| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH COMMUNICATION |
1 From the Centre for Paediatric Epidemiology and Biostatistics, Institute of Child Health, London (LL and CP), and the School of Public Health and Community Medicine, Hebrew University, Jerusalem (OM)
2 CP was supported as a Weston Fellow by the Canadian Institute for Advanced Research. 3 Address reprint requests to L Li, Centre for Paediatric Epidemiology and Biostatistics, ICH, 30 Guilford Street, London WC1N 1EH, United Kingdom. E-mail: L.Li{at}ich.ucl.ac.uk.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: The objective was to determine when parental height and factors early in offspring life start to affect offspring height, when these variables have the strongest effect, and whether these variables persist to adulthood.
Design: Longitudinal data from the 1958 British birth cohort (all of whom were born during 1 wk in March 1958), including height measurements at 7, 11, 16, and 33 y of age, were analyzed by using multivariate multilevel response models.
Results: Parental height, birth weight, maternal smoking during pregnancy, breastfeeding, parental divorce, and socioeconomic factors were all significantly associated with childhood height, but their effects differed thereafter. Parental height and birth weight were most strongly associated with offspring height, and their effects persisted (adjusted increase in adult height: 2 cm for 1 SD of maternal or paternal height, or 1 kg of birth weight). Socioeconomic disadvantage (manual social class, large family size, and overcrowded households) was associated with substantial deficits of 2–3 cm (adjusted estimates) in height at 7 y. Catch-up growth was apparent but was insufficient to overcome the initial insult on growth; the adjusted deficit was as high as 1 cm in adulthood.
Conclusions: Children from disadvantaged backgrounds have a delayed pattern of growth before the pubertal spurt, which is followed by catch-up growth. The health consequences of this pattern of growth need to be examined in future studies.
Key Words: Early environment • growth trajectory • height • cohort study • Britain
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We used longitudinal data from the 1958 British birth cohort (20) to investigate influences on growth trajectories from heights measured at ages 7, 11, 16, and 33 y. Our aim was to determine the role of parental height and early life factors during different growth phases to assess when their effects start and are the strongest, and the extent to which the effects persist through to adult height. The specific influences investigated were as follows: 1) parental height, as an indicator of genetic potential; 2) prenatal factors, namely birth weight and maternal smoking during pregnancy; and 3) postnatal factors, including breastfeeding (an indicator of nutrition in infancy), social class, family size, and level of household crowding (indicators of social disadvantage and standard of living), and parental separation or divorce (indicator of psychosocial distress).
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
17 000 live births was followed up at ages 7, 11, 16, 23, 33, and 41 y. We used information from birth to 33 y; the sample was found to be generally representative of the original birth cohort (20, 21).
Measures
Height
The height of each cohort member was measured to the nearest inch by trained medical personnel at ages 7, 11, and 16 y and during a home interview at 33 y of age. Of 16835 cohort members who had at least one height measure, 1512, 2842, 5051, and 7430 participants had 1, 2, 3, or all 4 measurements, respectively. Mother's height (in inches) measured in 1958 was used in the analysis, or reported maternal height in 1969 was used if the mother's height had not been measured. Father's height (in inches) was reported in 1969, in most cases by the mother. All height measures were converted from inches to centimeters.
Prenatal factors
Birth weight was measured in ounces and converted to grams. A measure of birth weight-for-gestational age was derived, ie, birth weight standardized within each gestational week. Maternal smoking during pregnancy was recorded at birth, and those who smoked 
1 cigarettes/d) after the fourth month of pregnancy were coded as smokers.
Postnatal factors
Infant-feeding method, reported in 1965, was categorized as never or ever breastfed; the latter included those who were breastfed for a short time (<1 mo) or longer, but exclusivity of breastfeeding was unknown. Family size at age 7 y was derived from information on the number of household residents aged <21 y, categorized as <3 or 3 children. Household crowding at age 7 y was categorized as <1.5 or 1.5 persons/room. Social class at age 7 y (or at birth if missing) was derived from the occupation of the male head of household and was classified as nonmanual or manual. Age of parental separation or divorce was reported in 1991 and was categorized as divorced or not divorced before age 7 y.
