mass: revised equations for calculating bone-free arm muscle area. Am. J. Clin. Nutr., 36, 680-690.
178 12 Himes, J.H., Roche, A.F. and Webb, P. (1980): Fat areas as estimates of total body fat. Am. J. Clin. Nutr., 33, 2093-2100. 13 Jelliffe, D.B. (1966): The assessment of the nutritional status of the community. WHO Monograph series No. 63, Geneva. 14 Jelliffe, D.B. and Jelliffe, E.F.P. (1969): The arm circumference as a public health index of protein-calorie malnutrition of early childhood. Monograph 8. J. Trop. Pediatr., 15, 177-188. 15 Lubchenco, L.D., Hansman, C. and Boyd, E. (1966): Intrauterine growth in length and head circumference as estimated from live birth at gestational ages from 26 to 42 weeks. Pediatrics, 37, 403-408. 16 McGowan, A., Jordan, M. and McGregor, J. (1975): Skinfold thickness in neonates. Biol. Neonate, 25, 66-84. 17 McLean, F. and Usher, R. (1970): Measurements of liveborn fetal malnutrition infants compared with similar gestation and with similar birth weight normal controls. Biol. Neonate, 16, 215-221. 18 Martorelli, R., Yarbrough, C., Lechtig, A., Delgado, H. and Klein, R.E. (1976): Upperarm anthropometric indicators of nutritional status. Am. J. Clin. Nutr., 29, 46-53. 19 Miller, H.C. and Hassanein, K. (1971): Diagnosis of impaired fetal growth in newborn infants. Pediatrics, 48, 511-522. 20 Thornton. C.J., Shannon, D.L., Hunter, M.A. and Brans, Y.W. (1982): Dynamic skinfold thickness measurements: a non-invasive estimate of neonatal extracellular water content. Pediatr. Res., 16, 989-994. 21 Trowbridge, F.L., Hinner, C.D. and Robertson, A.D. (1982): Arm muscle indicators and creatinine excretion in children. Am. J. Clin. Nutr., 36, 691-696. 22 Usher, R.H. (1970): Clinical and therapeutic aspect of fetal malnutrition. Pediatr. Clin. Am., 17, 169-183. 23 Usher, R.H. and McLean, F. (1969): Intrauterine growth of live-born Caucasian infants at sea level: Standard obtained from measurements in 7 dimensions of infants born between 25 and 44 weeks of gestation. J. Pediatr., 74, 901-910. 24 Widdowson, E.M. and Dickerson, Y.N.T. (1964): Chemical composition of the body. In: Mineral Metabolism: An Advanced Treatise, Vol. II, pp. 2-248. Editors: C.C. Comaz and F.B. Bronner. Academic Press, New York. 25 Widdowson, E.M. and Spray, C.M. (1951): Chemical development in utero. Arch. Dis. Child., 26, 205-214. 26 Widdowson, E.M., Crabb, D.E. and Milner, R.D.G. (1972): Cellular development of some human organs before birth. Arch. Dis. Child., 47, 652-655. 27 Ziegler, E.E., O’Donnell, A.M., Nelson, S.E. and Fomon, S.J. (1976): Body composition of the reference fetus. Growth, 40, 329-341.
