Objective
To compare maternal and neonatal outcomes in obese women according to weight change and obesity class.
Study Design
Cohort study from the Consortium on Safe Labor of 20,950 obese women with a singleton, term live birth from 2002-2008. Risk for adverse outcomes was calculated by multiple logistic regression analysis for weight change categories (weight loss [<0 kg], low [0-4.9 kg], normal [5.0-9.0 kg], high weight gain [>9.0 kg]) in each obesity class (I 30.0-34.9 kg/m 2 , II 35.0-39.9 kg/m 2 , and III ≥40 kg/m 2 ) and by predicted probabilities with weight change as a continuous variable.
Results
Weight loss was associated with decreased cesareans for class I women (nulliparas odds ratio [OR], 0.21; 95% confidence interval [CI], 0.11–0.42; multiparas OR, 0.61; 95% CI, 0.45–0.83) and increased small for gestational age infants (class I OR, 1.8; 95% CI, 1.3–2.5; class II OR, 2.2; 95% CI, 1.5–3.2; class III OR, 1.7; 95% CI, 1.1–2.6). High weight gain was associated with increased large for gestational age infants (class I OR, 2.4; 95% CI, 1.9–2.9; class II OR, 1.7; 95% CI, 1.3–2.1; class III OR, 1.6; 95% CI, 1.3–2.1). As weight change increased, the predicted probability for cesareans and large for gestational age infants increased. The predicted probability of low birthweight never exceeded 4% for all obesity classes, but small for gestational age infants increased with decreased weight change. The lowest average predicted probability of adverse outcomes (cesarean, postpartum hemorrhage, small for gestational age, large for gestational age, neonatal care unit admission) occurred when women (class I, II, III) lost weight.
Conclusion
Optimal maternal and neonatal outcomes appear to occur when weight gain is less than current Institute of Medicine recommendations for obese women. Further study of long-term outcomes is needed with respect to gestational weight changes.
Obesity has reached epidemic proportions, estimated at 35.7% for adults in the United States. The prevalence of obese reproductive age women (20-39 years) increased by 64% between 1988-1994 and 2007-2008, accounting for the greatest increase in obesity for women of any age category. The obesity epidemic and its associated obstetric and neonatal complications has highlighted the issue of gestational weight gain (GWG). Regardless of maternal weight status, high GWG has been associated with both maternal (eg, cesarean deliveries, long-term weight retention) and offspring risks (eg, larger infants, childhood obesity). Although the evidence is less consistent, an association between higher GWG and gestational diabetes and preeclampsia has also been reported. On the other hand, low GWG and weight loss have also been associated with maternal ketonemia and fetal growth restriction.
According to conventional wisdom, pregnancy is a time for weight gain, not for dieting or weight loss. Typically, an additional 300 calories per day is recommended for appropriate fetal growth and this was reflected in the 1990 Institute of Medicine report on Nutrition During Pregnancy. More recently, in 2009, the Institute of Medicine published revised guidelines for GWG. Important updates included a range of 5–9 kg (or 11-20 lbs) for GWG in obese women, defined by a prepregnancy body mass index (BMI) ≥30 kg/m 2 . This differed from the prior recommendation of “at least 15 pounds.” Notably, all obese women, were grouped into 1 category as a result of insufficient data from women in individual obesity classes (ie, obesity classes I-III) and the inability to draw statistically sound conclusions for GWG for the separate obesity classes. Behavioral interventions for women who are obese at conception have shown some success at meeting GWG recommendations, but limited evidence suggests that lower GWG or weight loss (ie, gestational weight change) in this population may improve maternal and neonatal outcomes. Given the nationwide rise in obesity as well as the influence of GWG on maternal and neonatal outcomes, we hypothesized that the current recommendations (up to a 20 pound weight gain) were too high for obese women and required closer scrutiny into each of the obesity classes. The objective of our investigation was to evaluate maternal and neonatal outcomes at birth in obese women by weight change and BMI class.
