Several studies of western populations have shown adult height to be negatively associated with morbidity and mortality, in particular from cardiovascular disease (CVD).1^–10 There is also substantial evidence of a positive association between height and the risk of a range of cancers.11 It is believed that experiences in early life underlie the link between height and health, as height is commonly thought to reflect circumstances in childhood.12 Interestingly, data from South Korea, a transitional economy, suggested that the relationship between height and CVD risk, previously assumed to be universal, may be context specific.13

In recent years the focus of research on western populations has turned to whether leg length has a stronger relationship with CVD than full height.14^–17 Both height and leg length, as well as CVD risk, have been shown in these populations to be associated with childhood socioeconomic circumstances.3 15 18^–31 It has been proposed that there are critical periods of leg growth that are particularly sensitive to socioeconomic deprivation32 33 and these may coincide with a critical period of CVD risk, either by a direct effect on CVD risk, or through learned behaviours. If this were the case, leg length would act as a relatively specific indicator of childhood socioeconomic conditions and as an indicator of CVD risk. In order to assess the use of height and its components as an indicator of CVD risk in different populations it is necessary to determine whether the relationships between anthropometry, CVD and childhood socioeconomic status are observed in a variety of settings.

Height in adulthood has been shown in western populations to be negatively associated with various measures of socioeconomic circumstances in childhood, including parent’s educational level,34 father’s socioeconomic status,35^–38 family income,39 family size35 39 and overcrowding in the home.34 35 Height in childhood and adolescence35 40 41 and the tempo of growth throughout childhood35 40 are both similarly associated with the socioeconomic position of origin. Research has shown leg length to be inversely associated with markers of childhood socioeconomic status, but the evidence as to whether it is a more specific correlate of childhood conditions than adult height is mixed.15 18 33^–39 41^–56

The vast majority of evidence on the effect of childhood socioeconomic circumstances on height so far derives from western countries, in particular the United Kingdom34 35 37 38 41 and Sweden.36 39 In these countries different aspects of socioeconomic circumstances, such as income, wealth and education, tend to be strongly mutually correlated, both cross-sectionally and across the life course,48 57 making it difficult to assess which aspect(s) of socioeconomic circumstances in childhood are the most important determinants of adult height. In contrast, the communist regimes in eastern Europe and the former Soviet Union in the second half of the 20th century aimed to narrow the gap in material conditions between different social and educational groups, resulting in relatively weak correlations between material conditions and education.58 Studying anthropometric and socioeconomic indices from people who grew up under these regimes may allow the relative importance of parental education and material conditions in childhood in determining adult height to be assessed.

This paper has two main objectives. First, to investigate the association between two self-reported indices of childhood conditions: parental education and the availability of household assets. Second, to examine whether these self-reported indices are associated with adult height and its components (leg length, trunk length and leg/trunk ratio), and whether leg length is associated with childhood conditions more specifically than full adult height.

METHODS

Study subjects

The data come from the baseline survey of the Health, Alcohol and Psychosocial factors In Eastern Europe (HAPIEE) study. Details of the study have been described elsewhere.59 Briefly, the study recruited nearly 29 000 urban-dwelling men and women aged 45–69 years. They were randomly selected from population registers in Novosibirsk (Russia), Krakow (Poland) and six towns of the Czech Republic. The overall response rate was 67%. This report is based on 24 174 subjects with non-missing data on all variables of interest.

Those who were excluded as a result of missing data were evenly distributed by sex and age. Russian individuals who had fewer amenities in childhood and those whose fathers had a lower level of education were less likely to have missing data. There was no clear association with mother’s education.

Measurements

Participants completed a structured questionnaire that included questions regarding their mother’s and father’s education (not available in the Czech Republic) and ownership of the following household items when they were 10 years old: cold tap water, hot tap water, radio, refrigerator, kitchen and toilet in the household. There were five categories for parent’s education: less than primary, complete primary, secondary, vocational and university. We calculated the number of assets for each individual (range from 0 to 6).

In a subsequent visit to a clinic, the participants underwent a short examination, during which measurements (to the nearest centimetre) were taken of participant’s standing and sitting height (the height from the seat of the chair to the top of the participant’s head while the participant was seated). Sitting height was used as an approximation of trunk length and was subtracted from height to give leg length. The leg to trunk length ratio was calculated. Extensive training and supervision of nurses performing the anthropometric measurements, along with random spot checks of individual nurses, ensured that inter-observer error in these measurements was minimised.

