SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Objective: Approximately one-third of US reproductive-aged women are obese, and prepregnancy obesity is a strong risk factor for adverse fetal and infant outcomes. The annual number of preventable adverse fetal and infant outcomes associated with prepregnancy obesity in the US was estimated.

Design and Methods: Adverse fetal and infant outcomes for which statistically significant associations with prepregnancy obesity had been reported by peer-reviewed meta-analyses, which included fetal deaths and nine different major birth defects, were assessed. The true prevalence of prepregnancy obesity was estimated by multiplying self-reported prepregnancy obesity by a bias factor based on the difference between measured and self-reported obesity in US adult women. A Monte Carlo simulation approach was used to model the attributable fraction and preventable number, accounting for uncertainty in the estimates for: [1] strength of the association with obesity, [2] obesity prevalence, and [3] outcome prevalence.

Results: Eliminating the impact of prepregnancy obesity would potentially prevent the highest numbers of four outcomes: fetal deaths (6,990; uncertainty interval [UI] 4,110-10,080), congenital heart defects (2,850; UI 1,035-5,065), hydrocephalus (490; UI 150-850), and spina bifida (405; UI 305-505). If 10% of women with prepregnancy obesity achieved a healthy weight before pregnancy or otherwise mitigated the impact of obesity, nearly 300 congenital heart defects and 700 fetal deaths per year could potentially be prevented.

Conclusion: This simulation suggests that effective prevention strategies to reduce prepregnancy obesity or the risk associated with obesity could have a measurable impact on infant health in the US.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Epidemiological studies have identified prepregnancy obesity as an important risk factor contributing to adverse fetal and infant outcomes including major birth defects and fetal deaths [1-3] Data are less clear for preterm birth, with systematic reviews suggesting an association between prepregnancy obesity and some types of preterm birth but not others [4]. Prepregnancy obesity is also associated with several adverse maternal outcomes including gestational diabetes, cesarean delivery, and preeclampsia and with increased health care utilization during pregnancy and at delivery [5-8].

About one-third of reproductive-aged women in the U.S. are obese, emphasizing the public health significance of adverse reproductive outcomes associated with this maternal risk factor [9]. Of great concern, self-reported prepregnancy obesity increased from 13% in 1993-1994 to 22% in 2002-2003, based on data from the pregnancy risk assessment monitoring system (PRAMS) [10], and the prevalence of measured obesity among women aged 20-39 in the US based on data from the National Health and Nutrition Examination Survey (NHANES) increased from 28% in 1999-2000 to 34% in 2007-2008 [9].

Adverse fetal and infant outcomes such as major birth defects can result in both lifelong disability and premature mortality. In the United States each year, ∼ 130,000 births are affected by a major, structural birth defect [11], and birth defects are the underlying cause for about 1 in 5 deaths in the first year of life [12]. In addition, the number of fetal deaths each year, 26,000, approaches the number of infant deaths—yet much less public health attention is focused on this serious reproductive outcome [12, 13]. The occurrence of neural tube defects has been significantly reduced by the implementation of folic acid fortification in the US [14]; however, much work remains to be done in decreasing the occurrence of neural tube defects not prevented by folic acid fortification and preventing other adverse fetal and infant outcomes by intervening on other modifiable risk factors such as maternal obesity. Estimating the number of adverse fetal and infant outcomes associated with prepregnancy obesity could help prioritize efforts to reduce or mitigate the impact of obesity before and during pregnancy.

Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Adverse fetal and infant outcomes

The main criterion for inclusion of fetal and infant outcomes was the availability of a statistically significant association based on a peer-reviewed meta-analysis or systematic review. Ten fetal/infant outcomes met this criterion (Table 1). A recent systematic review and meta-analysis found statistically significant associations between prepregnancy obesity and a number of specific major birth defects: spina bifida, anencephaly, congenital heart defects, cleft palate, cleft lip with cleft palate, anorectal atresia, hydrocephalus, and upper and lower limb deficiencies [2]. A recent meta-analysis also identified a statistically significant association between prepregnancy obesity and fetal deaths [1]. We used the odds ratios (ORs) and 95% confidence intervals (CI) from the respective meta-analyses in estimating the association with prepregnancy obesity and to incorporate appropriate levels of uncertainty in our modeling. Use of ORs to estimate the attributable fraction is appropriate for all 10 outcomes assessed because all are rare events; the most common outcome assessed was congenital heart defects which occur in ∼ 0.8% of births [15]. Most of the studies included in the meta-analyses used normal weight women as the referent category for estimating the effect of prepregnancy obesity, with some small variations in definition for normal weight; however, a few studies used a reference category that included overweight and normal weight women combined.

