- Email: [email protected]
Estimation of secular trends in adult height, and childhood socioeconomic circumstances in three Eastern European populations
Economics and Human Biology 6 (2008) 228–236
Contents lists available at ScienceDirect
Economics and Human Biology journal homepage: http://www.elsevier.com/locate/ehb
Estimation of secular trends in adult height, and childhood socioeconomic circumstances in three Eastern European populations Elizabeth Alice Webb a,*, Diana Kuh b, Andrzej Pajak c, Ruzena Kubinova d, Sofia Malyutina e, Martin Bobak a a
Department of Epidemiology and Public Health, University College London, 1-19 Torrington Place, London WC1E 6BT, UK The MRC National Survey of Health and Development, Department of Epidemiology and Public Health, University College London, UK Department of Epidemiology and Population Studies, Jagiellonian University, Krakow, Poland d Centre for Environmental Health, National Institute of Public Health, Prague, Czech Republic e Institute of Internal Medicine, Siberian Branch of the Russian Academy of Medical Sciences, Novosibirsk, Russia b c
A R T I C L E I N F O
Article history: Received 27 March 2008 Accepted 27 March 2008 Keywords: Secular trends Adult height Socioeconomic status Russia Poland Czech Republic
A B S T R A C T
The objective of these analyses was to estimate the strength and direction of secular trends in adult height and childhood socioeconomic circumstances in eight towns in three Eastern European countries in the mid-20th century, and to assess the extent to which childhood conditions might explain the height differences. We used cross-sectional data from the baseline survey of the Health, Alcohol and Psychosocial factors in Eastern Europe (HAPIEE) study, conducted in 2002–2005. The study examined 24,012 men and women born between 1933 and 1957, randomly selected from the general populations of Novosibirsk (Russia), Krakow (Poland) and six towns of the Czech Republic. To allow for age-related height loss we estimated maximum attained height. Parental education and household item ownership at age 10 were used as markers of childhood socioeconomic conditions. In all 5-year birth cohorts, Novosibirsk men and women were shortest. There were positive and statistically significant secular trends in childhood conditions and in maximum adult height. Adjustment for childhood conditions explained about one third of the trend in height. There appeared to be a small reduction in height of persons born during the Second World War which was, however, only significant in Novosibirsk. These results suggest that secular trends in height mirror, but are not wholly explained by, trends in socioeconomic circumstances in early life. ß 2008 Elsevier B.V. All rights reserved.
1. Introduction A trend of increasing adult height has been observed in European populations since the middle of the nineteenth century (Komlos, 1985; Floud, 1989), such that, when they reach adult height, children are on average taller than their same-sex parent (Cole, 2003). Such trends have been referred to as secular trends, and defined as ‘‘changes in growth and development of successive generations living in the same territories’’ (Ulijaszek, 1998). The major influences on these secular trends are the affluence and health of populations, therefore trends in height and health outcomes can be linked. In times of hardship, in particular during wars, the secular trend may slow or reverse and this may then be followed by a period of quickening of the trend when conditions improve (Relethford and Lees, 1980; Dubrova et al., 1995; Schmidt et al., 1995). This is not observed in all populations, however, including those who experienced the 1941–44 siege of Leningrad (Sparen et al., 2004).
* Corresponding author. Tel.: +44 20 7679 1680; fax: +44 20 7813 0242. E-mail address: [email protected] (E.A. Webb). 1570-677X/$ – see front matter ß 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.ehb.2008.03.001
E.A. Webb et al. / Economics and Human Biology 6 (2008) 228–236
229
Table 1 Description of cohort, measured 2002–5 Men Czech towns Number of participants Mean measured height (cm) [S.D.] Mean maximum height (cm) [S.D.]
3073 174.9 [6.5] 176.7 [6.3]
Women a
Novosib-irsk
Krakow
Czech townsa
Novosib-irsk
Krakow
3892 171.1 [6.4] 172.9 [6.2]
4295 172.2 [6.3] 173.9 [6.2]
3593 161.9 [6.1] 163.9 [5.8]
4643 158.2 [6.0] 160.2 [5.7]
4516 159.4 [5.9] 161.3 [5.7]
Year of birth 1933–37 (%) 1938–42 (%) 1943–47 (%) 1948–52 (%) 1953–57 (%)
21.2 22.4 19.8 19.3 17.2
17.6 23.3 18.8 20.5 19.9
18.1 19.9 21.5 19.8 20.8
17.4 23.2 19.4 20.7 19.3
17.3 23.1 18.3 20.0 21.3
16.1 19.7 19.8 21.4 23.1
Number of assets 0 (%) 1 (%) 2 (%) 3 (%) 4 (%) 5 (%) 6 (%) Mean [S.D.]
1.1 3.5 9.1 16.6 30.5 18.3 20.9 4.1 [1.4]
9.6 35.5 24.6 10.0 6.2 5.1 8.9 2.2 [1.7]
6.0 13.3 20.9 15.1 12.1 12.0 20.6 3.3 [1.9]
0.6 2.4 9.0 16.6 29.2 19.7 22.6 4.2 [1.4]
9.3 32.5 27.0 11.4 6.8 5.0 8.1 2.2 [1.7]
5.1 11.7 19.8 14.8 13.1 12.4 23.1 3.5 [1.9]
Father’s education Less than primary (%) Primary (%) Vocational (%) Secondary (%) University (%)
– – – – –
19.6 32.5 18.0 20.4 9.6
8.9 42.0 20.9 17.2 11.1
– – – – –
22.0 30.9 19.5 19.3 8.3
8.4 42.2 20.7 17.8 10.9
Mother’s education Less than primary (%) Primary (%) Vocational (%) Secondary (%) University (%)
– – – – –
25.1 31.7 15.9 21.7 5.7
10.5 51.9 13.0 20.9 3.8
– – – – –
28.2 30.3 17.1 20.0 4.4
9.9 52.1 13.6 20.3 4.2
a
Six towns in the Czech Republic (Havirov/Karvina, Hradec Kralove, Jihlava, Kromeriz, Liberec and Usti nad Labem).
Studies of adolescents in Eastern European populations suggest strong secular trends towards increasing height after the Second World War (Bielicki et al., 1992, 2005; Bielicki and Szklarska, 1999). These results are mirrored by studies of children, which have shown continuous increases in the height of Polish children from 1880 to 1990 (Krawczynski et al., 2003) and Czech children between 1950 and 2000 (Vignerova et al., 2006), for both sexes and all age groups. However, evidence of secular trends in height of adults in Eastern Europe, especially of persons born before and around the Second World War, is sparse. Given the social and economic history of the region, such trends could provide evidence of how rates of change are linked to changing socioeconomic circumstances. Perhaps the most extreme example of the dramatic effects of social change on health outcomes is the former Soviet Union. The 1930s were characterized by economic hardship and low life expectancy at birth (around 40 years for most of the decade); however, during the famine in 1933, life expectancy at birth dropped to only 10.3 years for men and 13.0 years for women (Haynes and Husan, 2003).1 World War II was another period of stark economic hardship; while data on life expectancy during the war is not available, in 1947 life expectancy at birth was 34.8 and 46.4 years for men and women respectively (Haynes and Husan, 2003). Estimates of the secular trend in adult height in those born in the post-war period in Russia vary 2 (Dubrova et al., 1995; Mironov, 2007; Brainerd, 2008) but generally suggest a steeper trend than those observed in other European populations over the same period (Relethford and Lees, 1980; Kuh et al., 1991; van Leer et al., 1992; Komlos and Kriwy, 2003). These trends were mirrored by a similarly dramatic increase in life expectancy over the same period (between 1930 and 1958 the combined life expectancy for men and women increased by more than 30 years (Haynes and Husan, 2003)). One would expect that the rapid secular trends to increasing height are at least partly due to the very poor socioeconomic conditions of Russians born in the 1930s and during World War II. A direct test of this hypothesis on an individual level is difficult due to
1 These estimates of life expectany are taken, by Haynes and Husan (2003) from E.M. Andreev, L.E. Darskii and T.L. Kharkova (1993) Naselenie Sovetskogo Soyuza 1922–1991 (Moscow: Nauka). Estimates of life expectancy at birth are particularly low because of the influence of high rates of infant mortality. 2 Dubrova et al. (1995) estimate a mean increase in women’s height of approximately 2 cm/decade between 1945 and 1970. Over the same period Brainerd (unpublished data) estimates men’s height increased by 1.8 cm/decade and women’s by 1.7 cm/decade. Between 1946 and 1972 Mironov (2007) estimates a 2.4 cm/decade mean increase in women’s height.
230
E.A. Webb et al. / Economics and Human Biology 6 (2008) 228–236
Table 2 Calculated cumulative loss of height, from maximum height (cm) Year of birth
Cumulative height change (mean [S.D.])