Statistical analysis

The associations between indicators of childhood socioeconomic circumstances (mother’s and father’s education and the number of assets at home at the age of 10 years) and anthropometric measures (height, leg length, trunk length and leg/trunk ratio) were analysed using linear regression. Analyses were conducted separately by country and sex because of the possible modifying effects. Regression analyses were adjusted for the year of birth, to control for age and for any secular trends in anthropometry or socioeconomic circumstances, and for the other measures of childhood socioeconomic circumstances.

RESULTS

Czech men and women were the tallest of the three nationalities and Russian men and women the shortest. The pattern was similar for leg length. Mean trunk length showed little between-country variation and consequently the mean leg/trunk ratio was lowest in Russia and highest in the Czech Republic for both men and women (table 1).

The mean number of assets in the home at the age of 10 years was highest in the Czech Republic and lowest in Russia (table 1). With respect to specific items, possession of all assets other than hot water and radio ownership was lowest in Russia and highest in the Czech Republic (data not shown). Proportions of parents with the upper three levels of education (vocational, secondary, university) were similar in Russia and Poland, but more Russian than Polish individuals had parents who had not completed their primary education (table 1).

In Russia and Poland, where data were available on both parental education and assets, there were strong positive associations between parental educational level and the number of assets in childhood (table 2). The associations were similar for father’s and mother’s education and were stronger in Poland than in Russia.

Table 3 shows the unadjusted association between childhood socioeconomic circumstances and adult height. In both genders, height was linearly associated with both assets in childhood (all countries) and parental education (in Russia and Poland). Similar associations are seen for leg and trunk length but not for leg/trunk ratio (data not shown in table 3).

After adjusting for age, childhood assets were positively and linearly associated with height, leg length and trunk length in both sexes and in all countries (table 4). Adjustment for parent’s education weakened the associations, more so in Poland than in Russia, but the trends, other than between assets and leg length among Polish women, were still statistically significant at the 95% level (table 4). In men, the associations of height, leg length and trunk length with childhood assets were similar across countries. Among women, associations of childhood assets with height and leg length were stronger among Polish individuals but associations with trunk length were stronger among Russian individuals. The leg/trunk ratio did not show a consistent relationship with assets in childhood, although there was some indication of an association among Polish and Czech women (table 4).

After adjusting for age both parents’ educational levels were associated with height, leg length and trunk length in both genders in Poland and Russia (table 4). Further adjustment for assets in childhood weakened the relationships but all, other than the association between leg length and father’s education in Russian women, were still statistically significant at the 95% level (table 4). There was no effect of parent’s education on the leg/trunk ratio in Russian men or women or Polish women, but the educational level of both parents showed an association in Polish men (table 4).

Age and asset-adjusted associations of mother’s education and anthropometric measures are similar to those of father’s education, although Russian women’s leg length remains statistically significantly related to mother’s education after adjustment for assets (table 4).

R2 values show that the contribution of childhood socioeconomic factors to the variation in anthropometric measures is small, but significant (table 4). The proportion of variation explained by childhood socioeconomic factors is greater for height and trunk length than for leg length. None of the variation in the leg/trunk ratio is explained by these factors.

Mother’s and father’s education are very closely linked (correlation coefficient 0.75) so adjusting the association of one parent’s education with an anthropometric measure for the other parent’s education is not appropriate as a result of colinearity. It is therefore not possible to determine whether there are any separate effects of mother’s and father’s education on anthropometric measures.

To assess which anthropometric indices were most closely associated with childhood circumstances, we calculated ratios of the age-adjusted regression coefficients (per unit change in assets or parental education as shown in table 4) to the standard deviation of the given anthropometric measure (fig 1). With the exception of assets in Russian women, height was associated with childhood conditions more strongly than leg length in all comparisons. The leg/trunk ratio was only weakly related to parental education and assets. Leg length appears to be more strongly influenced than trunk length by parents’ education in Poland, but the opposite is true for the Russian population.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Change in anthropometric variables by childhood conditions (relative to standard deviation).

DISCUSSION

To our knowledge, this is the first study of the association of childhood circumstances with adult height and its components in eastern Europe. We found that parental education was associated with household assets at the age of 10 years, and both measures of childhood conditions were associated with height, leg length and trunk length. The findings do not support the hypothesis that leg length is more sensitive to childhood conditions than full adult height.

