Objective
We sought to assess neonatal morbidity and mortality of elective cesarean section (CS) of uncomplicated twin pregnancies per week of gestation >35 +0 .
Study Design
We performed a retrospective cohort study in our nationwide database including all elective CS of twin pregnancies. Two main composite outcome measures were defined, ie, severe adverse neonatal outcome and mild neonatal morbidity.
Results
We report on 2228 neonates. More than 17% were born <37 +0 weeks of gestation. Adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for severe adverse neonatal outcome at 35 +0-6 , 36 +0-6 , and 37 +0-6 weeks were, OR, 9.4; 95% CI, 3.2–27.6; OR, 1.7; 95% CI, 0.5–5.3; and OR, 0.7; 95% CI, 0.2–2.0, respectively; and for mild neonatal morbidity, OR, 4.7; 95% CI, 2.6–8.7; OR, 4.9; 95% CI, 3.1–7.9; and 1.4; 95% CI, 0.9–2.1, respectively, compared to neonates born ≥38 +0 weeks of gestation.
Conclusion
In uncomplicated twin pregnancies elective CS can best be performed between 37 +0 and 39 +6 weeks of gestation.
The incidence of twin pregnancies has increased worldwide, especially in Western countries. The main cause is the use of assisted reproductive technology. Secondly there is an increase in spontaneous twin pregnancies associated with increasing maternal age and ethnicity. In The Netherlands, the incidence of cesarean section (CS) with twin pregnancies rose from 26.1% in 1993 to 36.9% in 2007.
In singleton pregnancies, it is known that the risk for respiratory morbidity is significantly higher after an elective (planned) cesarean delivery compared to a planned vaginal delivery and that this risk diminishes significantly with advancing gestational age until week 39 +0 . The definition of term pregnancy is being debated, suggesting introduction of new subcategories of term births (early term vs full term). In twin pregnancies, however, this discussion is more complicated. Most literature with regard to timing of twin deliveries is focused on the risk of intrauterine fetal demise with an increase of gestational age. However, no consensus is found on when this risk starts increasing, from ≥36-39 weeks. The risks for neonatal mortality and morbidity caused by iatrogenic preterm birth by elective CS of twin pregnancies are unclear as well. The aim of our study was to assess neonatal morbidity and mortality of elective CS for uncomplicated twin pregnancies per week of gestation >35 +0 .
Materials and Methods
Data were extracted from our nationwide database, The Netherlands Perinatal Registry (PRN), which includes 96% of approximately 180,000 deliveries per year at ≥16 completed weeks of gestation. The neonatal follow-up is registered in the PRN for approximately 68% of all hospitals. For more details with regard to content, process, and quality control of the data collection we refer to a previous article on singleton pregnancies. The current study was approved by the board of the PRN.
For this study, data from the PRN concerning 54,082 liveborn neonates of twin pregnancies from Jan. 1, 2000, through Dec. 31, 2007, were analyzed. We only included neonates born by an elective CS >35 +0 weeks of gestation and excluded neonates born by a planned CS registered with a maternal and/or fetal indication or born by an emergency CS. Also all twins of which 1 fetus was missing in the registration were excluded. Eventually we excluded all twins of which at least 1 fetus had a congenital anomaly, and twins of mothers with an adverse medical or obstetric history and/or a complication of pregnancy that could influence the risk for neonatal morbidity ( Figure ). The study was limited to hospitals that systematically registered neonatal follow-up.
Calculation of gestational age was, according to national guidelines, based on the date matching with assisted reproductive technology or based on the first day of the last menstrual period and verified by a first-trimester ultrasound; in case of discrepancy, gestational age was determined by the results of the first-trimester ultrasound. A first-trimester ultrasound is part of routine obstetric care, which is available for everyone in The Netherlands. Socioeconomic status was based on the mean household income level of the neighborhood, which was determined by the first 4 digits of the woman’s postal code. “Light” for gestational age was defined as a birthweight <10th percentile; “heavy” for gestational age was defined as a birthweight >90th percentile. In absence of validated growth curves specific for multiple pregnancies this was based on sex-, parity-, and race-specific growth curves developed for singleton pregnancies.