Materials and Methods
The Consortium on Safe Labor is a retrospective, observational, electronic database acquired from 12 institutions (19 hospitals) across 9 American College of Obstetricians and Gynecologists (ACOG) districts in the United States. The complete database contains 233,730 births resulting from 228,562 deliveries. Although the data were collected from 2002 to 2008, 87% of the births in the database occurred between 2005 and 2007. Extensive data were collected on each delivery including demographics, prenatal complications, labor and delivery information, and maternal and neonatal outcomes. Validation studies on 4 outcomes (shoulder dystocia, cesarean delivery for nonreassuring fetal heart rate, neonatal intensive care unit [NICU] admission for respiratory conditions, and neonatal asphyxia) were performed by hand-abstraction of eligible charts. Most variables reviewed were highly accurate when comparing data from the electronic database and the hand-abstraction. Further detail regarding the database is available.
Inclusion criteria for the current study were a prepregnancy BMI ≥30 kg/m 2 and known gestational weight change in a singleton, term (≥37.0 weeks), live-born gestation. If a woman contributed more than 1 pregnancy to the database, only the first pregnancy was analyzed to maintain the independence of the observations. Class I-III obesity was defined according to the World Health Organization (WHO) criteria as class I 30.0-34.9 kg/m 2 , class II 35.0-39.9 kg/m 2 , and class III ≥40 kg/m 2 . Weight change was defined as the difference between the self-reported prepregnancy weight and delivery weight. The weight change categories were defined as weight loss, low (0-4.9 kg), normal (5.0-9.0 kg), and high (> 9.0 kg). The weight change categories were chosen for their simplicity, ease of clinical use, and were also modeled after those of another investigation. Although the range of gestational weight change was −55 kg to 77 kg in the current cohort, this range was restricted to −20 kg to 50 kg to reflect a more clinically plausible value, also similar to the range chosen in another study. In doing so, only 70 women or 0.3% of the cohort was excluded.
Maternal demographics and characteristics included age, race/ethnicity, marital status, insurance, parity, smoking status, prior cesarean delivery, pregestational diabetes, chronic hypertension, and gestational age at delivery. The primary maternal outcomes were operative vaginal delivery, cesarean delivery, and postpartum hemorrhage. The primary neonatal outcomes were birthweight, shoulder dystocia, 5 minute Apgar score <7, and NICU admission. Small and large for gestational age infants (SGA, LGA) were defined by birthweights <10th% or >90th%, respectively, for the gestational age at birth. Low birthweight (LBW) and macrosomia were defined by birthweights <2500 g or >4500 g, respectively.
All analyses were stratified by obesity class (I, II, III). For the maternal demographics and characteristics, Pearson χ 2 , and analysis of variance tests were used to statistically compare the association between categorical and continuous variables, respectively, and weight change category. A P value < .05 was considered statistically significant. Through multiple logistic regression analysis, adjusted odds ratios (aORs) with 95% confidence intervals (CIs) were generated for each maternal and neonatal outcome in each of the weight change categories, using 5.0-9.0 kg as the referent, for each obesity class and adjusting for age, race/ethnicity, marital status, insurance, parity, smoking, and gestational age. To clarify the potential maternal and neonatal risks and benefits of weight change, logistic regression models were created with weight change as a continuous variable (kg) for each BMI class. The estimated logistic regression coefficients then determined the predicted probabilities for each of the maternal and neonatal outcomes, similar to the analysis approach of another study. The average predicted probability of 5 outcomes (cesarean, postpartum hemorrhage, SGA, LGA, and NICU) was then calculated over a weight change of −20 kg to +50 kg. These outcomes were chosen for their clinical relevance or the apparent association between gestational weight change in the logistic regression models and the individual predicted probabilities. Given that each of these 5 outcomes increased with increased weight change except for SGA, we performed a sensitivity analysis weighting SGA (2-5 times) to determine how varying the importance of this outcome would influence the average predicted probability.