Perhaps the most important limitation of this study is the retrospective collection of information on childhood circumstances. Although studies from the United Kingdom suggest that recall of childhood conditions is reasonably reliable over four to five decades,60^–62 no such data are available from eastern Europe. Recall bias cannot be excluded, but we tried to minimise it by asking about objective conditions, rather than subjective assessments. On the other hand, one would not expect a major systematic bias in recalling parental education, and it is reassuring that parental education and household assets were significantly associated with anthropometry.

The representativeness of the study samples should also be considered. The subjects were randomly selected from population registers, and are considered to be representative of urban populations. Non-responders were more likely to be male, younger, have lower levels of education, worse self-rated health and to smoke than those who responded.59 Responders and non-responders may also differ in their socioeconomic backgrounds. Responders who had missing values for the variables used in this analysis, and were therefore excluded from this analysis, were shorter, had fathers with higher education and had more assets in childhood. These differences were small and did not affect the results. Although those who were included in the analyses differ slightly from the general populations this does not affect the internal consistency of the findings, and therefore the results (on the association between childhood conditions and anthropometry) should be applicable to the whole (urban) populations. Further support for the representativeness of the HAPIEE samples is the fact that the mean height was similar (slightly taller) and the ranking of the countries was identical to the MONICA data from the same countries.63

A further limitation of the study is the lack of data on some potential confounding factors of the association between childhood socioeconomic circumstances and adult anthropometry. Genetic factors and parent’s height clearly have an influence on height but, unfortunately, data on these factors are not available. The potential for inaccuracies in recalled parental height (and diet and illness in childhood) was considered to be sufficiently great that such data would be unusable. All three study populations are comprised almost exclusively of Slavic people but there are no specific data to show an absence of genetic difference between the populations. Our colleagues in Novosibirsk, however, analysed a limited number of genetic markers related to CVD and found no differences in their distribution between Novosibirsk and other European populations (M Voevoda, personal communication, 2007).

Several of our findings deserve comment. First, the two aspects of childhood conditions on which we have data (parental education and household assets) were strongly associated. This is similar to findings in western populations57 and it suggests that, at least for the generation of parents of our study participants, the assumption of a weak relationship between different dimensions of socioeconomic status in eastern Europe is incorrect.

Second, adult height was smallest in Russia, virtually all the difference between Russia and the other two countries being caused by shorter leg length, and childhood circumstances were also least favourable in Russia. This ecological pattern is consistent with the hypothesis that childhood conditions are related to adult height, because it fits with the social histories of the three countries. The Czech Republic, formerly the more affluent part of Czechoslovakia, was among the most developed countries between the wars (infant mortality 123 per 100064 and life expectancy at birth for men and women combined 57.9 years (M Maly, personal communication, 2007) in 1935) and remained prosperous until 1948. Russia, as part of the former Soviet Union, on the other hand, struggled both before and after World War II, as illustrated by infant mortality of 198 per 1000 live births and life expectancy at birth of 39.6 years (men and women combined) in 1935.65 Poland, not as affluent as Czechoslovakia, occupied an intermediate position. The ranking of the countries, in terms of their development before and around World War II, thus corresponds with their ranking in terms of height. In our data, however, height and leg length were shorter in Russia than the other countries for people of equivalent measured socioeconomic position. This suggests that an important part of inter-country variation in childhood socioeconomic circumstances and growth was not captured by the measures used here.

Third, on the individual level, childhood conditions were associated with adult height, leg length and trunk length. The direction and magnitude of the associations were similar to those observed in most western studies.18 33 34 36 37 66 67 The presence of these associations in eastern European populations further supports the view that the effects of childhood circumstances on adult height are a genuine and general phenomenon not specific to western populations.

Fourth, although parental education and household assets in these data were linked more closely than anticipated, it seems that both types of disadvantage during childhood may independently negatively affect height.

Finally, our data did not confirm that leg length is correlated with childhood conditions more strongly than full height. Adult height seemed to reflect childhood circumstances better than leg length. The associations with the leg/trunk ratio were weak and inconsistent, indicating that the proportion of total height made up of leg and trunk length does not vary systematically across the socioeconomic spectrum. This result contradicts some British studies, which have suggested that leg length is the most plastic anthropometric measure and leg growth that which is most vulnerable to stunting where children are exposed to poor socioeconomic circumstances.18 32 33 These eastern European data suggest that both leg and trunk length are affected by poor socioeconomic circumstances in childhood, and that overall height is the most useful summary measure of the effects of childhood circumstances on growth.