The relationship between delivery mode and mortality in very low birthweight singleton vertex-presenting infants


Dr A. Riskin, Department of Neonatology, Bnai Zion Medical Centre, 47 Golomb Street, P.O.B. 4940, Haifa 31048, Israel.


Objective  To investigate the factors associated with caesarean delivery and the relationship between mode of delivery and mortality in singleton vertex-presenting very low birthweight (≤1500 g) live born infants.

Design  Observational population-based study.

Setting  Data collected from all 28 neonatal departments comprise the Israel National Very Low Birth Weight Infant Database.

Population  2955 singleton vertex-presenting very low birthweight infants registered in the database from 1995 to 2000, and born at 24–34 weeks of gestation.

Methods  The demographic, obstetric and perinatal factors associated with caesarean delivery and subsequent mortality were studied. The independent effect of the mode of delivery on mortality was tested by multiple logistic regression.

Main outcome measure  Mortality was defined as death prior to discharge.

Results  Caesarean delivery rate was 51.7%. Caesarean delivery was directly associated with increasing maternal age and gestational age, small for gestational age infants, maternal hypertensive disorders and antepartum haemorrhage, and was inversely related to premature labour and prolonged rupture of membranes. Factors associated with increased survival were increasing gestational age, antenatal corticosteroid therapy, maternal hypertensive disorders and no amnionitis. Mortality rate prior to discharge was lower after caesarean delivery (13.2%vs 21.8%), but in the multivariate analysis, adjusting for the other risk factors associated with mortality, delivery mode had no effect on infant survival (OR 1.00, 95% CI 0.74–1.33). In a subgroup with amnionitis, a protective effect of caesarean delivery was found.

Conclusions  Caesarean delivery did not enhance survival of vertex-presenting singleton very low birthweight babies. Caesarean delivery cannot be routinely recommended, unless there are other obstetric indications.


The optimal mode of delivery for very low birthweight infants is controversial. Claims that elective caesarean delivery reduces the chances of fetal or neonatal death and chronic morbidity have been met by counter claims that such a policy leads to increased risks of serious maternal morbidity and mortality with no benefit to the infant.1–11 However, data from prospective randomised studies are very limited due to recruitment difficulties.11–13 This dilemma is even greater at borderline viability at less than 26 weeks of gestation or birthweight <750 g that is associated with very high mortality rates and serious morbidities among survivors.14 Furthermore, in order to achieve an atraumatic delivery, elective caesarean section might require a vertical classic incision with long term influence on the obstetric future of the pregnant patient.15

In practice, however, the rate of elective caesarean deliveries in very low birthweight has markedly increased over the last decades.11,16–18 This further strengthens the need to try and determine whether this practice of elective caesarean deliveries is justified for a possibly better outcome of the infants, in face of potential serious morbidities among the mothers. The general clinical research question that awaits answer is: What is the optimal mode of delivery of very low birthweight infants?

This population-based observational study specifically aimed to determine the association between the mode of delivery (caesarean section vs vaginal delivery) and the rate of death prior to discharge in singleton vertex-presenting premature very low birthweight infants. In order to do this, we also studied the demographic and perinatal factors associated with caesarean delivery of these infants.


This study is based on analysis of data collected by the Israel Neonatal Network data on very low birthweight infants (birthweight ≤1500 g) born in Israel from 1995 to 2000. All 28 neonatal departments in Israel are included in data collection, which comprises the Israel National Very Low Birthweight Infant Database. Data are prospectively collected on a prestructured form and include information on the parents, maternal pregnancy history, antenatal care, details of the delivery, the infant's status at delivery, diagnoses, procedures and complications during hospital stay, and outcome at discharge. All live born infants in Israel receive a unique identification number at birth. Patient information received by the database coordinator is checked for missing items and logic errors, corrected, completed and then entered into a computerised database. Patient information is cross-checked with the national birth registry, and data from any missing infants are requested from the birth hospital. The database included 99% of all live born very low birthweight infants in Israel. Hospital of birth and patient identification remain confidential by consensus agreement of all participating centres. Data are collected on all infants until discharge home or death.

