Author’s Reply to a Letter to the Editor Regarding “The Association Between Inadequate Gestational Weight Gain and Infant Mortality Among U.S. Infants Born in 2002”

Abstract

We want to thank Drs. Bodnar, Hutcheon and Abrams for their comments on our paper, ‘‘The Association between Inadequate Gestational Weight Gain and Infant Mortality among U.S. Infants Born in 2002.’’ We were pleased to read that the correspondents agree that the link between gestational weight gain and infant mortality is an emerging and important research topic. On the overall question, previous studies have demonstrated an association between inadequate gestational weight gain and infant death using nationally representative samples [1, 2]. We do not believe there is any dispute over whether or not such an association exists. Rather, the issue raised appears to be one of the magnitude of the association. The correspondents appear to base this upon three main questions: (1) whether gestational age at delivery received adequate adjustment, (2) whether gestational weight gain was classified correctly as inadequate or adequate, and (3) whether important confounding influences were controlled. We provide a pointby-point response to their key comments below. On the question of whether there was adequate adjustment for gestational age, we adjusted for whether or not the birth was preterm (prior to 37 weeks or not) in model 4. After so doing, the association between inadequate gestational weight gain and infant mortality remained (AOR = 1.84, P \ 0.0001, 95 % CI: 1.51, 2.23). Additionally, we repeated the published analysis incorporating the influence of the suggested three categories of gestational age (\33, 33–36,[36 weeks). The sample size for gestational age of less than 33 weeks was small. The AOR of 1.84 attenuated to 1.57 (P \ 0.0001, 95 % CI: 1.30, 1.90), but statistical significance remained. The second question inquired about the potential for gestational weight gain misclassification because of lack of information on pre-pregnancy Body Mass Index (BMI). Although the present study uses the 2009 Institute of Medicine (IOM) guidelines (which now include a recommended weight gain range for obese women) as well as the BMI categories developed by the World Health Organization (which classify more women as overweight and fewer women as obese), the overweight and obese women who gave birth during 2002 would have been advised to gain weight according to the 1990 IOM guidelines. Those guidelines recommended a weight gain of at least 15 lbs with no upper range for obese women. According to data collected by the National Center for Health Statistics, during 1999–2004 nearly half of women were overweight [3]. Thus, we believe that our sensitivity analysis, in which weight gain of less than 15 lbs was considered inadequate, would have reduced the potential for misclassification of overweight or obese women who gained within the recommended range. Data from U.S. birth certificates in 2002 simply cannot provide information about gestational weight gain relative to pre-pregnancy BMI and we agree that the loss in precision and the degree of any bias due to self-reporting must be considered when interpreting the results. The present article has a lengthy discussion of these limitations. However, we do not agree that loss of precision necessarily limits the utility of the study. Science progresses by accumulation of evidence in small steps. R. R. Davis (&) S. L. Hofferth Department of Family Science, University of Maryland, College Park, MD 20742, USA e-mail: regina.davis@att.net