Objective
To compare maternal and neonatal outcomes between planned cesarean delivery and induction of labor in women with class III obesity (body mass index ≥40 kg/m 2 ).
Study Design
In this retrospective cohort study, we identified all women with a body mass index ≥40 kg/m 2 who delivered a singleton at our institution from January 2007 to February 2013 via planned cesarean or induction of labor (regardless of eventual delivery route) at 37-41 weeks. Patients in spontaneous labor were excluded. The primary outcome was a composite of maternal morbidity including death as well as operative, infection, and thromboembolic complications. The secondary outcome was a neonatal morbidity composite. Additional outcomes included individual components of the composites. Student t , χ 2 , and Fisher exact tests were used for statistical analysis. To calculate adjusted odds ratios, covariates were analyzed via multivariable logistic regression.
Results
There are 661 mother-infant pairs that met enrollment criteria—399 inductions and 262 cesareans. Groups were similar in terms of prepregnancy weight, pregnancy weight gain, and delivery body mass index. Of the 399 inductions, 258 had cervical ripening (64.7%) and 163 (40.9%) had a cesarean delivery. After multivariable adjustments, there was no significant difference in the maternal morbidity composite (adjusted odds ratio, 0.98; 95% confidence interval, 0.55–1.77) or in the neonatal morbidity composite (adjusted odds ratio, 0.81; 95% confidence interval, 0.37–1.77) between the induction and cesarean groups.
Conclusion
In term pregnant women with class III obesity, planned cesarean does not appear to reduce maternal and neonatal morbidity compared with induction of labor.
Over the last 2 decades, the rate of obesity has increased dramatically such that in the United States greater than 30% of reproductive-age women are obese (body mass index [BMI] ≥30) and approximately 8% are extremely obese (class III obesity – BMI ≥40 kg/m 2 ). The obesity epidemic affects all racial and ethnic groups and is highly associated with major medical comorbidities and poor pregnancy outcomes. Compared with normal-weight women, those with class III obesity have higher rates of labor induction and induction failure. In addition, increasing maternal obesity is associated with longer durations of labor, greater oxytocin requirements, and higher cesarean delivery rates. These longer durations of labor and greater oxytocin requirements result in increased rates of clinical chorioamnionitis, postpartum hemorrhage, and neonatal morbidity. Obese patients undergoing cesarean delivery also have an increased risk of endometritis and surgical complications, including wound infection, seroma, hematoma, skin disruption, and fascial dehiscence.
These prior studies demonstrate that there are increased risks with either labor induction or cesarean delivery in class III obese pregnant women. In addition, rates of postcesarean complications are significantly higher in those women who undergo cesarean delivery after an unsuccessful labor attempt. To date, few studies have evaluated the role of planned cesarean as opposed to labor induction in this high-risk population. Because of the increased risk of labor induction and the increased risk of morbidity in women with class III obesity who undergo labor induction, we hypothesized that planned cesarean delivery may result in less morbidity than induction in this high-risk group. Thus, our objective was to evaluate, in women with class III obesity at term, the differences in maternal, obstetric, and neonatal outcomes between planned cesarean delivery and induction of labor.
Materials and Methods
Following institutional review board approval, we identified all women from our validated obstetric research database (Obstetric Automated Record) with a singleton gestation and BMI ≥40kg/m 2 at delivery who underwent a planned cesarean (malpresentation, repeat, elective, suspected macrosomia) or induction of labor at 37 0/7 -41 6/7 weeks’ gestational age from January 2007 to February 2013. In this retrospective cohort study, we excluded patients in spontaneous labor as well as those with severe immunodeficiency (AIDS, chronic steroid use), multifetal gestations, fetal demise before presenting for delivery, and prenatally diagnosed congenital anomalies. Individual chart review was undertaken. Data were abstracted and entered into a relational database (Access 2003; Microsoft Corporation; Redmond, WA).
