Abstract
The mean infant birth weight in the United States increased for decades, but it might now be decreasing. Given race disparities in fetal growth, we explored race-specific trends in birth weight at Magee-Womens Hospital, Pittsburgh, Pennsylvania, from 1997 to 2011. Among singleton births delivered at 37–41 weeks (n = 70,607), we evaluated the proportions who were small for gestational age and large for gestational age and changes in mean birth weights over time. Results were stratified by maternal race/ethnicity. Since 1997, the number of infants born small for their gestational ages increased (8.7%–9.9%), whereas the number born large for their gestational ages decreased (8.9%–7.7%). After adjustment for gestational week at birth, maternal characteristics, and pregnancy conditions, birth weight decreased by 2.20 g per year (P < 0.0001). Decreases were greater for spontaneous births. Reductions were significantly greater in infants born to African-American women than in those born to white women (−3.78 vs. −1.88 per year; P for interaction = 0.010). Quantile regression models indicated that birth weight decreased across the entire distribution, but reductions among infants born to African-American women were limited to those in the upper quartile after accounting for maternal factors. Limiting the analysis to low-risk women eliminated birth weight reductions. Birth weight has decreased in recent years, and reductions were greater in infants born to African-American women. These trends might be explained by accumulation of risk factors such as hypertension and prepregnancy obesity that disproportionately affect African-American women. Our results raise the possibility of worsening race disparities in fetal growth.
Keywords: birth weight, growth restriction, macrosomia, obesity
Editor's note:An invited commentary on this article appears on page 24, and the authors’ response appears on page 26.
Mean infant birth weight among singletons delivered at term in the United States increased from 1950 through the 1990s (1), but it might now be decreasing (2). Reasons for this trend are not understood. In fact, there have been some reports that babies are getting bigger, although these are from other countries or were limited to births that occurred before 2000 (3–6). Others have noted that increasing use of induction associated with a higher proportion of full-term infants delivered at earlier gestational ages might be the cause of decreases in birth weight (7). In a recent study, however, Morisaki et al. (8) reported a decrease in birth weight even in a cohort in which gestational length did not change. Birth weight is a sentinel marker of newborn health, and small changes in its distribution might signal improved or worsening maternal, neonatal, and childhood health. This is of particular importance for African-American women and infants given the race disparities in fetal growth. The percentage of full-term infants with a low birth weight (<2,500 g) is 2-fold higher for African-American infants than for white infants (5.4% vs. 2.6), and race disparities in infant mortality among term newborns persist (3.8 vs. 2.3 per 1,000, respectively) (9, 10).
Background trends in maternal health could be causes of increases in both fetal overgrowth and fetal undergrowth. Rates of obesity among women of reproductive age have increased (11, 12) and smoking rates have decreased (13); both would lead to a trend of higher mean birth weight. However, lower rates of smoking might contribute to an increase in the incidence of preeclampsia (14, 15) and the excess growth restriction that can accompany this hypertensive disorder of pregnancy. Empirically, heavier women have heavier babies, and obesity is associated with increased risk of gestational diabetes and macrosomia. Obesity is also associated with excess risk of hypertension in pregnancy and the concomitant risk of placental dysfunction and growth restriction. Obesity and hypertension also disproportionately affect African-American women. Possible effects of these trends on race disparities in fetal growth have not been explored. A large, validated clinical registry is needed to evaluate changes in birth weight over time in order to adequately characterize clinical characteristics such as preeclampsia and gestational hypertension that are known to be highly misclassified in vital records and to vary by race/ethnicity (16, 17). The only clinical registry study in which trends of declining birth weight over time were reported included very few African-American women (8).
We utilized a perinatal hospital registry to explore trends in birth weight among term pregnancies delivered between 37 and 41 completed weeks of gestation to African-American and white women from 1997 to 2011. We hypothesized that lower rates of maternal smoking and higher maternal adiposity would be associated with excess fetal under- and overgrowth as assessed by the number of infants who were small for their gestational ages and large for their gestational ages, respectively, with heavier tails in the birth-weight distribution. We anticipated that these changes might be more pronounced in infants born to African-American women. We further considered whether these changes would persist among term births after excluding infants whose births were induced or who were delivered via cesarean delivery, and we examined whether changes in birth weight occurred in subgroups.