Data analysis
Height measures at ages 7, 11, 16, and 33 y of age were standardized to internally derived SD scores to account for the changes in variance by age. Accordingly, SD scores were comparable across age groups. Paternal and maternal height SD scores were also derived internally and were analyzed separately; in the graphic presentation we used midparental height SD scores that were calculated as the average height SD score of both parents. To examine the relations between early life factors and height at all ages simultaneously, we used multivariate multilevel modeling (22), treating height as a repeated outcome measure. The model can be considered to be 2-level, in which individual measurements are level-1 units and individual participants are level-2 units. In our models, age was treated as a fixed occasion to account for the within-individual correlations. This is an appropriate model when there are large time intervals and a small number of measurements. First, the unadjusted relation between each factor and height was examined. Next, the association was adjusted for parental height (mothers and fathers). To have a better insight into the separate contribution of different factors, we made further adjustments for birth weight, family size, and social class in the final model. We also considered the relations between height and interactions of the following pairs of factors: 1) midparental height and birth weight, 2) midparental height and social class, 3) birth weight and social class, and 4) family size and social class. Rather than assess all possible interactions, we focused on some of the major factors identified in the main-effect model. Furthermore, we calculated the percentage of variance explained by each factor as the change in variance after adding the factor to the model, divided by the total variance in height.
All analyses were restricted to subjects with one or more height measurements, and complete information on parental height, birth weight, family size, social class, and the factor of interest. We repeated the analysis by using the maximum available sample for each model (ie, with the maximum sample always for model 1). The results were similar to those based on one sample for each factor. Hence, the latter is presented here. The adjusted and unadjusted effects (SE) represent increments (or deficits for negative estimates) in height SD scores at each age for a unit increase in the factor of interest, eg, the increase of a parental height SD score, 1 kg of birth weight, or from nonsmoker to smoker. The estimates obtained from the models should be valid because height was found to be missing at random and the missing pattern did not affect the relations studied (data not shown).
The effects of each factor on height on 4 occasions were compared by using a joint contrast test. When the hypothesis of the same strength of the association at 4 ages was rejected, we further tested the differences between each 2 successive ages—ie, between 7 and 11 y, 11 and 16 y, and 16 and 33 y—and also between childhood and adulthood (7 and 33 y) with adjustment for multiple comparisons with the use of the Bonferroni method. The estimated effects of midparental height SD scores, birth weight, and overcrowding were plotted for boys and girls separately to illustrate trends in growth tempo. The analyses were performed by using SAS (version 8.2; SAS Institute Inc, Cary, NC) and MLwiN (IOE, London).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The strong positive correlations between height measures within individuals, which confirms the importance of using multivariate response models, is illustrated in Table 2. A significant association with height was evident at age 7 y for all factors examined (Tables 3 and 4); thereafter, however, the effect of each factor differed by growth stage, which indicated an influence on growth tempo. For boys, the effects of some factors were evident after age 16 y because they were still at different stages of puberty at this time. In contrast, most girls had achieved their final adult height by age 16 y; thus, the effects of early life factors were evident up to 16 y but less so beyond this age.
|
|
|
|
and 4). Similar effects seen in paternal height were 0.35 SD (boys) and 0.34 SD (girls) for height at age 7 y and 0.42 SD (men) and 0.39 SD (women) for adult height. The effect of maternal height attenuated after simultaneous adjustment for paternal height, but substantial effects remained. Likewise, effects of paternal height remained after adjustment for maternal height. Accordingly, midparental height captures the joint effects of both parents and shows stronger effects (0.55–0.72 SD) than does that of one parent (0.31–0.46 SD) (Figure 1). The association of midparental height SD scores with cohort member growth was affected little by adjustment for other factors (Figure 1).
