A pragmatic controlled trial to prevent childhood obesity within a risk group at maternity and child health-care clinics: results up to six years of age (the VACOPP study)

The prevalence of obesity even among pre-school-age children has increased in recent decades, and this is a global trend [1]. Overweight and obesity prevalence is significant already in the pre-school years: 16.1% and 3.9% of five-year-old girls and 7.5%/3.0% of boys of the same age in Finland are reported to have overweight and obesity, respectively [2]. Early adiposity rebound (AR) has been found to be a marker of higher risk for obesity in children and youth; AR is the point of minimal body mass index (BMI) before the second rise in the BMI curve in childhood, normally between five and 7 years of age (AR is considered to be early if it occurs before the age of 5 years) [3, 4]. Pre-schoolers with obesity tend to become schoolchildren and teenagers with obesity, which leads to increased risk of cardiovascular disease in adulthood and to an intergenerational cycle of these health problems [5, 6]. Results of obesity treatment are not encouraging, and prevention of excess weight gain is considered the most effective way to reduce obesity prevalence both during childhood and in adulthood. Early-childhood obesity has a multifactor origin [7, 8]. Prenatal modifiable factors suspected to promote obesity are mother’s obesity before pregnancy, gestational diabetes mellitus (GDM), and smoking during pregnancy, with another being excessive weight gain during pregnancy [7, 9‐14], and GDM appears to increase the risk of obesity in offspring even in cases of normal birth weight [15‐17]. Large-for-gestational-age newborns have been shown to have a higher risk of obesity; also, infant feeding, sleep duration, and rapid weight gain in the first few months have been shown to influence the risk of children gaining excess weight [18‐27].

In light of these potential risk factors, obesity prevention should start early in life. The widespread problem of obesity calls for preventive means that can be integrated into existing health-care settings and also for changes in society that contribute to healthy weight gain in the population [7, 8]. Pragmatic trials are aimed at finding effective preventive programmes that could be incorporated into the usual health-care system [28]. Pregnant mothers and families with small children visit child-welfare clinics regularly in primary health care. They are also interested in the wellbeing of their offspring and hence are receptive to lifestyle counselling. Dietary and physical-activity habits are modifiable during the pre-school years [29, 30]. With lifestyle counselling, a significant effect can be achieved when the target group are known to be at risk of gaining excessive weight. Mothers at risk of developing GDM and their offspring are one such risk group [7]. This group includes pregnant mothers with overweight or obesity, mothers with a history of GDM, a macrosomic newborn or close relatives with type 2 diabetes [31]. These mothers may also have a hereditary predisposition to obesity and type 2 diabetes, with a high risk of passing these risks to their offspring.

To the best of our knowledge, no previous results have been published from intervention studies that have aimed primarily at the prevention of obesity among the offspring and that have started during or before the first trimester of pregnancy. This is at odds with the growing evidence that the time before conception, the prenatal and perinatal periods, and early childhood are the critical windows for effective prevention. Some obesity-prevention studies targeted at infancy have been reported on, with most involving short intervention and follow-up periods [32, 33]. The effect on children’s adiposity or weight development, if any, has been found to be slightly positive [34‐36]. Some randomised studies are still in progress [37‐39]. There are a few studies, originally examining pregnancy outcomes such as excess weight gain during pregnancy, prevention of GDM, or postpartum weight retention, in which, additionally, the offspring’s weight development was evaluated for 1–7 years of age [40‐44]. Intensified counselling on diet and physical activity (PA) directed at mothers during the infant’s first year resulted in offspring’s slower weight gain by the age of 4 years in a cluster-randomised pilot study [45]. This intervention when applied during pregnancy did not have the same effect on offspring weight [41]. In a study by Gillman et al., treatment of mild GDM had no effect on the offspring’s weight gain by age 4–5 years [40], and likewise no effect on pre-school weight gain was found for the gestational lifestyle intervention of the NELLI Study, the Lifestyle in Pregnancy and Offspring (LiPO) study, or the study by Vesco et al. of a weight-management intervention for limiting gestational weight gain (GWG) a in a group of women with obesity [42‐44]. Evidence from the obesity-prevention programmes reported upon has shown that multifaceted intervention could be more effective than targeting a single behaviour [7, 32].