METHODS
Data were collected from the Magee Obstetric Medical and Infant database. The Magee Obstetric Medical and Infant database, which was established in 1995, routinely collects comprehensive maternal, fetal, and neonatal characteristics from the electronic and medical records of all women who give birth at Magee-Womens Hospital of the University of Pittsburgh Medical Center (18). The database is surveyed periodically by an honest broker to maintain accuracy by randomized direct comparison with patient charts and also by examining frequencies for variables that contain outliers on download, which, once identified, are verified or corrected via medical chart review. Personal identifying information in the database is then eliminated to ensure confidentially. The University of Pittsburgh Institutional Review Board approved this study.
Analysis was limited to singleton deliveries at 37–41 completed weeks of gestation from 1997 through 2011, the years for which we had the most complete data (70,607 births, 56,609 women). Birth weight was recorded in grams, and gestational age was reported in completed weeks based upon clinical evaluation, which included an early gestation ultrasound in most cases. We defined small for gestational age as a birth weight less than the 10th percentile and large for gestational age as greater than the 90th percentile at each completed week of gestation, as recommended by the Global Reference standard, which uses estimated fetal weights and accounts for mean differences in birth weight by maternal race (19, 20). Year of delivery was the primary exposure. Maternal race was self-reported as white, African American, Hispanic, Asian, or other. Numbers were too small to evaluate racial/ethnic groups other than white and African American (n = 2,123; 3.1%), so women in other groups were dropped from the analysis. Covariates included maternal age at delivery, which we categorized as older than 35 years of age or 35 years of age or younger because of evidence that pregnancy rates have increased for older mothers and because the risk of complications also increases with maternal age; smoking during pregnancy, which was assessed at delivery as any smoking during pregnancy; and primiparity. Maternal hypertension status was abstracted from the medical record using codes from the International Classification of Diseases, Ninth Revision, and reported as chronic hypertension (including cases of superimposed preeclampsia: codes 642.0, 642.1, 642.2, and 642.7); preeclampsia (de novo hypertension with proteinuria: codes 642.4, 642.5, and 642.6); or gestational hypertension (de novo hypertension without proteinuria: code 642.3). Gestational diabetes was defined according to the criteria of Carpenter and Coustan (21) and was abstracted from the medical records. Beginning in 2003, the Magee Obstetric Medical and Infant Database perinatal registry began collecting data on prepregnancy body mass index (BMI; weight (kg)/height (m)2), and therefore maternal BMIs were available for 36,801 births. We defined gestational weight gain as delivery weight minus prepregnancy weight and categorized it as adequate, inadequate, or excessive according prepregnancy BMI, as recommended by the Institute of Medicine guidelines. The recommended weight gain is 12.5 kg–18 kg for underweight women, 11.5 kg–16 kg for normal-weight women, 7 kg–11.5 kg for overweight women, and 5 kg–9 kg for obese women (22).
Statistical analysis
The proportions of liveborn infants who were small and large for their gestational ages each year were evaluated, and change was calculated relative to the rate in 1997. Poisson regression models were used to estimate risk ratios and 95% confidence intervals associated with delivery year. To assess the association between delivery year and birth weight as a continuous outcome, linear mixed-effect models were used to account for the correlation of repeated measures within the set of women who contributed more than 1 birth to the analysis (n = 13,998; 19.8%). Multivariable associations were examined using models adjusted for gestational week of delivery, maternal race, age, smoking, adequacy of gestational weight gain, prepregnancy BMI, parity, hypertension (chronic, preeclampsia, and gestational), and gestational diabetes. Quantile regression analysis was conducted to evaluate the effect of delivery year across the distribution of birth weights (25th, 50th, and 75th percentiles) (23). Models were repeated after limiting analyses to women with spontaneous labor by removing cesarean deliveries and medical inductions. Effect measure modification by maternal race was examined using likelihood ratio tests. The change in birth weight by year varied by race (P = 0.010), and we therefore stratified models by race. Results were also stratified by each maternal characteristic and condition to examine trends in birth weight over time in subgroups. All analyses were conducted using SAS, version 9.3 (SAS Institute, Inc., Cary, North Carolina).
RESULTS
On average, the proportion of mothers who delivered at our institution and were older than 35 years of age decreased over time (Table 1). The proportion of African-American women increased, as did the number of births to women with at least 4 years of education beyond high school. Rates of smoking during pregnancy declined. Although the proportion of overweight and obese women increased, rates of excessive gestational weight gain declined. Rates of chronic hypertension, preeclampsia, and gestational diabetes increased, whereas cases of gestational hypertension decreased. The proportion of infants who were small for their gestational ages at birth increased over this time period (from 8.7% to 9.9%; risk ratio =1.25, 95% confidence interval: 0.59, 2.65; P = 0.55), and the proportion of infants who were large for their gestational ages decreased (from 8.9% to 7.7%; risk ratio = 0.55, 95% confidence interval: 0.31, 1.00; P = 0.05) (Table 1). These changes were similar in infants born to African-American and white women.