1933–37 1938–42 1943–47 1948–52 1953–57
3.03 2.37 1.69 1.18 0.74
Men
Women [0.22] [0.22] [0.19] [0.16] [0.12]
3.49 2.68 1.86 1.26 0.75
[0.26] [0.26] [0.22] [0.19] [0.13]
Based on equations from Sorkin et al. (1999b).
lack of data. An indirect method, however, would be a comparison of the secular trends with other populations in Eastern Europe. This paper has two main objectives, the first of which is to quantify trends in childhood socioeconomic status of men and women born between 1933 and 1957 and living in eight urban centres in Russia, Poland and the Czech Republic.3 The secular trends in maximum adult height in these populations will be estimated by adjusting for age-related height loss. The second objective is to estimate how much of the observed trend in adult height can be explained by changes in childhood socioeconomic conditions, and whether World War II had an abrupt effect on the secular trend in adult height. 2. Methods 2.1. Study subjects The data are from the first survey of the HAPIEE (Health, Alcohol and Psychosocial factors In Eastern Europe) study, and were collected between 2002 and 2005. Details of the HAPIEE study have been described elsewhere (Peasey et al., 2006).4 Briefly, the study recruited nearly 29,000 men and women who were born between 1933 and 1957, aged 45–70 years at baseline. They were randomly selected from population registers in Novosibirsk (Russia), Krakow (Poland) and six towns of the Czech Republic which have populations of between 29,000 and 104,000 (Havirov/Karvina, Hradec Kralove, Jihlava, Kromeriz, Liberec and Usti nad Labem). The absolute response rates were 61% in Novosibirsk and Krakow and 55% in the six Czech towns and, after correcting mistakes in population registers, the overall response rate was 67%.5 This report is based on 24,012 participants with non-missing data on all variables of interest (Table 1). 2.2. Measurements Participants completed a structured questionnaire which included questions on demographics and socioeconomic conditions in childhood. In these analyses, we used two main socioeconomic variables: parental education and ownership household items when participants were 10 years old. The items included: cold tap water, hot tap water, radio, refrigerator, kitchen and toilet in the household.6 We used the sum of these items (range from 0 to 6). Parental education was categorised as follows: less than primary, complete primary, secondary, vocational and university. This measure was not available in the Czech part of the study. Participants attended a clinic and underwent a short physical examination, which included measurement of their height, to the nearest 0.1 cm, by a trained nurse using a stadiometer. Participants stood with their feet together and flat on the base of the stadiometer, with their heels centrally placed against the back plate, head tilted to the Frankfort plane position and arms held loosely by their side. 3 Novosibirsk in Russia, Krakow in Poland and six towns in the Czech Republic (Havirov/Karvina, Hradec Kralove, Jihlava, Kromeriz, Liberec and Usti nad Labem). All participants were living in urban centres, so the study is not representative of the whole populations of the countries in which they reside. 4 Data are held at the Department of Epidemiology and Public Health, University College London. Please contact M. Bobak ([email protected]) for further information. 5 Although participants were selected randomly from population registers, the response rate of 67% suggests that some selection bias may have been introduced. 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 (Peasey et al., 2006). Responders and non-responders may also differ in their childhood characteristics, although this cannot be ascertained from the data available. In the Krakow and Czech town study populations, there were no statistically significant differences in height or measures of SEC between responders included in the analyses and those who were excluded because of some missing data on covariates. In Novosibirsk, persons with some missing data on covariates were shorter and had fewer assets in childhood than those who had complete data; these differences are consistent with the pattern in the whole population and too small to affect the results. It is therefore likely that results are applicable to the whole urban populations from which the study participants were selected. 6 As no contemporaneous data were available, participants were asked about ownership of assets which would be easily remembered. A British study showed recall of similar information was reasonably accurate over 50 years (Berney and Blane, 1997) but there is no such study in an Eastern European setting and some inaccuracies of recall are inevitable. A systematic bias in recall of childhood conditions by age would render secular trends in number of household assets owned or in parental education unreliable. However, although a systematic bias is possible, more likely is a random misclassification which would underestimate the effect of childhood conditions on height and thus lead to residual confounding in the secular trends adjusted for childhood conditions.
E.A. Webb et al. / Economics and Human Biology 6 (2008) 228–236
231
Table 3 Mean measured height, mean maximum height and mean number of assets in childhood, by year of birth (S.D. in parentheses) Year of birth
Men Czech towns
Women Novosibirsk
Krakow
Czech towns
Novosibirsk
Krakow
Measured height (cm) 1933–37 172.7 [6.4] 1938–42 174.0 [6.0] 1943–47 175.2 [6.4] 1948–52 176.3 [6.2] 1953–57 176.9 [6.4] p for trend <0.001
169.1 [6.2] 169.4 [5.9] 171.4 [6.2] 172.4 [6.1] 173.1 [6.4] <0.001
170.0 [6.0] 171.0 [6.0] 171.8 [5.9] 173.1 [6.1] 174.9 [6.5] <0.001
159.2 [5.7] 161.0 [5.6] 161.6 [5.9] 163.3 [5.9] 164.3 [5.8] <0.001
155.6 [5.5] 156.5 [5.6] 158.2 [5.7] 159.5 [5.7] 160.8 [5.8] <0.001
157.1 [5.5] 158.2 [5.6] 159.3 [5.6] 160.2 [6.0] 161.4 [5.6] <0.001
Maximum height (cm) 1933–37 175.7 [6.4] 1938–42 176.3 [6.0] 1943–47 176.8 [6.4] 1948–52 177.5 [6.2] 1953–57 177.7 [6.4] p for trend <0.001
172.1 [6.2] 171.9 [5.9] 173.1 [6.2] 173.6 [6.1] 173.9 [6.4] <0.001
173.0 [6.0] 173.3 [6.0] 173.5 [5.9] 174.2 [6.1] 175.6 [6.4] <0.001
162.7 [5.7] 163.6 [5.6] 163.5 [5.9] 164.6 [5.9] 165.0 [5.8] <0.001
159.1 [5.5] 159.3 [5.6] 160.0 [5.7] 160.8 [5.7] 161.6 [5.8] <0.001
160.6 [5.5] 160.8 [5.6] 161.2 [5.6] 161.4 [5.9] 162.1 [5.6] <0.001
1.2 [1.0] 1.5 [1.2] 2.1 [1.5] 2.7 [1.7] 3.4 [2.0] <0.001
2.2 [1.6] 2.7 [1.7] 3.2 [1.8] 3.9 [1.9] 4.4 [1.7] <0.001
3.3 [1.3] 3.7 [1.2] 4.1 [1.2] 4.6 [1.2] 5.2 [1.0] <0.001
1.2 [1.0] 1.5 [1.1] 2.1 [1.4] 2.8 [1.7] 3.3 [1.8] <0.001
2.4 [1.6] 2.9 [1.7] 3.3 [1.8] 3.9 [1.8] 4.5 [1.8] <0.001
Number of assets 1933–37 1938–42 1943–47 1948–52 1953–57 p for trend
3.1 [1.4] 3.7 [1.2] 4.2 [1.2] 4.6 [1.3] 5.1 [1.1] <0.001
2.3. Statistical analysis We also calculated an estimate of maximum achieved adult height because participants’ heights were measured when they were aged between 45 and 70, by which point age-related loss of height would have begun (Cline et al., 1989; Davies et al., 1991; Sorkin et al., 1999a,b; Pini et al., 2001). Height loss begins at around age 40, has been shown to accelerate with increasing age and to be greater in women (Sorkin et al., 1999b). It is therefore appropriate for a study of secular trends in adult height which utilises cross-sectional data to make adjustments to avoid over-estimating the strength of a positive secular trend. Age at time of interview was used to calculate the amount of height lost as cumulative height change (chc) using the following equations (Sorkin et al., 1999b): Men :
chc ¼ 0:0021 age2 þ 0:1258 age 1:8829
Women :
chc ¼ 0:0027 age2 þ 0:1727 age 2:7616
These equations are based on data from 16 longitudinal studies, reviewed in Sorkin and colleagues’ article (Sorkin et al., 1999b), which measured height at least twice during adulthood. They surveyed more than 34,000 men and women aged 18– 97 years from 16 populations in Western Europe, Australia and the USA and another in the Czech Republic. The estimate of lost height was added to measured height to give an estimate of the maximum achieved adult height of participants. The associations between year of birth and both measured height and estimated maximum adult height were examined here using linear regression. Childhood socioeconomic conditions, which might explain part of the secular trend in height, were adjusted for. Childhood conditions have been shown in this and other populations to be positively associated with adult height (Kuh and Wadsworth, 1989; Power et al., 2002; Webb et al., 2008). Due to the rapid expansion of the population of Novosibirsk around WWII, the possible effects on any secular trend in height of migration to the region were also investigated. All analyses were conducted separately by sex and country.
Table 4 Beta coefficient [95% confidence interval] for being born in 1940–45 in a regression model of maximum height against year of birth (restricted to persons born before 1950)
Men Women Both gendersa a
Also adjusted for gender.
Czech towns
Novosibirsk
Krakow
0.14 [0.69, 0.42] 0.07 [0.54, 0.41] 0.10 [0.46, 0.26]
0.38 [0.88, 0.12] 0.34 [0.78, 0.09] 0.36 [0.69, 0.03]
0.06 [0.52, 0.40] 0.05 [0.48, 0.38] 0.05 [0.37, 0.26]
E.A. Webb et al. / Economics and Human Biology 6 (2008) 228–236
232
Table 5 Change [95% confidence interval] in maximum height associated with each 10 year increase in year of birth amongst persons born 1933–57 Gender
Model
Czech towns
Novosibirsk
Krakow
2.24 [1.92, 2.56] 1.79 [1.43, 2.16] –
2.33 [2.05, 2.61] 1.93 [1.62, 2.24] 1.71 [1.39, 2.03]
2.45 [2.19, 2.71] 1.92 [1.63, 2.20] 1.91 [1.63, 2.20]
Unadjusted Adjusted for childhood assets Adjusted for assets and parental education
2.52 [2.25, 2.79] 2.17 [1.85, 2.48] –
2.72 [2.48, 2.95] 2.46 [2.19, 2.72] 2.36 [2.09, 2.63]
2.20 [1.96, 2.43] 1.80 [1.54, 2.05] 1.81 [1.55, 2.06]
Maximum height (cm) Men Unadjusted Adjusted for childhood assets Adjusted for assets and parental education
1.05 [0.73, 1.37] 0.61 [0.24, 0.97] –
1.16 [0.88, 1.44] 0.76 [0.44, 1.07] 0.54 [0.22, 0.86]
1.27 [1.01, 1.53] 0.74 [0.45, 1.02] 0.73 [0.45, 0.10]
1.13 [0.86, 1.40] 0.77 [0.46, 1.08] –
1.33 [1.09, 1.56] 1.06 [0.80, 1.32] 0.96 [0.69, 1.23]
0.80 [0.57, 1.04] 0.40 [0.15, 0.66] 0.42 [0.16, 0.67]
Measured height (cm) Men Unadjusted Adjusted for childhood assets Adjusted for assets and parental education Women
Women
Unadjusted Adjusted for childhood assets Adjusted for assets and parental education
3. Results Men and women in Novosibirsk were the shortest, and men and women in the Czech towns were the tallest (Table 1). The mean number of assets in the home at age ten was also highest in the Czech towns and lowest in Novosibirsk (Table 1). Men and women born earlier lost more height (as calculated using the equations provided above) than those born later, with losses of more than 3 cm in men and nearly 3.5 cm in women aged 65–69 years (Table 2). Mean number of assets in childhood was positively and linearly associated with year of birth in both sexes in all countries (Table 3). The trend is of similar strength in each country and both sexes, with the oldest age group having a mean of 2 fewer assets than the youngest. Mean height increased linearly with year of birth in both genders and in all countries (Table 3). Using maximum adult height attenuated the association with year of birth (Table 3). The long-term linear trends in height were statistically significant but it is note-worthy that amongst Novosibirsk men, the mean maximum height actually decreased between the birth cohorts 1933–37 and 1938–42 (Table 3). We explored further the possible effect on height of being born during the war, restricting the dataset to persons born before 1950 (Table 4). In a linear regression of estimated maximum height against year of birth we included a dummy for being born in 1940–45. There were small negative departures from the longer term trends in maximum height for subjects born during the war, which, however, were only statistically significant in Novosibirsk participants (when men and women were combined). Table 5 shows linear associations of measured/maximum height with year of birth in (i) unadjusted models, and in models (ii) adjusted for household assets and (iii) adjusted for household assets and parental education. The associations of year of birth with maximum height are also illustrated in Figs. 1 and 2. In both sexes and all countries, adjustment for childhood assets led to a substantial attenuation of the associations between year of birth and height, and the relationship was further weakened after controlling for parent’s education. After adjustments, the strength of the secular trend in
Fig. 1. Secular trend in maximum height in men. Mean maximum height for 5 year age groups, OLS of height against year of birth, and regression adjusted for childhood SEC.
