Preterm delivery for maternal or fetal indications: maternal morbidity, neonatal outcome and late sequelae in infants


Correspondence: Dr M. Kurkinen-Räty, Department of Obstetrics and Gynaecology, Päijät-Häme Central Hospital, 15850 Lahti, Finland.


Objective To assess maternal morbidity, and neonatal outcome and especially long term sequelae in infants born preterm due to maternal or fetal indications.

Design Analysis of retrospective cohort.

Setting Oulu University Central Hospital, Finland.

Population One hundred and three women, who were between the 24th and the 33rd week of pregnancy, delivered by caesarean delivery because of maternal or fetal indications. They were matched with 103 women who had spontaneous preterm delivery at corresponding gestational weeks between 1990–1997.

Main outcome measures Maternal morbidity, reasons for caesarean delivery, neonatal mortality and morbidity rates, and later development of the infants.

Results Pre-eclampsia was diagnosed in 57% of the women in the indicated group and only in one woman in the control group. All infants in the indicated group and almost a third in the control group were born by caesarean birth; the main indication was threatening fetal asphyxia. There was a significant difference in neonatal mortality rates between the groups (175 vs 78 per thousand live births in the indicated vs control infants; RR 2.3, 95% CI 1.02, 4.9) and the main cause of death was respiratory insufficiency: 64% in the indicated group and 22% in the controls; RR 2.9, 95% CI 0.8, 10. Respiratory distress syndrome occurred more often (73%vs 53%, RR 1.4, 95% CI 1.1, 1.7) and it was more severe and more complicated in infants in the indicated group, compared with those in the control group. Symptomatic chronic lung disease at one year of age was more common in infants in the indicated group than in the control group (15%vs 3%; RR 4.6, 95% CI 1.4, 15.9).

Conclusions Not only the risks of neonatal mortality and morbidity but also long term pulmonary consequences, appear to be greater in infants born preterm by indicated delivery than in preterm infants born spontaneously at corresponding weeks.


Pre-term birth is one of the unsolved problems in perinatology and neonatology, and it is a leading cause of mortality in infants. Modern neonatal technology has improved the prognosis of these preterm infants, but the morbidity rates are still high. With better understanding of aetiologies, treatment strategies and preventative actions connected with preterm delivery it may be possible to diminish preterm birth rates. So far, however, they have been relatively stable at around 5%–10% in different populations in recent years1.

In addition to spontaneous preterm delivery, prematurity can be of intervention by obstetricians, if maternal or fetal indications require it. The proportion of deliveries resulting in such ‘indicated’ pre-maturity has increased recently2. The long term impact of preterm caesarean delivery on infants' later sequelae, however, has not been studied extensively3. Spontaneous and indicated preterm births have different profiles, according to a study by Meis et al.4. However the later consequences of premature indicated birth have been examined sparsely.

The aim of this study was to evaluate maternal morbidity, neonatal outcome and long term consequences in infants at one year of age in deliveries by caesarean birth because of maternal or fetal indications between gestational weeks 24 and 33.

We retrospectively analysed 103 singleton pregnancies, where the fetus was alive at the decision to carry out caesarean delivery at 24 to 33 weeks of gestational age because of maternal and/or fetal indications. The women were admitted between 1 January 1990 and October 1997 to Oulu University Hospital, where there are approximately 3500 deliveries per year. The control group consisted of 103 women with singleton pregnancies and spontaneous preterm delivery after regular contractions and/or preterm rupture of the membranes no more than 24 hours before. The mothers were matched one-to-one by gestational age at delivery plus or minus 1 week.

A diagnosis of pre-eclampsia was made if the systolic blood pressure was constantly > 140 mmHg, the diastolic blood pressure was > 90 mmHg, and proteinuria was > 300 mg per day. Medication (mainly labetalol and nifedipine) was given to women whose blood pressures were > 160/100 mmHg despite rest. Samples were taken at least twice a week for serum thrombocytes, liver function tests (serum aspartate aminotransferase and alanine aminotransferase) and renal function (serum urate, albumin).