Table 1.
Maternal Characteristics and Mean Birth Weights for Singleton Live Births at Magee-Womens Hospital, Pittsburgh, Pennsylvania, 1997–2011
| Characteristic | Years |
||||||||
|---|---|---|---|---|---|---|---|---|---|
| 1997–2001 (n = 20,442) |
2002–2006 (n = 26,283) |
2007–2011 (n = 23,882) |
|||||||
| No. of Births | % | Birth Weight, g (Mean (SD)) | No. of Births | % | Birth Weight, g (Mean (SD)) | No. of Births | % | Birth Weight, g (Mean (SD)) | |
| Overall | 3,453 (463) | 3,429 (463) | 3,403 (453) | ||||||
| Maternal characteristics | |||||||||
| Age >35 years | 3,079 | 15.1 | 3,495 (462) | 4,473 | 17.0 | 3,487 (469) | 3,179 | 13.3 | 3,458 (448) |
| Race/ethnicity | |||||||||
| White | 16,550 | 81.0 | 3,489 (457) | 20,457 | 77.8 | 3,472 (459) | 17,946 | 75.1 | 3,446 (452) |
| African American | 2,994 | 14.6 | 3,277 (459) | 4,425 | 16.8 | 3,255 (445) | 4,071 | 17.0 | 3,233 (439) |
| Education >16 yearsa | 8,114 | 42.6 | 3,420 (472) | 9,670 | 48.4 | 3,503 (449) | 9,395 | 50.7 | 3,342 (466) |
| Married | 13,993 | 68.5 | 3,503 (451) | 17,884 | 68.0 | 3,489 (451) | 14,372 | 60.2 | 3,464 (441) |
| Current smoker | 2,637 | 12.9 | 3,240 (450) | 3,326 | 12.7 | 3,234 (454) | 2,411 | 10.1 | 3,185 (464) |
| Hypertension | |||||||||
| Gestational hypertension | 1,238 | 6.1 | 3,449 (484) | 1,447 | 5.5 | 3,426 (478) | 1,229 | 5.1 | 3,395 (465) |
| Preeclampsia | 430 | 2.1 | 3,222 (558) | 931 | 3.5 | 3,270 (520) | 960 | 4.0 | 3,262 (484) |
| Chronic hypertension | 246 | 1.2 | 3,360 (488) | 409 | 1.6 | 3,332 (488) | 398 | 1.7 | 3,323 (484) |
| Gestational diabetes | 633 | 3.1 | 3,532 (501) | 1,079 | 4.1 | 3,463 (467) | 1,014 | 4.2 | 3,448 (436) |
| Nulliparous | 6,637 | 32.5 | 3,396 (455) | 8,976 | 34.2 | 3,391 (461) | 8,644 | 36.2 | 3,367 (451) |
| Adequacy of weight gainb | |||||||||
| Inadequate | 2,826 | 15.4 | 3,250 (447) | 2,680 | 15.9 | 3,221 (434) | |||
| Adequate | 5,533 | 30.1 | 3,367 (434) | 5,376 | 32.0 | 3,354 (426) | |||
| Excessive | 9,997 | 54.5 | 3,526 (462) | 8,750 | 52.1 | 3,507 (455) | |||
| Prepregnancy BMIc | |||||||||
| Underweight | 885 | 4.6 | 3,235 (439) | 754 | 4.3 | 3,227 (446) | |||
| Normal weight | 11,193 | 58.2 | 3,407 (448) | 9,857 | 56.1 | 3,381 (441) | |||
| Overweight | 4,145 | 21.5 | 3,474 (473) | 3,862 | 22.0 | 3,442 (468) | |||
| Obese | 3,018 | 15.7 | 3,491 (497) | 3,087 | 17.6 | 3,476 (476) | |||
| Cesarean delivery | 3,610 | 17.7 | 3,548 (527) | 6,131 | 23.3 | 3,506 (509) | 6,345 | 26.6 | 3,478 (492) |
| Induced labor | 12,293 | 60.1 | 3,472 (462) | 14,465 | 55.0 | 3,438 (461) | 11,849 | 49.6 | 3,411 (451) |
| Infant characteristics | |||||||||
| Female sex | 9,871 | 48.3 | 3,385 (448) | 12,834 | 48.8 | 3,360 (445) | 11,730 | 49.1 | 3,333 (442) |
| Small for gestational age | 1,768 | 8.6 | 2,701 (277) | 2,400 | 9.1 | 2,683 (270) | 2,370 | 9.9 | 2,690 (264) |
| Large for gestational age | 1,826 | 8.9 | 4,217 (313) | 2,324 | 8.8 | 4,196 (312) | 1,837 | 7.7 | 4,172 (318) |
Abbreviations: BMI, body mass index; SD, standard deviation.