Outcome measures
We studied 2 main composite outcome measures and 2 additional outcome measures. Our first main outcome was severe adverse neonatal outcome, defined as a composite measure of neonatal mortality until the 28th day, or neonatal morbidity: a 5-minute Apgar score <4, convulsions, intracranial hemorrhage, respiratory morbidity registered as pneumothorax or respiratory distress syndrome, respiratory support by intermittent positive pressure ventilation, severe resuscitation (defined as endotracheal artificial ventilation and/or administration of buffers), and/or sepsis, including both clinically suspected patients as well as proven infections with positive cultures.
Secondly, we studied mild neonatal morbidity, a composite of: respiratory morbidity registered as transient tachypnea of the newborn, respiratory support with continuous positive airway pressure or oxygen, or hypoglycemia (defined as a serum or plasma glucose level of <2.5 mmol/L). We also analyzed the parts of these composite outcome measures separately.
Our 2 additional outcome measures were admission to the neonatal intensive care unit (NICU) and admission to any neonatal ward for ≥5 days. Follow-up of neonates stopped at discharge from the hospital. If they were transferred to another hospital (eg, a university hospital), follow-up was continued.
Statistical analysis
All analyses were performed using software (SAS, version 9.1; SAS Institute, Cary, NC).
To evaluate differences in the baseline characteristics we used analysis of variance for continuous variables and the χ 2 test for categorical variables. Because several categorical variables have small numbers in the gestational age categories we also approximated the exact P value, using Monte Carlo simulations (n = 100,000). Probability values < .05 were considered statistically significant. The incidence of neonatal outcome was calculated for each week of gestation. The presence of trends in our outcome data was studied with an exact trend analysis (Cochran-Armitage). As we expected a decreasing morbidity with an increasing gestational age we used 1-sided P values with a statistical significance level of .025. Missing values occurred for 0.5% of all confounders and were imputed once for the regression analysis using R software (The R Foundation, Vienna, Austria). Univariate and multivariate logistic regression analyses were used to study the association between gestational age at delivery and neonatal outcomes with all elective CS performed >38 +0 weeks of gestation as a reference group. The odds ratio with 95% confidence interval was determined using generalized estimation equations for marginal models. The generalized estimation equations were used to account for the dependency of the twin children from the same pregnancy by using the mothers as a cluster variable. We adjusted the association for maternal age, ethnicity, parity, socioeconomic status, and fetal sex.
Two sensitivity analyses were performed. Since chorionicity is not registered in our database, we analyzed all twins with an unequal sex to validate the results for dichorionic twins. Secondly we repeated the regression analyses in which births with an uncertain gestational age (2.8%) were excluded.
To analyze the incidence of intrauterine fetal demise we selected a separate cohort of all twin pregnancies from 2000 through 2007 with a gestational age >35 +0 weeks, without congenital anomalies, independent of the (intended) mode of delivery. We measured the absolute incidence and calculated the incidence per 1000 fetus still in utero as presented by Yudkin et al.
Results
In the study period, 54,082 liveborn neonates of twin pregnancies were registered in the PRN. From this population, a study cohort was comprised ( Figure ). We excluded 12,448 neonates born <35 +0 weeks of gestation. We also excluded neonates of vaginal birth (n = 26,657), emergency CS (n = 7030), or missing mode of delivery (n = 66). In total 7881 liveborn neonates between 35 +0 and 41 +6 weeks of gestation were born by planned CS, 4577 electively. Furthermore we excluded 157 neonates because 1 neonate of the twin was not completely present in the registration. We also excluded all twins of which at least 1 neonate had a congenital anomaly (n = 103 twins, 206 neonates) and all twins with a mother with an adverse medical history or complication of pregnancy that could have affected the risk for neonatal morbidity ( Figure ) (n = 180 twins, 360 neonates). Finally, 1606 neonates had no registered neonatal follow-up. We therefore report on 2228 neonates of twin pregnancies born by an elective CS >35 +0 weeks of gestation.