|
and 4 and Figure 2). The association appeared to be constant at all ages, with increases of 0.44–0.55 SD (3 cm) for every 1-kg increase in birth weight; the association weakened slightly after adjustment for parental height, although further adjustment for family size and social class did not change the effect estimates (Tables 3 and 4). The relation with height persisted in the analyses of birth weight-for-gestational age. When we expressed birth weight measures as SD scores, we found that height increased by 1.3–1.7 cm for a 1-SD increase in birth weight compared with increases of 1.0–1.2 cm for every 1-SD increase in birth weight-for gestational age (data not shown).
|
and 4). Boys whose mothers smoked were shorter than were boys whose mothers did not smoke, by 0.19 SD (1.1 cm) at age 7 y; thereafter, however, the association weakened as the offspring of smokers started to catch up. In boys, the difference in height increment between successive ages (ie, 7-11, 11-16, and 16-33 y) was not significant; however, for the longer interval from age 7 y to adulthood, the gain in height for those whose mothers smoked was significantly greater than was that for those whose mothers did not smoke. In girls, the effect of maternal smoking was strongest at age 11 y. The girls whose mother smoked were shorter by a height SD score of 0.19 SD (1.1 cm), although these girls grew significantly faster between ages 11 and 16 y (16.6 compared with 16.1 cm for offspring of nonsmoking mothers). Thus, the relation had weakened by age 16 y. The effect of maternal smoking on height decreased substantially after adjustment for birth weight; 34% (boys) and 43% (girls) of the effect on height at age 7 y was mediated through birth weight. After adjustment for parental height, birth weight, family size, and social class, there was only a small deficit in height—0.08 (0.5 cm) at age 7 y for boys and 0.07 (0.4 cm) at age 11 y for girls whose mother smoked—whereas the association with adult height was abolished (Tables 3 and 4).
Postnatal factors
Cohort members who were breastfed were significantly taller at age 7 y than were those who were not breastfed by a height SD score of 0.17 (1.2 cm) for boys and 0.12 (0.9 cm) for girls, and they remained taller to adulthood (Tables 3 and 4). The effect of breastfeeding was substantially weakened after adjustment for parental height, birth weight, family size, and social class, especially among girls. Family size was significantly associated with height at all ages, with stronger associations for childhood high than for adult height. Children from larger families (3 children) were shorter at age 7 y than were those from smaller families; these children remained shorter until age 11 y, with a deficit of 0.34 SD in boys (2.3 cm) and of 0.29 SD in girls (2.2 cm) (Tables 3 and 4). After age 16 y, boys from larger families gained 0.8 cm more in height (P < 0.001) than did those from smaller families. This finding suggests that the boys continued growing, whereas the girls from larger families gained 1.1 cm more between ages 11 and 16 y (P < 0.001). By adulthood, the height deficit in children from larger families decreased to 0.20 SD (1.4 cm) in men and to 0.15 SD (1.0 cm) in women. The effect of family size on height decreased at all ages after adjustment for parental height, although it remained significant through to adult height.
Similarly, for household crowding, the relation with height was strongest at age 7 y and weakened with increasing age (Figure 3): those from overcrowded households (1.5 persons/room) were shorter than were those from less crowded households by 0.49 SD (2.8 cm) in boys and by 0.54 SD (3.2 cm) in girls (Tables 3 and 4). Thereafter, children from overcrowded households grew faster throughout childhood, thereby reducing their height deficit; boys grew an additional 1.2 cm after age 16 y (P < 0.001) and girls grew an additional 1.5 cm between 11 and 16 y (P < 0.001) compared with their counterparts from less-crowded households. By adulthood, the difference was 0.28 SD (1.9 cm) in men and 0.32 SD (2.1 cm) in women. The effect of household crowding decreased after adjustment for parental height, particularly its effect on adult height, which became nonsignificant in men (Table 4).
|
and 4). The growth deficit among boys from a manual class was 0.34 SD at age 7 y, which decreased significantly to 0.25 in adulthood (P < 0.001). For girls, the growth deficit fluctuated with age; it first decreased from ages 7 to 11 y, then increased to age 16 y, and then decreased thereafter. Effects of social class decreased after adjustment for parental height and, among men, became nonsignificant in adulthood (Table 4).
