Volume 94, Issue 11 p. 1203-1214
Original Research Article
Free Access

Impact of obstetric factors on outcome of extremely preterm births in Sweden: prospective population-based observational study (EXPRESS)

Karin Källén

Karin Källén

Centre for Reproductive Epidemiology, Lund University, Lund, Sweden

Search for more papers by this author
Fredrik Serenius

Fredrik Serenius

Women's and Children's Health, Section for Pediatrics, Uppsala University, Uppsala, Sweden

Department of Clinical Sciences, Pediatrics, Umeå University, Umeå, Sweden

Search for more papers by this author
Magnus Westgren

Magnus Westgren

Department of Obstetrics and Gynecology, Karolinska University Hospital Huddinge, Stockholm, Sweden

Search for more papers by this author
Karel Maršál

Corresponding Author

Karel Maršál

Department of Obstetrics and Gynecology, Clinical Sciences Lund, Lund University, Lund, Sweden

Correspondence

Karel Maršál, Department of Obstetrics and Gynecology Lund, Skåne University Hospital Lund, 221 85 Lund, Sweden.

E-mail: [email protected]

Search for more papers by this author
the EXPRESS Group

the EXPRESS Group

EXPRESS Group members are listed at the end of this article.Search for more papers by this author
First published: 06 August 2015
Citations: 52

Conflicts of interest:

The authors have stated explicitly that there are no conflicts of interest in connection with this article.

Abstract

Introduction

A population-based observational study investigated the contribution of obstetric factors to the survival and postnatal development of extremely preterm infants.

Material and methods

Mortality up to 1 year and neurodevelopment at 2.5 years (Bayley-III test, cerebral palsy, vision, hearing) were evaluated in infants born before 27 weeks of gestation in Sweden 2004–2007 (n = 1011), using logistic regression analyses of risk factors.

Results

Of 844 fetuses alive at admission, 8.4% died in utero before labor, 7.8% died intrapartum. Of 707 live-born infants, 15% died within 24 h, 70% survived ≥365 days, 64% were assessed at 2.5 years. The risk of death within 24 h after birth decreased with gestational age [odds ratio (OR) 0.3; 95% CI 0.2–0.4], antenatal corticosteroids (OR 0.3; 95% CI 0.1–0.6), and cesarean section (OR 0.4; 95% CI 0.2–0.9); it increased with multiple birth (OR 3.0; 95% CI 1.5–6.0), vaginal breech delivery (OR 2.3; 95% CI 1.0–5.1), 5-min Apgar score <4 (OR 50.4; 95% CI 28.2–90.2), and birth at a level II hospital (OR 2.6; 95% CI 1.2–5.3). The risk of death between 1 and 365 days remained significantly decreased for gestational age and corticosteroids. The risk of mental developmental delay at 2.5 years decreased with gestational age, birthweight and fetal growth; it increased with vaginal breech delivery (OR 2.0; 95% CI 1.2–7.4), male gender, low Apgar score and high Clinical Risk Index for Babies score.

Conclusion

Several obstetric factors, including abdominal delivery, influenced the risk of death within the first day of life, but not later. Antenatal corticosteroids and gestational age decreased the mortality up to 1 year. Mental developmental delay was related to vaginal breech delivery.

Abbreviations

  • BW
  • birthweight
  • CI
  • confidence interval
  • CP
  • cerebral palsy
  • CRIB
  • Clinical Risk Index for Babies
  • EXPRESS
  • Extremely Preterm Infants in Sweden Study
  • GA
  • gestational age
  • MDD
  • mental developmental delay
  • OR
  • odds ratio
  • PPROM
  • preterm prelabor rupture of membranes
  • SDS
  • standard deviation score
  • SD
  • standard deviation
  • SGA
  • small-for-gestational age
  • Key Message

    The overall survival of extremely preterm infants was related to pro-active obstetric policy with use of tocolytics, antenatal corticosteroids, and cesarean section. Obstetric factors, except for vaginal breech delivery, were not associated with neurodevelopmental delay at 2.5 years of age.

    Introduction

    It is difficult to evaluate to what extent individual obstetric measures contribute to the substantially increased survival of extremely preterm born infants that was noted in the past two decades 1, 2. The improved prognosis for these infants is mainly attributed to various improvements in neonatal care. The obstetric care differs between regions and between countries 3. This reflects the uncertainty of what is the optimal way of managing extremely preterm fetuses and how the management is influenced by medical care systems and funding in various societies. Finally, predominating attitudes among lay public generally favoring maintaining life under all circumstances may influence the obstetricians' opinion on these matters 4.

    We have previously reported mortality, morbidity, 1-year survival and neurodevelopmental outcome at 2.5 years of age in a population-based prospective observational study on all infants born before 27 weeks of gestation in Sweden during a 3-year period ending 2007 (Extremely Preterm Infants in Sweden Study; EXPRESS) 2, 5, 6. The present study is an evaluation of how obstetric factors and management influenced the survival at various time-points and the neurodevelopmental outcome at 2.5 years of age.

    Material and methods

    We collected perinatal data on all 1011 infants born before 27 completed gestational weeks from 1 April 2004 to 31 March 2007 in Sweden. Data on both the live-born and stillborn infants were registered, from the latter at gestational age 22+0 to 26+6 weeks. The details of the study design and data collection have been reported previously 2; here only a short account is given.