11 (1985) 169-178
169
EHD 00638
Anthropometric assessment of nutritional status in newborn infants. Discriminative value of mid arm circumference and of skinfold thickness Jean Louis Excler a, LCon Sam a, Yves Lasne b and Jacques Picard ’ 0Service de Nkonatologie (Professeur M. Bethenod), Hspital Debrousse, 29 rue Soeur Bouvier, 69005 Lyon, ’ Laboratoire des Isotopes, Hijpital E. Herriot, 69003 Lyon, and’ CRESSA, 69003 Lyon, France Accepted
for publication
16 December
Hbpital Militaire Desgenettes,
1984
Summary 74 appropriate-for-gestational age (AGA) and 22 small-for-gestational age (SGA) Caucasian infants were studied for anthropometric parameters: mid arm circumference (MAC), triceps and subscapular skinfold thickness (TSKF and SSKF) recorded at 15 and 60 s, chest circumference (cc), head circumference, birth weight and length. MAC is highly correlated with birth weight either in AGA (r = 0.936; P < 0.001) or in SGA infants (r = 0.860; P < 0.001). MAC is also correlated with gestational age in AGA (r = 0.850; P < 0.001) and SGA infants (r = 0.76; P < 0.001). Similar correlations were found between TSKF, SSKF and birth weight or gestational age. Arm muscle and fat areas are also positively correlated with birth weight and gestational age, in AGA and SGA infants. A multiple regression analysis of our data allowed a classification of the best discriminant anthropometric parameters between AGA and SGA infants. MAC, SSKFlS, SSKF60 and chest circumference were selected. An equation was established in AGA infants with these four parameters giving a predictive gestational age (weeks) = 1.216 MAC (cm) - 3.588 SSKF15 (mm) + 0.263 CC age : gestational (cm) + 17.9. The ratio of predicted gestational age to the real gestational age was 1.0 + 0.044 in AGA versus 0.896 & 0.034 in SGA infants. Our data suggest that MAC and SSKF provide a simple measure of body composition of neonates and a useful tool for determining the degree of maturity of a newborn independent of birth weight. anthropometry;
0378-3782/85/$03.30
nutritional
status;
maturity;
0 1985 Elsevier Science Publishers
newborn
infants
B.V. (Biomedical
Division)
170
Anthropometric parameters, such as mid arm circumference (MAC), skinfold thickness (SKF) and indicators derived from these measurements, such as arm muscle and fat areas, have been advocated and widely used to determine nutritional status and body composition in adults and in children [13,18]. The measures are easy to perform and non-invasive. Arm muscle indicators (MAC, arm muscle area) have been advocated as indices of protein nutritional status [14], whereas SKF and arm fat area estimate the energy status of children [8]. However, MAC, fat and muscle areas have been little studied in newborn infants, since only two papers report these parameters in newborn infants who were more than 34 weeks of gestational age 14,101.In addition, no distinction is clearly established for these parameters between appropriate-for-gestational age (AGA) and small-for-gestational age (SGA) infants. Therefore in the present study we analyze the relationship between MAC, SKF, upper arm fat, muscle areas and birth weight or gestational age, both in AGA and SGA infants.
Patients and Methods 96 Caucasian newborn infants (57 males and 39 females) were studied within 24 h of birth for anthropometric parameters. They were selected randomly from the population admitted into the neonatal unit. Infants with malformations were excluded from the study. Informed parental consent was obtained for all the infants and the protocol was approved by the human subjects committee of the hospital. Gestational age was determined from the menstrual history of the mother or from electroencephalographic criteria [6] when menstrual history was uncertain. A gestation of 37 weeks or less defined prematurity. According to Lubchenco’s classification [15] a birth weight below the 10th percentile for gestational age defined SGA. Birth weight was measured to the nearest 10 g. 74 infants were AGA (mean + S.D.): gestational age = 35.4 weeks $- 34 (range 29-42), birth weight = 2269 + 673 g (range: 880-3950) and 22 were SGA: gestational age = 37.4 * 2.4 weeks (range: 33-42) birth weight = 1852 + 375 g (range: 1040-2360). All measures were performed by the same person (J.L.E.). Length was taken to the nearest cm on a standard measuring table. Head (fronto-occipital), chest (at the level of nipples in expiration) and mid arm (mid-distance between acromion and tip of olecranon) circumferences were measured to the nearest mm with plastic tape. Subscapular skinfold thickness (SSKF) and triceps skinfold thickness (TSKF) were measured with the Harpenden caliper using the technique described by Brans et al. [3]. Measurements were obtained to the nearest 0.05 mm, exactly 15 and 60 s after application of the caliper. Dynamic skinfold (A% SKF) was calculated as follows [3]: A% SKF = (SSKF15 - SSKF60)/SSKFlS X 100. Upper arm area (A), upper arm muscle area (M) and upper arm fat area (F) were derived from measures of MAC, converted to mm (C) and SSKF60 (T) using the following procedure previously described [8,9]: A(mm2)=$xd2,whered=$
171
M
(mm*) =
(’---T)2 andF(mm*)=A=A-M
Ponderal index was calculated as follows [19]: ponderal index = birth weight (g)/length
(cm3)
x
100
All these estimates were calculated for each subject. As pointed out in previous publications [8,9], the calculations A and F are only approximations. First, the formulae assume that the upper arm is cylindrical in form, an assumption subject to some inaccuracy. Second, the estimates of muscle area do not take into account the humeral diameter and any variation in humeral diameter is therefore not taken into account. However, we measured the humeral diameter on X-ray films of the upper arm in 26 infants included in the study: the mean humeral diameter was 0.44 + 0.06 cm. There was no change of this diameter with gestational age (34.2 f 6.9 weeks, range: 29-41). The results were analyzed with a computer using linear correlations and discriminative and multiple regression analysis with significance at the 95% level.
