Multifactorial analysis of a mixed-longitudinal sample of Indian siblings: Age and sex effects on heritability.

The aim of the present study was to examine the heritability of 11 traits in a mixed-longitudinal sample of Indian siblings, and to determine whether heritability estimates vary during the growth period and whether they are influenced by sex. The sample consisted of 245 brothers and 213 sisters from 138 nuclear families living in a semi-urban area in Kolkata, India. The age ranged between 5 and 19 years. The traits were standardised for age and sex using standard deviation scores (SDS) produced by the LMS method (Cole, T.J., 1988. Fitting smoothed centile curves to reference data. J. R. Stat. Soc. A 151, 385-418). The standard deviation scores were analysed by PCA. The two factors with eigenvalues above 1 explained 77.3% of the variance; they showed a high level of pleiotropism present among the studied traits and represented body lengths (PC1) and body weight and breadths (PC2). The heritability between all types of siblings (irrespective of sex) for the PC1 and PC2 was estimated. The heritability between various pairs of siblings showed variations along the whole ontogenetic period studied. During the childhood and pre-pubertal period, heritability between brothers, brother-sister pairs and any sibling pairs was mostly constant, with small and non-significant variations. All the pairs showed the lowest degree of heritability during puberty for PC1 but not for PC2, with significant changes of heritability estimates between adolescence and adulthood, in most of the analysed sibling pairs and in both PC factors. The highest heritability was generally observed at the end of the examined growth period in all pairs. A significant effect of sex on heritability was only detected for PC2 at 11 years of age.

Heritability variations of morphometric traits in West Bengal (India) children aged 4-19 years: a mixed-longitudinal growth study.

BACKGROUND: Longitudinal and semi-longitudinal growth studies on siblings reflecting heritability changes during growth are very scarce. Moreover, studies of variables other than height, weight and BMI are virtually non-existent. AIM: The study compared changes in the heritability of six body lengths, four body breadths, and three indices between ages 4 and 19 years on the basis of a mixed-longitudinal sample of siblings, and examined whether heritability estimates change during the growth period. The data consisted of 238 brothers and 214 sisters from 134 middle-class nuclear families living in Kolkata (India). The analysis of sibling correlation was performed by maximum likelihood. The age-related patterns of heritabilities of the various traits were described by a cubic spline. RESULTS: The heritability was very high and significant in most traits, and at all considered ages. Mean heritability in the 10 morphometric traits was 69.3%, which was higher than the heritability values for the three indices. CONCLUSIONS: These results confirmed the existence of age-related trends in heritability of the considered morphometric traits. The sharp decline of the heritabilities at adolescence in most of the morphometric traits, and the acromio-iliac index in particular, may be due to the large inter-individual variation in the age at which the adolescent growth spurt is reached in both sexes.

Age trends of sibling resemblance for height, weight and BMI during growth in a mixed longitudinal sample from Sarsuna-Barisha, India.

BACKGROUND: Even though some studies have considered that sibling resemblance remains constant throughout the life cycle, several investigations emphasize the importance of age and its interactions with genetic and environmental factors in determining sibling similarity in several bodily traits. In fact, the study of age changes in familial resemblance is of great importance for the analysis of certain sources of variation observed in growth processes. AIM: The study examined sibling resemblance for height, weight and body mass index (BMI) in a mixed longitudinal sample from West Bengal, ages 2-19 years, in order to analyse the variations with age of the sibling resemblance for these phenotypes during growth. SAMPLE AND METHODS: Two hundred and forty-five brothers and 213 sisters from 138 middle-class nuclear families living in a semi-urban area of South Kolkata, India were analysed. The analysis of sibling resemblance was performed through correlations estimated by the maximum-likelihood method. The patterns of different trends of sibling resemblance with age were examined by fitting a cubic non-linear regression to the observed correlations. RESULTS: The results show clear variations with age in the sibling resemblance for the traits height and weight, though to a lesser extent for BMI. In general, we found the highest correlation values during the period of infancy, a remarkable decrease during puberty, and a trend of increase towards the end of the growth cycle. CONCLUSION: The study confirms the effect of age on the degree of similarity among siblings for height, weight and BMI in the sample. The sharp decline of correlation at adolescence can be interpreted in terms of the individual variation in age of reaching the adolescent growth spurt.

Longitudinal growth study of male Japanese junior high school athletes.