Data are collected by the neonatologists in each perinatal centre. All departments use operating manual and standard definitions, which were defined by the scientific committee before data collection and have remained unaltered since. Gestational age (in completed weeks) was determined by the best obstetric estimate of gestational age based on last menstrual period, obstetric history and examination, prenatal ultrasound and postnatal physical examination. Small for gestational age was defined as birthweight more than two standard deviations below the mean weight for gestational age, according to the intrauterine growth charts of Usher and McLean.19 Antenatal steroid treatment was defined as ‘partial’ if delivery occurred less than 24 hours after the first dose or more than one week after the last dose, and ‘complete’ if delivery occurred more than 24 hours and less than seven days after a complete course of treatment. Premature rupture of membranes was defined as membrane rupture more than 6 hours before the onset of regular spontaneous uterine contractions. Diagnosis of amnionitis was based on high maternal temperature (>37.8°C orally or 38°C rectally) recorded twice in 1 hour, during the rupture of membranes or during the first 6 hours after delivery, provided no other cause for the fever was found. Preterm labour was considered in the presence of uterine contractions occurring prior to 35 weeks of gestation together with cervical effacement and dilatation. Maternal hypertensive disorder was defined as either chronic hypertension (persistent elevation of blood pressure before 20 weeks of gestation or prior to pregnancy) or pregnancy-induced hypertension if blood pressure above 145/95 was first recorded after 20 weeks of gestation. Mortality was defined as death occurring before discharge from hospital.

During the six-year period, 8695 very low birthweight infants were registered in the database, representing 1.1% of the 761,000 live births in Israel. From this population, infants were excluded for the following reasons: 3732 were multiple births, 393 were delivered at less than 24 weeks or greater than 34 weeks of gestation, 1104 were non-vertex presentation, in 468 infants the presentation was not recorded, and 43 had lethal congenital anomalies. Thus, the final study population comprised 2955 singleton very low birthweight infants, of 24 to 34 weeks of gestation, of whom 1428 were delivered vaginally and 1527 by caesarean section.

The following perinatal and neonatal variables were included in the analyses: maternal age, ethnicity, fertility treatment, prenatal care, gestational age at beginning of prenatal care, pregnancy complications including preterm labour, premature rupture of membranes, amnionitis, antepartum vaginal haemorrhage due to either placenta praevia or placental abruption, and maternal hypertensive disorder, antenatal steroid treatment, delivery mode, gestational age at delivery in weeks, birthweight, small for gestational age and gender.

Logistic regression models were used to calculate both the unadjusted (univariate) and adjusted (multivariate) risk for caesarean section rate and for death prior to discharge. On completion of the univariate analyses, any variable for which the univariate test had a P value <0.15 was considered for inclusion in the multivariate model. The logistic regression models were used to assess the odds ratios (OR) for two outcomes: (1) mode of delivery—to determine factors associated with caesarean section, and (2) death prior to discharge—to assess the independent effect of mode of delivery on mortality adjusting for all other confounding variables. Because birthweight and gestational age were highly correlated (r= 0.7), birthweight was not entered into the multiple logistic regression analysis. Results of the logistic models are presented as unadjusted and adjusted OR with 95% confidence intervals (CI). The SAS statistical program (version 8.2; SAS Institute, Cary, North Carolina) was used for analysis of the data.

In a secondary analysis, the effect of mode of delivery on death prior to discharge was considered in the presence of five groups of pregnancy complications: maternal hypertensive disorders only (n= 625); premature labour only (n= 714); antepartum haemorrhage, with or without premature labour (n= 369); premature rupture of membranes, with or without premature labour (n= 388); and amnionitis, with or without premature rupture of membranes and/or premature labour (n= 240). A total of 619 subjects were not included in this analysis. In 319 various combinations of these conditions were reported and in 300 patients none of the five conditions were present.