Table 1. Input data on strength of association between outcomes and prepregnancy obesity, the prevalence of each adverse outcome, and the estimated annual number of each adverse infant outcome in the United States
Specific outcomeOdds ratio (OR)a95% Confidence interval for OR (used for upper and lower estimates of uncertaintyEstimated prevalence in the US (per 10,000 births)95% CI for prevalenceEstimated annual number of cases, based on 2004–2006 birth numbers95% CI for annual number of cases (used for uncertainty)
  1. a

    Pooled estimates for odds ratios are from the published meta-analyses and are based on crude odds ratios from contributing studies. For most studies included in the meta-analyses, normal weight women comprised the referent category; however, a few studies also included overweight women in the referent.

  2. b

    For hydrocephalus, pooled prevalence estimate developed with data from Arkansas, California, Georgia, Iowa, Massachusetts, North Carolina, New York, Texas, and Utah.

  3. c

    Assumes a triangular distribution ranging from the number of reported fetal deaths on fetal death certificates to 25% above this number.

Fetal deaths (1,13)2.071.59–2.7462.2Triangular distribution28,08525,982–31,386c
Congenital heart defects (2,15)1.301.12–1.5181.460.2–105.733,96025,115–44,098
Hydrocephalus (2,16)1.681.19–2.366.73b6.50–6.952,8062,712–2,900
Spina bifida (2,17)2.241.86–2.693.503.31–3.681,4601,383–1,537
Cleft lip with or without cleft palate (2,17)1.201.03–1.4010.6310.32–10.954,4374,304–4,570
Anorectal atresia (2,17)1.481.12–1.974.684.45–4.911,9521,855–2,048
Cleft palate (2,17)1.231.03–1.476.356.11–6.602,6512,549–2,754
Lower limb deficiencies (2,17)1.341.03–1.733.493.30–3.6714541,378–1,530
Anencephaly (2,17)1.391.03–1.872.061.92–2.20859800–918
Upper limb deficiencies (2,17)1.341.03–1.731.681.56–1.81701649–754

Prevalence of selected outcomes and estimated number affected each year in the US

The national prevalence and estimated annual number of US cases were available for seven of the nine selected birth defects (spina bifida, anencephaly, cleft palate, cleft lip with or without palate, anorectal atresia, and upper and lower limb deficiencies) from a recent publication of the National Birth Defects Prevention Network, pooling data for 2004 through 2006 from 14 state-based programs [17]. We used the standard errors from the published 95% CIs for these national estimates as a measure of the uncertainty. For congenital heart defects, the annual number of US cases was estimated based on an evaluation from the Centers for Disease Control and Prevention's (CDC) Metropolitan Atlanta Congenital Defects Program [15]. This evaluation used an updated review and classification of all congenital heart defects from 1998-2005 to develop an improved estimate of the prevalence. For hydrocephalus, we pooled data from 9 state-based birth defects programs (Arkansas, California, Georgia, Iowa, Massachusetts, New York, North Carolina, Texas, and Utah) for 2002 through 2006 and created a national prevalence estimate using the same methods as Parker et al. to adjust the prevalence to the racial/ethnic distribution of US live births [16, 17]. For CHDs and hydrocephalus, we used the standard errors from the 95% CIs for our prevalence estimates calculated as described above as a measure of the uncertainty. To obtain the estimated number of affected births per year for each of the included major birth defects, we multiplied the prevalence estimates by the average number of live births in the US in the period from 2004-2006 to match the time period of the national prevalence estimates [12, 17]. For fetal deaths, we used the number of fetal deaths registered by the US vital statistics system for 2005 as the point estimate [13]. Because we assumed very few of these reported cases were not true fetal deaths, but that there was significant under-ascertainment of fetal deaths, particularly those at the earlier gestational ages, we assumed a triangular distribution to reflect the uncertainty in the published estimate of fetal deaths. This triangular distribution ranged from the number of fetal deaths reported on fetal death certificates in 2005 (25,894) to 25% above that reported number.

Prevalence of prepregnancy obesity

Obesity is defined as having a body mass index (BMI) (weight in kilograms divided by height in meters squared) that is 30 or greater. Measured data on prepregnancy height and weight are not available for large population-based studies of women who have had either unaffected pregnancies or pregnancies affected by one of these adverse reproductive outcomes. While there might be some differences anticipated based on the association between BMI and impaired fertility [18], in general pregnant women would be expected to have similar levels of prepregnancy obesity as that measured in the National Health and Nutrition Examination Survey (NHANES) for all reproductive-aged women. However, studies of pregnant women have primarily relied on self-reported height and weight with a resulting misclassification of some obese women in other categories of BMI [10, 19]. For example, in a study of prepregnancy obesity and several types of major birth defects that included data from eight US study sites and had a racial/ethnic distribution similar to the US population, 28% of mothers of infants with spina bifida, 17% of all case-mothers, and 14.1% of control-mothers (mothers of infants without birth defects) were classified as having prepregnancy obesity based on their self-reported height and weight [3]. The prevalence estimates of prepregnancy obesity based on self-report in the birth defects study were all far below the 34% of reproductive-aged women who are classified as obese based on measured height and weight [3, 9].