The aim of this study was to establish the pattern of adolescent growth in height of male Japanese athletes, and to compare their growth with appropriate controls. The sample consisted of 126 baseball, 39 basketball, 83 soccer, and 53 volleyball players, and 36 nonathletes. The data were collected retrospectively in six public schools in Fukui Prefecture of Western Japan between 1970 and 1987. All subjects were measured at yearly intervals between the age of 6 and 18 years. The total number of measurement occasions for the 337 subjects amounted to 4,134. Athletes trained between 13 and 18 h/week. Individual growth patterns were determined by fitting Preece Baines model I to each individual's serial data. Various biological parameters were derived from the fitted curves. The results indicated that the average growth pattern of these adolescent athletes did not differ substantially from the control group. Volleyball players were significantly taller at 18 years (+3 cm) than the nonathletes (Student's t-test, P < 0.05). Soccer players showed a slight tendency towards late maturation with age at take-off and at peak velocity about half a year later than in nonathletes (Student's t-test, P < 0.05). The 2.3 cm gain in body height due to the later onset of the pubertal growth spurt in soccer players was counterbalanced by a 0.7 cm smaller prepubertal height and a 0.9 cm smaller adolescent gain, so that they ended up slightly, but not significantly, taller than the controls (+0.7 cm). The small differences in growth pattern observed between the nonathletes and the various sports suggest that, in contrast with top-level athletes, body size and maturation rate were not important selective factors in the various groups of athletes of the present study.

Secular changes in growth and maturation: an update.

Secular changes in growth and maturation in recent decades have been reviewed for various populations. The secular increase in attained height during the growth period is continuing in most countries, but has slowed down. The increase in adult stature over the past decades has varied between 0.3 and 3.0 cm/decade. The secular trend in the tempo of growth (earlier menarche and peak height velocity, and shortening of the growth cycle) has come to a halt in some populations, but is continuing or has been reversed in others. The secular trend in attained height and in the tempo of growth is usually more pronounced in children from low socioeconomic backgrounds, in those with poorly educated parents or in those from rural areas. It is concluded that updates of growth standards are required in all populations. More marked secular changes appear to occur in the lower height centiles, which may have direct implications on the future definition of 'short stature' in a population.

Socioeconomic factors and growth during childhood and early adolescence in Jena children.

The influence of socioeconomic parameters (number of children in the family, birth order of subject, professional status of the father, education of the mother) and background characteristics (birth length, birth weight and age of the mother at birth of the child) of growth in body height in children from 4 to 12 years of age was investigated. The analysis is based on data from a longitudinal growth survey, which started with 207 children (98 males and 109 females) from Jena (Germany) in 1985. Multiple regression analysis was used to estimate the degree of relationship between height and the social or background factors. In this study no differences in growth between children of different social groups were found. Variation in height of girls was mainly affected by the biological factor length at birth of the child.

Secular changes in growth.

Secular changes in growth have occurred in almost all industrialized countries during the last century. Almost all of the secular increase seen in adult height occurs during childhood. These changes can be considered to be indicators of the changes in the nutritional, hygienic and health status of a population. The secular trend in adult height has slowed down since the second world war, but is still continuing in most industrialized countries.

Determinants of growth in body length from birth to 6 years of age: A longitudinal study of Lublin children.

The effect of determinants of growth in body length from birth to 6 years of age were studied in a longitudinal sample of 59 male and 70 female infants from Lublin, Poland. Structural equation modeling (SEM) was used to study the effects of gender of the child, occupation of the parents, the educational level of the parents, per capita income, the stature of the parents, and the weight of the mother on body length at birth and at 1, 2, 3, and 6 years of age. Significant sex differences in length were observed at birth and during the first 2 years of postnatal life, but not in the period between 3 and 6 years of age. Socioeconomic status (SES), expressed as a latent variable in the SEM, was not related to body length at birth but was significantly related to body length during infancy and, to a lesser extent, to body length during childhood. Paternal stature was not related to body length at birth and during infancy, but was significantly related to body length from 3 years onwards. Maternal stature was significantly related to body length at birth and at 1 year of age, but not thereafter, while maternal weight was significantly related to body length at birth only. © 1996 Wiley-Liss, Inc.

Genetic variance in the pattern of the growth curve for height: a longitudinal analysis of male twins.