a Data were available for 19,047 births in 1997–2001, 19,979 births in 2002–2006, and 18,531 births in 2007–2011.
b Data were available for 18,356 births in 2002–2006 and 16,806 births in 2007–2011.
c Data were available for 19,241 births in 2002–2006 and 17,560 births in 2007–2011.
Mean birth weight at term decreased from 1997 to 2011 by 3.66 (standard error, 0.40) g per year on average (P < 0.0001; Table 2). Adjustment for gestational week attenuated this to a yearly reduction of 2.17 (standard error, 0.38) g (P < 0.0001), and additional adjustment for maternal age, smoking, educational level, hypertension, parity, race/ethnicity, marital status, gestational diabetes, and infant sex had minimal influence (decrease of 2.20 (standard error, 0.41) g per year; P < 0.0001). The per-year reduction was larger after we limited the analysis to spontaneous deliveries (2.99-g decrease; P < 0.0001). Data on prepregnancy BMIs were available beginning in 2003, and adjustment for this variable, along with adequacy of gestational weight gain, resulted in a 3.86-g decrease per year from 2003 through 2011 (P < 0.0001). As expected, overweight or obese status before pregnancy, higher educational level, being married, having excessive gestational weight gain, gestational diabetes, and infant male sex were all independently associated with higher birth weights. Smoking, African-American race, hypertension before or during pregnancy, inadequate gestational weight gain, and primiparity all were associated with lower birth weights. The decreasing trend in mean birth weight per year remained, however, even after we accounted for these factors.
Table 2.
Delivery Year and Maternal Factors Associated With Variationd in Birth Weight Among Term Deliveries at Magee-Womens Hospital, Pittsburgh, Pennsylvania, 1997–2011
| Characteristic | All Term Births, 1997–2011 (n = 70,607) |
Spontaneous Deliveries, 1997–2011 (n = 20,576)a |
All Term Births, 2003–2011 (n = 36,801) |
|||
|---|---|---|---|---|---|---|
| Estimated Change in Birth Weight, g | P Value | Estimated Change in Birth Weight, g | P Value | Estimated Change in Birth Weight, g | P Value | |
| Delivery yearb | −3.66 | <0.0001 | −3.20 | <0.0001 | −3.66 | <0.0001 |
| Delivery yearc | −2.17 | <0.0001 | −2.98 | <0.0001 | −2.17 | <0.0001 |
| Delivery yeard | −2.20 | <0.0001 | −2.99 | <0.0001 | −3.86 | <0.0001 |
| Gestational weeks | 140.22 | <0.0001 | 140.28 | <0.0001 | 131.47 | <0.0001 |
| Maternal age >35 years | 11.79 | 0.0218 | −3.70 | 0.6921 | 15.28 | 0.028 |
| Smoker | −183.62 | <0.0001 | −177.22 | <0.0001 | −173.78 | <0.0001 |
| Education >16 years | 24.30 | <0.0001 | 24.76 | 0.001 | 34.57 | <0.0001 |
| Hypertensione | <0.0001 | 0.1879 | <0.0001 | |||
| Chronic | −39.86 | 12.00 | −95.70 | |||
| Gestational | 10.05 | 8.64 | −33.75 | |||
| Preeclampsia | −73.38 | −55.51 | −88.70 | |||
| Primiparity | −107.25 | <0.0001 | −93.02 | <0.0001 | −103.26 | |
| Racef | <0.0001 | <0.0001 | <0.0001 | |||
| African American | −146.27 | −133.95 | −152.75 | |||
| Other | −135.68 | −146.78 | −131.95 | |||
| Married | 54.94 | <0.0001 | 53.36 | <0.0001 | 54.13 | |
| Gestational diabetes | 101.50 | <0.0001 | 47.27 | 0.0286 | 88.96 | <0.0001 |
| Male sex | 144.74 | <0.0001 | 126.92 | <0.0001 | 138.57 | <0.0001 |
| Prepregnancy BMIg,h | <0.0001 | |||||
| Overweight (24.0–29.9) | 36.10 | |||||
| Obese (≥30) | 101.38 | |||||
| Gestational weight gaini | <0.0001 | |||||
| Inadequate | −86.87 | |||||
| Excessive | 137.91 | |||||
Abbreviation: BMI, body mass index.