Thus, the two primary challenges in estimating the percent of women who are obese at the start of pregnancy are: [1] population-based estimates for measured BMI (which are more accurate) are only available for a random sample of all reproductive-age women, not case-mothers (mothers of infants with birth defects or delivering fetal deaths) or those delivering live-born infants, and [2] population-based estimates of self-reported prepregnancy BMI are available, but underestimate the true prevalence of obesity [20]. For 1999-2000, NHANES data for measured height and weight classified 33.2% of women 20 years and older as obese, whereas self-reported data from the Behavioral Risk Factor Surveillance System (BRFSS) for the same years and women in the same age range classified only 20.1% as obese [20]. We estimated the magnitude and uncertainty associated with the bias factor using a Monte Carlo approach in which we sampled from the assumed uncertainty distribution for true obesity prevalence among women aged 20 and older from NHANES and divided that value by a sample from the assumed uncertainty distribution of self-reported obesity from BRFSS (Figure 1). To do this, we assumed that the uncertainty in both the NHANES and BRFSS prevalence estimates could be described using Normal distributions with means and variances matching the reported values. Samples were restricted, however, to those in which the sampled BRFSS value was greater than or equal to the NHANES sample to reflect our belief that the true prevalence of obesity is always greater than or equal to the self-reported obesity prevalence [21].

image

Figure 1. Information sources used in a Monte Carlo simulation to first estimate a bias factor for self-reported obesity and then estimate the true prevalence of prepregnancy obesity in the United States. The simulation was repeated with 10,000 iterations. National Health and Nutrition Examination Survey (NHANES), Behavioral Risk Factor Surveillance System (BRFSS), Pregnancy Risk Assessment Monitoring System (PRAMS)

Download figure to PowerPoint

The Pregnancy Risk Assessment Monitoring System (PRAMS) is a survey of pregnancy risk factors that selects a sample of women who have recently delivered live infants, with oversampling of some higher risk groups. PRAMS data from 26 states showed a self-reported prepregnancy obesity prevalence of 18.7%, standard error = 0.2%, among women who had delivered live infants in 2004-2005 [19]. To estimate the true prevalence of prepregnancy obesity in the US, we assumed that the true prevalence of obesity equaled self-reported obesity (from PRAMS) multiplied by the bias factor estimated from the NHANES-BRFSS comparison (Figure 1). We assumed that the bias factor in reporting height and weight to determine obesity is the same for women reporting their prepregnancy weight as it is for all women aged 20 years or older. Prepregnancy obesity among women who would indeed become pregnant was estimated by multiplying a sampled value from the uncertainty distribution for the bias factor by a sampled value for the self-reported prepregnancy obesity prevalence where we assumed that uncertainty in the later value could be described using a Normal distribution with a mean of 18.7% and a standard error of 0.2% (Figure 1) [19].

Attributable fraction and estimated preventable number with complete elimination of prepregnancy obesity

We estimated the attributable fraction (AF) in the population related to prepregnancy obesity for selected adverse fetal and infant outcome. We used standard methods for the calculation of the attributable fraction, where the AF is a function of the prevalence of prepregnancy obesity (P[D]) and the OR estimating the strength of the association [22].

  • display math

The AF was then multiplied by the estimated annual number of births/pregnancies affected in the US for each outcome to estimate the preventable number for each outcome associated with prepregnancy obesity, assuming complete elimination of prepregnancy obesity if all women were able to reduce their BMI to the recommended healthy weight (typically defined as a BMI of 18.5-24.9) before pregnancy. Because complete elimination of prepregnancy obesity in the US from its current level reflects the best possible scenario, we also estimated a more achievable shorter term impact of a 10% reduction in the risk associated with prepregnancy obesity on these 10 outcomes. Preventable numbers were not summed across categories because infants could be in more than one category (e.g., a fetal death that was also affected by anencephaly).

We used a Monte Carlo simulation approach to model the AF and preventable number of outcomes, while incorporating appropriate levels of uncertainty into each model input parameter. We incorporated uncertainty into the model to account for both sampling error and uncertainty due to lack of knowledge about specific parameters. We used 10,000 iterations in the simulation which resulted in 10,000 possible values each for the AF and the preventable number. We summarized these estimates using the mean and created a 95% uncertainty interval (UI) defined as the 2.5th and 97.5th percentile of the distribution of the 10,000 possible values. All estimates of preventable numbers were rounded to the nearest multiple of 5 to avoid over-stating the precision of these estimates. All simulations were conducted using SAS 9.2 software (SAS Institute, Cary, NC).