Results
Inter- and intraindividual coefficients of variation were 8.9% for SKF15 and for SKF60 (standard deviation for either TSKF or SSKF = 0.20 mm, n = 18), and 2.46% for MAC (S.D. = 0.16 cm, n = 9). Sex did not contribute to the discrimination for the different anthropometric parameters. Boys and girls were therefore considered together in our study. The individual values of MAC with gestational age or birth weight are shown in Figs. 1 and 2, respectively. Table I gives the different correlations between anthropometric parameters and gestational age or birth weight. MAC was highly correlated with birth weight either in AGA (r = 0.936, P < 0.001) or in SGA infants (r = 0.86, P < 0.01). MAC was also well correlated with gestational age in AGA (r = 0.85, P < 0.001) and in SGA infants (Y = 0.76, P < 0.01). Although the regression lines of MAC in AGA and SGA had a tendency to converge with increasing birth weight (Fig. 1) and to diverge with increasing gestational age (Fig. 2), their slopes were not significantly different. The discrimination between AGA and SGA was poor when MAC and birth weight were considered together. However, the discrimination appeared better with gestational age, since all but five SGA infants were situated below the interval of prediction of AGA (Fig. 2). A close correlation was found between MAC and ponderal index in AGA (r = 0.612; P < 0.001) and in SGA infants (Y = 0.678; P < 0.001). In AGA infants, SSKFlS and SSKF60 correlated with birth weight, respectively r = 0.67 (P < 0.001) and r = 0.73 (P < 0.001). The same correlations were found between TSKFlS and TSKF60 with birth weight, respectively r = 0.70 (P < 0.001) and r = 0.73 (P < 0.001). In SGA infants, similar strong correlations were obtained between SSKF60 and birth weight (r = 0.67; P < 0.001). No correlation was obtained between birth weight and the difference between SKF15 and SKF60. However, in AGA infants, A% SKF
172
??AGA
n=74 y=o,OO169 x + 4,403 r = 0,936 P< 0,001
n=22 oSGAy=0,00209x r=0,865 P< 0,001
+ 3,571
4
L
I
1000
4000
3000
2000 6lRlHWElGHl
(9)
Fig. 1. Correlation between mid arm circumference and birth weight in appropriate-for-gestational (AGA) and small-for-gestational age (SGA) infants within the first 24 h of life. 0. ,.I-
ll-
age
SD
9.
7-
Ii-
5-
n=22 c SGA y= 0,2912x -3,065 r= 0.759 PC 0,001
<,/-
,’
d-
27
,’
/’
29
29
30
31
32
33
GESTATIONAL
34
35 AGE
36
37
39
39
40
41
, 12
(weeks)
Fig. 2. Correlation between mid arm circumference and gestational age in AGA and SGA infants, within 24 h of birth.