    The prospective collection of data was organized on a regional level (seven healthcare regions with seven level III perinatal centers). The perinatal data comprised 220 items regarding the demographic information, medical and obstetric history, pregnancy course, labor and delivery, neonatal data, and data on the postnatal course during hospitalization until discharge home. Information on survival at 1 year was assessed through linkage with the Swedish Population Register. The gestational age (GA) was based on ultrasound dating performed before 20 postmenstrual weeks in 95% of pregnancies and on the date of the last menstrual period in 16 pregnancies. In 28 pregnancies the dating method was not specified. In 10% of live-born and in 13% of stillborn infants there was a difference of ≥14 days between the expected dates of delivery according to ultrasound and according to the last menstrual period. The birthweights (BW) were evaluated using the national intrauterine growth standard based on fetal weights 7. The deviation from the expected BW was expressed as a standard deviation score (SDS), small-for-gestational age (SGA) BW being more than two standard deviations (SD) below the mean.

    Preeclampsia was defined as hypertension after 20 weeks of gestation (blood pressure ≥ 140/90 mmHg and proteinuria ≥ 0.03 g/L). Preterm prelabor rupture of membranes (PPROM) was defined as spontaneous rupture of membranes ≥1 h before the onset of contractions. The diagnosis of chorioamnionitis was made clinically. The following obstetric interventions were recorded per infant: antenatal use of tocolytics, antibiotics (any antibiotic drug administered to the mother during hospitalization that resulted in delivery), and corticosteroids (at least one dose of betamethasone), electronic fetal heart rate monitoring, delivery at level III hospital, and delivery by cesarean section. Spontaneous preterm labor was defined as a labor beginning with spontaneous uterine contractions or contractions after PPROM. The delivery on maternal and/or fetal indication was described as iatrogenic vaginal or cesarean delivery, either after induced labor or as a prelabor cesarean section. The condition of infants at birth was evaluated with Apgar scores at 1 and 5 min and the neonatal illness severity was estimated using the clinical risk index for babies (CRIB score) 8.

    At 2.5 years of corrected age, the children available for the follow up were subjected to a clinical examination including vision and hearing. The motor, cognitive and language development of children was evaluated using the Bayley Scales of Infant and Toddler Development, 3rd edition (Bayley-III) 9. In 41 cases the information about the children was obtained from their medical charts. Moderate or severe cerebral palsy (CP) and moderate or severe impairment regarding vision and hearing were characterized as neurosensory impairment. Mental developmental delay (MDD) was defined as the cognitive or language Bayley III scale <mean − 2SD of the term control group, or moderate or severe developmental delay according to chart review. Moderate or severe neurodevelopmental disability was present if there was moderate or severe neurosensory impairment, Bayley III cognitive, language or motor scores <mean−2SD of the controls or moderate/severe developmental delay according to chart review. The methodological details of examinations have been described previously 6.

    For analysis of the impact of obstetric factors, the study population (n = 1011) was divided into subgroups according to the outcome at given time-points: intrauterine death before admission to hospital (n = 167), antepartum or intrapartum death after admission (n = 137), live-born infants that died within 24 h after birth (n = 106), and infants that died between 1 and 365 days after birth (n = 104), infants who survived ≥365 days (n = 497), and children that were examined at 2.5 years of corrected age (n = 456).

    The Regional Research Ethics Board at Lund University approved the study (Registration No. 42/2004).

    Statistical methods

    Risk factor analyses for stillbirth, postnatal death within 24 h, death 1–364 days, and outcome at 2.5 years, were performed using logistic regression analyses. The details of each model are specified in the text and in table headings. For each outcome and evaluated potential risk factor, odds ratios (OR) with 95% CI were calculated: crude, adjusted for GA (entered as a linear continuous variable, and also as a second-grade polynomial), and, for mortality, adjusted for GA (linear term) and for BW SDS (second-grade polynomial). Variables with p-values <0.2 after adjustments for GA and BW SDS were entered into the final multiple models. When specifically studying the impact of BW SDS on survival, a cubic model of BW SDS was used to further improve the fit. The goodness of fit of each model was assessed by the Hosmer–Lemeshow test. For each model, the number of investigated factors never exceeded 1/10 of the number of cases. No adjustments were made for multiple comparisons. All statistical analyses were done using Gauss (Aptech Systems Inc., Maple Valley, WA, USA).

    Results

    Overall, 1011 infants were born to 887 mothers in 904 deliveries; 304 (30%) infants were stillborn and 707 (70%) were live-born. In all, 54% of all infants were males, 54% and 55% of the stillborns and the live-borns, respectively. There were 102 multiple births (11.3%); and of these, seven were triplet births. In 22% of 540 mothers with available information on the country of origin, the women came from non-Nordic countries.

    Mortality and survival

    Of the stillborns, 55% died before admission to hospital (Table 1). Of the 844 fetuses alive at admission, 8.4% died in utero before the onset of labor or before prelabor cesarean section, 7.8% died intrapartum and 84% were born alive. Of all live-born infants, 70% survived ≥365 days after birth. Half of the postnatal deaths occurred during the first day of life. Of 497 survivors at 1 year of age, six children had died before 2.5 years of corrected age.

