Recent data indicate a marked increase in the prevalence of obesity among school-aged children. Thus, efficacious programmes that prevent overweight development in children are urgently needed.
To evaluate the impact of repeatedly given, individualised dietary and lifestyle counselling on the prevalence of overweight during the first 10 years of life.
Design and participants:
This study was a part of the Special Turku Coronary Risk Factor Intervention Project for Children (STRIP), which is a prospective, randomised trial aimed at reducing the exposure of the intervention children to the known risk factors of atherosclerosis. At the child's age of 7 months, 1062 children were assigned to an intervention group (n=540) or to a control group (n=522). The intervention children received individualised counselling focused on healthy diet and physical activity biannually. Height and weight of the children were measured at least once a year.
Main outcome measure:
Prevalence of overweight and obesity among the intervention and control children by sex and age. Children were classified as overweight or obese if their weight for height was >20% or 40% above the mean weight for height of healthy Finnish children, respectively.
After the age of 2 years, there were continuously fewer overweight girls in the intervention group than in the control group. At the age of 10 years, 10.2% of the intervention girls and 18.8% of the control girls were overweight (P=0.0439), whereas 11.6% of the intervention boys and 12.1% of the control boys were overweight (P≈1.00). Only three children in the intervention group were obese at some age point, whereas 14 control children were classified as obese at some age point.
Individualised dietary and lifestyle counselling given twice a year since infancy decreases prevalence of overweight in school-aged girls even without any primary energy restrictions.
Childhood obesity is associated with serious short-term and long-term effects on health.1 The short-term consequences include psychological problems, like low self-esteem,2 and unwanted changes in cardiovascular risk factors.3 The most important long-term consequence of paediatric obesity is persistence of obesity and obesity-related changes of cardiovascular disease risk factors from childhood to adulthood.4, 5 Thus, increasing prevalence of obesity among children and adolescents is a major health threat. The number of overweight 6- to 19-year-old children has more than tripled in the US over the past three decades,6 and more than 20% of the 9- to 10-year-old girls participating in the National Heart, Lung, and Blood Institute Growth and Health Study were overweight.7 Comprehensive programmes with persisting efficacy are, therefore, needed for obesity prevention in childhood.
Previously used obesity preventing programmes have shown marginal efficacy at best during childhood. Nutrition and physical activity intervention targeted at children exposed to multiple cardiovascular disease risk factors produced no significant changes in body mass index (BMI) or body fat.8 Another school-based 3-year nutrition and physical activity programme aiming at primary prevention of obesity had no effect on percentage of body fat in American Indian children, although children's nutrition knowledge, attitudes and behaviours were positively changed by the intervention.9 Similar results have been achieved in other primary prevention studies. Several of these interventions have, to a variable degree, reduced dietary fat intake, increased physical activity during school time, increased consumption of fresh fruit and vegetables, and reduced television viewing, but have been unable to reduce prevalence of excessive body weight or adiposity.10, 11, 12, 13
Childhood obesity associates tightly with many risk factors of coronary heart disease.1 More than half of obese 5- to 10-year-old children are exposed to at least one and every fourth to two or more such risk factors.3 Overweight 9- to 10-year-old girls have higher serum triglyceride and low-density lipoprotein cholesterol (LDL-C) concentrations, lower serum high-density lipoprotein cholesterol (HDL-C) concentrations, higher LDL-C/HDL-C ratios, and higher systolic and diastolic blood pressure than girls of the same ethnic background with normal weight.7 The Special Turku Coronary Risk Factor Intervention Project for Children (STRIP) is a long-term randomised intervention trial starting at infancy and aiming at reduced exposure of the intervention children to the known risk factors of cardiovascular disease. The main focus of the intervention has been on reducing saturated fat intake without special emphasis on total energy intake. The aim of the present study was to examine whether the counselling given to the study children and their families would also decrease development of overweight and obesity.
Study design and subjects
Families were recruited into the prospective, randomised STRIP trial by nurses at the well-baby clinics in the city of Turku, Finland, at the infants' routine 5-month visit between February 1990 and June 1992 as described.14 At 7 months of age, the infants were randomly assigned either to an intervention group (n=540) or to a control group (n=522). The families of the intervention group received individualised dietary and lifestyle counselling at 1–3-month intervals until the child was 2 years old and twice a year thereafter. The families of the control group were seen by the counselling team twice a year until the child's age was 7 years and once a year after that age, but they only received similar basic health education as routinely given at Finnish well-baby clinics and school health care. During the trial, all children continued their regular visits at the communal well-baby clinics and school health care for vaccinations, growth and development follow-up, and basic health education. The few study children who had a chronic disease (e.g. diabetes) that might have an effect on body weight development were excluded from this analysis.