The primary outcome of this study was a composite of clinically relevant maternal morbidities including operative complications, infections, wound morbidity, venous thromboembolism (radiographically documented deep vein thrombosis or pulmonary embolism), and maternal death. Operative complications included hysterectomy, operative injury (uterine artery laceration requiring an O’Leary stitch; bladder, ureteral, or bowel injury; and/or extension of the hysterotomy into the contractile portion of the uterus), and postpartum hemorrhage (defined clinically as blood loss >500 mL after vaginal delivery, blood loss >1000 mL after cesarean delivery, or need for uterotonic medications other than oxytocin). Infections included clinical chorioamnionitis (temperature of at least 38°C before delivery and antibiotics administered for that indication), endometritis (temperature of at least 38°C postpartum and antibiotics administered for that indication), wound infection (defined as cellulitis or abscess), urinary tract infection (>100,000 colony forming units/mL of a single uropathogen on urine culture), radiographically diagnosed pneumonia, and culture-proven sepsis. Wound morbidity included cellulitis, wound disruption (hematoma or seroma), intraabdominal hematoma, need for wound exploration, debridement, packing, or antibiotic therapy.
The secondary outcome was a neonatal composite including intrapartum stillbirth, hypoxic ischemic encephalopathy (HIE) (defined as arterial cord pH <7.0 and seizures or other end organ dysfunction), neonatal intensive care unit admission for >72 hours, mechanical ventilation >24 hours, grade III or IV intraventricular hemorrhage, respiratory distress syndrome, necrotizing enterocolitis, suspected sepsis with antibiotics administered >72 hours during the first week of life, culture proven early-onset (first 7 days of life) sepsis, or neonatal death. Other neonatal outcomes evaluated but not included in the composite variable were 5-minute Apgar score <7, umbilical artery pH <7.1, transient tachypnea of the newborn, hyperbilirubinemia, and metabolic disturbances (hypo/hyperglycemia, hypo/hypernatremia). We considered neonatal diagnoses to have been made if they were included in the infant’s discharge summary or inpatient progress notes when available.
Individual components of composite variables were also analyzed independently. Owing to the imprecise nature of clinically diagnosed postpartum hemorrhage, we also analyzed the proportion of women in each group who had a decrease in the pre- to postdelivery hematocrit of at least 10% and those who received a blood transfusion. Patient demographics (age, race/ethnicity, parity, BMI) and maternal medical comorbidities were assessed as covariates. Intrapartum characteristics such as duration of labor, mode of delivery, indication for cesarean delivery, estimated blood loss, and use of cervical ripening were also ascertained.
The primary and secondary outcomes were compared between 2 groups: planned cesarean and induction of labor. We also stratified the induction of labor group into successful vaginal delivery and cesarean delivery to compare outcomes between 3 groups: planned cesarean, induction of labor with vaginal delivery, and induction of labor with cesarean delivery. Furthermore, in a prespecified secondary analysis, we compared the 2 original groups stratifying women into subsets with a BMI 40-49.9 and ≥50.
We used the unpaired Student t test for continuous variables. The χ 2 and Fisher exact tests were used for categorical data as appropriate. Odds ratios (ORs) were calculated. Covariates were chosen based on differences identified in the key characteristics between the 2 cohorts ( P < .05). Multivariable logistic regression was used to analyze the association between mode of delivery and outcomes, and adjustments were made for covariates. All statistical analyses were conducted using SAS version 9.2 (SAS Institute; Cary, NC).
With regards to obesity, Alabama perennially is in the top 3 or 4 states in the United States. In a study from our center over a 20-year period, the obesity rate in the obstetric population more than doubled from 16% in 1980 to 36% in 1999. The rate at our center now exceeds 40%, with the rate of class III obesity being 10%. Assuming that the primary outcome variable would occur in 30% of those in the labor induction group and that there would be a 2:1 ratio of inductions of labor to planned cesarean deliveries, we calculated that a sample size of 600 was needed to detect a 40% decrease in that outcome (to 18%) in the planned cesarean group (alpha = .05; 1-beta = 0.90). We thought that smaller differences might not justify choosing a cesarean delivery over attempting a vaginal delivery. Since there are approximately 4000 deliveries annually at our institution, we calculated that a 5-year study period would be sufficient as long as 3% of our patient population met enrollment criteria.