a Excludes cesarean deliveries and induced births.
b Unadjusted.
c Adjusted for delivery week.
d Adjusted for all factors.
e No hypertension was the reference category.
f White race was the reference category.
g Weight (kg)/height (m)2.
h Normal weight (BMI <24) was the reference category.
i Adequate weight gain was the reference group.
We then stratified the analysis by maternal race because there was evidence that the decline in birth weight varied by race (P = 0.010). Among infants born to white women, the change in birth weight was −1.88 g per year; among those born to African-American women, it was −3.78 g per year after adjustment for covariates (Table 3). Birth weight declined similarly among both groups when analyses were limited to spontaneous births (a change of −2.74 for infants born to white women vs. −3.98 for those born to African-American women; P for interaction = 0.271).
Table 3.
Delivery Year and Birth Weight, Stratified by Maternal Race/Ethnicity, Magee-Womens Hospital, Pittsburgh, Pennsylvania, 1997–2011
| Maternal Race/Ethnicity | All Term Births |
Spontaneous Term Births |
||||
|---|---|---|---|---|---|---|
| Estimated Change in Birth Weight,a g | P Value | P for Interaction | Estimated Change in Birth Weight,a g | P Value | P for Interaction | |
| White | −1.88 | <0.0001 | 0.01 | −2.74 | 0.0007 | 0.271 |
| African American | −3.78 | 0.0002 | −3.98 | 0.0181 | ||
a Adjusted for gestational week of delivery, maternal age, smoking, educational level, hypertension during pregnancy, parity, marital status, gestational diabetes, and infant sex.
When comparing the distribution of birth weight in 1997 with that in 2011, it appeared that the entire distribution shifted to the left (Figure 1A). Adjustment for maternal factors and pregnancy conditions did not explain the total temporal reduction in mean birth weight in the bottom quartile of the distribution (−37.68 g; P = 0.0.0040) or the middle (−45.03 g; P = 0.0005) or upper (−56.63 g; P < 0.0001) ranges. Removing induced births and cesarean deliveries attenuated the temporal reduction in birth weight at the lower end of the distribution but augmented the magnitude at the higher end (<25th percentile: −34.73, P = 0.0970; 25th–75th percentile: −47.27, P = 0.0125; >75th percentile: −65.67, P = 0.0085; Figure 1B). Results were similar for infants born to white women (Figure 1C). Among those born to African-American women, only birth weight reductions in the upper quartile were significant after accounting for maternal factors and pregnancy conditions (Figure 1D).
Figure 1.
Birth weight distribution in grams in 1997 (dashed line and light gray bars) and 2011 (solid line and medium gray bars) A) for all births; B) after excluding infants delivered after induction or those who underwent cesarean delivery; C) among white women; and D) among African-American women. Shown are quantile regression results for the differences in birth weight in the 25th, 50th, and 75th percentiles of the distribution, adjusted for gestational week, maternal age, smoking, educational level, hypertension, parity, race/ethnicity, marital status, gestational diabetes, and infant sex. Dark gray bars indicate overlapping distributions in 1997 and 2011.
We then stratified the analysis by maternal characteristics or conditions to determine whether reductions in birth weight were consistent across subgroups (Figure 2). Mean birth weight decreased each year in most subgroups, including smokers and nonsmokers. Mean birth weight did not change among those who were underweight or obese prior to pregnancy, but it did decline by 5.13 g and 4.84 g per year among normal-weight and overweight women, respectively. Per-year reductions in birth weight were similar according to gestational weight gain, offspring sex, and gestational diabetes status. Mean birth weight decreased each year among women with no hypertension or gestational hypertension; however, women with preeclampsia had larger babies on average over time (10.57-g increase per year; P = 0.0004). When stratified by parity, birth weights of first-born infants increased from 1997 to 2011 (3.63 g per year; P = 0.04), but birth weights of second- or higher-order births decreased (−2.59 g per year; P < 0.0001; P for interaction =0.11). Birth weight decreased at every gestational week of delivery, with the largest decrease among infants delivered at 40–41 weeks (−3.73 g per year, P < 0.0001).
Figure 2.
