Concerns with maternal obesity are relatively recent in conjunction with rising rates of obesity in U.S. society, but clinicians, dating back to the 1940s, have warned of the relationship between maternal obesity and poor infant outcomes.1 The most recent estimate of rates of obesity among adult women 20–39 years of age in the United States was 31.8%2; prepregnancy obesity rates in the United States have been estimated at 20% in recent years.3 Of particular concern is that high rates of overweight and obesity are also more common among groups that are already at higher risk for poor birth outcomes.3
Prior studies have examined aspects of the relationship between obesity and infant outcomes, both in developed4–8 and developing9 countries. A recent meta-analysis10 summarized the results from 11 studies that examined body mass index (BMI, calculated as weight [kg]/[height (m)]2) scores in the obese range (BMI greater than 30) and found a pooled odds ratio (OR) for the risk of infant death of 1.42 (95% confidence interval [CI] 1.24–1.63). However, prior studies have been limited in their applicability to a contemporary U.S. public health context by several constraints. Databases utilized to recently examine these questions include the Swedish Birth Register, but it is based on a largely homogeneous country with exceptionally low prepregnancy obesity and infant mortality rates.5,8 U.S.-based studies have been based on data from earlier periods when obesity rates were far lower4 or studies from single states.6,7
The issue of gestational weight gain and its relationship to prepregnancy obesity led the Institute of Medicine (IOM) to issue revised 2009 recommendations, which call for a range of 11–20 pounds for mothers with a prepregnancy BMI in the obese range with a recommendation for more research on the differential effect of obesity subclasses11 and compliance with the guidelines.12
The objective of this study is to address many of the limitations of prior studies by examining the relationships among maternal prepregnancy obesity, gestational weight gain, and infant mortality with 2 years of recent data from 38 U.S. states.
MATERIALS AND METHODS
This was a cohort study using data from the 2012–2013 national linked birth and infant death data sets. The 2003 revision of the U.S. Standard Certificate of Live Birth included information on the mother's height and prepregnancy weight that are needed to compute BMI. These data are primarily obtained from the first prenatal visit based on direct measurement of the mother's height and weight when the first visit occurred early in the pregnancy or from the mother's recall of her prepregnancy weight when the first visit occurred later in the pregnancy. Weight gain during pregnancy is computed by subtracting the mother's prepregnancy weight from the weight at the time of delivery as recorded in the hospital.13 State adoption of the 2003 revision has been gradual with 38 states and the District of Columbia in 2012 and 41 states in 2013 adopting the change.14 We examined the 38 states and the District of Columbia that had complete data for both years and accounted for 86% of all U.S. births in 2012. We used standard definitions for BMI categories (Table 1)15 and limited our analysis to singleton births. When examining weight gain, we limited the gestational age to 37–41 weeks to minimize the effect of gestation on weight gain. From an initial 7,885,022 overall U.S. births in 2012–2013, there were 6,419,836 singleton births in the 38 participating states and the District of Columbia and 36,691 infant deaths. Missing data were excluded from each variable before percentages were computed.
In the multivariate model we used logistic regression in models that included demographic and medical risk factors shown in the past to be related to infant mortality.16 The demographic factors included: maternal age, race–ethnicity, education, live birth order, gestational age, and source of payment for the birth. To address the possible influence of birth weight, we included a measure of small for gestational age.17 Maternal prenatal care, measured here as whether or not prenatal care began in the first trimester, was also included as were chronic diabetes and chronic hypertension. We tested for interactions between BMI and maternal race–ethnicity, age, live birth order, and medical risk factors (diabetes and hypertension). Of these, the only consistently significant interaction was with birth order and in a subanalysis we stratified our multivariate models by birth order and found no differences in the results. Given that the population studied is more than 6.4 million, tests of statistical significance are not presented in Table 1 because even small differences were significant. When we make specific comparisons of rates in the text, we cite the associated P value. Based on the multivariate logistic regression analyses, adjusted ORs and 95% CIs are presented.