Genetic aspects of the pattern of growth and of short-term variations in growth velocity for height were studied in a sample of 44 MZ and 42 DZ twin pairs from the Wroclaw Longitudinal Twin Study. The data consists of serial measurements of height, taken between 8.5 years of age and adulthood. The intra-pair resemblance of the pattern of attained height was quantified by means of the average Euclidean distance coefficient and the coefficient of shape difference, calculated on the raw height-for-age data. Comparison of these resemblance coefficients between the two types of twins indicated that the growth curves of MZ twins are closer to each other, and more similar in shape, than those of DZ twins. The shape of the growth curve was further characterized by a set of biological parameters, derived from Preece Baines model I (PB) fitted to each subject's serial growth data. Genetic analysis of these parameters, according to the model of Christian, Won Kang and Norton (1974), revealed a strong genetic component in the variance of size at particular milestones in the growth process (height at take-off, at peak velocity and at adulthood), and also in the timing of the growth process (age at take-off and at peak velocity). Height velocity at take-off and peak height velocity were also strongly genetically determined. Finally, short-term variations in growth velocity were analysed on the basis of the profiles of the residuals, obtained by the PB fits to each subject's serial measurements of height. Resemblance coefficients were calculated for the profiles of residuals. The results revealed a significantly greater similarity of profile shapes of the residuals in MZ twins than in DZ twins, strongly suggesting that there is a genetic component in the short-term variations of growth velocity.

[Reference values of the height and weight growth and growth rate of Belgian boys and girls 3-18 years of age]. / Normes de croissance staturale et pondérale et de vitesse de croissance staturale de garçons et de filles belges de 3 à 18 ans.

BACKGROUND: Belgian growth standards have until now been based on cross-sectional studies. This paper describes the first Belgian longitudinal growth standards for height and height velocity and recent cross-sectional standards of weight-for-age. POPULATION AND METHODS: The study was conducted between 1955 and 1975, initially on 259 Belgian boys and girls born in Brussels between 1955 and 1958. Only 48 boys and 50 girls were still being checked for height and weight at the end of the study. The distance charts show the classical centile lines; the height standards are provided with growth curves for the typical early average and late-maturing child in the population. The velocity charts show centiles for whole-year increments in height. Individual-type velocity curves are also provided. RESULTS: Comparison between these standards and those for British children shows that the Belgian children (essentially the girls) are taller than the British, probably due to more rapid maturation.

The pattern of growth and growth retardation of patients with hypophosphataemic vitamin D-resistant rickets: a longitudinal study.

Growth in height of 16 patients (5 boys and 11 girls) with hypophosphataemic rickets (HR) was studied in a longitudinal survey. The data shortly before and during puberty were analysed on the basis of Preece Baines curves, fitted to the original data; for the analysis at the age of 5 years, the original data were used. It appeared that the overall shape of the individual and average growth pattern could be adequately described by the Preece Baines method. The results further showed that from the age of 5 years onwards, average height was approximately two standard deviations below the normal mean for Dutch children. The patients showed a normal pubertal growth spurt which was, in general, insufficient to restore the growth retardation already established before adolescence. The four children who did show catch-up growth between the age of 5 years and adulthood had minimal rachitic lesions. The greater impact of the disease on growth in early childhood than on adolescent growth could be explained by the fact that HR mainly affects the growth of the legs, the major contributor to body size in early childhood. Finally, it was found that the difference between bone age, as determined by the Tanner Whitehouse (TW2)-method, and chronological age was not significant and the adult height in all patients except two could be adequately predicted from bone age and height.

Heights and weights of Swedish school children born in 1955 and 1967.

Samples of Swedish school children born in 1955 and 1967 were compared regarding average heights and weights from 10-15 years. The main findings were that both boys and girls had been gaining more weight than height, especially around the ages at which peak velocity generally occurs. Since the advantage in height of children born in 1967 gradually diminishes after age at peak height velocity, it seems that the height difference during puberty mainly reflects an earlier maturation of these children, compared to the children born in 1955.

Mathematical models for the study of individual growth patterns.

The analysis of longitudinal growth data requires specific methodological approaches. One of the main goals of longitudinal growth studies is to establish individual growth patterns and to estimate, so-called, biological parameters of the growth curve, such as the timing and intensity of the adolescent growth spurt, for example. These features are providing us with information about the shape of the growth curve, rather than telling us what size is attained at a particular age. A basic technique to establish the continuous growth process from a set of discrete measurements of size in function of age is provided by curve fitting. Various models have been proposed to achieve this goal. They can mainly be subdivided into nonstructural and structural models. This paper deals with a description of some of the most commonly used models in the analysis of human growth data, emphasizing on their applicability in certain age periods and on the merits and limitation of the various approaches. Attention is also paid to a special type of nonstructural models based on longitudinal principal components analysis.

Longitudinal study of growth of African babies: an analysis of seasonal variations in the average growth rate and the effects of infectious diseases on individual and average growth patterns.

The aim of the present study is to produce standards of weight for age and to analyse the effects of seasonal variation and of infectious disease on the rate of growth in weight of Central African babies. The study is based on longitudinal data of 4030 babies, aged 0-4 years, living in the North Western tropical forests of Zaire. The growth rate in weight shows a cyclic pattern which appears to be synchronized with the alterations of rainy and dry seasons, the rainy seasons slowing down the growth velocity below average. The interactions of genetic and nutritional factors are discussed. Infectious diseases such as measles and whooping cough tend to slow down growth in varying degrees. The effects were studied in terms of loss in weight and the amount of time needed to fully catch-up.

