Background: Physical activity plays an important role in prevention of chronic diseases. Animal studies have suggested that lifestyle and exercise habits may have a prenatal origin. Our aim was to assess the role of early growth on leisure time physical activity (LTPA) in later life among 57–70-years-old men and women. Methods: We examined 2003 individuals born in Helsinki, Finland between 1934 and 1944. Of them, 1967 individuals with adequate information on their LTPA in adult life were included in this study. LTPA was assessed by a validated exercise questionnaire (KIHD Study 12 month physical activity history). Subjects’ birth and serial growth measurements were obtained from birth, child welfare and school health records. Results: Participants with higher engagement in LTPA showed a more favourable adult anthropometric and body composition profile than those who were less active. LTPA was positively associated with adult social class. Higher weight and length at birth, and weight at 2 years after adult BMI adjustment, predicted higher intensity of total LTPA (P = 0.04, P = 0.01 and P = 0.03), respectively. Higher height at 2, 7 and 11 years predicted higher intensity of conditioning LTPA (P = 0.01, P = 0.04 and P = 0.004). Higher weight and height at 2, 7 and 11 years predicted higher energy expenditure (EE) of total LTPA (P-values being from 0.01 to 0.03). Furthermore, higher height at 2 and 11 years predicted higher EE of conditioning LTPA (P = 0.02 and P = 0.03). Conclusion: People who as children were taller and weighed more engage more in leisure time physical activity in late adulthood.
It is well accepted that many common non-communicable diseases develop as a consequence of prevailing adverse environmental conditions in combination with an individual’s genetic make-up. However, increasing evidence suggests that many chronic diseases, e.g. coronary heart disease, hypertension and type 2 diabetes, originate early in life through a phenomenon called programming.1–5 These outcomes may, however, be strongly modified by lifestyle.6,7 In a recent study, Andersen et al.8 found that the association between birth weight and physical activity (PA) in later life is weak in the normal birth weight range, but both low and high birth weights are associated with a lower probability of undertaking LTPA.
In general, PA plays an important role in prevention of chronic diseases, such as type 2 diabetes, hypertension, coronary heart disease and osteoporosis.9–11 Therefore it is of interest to study to what extent PA, in itself, is programmed early in life. Animal models have suggested that life style and exercise habits may have a prenatal origin; severe malnourishment of a pregnant rat leads to smaller body size at birth and inactivity among the offspring. Sedentary behaviour is further exacerbated by post-natal hypercaloric nutrition.12 However, human studies examining associations between early growth and exercise habits are scarce. While it seems clear that extreme groups such as young adults born severely preterm exercise less than those born at term,13–15 studies assessing the association of birth size with PA in childhood,16–18 as well as in adult life6–8 have been inconsistent.
It is obvious that more studies focusing upon prenatal factors and childhood growth in relation to PA are needed before any general conclusions can be drawn. Our aim was to assess the role of birth size and childhood growth on LTPA among participants in the Helsinki Birth Cohort Study (HBCS) (1934–44).
Subjects and study design
This study is a sub-study of the HBCS. The original cohort includes 8760 men and women born in Helsinki 1934–44, who attended child welfare clinics and who were resident in Finland in 1971 when every resident of Finland was assigned a unique identification number. The majority of the children went to school in the city (77%). Their birth records include measurements of weight and length at birth and information on gestational age. Child welfare and school health records include serial measurements of height and weight. Birth, child welfare and school health records have been described in detail previously.19
In order to achieve a sample size over 2000 people for the clinical part of the study, 2902 individuals of the original HBC were randomly selected to detailed clinical examinations. Of them 2003 were willing to participate, and 1967 (907 men and 1060 women) with adequate information on their leisure time physical activity (LTPA) were included in this study. The clinical study protocol was approved by the Ethics Committee of Epidemiology and Public Health at the Hospital District of Helsinki and Uusimaa and Ethics Committee of National Public Health Institute, Helsinki. Written informed consent was obtained from each participant before any study procedure.
Clinical study measurements were performed by a team of trained research nurses in the years 2001–04. Height was measured without shoes on to the nearest 0.1 cm and weight was measured in light indoor clothing to the nearest 0.1 kg. Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters. Waist circumference was measured midway between the lowest rib and the iliac crest. Lean and fat mass were measured by bioelectrical impedance analysis (BIA) using the InBody 3.0 eight-polar tactile electrode system, Biospace Co., Ltd, Seoul, Korea.20
At the clinical examination, the 57–70-year old subjects completed a validated exercise questionnaire (KIHD 12-month LTPA history).21,22 This questionnaire was used to assess duration, frequency and intensity of the most common forms of (LTPA) including indoor and outdoor activities performed during the previous 12 months. For each intensity grade, activity-specific metabolic equivalent (MET) values were used. MET is defined as the ratio of the metabolic rate during activity to the rate at rest [1 MET corresponds to an energy expenditure (EE) of 1 kcal × kg−1 × h−1]. Total exercise includes various types of leisure time activities, from house cleaning and gardening (non-conditioning) to e.g. jogging and resistance training (conditioning), during leisure time. Conditioning exercise is defined as more dedicated and vigorous PA. The rates of EE during PA vary depending on intensity, duration and frequency of the activity and on the body mass of the person, and hence EE of total and conditional LTPA was calculated accordingly.