Between 35 +0 and 35 +6 weeks of gestation, 104 (4.7%) of these neonates were born; 290 (13.0%) between 36 +0 and 36 +6 weeks, 984 (44.2%) between 37 +0 and 37 +6 weeks, and 850 (38.1%) between 38 +0 and 41 +6 weeks. Maternal and neonatal characteristics are presented in Tables 1 and 2 , respectively. Primiparous women tended to deliver premature more often than multiparous women. As can be expected, the mean birthweight increased with an increasing gestational age at delivery. As can be expected, the mean birthweight increased with an increasing gestational age at delivery; however, neonates <10th percentile tended to be delivered later.
| Characteristic | n (%) | n (%) | n (%) | n (%) | P value |
|---|---|---|---|---|---|
| Week of gestation | 35 +0-6 | 36 +0-6 | 37 +0-6 | 38 +0 -41 +6 | |
| Number of mothers | n = 52 | n = 145 | n = 492 | n = 425 | |
| (4.7%) | (13.0%) | (44.2%) | (38.1%) | ||
| Age at delivery, y | |||||
| Mean ± SD | 31.4 ± 5.0 | 32.2 ± 4.1 | 32.0 ± 4.4 | 32.4 ± 4.7 | .09 b |
| 0-35 | 38 (73.1) | 98 (67.6) | 347 (70.5) | 272 (64.0) | .02 c |
| >35 | 14 (26.9) | 47 (32.4) | 145 (29.5) | 153 (36.0) | |
| Race or ethnic group | |||||
| Western | 48 (92.3) | 132 (91.7) | 429 (87.6) | 366 (86.7) | .12 c |
| Other | 4 (7.7) | 12 (8.3) | 61 (12.4) | 56 (13.3) | |
| Missing | 0 | 1 | 2 | 3 | |
| Parity | |||||
| Primiparae | 30 (57.7) | 65 (44.8) | 228 (46.3) | 201 (47.3) | .14 c |
| Multiparae | 22 (42.3) | 80 (55.2) | 264 (53.7) | 224 (52.7) | |
| Socioeconomic status | |||||
| Very high | 10 (19.2) | 23 (16.3) | 91 (19.0) | 83 (19.8) | .58 c |
| High | 11 (21.2) | 33 (23.4) | 105 (21.9) | 84 (20.0) | |
| Normal | 7 (13.5) | 28 (19.9) | 90 (18.8) | 76 (18.1) | |
| Low | 13 (25.0) | 26 (18.4) | 82 (17.1) | 69 (16.5) | |
| Very low | 11 (21.1) | 31 (22.0) | 112 (23.3) | 107 (25.5) | |
| Missing | 0 | 4 | 12 | 6 |
a Numbers are presented as the number of mothers;
| Characteristic | n (%) | n (%) | n (%) | n (%) | P value |
|---|---|---|---|---|---|
| Week of gestation | 35 +0-6 | 36 +0-6 | 37 +0-6 | 38 +0 -41 +6 | |
| Number of neonates | n = 104 | n = 290 | n = 984 | n = 850 | |
| (4.7%) | (13.0%) | (44.2%) | (38.1%) | ||
| Sex | |||||
| Male | 51 (49.0) | 140 (48.3) | 484 (49.2) | 392 (46.1) | .61 b |
| Female | 53 (51.0) | 150 (51.7) | 500 (50.8) | 458 (53.9) | |
| Position | |||||
| Vertex | 51 (49.0) | 104 (35.9) | 300 (30.6) | 298 (35.1) | .003 b |
| Breech | 48 (46.2) | 164 (56.6) | 570 (58.2) | 471 (55.4) | |
| Other | 5 (4.8) | 22 (7.6) | 110 (11.2) | 81 (9.5) | |
| Missing | 0 | 0 | 4 | 0 | |
| Birthweight, g | |||||
| Mean ± SD | 2304 ± 334 | 2521 ± 379 | 2700 ± 375 | 2878 ± 401 | < .0001 c |
| <2500 | 69 (66.4) | 141 (48.6) | 281 (28.6) | 139 (16.4) | < .0001 b |
| Light for gestational age (<p10) d | 20 (19.2) | 57 (19.7) | 241 (24.5) | 286 (33.7) | < .0001 b |
| Heavy for gestational age (>p90) d | 0 (0.0) | 4 (1.4) | 8 (0.8) | 6 (0.7) | .57 e |
a Numbers are presented as the numbers of neonates;
d Derived from sex-, parity-, and race-specific growth curves;
e Approximation of the exact P value with Monte Carlo simulations.