The main results of the controlled lifestyle intervention designed to prevent obesity before school age (the VACOPP, or Vaasa Childhood Obesity Primary Prevention, study) are reported here [46]. The setting of the study was maternity and child health-care clinics in the city of Vaasa, in Western Finland. The intervention started at the end of the first trimester for pregnant mothers and continued with their families until the child was 5 years old. The outcomes presented here cover the offspring’s weight gain along with overweight and obesity incidence in the trial groups until the age of 6 years.

The study flow is described in Fig. 1. Roughly 700 women per year give birth in the city of Vaasa. In the intervention group, the offspring of 71 of the 127 mothers who started the intervention in pregnancy (56%) were still taking part in the study when the child was 6 years old (i.e., at the planned end of the study), with the corresponding figure for the control group being 76 out of 89 children (85%). Most of the dropouts were cases of moving to another city and hence being unable to remain in the study. We analysed baseline characteristics that might interfere with offspring weight development with regard to those children whose anthropometrics were available when they were 6 years old (n = 147) and found no statistically significant differences between the groups (Table 1). The baseline characteristics of children whose anthropometrics were available at age 1 year (n = 185) have already been reported [51].

According to the linear mixed-effects model, the BMI-SDS slopes did not differ significantly between the intervention and the control group (the p-values for linear and quadratic interactions were 0.89 and 0.81) (Table 2, Fig. 2). Adding gender and mother’s pre-pregnancy BMI to the model did not fundamentally affect the results. The proportions (expressed as percentage value deviation from the mean weight-for-height value in line with the Finnish definition of pre-school-age overweight and obesity) of children at the age of 6 years with at least overweight (≥ + 10% weight for height) or with obesity (≥ + 20% weight for height) were not significantly different between the groups. The result was the same when at least overweight and obesity were assessed as ISO-BMI (≥ 25 kg/m² and ≥30 kg/m², respectively) (Table 3). The difference in equivalent proportions of early adiposity rebound (< 5 years) between the two groups was not significant either (controls 34/45.3% vs. intervention 29/42.0%, p = 0.69) (Table 3).

The main result found for our pragmatic lifestyle intervention was a lower occurrence of GDM in the intervention group than in the control group, which result was reported earlier [51], [Table 1]. However, whether the intervention was effective in decreasing excessive weight gain among offspring remains an open question. The non-significant finding might be due also to the low power of the study causing failure to reveal differences between the groups. It has been shown that lower gestational glucose levels may be correlated with a child’s lower obesity and type 2 diabetes risk [11, 12, 17]. Rapid weight gain during the first year of life has been demonstrated to predict risk for later obesity [22]. In our study, the offspring’s weight gain up to 12 months of age did not differ significantly between groups, but there were slightly more children with overweight in the control group by 1 year of age [51]. Likewise, rapid weight gain in subsequent pre-school years seems to predict obesity in the school years [23]. In addition, early adiposity rebound has been shown to precede obesity in childhood and adulthood and to be a marker of cardiometabolic risk [3, 53]. In our study, no significant difference in the groups’ proportions of early vs. normal AR was found, but the proportion of children with early AR in both groups was high, predicting the offspring having the same metabolic risk as their mothers. The proportion of children at the age of 6 years with obesity in both group was high as well (defined as weight for height 12.9% and 13.2%) [2]. These results confirm that our target group for such an intervention may be appropriately chosen.

The offspring’s BMI was analysed and adjusted in accordance with the Finnish growth reference, for obtaining the SDS [50]. Weight gain was assessed with a linear mixed-effects model, which allows for a difference between the groups at baseline, intervention effects, and changes over time. No significant differences between the intervention and control group’s offspring weight gain during the first year or up to 6 years of age were found. Given that improvements in foetal conditions – such as the mother having a better glucose balance during pregnancy – have been shown to correlate with a good effect on offspring weight gain that emerges in the toddler years. Based on this our intervention had potential to diminish children’s overweight/obesity prevalence by age six [11, 12, 17]. However, as we have noted, the insufficient power of the study may have affected the results in this respect.