E.A. Webb et al. / Economics and Human Biology 6 (2008) 228–236
233
Table 6 Change [95% confidence interval] in maximum height associated with each 10 year increase in year of birth, comparing those born in and outside Novosibirsk Oblast Model
Measured height (cm) Men
Women
Maximum height (cm) Men
Women
a
Born outside Novosibirsk
Born in Novosibirsk
1850/2023 a
2042/2618a
Unadjusted Adjusted for childhood assets Adjusted for assets and parental education
2.05 [1.63, 2.47] 1.61 [1.14, 2.08] 1.45 [0.97, 1.93]
2.67 [2.29, 3.05] 2.31 [1.88, 2.74] 2.01 [1.56, 2.45]
Unadjusted Adjusted for childhood assets Adjusted for assets and parental education
2.84 [2.49, 3.19] 2.71 [2.31, 3.11] 2.57 [2.15, 2.98]
2.69 [2.38, 3.00] 2.34 [1.98, 2.69] 2.29 [1.92, 2.65]
Unadjusted Adjusted for childhood assets Adjusted for assets and parental education
0.88 [0.46, 1.29] 0.42 [0.04, 0.89] 0.26 [0.21, 0.73]
1.52 [1.13, 1.90] 1.15 [0.72, 1.58] 0.85 [0.40, 1.29]
Unadjusted Adjusted for childhood assets Adjusted for assets and parental education
1.43 [1.08, 1.79] 1.30 [0.89, 1.70] 1.14 [0.74, 1.56]
1.31 [1.00, 1.63] 0.95 [0.60, 1.31] 0.90 [0.54, 1.27]
Number of participants (men/women).
maximum height was similar for men in each country (around 0.7 cm per decade); in women, the height increase seemed steepest in Novosibirsk (1.1 cm per decade) and shallowest in Krakow (0.4 cm per decade). Further adjustment for current socioeconomic position did not substantially change these estimates (not shown in table). Even after adjustment for childhood conditions, however, men and women in Novosibirsk remained shortest in each birth cohort (Figs. 1 and 2). Because the population of Novosibirsk expanded rapidly before and during the World War II (Soboleva, 2005) we examined whether there were differences in height trends by migration status. Table 6 shows the difference in the strength of the trend in height by year of birth between those in the Novosibirsk study population who were born in Novosibirsk Oblast (region) and those who migrated there after birth. The trends were stronger in males born in Novosibirsk Oblast, but there were no differences for women. Despite this pattern, however, additional adjustment for migration status did not alter the figures for the Novosibirsk study population shown in Table 5 by more than 1 mm in either direction.
4. Discussion We found evidence of secular trends in adult height amongst persons born between 1933 and 1957 in urban areas of in Russia, Poland and the Czech Republic. The positive secular trend in measured height, derived from cross-sectional data, can be partially explained by loss of height in ageing and by improvements in social conditions in childhood over the birth cohorts covered by the study. However, even after adjustment for these factors, the trends remained positive and statistically significant. The trends were approximately linear and did not generally show variations amongst those born in the years
Fig. 2. Secular trend in maximum height in women. Mean maximum height for 5 year age groups, OLS of height against year of birth, and regression adjusted for childhood SEC.
234
E.A. Webb et al. / Economics and Human Biology 6 (2008) 228–236
immediately preceding, during or following WWII, with a possible exception was Russian subjects born around WWII whose estimated maximum heights were slightly shorter than expected. There are several note-worthy findings in this paper. Firstly, there is a linear trend in assets in the home at age 10. Participants were aged 10 between 1943 and 1967 so this trend indicates a steady improvement in living conditions in the post-war period in Novosibirsk, Krakow and the towns in the Czech Republic in which the study was based. This improvement in social conditions was reflected in Russia by an increase in life expectancy at birth between 1946 and 1956 from 41.5 to 61.0 years in men and 51.0 to 68.9 years in women (Haynes and Husan, 2003). In Poland and the Czech Republic, life expectancy was also improving in the post-war years until the mid 1960s (Zatonski, 1996). Secondly, the secular trend towards increasing height persists after controlling for both age-related loss of height and childhood socioeconomic circumstances. In all three samples studied both men and women show an increase in height of approximately 1 cm per decade, as is suggested by Tanner (1962). After adjustment, the secular trend was strongest amongst Novosibirsk women. The magnitude of these secular trends is within the range of those observed in other European populations (Sobral, 1990; Kuh et al., 1991; Bielicki and Waliszko, 1991; van Leer et al., 1992; Jacome de Castro et al., 1998; Padez and Johnston, 1999; Niewenweg et al., 2003); however, they seem weaker than have been shown previously in eastern European populations. Bielicki and colleagues’ studies of Polish conscripts suggest a secular trend of 2.5 cm per decade between those born in 1946 and 1957 (Bielicki et al., 1992; Bielicki and Szklarska, 1999) (although this may be confounded by a trend to earlier maturation) and Dubrova and colleagues have shown a 1.8 cm per decade increase in height in Russian women born between 1930 and 1965 (Dubrova et al., 1995). Third, although childhood conditions were strongly associated with height across all countries and birth cohorts, and although they improved markedly over time, they failed to explain fully the secular trend in height. Perhaps this is not surprising, given the possible problems associated with measuring childhood conditions retrospectively. The generally linear pattern of the trends in height observed in those born between 1933 and 1957 suggests an absence of major differences in height (and growth) amongst adults born in the period immediately before, during or after WWII. We observed only minor reductions in the estimated maximum heights of people born in the war years (N = 6662), most notably in Novosibirsk, but these differences were small and of low statistical significance. It is interesting that the strongest effect of WWII on growth appears to be amongst the Novosibirsk study population, as this centre was far removed from the violence experienced across much of Europe. There are three potential explanations for the lack of observed effect of WWII on trends in height. Firstly, this study may not have adequate statistical power to identify small effects, as only about one in five participants were born during the war. Secondly, during the wartime years of hardship, there may have been a delay in the onset of decreasing growth rates at the end of adolescence. Vignerova and colleagues noted growth of Czech children slowed and stopped at older ages during the period of food scarcity during and immediately after World War II (Vignerova et al., 2006). This would provide the opportunity for catch-up growth in those whose rate of growth had been slowed. Thirdly, hardship may not affect only a specific birth cohort. This was observed by Brundtland and colleagues, who investigated heights of children in Oslo between 1920 and 1975, and showed that all age groups were similarly affected by WWII and that there was a temporary reversal of the secular trend during that period (Brundtland et al., 1980). It is not clear, unfortunately, whether in those Oslo children the effects were equally long lasting in all age groups, or whether those who experienced the insult earlier in life were able to compensate with catch-up growth. This, however, seems an unlikely explanation for the absence of an interruption in the secular trend in these Eastern European cohorts, as some participants were born more than 10 years after the end of WWII. Although hardships began before, and persisted after the end of the war in most places, it is unlikely that the growth of the youngest participants, who were born in the later 1950s, would have been affected as severely as that of those who lived through the war or were born during it. Several features of this study require consideration when interpreting the results. The major potential limitation of the study is its cross-sectional design and the fact that participant’s height was measured at an age when age-related loss of height had begun. Height loss will be differential by age, with those oldest participants having lost the most height. This leads to an over-estimation of the strength of the secular trend in height. To minimize this problem, adjustments were made for the probable loss of height with age (Sorkin et al., 1999b). The equations used, however, were based on height loss observed in western populations, and it is possible that the age-related shrinkage is different in these Eastern European populations. It is also possible that factors other than age and height, which were not accounted for in the calculation of maximum height, have an effect on height lost in older age. To our knowledge there is no evidence that this is the case, and data from the British Regional Heart Study (Wannamethee et al., 2006) and the Whitehall II study (M. Shipley, unpublished) suggest that agerelated height loss is not related to occupation, education or other measures of adult SEC. The validity of the calculations is supported by the similarity between the estimated maximum heights of the Krakow participants born in 1952–53 and the measured heights of a cohort of 18–19 year olds born in the same years reported in another study (Golab et al., 2002). All participants were living in urban areas. In industrialised countries, urban–rural differences in height are well recognised, and have been consistently observed in Poland (Bielicki et al., 1981; Bielicki, 1986; Bielicki and Waliszko, 1991). Participants cannot, therefore, be considered representative of the whole populations of their countries. Secular trends in height may be of a steeper gradient in rural than urban communities (Bielicki and Waliszko, 1991) and therefore we suggest that the estimates of secular trends reported here may underestimate the overall secular trends of the three countries. A further point for consideration is that the population of Novosibirsk expanded rapidly around the time of WWII, mainly through in migration from other areas of the Soviet Union. In fact, 45% of the Novosibirsk study population was born outside
