Send letter to journal:
Re: Establishing modifiable risk factors for obesity in children

We read your paper with interest. Some of the risk factors for obesity mentioned may not be modifiable and not taking diet into account may have a confounding effect on risk factors found significant. We would like to bring following points to your notice:

1. It was interesting to note that though small for gestational age (SGA) infants showed higher growth quartiles, they had significantly lower percentage body fat (PBF) at 3.5y and at 7 years as compared with those born appropriate for gestational age (AGA), which actually puts them at a lower risk of being obese in later life. Other studies suggest that SGA infants grow rapidly in the first 1-2 years of life and this rapid postnatal growth links them to central and peripheral obesity in childhood (1).

2. You do not mention the method you used to calculate the hours of television viewing each day. There may be great variation in the hours of TV being watched each day which would be, based on memory, purely a guess work as to how much TV each child watched every day in last 7 years and to correlate it with a precise measurement such as PBF can be a gross exaggeration. Further though sedentary life style is modifiable, physical activity alone may not be effective in reducing body mass index (2).

3. The relative importance of growth in different periods of childhood linking to later obesity is still unclear. Other data show that fast-growing infants of normal weight are at increased risk of developing obesity later in childhood(3, 4). The trajectory to obesity by your own admission starts in early infancy. It follows that maternal obesity hardly remains a modifiable risk factor considering that there may be a genetic element to it and even if it is environmental, there is lack of effective treatment or preventive strategies at present.

4. Though you acknowledge that diet can potentially confound or mediate some of the reported findings, not taking dietary intake into account potentially questions the validity of risk factors found significant including increased maternal BMI, sedentary lifestyle and greater watching of TV. Dietary energy intake during infancy determines infant weight gain and may influence obesity risk during childhood in at least among formula- and mixed-fed infants (5). The effect of diet cannot be taken in isolation but has to be considered in context to interpret any of the above finding into clinical application. Further you have not mentioned if any dietary data was collected at any time during the study.

References:

1. Ong KK, Ahmed ML, Emmett PM, et al. Association between postnatal catch-up growth and obesity in childhood: prospective cohort study. BMJ 2000; 320:967–71.

2. Reilly JJ, Kelly L, Montgomery C et al. Physical activity to prevent obesity in young children: cluster randomised controlled trial. BMJ 2006; 333:1041.

3. Papadimitriou A, Gousi T, Giannouli O et al. The growth of children in relation to the timing of obesity development. Obesity 2006; 14:2173-2178.

4. Stettler N, Zemel BS, Kumanyika S et al. Infant Weight Gain and Childhood Overweight Status in a Multicenter, Cohort Study. Pediatrics 2002; 109:194-199.

Send letter to journal:
Re: Response to letter from Dr Gargh

Thank you for your comments.

1. As we stated in the Discussion research by several groups, including that by Ong, have found an association between rapid post natal growth and obesity in childhood (1). Although in our cohort the small for gestational age (SGA) children were more likely to grow rapidly in the postnatal period, they did start from a lower percentage body fat (PBF, measured by ponderal index) at birth and thus despite increasing significantly were still lower compared to the appropriate for gestational age (AGA) born children at 3.5 and 7 years. It will be interesting to see what happens to their PBF at 11 years (currently being assessed) and later in adolescence. We predict that by adolescence and adulthood those born SGA will have a higher PBF than those born AGA. From a public health perspective it is important to realize that it is the children born AGA that form the majority if the children whom are overweight or obese.

2. Space precluded detailed explanation of every variable. The amount of time viewing television was based on the question completed by the parent: “During a normal week, how many hours a day does your child watch television? less than 1 hour; between 1 and 3 hours; more than 3 hours.” There was a strong correlation with PBF and television viewing hours despite the potential measurement error which would reduce the ability to detect an association. This shows how important television viewing hours is as a risk factor. In the study you quote no improvement in BMI was seen following a physical activity programme in preschool children. However there is evidence that multi-faceted family-based programmes that involve parents, increase physical activity, provide dietary education and target reductions in sedentary behaviour may be effective for weight management (2). There is also some evidence that programmes that target reducing sedentary behaviour are more effective than those increasing physical activity (3).

3. We agree that the relative importance of growth in different periods of childhood to later obesity is still unclear. Our study is important as it is the first to show that each of the three growth periods (postnatal, early childhood and middle childhood) are equally critical in the development of obesity at 7. We disagree with your statement that maternal obesity is not modifiable. It should not be ignored just because it is difficult to treat or prevent. For a society level strategy to tackle the obesity epidemic a comprehensive approach needs to be taken and all points along the lifecourse need to be targeted, including maternal prenatal health.

4. We agree that dietary intake is important to complete the identification of risk factors. We indicated in the paper that we will report the influence of diet on obesity in a future publication. We have collected dietary information at both the 3.5 year and the 7 year assessments by both a 24 hour dietary recall and by a weekly semi- quantitative Food Frequency Questionnaire. The analysis of dietary data itself is a complex issue and is still in progress, hence its non inclusion in the current analysis. We fully agree that it will be important to consider this dietary data in addition to those variables already included when it is available. We reported infant feeding data but due to New Zealand’s high rate of breast feeding (approximately 97% at discharge), the numbers not breast fed are too small to show the protective effect of breast feeding. The protective effect is known to be small, in the order of 10% reduction of risk (4).

References

1. Ong KK, Ahmed ML, Emmett PM, et al. Association between postnatal catch-up growth and obesity in childhood: prospective cohort study. BMJ 2000; 320:967–71.

2. NHS Centre for Reviews and Dissemination. The Prevention and Treatment of Childhood Obesity. Effective Health Care Bulletin. 2002;7(6).

3. Epstein LH, Paluch RA, Gordy CC, Dorn J. Decreasing sedentary behaviors in treating pediatric obesity. Arch Pediatr Adolesc Med. 2000 Mar;154(3):220-6.