Average reduction in grams per year associated with maternal and delivery characteristics, 1997–2011. Results are from linear models stratified by each factor indicated and adjusted for all other factors. Body mass index (weight (kg)/height (m)2) was classified as follows: underweight, <18.5; normal weight, 18.5–23.9; overweight, 24.0–29.9; and obese, ≥30.0. DM, diabetes mellitus. Bars, 95% confidence intervals.
To further explore the reasons for these reductions, we restricted our fully adjusted model to low-risk white and African-American women who were nonsmokers, who did not have hypertension or gestational diabetes, who delivered at 37–40 weeks’ gestation, and who were primiparous. There was a modest reduction in mean birth weight among infants born to low-risk white women (−1.97 (standard error, 1.1) g) that was attenuated to no difference after further restricting this group to normal-weight women with adequate weight gain. There were no temporal reductions in birth weight among infants born to low-risk African-American women with adequate weight gain, which suggests that the accumulation of risk factors might explain the temporal changes we detected.
DISCUSSION
The results of our study suggest that mean infant birth weight has decreased modestly each year since 1997 independent of maternal characteristics and clinical conditions. Of note, these reductions appear to be larger in infants born to African-American women than in those born to white women, which raises the possibility that race disparities in fetal growth might be increasing. Several clues in our data suggest that these declines are linked to the accumulation of maternal risk factors: 1) the declines in birth weight were limited to second- and higher-order births, 2) induced births, which are more likely among high-risk mother/child dyads, accounted for the change in birth weight among births in the lowest quartile, and excluding medical inductions eliminated the race disparity in the decline in birth weight, and 3) sensitivity analysis limited to low-risk African-American and white primiparous women revealed no change in infant birth weight over time. Thus, the burden of accumulating risk factors such as obesity might converge after a first birth to impair fetal growth in subsequent pregnancies. Given that birth weight is a sentinel marker of infant health status and the indication that African-American mother/child dyads might be particularly susceptible to decreasing birth-weight trends, race disparities in maternal and child health might be on the increase. Our results are consistent with and extend those of the few contemporary, population-based studies of trends in US birth weights (2, 8). Our hospital registry allowed for a more detailed examination of clinical characteristics, such as hypertension in pregnancy, gestational diabetes, and induced deliveries, which are known to be misclassified in vital records. We were also able to account for prepregnancy BMI and pregnancy weight gain in a large subgroup. We were able to explore effect modification by race and found that reductions in birth weight were higher for infants born to African-American women than for those born to white women. Although this has not yet translated to higher rates of infants born small for their gestational age to African-American women in our hospital registry, the race-specific trends raise the possibility that race disparities in fetal growth might increasing. It is also possible that the shift in birth weight that we detected is a result of more high-risk pregnancies being managed appropriately, with delivery of smaller babies. However, this does not diminish the suggestion in these data that maternal health might be worsening and that this may lead to short- and long-term effects on fetal growth. Increasing maternal adiposity might provide enhanced glucose substrate but, concomitantly, might modestly impair placentation (24). The evidence that chronic placental inflammation is associated with increasing maternal prepregnancy BMI supports this possibility (25). In addition, modest elevations in blood pressure associated with increasing adiposity, even within the normotensive range, might be associated with fetal growth impairments (26). We were unable to directly examine these possibilities. Further studies of placental features, with serial blood pressure measurements across gestation and across serial pregnancies, are needed to disentangle the timing of growth decrements. Ultrasound studies might be useful in determining whether there have been temporal changes in first trimester growth or whether decrements occur later in gestation, during the time of rapid skeletal growth and fetal fat accretion. There is evidence that US women may be less healthy when entering pregnancy now than they were a decade ago, and our results suggest that the accumulation of maternal morbidities might account for our findings (27). Among infants born to African-American women, for example, only the decrease in birth weight in the upper quartile persisted after accounting for maternal comorbid conditions, which suggests that decrements in the lower half of the distribution might be explained by these conditions. In addition, birth weight does not accurately reflect neonatal body composition, lean mass, and subcutaneous fat. Catalano et al. (28) reported that obese women who gained less weight in pregnancy had offspring with lower birth weights and had a lower risk of having a child who was large for gestational age, which, on the surface, might appear to be a healthier outcome. However, these neonates had less lean mass relative to fat mass, which could set them up for a lifetime of metabolic disturbances. We unfortunately did not have data on newborn body composition, and thus our birth weight results might reveal only a portion of growth trends that are relevant to childhood health.