The overall dropout rate for the intervention group up to 6 years of age was 44% (Fig. 1). The most common reasons for dropping out were moving to a city out of reach of this intervention and the parents experiencing the study intervention or completing the questionnaires as too taxing. Furthermore, the recommendation to participate in blood tests every 2 years was felt to be too taxing for the child in many families, creating reluctance to take part in the study even despite the option of skipping the tests. The dropout rate in our study is acceptable in view of its longer-term intervention and follow-up. There were also dropouts in the control group (15% by age 6 years). It is possible that those families with the healthiest lifestyle and lowest risk of offspring’s excess weight gain were more likely to remain in the study, thereby diminishing the difference in proportions of children with overweight and obesity between groups. However, the baseline characteristics were comparable between groups at both 1 year and 6 years of age.

Our target group was mothers at risk of developing GDM and their offspring with a higher risk of unhealthy weight gain. The intervention extended across foetal, infant, and pre-school life, known times of risk for development of obesity. Almost 98% of the mothers in Finland visit municipal maternity health-care facilities, and the high participation percentage holds for child health-care clinic visits. If the intensified counselling is offered during these routine visits, the at-risk families are conveniently reached. However, those routine visits to child health-care clinics take place only once a year, which may entail too light an intervention for this risk group. Also, evidence is growing that intervention for this purpose should start even before pregnancy, to improve the mother’s metabolic health and hence a better prenatal environment in regard of the child receiving a healthier epigenetic heritage [54]. One marked problem is how to reach the risk group with childbearing potential for intensive counselling before pregnancy. Child health-care clinics may be a useful environment for targeting mothers with small children before pregnancy, but this is not true for first-time mothers. In addition, since obesity tends to begin in the early years, focusing more intensive lifestyle counselling also on offspring age 0–2 years within risk-showing families could be effective.

Our study had several limitations. It was not randomised, and the power may not have been sufficient to reveal statistically significant results. We believe also that, as the difference in study-group BMIs proved to be so small, precise primary power calculations would not have shown the number of participants to be sufficient for statistical significance in this intervention trial. An additional factor is that we wanted to perform the trial in this specific relatively large city in Finland, where the protocol is the same across all maternity health care, thereby primarily comparable in that regard. For this pragmatic trial, a randomised controlled design was not considered feasible, because the randomisation process would have been very likely to further reduce the rate of participation in the trial. A case-control study design is the choice in intervention studies when randomisation is not feasible and the study groups are matched as in our study (Table 1). The study design was discussed also in the protocol article [46]. The control group was prospective only from offspring age of 1 year, which may have caused some bias in the results; however, our choice may also have eliminated a possible Hawthorne effect on the control group during the intervention during pregnancy. As is the case with any pragmatic trial, the effectiveness of the counselling situation as a whole might have varied greatly. For example, the motivation of PHNs may vary, and the need for PHN deputies occasionally has an influence on counselling. The recruitment of the intervention group and the paperwork for the study were considered burdensome by some PHNs, mainly for reason of their busy work schedule. Allocating enough time for PHNs to manage the risk-group intervention appointments is crucial also.

One element of our study in its defence is its implementation in real-life practice, which demonstrates the counselling’s ability to be a sustainable part of municipal health care. Also, the maternity and child health-care clinics have a good opportunity to identify those at risk for childhood obesity at a stage in life when favourable lifestyle changes promote the offspring’s health most. Targeting the at-risk population in a setting that all families in this life situation visit eliminates the risk of stigmatisation. The costs of this study were quite moderate, and the results are generalisable to normal health care, because the study was realised as a part of usual practice at maternity and child health-care clinics.

Obesity with its expensive health effects and economic disadvantages challenges us to initiate solid preventive actions. Primary health-care, maternity, and child health-care clinics reach the beginning of the next generation. Preventive pragmatic trials in real-life settings are needed if we are to target obesity risk groups extensively and economically. In our study, the previously reported improved glucose tolerance during pregnancy demonstrated potential to have a good effect also on offspring weight gain. However, this effect could not be seen in the study. The offspring in both groups showed a high occurrence of early adiposity rebound and high prevalence of obesity, confirming their risk-group status. The knowledge now available suggests that preventive lifestyle interventions should start even before conception, to be able to influence the foetal environment effectively, and also focus on the child’s first 2 years, to cover this time with its special risk for obesity development. In addition to applying the right timing, there may be a need for putting more effort and time into the intervention if it is to result in obesity prevention in children in pragmatic settings in health care.