Using father’s occupation, which was available from the birth, child welfare clinic and school health records, we grouped the fathers into three groups (labourers/manual workers, lower middle and upper middle class) originally based upon a nine group social classification system by Statistics Finland. Childhood social class was based on father’s highest social class. Social class in adulthood, based on subject’s own occupation, was derived from Census data in 1980. Educational attainment and smoking habits were obtained from a self administered questionnaire. Educational attainment was categorized into three groups by approximate number of years studied: basic (≤9 years of education); secondary (10–12 years) and higher (≥13 years of education). The subjects were defined as smokers if they smoked one or more cigarette per day.
A multiple linear regression analysis was conducted to explore the relations of body size at birth, 2, 7 and 11 years, body composition, socio-economic and life style factors with frequency, duration, intensity and EE levels of total and conditioning LTPA. Variables with skewed distribution were log-transformed.
We converted each growth measurement for each individual to a z-score (SD-score). The z-score is the number of standard deviations by which an observation differs from the mean for the whole study group. Since the children were not measured exactly on their birthdays we obtained a z-score by interpolation. To measure how much size at any age differed from that predicted by the body size attained at an earlier age, we used the residuals from linear regression analysis, which we refer to as conditional growth. For consistency with our previous studies, data at birth and ages 2, 7 and 11 years were chosen for further analysis.
Although we present the adult body composition and LTPA measures separately for men and women in the table 1, we did not observe any interactions between sex and growth variables and therefore pooled analyses are presented in the tables 2, 3 and in the figure 1.
(A) Age- and sex-adjusted correlations between intensity of conditioning LTPA and height, weight and BMI annually from birth to age 11 years. The horizontal dashed line indicates a correlation of 0. (B) Age- and sex-adjusted correlations between EE of conditioning LTPA and height, weight and BMI annually from birth to age 11 years. The horizontal dashed line indicates a correlation of 0
b: Regression coefficients (with 95% CI) express percentage changes in adult EE associated with a 1 kg increase in weights and 1 cm in heights at birth, 2, 7 and 11 years.
c: Model is adjusted for adult age, sex and BMI.
*P < 0.05; **P ≤ 0.01.
CI, confidence interval. Values in italics indicate results of regression analyses. Bolded values indicate significant associations.
All analyses were adjusted for age, and in pooled analysis also for sex. Independent effects of childhood and adult social class were studied in relation to LTPA. In further analyses examining the relations of childhood body size and early growth with LTPA, childhood and adult social class, BMI, height, lean body mass, education and smoking were adjusted for, respectively. Analyses were performed using SPSS (Statistical Package for Social Sciences) version 17.
Characteristics of the study subjects are shown in table 1. Of the 2003 men and women, 1967 (98%) reported their total LTPA including both conditioning and non-conditioning LTPA during the past year, and are included in the present study. Of them, 1815 (92%) reported any conditioning LTPA. Women exercised more frequently and spent more time on LTPA than men (P < 0.001 for both), while men exercised with higher intensity and had higher EE of their LTPA (P = 0.001, P = 0.02). Participants with higher engagement to LTPA showed a more favourable adult anthropometric and body composition profile than those who were less active.
LTPA and social class
LTPA was examined in relation to childhood and adult social class. Childhood social class was not related to LTPA. However, adult social class was positively and independently associated with intensity of conditioning exercise (P < 0.001), particularly among men. Mean intensity was 5.65 METs in men belonging to the highest social class, 5.23 in the second, 4.94 in the third and 4.86 METs in the lowest group. Associations between total intensity and adult social class were similar and statistically significant among men (P = 0.004). In women, the results paralleled to those of the men, but did not reach statistical significance (P = 0.4 and P = 0.2). Further adjustments for childhood growth measurements or other confounders did not attenuate the results.
A similar trend was observed for EE of conditioning LTPA (P = 0.02), but for total EE of LTPA it was not statistically significant (P = 0.52). Childhood social class was not associated with EE of LTPA. In men, childhood social class and educational attainment were positively associated (P ≤ 0.01 for both) with frequency of total LTPA. Smoking was negatively associated with all of the components of LTPA in both genders, except for duration of total LTPA.