The fetus was monitored daily by cardiotocography and at least once a week by ultrasonography involving Doppler equipment (Toshiba SSA-270 with a 3.5 mHz abdominal convex probe and pulsed ultrasound with a 120 kHz filter) for blood flow velocities. The systolic to diastolic ratio of umbilical artery blood flow was measured, and absent end-diastolic or reversed end-diastolic velocity were detected visually. A constant decrease in fetal heart rate variability (short term variation < 4 ms) or the appearance of late (after uterine contraction) or variable decelerations (> 20 lost beats) in cardiotocography either by computer analysis (Team 8002 or Team Care, Sonicaid, Oxford), or mainly by visual evaluation, were considered as possible signs of fetal distress, as was a high systolic/diastolic ratio (values > 4) in the umbilical artery, or absent or reversed end-diastolic velocity in the same artery. The amount of amniotic fluid was estimated when evaluating the fetal biophysical profile. A vertical length of < 2 cm of the highest amniotic fluid column was the criterion of oligohydramnios.

Corticosteroids (dexamethasone) were administered to the women over two days (first 15 mg, and 10 mg after 24 hours) according to individual indications from the 24th week until the 32nd week in order to reduce the incidence of respiratory distress syndrome. Among those in the control group, tocolytics (mainly intravenous infusions of physiologic saline, ritodrine infusion or peroral sulindac) were used to treat regular premature contractions.

A diagnosis of respiratory distress syndrome was made on the basis of need for respiratory support, clinical assessment and radiologic findings. Bronchopulmonary dysplasia was diagnosed (according to the criteria of Northway et al.5) in infants with typical radiologic findings who continued to require supplemental oxygen and who had respiratory distress symptoms or signs at four weeks of age.

A diagnosis of infection in infants was based on positive blood culture results and clinical manifestations; and in those with negative blood culture, on clinical manifestations, radiological findings and laboratory indicators (i.e. serum C-reactive protein level > 20 mg/L)6. Infants with signs of maternal infection, and all infants with respiratory distress, were treated with antibiotics after birth until an infection was proven or excluded. A blood glucose concentration < 2.5 mmol/L (45 mg/100 mL) more than once was defined as hypoglycaemia, and its duration was registered as the complete period until the day the last low blood glucose value was recorded7.

Central nervous system abnormalities of the newborns were identified by serial cranial ultrasonography or at autopsy. Intraventricular-periventricular haemorrhage was graded into four categories, according to Papile et al.8. Periventricular leukomalacia was considered to exist when periventricular white matter cysts were seen on ultrasound. Ophthalmological examinations were performed in infants born before 32 weeks of gestation and in all infants with respiratory distress syndrome at 34 to 36 weeks of age and later as needed. Hearing screening was performed before discharge. Both groups of infants were evaluated at the outpatient clinic of the hospital at intervals of one to three months, up to at least one year of corrected age (calculated from the estimated and actual birth date). In 25 infants (15 in the indicated group and 10 in the control group) whose parents had moved from the Oulu region and lived near their central hospital, data were collected from central hospital records or from family health care centres. All infants, except two in the control group who were confirmed as in good condition at three months of age, remained in follow up until one year.

A diagnosis of symptomatic chronic lung disease was made in infants who needed supplemental oxygen, continuous bronchodilatation or steroid treatment because of respiratory symptoms or signs at one year of corrected age.

Neurological status was evaluated by a paediatrician together with a physiotherapist. The infants were grouped into four categories of neurological outcome according to the results of their examination at one year of corrected age: good neurological outcome = normal neurological status; delayed motor development = abnormalities of tone or reflexes but functionally normal or borderline; cerebral palsy = spastic diplegia or hemiplegia or spastic tetraplegia; other impairment = for example those with malformations. Measurements of weight, length and occipitofrontal circumference were made by a trained nurse, using standardised instruments. After correction for gestational age, a paediatrician recorded the measurements on growth charts officially used in Finland. Length and head circumference were expressed as standard deviations from normal growth, and weight was expressed relative to length as a percentage of normal.