E.A. Webb et al. / Economics and Human Biology 6 (2008) 228–236
235
the Novosibirsk Oblast (region). As the majority of inhabitants of Novosibirsk originally came from European Russia and are ethnic Russians, genetic differences are unlikely to be an important factor here. The trend to increasing height was stronger amongst men born in the Novosibirsk region, while there were no differences amongst women. However, adjustment for migration status did not suggest that it was an important confounding factor in the relationship between height and year of birth. Admittedly, this is necessarily only a crude assessment of the effect of migration, as the date of, and reasons for, migration are not known. In summary, there are observable secular trends towards increasing height amongst persons born between 1933 and 1957 and living in Novosibirsk, Krakow and six towns in the Czech Republic. These trends are not explained by parallel trends towards improving childhood SEC or by loss of height in ageing, and being born during WWII had only small negative effect, most notably in the Novosibirsk population. Acknowledgements The HAPIEE study was funded by the Wellcome Trust, the MacArthur Foundation initiative on Social Upheaval and Health, and the National Institute of Aging. EW is supported by the UK Medical Research Council studentship. References Berney, L.R., Blane, D.B., 1997. Collecting retrospective data: accuracy of recall after 50 years judged against historical records. Social Science & Medicine 45, 1519–1525. Bielicki, T., 1986. Physical growth as a measure of the economic well-being of populations: the twentieth century. In: Falkner, F., Tanner, J.M. (Eds.), Human Growth, vol. 3. Plenum Press, New York, pp. 283–305. Bielicki, T., Malina, R.M., Waliszko, A., 1992. Monitoring the dynamics of social stratification: statural variation among Polish conscripts in 1976 and 1986. American Journal of Human Biology 4, 345–352. Bielicki, T., Szczotka, H., Charzewski, J., 1981. The influence of three socio-economic factors on body height in Polish military conscripts. Human Biology 53, 543–555. Bielicki, T., Szklarska, A., 1999. The stratifying force of family size, urbanization and parental education in Socialist-era Poland. Journal of Biosocial Science 31, 525–536. Bielicki, T., Szklarska, A., Koziel, S., Ulijaszek, S.J., 2005. Changing patterns of social variation in stature in Poland: effects of transition from a command economy to the free-market system? Journal of Biosocial Science 37, 427–434. Bielicki, T., Waliszko, H., 1991. Urbanization-dependent gradients in stature among Polish conscripts in 1976 and 1986. American Journal of Human Biology 3, 419–424. Brainerd, E., 2008. Reassessing the standard of living in the Soviet Union: an analysis using archival and anthropometric data. Brundtland, G.H., Liestol, K., Walloe, L., 1980. Height, weight and menarcheal age of Oslo schoolchildren during the last 60 years. Annals of Human Biology 7, 307–322. Cline, M.G., Meredith, K.E., Boyer, J.T., Burrows, B., 1989. Decline of height with age in adults in a general-population sample—estimating maximum height and distinguishing birth cohort effects from actual loss of stature with aging. Human Biology 61, 415–425. Cole, T.J., 2003. The secular trend in human physical growth: a biological view. Economics & Human Biology 1, 161–168. Davies, K.M., Recker, R.R., Stegman, M.R., Heaney, R.P., 1991. Tallness versus shrinkage—do women shrink with age or grow taller with recent birth date. Journal of Bone and Mineral Research 6, 1115–1120. Dubrova, Y.E., Kurbatova, O.L., Kholod, O.N., Prokhorovskaya, V.D., 1995. Secular growth trend in two generations of the Russian population. Human Biology 67, 755–767. Floud, R., 1989. Measuring European inequality, the use of height data. In: Fox, A.J. (Ed.), Health Inequalities in European Countries. Gower, Aldershot. Golab, S., Chrzanowska, M., Zarow, J., Sobiecki, J., Brudecki, J., 2002. Zmiany sekularne w rozwoju fizycznym dzieci i mlodziezy krakowskiej w ostatnim 30leciu XX wieku. Wychowanie Fizyczne i Sport 46, 301–314. Haynes, M., Husan, R., 2003. A Century of State Murder? Death and Policy in Twentieth-Century Russia. Pluto Press, London. Jacome de Castro, J., Aleixo Dias, J., Baptista, F., Garcia e Costa, J., Galvao-Teles, A., Camilo-Alves, A., 1998. Secular trends of weight, height and obesity in cohorts of young Portuguese males in the District of Lisbon: 1960–1990. European Journal of Epidemiology 14, 299–303. Komlos, J., 1985. Stature and nutrition in the Habsburg Monarchy: the standard of living and economic development in the Eighteenth Century. American Historical Review 90, 1149–1161. Komlos, J., Kriwy, P., 2003. The biological standard of living in the two Germanies. German Economic Review 4, 459–473. Krawczynski, M., Walkowiak, J., Krzyzaniak, A., 2003. Secular changes in body height and weight in children and adolescents in Poznan, Poland between 1880 and 2000. Acta Paediatrica 92, 277–282. Kuh, D.L., Power, C., Rodgers, B., 1991. Secular trends in social-class and sex-differences in adult height. International Journal of Epidemiology 20, 1001– 1009. Kuh, D.L., Wadsworth, M., 1989. Parental height—childhood environment and subsequent adult height in a national birth cohort. International Journal of Epidemiology 18, 663–668. Mironov, B., 2007. Birth weight and physical stature in St. Petersburg: living standards of women in Russia, 1980–2005. Economics & Human Biology 5, 123– 143. Niewenweg, R., Smit, M.L., Walenkamp, M.J., Wit, J.M., 2003. Adult height corrected for shrinking and secular trend. Annals of Human Biology 30, 563–569. Padez, C., Johnston, F.E., 1999. Secular trends in male adult height 1904–1996 in relation to place of residence and parent’s educational level in Portugal. Annals of Human Biology 26, 287–298. Peasey, A., Bobak, M., Kubinova, R., Malyutina, S., Pajak, A., Tamosiunas, A., Pikhart, H., Nicholson, A., Marmot, M., 2006. Determinants of cardiovascular disease and other non-communicable diseases in Central and Eastern Europe: rationale and design of the HAPIEE study. BMC Public Health 6, 255–264. Pini, R., Tonon, E., Cavallini, M.C., Bencini, F., Di Bari, M., Masotti, G., Marchionni, N., 2001. Accuracy of equations for predicting stature from knee height, and assessment of statural loss in an older Italian population. Journals of Gerontology Series A-Biological Sciences and Medical Sciences 56, B3–B7. Power, C., Manor, O., Li, L., 2002. Are inequalities in height underestimated by adult social position? Effects of changing social structure and height selection in a cohort study. British Medical Journal 325, 131–134. Relethford, J.H., Lees, F.C., 1980. The effects of ageing and secular trend on stature in rural western Ireland. American Journal of Physical Anthropology 55, 81–88. Schmidt, I.M., Jorgensen, M.H., Michaelsen, K.F., 1995. Height of conscripts in Europe: is postneonatal mortality a predictor? Annals of Human Biology 22, 57–67. Soboleva, S., 2005. Economic migration to Western Siberia. Russian Regional Perspectives Journal 1, 45–47. Sobral, F., 1990. Secular changes in stature in Southern Portugal between 1930 and 1980 according to conscript data. Human Biology 62, 491–504.
236
E.A. Webb et al. / Economics and Human Biology 6 (2008) 228–236
Sorkin, J.D., Muller, D.C., Andres, R., 1999a. Longitudinal change in height of men and women: implications for interpretation of the body mass index—the Baltimore longitudinal study of aging. American Journal of Epidemiology 150, 969–977. Sorkin, J.D., Muller, D.C., Andres, R., 1999b. Longitudinal change in the heights of men and women: consequential effects on body mass index. Epidemiologic Reviews 21, 247–260. Sparen, P., Vagero, D., Shestov, D., Plavinskaja, S., Parfenova, N., Hoptiar, V., Paturot, D., Galanti, M.R., 2004. Long term mortality after severe starvation during the siege of Leningrad: prospective cohort study. British Medical Journal 328, 11. Tanner, J.M., 1962. Growth at Adolescence. Blackwell, Oxford. Ulijaszek, S.J., 1998. The secular trend. In: Ulijaszek, S.J., Johnston, F.E., Preece, M.A. (Eds.), The Cambridge Encyclopaedia of Human Growth and Development. Cambridge University Press, Cambridge, pp. 395–398. van Leer, E.M., van Noord, P.A., Seidell, J.C., 1992. Components of adult height and height loss. Secular trend and effects of ageing in women in the DOM project. Annals of Epidemiology 2, 611–615. Vignerova, J., Brabec, M., Blaha, P., 2006. Two centuries of growth among Czech children and youth. Economics & Human Biology 4, 237–252. Wannamethee, S.G., Shaper, A.G., Lennon, L., Whincup, P.H., 2006. Height loss in older men: associations with total mortality and incidence of cardiovascular disease. Archives of Internal Medicine 166, 2546–2552. Webb, E., Kuh, D., Peasey, A., Pajak, A., Malyutina, S., Kubinova, R., Topor-Madry, R., Denisova, D., Capkova, N., Marmot, M., Bobak, M., 2008. Childhood socioeconomic circumstances and adult height and leg length in central and eastern Europe. Journal of Epidemiology and Community Health 62, 351– 357. Zatonski, W., 1996. Evolution of Health in Poland Since 1988. The Institute of Oncology, Warsaw, Poland.