Association between birth and childhood body size and frequency, duration and intensity of LTPA
Table 2 shows that birth size was positively associated with intensity of total LTPA. The results remained the same when adult social class and BMI were taken into account. After adjustment for adult lean body mass, these associations became non-significant. There were no associations between childhood weight or height and frequency or duration of LTPA. Height during infancy and childhood was not associated with intensity of total LTPA. However, height at 2, 7 and 11 years was positively associated with intensity of conditioning LTPA. Adjustments for lean body mass did not affect the results, but adjustments for adult height did attenuate the associations. Adjustments for social class and educational attainment influenced the results only little. Adjustment for smoking did not affect the associations. Figure 1A shows age and sex adjusted correlation between childhood body size from birth to 11 years and intensity of conditioning LTPA.
Childhood body size and EE of LTPA
After childhood growth measurements were introduced adult social class was no longer associated with EE of conditioning LTPA. Weight and height at 2, 7 and 11 years of age were positively associated with EE of total LTPA. Height at 2 and 11 years were also associated with EE of conditioning LTPA (table 3 and figure 1B). Adjustments for adult BMI slightly attenuated the associations between weight and EE of total LTPA, but did not affect associations between height and EE of LTPA. Adjustments for adult lean body mass and height made all the associations non-significant. Further adjustments for childhood and adult social class, educational attainment and smoking habits had no significant impact on the associations between body size and EE of LTPA.
Childhood conditional growth and LTPA in later life
We examined the effects of change in weight, height and BMI during the three periods of growth (0–2, 2–7, 7–11 years) on intensity and EE of total and conditioning LTPA. An increase in height during infancy (between birth and 2 years), and from 7 to 11 years predicted higher intensity of conditioning LTPA (P = 0.01 and P = 0.004), respectively. Adult social class remained strongly associated with intensity of conditioning LTPA (P < 0.001). Gain in height during infancy was associated with higher EE of both total and conditioning LTPA (P < 0.01 and P = 0.01), and gain in height from 7 to 11 years was associated with higher EE of conditioning LTPA (P = 0.03).
We examined whether early growth predicts adult behaviour in terms of LTPA. We found that children, who were heavier and taller at birth, and heavier during infancy, reported higher intensity levels of total LTPA in adult life. Those who were taller at 2, 7 and 11 years reported as adults higher intensity levels of conditioning LTPA. Children, who were heavier and taller from 2 to 11 years, reported higher EE of total LTPA. Those with higher EE of conditioning LTPA were taller at 2 and 11 years of age. Increases in height during infancy and from 7 to 11 years were associated with higher intensity and EE of conditioning LTPA, and increases in height and weight during infancy were associated with EE of total LTPA. Those, who belonged to the highest social class as adults where physically more active than those, who belonged to the lower social classes.
Physically inactive individuals are well known to have an increased risk for several common non-communicable diseases including type 2 diabetes, coronary heart disease and osteoporosis. This risk can be reduced by increasing PA.9–11 More recently a small body size at birth has been identified as a risk factor for several of these diseases. Also childhood growth modifies the disease risk.2,3 Not only physiological outcomes but also behavioural ones have been suggested to be programmed early in life.23 However, only few studies have been focusing upon the relationship between early growth and measures of exercise in adult life. Exercise habits are modifiable and furthermore they could mediate part of the effect of birth size on later health. Even though the associations observed between childhood height and weight and LTPA are relatively weak compared to the previously reported associations with coronary heart disease2 and stroke,24 they seem to be similar to or stronger than the associations with other risk factors such as serum lipids.25
The results of those few studies which have studied the associations between birth size and PA in adult life have been inconsistent. Laaksonen et al.6 did not find any associations between birth size and cardio-respiratory fitness or duration of strenuous LTPA in adult life. Davies and co-workers have reported that a higher birth weight was positively associated with regular PA,26 while others did not find significant relationships between early growth and PA in later life.27 We have previously reported in an older cohort a strong and inverse association between birth size and exercise frequency in men.7 The recent findings from the extensive NordNet Study showed that the associations found between birth weight and LTPA are weak within the normal birth weight range. However, both low and high birth weights were associated with lower LTPA.8 The measurement of PA is complex and this may certainly explain some of the contradictory findings. Furthermore PA consists of several components.