Statistical analysis was performed using SPSS for Windows release 7.0 (SPSS, Inc, Chigaco, Illinois, USA) and by confidence interval analysis. Continuous variables were tested by using the paired t test with matched paired data and with the independent t test within unpaired data, and a Mann-Whitney U test or for paired data a sign test were used in skewed distributions. Categorical data were analysed by using Fisher's two-tailed exact test. Forward stepwise logistic regression analysis was used for testing the confounding factors.

Odds ratios, in the regression model, and relative risks and 95% confidence intervals are reported. P < 0.05 was used as the level of significance, and a 95% CI not including 1 (in relative risks and in odd ratios) or 0 in differences were considered statistically significant.


Because women indicated and spontaneously delivered groups were matched one-to-one only by the gestational age at the delivery, the other maternal and neonatal parameters differed in many aspects. The women in the indicated group had more chronic illnesses and were older than the women in the control group (Table 1). The illnesses were mainly of essential hypertension (16/33) and nephropathy (two diabetic and five of renal insufficiency). Gastric carcinoma was diagnosed in one woman in the indicated group during pregnancy and she died one month after delivery. Fetal chromosomal analyses were performed three times more frequently in women in the indicated group than in the control group and consisted mainly of amniocenteses. One fetus with trisomy 21 was found in both groups.

Table 1.  Maternal characteristics. Values are given as n (%) or mean [SD] {range} unless otherwise indicated. RR = relative risk.
 Indicated (n= 103)Spontaneous (n= 103)RR95% CI for RR
  1. *Statistically significant.

  2. P < 0.001.

  3. 95% CI for difference of mean.

History before current pregnancy    
 Chronic illness33(32)13(13)2.51.4, 4.5*
 Smoking11(11)17(17)0.60.3, 1.3
 Infertility11(11)9(9)1.20.5, 2.8
 Induced abortions6(6)17(17)0.40.1, 0.9*
 Previous preterm deliveries25(24)24(23)1.00.6, 1.7
 Previous spontaneous abortions35(34)21(20)1.71.04, 2.7*
Current pregnancy    
 Maternal age (years)32.1 [6.6] {17–46}27.9 [6.0] {15–43} 2.5, 5.9
 Nulliparous44(43)39(38)1.10.8, 1.6
 Bleeding in 1st trimester8(8)17(17)0.50.2, 1.0
 Bleeding in later pregnancy20(19)38(37)0. 50.3, 0.8*
 Chromosomal analyses performed31(30)10(10)3.11.6, 6.0*

Signs of chronic fetal asphyxia were detected by Doppler ultrasonography or cardiotocography in half of the cases in the indicated group (Table 2). Thirteen fetuses in whom absent end-diastolic velocity was detected in the umbilical artery also had a reverse end-diastolic velocity in the same artery. If an absent or reversed end-diastolic velocity (AREDV) was seen, 35/51 of the fetuses (69%) also had pathology in the cardiotocography. These cardiotocography pathologies consisted of decelerations in 30 fetuses and decreased variability in 32. Over half of the women in the indicated group had pre-eclampsia, compared with only one in the control group. Maternal indications consisted of subjective symptoms (e.g. headache, visual disturbances, upper abdominal discomfort or pain), or laboratory results showing worsening of the maternal status. Four women had convulsions. Serum urate was elevated in 24 mothers, in whom 16 had thrombocytopenia (values < 140 × 109/L) and 20 had elevated liver enzymes. Daily proteinuria over 5 g was seen in 35% of women. The primary indication for caesarean delivery was fetal in both the indicated group (63/103) and in the control group (28/29). There were nine cases of placental abruption and four of placenta praevia in the indicated group.

Table 2.  Perinatal characteristics. Values are given as n (%) and median [range] unless otherwise indicated. RR = relative risk; AREDV = absent or reversed end-diastolic velocity.
 Indicated (n= 103)Spontaneous (n= 103)RR95% CI for RR
  1. *Statistically significant.