    Table 1. Mortality and survival according to gestational age of all infants born before 27 weeks of gestation in Sweden during 3 years (2004–2007); values are n (%)
    Gestational age (completed weeks)
    22a 23 24 25 26 Total <27 weeks
    All infants (n = 142) (n = 183) (n = 191) (n = 250) (n = 245) (n = 1011)
    Death before admission 39 (27.5) 43 (23.5) 30 (15.7) 25 (10.0) 30 (12.2) 167 (16.5)
    Alive at admission (n = 103) (n = 140) (n = 161) (n = 225) (n = 215) (n = 844)
    Death before onset of labor or before elective cesarean section 12 (11.7) 20 (14.3) 15 (9.3) 17 (7.6) 7 (3.3) 71 (8.4)
    Intrapartum death 40 (38.8) 19 (13.6) 2 (1.2) 3 (1.3) 2 (0.9) 66 (7.8)
    Live-born (= 51) (= 101) (= 144) (= 205) (= 206) (= 707)
    Death <24 h after birth 39 (76.5) 30 (29.7) 21 (14.6) 10 (4.9) 6 (2.9) 106 (15.0)
    Death 1–6 days 2 (3.9) 9 (8.9) 9 (6.2) 13 (6.3) 13 (6.3) 46 (6.5)
    Death 7–27 days 4 (7.8) 7 (6.9) 11 (7.6) 6 (2.9) 7 (3.4) 35 (5.0)
    Death 28–364 days 1 (2.0) 2 (2.0) 7 (4.9) 9 (4.4) 4 (1.9) 23 (3.3)
    Alive at 365 days 5 (9.8) 53 (52.5) 96 (66.7) 167 (81.5) 176 (85.4) 497 (70.3)
    Alive at 2.5 years corrected age 5 (9.8) 52 (51.5) 95 (66.0) 166 (81.0) 173 (84.0) 491 (69.4)
    Assessed at 2.5 yearsb 5 (9.8) 47 (46.5) 86 (59.7) 151 (73.7) 167 (81.1) 456 (64.5)
    • aIncludes two infants born alive at 21+5 and 21+6 (weeks+days), respectively.
    • bIncludes 41 children evaluated according to patient charts.

    The maternal chronic diseases, and pregnancy and labor complications are presented in the Supplementary material (Table S1). In the toal of 1011 cases, frequencies of essential hypertension, preeclampsia and abruption of placenta including placenta praevia were 4.2%, 10.9% and 12.4%, respectively. The corresponding rates for PPROM and chorioamnionitis were 14.6% and 15.3%, respectively. Diabetes mellitus is not presented because it was reported in only 11 cases (1.1%). Abruption of placenta and placenta praevia were most frequent in pregnancies with live-born infants and at higher GA (25–26 weeks).

    Of all infants, 70% were born at level III hospitals, the corresponding figure for the live-born infants was 79% (Table 2). Antibiotics were given to mothers of 47% of infants, most often to the mothers of infants born alive (55%; in 89% of PPROM cases). In 74% of pregnancies with spontaneous preterm labor, tocolytic treatment was initiated; in 86% of survivors without any differences between the GA subgroups. Antenatal corticosteroids were given to 65% of all infants, to 84% of those born alive, and to 90% of survivors. In live-born infants, the use of electronic fetal heart rate monitoring increased with increasing GA.

    Table 2. Obstetric interventions according to outcome groups and gestational age; values are n (%)
    Gestational age (weeks)
    22 23 24 25 26 Total < 27 weeks
    Alive at admission, stillbirths (n = 52) (n = 39) (n = 17) (n = 20) (n = 9) (n = 137)
    Antibiotics 22 (42.3) 20 (51.3) 11 (64.7) 6 (30.0) 2 (22.2) 61 (44.5)
    Tocolysis, n/Na 19/43 (44.2) 15/32 (46.9) 8/13 (61.5) 5/9 (55.6) 2/5 (40.0) 49/102 (48.0)
    Corticosteroids 7 (13.4) 29 (74.4) 17 (100) 9 (45.0) 2 (22.2) 64 (46.7)
    Electronic FHR monitoring (EFM) 7 (13.5) 4 (10.3) 4 (23.5) 10 (50.0) 2 (22.2) 27 (19.7)
    No information on EFM 8 (15.4) 16 (41.0) 6 (35.3) 3 (15.0) 2 (22.2) 35 (25.5)
    Cesarean section 1 (1.9) 1 (2.6) 1 (5.9) 1 (5.0) 1 (11.1) 5 (3.6)
    Delivery at level III hospital 24 (46.2) 30 (76.9) 15 (88.2) 12 (60.0) 4 (44.4) 85 (62.0)
    Live-born, death <24 h after birth (n = 39) (n = 30) (n = 21) (n = 10) (n = 6) (n = 106)
    Antibiotics 23 (59.0) 19 (63.3) 14 (66.7) 4 (40.0) 4 (66.7) 64 (60.4)
    Tocolysis, n/Na 19/33 (57.6) 17/27 (63.0) 16/18 (88.9) 4/7 (57.1) 3/4 (75.0) 59/89 (66.3)
    Corticosteroids 10 (25.7) 21 (70.0) 19 (90.4) 7 (70.0) 0 (0.0) 62 (58.5)
    Electronic FHR monitoring (EFM) 0 (0.0) 8 (26.7) 11 (52.4) 4 (40.0) 4 (66.7) 27 (25.5)
    No information on EFM 6 (15.4) 10 (33.3) 1 (4.8) 0 (0.0) 1 (16.7) 18 (17.0)
    Cesarean section 3 (7.7) 1 (3.3) 6 (28.6) 6 (60.0) 4 (66.7) 20 (18.9)
    Delivery at level III hospital 14 (35.9) 18 (60.0) 21 (100) 6 (60.0) 4 (66.7) 63 (59.4)
    Live-born, death 1–364 days (n = 7) (n = 18) (n = 27) (n = 28) (n = 24) (n = 104)
    Antibiotics 4 (57.1) 8 (44.4) 11 (40.7) 13 (46.4) 10 (41.7) 46 (44.2)
    Tocolysis, n/Na 5/7 (71.4) 14/17 (82.4) 14/18 (77.8) 10/17 (58.8) 8/11 (72.7) 51/70 (72.9)
    Corticosteroids 5 (71.4) 16 (88.9) 24 (88.9) 20 (71.4) 17 (70.8) 82 (78.8)
    Electronic FHR monitoring (EFM) 0 (0.0) 4 (22.2) 16 (59.3) 18 (64.3) 16 (66.7) 54 (51.9)
    No information on EFM 1 (14.3) 5 (27.8) 5 (18.5) 5 (17.9) 0 (0.0) 16 (15.4)
    Cesarean section 0 (0.0) 3 (16.7) 15 (55.6) 19 (67.9) 18 (75.0) 55 (52.9)
    Delivery at level III hospital 5 (71.4) 14 (77.8) 25 (92.6) 19 (67.9) 19 (79.2) 82 (78.8)
    Survivors, alive at 365 days (n = 5) (n = 53) (n = 96) (n = 167) (n = 176) (n = 497)
    Antibiotics 5 (100) 35 (66.0) 56 (58.3) 94 (56.3) 88 (50.0) 278 (55.9)
    Tocolysis, n/Na 4/5 (80.0) 40/47 (85.1) 63/74 (85.1) 103/118 (87.3) 93/107 (86.9) 303/351 (86.3)
    Corticosteroids 5 (100) 48 (90.6) 87 (90.6) 149 (89.3) 158 (89.8) 447 (89.9)
    Electronic FHR monitoring 2 (40.0) 18 (34.0) 54 (56.2) 122 (73.1) 149 (84.7) 345 (69.4)
    No information on EFM 1 (20.0) 10 (18.9) 12 (12.5) 17 (10.2) 16 (9.1) 56 (11.3)
    Cesarean section 0 (0.0) 13 (24.5) 46 (47.9) 103 (61.7) 119 (67.6) 281 (56.5)
    Delivery at level III hospital 5 (100) 47 (88.7) 83 (86.5) 141 (84.4) 137 (77.8) 413 (83.1)
    • EFM, electronic FHR monitoring; FHR, fetal heart rate.
    • aCalculated for cases of spontaneous preterm labor only (N).