Contents lists available at ScienceDirect
Economics and Human Biology journal homepage: http://www.elsevier.com/locate/ehb
Estimation of secular trends in adult height, and childhood socioeconomic circumstances in three Eastern European populations Elizabeth Alice Webb a,*, Diana Kuh b, Andrzej Pajak c, Ruzena Kubinova d, Sofia Malyutina e, Martin Bobak a a
Department of Epidemiology and Public Health, University College London, 1-19 Torrington Place, London WC1E 6BT, UK The MRC National Survey of Health and Development, Department of Epidemiology and Public Health, University College London, UK Department of Epidemiology and Population Studies, Jagiellonian University, Krakow, Poland d Centre for Environmental Health, National Institute of Public Health, Prague, Czech Republic e Institute of Internal Medicine, Siberian Branch of the Russian Academy of Medical Sciences, Novosibirsk, Russia b c
A R T I C L E I N F O
Article history: Received 27 March 2008 Accepted 27 March 2008 Keywords: Secular trends Adult height Socioeconomic status Russia Poland Czech Republic
A B S T R A C T
The objective of these analyses was to estimate the strength and direction of secular trends in adult height and childhood socioeconomic circumstances in eight towns in three Eastern European countries in the mid-20th century, and to assess the extent to which childhood conditions might explain the height differences. We used cross-sectional data from the baseline survey of the Health, Alcohol and Psychosocial factors in Eastern Europe (HAPIEE) study, conducted in 2002–2005. The study examined 24,012 men and women born between 1933 and 1957, randomly selected from the general populations of Novosibirsk (Russia), Krakow (Poland) and six towns of the Czech Republic. To allow for age-related height loss we estimated maximum attained height. Parental education and household item ownership at age 10 were used as markers of childhood socioeconomic conditions. In all 5-year birth cohorts, Novosibirsk men and women were shortest. There were positive and statistically significant secular trends in childhood conditions and in maximum adult height. Adjustment for childhood conditions explained about one third of the trend in height. There appeared to be a small reduction in height of persons born during the Second World War which was, however, only significant in Novosibirsk. These results suggest that secular trends in height mirror, but are not wholly explained by, trends in socioeconomic circumstances in early life. ß 2008 Elsevier B.V. All rights reserved.
1. Introduction A trend of increasing adult height has been observed in European populations since the middle of the nineteenth century (Komlos, 1985; Floud, 1989), such that, when they reach adult height, children are on average taller than their same-sex parent (Cole, 2003). Such trends have been referred to as secular trends, and defined as ‘‘changes in growth and development of successive generations living in the same territories’’ (Ulijaszek, 1998). The major influences on these secular trends are the affluence and health of populations, therefore trends in height and health outcomes can be linked. In times of hardship, in particular during wars, the secular trend may slow or reverse and this may then be followed by a period of quickening of the trend when conditions improve (Relethford and Lees, 1980; Dubrova et al., 1995; Schmidt et al., 1995). This is not observed in all populations, however, including those who experienced the 1941–44 siege of Leningrad (Sparen et al., 2004).
* Corresponding author. Tel.: +44 20 7679 1680; fax: +44 20 7813 0242. E-mail address: [email protected] (E.A. Webb). 1570-677X/$ – see front matter ß 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.ehb.2008.03.001
E.A. Webb et al. / Economics and Human Biology 6 (2008) 228–236
229
Table 1 Description of cohort, measured 2002–5 Men Czech towns Number of participants Mean measured height (cm) [S.D.] Mean maximum height (cm) [S.D.]
3073 174.9 [6.5] 176.7 [6.3]
Women a
Novosib-irsk
Krakow
Czech townsa
Novosib-irsk
Krakow
3892 171.1 [6.4] 172.9 [6.2]
4295 172.2 [6.3] 173.9 [6.2]
3593 161.9 [6.1] 163.9 [5.8]
4643 158.2 [6.0] 160.2 [5.7]
4516 159.4 [5.9] 161.3 [5.7]
Year of birth 1933–37 (%) 1938–42 (%) 1943–47 (%) 1948–52 (%) 1953–57 (%)
21.2 22.4 19.8 19.3 17.2
17.6 23.3 18.8 20.5 19.9
18.1 19.9 21.5 19.8 20.8
17.4 23.2 19.4 20.7 19.3
17.3 23.1 18.3 20.0 21.3
16.1 19.7 19.8 21.4 23.1
Number of assets 0 (%) 1 (%) 2 (%) 3 (%) 4 (%) 5 (%) 6 (%) Mean [S.D.]
1.1 3.5 9.1 16.6 30.5 18.3 20.9 4.1 [1.4]
9.6 35.5 24.6 10.0 6.2 5.1 8.9 2.2 [1.7]
6.0 13.3 20.9 15.1 12.1 12.0 20.6 3.3 [1.9]
0.6 2.4 9.0 16.6 29.2 19.7 22.6 4.2 [1.4]
9.3 32.5 27.0 11.4 6.8 5.0 8.1 2.2 [1.7]
5.1 11.7 19.8 14.8 13.1 12.4 23.1 3.5 [1.9]
Father’s education Less than primary (%) Primary (%) Vocational (%) Secondary (%) University (%)
– – – – –
19.6 32.5 18.0 20.4 9.6
8.9 42.0 20.9 17.2 11.1
– – – – –
22.0 30.9 19.5 19.3 8.3
8.4 42.2 20.7 17.8 10.9
Mother’s education Less than primary (%) Primary (%) Vocational (%) Secondary (%) University (%)
– – – – –
25.1 31.7 15.9 21.7 5.7
10.5 51.9 13.0 20.9 3.8
– – – – –
28.2 30.3 17.1 20.0 4.4
9.9 52.1 13.6 20.3 4.2
a
Six towns in the Czech Republic (Havirov/Karvina, Hradec Kralove, Jihlava, Kromeriz, Liberec and Usti nad Labem).
Studies of adolescents in Eastern European populations suggest strong secular trends towards increasing height after the Second World War (Bielicki et al., 1992, 2005; Bielicki and Szklarska, 1999). These results are mirrored by studies of children, which have shown continuous increases in the height of Polish children from 1880 to 1990 (Krawczynski et al., 2003) and Czech children between 1950 and 2000 (Vignerova et al., 2006), for both sexes and all age groups. However, evidence of secular trends in height of adults in Eastern Europe, especially of persons born before and around the Second World War, is sparse. Given the social and economic history of the region, such trends could provide evidence of how rates of change are linked to changing socioeconomic circumstances. Perhaps the most extreme example of the dramatic effects of social change on health outcomes is the former Soviet Union. The 1930s were characterized by economic hardship and low life expectancy at birth (around 40 years for most of the decade); however, during the famine in 1933, life expectancy at birth dropped to only 10.3 years for men and 13.0 years for women (Haynes and Husan, 2003).1 World War II was another period of stark economic hardship; while data on life expectancy during the war is not available, in 1947 life expectancy at birth was 34.8 and 46.4 years for men and women respectively (Haynes and Husan, 2003). Estimates of the secular trend in adult height in those born in the post-war period in Russia vary 2 (Dubrova et al., 1995; Mironov, 2007; Brainerd, 2008) but generally suggest a steeper trend than those observed in other European populations over the same period (Relethford and Lees, 1980; Kuh et al., 1991; van Leer et al., 1992; Komlos and Kriwy, 2003). These trends were mirrored by a similarly dramatic increase in life expectancy over the same period (between 1930 and 1958 the combined life expectancy for men and women increased by more than 30 years (Haynes and Husan, 2003)). One would expect that the rapid secular trends to increasing height are at least partly due to the very poor socioeconomic conditions of Russians born in the 1930s and during World War II. A direct test of this hypothesis on an individual level is difficult due to
1 These estimates of life expectany are taken, by Haynes and Husan (2003) from E.M. Andreev, L.E. Darskii and T.L. Kharkova (1993) Naselenie Sovetskogo Soyuza 1922–1991 (Moscow: Nauka). Estimates of life expectancy at birth are particularly low because of the influence of high rates of infant mortality. 2 Dubrova et al. (1995) estimate a mean increase in women’s height of approximately 2 cm/decade between 1945 and 1970. Over the same period Brainerd (unpublished data) estimates men’s height increased by 1.8 cm/decade and women’s by 1.7 cm/decade. Between 1946 and 1972 Mironov (2007) estimates a 2.4 cm/decade mean increase in women’s height.
230
E.A. Webb et al. / Economics and Human Biology 6 (2008) 228–236
Table 2 Calculated cumulative loss of height, from maximum height (cm) Year of birth
Cumulative height change (mean [S.D.])