The overall caesarean section rate was 51.7%, and it was directly related to gestational age (Fig. 1) increasing from 19.1% at 24 weeks, to 53.9% at 29 weeks, and 69.3% at 34 weeks of gestation (P < 0.001 for linear trend). The caesarean section rate increased with increasing maternal age, and was significantly higher in Jewish mothers and following infertility treatment. Almost 90% of mothers with hypertensive disorders were delivered by caesarean section. Conversely, caesarean section rates were much lower among mothers with premature contractions (26.4%), premature rupture of membranes (34.4%) and amnionitis (36.6%). The unadjusted and adjusted OR for caesarean delivery are shown in Table 1. In the multivariate logistic regression analysis, delivery by caesarean section among vertex-presenting very low birthweight singleton infants was independently associated with increasing maternal age (OR = 1.24 for each increment of five years), maternal hypertensive disorders (OR = 4.24), antepartum haemorrhage (OR = 2.11), increasing gestational age (OR = 1.12 for each increase of one week in gestational age) and small for gestational age infants (OR = 1.98). The rate of caesarean deliveries was significantly decreased in mothers with preterm labour (OR = 0.20) and premature rupture of membranes (OR = 0.53).

Figure 1.

Relationship between gestational age, caesarean delivery rate and the rate of death prior to discharge according to delivery mode, in very low birthweight singleton vertex-presenting infants.

Table 1.  Unadjusted and adjusted odds ratio for factors associated with caesarean section delivery among very low birthweight singleton vertex-presenting infants.
Risk factorUnadjusted OR (95% CI)Adjusted OR (95% CI)
Maternal age (increment of 5 years)1.30 (1.22–1.38)1.24 (1.15–1.34)
Ethnicity (non-Jews vs Jews)0.69 (0.59–0.81)0.90 (0.73–1.11)
Fertility treatment2.05 (1.58–2.67)1.31 (0.95–1.82)
Week starting prenatal care
13–200.82 (0.66–1.02)0.88 (0.67–1.16)
>20 or no prenatal care0.60 (0.46–0.78)0.70 (0.50–0.99)
Maternal hypertensive disorders11.1 (8.9–13.9)4.24 (3.26–5.56)
Preterm contractions0.11 (0.09–0.13)0.20 (0.16–0.24)
Premature rupture of membranes >6 hours0.39 (0.33–0.46)0.53 (0.41–0.67)
Amnionitis0.50 (0.39–0.64)1.24 (0.91–1.69)
Antepartum haemorrhage1.21 (1.01–1.46)2.11(1.68–2.66)
Intrauterine growth (small for gestational age)4.60 (3.80–5.60)1.98 (1.54–2.54)
Antenatal steroid therapy
Full vs none1.20 (1.02–1.41)1.12 (0.91–1.38)
Partial vs none0.82 (0.66–1.01)0.75 (0.57–0.99)
Gestational age (increase of 1 week)1.26 (1.22–1.30)1.12 (1.08–1.17)

The overall mortality rate until discharge was 17.4%. The mortality rate was 13.2% for infants delivered by caesarean section compared with 21.8% among infants delivered vaginally (P < 0.001). The mortality rates decreased significantly with increasing gestational age, regardless of delivery mode (Fig. 1). The risk of mortality among infants delivered by caesarean section compared with vaginal deliveries after adjustment for gestational age only was 1.13 (95% CI 0.90–1.42). The unadjusted and adjusted OR for mortality prior to discharge are shown in Table 2. In the univariate analysis (unadjusted OR), factors significantly associated with excess mortality included non-Jewish ethnicity, no prenatal care, preterm labour, amnionitis, antepartum haemorrhage and decreasing birthweight and gestational age. Lower mortality occurred with maternal hypertensive disorders, and among infants who received antenatal steroid therapy. In the multivariate model, delivery by caesarean section had no effect on mortality (OR = 1.00, 95% CI 0.74–1.33). Additional risk factors for excess mortality were decreasing gestational age (OR = 1.79 for each week decrease), small for gestational age infants (OR = 4.91), presence of amnionitis (OR=1.66) and late or no prenatal care (OR = 1.89), whereas maternal hypertensive disorder and antenatal steroid therapy had a protective effect on mortality. The addition of birthweight to this model showed that both birthweight (OR = 1.48 for each 100 g decrease) and gestational age (OR = 1.25 for each week decrease) were significantly associated with mortality. In this model too, caesarean delivery did not have a significant effect on mortality (OR = 0.87, 95% CI 0.64–1.17), but furthermore small for gestational age was not significant (OR = 1.43, 95% CI 0.92–2.33). This model however did not fit well. Multivariate analyses in three stratified gestational age groups (24–27, 28–30 and 31–34 weeks) showed a similar pattern (data not shown).