Results
A total of 661 patients met enrollment criteria and were included for analysis. Of these, 262 (39.6%) patients had a planned cesarean delivery, whereas 399 (60.4%) patients underwent induction of labor. For patients undergoing planned cesarean delivery, indications for cesarean included repeat (n = 219), malpresentation (n = 22), macrosomia (n = 10), elective (patient demand or inability to perform intrapartum monitoring) (n = 6), and other (n = 5). In women undergoing labor inductions, there were 236 vaginal deliveries (59.1%) and 163 cesarean deliveries (40.9%). Of the women undergoing induction, 258 (64.7%) had cervical ripening with a Foley catheter (majority) or prostaglandins. The most common indications for cesarean in the induction group included failed induction (failure to enter active phase of labor within 24-48 hours) (n = 72), nonreassuring fetal heart rate tracing (n = 97), and active phase arrest (arrest of dilation for >4-6 hours after a cervical dilation of at least 5 cm) (n = 19). Baseline demographics and other characteristics of these 2 groups are presented in Table 1 . Groups were similar in terms of height, prepregnancy weight, pregnancy weight gain, delivery BMI, smoking, payment source, diabetes, gestational age at delivery, and neonatal gender. However, the groups differed regarding age, parity, race/ethnicity, marital status, hypertension (both chronic and pregnancy related), history of cesarean delivery, history of vaginal delivery, suspected fetal growth restriction, oligohydramnios, abnormal antenatal testing, and neonatal birthweight.
| Description | Induction of labor, n = 399 | Planned cesarean, n = 262 | P value |
|---|---|---|---|
| Age, y | 27.6 ± 5.8 | 29.1 ± 5.6 | .001 |
| Parity | 1 (0–2) | 2 (1–2) | < .001 |
| Nulliparity | 180 (45.1%) | 14 (5.3%) | < .001 |
| Race/Ethnicity | < .001 | ||
| Black | 302 (75.9%) | 167 (63.7%) | |
| White | 79 (19.9%) | 65 (24.8%) | |
| Hispanic | 13 (3.3%) | 27 (10.3%) | |
| Other | 4 (1.0%) | 3 (1.2%) | |
| Height, in | 64.9 ± 2.9 | 64.7 ± 2.9 | .23 |
| Prepregnancy weight, lbs | 261.4 ± 45.4 | 263.9 ± 55.6 | .60 |
| Pregnancy weight gain, lbs | 25.6 ± 19.8 | 24.5 ± 24.0 | .56 |
| Delivery BMI | .88 | ||
| 40-49.9 | 275 (68.9%) | 182 (69.5%) | |
| ≥50 | 124 (31.1%) | 80 (30.5%) | |
| Cigarette smoking | 45 (11.3%) | 40 (15.3%) | .13 |
| Alcohol use | 2 (0.5%) | 0 (0%) | .52 |
| Illicit drug use | 12 (3.0%) | 2 (0.8%) | .06 |
| Payment source | .57 | ||
| Private insurance | 60 (15.2%) | 33 (12.8%) | |
| Medicaid | 295 (74.7%) | 193 (75.1%) | |
| Uninsured/self pay | 40 (10.1%) | 31 (12.1%) | |
| Marital status | < .001 | ||
| Single | 296 (75.1%) | 170 (65.1%) | |
| Married | 85 (21.6%) | 88 (33.7%) | |
| Divorced | 13 (3.3%) | 3 (1.2%) | |
| Chronic HTN | 144 (36.1%) | 64 (24.4%) | .002 |
| Diabetes mellitus | 135 (33.8%) | 76 (29.0%) | .20 |
| Pregestational | 72 (18.1%) | 39 (14.9%) | .30 |