The only groups that showed increasing birth weights were first births and those following preeclampsia. The reasons are unclear, but a possibility is that obesity may be contributing to a larger proportion of preeclampsia cases over time and that the rate preeclampsia among first births may be increasing (12). The tradeoff between fetal over- and undergrowth related to preeclampsia in the setting of rising obesity rates warrants investigation (29).
Limitations of our study must be considered. Although we had access to extensive clinical diagnostic information, many relevant features of pregnancy were unavailable, including serial blood pressure measurements, placental characteristics, and biomarkers of pathways that may play a role, such as insulin resistance and inflammation. Other important characteristics, such as whether the pregnancy was intentional, were not available, and we were not able to determine whether gestational age dating via ultrasound differed by maternal race. The patient population that we studied changed over the study period, and the generalizability of our findings to the larger population is not clear. For example, the number of women who were older than 35 years of age when they gave birth decreased modestly from 1997 to 2011. This is likely due to a higher proportion of births to minority women, who have, on average, lower mean maternal ages at delivery. For example, from 2007 to 2011, 15.24% of white women in our study were older than 35 years of age at delivery compared with 5.51% of African-American women. We also were limited to the racial/ethnic profile of our region and could not explore birth weight trends in Hispanic and Asian women. Quantity of cigarettes smoked per day was not reported in our registry, and thus we could not explore the linear association of smoking with birth weight declines or determine whether results were different among women who quit smoking during pregnancy. We were also unable to examine other possible explanations for decrements in birth weight, such as nutritional and environmental exposures.
Our results indicate that, on average, birth weight has decreased each year since 1997 and that infants born to African-American women were more strongly affected than were infants born to white women. The reasons for these reductions might be explained in part by increasing accumulation of risk factors, such as hypertension and prepregnancy overweight and obesity, that are known to disproportionately affect African-American women.
ACKNOWLEDGMENTS
Author affiliations: Department of Obstetrics, Gynecology and Reproductive Sciences, School of Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania (Janet M. Catov, James M. Roberts, Hyagriv N. Simhan); Department of Obstetrics, Gynecology and Reproductive Sciences, Magee-Womens Research Institute, Pittsburgh, Pennsylvania (Janet M. Catov, James M. Roberts, Hyagriv N. Simhan); Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania (Janet M. Catov, Jia Xu); Biostatistics/Epidemiology/Research Design Core, Center for Clinical Translational Sciences, The University of Texas Health Science Center at Houston, Houston, Texas (MinJae Lee); and Division of Epidemiology, Human Genetics and Environmental Sciences, The University of Texas, Houston, Texas (MinJae Lee).
This work was supported by grant 5K12HD43441-09 from the National Institutes of Health's National Institute of Child Health and Human Development and a Building Interdisciplinary Research Careers in Women's Health Award.
Conflict of interest: none declared.
REFERENCES
- 1.Chike-Obi U, David RJ, Coutinho R et al. Birth weight has increased over a generation. Am J Epidemiol. 1996;1446:563–569. [DOI] [PubMed] [Google Scholar]
- 2.Donahue SM, Kleinman KP, Gillman MW et al. Trends in birth weight and gestational length among singleton term births in the United States: 1990–2005. Obstet Gynecol. 2010;115(2 pt 1):357–364. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ananth CV, Wen SW. Trends in fetal growth among singleton gestations in the United States and Canada, 1985 through 1998. Semin Perinatol. 2002;264:260–267. [DOI] [PubMed] [Google Scholar]
- 4.Kramer MS, Morin I, Yang H et al. Why are babies getting bigger? Temporal trends in fetal growth and its determinants. J Pediatr. 2002;1414:538–542. [DOI] [PubMed] [Google Scholar]
- 5.Lahmann PH, Wills R-A, Coory M. Trends in birth size and macrosomia in Queensland, Australia, from 1988 to 2005. Paediatr Perinat Epidemiol. 2009;236:533–541. [DOI] [PubMed] [Google Scholar]