Even with 2 years of data from 38 states, analyzing infant mortality by cause of death and BMI can result in small numbers of events. For this reason we analyzed cause-of-death data for three composite categories that are widely used in the literature: preterm-related causes, congenital anomalies, and sudden unexpected infant death. The derivation of these composite categories has been described in detail elsewhere.18,19 The study was based on the publicly available linked Birth-Infant Death data set (http://www.cdc.gov/nchs/data_access/Vitalstatsonline.htm) from the National Center for Health Statistics and was exempt from institutional review board review.
Table 1 presents the infant mortality rates by BMI category and key subgroups for singleton births. The exclusion of multiple births is the reason the rates in Table 1 are lower than overall national averages. The table identifies both subgroups at greatest risk and the pattern of change within subgroups across BMI categories and because the former is available from other sources,16 we focus on the changes associated with BMI. An overall “J pattern” is seen across subgroups with underweight births having higher infant mortality rates (5.4/1,000) than normal-weight births (4.2) (P<.001), but the infant mortality rate begins to increase rapidly across the obesity categories: 5.9 for obese I (P<.001) to 8.2 for obese III (P<.001) when compared with normal births. The familiar pattern of higher infant mortality rates in births to non-Hispanic black mothers when compared with non-Hispanic white mothers is seen in higher infant mortality rates in every BMI category (P<.001 in each category). Births to normal-weight non-Hispanic black mothers have a higher infant mortality rate (8.0) than births to non-Hispanic white (6.65; P<.001) or Hispanic (6.6; P<.001) mothers in the obese III category. Other subgroups in which births to mothers in the obese III category have greater than 1% infant deaths (infant mortality rate greater than 10 per 1,000) are mothers younger than 20 years of age, those with less than a high school education, first-time mothers, smokers, those with chronic diabetes, mothers who reported paying for the birth themselves as well as premature and small-for-gestational-age births.
Table 2 presents infant mortality rates per 100,000 live births for the three major cause-of-death groupings by BMI category. These three major groupings account for 70.5% of all infant deaths. Nearly one third of infant deaths were the result of preterm-related causes, which include disorders related to short gestation and low birth weight not elsewhere classified and most maternal complications of pregnancy as well as more specific categories such as respiratory distress of the newborn, bacterial sepsis of the newborn, and others.16 The infant mortality rate from preterm-related causes increased most rapidly with increasing BMI and was twice as high for obese (228) as for normal-weight women (114; P<.001). Deaths from congenital anomalies were also more common in the obese category (130/100,000) than among normal-weight births (108/100,000; P<.001), a finding consistent with past research.20 “Sudden unexpected infant death” is a category that includes sudden infant death syndrome, an ill-defined and unknown cause of mortality, and “accidental suffocation and strangulation in bed,” and the rate of deaths for this cause was higher in obese (106/100,000) than in normal-weight (74/100,000; P<.001) categories.
Table 3 presents the IOM recommended weight gains by BMI category,11 the mean reported weight gains and distributions by category as well as the associated infant mortality rates. We limited this analysis to singleton births between 37 and 41 weeks of gestation, which accounts for the lower infant mortality rates. Only in the case of underweight and normal-weight mothers did the average weight gain fall within the recommended ranges. For the overweight and obese categories, mothers on average gained more than the recommended amount. However, a majority of mothers in the obese III category gained either less than (33%) or within (25%) the recommended range. In four of the six BMI categories (underweight, normal weight, overweight, and obese I), births to mothers whose weight gains were within the recommended range had a lower infant mortality rate than mothers who gained less than the recommend amount (P<.001). Births to mothers in the overweight (P<.001) category had a lower infant mortality rate (P<.001) when more than the recommended amount was gained compared with those who gained weight in the recommended range.