    Of the 707 live-born infants, 50% were delivered by cesarean section (at 22–23 weeks 13%, at 24–26 weeks 61%) (Table 2). Cesarean section was performed in nine cases resulting in stillbirth, equally distributed among the gestational weeks. Of those, six were made on maternal indication (one severe preeclampsia, three placenta abruptions, two uterine pathology – multiple scars and large myoma, respectively), and one as a part of maternal resuscitation after cardiac arrest. Two cesarean sections were performed on fetal indication in twin pregnancies where one fetus had died in utero.

    Supporting Information Table S2 presents the infant characteristics at birth according to outcome groups and GA. The BW ranged from 100 to 1130 g in the stillborns and from 266 to 1500 g in the live-born infants. The median BW of the 497 survivors was 770 g (range 348–1315 g). Of the stillborns, 39.5% were SGA; of the infants born alive, 16% were SGA. Nine (1.3%) of the live-born infants were large for gestational age (three with congenital anomalies, one hydrops, five with no concomitant diagnosis). Structural congenital anomalies were diagnosed in 4.6% of stillborns and in 11% of live-born infants. Apgar score <4 at 5 min was most common among 106 infants who died within the first day of life (79%), ranging from 92% at 22 weeks to 17% at 26 weeks. Among infants surviving the first 24 h, 7.5% had Apgar score at 5 min <4 with no significant differences between the gestational weeks or between the infants who died within 1 year and the survivors ≥365 days. The frequency of a CRIB score >10 among infants who died between 1 and 365 days after birth (22%) was about twice that of the survivors (10%).

    Of the children assessed at 2.5 years of corrected age, 34 of 456 (7.5%) had neurosensory impairment, 88 of 440 (20%) had MDD, and 124 of 456 (27.2%) had any moderate or severe disability.

    Analyses of obstetric factors

    Table 3 shows the associations between obstetric factors and stillbirth among fetuses admitted alive, neonatal death within 24 h, and infant death between 1 and 365 days, respectively, as compared with infants born alive and surviving ≥365 days. The results shown are based on univariate analyses and on multiple models including the variables with p-values <0.2 after adjustments for GA and BW SDS. The Supplementary material, Tables S3–S5, presents the complete results including the intermediate multiple models with GA, and GA and BW SDS, respectively.