To further determine the potential immediate neonatal risk associated with either weight loss or low GWG, we analyzed neonatal outcomes for SGA infants born to women with weight loss or low (0-4.9 kg) weight gain and compared them with SGA infants born to women with normal (5.0-9.0 kg) weight gain using χ 2 or Fisher exact tests. The following neonatal outcomes were reported for this analysis: respiratory distress syndrome (RDS), transient tachypnea of the newborn (TTN), use of ventilators, oxygen or CPAP (continuous positive airway pressure), pneumonia, meconium aspiration, anemia, sepsis, asphyxia, congenital anomalies, NICU admission, and death. All statistical analyses were performed with SAS software (version 9.2; SAS Institute, Cary, NC) using primarily LOGSTIC and GLM procedures. Appropriate institutional review board approval was obtained from all the participating institutions.
Results
From the 228,562 deliveries in the entire database, 20,950 obese women (11,984 class I, 5307 class II, and 3659 class III) were studied ( Figure 1 ). Of the 57 stillbirths, which were excluded from the remainder of the analysis, 70% occurred in those with high weight gain and none occurred in those who lost weight. There were differences among the weight change categories with respect to age, race/ethnicity, marital status, insurance, parity, smoking, and prior cesarean, P ≤ .001 ( Table 1 ). This analysis was stratified by obesity class and the differences in the comparisons persisted for each obesity class, P < .02 (data not shown). The mean (± SD) GWG was 12.5 ± 7.6 kg class I, 10.6 ± 8.1 kg class II, and 8.9 ± 8.9 kg class III, P < .001. Weight loss was most common in class III (12%) and high weight gain was most common in class I women (69%). For women who lost weight, the mean (± SD) weight loss was −4.8 ± 4.5 kg, −4.6 ± 4.3 kg, and −5.6 ± 4.2 kg for class I, II, and III, respectively. The proportion of women gaining within the 2009 Institute of Medicine recommended guidelines (5-9 kg) was 17% for all obesity classes.
| Variable n (%) or mean (SD) | Weight change categories a | P value | |||
|---|---|---|---|---|---|
| Loss n = 1182 | Low n = 3028 | Normal n = 3613 | High n = 13,127 | ||
| Age, y | 28.1 ± 5.7 | 28.8 ± 5.7 | 28.7 ± 5.8 | 27.7 ± 5.9 | < .001 b |
| Race/Ethnicity | < .001 c | ||||
| Non-Hispanic white | 555 (48.3) | 1419 (48.38) | 1582 (44.9) | 5710 (44.8) | |
| Non-Hispanic black | 405 (35.2) | 848 (28.9) | 988 (28.0) | 3943 (30.9) | |
| Hispanic | 161 (14.0) | 556 (18.9)` | 826 (23.4) | 2555 (20.0) | |
| Other | 28 (2.4) | 114 (3.9) | 128 (3.6) | 544 (4.3) | |
| Insurance | < .001 c | ||||
| Private | 576 (48.7) | 1597 (52.7) | 1798 (49.8) | 6412 (48.8) | |
| Public | 540 (45.7) | 1167 (38.5) | 1416 (39.2) | 5288 (40.3) | |
| Other | 66 (5.6) | 264 (8.7) | 399 (11.0) | 1427 (10.9) | |
| Married | 637 (53.9) | 1808 (59.7) | 2124 (58.9) | 7011 (53.4) | < .001 c |
| Parity | < .001 c | ||||
| 0 | 305 (25.8) | 743 (24.5) | 983 (27.2) | 4908 (37.4) | |
| ≥1 | 877 (74.2) | 2285 (75.5) | 2630 (72.8) | 8219 (62.6) | |
| Smoker | 138 (11.7) | 248 (8.2) | 259 (7.2) | 1025 (7.8) | < .001 c |
| Prior cesarean | 232 (21.0) | 641 (22.6) | 826 (24.2) | 2552 (20.4) | < .001 c |
| Pregestational diabetes | 59 (5.4) | 143 (5.0) | 153 (4.6) | 505 (4.1) | .05 c |
| Chronic hypertension | 69 (6.7) | 173 (6.6) | 210 (6.6) | 872 (7.6) | .11 c |
| Weight change, kg | −5.0 ± 4.4 | 2.8 ± 1.6 | 7.1 ± 1.1 | 16.0 ± 5.8 | < .001 b |
| Gestational age at delivery, wks | 39.1 ± 1.1 | 39.1 ± 1.1 | 39.1 ± 1.1 | 39.2 ± 1.1 | < .001 b |
a Weight loss (<0 kg), low (0-4.9 kg weight gain), normal (5.0-9.0 kg weight gain), and high (>9.0 kg weight gain)
The aORs for the maternal and neonatal outcomes from the multiple logistic regression analysis are presented in Tables 2 and 3 , respectively. Cesareans decreased for class I women who lost weight, but increased in women with high weight gain in most BMI classes, compared with normal weight gain. Operative vaginal delivery and postpartum hemorrhage did not differ from normal weight gain in all obesity classes. SGA infants increased for women in all obesity classes who lost weight, whereas LGA and macrosomia increased with high weight gain in most obesity classes, compared with normal weight gain. Shoulder dystocia and 5 minute Apgar <7 did not differ from normal weight gain in all obesity classes.