Incidence rates for all outcome measures are presented per week of gestation in Table 3 . The absolute risks for severe adverse neonatal outcome were 8.7% between 35 +0 and 35 +6 , 1.7% between 36 +0 and 36 +6 , and 0.7% between 37 +0 and 37 +6 weeks compared with 1.1% between 38 +0 and 41 +6 weeks of gestation ( P for trend exact < .0001). For mild neonatal morbidity the absolute risks were 22.1% between 35 +0 and 35 +6 , 22.1% between 36 +0 and 36 +6 , and 7.6% between 37 +0 and 37 +6 weeks compared to 5.5% between 38 +0 and 41 +6 weeks of gestation ( P for trend exact < .0001). Admission to the NICU demonstrated risks of 4.8% between 35 +0 and 35 +6 , 1.0% between 36 +0 and 36 +6 , and 0.5% between 37 +0 and 37 +6 weeks compared with 0.2% between 38 +0 and 41 +6 weeks of gestation ( P for trend exact < .0001). Admission to any neonatal ward demonstrated risks of 60.6% between 35 +0 and 35 +6 , 36.9% between 36 +0 and 36 +6 , and 19.2% between 37 +0 and 37 +6 weeks compared with 15.3% between 38 +0 and 41 +6 weeks of gestation ( P for trend exact < .0001). Compared to neonates born between 38 +0 and 41 +6 weeks of gestation, neonates born between 35 +0 and 35 +6 weeks were at significantly higher risk for all our outcome measures and between 36 +0 and 36 +6 weeks at significantly higher risk for mild neonatal morbidity and hospitalization ≥5 days. Between 37 +0 and 37 +6 weeks of gestation there were no significantly higher risks ( Table 4 ). After excluding all neonates with equal sex (n = 1278), similar results were found (results not shown).
| Outcome | n (%) | n (%) | n (%) | n (%) | P for trend exact b |
|---|---|---|---|---|---|
| Week of gestation | 35 +0-6 | 36 +0-6 | 37 +0-6 | 38 +0 -41 +6 | |
| Number of neonates | n = 104 | n = 290 | n = 984 | n = 850 | |
| (4.7) | (13.0) | (44.2) | (38.1) | ||
| Severe adverse neonatal outcome c | 9 (8.7) | 5 (1.7) | 7 (0.7) | 9 (1.1) | < .0001 |
| Neonatal death <24 h | 0 (0.0) | 0 (0.0) | 1 (0.10) | 2 (0.24) | .25 |
| Neonatal death day 2-7 | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | N/A |
| Neonatal death day 8-28 | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | N/A |
| Apgar score <4 d | 1 (1.0) | 0 (0.0) | 1 (0.1) | 1 (0.1) | .23 |
| Convulsions | 0 (0.0) | 0 (0.0) | 2 (0.2) | 0 (0.0) | .52 |
| Intracranial hemorrhage | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | N/A |
| Pneumothorax | 0 (0.0) | 2 (0.7) | 0 (0.0) | 0 (0.0) | .07 |
| Respiratory distress syndrome | 4 (3.9) | 4 (1.4) | 1 (0.1) | 1 (0.1) | .002 |
| IPPV | 1 (1.0) | 1 (0.3) | 2 (0.2) | 0 (0.0) | .04 |
| Severe resuscitation e | 1 (1.0) | 1 (0.3) | 2 (0.2) | 2 (0.2) | .23 |
| Sepsis | 4 (3.9) | 0 (0.0) | 3 (0.3) | 3 (0.4) | .01 |
| Mild neonatal morbidity f | 23 (22.1) | 64 (22.1) | 75 (7.6) | 47 (5.5) | < .0001 |
| Transient tachypneu of the newborn | 4 (3.9) | 12 (4.1) | 21 (2.1) | 10 (1.2) | .002 |
| CPAP | 4 (3.9) | 5 (1.7) | 3 (0.3) | 1 (0.1) | < .0001 |
| Oxygen | 10 (9.6) | 14 (4.8) | 15 (1.5) | 11 (1.3) | < .0001 |
| Hypoglycemia | 13 (12.5) | 45 (15.5) | 48 (4.9) | 30 (3.5) | < .0001 |
| Admission | |||||
| To the NICU | 5 (4.8) | 3 (1.0) | 5 (0.5) | 2 (0.2) | < .0001 |
| To any neonatal ward ≥5 days | 63 (60.6) | 107 (36.9) | 189 (19.2) | 130 (15.3) | < .0001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