Table 2.  Unadjusted and adjusted odds ratio for delivery mode and other risk factors associated with mortality prior to discharge among very low birthweight singleton vertex-presenting infants.
Risk factorUnadjusted OR (95% CI)Adjusted OR (95% CI)
Delivery mode (caesarean section vs vaginal)0.54 (0.45–0.66)1.00 (0.74–1.33)
Maternal age (5 years increase)0.98 (0.91–1.06)Excluded
Ethnicity (non-Jews vs Jews)1.41 (1.15–1.71)1.12 (0.87–1.43)
Fertility treatment0.71 (0.49–1.00)0.84 (0.53–1.28)
Week starting prenatal care
13–201.09 (0.82–1.45)1.12 (0.79–1.58)
>20 or no prenatal care1.66 (1.21–2.25)1.89 (1.27–2.79)
Maternal hypertensive disorders0.41 (0.32–0.53)0.57 (0.39–0.82)
Preterm contractions1.60 (1.32–1.94)0.91 (0.69–1.21)
Premature rupture of membranes >6 hours1.03 (0.82–1.28)Excluded
Amnionitis2.29 (1.75–2.99)1.66 (1.18–2.33)
Antepartum haemorrhage1.42 (1.13–1.77)1.14 (0.86–1.50)
Intrauterine growth (small for gestational age)1.17 (0.94–1.45)4.91 (3.54–6.86)
Antenatal steroid therapy
Full vs no0.45 (0.36–0.55)0.50 (0.38–0.64)
Partial vs no0.48 (0.35–0.64)0.40 (0.28–0.56)
Gestational age (decrease of 1 week)1.70 (1.62–1.79)1.79 (1.69–1.90)
Gender (male vs female)1.14 (0.94–1.38)1.10 (0.87–1.38)

Pregnancy complications associated with preterm delivery greatly influence delivery mode and neonatal outcome. Thus, we performed a secondary analysis of the rate of death prior to discharge by mode of delivery, according to five selected groups of pregnancy complications. Two multiple logistic regression models were used: the first adjusted for gestational age only, and a second model adjusted for gestational age, small for gestational age and antenatal steroid treatment (Table 3). Caesarean delivery rates ranged from 19.2% in patients with premature labour only to 91.7% in the presence of hypertensive disorders. These analyses further supported the finding that the mode of delivery had no significant effect on subsequent mortality of very low birthweight infants during their hospitalisation. However, in the subgroup of patients in whom a diagnosis of amnionitis was reported, there was a protective effect of caesarean delivery on infant mortality (OR 0.30, 95% CI 0.11–0.77).

Table 3.  Unadjusted and adjusted odds ratios (95% CI) for mortality among caesarean section versus vaginally delivered very low birthweight singleton vertex-presenting infants in selected pregnancy complications.
Pregnancy complicationnCaesarean section (%)Mortality
UnadjustedAdjusted for gestational ageAdjusted for gestational age, small for gestational age and antenatal steroids
Hypertensive disorders only62591.72.44 (0.73–15.20)1.91 (0.46–13.55)1.68 (0.38–12.01)
Premature labour only71419.20.76 (0.46–1.20)1.71 (0.97–2.96)1.34 (0.72–2.44)
Antepartum haemorrhage (±premature labour)36951.20.83 (0.51–1.35)1.49 (0.84–2.64)1.15 (0.62–2.17)
Premature rupture of membranes (±premature labour)38827.61.47 (0.75–2.79)1.60 (0.80–3.13)1.58 (0.77–3.16)
Amnionitis (±premature rupture of membranes ±premature labour)24032.90.23 (0.10–0.48)0.46 (0.19–1.03)0.30 (0.11–0.77)