    Table 3. Factors associated with stillbirth among fetuses admitted to hospital alive, with neonatal death within 24 h and with infant death between 1 and 365 days, respectively, vs. infants born alive and surviving at least 365 days
    Stillbirth among fetuses admitted alive Neonatal death within 24 h Infant death between 1 and 365 days
    Univariate analysis Multiple model Univariate analysis Multiple model Univariate analysis Multiple model
    OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI
    Gestational age and intrauterine growth
    GA, 1-week increment 0.4 0.4–0.5*** 0.4 0.3–0.5*** 0.3 0.2–0.3*** 0.3 0.2–0.4*** 0.7 0.6–0.8*** 0.6 0.5–0.8***
    SD (birthweight), one-step increment 0.7 0.6–0.8*** 1.3 1.1–1.6** 0.8 0.7–0.9** 0.8 0.6–1.0*
    SD (birthweight), second grade model
    Linear model 0.8 Simultaneous 0.6 Simultaneous 1.6 Simultaneous 1.5 Simultaneous
    Quadratic term 1.1 p < 0.001 1.0 p < 0.001 1.1 p < 0.001 1.3 p < 0.001
    Pregnancy and delivery
    Preeclampsia/Essential hypertension 1.2 0.7–2.1 0.4 0.1–0.9* 0.7 0.1–2.8 1.2 0.7–2.3
    Chorioamnionitis/PPROM 1.7 1.1–2.6* 1.9 1.2–3.0** 1.3 0.8–2.0 2.5 1.3–4.8** 0.9 0.5–1.6
    Multiple birth 1.9 1.2–3.0* 3.0 1.5–6.0** 1.5 0.9–2.5 1.4 0.8–2.4
    Abruption of placenta 0.8 0.4–1.6 0.8 0.4–1.7 0.9 0.5–1.9
    Breech, vaginal delivery 4.0 2.5–6.4*** 3.0 1.7–5.2*** 4.4 2.4–7.8*** 2.3 1.0–5.1* 2.5 1.2–5.1* 1.7 0.8–3.9
    Birth at level II or level I hospital 2.3 1.6–3.4*** 1.8 1.1–2.9* 3.4 2.1–5.3*** 2.6 1.2–5.3* 1.3 0.8–2.2 1.2 0.6–2.3
    Cesarean section 0.2 0.1–0.3*** 0.4 0.2–0.9* 0.9 0.6–1.3
    Antenatal antibiotics 1.2 0.8–2.0 0.6 0.4–0.9* 0.8 0.5–1.3
    Antenatal corticosteroids 0.1 0.1–0.2*** 0.3 0.1–0.6** 0.3 0.2–0.6*** 0.3 0.2–0.8**
    Tocolysisa 0.3 0.2–0.5*** 0.8 0.3–1.9 0.4 0.2–0.8* 0.6 0.3–1.2
    Newborn characteristics
    Male gender (vs. female) 1.0 0.7–1.5 1.0 0.6–1.5 1.1 0.7–1.7
    Birth weight (100 g increment) 0.4 0.4–0.5*** Not included 0.6 0.5–0.6*** Not included 0.7 0.6–0.8*** Not included
    SGA 2.1 1.4–3.3*** Not included 0.5 0.3–1.1 Not included 1.6 0.9–2.6 Not included
    Apgar score <4 at 1 min 13.2 7.8–22.3*** Not included 1.6 1.0–2.6* Not included
    Apgar score <4 at 5 min 50.4 28.2–90.2*** Not included 1.4 0.7–2.9 Not included
    CRIB score >10 2.6 1.5–4.5*** Not included
    CRIB score (continuous), one-step increment 1.2 1.1–1.3*** Not included
    • Results from univariate analyses and multiple logistic regression analyses. The multiple models include all variables with p < 0.2 after adjustment for gestational age and fetal growth (data not shown), with exception of Apgar score and CRIB score that were considered to be intermediate variables.
    • CI, confidence interval; CRIB, Clinical Risk Index for Babies; GA, gestational age; OR, Odds ratio; PPROM, preterm prelabor rupture of membranes; SD, standard deviation of birth weight according to the Swedish standard curve for intrauterine growth 7; SGA, small-for-gestational age, i.e. birth weight below the mean −2 SD of the standard 7.
    • *p < 0.05; **p < 0.01; ***p < 0.001.
    • aOnly cases with spontaneous preterm labor.

    The risk of stillbirth decreased with advancing GA, higher BW and increasing BW SD, whereas chorioamnionitis/PPROM, SGA, vaginal breech delivery, and birth at level II or level I hospital were associated with an increased risk in univariate analyses. In the multiple models, the associations remained significant.

    In univariate analyses, the risk of neonatal death within 24 h decreased with advancing GA, higher BW, preeclampsia/essential hypertension, tocolysis, antenatal corticosteroids and cesarean delivery. Multiple birth, vaginal breech delivery, birth at level II or I hospital, and Apgar score <4 both at 1 and 5 min were adversely associated with outcome. In multiple models the associations remained significant with the exception of preeclampsia and tocolysis. Administration of antenatal corticosteroids (risk decrease), Apgar score <4 at 5 min and birth at level II or I hospital (risk increase) were the most significant predictors. Adjusting for all variables led to chorioamnionitis/PPROM becoming significantly associated with increased risk of death within 24 h.

    The risk of infant death between 1 and 365 days decreased with advancing GA, higher BW, antenatal antibiotics and tocolysis, whereas vaginal breech delivery, Apgar score <4 at 1 min, and high or increasing CRIB score were associated with an increased risk. The strongest associations between the risk factors and death between 1 and 365 days of life were found for advancing GA and for antenatal corticosteroids (risk decrease), and for increasing CRIB score (risk increase). This did not change after adjusting for all variables in the multiple models.

    Maternal demographic factors and fetal gender did not show any significant associations with the investigated outcomes, i.e. stillbirth, neonatal death within 24 h, and infant death between 1 and 365 days.

    Using a cubic model for BW SDS and a quadratic model for GA, very high goodness of fit for the logistic regression analyses was achieved (Hosmer and Lemeshow = 0.72 and = 0.29 for stillbirth and infant death, respectively). The results are presented in Figure 1. Although the risk for stillbirth steadily decreased with increasing BW SDS, a U-shaped relation was found between BW SDS and infant death.

    Details are in the caption following the image
    Risk for stillbirth and infant death in relation to fetal growth. The fetal growth is expressed as birthweight standard deviation scores according to the Swedish intrauterine growth standard 7. Only fetuses alive at admission to the delivery unit were included. Odds ratios (with 95% CI as vertical bars) were obtained from multiple logistic regression analyses using a cubic model for birthweight standard deviation scores and a quadratic model for gestational age.