  2. P < 0.001.

Maternal hospitalisation (days)†*8[0.55]1[0.40]  
Use of tocolysis4(4)66(64)0.060.02, 0.2*
Use of corticosteroids21(20)27(26)0.80.5, 1.3
Sonographically defined oligohydramnios31(30)13(13)2.41.3, 4.3*
Pathology in cardiotocography51(50)20(19)2.61.6, 4.0*
AREDV in umbilical artery51(50)0  
Clinical amnionitis1(1)17(17)0.060.01, 0.4*
Puerperal infections8(8)9(9)0.90.4, 2.2
Caesarean delivery103(100)29(28)3.62.6, 4.8*
Main reason for caesarean delivery    
 Fetal63(61)28(97)0.60.5, 0.8*
 Maternal40(39)1(3)111.6, 79*
Placental histopathology64(62)41(40)1.61.2, 2.1*
 None23(36)10(24)1.50.8, 2.8
 Infarct/fibrosis38(59)16(39)1.51.0, 2.4
 Inflammation3(5)15(37)0.10.04, 0.4*

According to logistic regression analysis indicated, preterm birth itself was the strongest determinant of major neonatal and later infant parameters. Maternal illnesses, signs of fetal asphyxia (AREDV in the umbilical artery and cardiotocography pathologies) and the history of previous preterm deliveries were more frequent in the indicated versus control group (Table 1 and 2). Maternal age, however, was not a confounding factor on the infant parameters. AREDV, cardiotocography pathologies and pre-eclampsia, as potential explanatory variables, were dependent on each other and had an effect on the outcome of the infant. The incidence of respiratory distress syndrome was not different if the woman had pre-eclampsia (43/60) or not (86/141), relative risk 1.4, 95% CI 0.9, 2.3. AREDV in the umbilical artery decreased the risk of neonatal respiratory distress syndrome, but not significantly (RR 0.9, 95% CI 0.5, 1.4). The use of prenatal corticoids had no effect on the incidence of respiratory distress syndrome, but instead appeared to increase the incidence of chronic lung disease at one year of age, which was more common (18%vs 6%) in the group that received corticoids (OR 2.2, 95% CI 1.2, 4.2). Indication for delivery itself was the major determinant on the incidence of small for gestational age fetuses, the incidence of which was nine times greater among the indicated compared with control infants.

The infants in the indicated group were not only significantly smaller but their condition was poorer after delivery than in the controls (Table 3). The incidence of small for gestational age fetuses was higher in the AREDV group (40/51) compared with those in the non-AREDV group (23/124) (P < 0.001). The neonatal mortality rate was significantly higher in the indicated group than in the controls (175 vs 78 per 1000 live births). The main causes of death were respiratory distress syndrome (n= 11) and bronchopulmonary dysplasia (n= 3) in the indicated group, and two cases of respiratory distress syndrome in the control group (Table 4).

Table 3.  Perinatal characteristics of the infants. Values are given as mean (SD) and n [%] unless otherwise indicated. RR = relative risk.
 Indicated (n= 103)Spontaneous (n= 103)RR or P95% CI for difference of mean
  1. *Statistically significant.

  2. P value.

  3. 95% CI for RR =4.1, 18,

  4. §0.8, 1.3, 0.4, 1.3.

Gestational age at birth (wk)30.5(2.1)30.4(2.1)0.8−0.04, 0.05
Birth weight (g)1294(469)1605(427)< 0.001−382, −240*
Length (cm)38.2(4.3)40.5(3.6)< 0.001−3.0, −1.7*
Small for gestational age (< 2 SD)60[58]7[7]8.6‡*
Head circumference (cm)27.9(2.7)28.9(2.7)< 0.001−1.4, −0.5*
Apgar score < 7    
 At 1 min64[62]50[49]10§
 At 5 min37[36]29[29]10§
Umbilical artery pH7.23(0.1)7.29(0.1)0.002−0.1, −0.02*
Table 4.  Infant mortality and main causes of death. Values are given as n(%) unless otherwise indicated. RR = relative risk.
 Indicated (n= 103)Spontaneous (n= 103)RR95%CI for RR
  1. *Statistically significant.