The main finding in this study is that delivery mode did not affect neonatal mortality of singleton vertex-presenting preterm very low birthweight infants when other perinatal risk factors for neonatal mortality were controlled. Jonas and her co-investigators similarly suggested, on the basis of their population-based studies both in Washington State, USA8 and in Victoria, Australia,6 that caesarean delivery did not enhance the neonatal survival of vertex-presenting very low birthweight babies. This is further supported by the works of Malloy et al.3 and others.2,4,5,9 We also found that decreased gestational age, small for gestational age and amnionitis were associated with excess neonatal mortality, whereas maternal hypertensive disorders and antepartum corticosteroid use were protective, as previously reported.10,14,20–24

One of the advantages of the database used in this study was the ability to use gestational age at delivery and not just birthweight, as gestational age is considered a better predictor of neonatal survival than birthweight. Gestational age estimation in our population is generally accurate due to early prenatal care in the majority of patients (71% before 12 weeks and 86% before 20 weeks) and the widespread use of prenatal ultrasound.

Delivery mode of a vertex-presenting preterm infant depends on many complex factors. In patients with pre-eclampsia when maternal condition dictates delivery, in the absence of labour, many will undergo elective caesarean section. On the other hand, most of the women in preterm labour or with premature rupture of membranes deliver vaginally, unless there is evidence of fetal distress mandating an emergency caesarean section. Thus, in this work, we also performed subgroup analysis of the risk for death prior to discharge after delivery by caesarean section, according to maternal pregnancy complications that are associated with preterm delivery. Overall, this analysis further supported the finding that the mode of delivery had no effect on subsequent mortality of very low birthweight infants during their hospitalisation.

Based on our results, it may be argued that in the presence of amnionitis, caesarean delivery may have had some protective effect, although the group was relatively small. This may be related to the length of exposure to infection and inflammatory mediators, which may affect the outcome of the newborn.25 However, data in the literature are contradictory,26 and caesarean section is also associated with increased morbidity to the mother,27 especially when amnionitis is present. When there is clinical evidence of amnionitis, immediate delivery is indicated, however, the mode of delivery, caesarean section versus induction of vaginal delivery, is based on obstetric considerations. Our finding about the possible protective effect of caesarean delivery in the presence of amnionitis, although interesting, cannot currently justify change in this clinical practice, unless supported by further studies.

For the sake of uniformity, we limited our analysis to singleton vertex-presenting neonates, because mode of delivery and outcome in breech presentation and multiple gestation might be influenced by other factors.8,28,29 The type of breech (frank vs non-frank) and the experience of the physician in vaginal breech deliveries influence the mode of delivery of a preterm breech-presenting neonate. Also, during delivery, these infants are at an increased risk for birth asphyxia due to cord prolapse or head entrapment. Although the Term Breech Trial30 did not state that its results should be applied to preterm breech infants, in practice, many centres moved to elective caesarean delivery of all breech-presenting preterm infants when delivery is indicated.11 In multiple gestation, the mode of delivery is influenced by the presentation of both infants and there is a tendency toward worse outcome in the second-born twin. Furthermore, we excluded very low birthweight infants born at ≥35 weeks as they represented a group of severe intrauterine growth restricted infants with different morbidities than preterm appropriately grown very low birthweight infants.