| Gestational | 63 (15.8%) | 37 (14.2%) | .57 |
| Pregnancy-related hypertension | 210 (52.9%) | 58 (22.1%) | < .001 |
| With magnesium | 88 (22.5%) | 25 (9.6%) | < .001 |
| Without magnesium | 303 (75.9%) | 235 (90.4%) | |
| Prior cesarean delivery | 43 (10.8%) | 228 (87.0%) | < .001 |
| Prior vaginal delivery | 195 (48.9%) | 59 (22.5%) | < .001 |
| Gestational age at delivery, wk | 38.6 ± 1.0 | 38.6 ± 0.9 | .61 |
| Fetal growth restriction | 14 (3.5%) | 2 (0.8%) | .04 |
| Oligohydramnios | 21 (5.3%) | 3 (1.2%) | .005 |
| Abnormal antepartum testing | 57 (14.3%) | 18 (6.9%) | .003 |
| Neonatal birthweight, g | 3330 ± 535 | 3520 ± 594 | < .001 |
Both unadjusted and adjusted results for the primary outcome and its components are presented in Table 2 . Overall, there was no significant difference in the primary outcome—total maternal morbidity composite—between the induction of labor and planned cesarean groups (OR, 1.3; 95% confidence interval [CI], 0.90–1.83). After multivariable adjustments for baseline differences in the groups, there still was no significant difference in the maternal morbidity composite (adjusted OR [aOR], 0.98; 95% CI, 0.55–1.77).
| Description | Induction of labor, n = 399 | Planned cesarean, n = 262 | OR (95% CI) | aOR a |
|---|---|---|---|---|
| Maternal morbidity composite, n (%) b | 117 (29.3) | 64 (24.4) | 1.3 (0.90–1.83) | 0.98 (0.55–1.77) |
| Operative complications, n (%) c | 50 (12.5) | 42 (16.0) | 0.75 (0.48–1.17) | 0.60 (0.28–1.25) |
| Hysterectomy | 0 (0) | 2 (0.8) | — | — |
| Operative Injury | 17 (4.3) | 39 (14.9) | 0.25 (0.14–0.46) | 0.30 (0.12–0.78) |
| Uterine artery laceration | 5 (1.3) | 1 (0.4) | 3.32 (0.39–28.58) | 5.7 (0.40–82.33) |
| Bladder/ureteral/bowel injury | 0 (0) | 1 (0.4) | — | — |
| Hysterotomy extension | 12 (3.0) | 37 (14.1) | 0.19 (0.10–0.37) | 0.19 (0.07–0.54) |
| Postpartum hemorrhage | 29 (7.3) | 9 (3.4) | 2.20 (1.03–4.73) | 0.67 (0.22–2.03) |
| HCT decrease ≥10% | 13 (3.3) | 5 (1.9) | 1.73 (0.61–4.91) | 0.90 (0.18–4.39) |
| Medication other than oxytocin | 17 (4.3) | 5 (1.9) | 2.29 (0.83–6.28) | 0.90 (0.20–4.18) |
| Blood transfusion | 4 (1.0) | 3 (1.2) | 0.88 (0.19–3.95) | 0.46 (0.05–4.02) |
| Infections, n (%) | 71 (17.8) | 21 (8.0) | 2.48 (1.49–4.16) | 1.62 (0.75–3.52) |
| Chorioamnionitis | 51 (12.8) | 2 (0.8) | 19.05 (4.60–78.97) | 10.6 (2.05–54.51) |
| Endometritis | 11 (2.8) | 6 (2.3) | 1.21 (0.44–3.31) | 1.56 (0.34–7.16) |
| Urinary tract infection | 6 (1.5) | 3 (1.2) | 1.32 (0.33–5.31) | 2.5 (0.34–18.81) |
| Pneumonia | 4 (1.0) | 2 (0.8) | 1.31 (0.24–7.23) | 1.03 (0.08–13.63) |
| Sepsis | 0 (0) | 0 (0) | — | — |
| Abdominal abscess | 1 (0.3) | 0 (0) | — | — |
| Wound infection | 11 (2.8) | 8 (3.1) | 0.90 (0.36–2.27) | 0.31 (0.08–1.28) |
| Wound morbidity, n (%) | 25 (6.3) | 21 (8.0) | 0.77 (0.42–1.40) | 0.31 (0.12–0.82) |
| Cellulitis | 14 (3.5) | 11 (4.2) | 0.83 (0.37–1.86) | 0.45 (0.13–1.61) |