The formula we have used to estimate the AF is unbiased when the measure of association, in our case the OR, is not adjusted for potential confounders. We used summary OR estimates from published meta-analyses that are based on unadjusted effect estimates (crude ORs). The AF estimates based on Equation [1] might be biased if there is confounding of the association between exposure (prepregnancy obesity) and the outcome (10 adverse fetal/infant outcomes assessed) [22]. However, there is little evidence of confounding of the association between prepregnancy obesity and these adverse fetal/infant outcomes from the input studies used in the meta-analyses. The primary meta-analyses used crude ORs, but sensitivity analyses were conducted for neural tube defects and congenital heart defects using adjusted ORs to assess the potential role of confounding [2]. For neural tube defects, the summary OR using crude ORs (1.87, 95% CI 1.62-2.15) was identical to the summary OR using adjusted ORs (1.87, 95% CI 1.47-2.37) but the confidence interval was slightly wider. For congenital heart defects, the summary OR using crude ORs (1.30, 95% CI 1.12-1.51) was very similar to the summary OR using adjusted ORs (1.27, 95% CI 1.11-1.46) as was the confidence interval. In Waller et al., 2007 study which was the largest study contributing to the Stothard et al. meta-analysis, adjusted estimates (adjusted for maternal age, ethnicity, education, parity, periconceptional smoking, and periconceptional folic acid use) were also quite similar to crude estimates. For example, the crude OR for spina bifida was 2.25 and the adjusted OR was 2.10; the crude OR for congenital heart defects was 1.43 and the adjusted OR was 1.40 [3]. Similarly for fetal deaths, the primary meta-analysis pooled estimate was based on crude ORs (summary OR = 2.07, 95% CI 1.59-2.74), and was very similar to a pooled estimate based on adjusted ORs (summary OR = 2.08, 95% CI 1.30-3.63), suggesting a minimal effect of confounding [1].

An alternative estimator for the AF does not have this bias limitation but requires estimation of the prevalence of obesity among mothers of children with the outcome of interest [22]. To assess the robustness of the initial AF estimates, we calculated an alternative estimator of the AF for two outcomes, spina bifida and congenital heart defects, using the estimator that uses the prevalence of the exposure among case-infants and adjusted effect estimates. Independent estimates of the prevalence of obesity among case mothers were available in the published literature from the National Birth Defects Prevention Study for two of the defects we considered, spina bifida and congenital heart defects [3]. For these outcomes, we calculated alternative estimates of AF and number of prevented cases, and compared these to the estimates based on the meta-analyses.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The input data from meta-analyses of adverse fetal and infant outcomes associated with prepregnancy obesity included ORs ranging from 1.20 (for cleft lip with or without cleft palate) to 2.24 (for spina bifida). The strength of association for fetal deaths (OR = 2.07) was similar to that for spina bifida (Table 1). The estimated number of affected infants per year ranged from 701 (for upper limb deficiencies) to 33,960 (for congenital heart defects). After applying the bias factor to the sampling distribution of prepregnancy obesity, the estimated mean prevalence of prepregnancy obesity in the US was estimated as 30.9%.

The AF for prepregnancy obesity in the population ranged from 0.06 (95% Uncertainty Interval [UI] 0.01-0.11) for cleft lip with or without cleft palate to 0.28 (95% UI 0.21-0.34) for spina bifida (Table 2). Reducing the risk for obese women to the risk found for women in the normal weight range could potentially prevent congenital heart defects in 2,850 infants (95% UI 1,035-5,065); hydrocephalus in 490 infants (95% UI 150-850); spina bifida in 405 infants (95% UI 305-505), as well as smaller numbers of orofacial clefts, anencephaly, anorectal atresia, and limb deficiencies each year in the US. And, if all risk associated with prepregnancy obesity were eliminated it could potentially prevent 6,990 fetal deaths (95% UI 4,110-10,080) each year in the US. Even a more modest reduction in prepregnancy obesity (10%) could potentially prevent 700 fetal deaths and result in 285 infants born without congenital heart defects and 40 infants born without spina bifida each year in the US.

Table 2. Attributable fraction and preventable number of adverse fetal and infant outcomes associated with prepregnancy obesity annually in the United States
Specific outcomeAttributable fraction (95% uncertainty interval)Preventable number (95% uncertainty interval)a
100% elimination of the risk associated with prepregnancy obesity10% reduction in the risk associated with prepregnancy obesity
  1. a

    Rounded to the nearest 5.