173 TABLE
I
Equation of regression between the different following anthropometric parameters in AGA and SGA infants: MAC (cm), SSKF (mm), TSKF (mm), chest circumference (cm), birth weight (g), gestational age (weeks), ponderal index. n
Equations
r
AGA = 74
MAC = 0.00189 BW + 4.403 BW = 466.71 MAC- 1179.57 MAC = 0.3527 GA - 3.798 SSKF15 = 0.0012 BW + 0.8146 SSKF60 = 0.0013 BW + 0.8146 TSKF15 = 0.0009 BW + 1.631 TSKF60 = 0.0009 BW + 1.176 SSKFl5 = 0.1763 GA- 2.172 SSKF60 = 0.2088 GA - 3.660 TSKFIS = 0.1536 GA - 1.7548 TSKF60 = 0.1599 GA - 2.4367 CC = 0.0041 BW + 19.744 CC = 0.906 GA - 3.223
0.936 0.936 0.853 0.674 0.726 0.697 0.733 0.516 0.644 0.600 0.657 0.874 0.878
* * * * * * * * * * * * *
SGA = 22
MAC = 0.00209 BW + 3.571 MAC = 0.2812 GA- 3.065 MAC = 3.00 PI + 0.696 SSKFl5 = 0.00157 BW -0.0152 SSKF60 = 0.00135 BW + 0.0664 TSKF15 = 0.00121 BW + 0.4418 TSKF60 = 0.00114 BW + 0.2323 SSKF15 = 0.1897 GA - 4.2444 SSKF60 = 0.1579 GA- 3.3323 TSKFl5 = 0.1407 GA - 2.5733 TSKF60 = 0.1329 GA - 2.6236 CC = 0.0066 BW + 14.903 CC = 0.814 GA- 3.127
0.865 0.758 0.678 0.688 0.678 0.651 0.680 0.541 0.516 0.492 0.516 0.902 0.816
* * * * * * * ** ** *** **
GA, gestational age; BW, birth weight; PI, ponderal * P < 0.001; ** P < 0.01; *** P < 0.02. TABLE
index; CC, chest circumfence.
II
Equations of regression AGA and SGA infants.
between
upper arm areas (mm’),
birth weight (g) and gestational
n
Equations
r
AGA = 74
Muscle area = 0.1816 BW + 69.514 Muscle area = 38.4610 GA - 890.64 Fat area = 0.0729 BW - 26.980 Fat area = 13.5296 GA - 341.670
0.862 0.765 0.831 0.646
* * * *
SGA = 22
Muscle area = 0.1770 BW + 34.467 Muscle area = 24.6228 GA - 557.69 Fat area = 0.060 BW - 26.752 Fat area = 7.4687 GA - 194.568
0.806 0.730 0.173 0.625
* * * *
BW, birth weight; * P < 0.001.
GA, gestational
age.
age (weeks) in
174
SGA
500-
“E &
n-22
400.
a w : 5 a
300.
zoo-
100. I
I 1000
I
2000
BIRTHWEIGHT
3000
4000
(9)
Fig. 3. Correlations of arm muscle (AMA) and fat areas (AFA) with birth weight (BW) in AGA and SGA infants, within 24 h of birth. AGA: AMA = 0.1816 BW +69.514; r = 0.862 *. AFA = 0.0729 BW - 26.980; r = 0.831 *. SGA: AMA = 0.177 BW + 34.467; r = 0.806 *. AFA = 0.060 BW - 26.752; r = 0.773 *. * P < 0.001.