    The results of analyses regarding the neurodevelopmental outcome at 2.5 years of corrected age are presented in Table 4. Of the obstetric factors, vaginal breech delivery was the only one significantly associated with the three categories of neurodevelopmental impairment. For the MDD, the increased risk remained significant after adjustment for GA (< 0.05). Male gender increased the risk for MDD and any moderate or severe disability; the increase remained significant after adjustment for GA (< 0.01 and < 0.05, respectively). Increase in the BW was associated with a risk decrease in all three categories of impairment. The strongest association with the neurodevelopmental outcome was found for neonatal condition, characterized by the Apgar score and CRIB score. Apgar score <4 at 1 min and Apgar score <7 at 5 min were significantly related to an increased risk for MDD and any moderate or severe disability (< 0.01 in both cases after adjustment for GA). CRIB score >10 was significantly associated with an increase in the risk for all three neurodevelopmental categories (< 0.001).

    Table 4. Factors associated with outcome at 2.5 years of corrected age
    n Neurosensory impairmenta Mental developmental delayb Any moderate or severe disabilityc
    (n = 34/456; 7.5%) (n = 88/440; 20%) (n = 124/456; 27.2%)
    % OR crude OR adj.d 95% CI % OR crude OR adj.d 95% CI % OR crude OR adj.d 95% CI
    GA, 1-week increment 0.6** 0.6*** 0.6***
    SD (birthweight), one-step increment 1.0 0.9 0.7–1.3 0.8 0.8** 0.6–0.9 1.0 0.9 0.7–1.0
    Preeclampsia/Essential hypertension 62 6.5 0.8 1.0 0.3–3.0 16.7 0.8 1.0 0.5–2.0 25.8 0.9 1.1 0.6–2.1
    Chorioamnionitis/PPROM 112 6.2 0.7 0.8 0.3–2.0 18.3 0.8 0.9 0.5–1.7 25.0 0.7 0.8 0.5–1.5
    Abruption of placenta 67 4.5 0.5 0.6 0.2–1.9 15.4 0.7 0.7 0.4–1.5 23.9 0.8 0.9 0.5–1.6
    Multiple birth 86 5.8 0.7 0.8 0.3–2.1 24.7 1.4 1.5 0.8–2.7 31.4 1.3 1.4 0.8–2.4
    Breech/vaginal delivery 29/202 17.2 3.5* 2.4 0.7–8.7 46.4 4.5*** 2.0* 1.2–7.4 55.2 3.6** 2.1 0.9–5.1
    Birth at level II or level I hospital 77 6.9 1.6 1.9 0.8–4.4 18.6 0.9 1.1 0.5–2.1 27.4 0.9 1.1 0.6–1.9
    Cesarean section 254 7.5 1.0 1.5 0.7–3.2 19.4 0.9 1.3 0.8–2.2 25.2 0.8 1.1 0.7–1.8
    Antenatal antibiotics 252 7.9 1.1 1.0 0.5–2.1 19.4 0.9 0.8 0.5–1.3 28.6 1.1 1.0 0.6–1.6
    Antenatal corticosteroids 411 7.5 1.0 1.1 0.3–4.8 20.0 0.7 0.7 0.3–1.9 27.7 1.1 1.2 0.5–2.9
    Tocolysise 256/325 8.2 1.0 1.0 0.3–2.7 20.5 0.9 0.9 0.4–1.8 29.3 1.0 0.9 0.5–1.8
    Male gender 248 8.9 1.6 1.7 0.8–3.5 24.2 1.8* 2.0** 1.2–3.3 31.0 1.6* 1.7* 1.1–2.6
    Birthweight, 100-g increment 0.8* 1.0 0.7–1.3 0.7*** 0.8** 0.6–0.9 0.7*** 0.9 0.7–1.0
    SGA 73 6.8 0.9 1.1 0.4–3.0 22.2 1.2 1.5 0.8–2.8 27.4 1.0 1.3 0.7–2.3
    Apgar score <4 at 1 min 109 7.3 1.0 0.9 0.4–2.0 28.6 1.9* 1.7* 1.0–2.8 36.7 1.8* 1.6* 1.0–2.6
    Apgar score <4 at 5 min 33 3.0 0.4 0.3 0.0–2.7 30.3 1.8 1.8 0.8–3.9 36.4 1.6 1.5 0.7–3.3
    Apgar score <7 at 5 min 149 6.7 0.8 0.8 0.3–1.6 29.5 2.3*** 2.1** 1.3–3.5 36.9 2.0** 1.9** 1.2–2.9
    CRIB score >10 42 23.8 5.1*** 3.7** 1.6–8.9 50.0 4.8*** 3.5*** 1.7–7.2 61.9 5.2*** 3.7*** 1.8–7.3
    CRIB score (continuous), one-step increment 1.2*** 1.2* 1.0–1.3 1.2*** 1.2*** 1.1–1.3 1.2*** 1.1*** 1.1–1.2
    • Results from univariate analyses and multiple logistic regression analyses adjusting for gestational age. Children who were not evaluated at 2.5 years were excluded.
    • CI, confidence interval; CRIB, Clinical Risk Index for Babies; GA, gestational age; OR, Odds ratio; PPROM, preterm prelabor rupture of membranes; SD, standard deviation of birth weight according to the Swedish standard curve for intrauterine growth 7; SGA, small-for-gestational age, i.e. birth weight below the mean −2 SD of the standard 7.
    • *p < 0.05; **p < 0.01; ***p < 0.001.
    • aModerate or severe impairment regarding vision, hearing, or cerebral palsy (CP).
    • bMDD: mental developmental delay = cognitive or language Bayley III scale < mean − 2SD, or moderate/severe developmental delay according to chart review.
    • cVisual or hearing impairment, CP, low Bayley-III composite cognitive, language or motor score or moderate/severe developmental delay according to chart review.
    • dAdjusted for gestational age (continuous, linear).
    • eOnly cases with spontaneous preterm labor.

