Letters to the Editor

Indian Pediatrics 1999; 36:613-615

Reply

There has been little agreement over the criteria or cut-off for the diagnosis of over- weight and obesity in children and adolescents. The WHO recently recommended the use of appropriate weight-for-height in children based on the NCHS growth standards and considered > +2 Z scores as a cut-off for overweight in children (1). This anthropometric indicator is recommended for screening overweight children in population samples while changes in weight-for-height or weight-for-age as well as absolute weight loss over a period of time may be used for the clinical monitoring of overweight children. In the case of adolescents, the same Expert Committee (1) recommends the use of both BMI-for-age and skinfolds-for-age. A cut-off of > 85th percentile of BMI-for-age is considered to be indicative of the risk of overweight in an adolescent while a >85th percentile of BMI-for-age plus a >90th percentile triceps skinfold-for-age and a >90th percentile sub- scapular skin-fold-age should be diagnostic of adolescent obesity.

Several controversial issues continue to persist despite these recent recommendations by the WHO. The use of BMI which is a clearly useful and robust indicator in adults, when applied to epidemiological studies is certainly not ideal in clinical practice. The use in population studies of BMI in adults has limitations when applied to individuals(2). How do we distinguish someone who is muscular from someone who is overweight if both have BMI in the overweight range (BMI 25-29.9). There is even more difficulty to decide at the individual level whether someone is likely to be undernourished because he has a BMI of 18 or because he is thin and physically very active. Despite these reservations BMI is an exceptionally robust index of nutritional status in adults (both under- and over nutrition) for epidemiological purposes. In children however, the problems with the universal use of BMI as an indicator of obesity, both for epidemiological purposes and in clinical practice, increase even further(3).

Firstly, which population might be best used as a reference population? Are we going to continue to rely on the data collected in children from USA whose National Center for Health Statistics (NCHS) growth references have been accepted by the WHO as Inter- national Reference standards? Do they represent the very different patterns in the body mass of different population groups and do they account for the morphometric differences in the timing of the growth spurts in weight and height in different population groups globally? On a positive note, the WHO is in the process of collecting representative growth data in children from different population groups to generate its own International Reference standards for wider acceptability and universal use. Secondly, the secular trends in growth evident in several developing countries is likely to influence the BMI and need to be interpreted with reference to some reference standard obtained at a specific time point. Thus, the monitoring of these changes over time become crucial and will limit the ready use of BMI in children ideally until the secular trends stabilize as they tended to do in the developed West. Thirdly, in developing societies like India there is a further problem with BMI being used to diagnose childhood obesity. There is increasing evidence that malnutrition will increase both the risk of the development of obesity and might even exaggerate its prevalence. Stunting in children is likely to alter the relationship of the appropriateness of weight for height (although on theoretical grounds the assumption is that BMI is an ideal indicator since it is independent of height certainly in adults). Thus, in children with nutritional stunting one may exaggerate the apparent prevalence of obesity since gain in weight may occur with little change if any, in stature on improvement in nutrition. Further, irrespective of the method of classification used, the commonly used indicator of pre- school childhood malnutrition (Wt for Ht) has been shown to exaggerate the problem of obesity among the stunted children in developing societies(4). Repeated episodes of malnutrition followed by nutritional rehabilitation is also known to increase the amount of adipose tissue laid down during the recovery process and increase the risk of obesity. This discordance between linear growth and adipocyte development will enhance adipocyte deposition when linear growth is affected by malnutrition. Given the enormous numbers of stunted children in developing societies and the persistence of malnutrition even in countries in rapid developmental transition, it is not unlikely that these factors will contribute to increase in the apparent prevalence of obesity in developing countries like India.

Hitherto, no consensus has emerged on the application of BMI to young children. This is largely the result of lack of data on the validity of BMI as a measure of adiposity in children, the absence of a universally acceptable reference population, lack of agreement on which cut-off point to use and more importantly, the fact that few studies have looked at the sensitivity, specificity and predictive value of the persistence of obesity into adult- hood or the development of its co-morbidities(5). It would hence be rather premature to rush into recommending the universal use of BMI in children. BMI has a number of advantages. Heights and weights are routinely measured and can be readily used to calculate BMI both in clinical practice and in epidemiological studies in the community. The important message however, I believe, that is highlighted by the letter from Dr. Adhikari is the recognition that there is a pressing need for a simple and universally applicable definition of obesity. in children and adolescents. On an optimistic note several committees are seized of this important issue. They include the International Obesity Task Force and the European Childhood Obesity Group and we eagerly await the recommendations that will follow.
 

P.S. Shetty,
Professor of Human Nutrition,
Head Public Health Nutrition Unit,
Department of Epidemiology and
Population Health, London Schools oj Hygiene and Tropical Medicine,
49-51, Bedford Square, London,
WCIB 3 DP, UK.