was poorly but significantly correlated with birth weight (r = - 0.315; P < 0.01) and with gestational age (I = - 0.306; P < 0.01). As shown in Table II, muscle areas and fat areas showed strong correlations with birth weight, respectively in AGA (r = 0.86; P < 0.001 & r = 0.83; P < 0.001) and in SGA infants (r = 0.80; P < 0.001 & r = 0.77; P < 0.001). However, both muscle and fat areas were less correlated with gestational age either in AGA or in SGA infants. Muscle area regression lines in AGA and in SGA infants were parallel, whereas fat area lines had a tendency to diverge with increasing birth weight (Fig. 3) and with increasing gestational age (Fig. 4). However, the slopes of these different regression lines were not significantly different between AGA and SGA infants. Chest circumference also showed a good correlation with birth weight in AGA (r = 0.87; P < 0.001) and SGA infants (r = 0.90; P < 0.001). Multiple regression allowed a classification of the best discriminant anthropometric parameters between AGA and SGA infants. It was found that parameters such as birth weight, length and head circumference were poorly discriminative compared with MAC, SSKF and chest circumference between AGA and SGA infants. An equation was established in AGA infants between gestational age and the best discriminant anthropometric parameters: gestational age (weeks) = 1.216 MAC (cm) - 3.588 SSKFlS (mm) + 3.963 SSKF60 (mm) + 0.263 chest circumference (cm) + 17.9
175
600-
+ AGA
n=74
SGA
n=22
500-
400-
300-
zoo-
100.
L,,.,..,....,,., 21 28
29
30
31
32
33
GESTATIONAL
34
35 AGE
36
31
38
39
40
41
(weeks)
Fig. 4. Correlations of arm muscle (AMA) and fat areas (AFA) with gestational age (GA) in AGA and SGA infants, within 24 h of birth. AGA: AMA = 38.461 GA- 890.64; r = 0.765 *. AFA = 13.523 GA - 341.67; r = 0.646 *. SGA: AMA = 24.623 GA - 551.69; i- = 0.730 *. AFA = 7.469 GA - 194.569: r = 0.625 **. * P < 0.001; ** P < 0.1.
The ratio between this calculated gestational age and the known gestational age was (mean f 1 SD.): 1.0 &-0.044 in AGA infants versus 0.896 * 0.034 in SGA infants. Discussion
Measurement of MAC is non-invasive, painless and extremely easy to perform in neonates. The method is reproducible, since the intra- and interindividual coefficients of variation were low, similar to those found in the literature for children [18,21] and neonates [10,16]. MAC is highly correlated with gestational age and birth weight. Since humeral diameter showed no variance with gestational age in our study, three other factors could contribute to the variations of MAC: muscle, fat and extracellular interstitial water. All but five SGA infants had a MAC situated below the interval of prediction of AGA infants. The ponderal index of these five infants did not differ from the other SGA’s ponderal index. Although correlated with MAC, ponderal index seems to be of poor discriminative value compared to MAC. In our study, arm muscle area is well correlated with birth weight (Fig. 3). Arm muscle area is closely related to creatinine excretion in children and offers a useful index of body muscle mass [21]. In newborn infants, the muscle mass accounts for 25% of body weight, irrespective of gestational age [24] and represents a major
176
protein reservoir. The regression lines of arm muscle areas in our study increase linearly with birth weight and are parallel to the total body protein content in neonates given by Brans et al. [l]. Moreover, the ratio of protein to total body weight is also approximately constant (12%) during the third trimester of gestation [1,27]. These data suggest that arm muscle area could be a valuable index of protein body composition in newborn infants and to a lesser extent of muscle nitrogen since 90% of muscle nitrogen belongs to proteins [24]. The separation between the regression lines of arm muscle areas of AGA and SGA infants either with birth weight or gestational age suggests that SGA infants have a decreased muscle mass compared to AGA infants [26]. On the other hand, the parallelism of these regression lines is in agreement with the findings of protein accretion in SGA similar to their normal growth controls [2]. However, we must be aware of the limits of arm muscle area. As suggested by Heymsfield et al. [ll], connective tissues represented by neurovascular bundles can lead to a significant over-estimation (5-10%) for muscle area. This factor was very difficult to evaluate and has not been studied in our work. According to Heymsfield [ll], it could be evaluated by ultrasound and might deserve further study in newborn infants. As previously described by Brans et al. [3], we found a positive correlation between TSKF60, birth weight and gestational age. Similar results were found at the subscapular site at 60 s. Correlations were better with birth weight than with gestational age, and much weaker in SGA than in AGA infants. These results suggest that subcutaneous fat deposition estimated by SKF60 measurements