    Discussion

    During 2004–2007, 70% of extremely preterm infants born alive before 27 gestational weeks in Sweden survived at least 1 year. Most obstetric interventions, including transport to level III perinatal centers, tocolysis and cesarean section, decreased the risk of death within the first day of life, but not thereafter. Antenatal corticosteroids diminished the mortality risk both during the first 24 h and up to the age of 1 year. Vaginal breech delivery was associated with increased risk both for mortality and for neurodevelopmental delay at 2.5 years of age.

    Stillbirths after admission, but before the onset of labor (8.4%) were more common in association with chorioamnionitis/PPROM and at level II or I hospitals. It can be speculated that the obstetrician has refrained from actively delivering the mother in many of these cases due to anticipated poor survival. The number of intrapartum deaths was rather low compared with other studies on extremely preterm deliveries 10, 11. Intrapartum electronic fetal heart rate monitoring was applied often if the fetus was judged viable, however, the present material does not permit any conclusions regarding fetal monitoring during very preterm labor.

    Several classical obstetric complications associated with preterm delivery, e.g. multiple pregnancy, PPROM, related to infection and inflammation 12, or vaginal breech delivery, were associated with an increased risk for neonatal deaths within the first 24 h. It has been claimed that the neonatal deaths occurring during the first 12 h after birth do reflect the obstetric management and the degree of activity at primary resuscitation 13. Later deaths probably reflect more the standard of neonatal care. Our results show that if the child survived the first 24 h, continuing survival was likely.

    The cesarean section rate was 50%, comparable to that in the Finnish population-based study (58%) 14, but much higher than reported from the EPICure study for infants born before 26 gestational weeks (15.6%) 15. The optimal mode of delivery for extremely preterm infants is not established. Retrospective register-based studies reported that a liberal use of cesarean section on fetal indication in very preterm deliveries is associated with increased survival rate also when controlling for BW 16. Data from prospective studies are very limited and inconclusive due to methodological difficulties 3, 17. There are some observational and registry studies 18, 19 indicating that cesarean delivery is preferable for preterm fetuses in breech presentation. For vertex presentation in very preterm deliveries there are no conclusive data supporting routine use of cesarean section 19. In the present study, cesarean section was associated with increased survival within the first 24 h after birth, but not thereafter.

    Delivery at a level II or I hospital was associated with higher rates of stillbirth and death within 24 h after birth. This relationship concurs with the previous observational study in Sweden 20. An important goal of centralization is to provide the level III units with a higher number of cases, hence improving the skills and experience of the specialized perinatal teams.

    The antenatal treatment with tocolytics and corticosteroids was more frequent than reported in the literature 10, 14, 15 and the results indicated a high efficacy with decreased mortality. In contrast to tocolytics, the significant effect of corticosteroids remained after adjustment for other variables. A plausible explanation of the beneficial effect of tocolysis might be the time gain enabling corticosteroid treatment and transport to the level III hospital. Another possibility might be that our finding just reflects selection of a group of mothers in best health. In agreement with many other studies the present study confirmed the strong association between antenatal corticosteroids and the improved neonatal outcome 21. We did not find any difference in the steroid effect when comparing the estimates for 23 weeks with those for 24–26 weeks (p = 0.75), thus suggesting that the antenatal steroids are efficient already at that early GA. There were too few cases at 22+0 to 22+6 weeks to allow corresponding comparison.

    At 2.5 years of corrected age, low 1-min and 5-min Apgar scores, and high CRIB scores were strongly associated with neurodevelopmental disability after adjustment for GA. These indicators, besides acute and chronic fetal conditions, reflect intrapartum obstetric management, resuscitation and initial stabilization after birth. Previously, CRIB score was shown to have little value for predicting neurodevelopmental outcome 22. In our study CRIB score >10 was a better predictor than BW or GA. Although Apgar scores are associated with outcome at group level, it must be emphasized that individual Apgar scores cannot discriminate between infants who eventually survive unimpaired or who survive with impairment 23.

    In contrast to several studies, male gender was not associated with higher mortality. At follow up, however, the neurodevelopmental outcome of boys was poorer than that of girls which is also reported by others 24. We were not able to show an association between 2.5-year outcome and SGA, which might be due to the early age at follow up. However, a recent large study from the National Institute of Child Health and Human Development reported significantly increased risks for low Bayley-III cognitive and language scores at 18–24 months in SGA infants born at 23–26 weeks 25. Severe intrauterine growth restriction has been also associated with greater risk for cognitive impairment at 5–8 years in boys born at median 26.9 weeks 26.

    In this study, among obstetric factors, only vaginal breech delivery was significantly associated with adverse outcome (MDD) at 2.5 years corrected age – the unadjusted OR for neurosensory impairment (moderate and severe CP, moderate or severe impairment regarding vision and hearing) was significantly increased. In the EPICure study, breech delivery was associated with increased risk for CP at 30 months of age in infants born at <26 weeks 27 and a recent Norwegian registry study reported an increased risk for CP at 4 years in preterm children after breech delivery 28. The fact that term infants with breech presentation exhibit greater risk for CP than those with cephalic presentation regardless of mode of delivery 28 indicates that underlying conditions causing breech presentation contribute to the outcome. The previously not reported finding of the association with MDD might reflect a pathophysiology different from that associated with CP. This merits further exploration in our cohort at later follow up at 6.5 and 11 years.

    Proper counseling of the parents and the obstetric decision-making require access to accurate and up-to-date information not only on mortality, but also on morbidity. Considering the rather high long-term morbidity of extremely preterm infants 29, it should go without saying that institutions providing care for these mothers and infants should continuously perform long-term follow up into childhood and behind.