We examined the relationship between BMI and infant mortality rate overall and for neonatal and postneonatal mortality controlling for multiple confounders and the results are presented in Figure 1 and Table 4. In the multivariate analysis, the “J pattern” seen in the unadjusted analysis disappears with a lower risk for infant mortality among births to mothers in the lowest BMI category (Fig. 1). In the adjusted models, there remains a consistent increase in the risk of infant death across BMI categories for overall infant mortality as well as neonatal and postneonatal mortality. Births to mothers in the obese III category as pregnancy began were associated with a significantly higher risk of death in the first 28 days (74% higher), from 28 days to 1 year (51% higher), and overall in the first year (73% greater).
Stratifying the analysis to look at the relationships specifically by race–ethnicity, the familiar pattern is seen in Table 4, although in the case of non-Hispanic black mothers, the increases in infant mortality are generally more pronounced. To better understand the possible effect of adherence to weight gain recommendations, we added adherence to the model, but given the relationship between length of gestation and weight gain, we limited the analysis to full-term (37–41 weeks of gestation) births in Table 4. Although the actual infant mortality rates were generally lower, the pattern of escalating risk across higher BMI categories remained the same for full-term births in the adjusted analysis without including gestational weight gain. Adding adherence to the weight gain recommendations had no effect on the pattern.
We also stratified the analysis by adherence to weight gain recommendations. With gaining the recommended weight as a reference, the adjusted ORs for the risk of an infant death for those gaining less than the recommended weight was 1.07 (95% CI 1.01–1.12); it was 1.04 (95% CI 0.99–1.09) for those gaining more than the recommended weight.
We did a series of subanalyses to assess whether they would have an effect on the relationships examined. Limiting the sample to women who did not have either diabetes or hypertension had no effect on the familiar pattern. Stratified analyses by payer source and parity also had no effect on the finding of increasing rates of infant death across prepregnancy BMI categories nor did stratifying by smoking status.
We used a data set of more than 6.4 million births to examine the relationship among prepregnancy BMI, gestational weight gain, and infant mortality. We found a strong relationship between prepregnancy BMI and infant mortality, which increased substantially across categories of obesity even after controlling for demographic and medical risk factors that have been shown to be related to infant mortality. These differences were consistent for both neonatal and postneonatal mortality and across the major causes of infant death. Many mothers whose gestational weight gain was within the ranges recommended by the IOM had births with slightly lower infant mortality rates than those who did not, but after adjustment, the differences largely disappeared.
These findings are generally consistent with past research although most prior studies focused on neonatal mortality and did not include either cause of death or weight gain data. Both a recent meta analysis10 and a large population-based study from Sweden8 found increasing rates of infant mortality across categories of BMI. Johansson et al found even larger increases in the adjusted ORs across BMI categories,8 although its applicability to a U.S. context may be limited. Sweden's rate of prepregnancy obesity is half that of the U.S. (9.9% compared with 20%) and its infant mortality rate is less than half that of the United States (2.6 compared with 6.0/1,000 in 2012).21 Chen et al, using data from the 1988 National Maternal and Infant Health Survey, found a clear relationship between prepregnancy BMI and neonatal death but a weak relationship with postneonatal death. They found a mixed effect of weight gain, with mothers in the underweight category who gained the least weight and the mothers in the obese category who gained the most weight, at greatest risk for infant death.4 Salihu et al,7 using Missouri data from 1978–1997 and limiting analysis to neonatal mortality, found a relationship between BMI and early neonatal but not late neonatal death for white mothers and a relationship with all neonatal death for black mothers.