  2. Includes one infant who died during delivery.

 Death ≤day 717(17)5(5)3.41.3, 8.9*
 Death day 0–2818(17)8(8)2.31.02, 4.9*
 Neonatal mortality (per 1000 live births)175782.31.02, 4.9*
 Postneonatal death4(5)1(1)4.50.5, 39
 Infant mortality day 0–356 (per 1000 live births)214872.41.2, 5.1*
 Causes of death    
 Respiratory insufficiency14(64)2(22)2.90.8, 10
 Other3(14)6(67)0.20.06, 0.6*
 TOTAL22(100)9(100)2.41.2, 5.1*

Respiratory distress syndrome occurred more often and it was more severe and more complicated in infants in the indicated group compared with controls (Table 5). Among infants in the indicated group there was a significant difference with regard to respiratory distress syndrome according to whether caesarean delivery was carried out on the basis of fetal or maternal indications (65%vs 88%), P= 0.02, respectively. The occurrence of bronchopulmonary dysplasia was more common in the infants in the indicated group than in the controls. Hypoglycaemia (79%) was more common and its duration longer in infants in the indicated group compared with those in the control group, although half of the infants in the latter group also had low blood glucose values.

Table 5.  Morbidity of the infants during initial hospitalisation. Values are given as n (%) or median [range] unless otherwise indicated. FiO2 = fraction of inspired oxygen; RDS = respiratory distress syndrome; RR = relative risk.
 Indicated (n= 102)Spontaneous (n= 103)RR95% CI for RR
  1. *Statistically significant.

  2. P: 0.13; ≤ 0.001; §0.28.

RDS74(73%)55(53%)1.41.1, 1.7*
Pulmonary air leak18(18%)5(5%)3.61.4, 9.4*
Duration of FiO2 ≥ 21% (days)9[0.454]7[0.127]  
Bronchopulmonary dysplasia    
 At 28 days31(36%)19(20%)1.91.1, 3.0*
 At term18(18%)5(5%)4.21.6, 11*
Infection by day one1(1%)11(11%)0.10.01, 0.7*
No. of cases with hypoglycaemia81(79%)50(49%)1.61.3, 2.0*
Duration of hypoglycaemia (days)†*3[0.96]0.5[0.20]  
Intraventricular haemohage    
 Grades 1–212(12%)12(12%)1.00.5, 2.1*
 Grades 3–46(6%)9(9%)0.70.2, 1.8
Periventricular leukomalacia3(3%)9(9%)0.30.1, 1.2
Primary hospital stay (days)§44[0.304]40[0.151]  

At one year of age there was a significantly higher proportion of cases of chronic lung disease in the indicated group infants than in the controls (15%vs 3%; RR 4.6, 95% CI 1.4, 16). The incidence of chronic lung disease was significantly higher in AREDV group (8/37) compared with non-AREDV group (6/112) (RR 2.7, 95% CI 1.5, 4.6). There were no differences in long term neurologic findings at one year follow up, but the size of infants in the indicated group was still significantly lower than in the controls (Table 6).

Table 6.  Outcome of infants at one year of corrected age. Values are given as n (%) or median [range] unless otherwise indicated. RR = relative risk.
 Indicated (n= 94)Spontaneous (n= 81)RRP95% CI for RR
  1. *Statistically significant.

  2. 95% CI for difference of mean.

  3. Relative to length as percentage of normal.

  4. §As standard deviations of normal growth.

Pulmonary outcome     
 Chronic lung disease12(15)3(3)4.6 1.4, 16*
 Steroid treatment14(17)4(4)4.1 1.4, 12*
 Bronchodilatators9(11)4(4)2.6 0.8, 8.2
Duration of rehospitalisation (days)0[0.75]0[0.11] 0.04* 
Neurosensory development     
 Cerebral palsy5(6)10(11)0.6 0.2, 1.6
 Delayed motor development8(10)8(9)1.2 0.5, 3.0
 Visual disability5(6)3(3)1.9 0.5, 7.8
 Hearing loss5(6)3(3)1.9 0.5, 7.8
 Weight−4.1(8.1)−1.4(8.3) 0.03−5.3, −0.3*†
Length§−1.2[1.4]−0.6[1.2] 0.002−1.0, −0.2*†
 Head size§−0.4[1.1]−0.003[1.0] 0.03−0.7, −0.3*†


In this cohort study the objective was to evaluate the neonatal outcome, and especially long term pulmonary sequelae, in infants who required delivery (indicated preterm) as a result of maternal or fetal reasons. The results reveal the central role of respiratory problems in neonatal mortality and morbidity as well as in later morbidity, of infants. It was unexpected that the infants born for reasons indicated preterm had worse pulmonary outcome at the age of one year.