1933–37 1938–42 1943–47 1948–52 1953–57
3.03 2.37 1.69 1.18 0.74
Men
Women [0.22] [0.22] [0.19] [0.16] [0.12]
3.49 2.68 1.86 1.26 0.75
[0.26] [0.26] [0.22] [0.19] [0.13]
Based on equations from Sorkin et al. (1999b).
lack of data. An indirect method, however, would be a comparison of the secular trends with other populations in Eastern Europe. This paper has two main objectives, the first of which is to quantify trends in childhood socioeconomic status of men and women born between 1933 and 1957 and living in eight urban centres in Russia, Poland and the Czech Republic.3 The secular trends in maximum adult height in these populations will be estimated by adjusting for age-related height loss. The second objective is to estimate how much of the observed trend in adult height can be explained by changes in childhood socioeconomic conditions, and whether World War II had an abrupt effect on the secular trend in adult height. 2. Methods 2.1. Study subjects The data are from the first survey of the HAPIEE (Health, Alcohol and Psychosocial factors In Eastern Europe) study, and were collected between 2002 and 2005. Details of the HAPIEE study have been described elsewhere (Peasey et al., 2006).4 Briefly, the study recruited nearly 29,000 men and women who were born between 1933 and 1957, aged 45–70 years at baseline. They were randomly selected from population registers in Novosibirsk (Russia), Krakow (Poland) and six towns of the Czech Republic which have populations of between 29,000 and 104,000 (Havirov/Karvina, Hradec Kralove, Jihlava, Kromeriz, Liberec and Usti nad Labem). The absolute response rates were 61% in Novosibirsk and Krakow and 55% in the six Czech towns and, after correcting mistakes in population registers, the overall response rate was 67%.5 This report is based on 24,012 participants with non-missing data on all variables of interest (Table 1). 2.2. Measurements Participants completed a structured questionnaire which included questions on demographics and socioeconomic conditions in childhood. In these analyses, we used two main socioeconomic variables: parental education and ownership household items when participants were 10 years old. The items included: cold tap water, hot tap water, radio, refrigerator, kitchen and toilet in the household.6 We used the sum of these items (range from 0 to 6). Parental education was categorised as follows: less than primary, complete primary, secondary, vocational and university. This measure was not available in the Czech part of the study. Participants attended a clinic and underwent a short physical examination, which included measurement of their height, to the nearest 0.1 cm, by a trained nurse using a stadiometer. Participants stood with their feet together and flat on the base of the stadiometer, with their heels centrally placed against the back plate, head tilted to the Frankfort plane position and arms held loosely by their side. 3 Novosibirsk in Russia, Krakow in Poland and six towns in the Czech Republic (Havirov/Karvina, Hradec Kralove, Jihlava, Kromeriz, Liberec and Usti nad Labem). All participants were living in urban centres, so the study is not representative of the whole populations of the countries in which they reside. 4 Data are held at the Department of Epidemiology and Public Health, University College London. Please contact M. Bobak ([email protected]) for further information. 5 Although participants were selected randomly from population registers, the response rate of 67% suggests that some selection bias may have been introduced. 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 (Peasey et al., 2006). Responders and non-responders may also differ in their childhood characteristics, although this cannot be ascertained from the data available. In the Krakow and Czech town study populations, there were no statistically significant differences in height or measures of SEC between responders included in the analyses and those who were excluded because of some missing data on covariates. In Novosibirsk, persons with some missing data on covariates were shorter and had fewer assets in childhood than those who had complete data; these differences are consistent with the pattern in the whole population and too small to affect the results. It is therefore likely that results are applicable to the whole urban populations from which the study participants were selected. 6 As no contemporaneous data were available, participants were asked about ownership of assets which would be easily remembered. A British study showed recall of similar information was reasonably accurate over 50 years (Berney and Blane, 1997) but there is no such study in an Eastern European setting and some inaccuracies of recall are inevitable. A systematic bias in recall of childhood conditions by age would render secular trends in number of household assets owned or in parental education unreliable. However, although a systematic bias is possible, more likely is a random misclassification which would underestimate the effect of childhood conditions on height and thus lead to residual confounding in the secular trends adjusted for childhood conditions.
E.A. Webb et al. / Economics and Human Biology 6 (2008) 228–236
231
Table 3 Mean measured height, mean maximum height and mean number of assets in childhood, by year of birth (S.D. in parentheses) Year of birth
Men Czech towns
Women Novosibirsk
Krakow
Czech towns
Novosibirsk
Krakow
Measured height (cm) 1933–37 172.7 [6.4] 1938–42 174.0 [6.0] 1943–47 175.2 [6.4] 1948–52 176.3 [6.2] 1953–57 176.9 [6.4] p for trend <0.001
169.1 [6.2] 169.4 [5.9] 171.4 [6.2] 172.4 [6.1] 173.1 [6.4] <0.001
170.0 [6.0] 171.0 [6.0] 171.8 [5.9] 173.1 [6.1] 174.9 [6.5] <0.001
159.2 [5.7] 161.0 [5.6] 161.6 [5.9] 163.3 [5.9] 164.3 [5.8] <0.001
155.6 [5.5] 156.5 [5.6] 158.2 [5.7] 159.5 [5.7] 160.8 [5.8] <0.001
157.1 [5.5] 158.2 [5.6] 159.3 [5.6] 160.2 [6.0] 161.4 [5.6] <0.001
Maximum height (cm) 1933–37 175.7 [6.4] 1938–42 176.3 [6.0] 1943–47 176.8 [6.4] 1948–52 177.5 [6.2] 1953–57 177.7 [6.4] p for trend <0.001
172.1 [6.2] 171.9 [5.9] 173.1 [6.2] 173.6 [6.1] 173.9 [6.4] <0.001
173.0 [6.0] 173.3 [6.0] 173.5 [5.9] 174.2 [6.1] 175.6 [6.4] <0.001
162.7 [5.7] 163.6 [5.6] 163.5 [5.9] 164.6 [5.9] 165.0 [5.8] <0.001
159.1 [5.5] 159.3 [5.6] 160.0 [5.7] 160.8 [5.7] 161.6 [5.8] <0.001
160.6 [5.5] 160.8 [5.6] 161.2 [5.6] 161.4 [5.9] 162.1 [5.6] <0.001
1.2 [1.0] 1.5 [1.2] 2.1 [1.5] 2.7 [1.7] 3.4 [2.0] <0.001
2.2 [1.6] 2.7 [1.7] 3.2 [1.8] 3.9 [1.9] 4.4 [1.7] <0.001
3.3 [1.3] 3.7 [1.2] 4.1 [1.2] 4.6 [1.2] 5.2 [1.0] <0.001
1.2 [1.0] 1.5 [1.1] 2.1 [1.4] 2.8 [1.7] 3.3 [1.8] <0.001
2.4 [1.6] 2.9 [1.7] 3.3 [1.8] 3.9 [1.8] 4.5 [1.8] <0.001
Number of assets 1933–37 1938–42 1943–47 1948–52 1953–57 p for trend
3.1 [1.4] 3.7 [1.2] 4.2 [1.2] 4.6 [1.3] 5.1 [1.1] <0.001
2.3. Statistical analysis We also calculated an estimate of maximum achieved adult height because participants’ heights were measured when they were aged between 45 and 70, by which point age-related loss of height would have begun (Cline et al., 1989; Davies et al., 1991; Sorkin et al., 1999a,b; Pini et al., 2001). Height loss begins at around age 40, has been shown to accelerate with increasing age and to be greater in women (Sorkin et al., 1999b). It is therefore appropriate for a study of secular trends in adult height which utilises cross-sectional data to make adjustments to avoid over-estimating the strength of a positive secular trend. Age at time of interview was used to calculate the amount of height lost as cumulative height change (chc) using the following equations (Sorkin et al., 1999b): Men :
chc ¼ 0:0021 age2 þ 0:1258 age 1:8829
Women :
chc ¼ 0:0027 age2 þ 0:1727 age 2:7616
These equations are based on data from 16 longitudinal studies, reviewed in Sorkin and colleagues’ article (Sorkin et al., 1999b), which measured height at least twice during adulthood. They surveyed more than 34,000 men and women aged 18– 97 years from 16 populations in Western Europe, Australia and the USA and another in the Czech Republic. The estimate of lost height was added to measured height to give an estimate of the maximum achieved adult height of participants. The associations between year of birth and both measured height and estimated maximum adult height were examined here using linear regression. Childhood socioeconomic conditions, which might explain part of the secular trend in height, were adjusted for. Childhood conditions have been shown in this and other populations to be positively associated with adult height (Kuh and Wadsworth, 1989; Power et al., 2002; Webb et al., 2008). Due to the rapid expansion of the population of Novosibirsk around WWII, the possible effects on any secular trend in height of migration to the region were also investigated. All analyses were conducted separately by sex and country.
Table 4 Beta coefficient [95% confidence interval] for being born in 1940–45 in a regression model of maximum height against year of birth (restricted to persons born before 1950)
Men Women Both gendersa a
Also adjusted for gender.
Czech towns
Novosibirsk
Krakow
0.14 [0.69, 0.42] 0.07 [0.54, 0.41] 0.10 [0.46, 0.26]
0.38 [0.88, 0.12] 0.34 [0.78, 0.09] 0.36 [0.69, 0.03]
0.06 [0.52, 0.40] 0.05 [0.48, 0.38] 0.05 [0.37, 0.26]
E.A. Webb et al. / Economics and Human Biology 6 (2008) 228–236
232
Table 5 Change [95% confidence interval] in maximum height associated with each 10 year increase in year of birth amongst persons born 1933–57 Gender
Model
Czech towns
Novosibirsk
Krakow
2.24 [1.92, 2.56] 1.79 [1.43, 2.16] –
2.33 [2.05, 2.61] 1.93 [1.62, 2.24] 1.71 [1.39, 2.03]
2.45 [2.19, 2.71] 1.92 [1.63, 2.20] 1.91 [1.63, 2.20]
Unadjusted Adjusted for childhood assets Adjusted for assets and parental education
2.52 [2.25, 2.79] 2.17 [1.85, 2.48] –
2.72 [2.48, 2.95] 2.46 [2.19, 2.72] 2.36 [2.09, 2.63]
2.20 [1.96, 2.43] 1.80 [1.54, 2.05] 1.81 [1.55, 2.06]
Maximum height (cm) Men Unadjusted Adjusted for childhood assets Adjusted for assets and parental education
1.05 [0.73, 1.37] 0.61 [0.24, 0.97] –
1.16 [0.88, 1.44] 0.76 [0.44, 1.07] 0.54 [0.22, 0.86]
1.27 [1.01, 1.53] 0.74 [0.45, 1.02] 0.73 [0.45, 0.10]
1.13 [0.86, 1.40] 0.77 [0.46, 1.08] –
1.33 [1.09, 1.56] 1.06 [0.80, 1.32] 0.96 [0.69, 1.23]
0.80 [0.57, 1.04] 0.40 [0.15, 0.66] 0.42 [0.16, 0.67]
Measured height (cm) Men Unadjusted Adjusted for childhood assets Adjusted for assets and parental education Women
Women
Unadjusted Adjusted for childhood assets Adjusted for assets and parental education
3. Results Men and women in Novosibirsk were the shortest, and men and women in the Czech towns were the tallest (Table 1). The mean number of assets in the home at age ten was also highest in the Czech towns and lowest in Novosibirsk (Table 1). Men and women born earlier lost more height (as calculated using the equations provided above) than those born later, with losses of more than 3 cm in men and nearly 3.5 cm in women aged 65–69 years (Table 2). Mean number of assets in childhood was positively and linearly associated with year of birth in both sexes in all countries (Table 3). The trend is of similar strength in each country and both sexes, with the oldest age group having a mean of 2 fewer assets than the youngest. Mean height increased linearly with year of birth in both genders and in all countries (Table 3). Using maximum adult height attenuated the association with year of birth (Table 3). The long-term linear trends in height were statistically significant but it is note-worthy that amongst Novosibirsk men, the mean maximum height actually decreased between the birth cohorts 1933–37 and 1938–42 (Table 3). We explored further the possible effect on height of being born during the war, restricting the dataset to persons born before 1950 (Table 4). In a linear regression of estimated maximum height against year of birth we included a dummy for being born in 1940–45. There were small negative departures from the longer term trends in maximum height for subjects born during the war, which, however, were only statistically significant in Novosibirsk participants (when men and women were combined). Table 5 shows linear associations of measured/maximum height with year of birth in (i) unadjusted models, and in models (ii) adjusted for household assets and (iii) adjusted for household assets and parental education. The associations of year of birth with maximum height are also illustrated in Figs. 1 and 2. In both sexes and all countries, adjustment for childhood assets led to a substantial attenuation of the associations between year of birth and height, and the relationship was further weakened after controlling for parent’s education. After adjustments, the strength of the secular trend in
Fig. 1. Secular trend in maximum height in men. Mean maximum height for 5 year age groups, OLS of height against year of birth, and regression adjusted for childhood SEC.