| Wound infection | 11 (2.8) | 8 (3.1) | 0.90 (0.36–2.27) | 0.31 (0.08–1.28) |
| Subcutaneous hematoma | 1 (0.3) | 4 (1.5) | 0.16 (0.02–1.46) | 0.14 (0.01–1.52) |
| Seroma | 5 (1.3) | 4 (1.5) | 0.82 (0.22–3.08) | 0.25 (0.04–1.75) |
| Exploration | 8 (2.0) | 8 (3.1) | 0.65 (0.24–1.76) | 0.13 (0.03–0.57) |
| Surgical debridement | 4 (1.0) | 3 (1.2) | 0.88 (0.20–3.95) | 0.34 (0.06–1.83) |
| Wound packing | 10 (2.5) | 9 (3.4) | 0.72 (0.29–1.81) | 0.17 (0.04–0.69) |
| Antibiotics | 19 (4.8) | 16 (6.1) | 0.77 (0.39–1.53) | 0.44 (0.14–1.37) |
| Intraabdominal hematoma | 3 (0.8) | 1 (0.4) | 2.0 (0.21–19.21) | 1.19 (0.01–108.07) |
| Venous thromboembolism, n (%) | 2 (0.5) | 0 (0) | — | — |
| DVT | 2 (0.5) | 0 (0) | — | — |
| PTE | 0 (0) | 0 (0) | — | — |
| Maternal death, n (%) | 0 (0) | 0 (0) | — |
a Controlled for age (continuous), nulliparity, oligohydramnios, prior cesarean delivery, pregnancy-related hypertension, chronic hypertension, married, black, and neonatal birthweight (continuous)
b Total complication comprised of operative complications, infections, wound morbidity, and venous thromboembolism
c Comprised of hysterectomy, operative injury, and postpartum hemorrhage.
Although there was no significant difference in the total operative complications between groups (aOR, 0.60; 95% CI, 0.28–1.25), there were lower rates of operative injury from hysterotomy extension into the contractile portion of the uterus in the induction group than in the cesarean group (aOR, 0.19; 95% CI, 0.07–0.54). Other individual operative complications were not significantly different between groups.
In the unadjusted comparison, infections were more common in the induction group (OR, 2.48; 95% CI, 1.49–4.16), largely attributable to increased rates of chorioamnionitis (OR, 19.05; 95% CI, 4.60–78.97). However, after multivariable adjustments, the overall infection rate was not significantly different between groups (aOR, 1.62; 95% CI, 0.75–3.52), but the increased odds of chorioamnionitis in the induction group persisted (aOR, 10.6; 95% CI, 2.05–54.51). None of the other individual infections evaluated were significantly different between groups.
In unadjusted analysis, there was no difference in composite wound morbidity between the 2 groups (OR, 0.77; 95% CI, 0.42–1.40). However, after multivariable adjustments, wound morbidity was less likely in the induction of labor group than in the planned cesarean group (aOR, 0.31; 95% CI, 0.12–0.82). Wound exploration (aOR, 0.13; 95% CI, 0.03–0.57) and wound packing (aOR, 0.17; 95% CI, 0.04–0.69) also were less likely in the induction of labor group. Other individual wound morbidities were not significantly different between groups.
Two women, both of whom had an induction of labor, developed deep vein thrombosis. In the entire cohort, there were no cases of pulmonary thromboembolism or maternal death.