Among the interaction terms examined, only the interaction between family size and social class had a strong effect on childhood height (P < 0.05), which diminished with increasing age and was no longer significant on adult height (P > 0.10) (Figure 4). At age 7 y, the difference between children from small and large families was 0.12 SD in boys (0.7 cm) and 0.18 SD in girls (1.1 cm) in the nonmanual class, whereas a greater difference of 0.37 SD in boys (2.1 cm) and of 0.32 SD in girls (1.9 cm) was found in the manual class. None of the other interactions examined were significant.
|
). The effect of parental divorce on growth trajectories for boys was short term and changed little after adjustment for parental height, birth weight, family size, and social class. In contrast, a growth deficit in girls whose parents separated or divorced by age 7 y persisted to adulthood (0.21 SD, 1.3 cm) but attenuated and was no longer significant after adjustment for parental height. Adjustments for early life factors further reduced the effect (Table 4). For boys at least, the influence of parental divorce on height depended on the timing of the separation. Growth trajectories in those children whose parents divorced before age 4 y were similar to those whose parents did not divorce, but boys whose parents divorced between ages 4 and 7 y were significantly shorter than their counterparts in childhood and in adulthood (data not shown).
Finally, the percentage of variance in height explained by each factor was examined for ages 7 and 33 y (Table 5). As expected, the greatest percentage of variance in childhood and adult height was explained by parental height, with an increase in the variance explained between the 2 ages. Similarly, the percentage of variance explained by birth weight did not diminish with increasing age, whereas a decrease was seen for indicators of childhood socioeconomic conditions (social class, family size, and household crowding).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1–2 cm were evident for adult height. Third, for parental height and birth weight, associations were evident throughout childhood and were undiminished and even possibly strengthened by adulthood. There was only slight attenuation of the effects after adjustment for other factors, which suggested a primarily genetic influence of parental height and birth weight. The long-term effect of early conditions on growth in height has been neglected in previous work and evidence is therefore scarce, although influences on growth are thought to start in early childhood when growth is particularly sensitive to environmental insults (23). Our study has many strengths and some limitations with respect to influences on growth. The 1958 birth cohort had undergone repeated measurements of height from early childhood through to adulthood and of several early life influences on growth. It is therefore an ideal sample in which to examine how prenatal and early childhood factors influence the entire growth trajectory. Such data allow the use of multivariate multilevel models, which take into account the strong correlation of repeated height measures (correlation coefficients: 0.66–0.92). These models are practical for the number of repeated measurements available in our data. Importantly, in longitudinal studies with sample attrition, the multilevel model has the flexibility to include all subjects with one height measure or more. Thus, most of the study participants are included in the analysis. A limitation of this study, however, is that some (although not all) exposure measures are indicator variables. Information on direct exposures, such as dietary quality and quantity, was not recorded.
Prenatal influences on growth trajectories
Our finding that birth weight is related to subsequent height is consistent with the literature (24, 25). This analysis of the influence of birth weight and child-to-adult growth trajectories in a large cohort showed that instead of diminishing over time, the effect of birth weight remains strong. In this study, we observed an increase of 2 cm in adult height with an increase of 1 kg in birth weight. Effects on height were also seen, though weakened, with a measure of birth weight-for-gestational age, suggesting that birth weight and prematurity both influence height trajectories. The effect of birth weight on height was only slightly attenuated with adjustment for other factors, which suggests a genetic effect of birth weight or that growth and final height are programmed during fetal life. We also found, as did others (26), that even though birth weight is influenced by parental height, its effect on final height is independent of this factor (26). This is consistent with the findings of a study of monozygotic twins (thus controlling for genetic influences), which showed that the twin who was heavier at birth was also taller as an adult (27). The findings for maternal smoking suggest that the fetal environment affects offspring growth. Children prenatally exposed to tobacco are smaller at birth (28–30) and shorter at subsequent ages (31–33), although the growth deficit lessens with increasing age (11, 34, 35). In the 1958 cohort, the rate of height growth was affected by tobacco exposure in utero, as was observed in the current study and in previous work (11). Cohort members born to mothers who smoked were smaller at birth than were those born to mothers who did not smoke (36), but subsequently they grew faster through to the achievement of adult height. Catch-up growth may also have occurred before age 7 y, when the cohort was first measured, as suggested by a British study in which children whose mothers smoked had completed their catch-up growth over the first 2 y of life (37). The mechanisms regulating postnatal catch-up growth are unclear, although there appears to be a strong drive to compensate for an adverse fetal environment, as suggested by others (37).