Although this study relies on a large data set, it does have several limitations. Although the data account for 86% of all births in the United States, it does not include data from 12 states and thus is not fully representative of all U.S. births.14 As noted earlier, the information on prepregnancy weight as recorded on the birth certificate comes from measurement at the first visit or maternal recall. Birth certificate data do not permit stratification by these sources, so we were unable to examine the implications of the alternative bases for the measure of BMI. The measurement of BMI from the birth certificate is relatively new and there have been limited studies of the accuracy of birth certificate reporting of BMI. Park and colleagues22 found the BMI birth certificate measure to be generally valid and reliable for population-based research. Bodnar et al23 found BMI reporting most reliable for normal, overweight, and obese 2 and 3 categories but less accurate for underweight and obese 1 categories. Chen and colleagues4 also found a strong correlation between self-reported and measured prepregnancy BMI. A National Center for Health Statistics study of the quality of reporting on the revised 2003 birth certificate found reasonable data quality for the variables used here in our model such as gestational age, trimester prenatal care began, and source of payment; however, diabetes and hypertension appear to be underreported in birth certificate data.24 Like most published studies from the United States,4,25 we used the birth certificate measure of gestational weight gain. The limited studies on the accuracy of the birth certificate weight gain measure have found it to not be as reliable as the BMI measure,4 although Oken et al26 found a high correlation between self-reported and clinically measured weight gain.
Although the association between high BMI and infant death is robust, that is not proof that prepregnancy weight loss, whether by lifestyle changes or bariatric surgery, will change this relationship. It is possible that women who are predisposed to obesity have an underlying pathology that leads to the poor outcomes even if they lose weight. Studies of bariatric surgery suggest overall improved birth outcomes after weight loss,27 although these studies were not sufficiently powered to examine infant death.
We included small for gestational age in our analysis with little effect on the overall pattern of relationship between BMI and infant mortality. However, one exception was the case of underweight mothers who had a slightly higher risk of infant mortality in the initial model, but after the small-for-gestational-age variable was added to the final model, that relationship disappeared. The role of maternal underweight on infant outcomes, although not the focus of this study, deserves further attention.
This study, in conjunction with others that have demonstrated an association of maternal obesity with increased risk of perinatal death and infant mortality, builds a convincing foundation for future studies of perinatal outcomes for women who were obese and attained a normal or more normal BMI before conception. If such studies demonstrate that preconception weight loss is associated with improved outcome, even if the mechanism underlying that improvement is unknown, a public health campaign encouraging preconception weight loss to benefit not only the maternal but also the fetal and neonatal outcomes would be scientifically grounded, feasible, and compelling.
The nature of the weight gain recommendations once a woman is pregnant also deserves additional attention. The focus of the IOM guidelines was on general maternal and child health and not simply infant mortality, but the findings give pause concerning both the lack of adherence and the apparent limited benefits of compliance, at least in terms of infant mortality. The IOM's recommendation concerning more research on the effect of the subclasses of obesity seems particularly well-founded given these findings.
Concern about the effect of obesity on the nation's health has drawn the attention of providers, policymakers, and the public, most notably the First Lady's “Let's Move” campaign targeting childhood obesity.28 This study suggests the need for more research into the benefits of a systemwide effort to prevent obesity among women of reproductive age.
1. Gilbert JA. Maternal obesity, large babies, and diabetes. Br Med J 1949;1:702–4.
2. Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of childhood and adult obesity in the United States, 2011–2012. JAMA 2014;311:806–14.
3. Fisher SC, Kim SY, Sharma AJ, Rochat R, Morrow B. Is obesity still increasing among pregnant women? Prepregnancy obesity trends in 20 states, 2003–2009. Prev Med 2013;56:372–8.
4. Chen A, Feresu SA, Fernandez C, Rogan WJ. Maternal obesity and the risk of infant death in the United States. Epidemiology 2009;20:74–81.
5. Nohr EA, Vaeth M, Bech BH, Henriksen T, Cnattingius S, Olsen J. Maternal obesity and neonatal mortality according to subtypes of preterm birth. Obstet Gynecol 2007;110:1083–90.
6. Thompson D, Clark C, Wood B, Zeni MB. Maternal obesity and risk of infant death based on Florida birth records for 2004. Public Health Rep 2008;123:487–93.
7. Salihu HM, Alio AP, Wilson RE, Sharma PP, Kirby RS, Alexander GR. Obesity and extreme obesity: new insights into the black-white disparity in neonatal mortality. Obstet Gynecol 2008;111:1410–6.