E.A. Webb et al. / Economics and Human Biology 6 (2008) 228–236
233
Table 6 Change [95% confidence interval] in maximum height associated with each 10 year increase in year of birth, comparing those born in and outside Novosibirsk Oblast Model
Measured height (cm) Men
Women
Maximum height (cm) Men
Women
a
Born outside Novosibirsk
Born in Novosibirsk
1850/2023 a
2042/2618a
Unadjusted Adjusted for childhood assets Adjusted for assets and parental education
2.05 [1.63, 2.47] 1.61 [1.14, 2.08] 1.45 [0.97, 1.93]
2.67 [2.29, 3.05] 2.31 [1.88, 2.74] 2.01 [1.56, 2.45]
Unadjusted Adjusted for childhood assets Adjusted for assets and parental education
2.84 [2.49, 3.19] 2.71 [2.31, 3.11] 2.57 [2.15, 2.98]
2.69 [2.38, 3.00] 2.34 [1.98, 2.69] 2.29 [1.92, 2.65]
Unadjusted Adjusted for childhood assets Adjusted for assets and parental education
0.88 [0.46, 1.29] 0.42 [0.04, 0.89] 0.26 [0.21, 0.73]
1.52 [1.13, 1.90] 1.15 [0.72, 1.58] 0.85 [0.40, 1.29]
Unadjusted Adjusted for childhood assets Adjusted for assets and parental education
1.43 [1.08, 1.79] 1.30 [0.89, 1.70] 1.14 [0.74, 1.56]
1.31 [1.00, 1.63] 0.95 [0.60, 1.31] 0.90 [0.54, 1.27]
Number of participants (men/women).
maximum height was similar for men in each country (around 0.7 cm per decade); in women, the height increase seemed steepest in Novosibirsk (1.1 cm per decade) and shallowest in Krakow (0.4 cm per decade). Further adjustment for current socioeconomic position did not substantially change these estimates (not shown in table). Even after adjustment for childhood conditions, however, men and women in Novosibirsk remained shortest in each birth cohort (Figs. 1 and 2). Because the population of Novosibirsk expanded rapidly before and during the World War II (Soboleva, 2005) we examined whether there were differences in height trends by migration status. Table 6 shows the difference in the strength of the trend in height by year of birth between those in the Novosibirsk study population who were born in Novosibirsk Oblast (region) and those who migrated there after birth. The trends were stronger in males born in Novosibirsk Oblast, but there were no differences for women. Despite this pattern, however, additional adjustment for migration status did not alter the figures for the Novosibirsk study population shown in Table 5 by more than 1 mm in either direction.
4. Discussion We found evidence of secular trends in adult height amongst persons born between 1933 and 1957 in urban areas of in Russia, Poland and the Czech Republic. The positive secular trend in measured height, derived from cross-sectional data, can be partially explained by loss of height in ageing and by improvements in social conditions in childhood over the birth cohorts covered by the study. However, even after adjustment for these factors, the trends remained positive and statistically significant. The trends were approximately linear and did not generally show variations amongst those born in the years
Fig. 2. Secular trend in maximum height in women. Mean maximum height for 5 year age groups, OLS of height against year of birth, and regression adjusted for childhood SEC.
234
E.A. Webb et al. / Economics and Human Biology 6 (2008) 228–236
immediately preceding, during or following WWII, with a possible exception was Russian subjects born around WWII whose estimated maximum heights were slightly shorter than expected. There are several note-worthy findings in this paper. Firstly, there is a linear trend in assets in the home at age 10. Participants were aged 10 between 1943 and 1967 so this trend indicates a steady improvement in living conditions in the post-war period in Novosibirsk, Krakow and the towns in the Czech Republic in which the study was based. This improvement in social conditions was reflected in Russia by an increase in life expectancy at birth between 1946 and 1956 from 41.5 to 61.0 years in men and 51.0 to 68.9 years in women (Haynes and Husan, 2003). In Poland and the Czech Republic, life expectancy was also improving in the post-war years until the mid 1960s (Zatonski, 1996). Secondly, the secular trend towards increasing height persists after controlling for both age-related loss of height and childhood socioeconomic circumstances. In all three samples studied both men and women show an increase in height of approximately 1 cm per decade, as is suggested by Tanner (1962). After adjustment, the secular trend was strongest amongst Novosibirsk women. The magnitude of these secular trends is within the range of those observed in other European populations (Sobral, 1990; Kuh et al., 1991; Bielicki and Waliszko, 1991; van Leer et al., 1992; Jacome de Castro et al., 1998; Padez and Johnston, 1999; Niewenweg et al., 2003); however, they seem weaker than have been shown previously in eastern European populations. Bielicki and colleagues’ studies of Polish conscripts suggest a secular trend of 2.5 cm per decade between those born in 1946 and 1957 (Bielicki et al., 1992; Bielicki and Szklarska, 1999) (although this may be confounded by a trend to earlier maturation) and Dubrova and colleagues have shown a 1.8 cm per decade increase in height in Russian women born between 1930 and 1965 (Dubrova et al., 1995). Third, although childhood conditions were strongly associated with height across all countries and birth cohorts, and although they improved markedly over time, they failed to explain fully the secular trend in height. Perhaps this is not surprising, given the possible problems associated with measuring childhood conditions retrospectively. The generally linear pattern of the trends in height observed in those born between 1933 and 1957 suggests an absence of major differences in height (and growth) amongst adults born in the period immediately before, during or after WWII. We observed only minor reductions in the estimated maximum heights of people born in the war years (N = 6662), most notably in Novosibirsk, but these differences were small and of low statistical significance. It is interesting that the strongest effect of WWII on growth appears to be amongst the Novosibirsk study population, as this centre was far removed from the violence experienced across much of Europe. There are three potential explanations for the lack of observed effect of WWII on trends in height. Firstly, this study may not have adequate statistical power to identify small effects, as only about one in five participants were born during the war. Secondly, during the wartime years of hardship, there may have been a delay in the onset of decreasing growth rates at the end of adolescence. Vignerova and colleagues noted growth of Czech children slowed and stopped at older ages during the period of food scarcity during and immediately after World War II (Vignerova et al., 2006). This would provide the opportunity for catch-up growth in those whose rate of growth had been slowed. Thirdly, hardship may not affect only a specific birth cohort. This was observed by Brundtland and colleagues, who investigated heights of children in Oslo between 1920 and 1975, and showed that all age groups were similarly affected by WWII and that there was a temporary reversal of the secular trend during that period (Brundtland et al., 1980). It is not clear, unfortunately, whether in those Oslo children the effects were equally long lasting in all age groups, or whether those who experienced the insult earlier in life were able to compensate with catch-up growth. This, however, seems an unlikely explanation for the absence of an interruption in the secular trend in these Eastern European cohorts, as some participants were born more than 10 years after the end of WWII. Although hardships began before, and persisted after the end of the war in most places, it is unlikely that the growth of the youngest participants, who were born in the later 1950s, would have been affected as severely as that of those who lived through the war or were born during it. Several features of this study require consideration when interpreting the results. The major potential limitation of the study is its cross-sectional design and the fact that participant’s height was measured at an age when age-related loss of height had begun. Height loss will be differential by age, with those oldest participants having lost the most height. This leads to an over-estimation of the strength of the secular trend in height. To minimize this problem, adjustments were made for the probable loss of height with age (Sorkin et al., 1999b). The equations used, however, were based on height loss observed in western populations, and it is possible that the age-related shrinkage is different in these Eastern European populations. It is also possible that factors other than age and height, which were not accounted for in the calculation of maximum height, have an effect on height lost in older age. To our knowledge there is no evidence that this is the case, and data from the British Regional Heart Study (Wannamethee et al., 2006) and the Whitehall II study (M. Shipley, unpublished) suggest that agerelated height loss is not related to occupation, education or other measures of adult SEC. The validity of the calculations is supported by the similarity between the estimated maximum heights of the Krakow participants born in 1952–53 and the measured heights of a cohort of 18–19 year olds born in the same years reported in another study (Golab et al., 2002). All participants were living in urban areas. In industrialised countries, urban–rural differences in height are well recognised, and have been consistently observed in Poland (Bielicki et al., 1981; Bielicki, 1986; Bielicki and Waliszko, 1991). Participants cannot, therefore, be considered representative of the whole populations of their countries. Secular trends in height may be of a steeper gradient in rural than urban communities (Bielicki and Waliszko, 1991) and therefore we suggest that the estimates of secular trends reported here may underestimate the overall secular trends of the three countries. A further point for consideration is that the population of Novosibirsk expanded rapidly around the time of WWII, mainly through in migration from other areas of the Soviet Union. In fact, 45% of the Novosibirsk study population was born outside