At the age of 10 years, 585 children still remained in the trial. The dropout rates were closely similar in the intervention and control groups through the years. At the infants' age of 7 months, 1056 mothers and 983 fathers participated in the baseline visit. At the children's age of 10 years, 484 mothers and 331 fathers participated in the respective trial visit.
The Joint Commission on Ethics of the Turku University and the Turku University Central Hospital approved the STRIP study. Informed consent was obtained from the parents of the children at the beginning of the trial.
The counselling of the intervention families was given by a nutritionist and a physician. It was based on a constructivist theory of learning: the child and the family had an active role in counselling and their previous knowledge and experiences influenced the counselling session. Child's food consumption was recorded in detail in food diaries for 3–4 days before each visit. The nutrient compositions of diets were analysed using the Micro-Nutrica PC Programme, which was updated continuously (Research and Development Unit of the Social Insurance Institution, Turku, Finland). At each visit, the child's recent food record and physical activity habits were used as a basis for suggestions given. A predetermined diet or physical activity pattern were never ordered. Instead, the families were encouraged to gradually change the child's eating and physical activity habits towards healthier lifestyle. During the first years, the counselling was mainly focused on the child's diet and was given to the parents. However, from the age 7.5 years onwards, progressively more information and suggestions were given directly to the child, as the child was also met alone without parents. The counselling was based on the age and cognitive ability of the child.
At the beginning of the trial, the nutrition counselling mainly aimed at reduction of the child's intake of saturated fat. The optimal diet was defined to contain energy without any restrictions, protein 10–15% of energy (E%), carbohydrates 50–60 E%, and fat 30 E% after the age of 2 years (30–35 E% between 1 and 2 years) with unsaturated to saturated fatty acid ratio (U:S) of 2:1. Breastfeeding was encouraged as long as the mother felt it feasible. If the child was not breast-fed, commercial cow milk-based formulas were recommended until the age of 12 months. Thereafter, 0.5–0.6 L skim milk daily was recommended. To maintain adequate energy intake, the parents were counselled to add 2 or 3 teaspoonfuls (∼10 g) of soft margarine or vegetable oil daily to the child's food between 12 to 24 months of age. Type of milk was the major subject matter of the counselling during the first months of the trial, but suggestions were also made to change from products containing large amounts of saturated fat to products with either less saturated fat or to products with more unsaturated fat instead of saturated fat. During the course of the trial, also ample use of vegetables, fruits and berries as well as whole-grain products was encouraged.
A physician gave the counselling regarding the physical activity pattern of the intervention children based on the child's previous habits. Hobbies of the child were recorded at each visit and suggestions were made to increase the amount of everyday physical activity, like walking or cycling to the school if possible. No special physical activity programmes were offered for the children of the intervention group.
The children in the control group received similar basic health education as routinely given at Finnish well-baby clinics and school health care. Breast feeding was encouraged as to the children in the intervention group. At the age of 12 months, cow's milk with 1.9% (1.5% after May 1995) fat was recommended for daily use. In the control group, no suggestions on use of fats were made and dietary issues were discussed only superficially. Hobbies of the children were recorded but no suggestions were made as to engaging them into any physical activity. Obese children (weight for height 40% above the Finnish mean) of the control group were an exception since counselling regarding weight management was given to them as a part of basic health education.
Weights of the children were measured to the nearest 0.1 kg with an electronic scale (S10; Soehnle, Murrhardt, Germany) at each visit. Recumbent lengths of the children younger than 2 years were recorded and thereafter standing heights were measured to the nearest millimetre with a wall-mounted Harpenden stadiometer (Holtain, Crymych, UK). Absolute height and height s.d. on the Finnish growth charts as well as absolute weight and weight for height were recorded. Weight for height was expressed as percentual deviation from the mean values of healthy Finnish children of the same sex.15 As reference values for BMI have not been determined for Finnish children, the above method is customarily used to estimate body weight status. Children were classified as overweight if their weight for height was >20% above the mean and as obese if their weight for height was 40% above the mean. This primary classification of overweight and obesity was also applied in multivariate models. On the other hand, children were classified as slim if their weight for height was more than 15% below the mean value. To facilitate international comparisons, a secondary classification of overweight based on BMI was used. BMI was calculated for the 10-year-old children as the weight (kg) divided by the square of the height (m). Children were defined to be overweight if their BMI exceeded the cutoff points provided by Cole et al. (19.86 and 19.84 for girls and boys, respectively).16 The BMIs of our study children were also compared to Centers for Disease Control and Prevention Growth Charts for the United States and the children were defined to be overweight if the BMI was at or above 85th percentile.17 Birth weights of the study children were collected from the well-baby clinic records. Weights of the parents were measured annually and the BMI was calculated as described above.