Postnatal influences on growth trajectories
The literature on infant feeding shows variable effects on height. A British study found that infants who were breastfed during the 1920s to 1940s were taller in childhood and in adulthood than were those who were not breastfed (6), but in more recent generations, such as the 1970 British birth cohort, an effect of breastfeeding on height was explained by social factors (38). In the 1958 cohort, we also found that an effect of breastfeeding appeared to be largely, though not entirely, due to confounding with other influences on height. There was no consistent evidence for a long-term effect of breastfeeding on growth trajectory. Whether there is an effect of breastfeeding on height is likely to depend on the nutritional adequacy of the alternative method of infant feeding. In early British cohorts, breastfeeding may have conferred a benefit, but nowadays differences in growth may have decreased because of improvements in the nutritional adequacy of infant formulas. However, there may be other aspects of diet that have influenced the growth trajectories of this cohort of children born in 1958. The adequacy of diet across a longer period in childhood may well have affected the children's growth, as suggested by our findings for socioeconomic indicators—including social class, family size, and household crowding. All of these factors influenced height, as apparent from the earliest age of measurement, 7 y, and children from manual backgrounds, large families, and overcrowded households had height deficits. Previous studies showed that the effect of family size on childhood height varied by social class (39), and our results confirmed this finding. However, we show here that this relation was no longer evident at later ages. Effects of social class, family size, and overcrowding attenuated slightly from childhood (a deficit of 2–3 cm) through to adulthood, with catch-up growth following on from the period of early delay. Thus, deficits associated with disadvantaged early socioeconomic circumstances had lessened by adulthood, although they were not entirely eliminated (a deficit of 1–2 cm remained in adult height). Early life conditions therefore appeared to influence height partly through an influence on the tempo of growth (40). Poverty, inadequate diet, and childhood infections are likely to underlie the association seen here with social class, family size, and household crowding, but there may have been other factors involved, for example, sleep disturbance associated with overcrowding, which is known to reduce growth hormone secretion (41).
The postnatal environment also includes psychosocial influences, and as a marker of distress, we examined the influence of parental separation or divorce on the child-to-adult growth trajectory. Our study showed that children whose parents separated were shorter than were other children; however, in girls, the deficit in growth was attributable to other factors—possibly to the sharp drop in household income experienced by divorced families (42). There was a stronger effect on childhood stature than on adult stature among boys. Boys whose parents separated between ages 4 and 7 y had delayed growth in childhood, whereas boys whose parents separated before 4 y of age were unaffected. The explanation for sex differences in effect is uncertain, although it is notable that another study based on this cohort (but that used a different psychosocial measure) also suggested that growth was affected only in boys (16). It is known that children with growth deficits can experience spontaneous catch-up growth when they are removed from stressful home circumstances (43), but how emotional distress influences growth is not well understood.