Fetal deaths0.25 (0.15–0.35)6,990 (4,110–10,080)700 (410–1,000)
Congenital heart defects0.08 (0.03–0.14)2,850 (1,035–5,065)285 (105–510)
Hydrocephalus0.17 (0.05–0.30)490 (150–850)50 (15–85)
Spina bifida0.28 (0.21–0.34)405 (305–505)40 (30–50)
Cleft lip with or without cleft palate0.06 (0.01–0.11)260 (35–500)25 (5–50)
Anorectal atresia0.13 (0.03–0.24)255 (60–465)25 (5–45)
Cleft palate0.07 (0.01–0.13)180 (30–340)20 (5–35)
Lower limb deficiencies0.10 (0.01–0.19)140 (15–270)15 (0–25)
Anencephaly0.11 (0.01–0.21)95 (10–180)10 (0–20)
Upper limb deficiencies0.10 (0.01–0.19)65 (10–130)5 (0–15)
Table 3. Alternative estimates of attributable fraction (AF) and the preventable number of casesa for congenital heart defects and spina bifida based on defect specific estimates of the prevalence of obesity among case mothers and adjusted effect estimates from the National Birth Defects Prevention Study using an alternative AF estimator (3,22)
Specific birth defectAdjusted Odds ratio (95% confidence interval)Self-reported obesity prevalence among case-mothersEstimated obesity among case-mothers (95% uncertainty interval) after accounting for bias in self-reported BMIAttributable fraction (95% uncertainty interval)Preventable number (95% uncertainty interval)
  1. Rounded to the nearest 5.

Congenital heart defects1.40 (1.24–1.59)19.0%31.4% (27.5–35.4%)0.09 (0.06–0.12)2,990 (1,830–4,385)
Spina bifida2.10 (1.63–2.71)27.5%45.4% (37.0–54.4%)0.24 (0.17–0.31)345 (240–450)

Alternative estimates of AF and the preventable number of cases for spina bifida and congenital heart defects based on estimates of the prevalence of obesity among case mothers and an alternative AF estimator are presented in Table 3. For spina bifida, the alternative AF estimator yielded an AF (24%) and preventable number (345 cases) that were slightly lower than the AF (28%) and preventable number (405) based on the meta-analyses pooled estimates. For congenital heart defects, the alternative AF estimator yielded a very similar estimate of the AF (9% vs. 8%) and a slightly higher estimate of the preventable number (2,990 vs. 2,850) than the estimate based on the meta-analyses.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Prepregnancy obesity has a major impact on fetal and infant health in the US. Interventions that successfully reduce the risk of major birth defects and fetal deaths associated with prepregnancy obesity could significantly improve the health of infants. For example, assuming that prepregnancy obesity is a sufficient cause for these outcomes, nearly 3,000 infants would be born without congenital heart defects if the risk associated with prepregnancy obesity could be completely eliminated. Further, about 7,000 fetal deaths per year could be prevented if the risk associated with prepregnancy obesity could be eliminated either by women reducing their BMI to the normal weight range before pregnancy or possibly through other interventions (e.g., physical activity or diet quality) that might reduce the risk of adverse fetal and infant outcomes among obese women. Even a more modest 10% reduction in prepregnancy obesity or the risk associated with prepregnancy obesity could have a measurable impact on fetal and infant health by preventing nearly 700 fetal deaths and resulting in almost 300 fewer infants born with congenital heart defects each year in the US.

While reducing BMI in reproductive-aged women to the normal weight category before pregnancy or otherwise reducing the risk associated with prepregnancy obesity is a complex and difficult task, effective interventions on this front could compare very favorably with the magnitude of the impact for other successful public health interventions. For example, folic acid fortification of enriched cereal grains in the United States has prevented neural tube defects (spina bifida and anencephaly) in ∼ 1,000 infants per year, a tremendous public health success and one that provides significant savings in terms of both direct medical costs and other costs [14, 23]. And, perinatal transmission of human immunodeficiency virus in the US was reduced by maternal antiretroviral treatment from ∼ 1,700 pediatric patients with acquired immunodeficiency syndrome in 1992 to less than 50 affected in 2005 [24].

In addition to the fetal and infant outcomes considered in this analysis, prepregnancy obesity is also strongly associated with a number of adverse maternal outcomes including pregestational and gestational diabetes, preeclampsia, some types of preterm birth, and cesarean delivery [4-7, 25, 26]. Adverse maternal outcomes associated with prepregnancy obesity also pose a significant health burden. Gestational diabetes affects about 4% of births in the US each year, with considerable variation by race/ethnicity [27]. It has been estimated that half of gestational diabetes is associated with being overweight or obese [28]. Preeclampsia is estimated to affect about 3% of pregnancies annually [29], and remains a leading cause of maternal mortality in the US. Both diabetes and preeclampsia pose immediate and future health risks to both the mother and the fetus/infant. Cesarean delivery has increased 53% from 21% of deliveries in 1996 to 32% of deliveries in the US in 2007 [30]. There is a strong association between maternal obesity and cesarean delivery, and the increasing trends in obesity have likely contributed in part to the large increase in cesarean deliveries [6].