- 6.Schack-Nielsen L, Mølgaard C, Sørensen TIA et al. Secular change in size at birth from 1973 to 2003: National Data from Denmark. Obesity (Silver Spring). 2006;147:1257–1263. [DOI] [PubMed] [Google Scholar]
- 7.Zhang X, Joseph KS, Kramer MS. Decreased term and postterm birthweight in the United States: impact of labor induction. Am J Obstet Gynecol. 2010;2032:124.e1–124.e7. [DOI] [PubMed] [Google Scholar]
- 8.Morisaki N, Esplin MS, Varner MW et al. Declines in birth weight and fetal growth independent of gestational length. Obstet Gynecol. 2013;1211:51–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Rossen LM, Schoendorf KC. Trends in racial and ethnic disparities in infant mortality rates in the United States, 1989–2006. Am J Public Health. 2014;1048:1549–1556. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Martin JA, Hamilton BE, Osterman MJ et al. Births: final data for 2012. Natl Vital Stat Rep. 2013;629:1–68. [PubMed] [Google Scholar]
- 11.Fisher SC, Kim SY, Sharma AJ et al. Is obesity still increasing among pregnant women? Prepregnancy obesity trends in 20 states, 2003–2009. Prev Med. 2013;566:372–378. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Moniz MH, Chang T, Choi H et al. Obesity prevalence among reproductive-aged women by number of live births: United States, 2001–2010. Obstet Gynecol. 2014;123(suppl 1):158S. [Google Scholar]
- 13.Hayes DK, Fan AZ, Smith RA et al. Trends in selected chronic conditions and behavioral risk factors among women of reproductive age, behavioral risk factor surveillance system, 2001–2009. Prev Chronic Dis. 2011;86:A120. [PMC free article] [PubMed] [Google Scholar]
- 14.Conde-Agudelo A, Althabe F, Belizán JM et al. Cigarette smoking during pregnancy and risk of preeclampsia: a systematic review. Am J Obstet Gynecol. 1999;1814:1026–1035. [DOI] [PubMed] [Google Scholar]
- 15.Engel SM, Janevic TM, Stein CR et al. Maternal smoking, preeclampsia, and infant health outcomes in New York City, 1995–2003. Am J Epidemiol. 2009;1691:33–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Breathett K, Muhlestein D, Foraker R et al. Differences in preeclampsia rates between African American and Caucasian women: trends from the National Hospital Discharge Survey. J Womens Health (Larchmt). 2014;2311:886–893. [DOI] [PubMed] [Google Scholar]
- 17.Zollinger TW, Przybylski MJ, Gamache RE. Reliability of Indiana birth certificate data compared to medical records. Ann Epidemiol. 2006;161:1–10. [DOI] [PubMed] [Google Scholar]
- 18.Bodnar LM, Simhan HN. The prevalence of preterm birth and season of conception. Paediatr Perinat Epidemiol. 2008;226:538–545. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Ding G, Tian Y, Zhang Y et al. Application of a global reference for fetal-weight and birthweight percentiles in predicting infant mortality. BJOG. 2013;12013:1613–1621. [DOI] [PubMed] [Google Scholar]
- 20.Mikolajczyk RT, Zhang J, Betran AP et al. A global reference for fetal-weight and birthweight percentiles. Lancet. 2011;3779780:1855–1861. [DOI] [PubMed] [Google Scholar]
- 21.Carpenter MW, Coustan DR. Criteria for screening tests for gestational diabetes. Am J Obstet Gynecol. 1982;1447:768–773. [DOI] [PubMed] [Google Scholar]
- 22.Rasmussen K, Yaktine A, eds. Weight Gain During Pregnancy: Reexamining the Recommendations. Washington, DC: The National Academies Press; 2009. [PubMed] [Google Scholar]
- 23.Koenker R. Quantile Regression. New York, NY: Cambridge University Press; 2005. [Google Scholar]
- 24.Catalano PM, Ehrenberg HM. The short- and long-term implications of maternal obesity on the mother and her offspring. BJOG. 2006;11310:1126–1133. [DOI] [PubMed] [Google Scholar]
- 25.Challier JC, Basu S, Bintein T et al. Obesity in pregnancy stimulates macrophage accumulation and inflammation in the placenta. Placenta. 2008;293:274–281. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Macdonald-Wallis C, Tilling K, Fraser A et al. Associations of blood pressure change in pregnancy with fetal growth and gestational age at delivery: findings from a prospective cohort. Hypertension. 2014;641:36–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Fridman M, Korst LM, Chow J et al. Trends in maternal morbidity before and during pregnancy in California. Am J Public Health. 2014;104(suppl 1):S49–S57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Catalano PM, Mele L, Landon MB et al. Inadequate weight gain in overweight and obese pregnant women: what is the effect on fetal growth? Am J Obstet Gynecol. 2014;2112:137.e1–137.e7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Rasmussen S, Irgens LM, Espinoza J. Maternal obesity and excess of fetal growth in pre-eclampsia. BJOG. 2014;12111:1351–1357. [DOI] [PubMed] [Google Scholar]