Unadjusted and adjusted analyses for the secondary outcome of composite neonatal morbidity and its components are presented in Table 3 . Between the 2 groups, there was no significant difference in the neonatal morbidity composite (OR, 1.33; 95% CI, 0.82–2.18) or any of its components. After multivariable adjustments for baseline differences, there remained no differences in composite neonatal morbidity (aOR, 0.81; 95% CI, 0.37–1.77) or any of its components.
| Description | Induction of labor, n = 399 | Planned cesarean, n = 262 | OR (95% CI) | aOR (95% CI) a |
|---|---|---|---|---|
| Neonatal composite, n (%) | 53 (13.3) | 27 (10.3) | 1.33 (0.82–2.18) | 0.81 (0.37–1.77) |
| Intrapartum fetal death | 0 (0) | 0 (0) | — | — |
| NICU admission >72 h | 39 (9.8) | 23 (8.8) | 1.12 (0.65–1.93) | 0.77 (0.33–1.83) |
| Mechanical ventilation >24 h | 9 (2.3) | 3 (1.2) | 1.98 (0.53–7.40) | 0.71 (0.09–5.36) |
| Suspected sepsis | 21 (5.3) | 9 (3.4) | 1.56 (0.70–3.47) | 1.0 (0.28–3.43) |
| Culture proven early-onset sepsis | 1 (0.3) | 0 (0) | — | — |
| Neonatal death | 0 (0) | 0 (0) | — | — |
| Grade III/IV IVH | 1 (0.3) | 0 (0) | — | — |
| HIE | 1 (0.3) | 1 (0.4) | 0.65 (0.04–10.49) | — |
| RDS | 14 (3.5) | 9 (3.4) | 1.02 (0.44–2.40) | 0.76 (0.20–2.81) |
| Seizures | 1 (0.3) | 0 (0) | — | — |
| CPR | 1 (0.3) | 0 (0) | — | — |
| NEC | 0 (0) | 0 (0) | — | — |
| Other neonatal outcomes, n (%) | ||||
| 5-min Apgar <7 | 18 (4.5) | 4 (1.5) | 3.05 (1.02–9.11) | 3.26 (0.74–14.32) |
| Umbilical Artery pH <7.1 | 22 (5.7) | 19 (7.4) | 0.75 (0.40–1.42) | 0.63 (0.23–1.72) |
| TTN | 13 (3.3) | 18 (6.9) | 0.46 (0.22–0.95) | 0.70 (0.22–2.25) |
| Hyperbilirubinemia | 28 (7.1) | 15 (5.7) | 1.25 (0.65–2.39) | 1.44 (0.51–4.12) |
| Metabolic disturbances | 32 (8.0) | 20 (7.6) | 1.06 (0.59–1.89) | 1.05 (0.43–2.59) |
a Controlled for age (continuous), nulliparity oligohydramnios, prior cesarean delivery, pregnancy-related hypertension, chronic hypertension, married, black, and neonatal birthweight (continuous).
In the unadjusted analysis, infants in the induction cohort had a significantly increased risk of a 5-minute Apgar score <7 (OR, 3.05; 95% CI, 1.02–9.11) and a decreased risk of transient tachypnea of the newborn (OR, 0.46; 95% CI, 0.22–0.95). However, with multivariable adjustments, these differences no longer were statistically significant—aOR 3.26 (95% CI, 0.74–14.32) and 0.70 (95% CI, 0.22–2.25), respectively. In addition, there were no differences in rates of umbilical artery pH <7.1 or other minor comorbidities (hyperbilirubinemia and other metabolic disturbances) between the 2 groups.
Results of further stratifying the induction of labor group by delivery method (successful vaginal delivery or cesarean delivery) and comparing these patients with those who underwent planned cesarean are presented in Table 4 . The rate of total complications was higher in the failed induction of labor group (45.4%) than in the successful induction group (18.2%) or the planned cesarean group (24.4%) ( P < .001). The failed induction of labor group similarly had the highest rate of operative complications (20.9%), infections (24.5%), and wound morbidity (15.3%) (all P < .001). Moreover, the failed induction group had higher rates of total composite neonatal morbidity as compared with the successful induction and planned cesarean groups (20.9%, 8.1%, and 10.3%, respectively; P = .0003). ORs of both maternal and neonatal morbidity for the induction with cesarean and planned cesarean groups in relation to a referent group of successful induction of labor are similarly presented in Table 4 .