is affected both by impaired intrauterine growth and by dysmaturity. Usher [22] and Usher and McLean [23] have shown the same effect of intrauterine growth retardation on the abdominal skinfold. Arm fat area is better correlated with birth weight than SKF in AGA and SGA infants (Fig. 3). The present results in preterm and full-term neonates confirm the situation observed in children and in adults that fat areas are systematically better estimations than the corresponding SKF for estimating fat body weight [9,12]. Arm fat area is also correlated to a lesser extent with gestational area. These findings are similar to those described in the study of carcasses of infants [25] and the study of Enzi et al. [7]: in AGA newborns, SKF and body fat mass progressively increase from the 30th to the 41st week of gestation showing a significant correlation with gestational age. In SGA infants, the fat accretion is reduced with increasing birth weight [7]. Thus, SKF and particularly arm fat area constitute a valuable approach to body fat mass in neonates. Brans et al. [3] suggested that the dynamic skinfold thickness (A% SKF) reflects the extracellular water content and this was recently confirmed by Thornton et al. 1201.In the present study, a weak but significant negative correlation was observed in AGA infants between A% SKF and gestational age and birth weight. This finding agrees with the previous results of Brans et al. [3], suggesting that the proportion of body water decreases with maturity. However, the correlation was much weaker than in their study but similar to the results of Bustamente et al. [5]. The absence of correlation in SGA infants was not unexpected since the body water content is much larger in these infants regardless of maturity [l].
171
When a multiple regression analysis was carried out, it appeared that four parameters showed the greatest value for discriminating AGA and SGA infants: MAC, SSKFlS. SSKF60 and chest circumference. This finding was not surprising for the skinfold thickness and chest circumference since they agree with previous investigations [3,17]. It also agrees with the marked differences between AGA and SGA infants observed with MAC determinations in the present study. An equation could be established in AGA infants between these parameters and gestational age, showing the influence of maturity on these parameters. However, in SGA infants the application of the same equation resulted in a lower calculated gestational age, suggesting the influence of dysmaturity and poor fetal nutrition on the maturation of these parameters of body composition. In conclusion, the present data suggest that MAC and SKF thickness provide a simple approach to the body composition of neonates. The multiple regression analysis offers a better utilization of the parameters for their relation to maturity. Acknowledgements
We are indebted to Y. Ingenbleek, M.D., Ph.D. (Nestle Research Department), and to A.G.S. Philip, M.D. (Division of Neonatology, Maine Medical Center, Portland, Maine, U.S.A.) for their helpful comments. References 1 Brans, Y.W. and Cassady, G. (1975): Fetal nutrition and body composition. In: Total parenteral nutrition, pp. 301-333. Editors: H. Ghadimi, John Wiley, New York. 2 Brans, Y.W. and Shannon, D.L. (1981): Chemical changes in human skeletal muscle during fetal development. Biol. Neonate, 40, 21-28. 3 Brans, Y.W., Sumners, J.E., Dweck, H.S. and Cassady, G. (1974): A non invasive approach of body composition in the neonate: dynamic skinfold measurements. Pediatr. Res., 8, 215-222. 4 Brook, O.G., Butters, F., Wood, C., Bailey, P. and T&ma&i, F. (1981): Size at birth from 37-41 weeks gestation: ethnic standards for British infants of both sexes. J. Hum. Nutr., 35, 415-430. 5 Bustamente, S., Jacobs, P. and Gaines, J. (1982): Body weight, static and dynamic skinfold thickness in small premature infants during first month of life. Pediatr. Res., 16, 109. 6 Dreyfus-Brissac, C. (1979): Ontogenesis of brain bioelectrical activity and sleep. Organisation in neonates and infants. In: Human Growth, pp. 157-178. Editors: F. Falkner and J.M. Tanner. Plenum Publishing Corp., New York. 7 Enzi, G., Zanardo, V., Caretta, F., Inelmen, E.M. and RubaltelIi, F. (1981): Intrauterine growth and adipose tissue development. Am. J. Clin. Nutr., 34, 1785-1790. 8 Frisancho, A.R. (1974): Triceps skinfold and upper arm muscle size norms for assessment of nutritional status. Am. J. Clin. Nutr., 27, 1052-1057. 9 Frisancho, A.R. (1981): New norms of upperlimb fat and muscle areas for assessment of nutritional status. Am. J. CIin. Nutr., 34, 2540-2545. 10 Frisancho, A.R., KIayman, J.E. and Matos, J. (1977): Newborn body composition and its relation to linear growth. Am. J. Clin. Nutr., 30, 704-711. 11 Heymsfield, S.B., McManus, C., Smith, J., Stevens, V. and Nixon, D. (1982): Anthropometric measurement of muscle mass: revised equations for calculating bone-free arm muscle area. Am. J. Clin. Nutr., 36, 680-690.