| Variable | Weight change categories a | ||||
|---|---|---|---|---|---|
| Loss | Low | Normal | High | ||
| Operative vaginal delivery | |||||
| Class I | n (%) | 24 (6.2) | 62 (4.6) | 80 (4.1) | 460 (5.5) |
| aOR (95% CI) | 1.5 (0.93–2.5) | 1.1 (0.80–1.6) | 1 | 1.3 (0.98–1.6) | |
| Class II | n (%) | 16 (4.6) | 43 (4.7) | 41 (4.0) | 133 (4.4) |
| aOR (95% CI) | 1.2 (0.64–2.2) | 1.1 (0.68–1.7) | 1 | 1.0 (0.70–1.4) | |
| Class III | n (%) | 18 (4.0) | 37 (4.9) | 27 (4.1) | 82 (4.6) |
| aOR (95% CI) | 0.97 (0.52–1.8) | 1.2 (0.69–1.9) | 1 | 1.1 (0.73–1.8) | |
| Nulliparas cesarean delivery | |||||
| Class I | n (%) | 15 (14.7) | 111 (34.6) | 195 (37.7) | 1279 (41.8) |
| aOR (95% CI) | 0.21 (0.11–0.42) | 0.85 (0.62–1.2) | 1 | 1.2 (1.0–1.5) | |
| Class II | n (%) | 30 (34.9) | 90 (41.7) | 115 (39.4) | 565 (48.2) |
| aOR (95% CI) | 0.81 (0.48–1.4) | 1.1 (0.74–1.6) | 1 | 1.5 (1.1–2.0) | |
| Class III | n (%) | 58 (43.9) | 105 (51.0) | 79 (45.4) | 401 (58.6) |
| aOR (95% CI) | 0.79 (0.49–1.3) | 1.1 (0.71–1.7) | 1 | 1.7 (1.2–2.4) | |
| Multiparas cesarean delivery | |||||
| Class I | n (%) | 64 (21.6) | 292 (28.1) | 425 (30.1) | 1816 (34.5) |
| aOR (95% CI) | 0.61 (0.44–0.83) | 0.88 (0.74–1.1) | 1 | 1.3 (1.1–1.4) | |
| Class II | n (%) | 86 (32.0) | 237 (33.8) | 277 (38.1) | 754 (40.6) |
| aOR (95% CI) | 0.82 (0.60–1.1) | 0.82 (0.66–1.0) | 1 | 1.1 (0.93–1.3) | |
| Class III | n (%) | 144 (41.0) | 222 (40.7) | 233 (47.5) | 562 (50.5) |
| aOR (95% CI) | 0.76 (0.56–1.0) | 0.77 (0.59–.99) | 1 | 1.1 (0.94–1.5) | |
| Postpartum hemorrhage | |||||
| Class I | n (%) | 4 (1.5) | 25 (2.7) | 38 (2.7) | 185 (3.1) |
| aOR (95% CI) | 0.55 (0.19–1.6) | 0.96 (0.57–1.6) | 1 | 1.1 (0.75–1.5) | |
| Class II | n (%) | 6 (2.6) | 20 (3.2) | 33 (4.5) | 81 (3.7) |
| aOR (95% CI) | 0.54 (0.22–1.3) | 0.67 (0.38–1.2) | 1 | 0.74 (0.49–1.1) | |
| Class III | n (%) | 15 (4.5) | 20 (3.8) | 19 (3.8) | 90 (6.6) |
| aOR (95% CI) | 1.1 (0.55–2.2) | 0.94 (0.49–1.8) | 1 | 1.6 (0.97–2.7) | |
a Weight loss (<0 kg), low (0-4.9 kg weight gain), normal (5.0-9.0 kg weight gain, referent), and high (>9.0 kg weight gain).