To reduce the risk of adverse fetal and infant outcomes associated with prepregnancy obesity, ideally, all reproductive-aged women would achieve a healthy weight before pregnancy. Other interventions might also reduce some of the risk associated with prepregnancy obesity possibly by improving metabolic status. Some data have suggested that physical activity reduces the risk of neural tube defects when compared to women that are physically inactive during pregnancy [31]. Another study specifically assessed swimming during pregnancy and reported that it modestly reduced the risk of birth defects compared to a group reporting no exercise [32]. However, these studies have not reported results stratified by maternal BMI.

Improvements in diet quality might also reduce the risk of adverse fetal and infant outcomes associated with prepregnancy obesity. Dieting to reduce weight is not recommended for women of any BMI during pregnancy because of concerns about the potential for negative impact on the fetus, and dieting during the first trimester has been associated with an increased risk of neural tube defects [33, 34]. Institute of Medicine guidelines suggest that obese women gain 11-20 pounds during pregnancy, which is less than women with lower BMI. Higher diet quality has been associated with a reduced risk of neural tube defects and orofacial clefts [35]. Furthermore, poor diet quality during pregnancy is more frequently reported by women with pre-pregnancy obesity compared to women who were not obese prior to pregnancy [36]. High levels of dietary glycemic intake have also been associated with an increased risk of neural tube defects [37].

Prepregnancy diabetes is strongly associated with birth defects. This risk increases with increasing BMI and can be mitigated by good glycemic control both early and throughout pregnancy [38]. Some women with prepregnancy obesity are likely to have undiagnosed diabetes or prediabetes, given that about 28% of diabetes is undiagnosed in the US [39]. Thus, it is possible that undiagnosed diabetes is responsible for some of the risk observed in association with maternal obesity [39]. Obese adults account for more than 50% of the estimated 5.2 million adults in the US with undiagnosed diabetes and improving the metabolic status of this population could significantly improve health [39].

This analysis used prevalence data that were based primarily on national estimates and incorporated various levels of uncertainty into the model, providing the most rigorous available estimate of the number of affected individuals per year in the US. However, as with any simulation study, there are important limitations to consider. This method assumes that there is a causal relationship between prepregnancy obesity and each outcome, but the associations observed between these variables might not be causal. Also, the causes of these adverse outcomes are believed to be multifactorial, but this approach assumes that one exposure (prepregnancy obesity) is sufficient for the outcome. It is also important to recognize that some individuals might be affected by more than one of these outcomes (e.g., a fetal death in a fetus with anencephaly); thus, it is not appropriate to sum the preventable numbers across the outcomes. The model also assumes that decreasing prepregnancy BMI to a level categorized as normal weight would decrease the risk of these adverse fetal and infant outcomes. However, it is possible that some genetic or nongenetic risk factors are associated with both prepregnancy BMI and with adverse reproductive outcomes. Limited data available on bariatric surgery suggest that reductions in prepregnancy BMI improve some pregnancy outcomes, but these assessments have not had sufficient power to assess less frequent outcomes including birth defects [40]. Furthermore, the lack of measured prepregnancy BMI data on a population-based sample of women who will become pregnant is an important limitation, and causes us to rely on self-reported estimates and adjust these estimates for reporting bias. We applied the same bias factor to all women, but the bias in reporting likely varies by the specific BMI and age of the women; however, this should have had little impact on the analysis since we were only estimating the percent that were obese at the time of conception rather than estimating the number of women in each BMI category. Finally, the comparison using the alternative AF estimator suggests these estimates from the main analyses might have small amounts of bias, but that the magnitude of the bias is likely to be quite small.