178 12 Himes, J.H., Roche, A.F. and Webb, P. (1980): Fat areas as estimates of total body fat. Am. J. Clin. Nutr., 33, 2093-2100. 13 Jelliffe, D.B. (1966): The assessment of the nutritional status of the community. WHO Monograph series No. 63, Geneva. 14 Jelliffe, D.B. and Jelliffe, E.F.P. (1969): The arm circumference as a public health index of protein-calorie malnutrition of early childhood. Monograph 8. J. Trop. Pediatr., 15, 177-188. 15 Lubchenco, L.D., Hansman, C. and Boyd, E. (1966): Intrauterine growth in length and head circumference as estimated from live birth at gestational ages from 26 to 42 weeks. Pediatrics, 37, 403-408. 16 McGowan, A., Jordan, M. and McGregor, J. (1975): Skinfold thickness in neonates. Biol. Neonate, 25, 66-84. 17 McLean, F. and Usher, R. (1970): Measurements of liveborn fetal malnutrition infants compared with similar gestation and with similar birth weight normal controls. Biol. Neonate, 16, 215-221. 18 Martorelli, R., Yarbrough, C., Lechtig, A., Delgado, H. and Klein, R.E. (1976): Upperarm anthropometric indicators of nutritional status. Am. J. Clin. Nutr., 29, 46-53. 19 Miller, H.C. and Hassanein, K. (1971): Diagnosis of impaired fetal growth in newborn infants. Pediatrics, 48, 511-522. 20 Thornton. C.J., Shannon, D.L., Hunter, M.A. and Brans, Y.W. (1982): Dynamic skinfold thickness measurements: a non-invasive estimate of neonatal extracellular water content. Pediatr. Res., 16, 989-994. 21 Trowbridge, F.L., Hinner, C.D. and Robertson, A.D. (1982): Arm muscle indicators and creatinine excretion in children. Am. J. Clin. Nutr., 36, 691-696. 22 Usher, R.H. (1970): Clinical and therapeutic aspect of fetal malnutrition. Pediatr. Clin. Am., 17, 169-183. 23 Usher, R.H. and McLean, F. (1969): Intrauterine growth of live-born Caucasian infants at sea level: Standard obtained from measurements in 7 dimensions of infants born between 25 and 44 weeks of gestation. J. Pediatr., 74, 901-910. 24 Widdowson, E.M. and Dickerson, Y.N.T. (1964): Chemical composition of the body. In: Mineral Metabolism: An Advanced Treatise, Vol. II, pp. 2-248. Editors: C.C. Comaz and F.B. Bronner. Academic Press, New York. 25 Widdowson, E.M. and Spray, C.M. (1951): Chemical development in utero. Arch. Dis. Child., 26, 205-214. 26 Widdowson, E.M., Crabb, D.E. and Milner, R.D.G. (1972): Cellular development of some human organs before birth. Arch. Dis. Child., 47, 652-655. 27 Ziegler, E.E., O’Donnell, A.M., Nelson, S.E. and Fomon, S.J. (1976): Body composition of the reference fetus. Growth, 40, 329-341.