The results of a Brazilian study consisting of adolescents aged 10–12 years showed that genetic factors or habit formation during early childhood seem to be of greater importance in determining PA than physiological factors as consequences of programming. Interestingly the majority of the children with sedentary lifestyle belonged to the lowest birth weight group, but the association between birth weight and life style was not statistically significant.16 In the Avon longitudinal study of parents and children (ALSPAC), birth size did not associate with PA in 11–12-years old, however, parental activity levels did predict activity level of their child.17 Inconsistently, results from the Northern Ireland Young Hearts Project have shown positive associations between birth weight and aerobic fitness in 12 year-old boys and girls.18
In our previous studies, birth size as well as size during infancy have been shown to be associated with adult body composition.19 The association was found to be strong and positive with adult lean body mass. Since lean body mass consists mainly of muscle, birth weight was associated with grip strength as well, which might well be a marker of functional capacity in elderly individuals.28
Several mechanisms may explain the positive association between childhood body size and LTPA in later life. Intensity of LTPA, especially intensity of conditioning LTPA which reflects the more dedicated and vigorous form of PA can be due to higher muscle growth during pre-pubertal age and tracking of muscle size. Another explanation could be the quality of muscle mass, including the number and composition of muscle fibres. There is evidence that number of muscle fibres are not fixed by the time of birth as previously believed. The role of satellite cells in post-natal growth and regeneration is increasingly recognized.29 Moreover, Jensen and co-workers have suggested that adverse fetal environment may induce primary changes in muscle fiber composition.30–32
Study strengths and limitations
In addition to the present study, there are only a few studies examining strictly the relationship between early growth and LTPA. Compared to our study, no other study in this field of research has included such detailed childhood growth data with follow-up period of similar duration. Both men and women were included in the study. Birth and childhood growth data were obtained from reliable records and not based on recalled values. Body composition was measured with a validated bioimpedance method suitable for analyses on epidemiological studies.19,28 LTPA was assessed with a validated questionnaire.21,22
Limitations of the data have been discussed elsewhere.2 Our study was restricted to individuals, who were born in the Helsinki University Central Hospital, attended child welfare clinics, which were free and attendance was voluntary, were alive and living in Finland in the year 2003, and were willing to participate in the clinical examination. Therefore, the people in our study may not be representative of the population now living in Helsinki. However, at birth the distribution of social class, which is based on fathers’ occupations, was similar to that in the city as whole, where at that time ∼60% of men were labourers. The mechanisms suggested here to explain our results are speculative, since we do not have body composition measurements during childhood in detail, and the adult LTPA measurements are all cross-sectional, which precludes conclusions on causality. In addition to the childhood weight gains, it would be valuable to evaluate other factors such as PA and dietary patterns during childhood, and their effects on adult body composition and PA. Unfortunately such data are not available for our cohort.
In conclusion, individuals with higher engagement to LTPA, showed a more favourable anthropometric and body composition profile than those who were less active. Even more importantly, we have shown that body size and growth from birth to 11 years, whether through foetal programming or post-natal events, or through combination of these two, predicts LTPA patterns in later life. The children who were heavier at birth and in infancy exercised with higher intensity, and the children who were heavier from 2 to 11 years, had higher EE of total LTPA. However, these children were not overweight or obese according to today’s standards. Their higher weight and BMI is more an expression of the higher muscle mass. Taller children exercised more vigorously; taller stature predicted higher intensity levels of conditioning LTPA, and also higher EE levels of total LTPA. Furthermore, EE of conditioning LTPA was higher in taller children. Adult social class was independently associated with intensity of LTPA. Our results are consistent with the suggestion that both prenatal and factors affecting childhood growth may play an important role in determining levels of LTPA in later life. More studies are required before more general conclusion can be drawn.
The British Heart Foundation; The Academy of Finland; the Päivikki and Sakari Sohlberg Foundation; Finska Läkaresällskapet; the Finnish Cardiovascular Research Foundation; the Juho Vainio Foundation; the Jalmari and Rauha Ahokas Foundation; the Yrjö Jahnsson Foundation; Samfundet Folkhälsan; the Signe and Arne Gyllenberg Foundation.
Conflicts of interest: None declared.
Body size and growth in height from birth to 11 years predicts LTPA patterns in later life.
Since the risk of CVD-related diseases can be reduced already at the moderate level of LTPA, it is important to examine, whether PA, in itself, has foetal or childhood origins.
Socio-economic status seems to be an important determinant of exercise habits.
Long term future studies are needed to confirm our suggestion that LTPA may be part of the underlying mechanism of the association between early growth and later disease.
. Aerobic capacity, strength, flexibility, and activity levels in unimpaired extremely low birth weight ( < or = 800g) survivors at 17 years of age compared with term-born control subjects. Pediatrics 2005;116:e58-65.
Minna K.Salonen, EeroKajantie, CliveOsmond, TomForsén, HilkkaYlihärsilä, MariaPaile-Hyvärinen, David J.P.Barker, Johan G.ErikssonEur J Public Health(2011)21 (6):
719-724DOI: http://dx.doi.org/10.1093/eurpub/ckq176First published online: 1 December 2010 (6 pages)