Pubertal status of the study children was recorded from the age of 9 years onwards. The physicians examining the study children were carefully trained in pubertal staging by an experienced paediatric endocrinologist. Testicular length was measured with a ruler and breast tissue diameter and pubic hair development were estimated visually. The signs of puberty were recorded according to Tanner staging:18 stages M1/P1 in girls and G1/P1 in boys were considered prepubertal and all other stages pubertal. At the age of 9 years, the pubertal status was recorded from all study children but at the age of 10 years, one girl and 14 boys refused proper examination. These children were all, except for one boy, of normal weight.
The SAS for Windows 9.1 (SAS Institute, Cary, NC, USA) statistical software was used to carry out statistical analyses and the genders were analysed separately. Results are shown as percentages, odds ratios with 95% confidence intervals (95% CI), or means with standard deviations. The Fisher's Exact Tests were used to compare the proportions of overweight or slim children in the intervention and control groups at the age of 10 years. Possible factors predicting the development of overweight were assessed using logistic regression analysis with the generalised estimating equations (GEE) methodology.19 For this analysis, the children were divided into two groups: to those, whose weight for height exceeded the Finnish mean by more than 20% at any age point and to those whose weight for height was constantly +20% or less. The possible predictors used in the model included study group, birth weight, age, mother's BMI, father's BMI and pubertal status at the age of 10 years. Repeated measures ANOVA was used to study the main effects and the interactions between the study group and time in height s.d. and weight for height. Two-sample t-tests were run at each age point to compare the weights for height of the intervention and control girls since an interaction was found between the study group and time factors. P<0.05 was considered statistically significant.
There were continuously fewer overweight girls (weight for height more than 20% above the Finnish mean) in the intervention group of the STRIP trial than in the control group after the age of 2 years (Figure 1). At the age of 10 years, 10.2% of the girls in the intervention group and 18.8% of the girls in the control group were overweight (P=0.0439). However, the difference between the boys in the intervention group and in the control group was not significant. At the age of 10 years, 11.6% of the boys in the intervention group and 12.1% of the boys in the control group were overweight (P≈1.00). Only two girls and one boy in the intervention group were classified as obese (the weight for height 40% or more above the Finnish mean) at some age point, while eight girls and six boys in the control group were classified as obese at least once during the trial (Figure 2). The weight for height of four of these girls and five of these boys increased continuously throughout the trial.
To facilitate international comparisons, the prevalence of overweight was also assessed among the 10-year-old children in the control group using international cutoff points and American growth charts (Table 1). According to the inter-national cutoff points, 23.6 and 17.8% of the 10-year-old girls and boys, respectively, were overweight. When the BMIs of our study children were compared to the American growth charts, 22.9 and 19.1% of the 10-year-old girls and boys, respectively, were classified as overweight (at or above the 85th percentile).
According to the model used to study possible predictors of overweight, each year in child's age increased the risk of becoming overweight by 39 and 41% for girls and boys, respectively (P<0.0001, 95% CI 25 to 55% and P=0.0005, 95% CI 24 to 61% for girls and boys, respectively). One unit increase in mother's BMI increased the risk of becoming overweight by 16% in girls (P=0.0248, 95% CI 6 to 26%) but mother's BMI was not a significant predictor of overweight in boys. One unit increase in father's BMI increased the risk by 10 and 14% in girls and boys, respectively (P=0.0175, 95% CI 5 to 14% and P=0.0399, 95% CI 1 to 30%, for girls and boys, respectively). The study group, birth weight, and pubertal status were not statistically significant predictors of overweight in this model.
Growth of the children was closely similar in the intervention and control groups during the first 10 years of life (Figure 3). The height s.d.s of the girls and the boys in both study groups were constantly slightly above the mean values on the Finnish growth charts15 (for comparison between the intervention group and the control group, P=0.17 and P=0.47 for girls and boys, respectively). The weights for height of the boys in the intervention group and in the control group were closely similar (P=0.30). The weights for height of girls were above the mean Finnish values with only one exception (3-year-old control girls) indicating that the children in the intervention group as well as in the control group were to some extent heavier than the average Finnish children. The girls in the intervention group and control group behaved differently over time as regards the weight for height (P=0.0108). Until the age of 5 years, the girls in the intervention group were somewhat heavier than those in the control group, but thereafter the reverse was true. However, the difference between the groups was not statistically significant at any single age point.