| Variable | Weight change categories a | ||||
|---|---|---|---|---|---|
| Loss | Low | Normal | High | ||
| SGA infants | |||||
| Class I | n (%) | 66 (17.3) | 135 (10.0) | 187 (9.7) | 549 (6.6) |
| aOR (95% CI) | 1.8 (1.3–2.5) | 1.0 (0.82–1.3) | 1 | 0.60 (0.50–0.72) | |
| Class II | n (%) | 57 (16.6) | 100 (10.9) | 85 (8.4) | 193 (6.4) |
| aOR (95% CI) | 2.2 (1.5–3.2) | 1.4 (1.0–1.9) | 1 | 0.66 (0.51–0.87) | |
| Class III | n (%) | 57 (12.8) | 72 (9.6) | 51 (7.7) | 119 (6.7) |
| aOR (95% CI) | 1.7 (1.1–2.6) | 1.4 (0.93–2.0) | 1 | 0.73 (0.51–1.0) | |
| LGA infants | |||||
| Class I | n (%) | 13 (3.4) | 76 (5.6) | 119 (6.2) | 1029 (12.5) |
| aOR (95% CI) | 0.59 (0.33–1.1) | 0.92 (0.68–1.2) | 1 | 2.4 (1.9–2.9) | |
| Class II | n (%) | 17 (5.0) | 67 (7.3) | 101 (9.9) | 435 (14.5) |
| aOR (95% CI) | 0.51 (0.30–0.86) | 0.71 (0.51–0.98) | 1 | 1.7 (1.3–2.1) | |
| Class III | n (%) | 31 (6.9) | 77 (10.3) | 85 (12.8) | 315 (17.7) |
| aOR (95% CI) | 0.48 (0.31–0.75) | 0.76 (0.54–1.1) | 1 | 1.6 (1.3–2.1) | |
| LBW | |||||
| Class I | n (%) | 17 (4.4) | 28 (2.1) | 46 (2.4) | 133 (1.6) |
| aOR (95% CI) | 1.6 (0.85–3.0) | 0.90 (0.55–1.5) | 1 | 0.65 (0.46–0.94) | |
| Class II | n (%) | 15 (4.4) | 22 (2.4) | 19 (1.8) | 61 (2.1) |
| aOR (95% CI) | 2.4 (1.1–4.9) | 1.3 (0.69–2.5) | 1 | 1.0 (0.59–1.7) | |
| Class III | n (%) | 9 (2.0) | 16 (2.1) | 12 (1.8) | 33 (1.9) |
| aOR (95% CI) | 1.1 (0.45–2.8) | 1.3 (0.60–3.0) | 1 | 0.91 (0.45–1.9) | |
| Macrosomia | |||||
| Class I | n (%) | 2 (0.52) | 7 (0.52) | 15 (0.78) | 191 (2.3) |
| aOR (95% CI) | 0.83 (0.19–3.6) | 0.73 (0.30–1.8) | 1 | 3.1 (1.9–5.4) | |
| Class II | n (%) | 1 (0.29) | 8 (0.88) | 13 (1.3) | 87 (2.9) |
| aOR (95% CI) | 0.24 (0.03–1.9) | 0.66 (0.27–1.6) | 1 | 2.4 (1.3–4.4) | |
| Class III | n (%) | 7 (1.6) | 11 (1.5) | 19 (2.9) | 66 (3.7) |
| aOR (95% CI) | 0.58 (0.24–1.4) | 0.51 (0.24–1.1) | 1 | 1.5 (0.86–2.5) | |
| Shoulder dystocia | |||||
| Class I | n (%) | 1 (0.28) | 13 (1.0) | 32 (1.8) | 156 (2.0) |
| aOR (95% CI) | 0.16 (0.02–1.2) | 0.60 (0.31–1.1) | 1 | 1.1 (0.76–1.7) | |
| Class II | n (%) | 3 (0.93) | 17 (2.0) | 15 (1.6) | 43 (1.5) |
| aOR (95% CI) | 0.58 (0.17–2.0) | 1.3 (0.64–2.6) | 1 | 0.91 (0.50–1.7) | |
| Class III | n (%) | 6 (1.5) | 6 (0.9) | 8 (1.3) | 30 (1.8) |