    In the EXPRESS study the 1-year survival was high compared with other national figures 10, 14, 30. When evaluating the impact of obstetric factors and interventions, several possibly contributing circumstances specific for the Swedish healthcare system should be considered. Standardized antenatal and perinatal care is free of charge for all inhabitants. Almost all pregnancies are dated with ultrasound in the first or early second trimester. Generally, the ultrasound pick-up rate of severe malformations is high. Hence pregnancies with severe malformations are frequently terminated, which is the most likely explanation for the low rate of severe malformations in the present material. Teenage pregnancies are very few (<4%) depending on liberal use of contraception and a rather high rate of terminations of pregnancy among young women. Single embryo transfer is generally implemented in all in vitro fertilization programs since 2003, which significantly decreased the proportion of preterm deliveries due to multiple pregnancies.

    To provide evidence for obstetric management of extremely preterm births, randomized controlled trials would be desirable. Unfortunately, to perform randomized trials in this group of patients is very difficult. From this perspective, the strength of the prospective observational EXPRESS study is that it is truly population-based and that it includes antenatal deaths and early delivery-room deaths. The weakness of this type of study is the difficulty or even impossibility distinguishing the effect on survival of the underlying pathology from the effects of preterm labor and delivery per se. Furthermore, the ultrasound dating of some of these pregnancies in the second trimester can be misleading because of early growth restriction and possibility of systematic GA underestimation. The resulting p-values should be interpreted with caution as some might be due to multiple comparisons. The found beneficial effect of cesarean section on short-term survival might include a risk of selection bias – it cannot be excluded that mothers with more favorable prospects for infant survival were those most often selected for cesarean section.

    In conclusion, pro-active obstetric policy with antenatal use of tocolytics and corticosteroids, transport to level III perinatal centers and liberal use of cesarean section, is associated with improved survival in extremely preterm infants. Vaginal breech delivery increases risk for mortality and for neurodevelopmental delay at 2.5 years of age, so supporting the view that cesarean section in these cases should be the method of choice.

    EXPRESS Group members

    Obstetricians

    Tomas Fritz, MD, Sahlgrenska University Hospital, Göteborg; Per Åke Holmgren MD, PhD, Umeå University Hospital, Umeå; Annika Jeppsson, MD, PhD, Sahlgrenska University Hospital, Göteborg, previously Linköping University Hospital; Karel Maršál, MD, PhD (principal investigator), Skåne University Hospital, Lund; Solveig Nordén-Lindeberg, MD, PhD, Uppsala University Hospital, Uppsala; Ingrid Östlund, MD, PhD, Örebro University Hospital, Örebro; Marija Simic, MD, PhD, Karolinska University Hospital Solna, Stockholm; the late Margareta Wennergren, MD, PhD, Sahlgrenska University Hospital, Göteborg; Magnus Westgren, MD, PhD, Karolinska University Hospital Huddinge, Stockholm.

    Neonatologists

    Mats Blennow, MD, PhD, Karolinska University Hospital Huddinge, Stockholm; Uwe Ewald, MD, PhD, Uppsala University Hospital, Uppsala; Vineta Fellman, MD, PhD, Skåne University Hospital, Lund; Lena Hellström-Westas, MD, PhD, Uppsala University Hospital, Uppsala; Hugo Lagercrantz, MD, PhD, Astrid Lindgren Children's Hospital, Karolinska Institute, Stockholm; Eva Lindberg, MD, PhD, Örebro University Hospital, Örebro; Mikael Norman, MD, PhD, Karolinska University Hospital, Stockholm; Elisabeth Olhager, MD, PhD, Skåne University Hospital, Lund; Fredrik Serenius, MD, PhD, Uppsala University Hospital, Uppsala, previously Umeå University Hospital, Umeå; Gunnar Sjörs, MD, PhD, Uppsala University Hospital, Uppsala; Lennart Stigson, MD, Sahlgrenska University Hospital, Göteborg.

    Other experts

    Maria Hafström, MD, PhD (pediatric neurology), Sahlgrenska University Hospital, Göteborg; Gerd Holmström, MD, PhD (ophthalmology), Uppsala University Hospital, Uppsala; Karin Källén, PhD (reproductive epidemiology), Lund University, Lund; the late Ricardo Laurini, MD, PhD (pathology), Bodö Central Hospital, Bodö, Norway; Anita Lundqvist, PhD (data collection coordinator), Lund University, Lund; Pia Lundqvist, RN, PhD (patient interviews), Lund University, Lund; Tore Nilstun, PhD (medical ethics), Lund University, Lund; Karin Stjernqvist, PhD (psychology), Lund University, Lund; Bo Strömberg, MD, PhD (pediatric neurology), Uppsala University Hospital, Uppsala; Brigitte Vollmer, MD, PhD (pediatric neurology), Astrid Lindgren Children's Hospital, Karolinska Institute, Stockholm.

    Acknowledgments

    The professional help in the collection of data and follow-up examinations of children by the following colleagues is gratefully acknowledged: Dr. Dordi Austeng, Dr. Kristina Forsblad, and Dr. Ulla Lindskog. The technical assistance at data collection by Ms. Grozda Pajic, Lund University, is greatly appreciated. The authors thank Marius Kublickas, MD, PhD (MedSciNet AB) for the design and maintenance of study database.

      Funding

      This study was supported by the Swedish Research Council (grants 2006-3855 and 2009-4250), the Swedish National Board of Health and Welfare, Grants to Researchers in the Public Health Care from the Swedish Government, the Uppsala-Örebro Regional Research Council grant RFR-10324, and grants from the Research Council South East Region of Sweden, from the Evy and Gunnar Sandberg Foundation, and from the Birgit and Håkan Ohlsson Foundation.