Comparison of biometric data of children with high and low levels of lead in the blood.

This paper deals with a biometric study of 312 boys and girls, aged 2.5-16 years, living in an area with a long history of pollution by lead. The aim was to search for eventual relationships between ten biometric variables and measures of lead absorption in the bodies, i.e. the amount of lead in the blood (PbB), of these children. Standardized values of the biometric variables were compared in the high-PbB and low-PbB categories, by multivariate analysis of variance. Comparison of the vectors of the ten biometric variables reveals a significant difference between the two categories of PbB levels. We found some evidence that the younger children (below 8 years of age) are more likely to absorb lead in the body and are more vulnerable to the effects of subclinical lead intoxication than their older counterparts. The differences between the averages of biometric variables in the two PbB categories are consistently (although not significantly) greater among younger children. This trend disappeared in the older age group. These results confirm data from the literature that young children are especially at risk. It can be concluded that there is a subtle, but significant, influence of lead absorption on the biometric profiles of children and that this effect is probably more important in children below 8 years of age.

Testing for the presence of genetic variance in factors of face measurements of Belgian twins.

Factor analysis with VARIMAX rotation was used to analyse 15 face measurements in Belgian same-sexed twins, aged 18-25 years: 39 dizygotic and 57 monozygotic male pairs and 42 dizygotic and 67 monozygotic female paris. According to Christian's model, we used the ratio of the within-mean squares of dizygotic and monozygotic pairs to test for the presence of a genetic component in the variance of the facial dimensions and of all the rotated factors were statistically significant (P less than 0.05), suggesting a genetic component in the variance. The probabilities of the F values were generally lower in males than in females. The factor analysis yielded five main factors of which three were well separated: face height, ear size and lips. The two others were breadth factors, but were less clearly defined, probably due to a bad selection of variables. A comparison of the F values of the factors with those of their contributing variables seemed to indicate that well-defined factors may better describe genetically determined structures than the original variables can.

Degree of resemblance of the pattern of growth among sibs in families of West Bengal (India).

The Preece-Baines Model 1 curve has been fitted to longitudinal data on growth in height of 105 boys and girls in 70 Bengal families: 60 of these were sibs distributed in 25 families. For a number of growth characteristics, such as age at peak height velocity, the proportion of the total population variance that was due to variation between, as opposed to within, families was estimated by Smith's (1980) method. The proportions for age at take-off of the adolescent spurt, of age at PHV, and of PHV itself were 22%, 26% and 33% (where adult height gives a value of about 40%). The sample is small and these estimates have high standard errors and need confirmation.

A longitudinal study of the growth in height of boys and girls of West Bengal (India) aged six months of 20 years.

This study is the first Indian longitudinal growth survey from early childhood to maturity. The heights of 303 boys and 260 girls, from middle-class families in a semi-urban area south of Calcutta, were measured at regular intervals over periods of up to 14 years (between 1952 and 1966). The data were analysed using appropriate mixed longitudinal and curve-fitting techniques. Growth in height of these middle-class Bengali children, who are not a representative sample of the Indian population, is slightly above the national Indian Council of Medical Research Standards. In both sexes, mean heights are below the 10th centile line of the British standards from an early age onwards, mainly due to a smaller prepubertal growth. The adolescent growth spurt in the Indians similar to that seen in British children, as is the age at which it occurs (peak height velocity at 14.0 years in boys, 12.5 years in girls). The sex difference of 14.0 cm in adult stature is attributable to a greater adolescent gain in the boys of 6.0 cm, a greater height in boys at the girls' age at take-off of 3.3 cm and a gain in height by the boys of 4.7 cm between the girls' and boys' ages at take-off.

A comparative study of the fit of four different functions to longitudinal data of growth in height of Belgian girls.

We have fitted Preece-Baines model 1, double logistic, logistic and Gompertz functions to longitudinal data on the growth in height of 35 Belgian girls, followed from birth to 18.0 years. The Preece-Baines model 1 showed significantly lower residual mean squares than the double logistic function, when fitted to data beyond the age of 1.0 year. The former model was also most robust towards variations in the lower age bound by the subject's data series and always described the adolescent spurt better than the latter. Both models fitted the data badly when measurements before the age of 1.0 year were included, and they usually estimated the point at take-off too early. Over the adolescent cycle, only, the logistic function fitted our data slightly better than the Gompertz function, with significantly lower pooled residual mean squares, though a slightly worse performance in the runs-test.