Original ArticleA longitudinal study of pediatric body mass index values predicted health in middle age
Introduction
Obesity is related to various chronic diseases in adults, such as cardiovascular diseases, diabetes mellitus, and some cancers [1], [2], [3]. It is reaching epidemic levels in many countries, and preventive strategies are needed [4]. Childhood overweight and obesity may persist into adulthood [5], [6], [7]. It is easier to induce weight control among overweight children than overweight adults [8], [9], [10], [11]. Early identification of children who are at high risk of becoming obese adults is therefore important [12].
Some research has revealed a relationship between overweight in childhood and young adulthood. For instance, Whitaker et al. [13] found an association between obesity in the age range 21 to 29 and childhood and parental obesity in 854 persons from the United States. Williams [14] showed that body mass index (BMI) tracked from childhood to age 21 in a cohort of New Zealanders. Useful though they are, these studies tended to focus on the issue of association rather than the diagnostic features of using childhood BMI as a tool for the early identification of people at health risk. Furthermore, there is no commonly agreed definition of child overweight or obesity based on BMI. Cole et al. [15] developed an international reference of childhood BMI values by extrapolating the BMI centile values at age 18 years backward into childhood. Guo et al. [16] and He and Karlberg [17], [18] suggested defining childhood BMI according to the risk of becoming overweight in adults. Recently, a cross-sectional study of Canadian children and adolescents confirmed that the international cutoff values were associated with coronary heart disease risk factors [19]; however, the authors rightly pointed out that, despite the association, the cutoff values might not be optimal. They recommended further studies using longitudinal data and receiver operating characteristics (ROC) analysis to determine the cutoff points for identifying children at risk, taking into account both sensitivity and specificity.
Drawing on data from a large-scale, population-based study, our objective was to characterize the various aspects of diagnostic values of using BMI measured at pediatric ages to predict obesity, overweight, and diseases in adulthood. Furthermore, we compared the choice of different pediatric BMI cutoff points for predicting the adult outcomes, and examined whether there is a recommendable age for making the prediction and consequent intervention.
Section snippets
Participants and materials
The National Child Development Study (NCDS) is a cohort study of about 17,000 people born in England, Wales, and Scotland in a week in March 1958. The NCDS included medical examinations at ages 7, 11, and 16 (pediatric ages), when weights and heights were measured by medical personnel. In the face-to-face survey at age 33, weights and heights were measured by trained interviewers, using a standardized method with a Salter portable scale and stadiometer. At age 42, the respondents self-reported
Results
A total of 12,327 cohort members were included in the ROC analysis; all had nonmissing BMI values in at least one pediatric age and participated at ages 33 and/or 42. Table 1 shows the mean, standard deviation (SD), and 85th and 95th percentiles of BMI at various ages, by gender. At ages 33, 11% of the men and 12% of the women were obese; 51% of men and 35% of women were overweight. Table 1 also gives the percentages of children whose BMI exceeded the international standards for obesity and
Discussion
We have shown the value of using BMI measured at pediatric ages to predict obesity and overweight in middle age. BMI values at ages 11 and 16 could predict adult obesity, with areas under ROC curve of about 0.80. They could also predict adult overweight, with areas under ROC curve of about 0.75. Childhood BMI values are also predictive of diabetes and hypertension in middle age, but the areas under ROC curves were only around 0.60. We have also shown the very different diagnostic features of
Acknowledgments
The authors thank the UK Data Archive, and the Centre for Longitudinal Studies, London, UK, for providing the data. The authors also thank the anonymous reviewers for useful comments on an early version of this article.
References (42)
Pathophysiology of obesity
Am J Clin Nutr
(1992)Obesity: adverse effects on health and longevity
Am J Clin Nutr
(1979)- et al.
The predictive value of childhood body mass index values for overweight at age 35 y
Am J Clin Nutr
(1994) - et al.
Do obese children become obese adults? A review of the literature
Prev Med
(1993) - et al.
Predicting overweight and obesity in adulthood from body mass index values in childhood and adolescence
Am J Clin Nutr
(2002) - et al.
The utility of the international child and adolescent overweight guidelines for predicting coronary heart disease risk factors
J Clin Epidemiol
(2003) - et al.
The agreement between self-reporting and clinical diagnosis for selected medical conditions among the elderly in Taiwan
Public Health
(2000) - et al.
The validity of self-reported hypertension and correlates of hypertension awareness among blacks and whites within the stroke belt
Am J Prev Med
(1995) - et al.
Comparison of survey and physician claims data for detecting hypertension
J Clin Epidemiol
(1997) - et al.
Validity of adolescent and young adult self-report of Papanicolaou smear results
Obstet Gynecol
(2000)