The proportion of slim girls (weight for height more than 15% below the mean) varied from 0 to 2.9% in the intervention group and from 0 to 4.1% in the control group (Figure 1). At the age of 10 years, 2.9% of the girls in the intervention group and 3.5% of the girls in the control group were slim (P≈1.00). The proportion of slim boys ranged from 0.4 to 4.1% and from 0 to 8.3% in the intervention and control groups, respectively. At the age of 10 years, 4.1% of the boys in the intervention group and 8.3% of the boys in the control group were slim (P=0.16).
By the age of 9 years, 12.4% of the girls in the intervention group and 11.6% of the girls in the control group had entered puberty. No pubertal changes were observed in any of the boys of that age. By 10 years of age, one-third of the girls had entered puberty, whereas 12.2% of the boys in the intervention group and 14.4% of the boys in the control group showed signs of puberty. There were no differences between the intervention group and the control group in the proportions of children who had entered puberty by the ages of 9 and 10 years. At the age of 9 years, 21.6 and 10.4% of the overweight girls and of the normal-weight girls had entered puberty, respectively (P=0.057) (Figure 4). At the age of 10 years, 63.4 and 32.1% of the overweight girls and of the normal-weight girls had entered puberty, respectively (P<0.001). No difference was seen in the start of puberty in normal-weight and overweight 10-year-old boys as 13.9 and 13.4% of the overweight boys and of the normal-weight boys showed signs of puberty, respectively (P≈1.00).
The effect of dropouts to these results was assessed at all studied age points by comparing the proportion of those overweight children who dropped out by the next follow-up visit with the proportion of those normal-weight children who dropped out by the next follow-up visit. No difference was found in the proportions of dropouts among overweight and normal-weight children at any age point (data not shown). For example, 6.5% of the overweight 9-year-old children dropped out by the 10-year follow-up visit compared to 7.1% of the normal-weight 9-year-old children (P≈1.00).
The main finding of our study is that repeated and individualised nutrition and lifestyle counselling starting already in infancy may prevent the development of overweight and obesity in prepubertal children. The effect of the counselling was clear especially in girls since the proportion of overweight girls increased much more rapidly in the control group than in the intervention group. There were more obese girls and obese boys in the control group than in the intervention group even though counselling regarding weight management had been given to all those children who had become obese regardless whether they belonged to the intervention group or the control group.
Half of our study children belonged to the intervention arm of the STRIP trial and the families of these children had received repeatedly nutritional and lifestyle counselling since the children were 8 months old. The aim of the counselling was to reduce the exposure of these children to the known environmental risk factors of coronary heart disease. A physician and a nutritionist met the children of the intervention group twice a year. The counselling given was individualised and based on the child's recent food record and previous physical activity pattern. The families were encouraged to gradually change their child's eating and physical activity patterns towards healthier lifestyle. No predetermined programmes were used. As considerable concern has been raised over the increasing prevalence of childhood obesity in the Western countries, there is a clear demand for clinical and public health actions that are widely usable. Population-based preventive strategies seem to be more justified than strategies targeted at high-risk children also because factors that determine the susceptibility of an individual to obesity are poorly understood. This study shows that long-term nutrition and lifestyle counselling reduces the risk of overweight and obesity of girls even if the intervention is not very intense and the primary target of the counselling is reduction of exposure of the children to the risk factors of coronary heart disease in general and not the prevention of obesity. Our findings are, thus, in contradiction to most previous primary prevention studies, which were able to reduce dietary fat intake, increase physical activity during school time, increase consumption of fresh fruit and vegetables, and reduce television viewing but were not able to reduce the body weight or adiposity.10, 11, 12, 13, 20 Similarly to our findings, the intervention given in the Planet Health Study was able to reduce the prevalence of overweight in girls.21 It remains unclear why lifestyle intervention is more effective in girls than in boys.