We limited this modeling to adverse fetal and infant outcomes with peer-reviewed meta-analyses showing statistically significant associations with prepregnancy obesity, and did not assess other outcomes for which individual studies suggested an association with prepregnancy obesity. Thus, additional adverse fetal and infant outcomes associated with this exposure likely could be prevented with effective interventions but are not included in our estimates. These meta-analyses were based on extensive literature reviews that included both US and non-US studies. For example, the fetal deaths meta-analysis included one US study and eight studies from outside the US. While the US study findings were consistent with the other findings, it is important to note that the population on which these pooled estimates for the meta-analysis is based (mostly outside the US) is very different from the target population (the US) to which we are applying these estimates. Both meta-analyses had clearly defined inclusion and exclusion criteria to limit the included studies. In addition, the meta-analysis of major birth defects included a sensitivity analyses for the two groups of birth defects with the most published studies examining the relation between prepregnancy obesity and the defect: neural tube defects and congenital heart defects [2]. When limited to the higher quality studies, the effect estimates for prepregnancy obesity were slightly higher, but with very similar confidence intervals to the main summary estimates. The meta-analysis of stillbirths (fetal deaths) included a sensitivity analysis that was limited to those studies that provided adjusted effect estimates, and found a summary odds ratio that was nearly identical to the effect estimate for all studies using unadjusted estimates [1].

The modeling in this article was limited to the impact of prepregancy obesity and thus does not fully capture the impact of changes in BMI. We did not attempt to estimate the preventable numbers associated with prepregnancy BMI in the overweight category, which has also been suggested to increase the risk of some adverse fetal and infant outcomes but for which the data are less clear [1, 2]. Because of the limited data available on the association with adverse fetal outcomes, we did not separately estimate the impact of severe obesity or assess the impact of BMI as a continuous variable.

Findings from this simulation suggest that effective prevention strategies to reduce prepregnancy obesity or the risk associated with obesity could have a measurable impact on infant health. These findings also highlight the need to prioritize obesity treatment as well as prevention and control interventions among reproductive-aged women. Strategies will need to focus on having women achieve a healthy weight before pregnancy or improve their metabolic status before or during pregnancy with other interventions such as increased physical activity, because dieting to achieve weight loss during pregnancy is not recommended and might pose additional risks to the fetus. Achieving a healthy weight or improved metabolic status before pregnancy will also ensure that risk is lowered in the first few weeks of gestation when many structural birth defects occur.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