The optimal mode of delivery for very low birthweight infants has not yet been fully ascertained. The best way to try and answer this question would be an appropriately powered randomised controlled trial. However, it is almost two decades since the randomised studies of Wallace et al.1 and Lumley et al.31 tried to compare caesarean delivery with expectant management of cephalic presenting immature infants. Both studies failed to reach any conclusion due to recruitment problems. Data in the literature are so far inconclusive and associated with problems of randomisation, selection criteria and the effects of multiple confounding variables.12,13,27 Conducting an appropriately controlled study designed to answer this important clinical issue should still be the ultimate goal. However, it seems increasingly unlikely that such a study will be undertaken and completed in the near future.

In the absence of such evidence, we undertook a multivariate analysis of observational data based on a population of almost all very low birthweight infants born in Israel. During the study period, 1995–2000, recent technologies and treatments including antenatal corticosteroids and surfactant were widely used. The data were collected in a uniform, consistent and reliable database with the use of standard definitions by all participating centres. The multivariate analysis attempted to control for confounding variables, such as the fact that the mode of delivery may be influenced by the indication, such as hypertension, bleeding or preterm labour, which our methodology attempted to even out.

A number of limitations however need to be considered. Firstly, because this was an observational study, some unknown biases and confounding variables may not be accounted for. The specific circumstances influencing the obstetric decision-making process for the individual deliveries, and factors that might have led the obstetrician to perform a caesarean section, are unknown to us. The presence of, for example, fetal distress, cord prolapse, the stage of labour at the time of admission and other complications, may influence both the management and outcome of these pregnancies. Furthermore, data collection, as much as we tried to avoid it, might still have been incomplete, misclassified or subject to temporal changes in data input or clinical practice in any of the participating hospitals of the network. Despite these reservations, in the absence of a prospective randomised control study, which may not be feasible, we believe that these are the best available data.


Based on our findings, we reject the assumption that caesarean section may improve the survival of vertex-presenting very low birthweight singleton immature babies. Prematurity is the primary cause of death of very low birthweight infants. The obstetrician may aid the preterm fetus by giving the mother steroids to accelerate fetal lung maturity. Our data are consistent with the hypothesis that mode of delivery does not influence mortality in very low birthweight infants, if there are no obstetric indications favouring a particular mode. We believe that caesarean section cannot be recommended as the routine mode of delivery for vertex-presenting singleton preterm very low birthweight babies, unless there are other recognised maternal or fetal indications. Our results do suggest that in the presence of amnionitis, caesarean delivery may be associated with a decreased mortality among these infants. However, all these conclusions are guarded due to the limits of the methodology employed.


The Israel National Very-Low-Birth-Weight Infant Database is funded by the Israel Centre for Disease Control and the Ministry of Health.

The Israel Neonatal Network participating centres in the Israel National VLBW infant database are

Assaf Harofeh Medical Centre, Rishon Le Zion; Barzilay Medical Centre, Ashkelon; Bikur Holim Hospital, Jerusalem; Bnai Zion Medical Centre, Haifa; Carmel Medical Centre, Haifa; English (Scottish) Hospital, Nazareth; French Hospital, Nazareth; Hadassah University Hospital Ein-Karem, Jerusalem; Hadassah University Hospital Har Hazofim, Jerusalem; Haemek Medical Centre, Afula; Hillel Yafe Medical Centre, Hadera; Italian Hospital, Nazareth; Kaplan Hospital, Rehovot; Laniado Hospital, Netanya; Maayanei Hayeshua Hospital, Bnei-Brak; Meir Medical Centre, Kefar Saba; Misgav Ladach Hospital, Jerusalem; Naharia Hospital, Naharia; Poria Hospital, Tiberias; Rambam Medical Centre, Haifa; Rivka Ziv Hospital, Zefat; Schneider Children's Medical Centre of Israel and Rabin Medical Centre (Belinson Campus), Petach-Tikva; Shaare-Zedek Hospital, Jerusalem; Sheba Medical Centre, Tel-Hashomer; Soroka Medical Centre, Beer-Sheva; Sourasky Medical Centre, Tel-Aviv; Wolfson Medical Centre, Holon; Yoseftal Hospital, Eilat.

Coordinating centre

Women and Children's Health Research Unit, Gertner Institute, Tel-Hashomer.

Accepted 29 March 2004