8. Johansson S, Villamor E, Altman M, Bonamy AK, Granath F, Cnattingius S. Maternal overweight and obesity in early pregnancy and risk of infant mortality: a population based cohort study in Sweden. BMJ 2014;349:g6572.
9. Cresswell JA, Campbell OM, De Silva MJ, Filippi V. Effect of maternal obesity on neonatal death in sub-Saharan Africa: multivariable analysis of 27 national datasets. Lancet 2012;380:1325–30.
10. Meehan S, Beck CR, Mair-Jenkins J, Leonardi-Bee J, Puleston R. Maternal obesity and infant mortality: a meta-analysis. Pediatrics 2014;133:863–71.
11. Institute of Medicine and National Research Council. Weight gain during pregnancy: reexamining the guidelines. Washington, DC: National Academies Press; 2009.
12. Deputy NP, Sharma AJ, Kim SY, Hinkle SN. Prevalence and characteristics associated with gestational weight gain adequacy. Obstet Gynecol 2015;125:773–81.
13. National Center for Health Statistics. Weight gain computation reference. Guide to completing the facility worksheets for the certificate of live birth and report of fetal death. Hyattsville (MD): National Center for Health Statistics; 2003.
14. Martin J, Hamilton B, Osterman J, Curtin S, Matthews T. Births: final data for 2013. Hyattsville (MD): National Center for Health Statistics; 2015.
15. National Center for Health Statistics. User guide to the 2013 natality public use file. Hyattsville (MD): National Center for Health Statistics; 2015.
16. Mathews TJ, MacDorman M, Thoma M. Infant mortality statistics from the 2013 period linked birth/infant death data set. Hyattsville (MD): National Center for Health Statistics; 2015.
17. Oken E, Kleinman KP, Rich-Edwards J, Gillman MW. A nearly continuous measure of birth weight for gestational age using a United States national reference. BMC Pediatr 2003;3:6.
18. Callaghan WM, MacDorman MF, Rasmussen SA, Qin C, Lackritz EM. The contribution of preterm birth to infant mortality rates in the United States. Pediatrics 2006;118:1566–73.
19. Nashelsky M, Pinckard J. The death of SIDS. Acad Forensic Pathol 2011;1:92–8.
20. Gilboa SM, Correa A, Botto LD, Rasmussen SA, Waller DK, Hobbs CA, et al.. Association between prepregnancy body mass index and congenital heart defects. Am J Obstet Gynecol 2010;202:51.e51–10.
21. OECD health data 2015: statistics and indicators for 34 countries. Paris (France): OECD; 2015.
22. Park S, Sappenfield W, Bish C, Bensyl D, Goodman D, Menges J. Reliability and validity of birth certificate prepregnancy weight and height among women enrolled in prenatal WIC Program: Florida, 2005. Matern Child Health J 2011;15:851–9.
23. Bodnar LM, Siega-Riz AM, Simhan HN, Diesel JC, Abrams B. The impact of exposure misclassification on associations between prepregnancy BMI and adverse pregnancy outcomes. Obesity 2010;18:2184–90.
24. Martin J, Wilson E, Osterman M, Saadi E, Sutton S, Hamilton B. Assessing the quality of medical and health data from the 2003 birth certificate revision: results from two states. Hyattsville (MD): National Center for Health Statistics; 2013.
25. Kiel DW, Dodson EA, Artal R, Boehmer TK, Leet TL. Gestational weight gain and pregnancy outcomes in obese women: how much is enough? Obstet Gynecol 2007;110:752–8.
26. Oken E, Taveras EM, Kleinman KP, Rich-Edwards JW, Gillman MW. Gestational weight gain and child adiposity at age 3 years. Am J Obstet Gynecol 2007;196:322.e1–8.
27. Yi XY, Li QF, Zhang J, Wang ZH. A meta-analysis of maternal and fetal outcomes of pregnancy after bariatric surgery. Int J Gynaecol Obstet 2015;130:3–9.
28. Obama M. Let's move! Raising a healthier generation of kids. Child Obes 2012;8:1.