All authors completed this work in their roles as employees of the Centers for Disease Control and Prevention (CDC).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • 1
    Chu SY, Kim SY, Lau J, et al. Maternal obesity and risk of stillbirth: a metaanalysis. Am J Obstet Gynecol 2007;197:223228.
  • 2
    Stothard KJ, Tennant PW, Bell R, et al. Maternal overweight and obesity and the risk of congenital anomalies: a systematic review and meta-analysis. JAMA 2009;301:636650.
  • 3
    Waller DK, Shaw GM, Rasmussen SA, et al. Prepregnancy obesity as a risk factor for structural birth defects. Arch Pediatr Adolesc Med 2007;161:74550.
  • 4
    McDonald SD, Han Z, Mulla S, et al. Overweight and obesity in mothers and risk of preterm birth and low birth weight infants: systematic review and meta-analyses. BMJ 2010;341:c3428.
  • 5
    Chu SY, Callaghan WM, Kim SY, et al. Maternal obesity and risk of gestational diabetes mellitus. Diabetes Care 2007;30:20702076.
  • 6
    Chu SY, Kim SY, Schmid CH, et al. Maternal obesity and risk of cesarean delivery: a meta-analysis. Obes Rev 2007;8:385394.
  • 7
    O'Brien TE, Ray JG, Chan WS. Maternal body mass index and the risk of preeclampsia: a systematic overview. Epidemiology 2003;14:368374.
  • 8
    Chu SY, Bachman DJ, Callaghan WM, et al. Association between obesity during pregnancy and increased use of health care. N Engl J Med 2008;358:14441453.
  • 9
    Flegal KM, Carroll MD, Ogden CL, et al. Prevalence and trends in obesity among US adults, 1999-2008. JAMA 2010;303:235241.
  • 10
    Kim SY, Dietz PM, England L, et al. Trends in pre-pregnancy obesity in nine states, 1993-2003. Obesity (Silver Spring) 2007;15:986993.
  • 11
    Centers for Disease Control and Prevention (CDC).Update on overall prevalence of major birth defects—Atlanta, Georgia, 1978-2005. MMWR Morb Mortal Wkly Rep 2008;57:15.
  • 12
    Mathews TJ, MacDorman MF. Infant mortality statistics from the 2006 period linked birth/infant death data set. Natl Vital Stat Rep 2010;58:131.
  • 13
    MacDorman MF, Kirmeyer S. Fetal and perinatal mortality, United States, 2005. Natl Vital Stat Rep 2009;57:119.
  • 14
    Centers for Disease Control and Prevention (CDC).Spina bifida and anencephaly before and after folic acid mandate—United States, 1995-1996 and 1999-2000. MMWR Morb Mortal Wkly Rep 2004;53:362365.
  • 15
    Reller MD, Strickland MJ, Riehle-Colarusso T, et al. Prevalence of congenital heart defects in metropolitan Atlanta, 1998-2005. J Pediatr 2008;153:807813.
  • 16
    Parker SE, Mai CT, Canfield MA, et al. Updated national birth prevalence estimates for selected birth defects in the United States, 2004-2006. Birth Defects Res A Clin Mol Teratol 2010;88:10081016.
  • 17
    NBDPN. 2009.Congenital malformation surveillance report: a report from the national birth defects prevention network PART II. Birth Defects Res A Clin Mol Teratol 2009;85:9391004.
  • 18
    Metwally M, Li TC, Ledger WL. The impact of obesity on female reproductive function. Obes Rev 2007;8:515523.
  • 19
    Chu SY, Kim SY, Bish CL. Prepregnancy obesity prevalence in the United States, 2004-2005. Matern Child Health J 2009;13:614620.
  • 20
    Yun S, Zhu BP, Black W, et al. A comparison of national estimates of obesity prevalence from the behavioral risk factor surveillance system and the National Health and Nutrition Examination survey. Int J Obes (Lond) 2006;30:164170.
  • 21
    Gorber SC, Tremblay M, Moher D, et al. A comparison of direct vs. self-report measures for assessing height, weight and body mass index: a systematic review. Obes Rev 2007;8:307326.
  • 22
    Rockhill B, Newman B, Weinberg C. Use and misuse of population attributable fractions. Am J Public Health 1998;88:1519.
  • 23
    Yi Y, Lindemann M, Colligs A, et al. Economic burden of neural tube defects and impact of prevention with folic acid: a literature review. Eur J Pediatr 2011;170:13911400.
  • 24
    Fowler MG, Gable AR, Lampe MA, et al. Perinatal HIV and its prevention: progress toward an HIV-free generation. Clin Perinatol 2010;37:699719.
  • 25
    Torloni MR, Betran AP, Daher S, et al. Maternal BMI and preterm birth: a systematic review of the literature with meta-analysis. J Matern Fetal Neonatal Med 2009;22:957970.
  • 26
    Torloni MR, Betran AP, Horta BL, et al. Prepregnancy BMI and the risk of gestational diabetes: a systematic review of the literature with meta-analysis. Obes Rev 2009;10:194203.
  • 27
    Martin JA, Menacker F. Expanded health data from the new birth certificate, 2004. Natl Vital Stat Rep 2007;55:122.
  • 28
    Kim SY, England L, Wilson HG, et al. Percentage of gestational diabetes mellitus attributable to overweight and obesity. Am J Public Health 2010;100:10471052.
  • 29
    Berg CJ, Mackay AP, Qin C, et al. Overview of maternal morbidity during hospitalization for labor and delivery in the United States: 1993-1997 and 2001-2005. Obstet Gynecol 2009;113:10751081.
  • 30
    Menacker F, Hamilton BE. Recent trends in cesarean delivery in the United States. NCHS Data Brief 2010;35:18.
  • 31
    Carmichael SL, Shaw GM, Neri E, et al. Physical activity and risk of neural tube defects. Matern Child Health J 2002;6:151157.
  • 32
    Juhl M, Kogevinas M, Andersen PK, et al. Is swimming during pregnancy a safe exercise? Epidemiology 2010;21:253258.
  • 33
    Dewey KG, McCrory MA. Effects of dieting and physical activity on pregnancy and lactation. Am J Clin Nutr 1994;59(2 Suppl):446S452S; discussion 452S-453S.
  • 34
    Carmichael SL, Shaw GM, Schaffer DM, et al. Dieting behaviors and risk of neural tube defects. Am J Epidemiol 2003;158:11271131.
  • 35
    Carmichael SL, Yang W, Feldkamp ML, et al. Reduced risks of neural tube defects and orofacial clefts with higher diet quality. Arch Pediatr Adolesc Med 2012;166:121126.
  • 36
    Rifas-Shiman SL, Rich-Edwards JW, Kleinman KP, et al. Dietary quality during pregnancy varies by maternal characteristics in Project Viva: a US cohort. J Am Diet Assoc 2009;109:10041011.
  • 37
    Yazdy MM, Liu S, Mitchell AA, et al. Maternal dietary glycemic intake and the risk of neural tube defects. Am J Epidemiol 2010;171:407414.
  • 38
    Correa A, Gilboa SM, Besser LM, et al. Diabetes mellitus and birth defects. Am J Obstet Gynecol 2008;199:237 e19.
  • 39
    Wee CC, Hamel MB, Huang A, et al. Obesity and undiagnosed diabetes in the U.S. Diabetes Care 2008;31:18131815.
  • 40
    Guelinckx I, Devlieger R, Vansant G. Reproductive outcome after bariatric surgery: a critical review. Hum Reprod Update 2009;15:189201.