Finally, our study also indicated well-established associations in height between parents and offspring (10, 14). Offspring height was influenced by both mothers and fathers, and the effects were undiminished with increasing age, which appeared to be strongest for adult height than for childhood height. Although the heights of the parents would have been affected by their own childhood environment, it is thought to mainly reflect genotype (7, 10, 39, 44, 45). Our study confirmed the salience of parental measures. In summary, we showed that whereas factors such as parental height and birth weight influence height at each stage from childhood to adulthood, for other factors the effects appear to depend on the life stage examined. Children with growth deficits resulting from unfavorable conditions early in life can experience catch-up growth subsequently, and studies that only examine height in the later phases of growth may fail to detect the full effect of adversity in early life. It is uncertain whether delays in growth have a life-long effect on later health outcomes, but our study suggests that future research should consider growth trajectories when examining the stature-disease relation.
|
ACKNOWLEDGMENTS |
|---|
All of the authors designed the study and wrote the paper. LL conducted the data analysis. None of the authors had a conflict of interest.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. C. Moore, P. Rajaraman, R. Dubrow, A. S. Darefsky, C. Koebnick, A. Hollenbeck, A. Schatzkin, and M. F. Leitzmann Height, Body Mass Index, and Physical Activity in Relation to Glioma Risk Cancer Res., November 1, 2009; 69(21): 8349 - 8355. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. W Lawson and R. Mace Sibling configuration and childhood growth in contemporary British families Int. J. Epidemiol., December 1, 2008; 37(6): 1408 - 1421. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. C. Maty, J. W. Lynch, T. E. Raghunathan, and G. A. Kaplan Childhood Socioeconomic Position, Gender, Adult Body Mass Index, and Incidence of Type 2 Diabetes Mellitus Over 34 Years in the Alameda County Study Am J Public Health, August 1, 2008; 98(8): 1486 - 1494. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E Webb, D Kuh, A Peasey, A Pajak, S Malyutina, R Kubinova, R Topor-Madry, D Denisova, N Capkova, M Marmot, et al. Childhood socioeconomic circumstances and adult height and leg length in central and eastern Europe J Epidemiol Community Health, April 1, 2008; 62(4): 351 - 357. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. B. Terry, Y. Wei, and D. Esserman Maternal, Birth, and Early-Life Influences on Adult Body Size in Women Am. J. Epidemiol., July 1, 2007; 166(1): 5 - 13. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Kuh, R. Hardy, S. Butterworth, L. Okell, M. Richards, M. Wadsworth, C. Cooper, and A. A. Sayer Developmental Origins of Midlife Physical Performance: Evidence from a British Birth Cohort Am. J. Epidemiol., July 15, 2006; 164(2): 110 - 121. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Kuh, R. Hardy, S. Butterworth, L. Okell, M. Wadsworth, C. Cooper, and A. Aihie Sayer Developmental origins of midlife grip strength: findings from a birth cohort study. J. Gerontol. A Biol. Sci. Med. Sci., July 1, 2006; 61(7): 702 - 706. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. P. Gigante, B. L. Horta, R. C. Lima, F. C. Barros, and C. G. Victora Early Life Factors Are Determinants of Female Height at Age 19 Years in a Population-Based Birth Cohort (Pelotas, Brazil) J. Nutr., February 1, 2006; 136(2): 473 - 478. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Power and J. Elliott Cohort profile: 1958 British birth cohort (National Child Development Study) Int. J. Epidemiol., February 1, 2006; 35(1): 34 - 41. [Full Text] [PDF]
|
|
|
|
 |
|
 A. K Adams, H. E Harvey, and R. J Prince Association of maternal smoking with overweight at age 3 y in American Indian children Am. J. Clinical Nutrition, August 1, 2005; 82(2): 393 - 398. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Power, E. Hypponen, and G. Davey Smith Socioeconomic Position in Childhood and Early Adult Life and Risk of Mortality: A Prospective Study of the Mothers of the 1958 British Birth Cohort Am J Public Health, August 1, 2005; 95(8): 1396 - 1402. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. R. Weinberg Invited Commentary: Barker Meets Simpson Am. J. Epidemiol., January 1, 2005; 161(1): 33 - 35. [Full Text] [PDF]
|
|
|
|
|
|
L. Li and C. Power Influences on childhood height: comparing two generations in the 1958 British birth cohort Int. J. Epidemiol., December 1, 2004; 33(6): 1320 - 1328. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Adjusted for birth weight, family size, and social class. unadj, unadjusted; adj, adjusted.