E.A. Webb et al. / Economics and Human Biology 6 (2008) 228–236
235
the Novosibirsk Oblast (region). As the majority of inhabitants of Novosibirsk originally came from European Russia and are ethnic Russians, genetic differences are unlikely to be an important factor here. The trend to increasing height was stronger amongst men born in the Novosibirsk region, while there were no differences amongst women. However, adjustment for migration status did not suggest that it was an important confounding factor in the relationship between height and year of birth. Admittedly, this is necessarily only a crude assessment of the effect of migration, as the date of, and reasons for, migration are not known. In summary, there are observable secular trends towards increasing height amongst persons born between 1933 and 1957 and living in Novosibirsk, Krakow and six towns in the Czech Republic. These trends are not explained by parallel trends towards improving childhood SEC or by loss of height in ageing, and being born during WWII had only small negative effect, most notably in the Novosibirsk population. Acknowledgements The HAPIEE study was funded by the Wellcome Trust, the MacArthur Foundation initiative on Social Upheaval and Health, and the National Institute of Aging. EW is supported by the UK Medical Research Council studentship. References Berney, L.R., Blane, D.B., 1997. Collecting retrospective data: accuracy of recall after 50 years judged against historical records. Social Science & Medicine 45, 1519–1525. Bielicki, T., 1986. Physical growth as a measure of the economic well-being of populations: the twentieth century. In: Falkner, F., Tanner, J.M. (Eds.), Human Growth, vol. 3. Plenum Press, New York, pp. 283–305. Bielicki, T., Malina, R.M., Waliszko, A., 1992. Monitoring the dynamics of social stratification: statural variation among Polish conscripts in 1976 and 1986. American Journal of Human Biology 4, 345–352. Bielicki, T., Szczotka, H., Charzewski, J., 1981. The influence of three socio-economic factors on body height in Polish military conscripts. Human Biology 53, 543–555. Bielicki, T., Szklarska, A., 1999. The stratifying force of family size, urbanization and parental education in Socialist-era Poland. Journal of Biosocial Science 31, 525–536. Bielicki, T., Szklarska, A., Koziel, S., Ulijaszek, S.J., 2005. Changing patterns of social variation in stature in Poland: effects of transition from a command economy to the free-market system? Journal of Biosocial Science 37, 427–434. Bielicki, T., Waliszko, H., 1991. Urbanization-dependent gradients in stature among Polish conscripts in 1976 and 1986. American Journal of Human Biology 3, 419–424. Brainerd, E., 2008. Reassessing the standard of living in the Soviet Union: an analysis using archival and anthropometric data. Brundtland, G.H., Liestol, K., Walloe, L., 1980. Height, weight and menarcheal age of Oslo schoolchildren during the last 60 years. Annals of Human Biology 7, 307–322. Cline, M.G., Meredith, K.E., Boyer, J.T., Burrows, B., 1989. Decline of height with age in adults in a general-population sample—estimating maximum height and distinguishing birth cohort effects from actual loss of stature with aging. Human Biology 61, 415–425. Cole, T.J., 2003. The secular trend in human physical growth: a biological view. Economics & Human Biology 1, 161–168. Davies, K.M., Recker, R.R., Stegman, M.R., Heaney, R.P., 1991. Tallness versus shrinkage—do women shrink with age or grow taller with recent birth date. Journal of Bone and Mineral Research 6, 1115–1120. Dubrova, Y.E., Kurbatova, O.L., Kholod, O.N., Prokhorovskaya, V.D., 1995. Secular growth trend in two generations of the Russian population. Human Biology 67, 755–767. Floud, R., 1989. Measuring European inequality, the use of height data. In: Fox, A.J. (Ed.), Health Inequalities in European Countries. Gower, Aldershot. Golab, S., Chrzanowska, M., Zarow, J., Sobiecki, J., Brudecki, J., 2002. Zmiany sekularne w rozwoju fizycznym dzieci i mlodziezy krakowskiej w ostatnim 30leciu XX wieku. Wychowanie Fizyczne i Sport 46, 301–314. Haynes, M., Husan, R., 2003. A Century of State Murder? Death and Policy in Twentieth-Century Russia. Pluto Press, London. Jacome de Castro, J., Aleixo Dias, J., Baptista, F., Garcia e Costa, J., Galvao-Teles, A., Camilo-Alves, A., 1998. Secular trends of weight, height and obesity in cohorts of young Portuguese males in the District of Lisbon: 1960–1990. European Journal of Epidemiology 14, 299–303. Komlos, J., 1985. Stature and nutrition in the Habsburg Monarchy: the standard of living and economic development in the Eighteenth Century. American Historical Review 90, 1149–1161. Komlos, J., Kriwy, P., 2003. The biological standard of living in the two Germanies. German Economic Review 4, 459–473. Krawczynski, M., Walkowiak, J., Krzyzaniak, A., 2003. Secular changes in body height and weight in children and adolescents in Poznan, Poland between 1880 and 2000. Acta Paediatrica 92, 277–282. Kuh, D.L., Power, C., Rodgers, B., 1991. Secular trends in social-class and sex-differences in adult height. International Journal of Epidemiology 20, 1001– 1009. Kuh, D.L., Wadsworth, M., 1989. Parental height—childhood environment and subsequent adult height in a national birth cohort. International Journal of Epidemiology 18, 663–668. Mironov, B., 2007. Birth weight and physical stature in St. Petersburg: living standards of women in Russia, 1980–2005. Economics & Human Biology 5, 123– 143. Niewenweg, R., Smit, M.L., Walenkamp, M.J., Wit, J.M., 2003. Adult height corrected for shrinking and secular trend. Annals of Human Biology 30, 563–569. Padez, C., Johnston, F.E., 1999. Secular trends in male adult height 1904–1996 in relation to place of residence and parent’s educational level in Portugal. Annals of Human Biology 26, 287–298. Peasey, A., Bobak, M., Kubinova, R., Malyutina, S., Pajak, A., Tamosiunas, A., Pikhart, H., Nicholson, A., Marmot, M., 2006. Determinants of cardiovascular disease and other non-communicable diseases in Central and Eastern Europe: rationale and design of the HAPIEE study. BMC Public Health 6, 255–264. Pini, R., Tonon, E., Cavallini, M.C., Bencini, F., Di Bari, M., Masotti, G., Marchionni, N., 2001. Accuracy of equations for predicting stature from knee height, and assessment of statural loss in an older Italian population. Journals of Gerontology Series A-Biological Sciences and Medical Sciences 56, B3–B7. Power, C., Manor, O., Li, L., 2002. Are inequalities in height underestimated by adult social position? Effects of changing social structure and height selection in a cohort study. British Medical Journal 325, 131–134. Relethford, J.H., Lees, F.C., 1980. The effects of ageing and secular trend on stature in rural western Ireland. American Journal of Physical Anthropology 55, 81–88. Schmidt, I.M., Jorgensen, M.H., Michaelsen, K.F., 1995. Height of conscripts in Europe: is postneonatal mortality a predictor? Annals of Human Biology 22, 57–67. Soboleva, S., 2005. Economic migration to Western Siberia. Russian Regional Perspectives Journal 1, 45–47. Sobral, F., 1990. Secular changes in stature in Southern Portugal between 1930 and 1980 according to conscript data. Human Biology 62, 491–504.
236
E.A. Webb et al. / Economics and Human Biology 6 (2008) 228–236
Sorkin, J.D., Muller, D.C., Andres, R., 1999a. Longitudinal change in height of men and women: implications for interpretation of the body mass index—the Baltimore longitudinal study of aging. American Journal of Epidemiology 150, 969–977. Sorkin, J.D., Muller, D.C., Andres, R., 1999b. Longitudinal change in the heights of men and women: consequential effects on body mass index. Epidemiologic Reviews 21, 247–260. Sparen, P., Vagero, D., Shestov, D., Plavinskaja, S., Parfenova, N., Hoptiar, V., Paturot, D., Galanti, M.R., 2004. Long term mortality after severe starvation during the siege of Leningrad: prospective cohort study. British Medical Journal 328, 11. Tanner, J.M., 1962. Growth at Adolescence. Blackwell, Oxford. Ulijaszek, S.J., 1998. The secular trend. In: Ulijaszek, S.J., Johnston, F.E., Preece, M.A. (Eds.), The Cambridge Encyclopaedia of Human Growth and Development. Cambridge University Press, Cambridge, pp. 395–398. van Leer, E.M., van Noord, P.A., Seidell, J.C., 1992. Components of adult height and height loss. Secular trend and effects of ageing in women in the DOM project. Annals of Epidemiology 2, 611–615. Vignerova, J., Brabec, M., Blaha, P., 2006. Two centuries of growth among Czech children and youth. Economics & Human Biology 4, 237–252. Wannamethee, S.G., Shaper, A.G., Lennon, L., Whincup, P.H., 2006. Height loss in older men: associations with total mortality and incidence of cardiovascular disease. Archives of Internal Medicine 166, 2546–2552. Webb, E., Kuh, D., Peasey, A., Pajak, A., Malyutina, S., Kubinova, R., Topor-Madry, R., Denisova, D., Capkova, N., Marmot, M., Bobak, M., 2008. Childhood socioeconomic circumstances and adult height and leg length in central and eastern Europe. Journal of Epidemiology and Community Health 62, 351– 357. Zatonski, W., 1996. Evolution of Health in Poland Since 1988. The Institute of Oncology, Warsaw, Poland.