| aOR (95% CI) | 1.1 (0.37–3.2) | 0.69 (0.24–2.0) | 1 | 1.4 (0.63–3.1) | |
| 5 min Apgar <7 | |||||
| Class I | n (%) | 2 (0.52) | 5 (0.37) | 16 (0.83) | 78 (0.94) |
| aOR (95% CI) | 0.56 (0.13–2.4) | 0.43 (0.16–1.2) | 1 | 0.98 (0.57–1.7) | |
| Class II | n (%) | 3 (0.87) | 8 (0.87) | 9 (0.89) | 28 (0.93) |
| aOR (95% CI) | 1.0 (0.28–3.9) | 1.0 (0.40–2.7) | 1 | 0.96 (0.45–2.1) | |
| Class III | n (%) | 4 (0.89) | 3 (0.40) | 6 (0.90) | 23 (1.3) |
| aOR (95% CI) | 0.91 (0.26–3.3) | 0.15 (0.02–1.2) | 1 | 1.3 (0.52–3.2) | |
| NICU admission | |||||
| Class I | n (%) | 33 (8.5) | 79 (5.8) | 146 (7.6) | 715 (8.6) |
| aOR (95% CI) | 1.1 (0.72–1.6) | 0.75 (0.56–1.0) | 1 | 1.1 (0.92–1.3) | |
| Class II | n (%) | 28 (8.1) | 92 (10.0) | 81 (8.0) | 307 (10.1) |
| aOR (95% CI) | 1.0 (0.64–1.6) | 1.3 (0.92–1.7) | 1 | 1.3 (0.99–1.7) | |
| Class III | n (%) | 51 (10.8) | 57 (7.7) | 83 (12.8) | 196 (11.0) |
| aOR (95% CI) | 0.83 (0.57–1.2) | 0.56 (0.39–0.80) | 1 | 0.85 (0.64–1.1) | |
a Weight loss (<0 kg), low (0-4.9 kg weight gain), normal (5.0-9.0 kg weight gain, referent), and high (>9.0 kg weight gain).
The predicted probability of cesarean increased linearly as weight change increased for all obesity classes, whereas operative vaginal delivery and postpartum hemorrhage did not vary significantly with weight change ( Figure 2 , A-C). The predicted probabilities of SGA decreased as weight change increased, whereas LGA and macrosomia increased exponentially with increased weight change. Shoulder dystocia, 5 minute Apgar <7, and NICU admissions showed a less pronounced increase as weight change increased for all obesity classes ( Figure 3 , A-C). The predicted probability of LBW was always <4% for all the obesity classes. The average predicted probability of 5 maternal and neonatal outcomes ( Figure 4 ) shows that the lowest probability of these outcomes occurred at weight loss in all obesity classes. In the sensitivity analysis of varying weights for SGA, weight gain was associated with the lowest average predicted probability only when SGA was weighted 4 times for class I (lowest predicted probability of 10.6% at 5.0 kg), 3 times for class II (lowest predicted probability of 12.4% at 4.5 kg), and 5 times for class III (lowest predicted probability of 13.1% at 2.9 kg) obese women ( Figure 5 ).