The percentage of overweight 10-year-old girls in the control group of our study was slightly lower than the prevalence of overweight in American non-Hispanic white girls of the same age (22.9% compared to 27.7%).22 In boys, the prevalence of overweight was markedly higher in the US (19.1% in our study compared to 29.3% in the US).22 Thus, the prevalence of obesity in prepubertal and early pubertal children is probably somewhat lower in Finnish than in American children. In a previous Finnish study,23 the prevalence of overweight was 12.6 and 19.0% among 12-year-old girls and boys, respectively, but a different definition of overweight was used in that study. The comparison of the prevalences of the childhood overweight and obesity is overall hampered by the fact that definitions of childhood overweight and obesity vary a lot from one study to another. An effort has been made to develop definitions suitable for international comparisons16 but they have not been validated on Finnish population.
The family history of obesity is of special importance in obesity prevention. Our results indicate that the BMI's of the parents are strong predictors of a child's overweight. Previous studies show that parental obesity is a stronger predictor of obesity than the child's weight status before the age of 3 years.4, 24 Thus, family-based approach may be beneficial in obesity prevention. In the STRIP project, the counselling was mainly delivered to the parents until the child had reached the age of 7 years. Thereafter, the parents were met at each visit but increasing proportion of the information was targeted directly to the child. Weight management of the parents was also discussed when indicated but only superficially. Development of eating and physical activity habits of a child is a complex process of learning, in which social influences play an important role. Thus, our findings support the hypothesis that these healthy habits should be learned early in childhood to achieve long-term beneficial effects described above.25
Our previous analysis of the growth patterns of the study children26 and our current results clearly show that the counselling had no adverse effects on the growth of the children, as the height s.d.s and the weights for height of the girls and the boys in both study groups were constantly slightly above the mean values on the Finnish growth charts15 indicating that the children in the intervention group as well as in the control group were to some extent taller and heavier than the average Finnish children, and no difference was seen between the intervention group and the control group in height s.d.s and weights for height. The counselling given to the children indeed has led to marked beneficial nutritional effects as the relative intakes of fat and saturated fatty acids have constantly been lower and the relative intake of polyunsaturated fatty acids has been higher in the intervention children than in the control children once the counselling had started.26, 27 The higher intake of polyunsaturated fatty acids of the intervention children may be favourable for the weight management as it has been shown that polyunsaturated fatty acids exert a relatively stronger control over appetite than monounsaturated fatty acids or saturated fatty acids.28
Lifestyle education targeting to weight management should incorporate issues beyond healthy eating and physical activity because of the risk of promoting anorectic behaviour. In a study of schoolchildren in Hamburg, 30% of the boys and 50% of the girls evaluated themselves as being either ‘too fat’ or ‘much too fat’, and more than half of the girls aged 15 years and above had dieted at least once.29 In the STRIP project, none of the study children has thus far been diagnosed to have an eating disorder. The counselling given in the project is thus probably safe, since the proportions of slim children in the intervention group and the control group showed no difference.
The majority of the 9-year-old girls and all 9-year-old boys in this study were prepubertal. By the age of 10 years, approximately one-third of the girls and around 10% of the boys had entered puberty in both study groups implicating that the intervention did not affect the onset of pubertal development. Interestingly, two-thirds of the overweight girls had entered puberty by the age 10 years, whereas only one-third of the normal-weight girls showed physical signs of puberty. A limitation of our study is that the breast development was staged by inspection, which could cause a bias of overestimating the pubertal onset of overweight girls. However, similar findings have been reported in previous studies showing that overweight girls enter puberty earlier than normal-weight children.30, 31, 32, 33 The association between overweight and sexual maturation seems to be more complicated in boys since some studies have found a negative association and others (including our study) have found no association.34, 35
Weight management issues are an extremely delicate topic and in a study like ours one would expect that the overweight children would be more prone to dropout. However, this was not the case, as careful comparison of the previous features of the children and families that stayed in the study or dropped out show no significant differences. Almost equal numbers of children in the intervention group and in the control group dropped out of the trial before the 10-year follow-up visit. The same holds true for both genders.
In conclusion, our data show that beginning of promotion of healthy eating and physical activity patterns at a young age by delivering age-adjusted counselling to the children and their families may reduce development of overweight, in addition to diminishing the children's exposure to known environmental risk factors of coronary heart disease. The prevalence of overweight may thus be reduced with individualised and long-term counselling which does not have to be extremely intense. It remains to be seen whether the results achieved in weight management in our study will persist over puberty.
This study was supported by Grants from Finnish Cardiac Research Foundation; Foundation for Pediatric Research, Finland; Academy of Finland (Grant 206374); Special Federal Grants for University Hospitals in Finland; Yrjö Jahnsson Foundation; Sigrid Juselius Foundation; Turku University Foundation; Juho Vainio Foundation; and Finnish Cultural Foundation.