The greater age of women in the indicated group places them at risk of chronic diseases, such as essential hypertension. This has its ramifications in pregnancy complications, mainly pre-eclampsia. Age did not, however, appear to be a confounding factor in these women who were matched by gestational age. The defect of trophoblast invasion, an aetiological factor of pre-eclampsia, was seen histopathologically as placental fibrosis and infarction in both groups. This might indicate that the underlying aetiology of both pre-eclampsia and preterm birth really begins in premature failure of placentation9,10. Although half the women in the indicated group had pre-eclampsia, this was not the strongest determinant on later neonatal outcome.

Bleeding during the first trimester is a risk factor for both spontaneous and indicated preterm birth according to this study4. The fact that clinical infection at the time of delivery was seen significantly more often in women in the control group can be explained by the theory of an ascending infection leading to preterm birth. Infectious morbidity during the puerperium, however, was equal in both groups. The fact that there were more caesarean deliveries in the indicated group led to the risk of more post-operative infections in this group. Increased maternal age certainly made the frequency of chromosomal analyses and previous spontaneous abortions higher in women in the indicated group than in the control group. The higher neonatal mortality rate in the indicated group than in the control group was expected, because they were all high risk pregnancies, and the decision to deliver was made on the basis of the fact that the fetus was at a major risk of intrauterine death. In some cases the decision was made on an emergency basis. In a study carried out by Wolff et al.11 there were no significant differences in pre-discharge mortality rates among infant groups born preterm after various pregnancy complications. Although most of the pregnancies were complicated by pre-eclampsia in our study, they were not classified into groups according to pregnancy complications, but according to whether preterm delivery was indicated or spontaneous. The comparison between groups is somewhat troublesome because of the unequal distribution of the route of the delivery. It was impossible to find an indication to vaginal delivery in a situation were the fetus was already under extreme stress.

Early neonatal morbidity differed between the groups particularly as regards respiratory problems and hypoglycaemia. The higher rate of respiratory problems in the infants born to women in the indicated group, compared with those in the control group, was somewhat surprising, although it was at the same level as in the study by Wolff et al.11. Iannucci et al.12 found no differences in survival or frequency of respiratory distress syndrome in preterm infants in different aetiological groups, but in their study the infants were more immature (birthweight < 800 g) than in our study. Because the commonest cause of death in our series was a respiratory problem during the neonatal period, the question arises of whether in some cases it might have been possible to wait at least a few days for corticoids to act. In the early 1990s the use of steroids was not as common as at present.

Most fetuses in the indicated group had signs of chronic fetal distress, which is known to enhance lung maturation13. All infants in the indicated group were born by caesarean delivery without their mothers having spontaneous contractions. Although the presence of fetal lung fluid is essential for normal lung development, its clearance is equally essential for normal respiratory adaptation. The infants born by caesarean delivery without preceding spontaneous contractions lacked this stress reaction and thoracic compression effect during the second phase of labour, potentially leading to respiratory problems more commonly than in the control group. It is noteworthy that the most compromised fetuses (i.e. those having AREDV in the umbilical artery) had less respiratory distress syndrome but most of them were small for gestational age fetuses. Antenatal use of corticoids seemed to have no effect on the frequency of respiratory distress syndrome but even appeared to increase the incidence of chronic lung disease later. This might be explained by the fact that corticoids were administered more easily in these very preterm pregnancies.

Oligohydramnios detected by ultrasonography was a common finding in fetuses in the indicated group. This may result in compression of the fetal thorax and/or alteration of the dynamics of lung fluid, thus restricting lung growth and delaying maturation14. Intrapartal asphyxia, which was seen more often in the indicated group, inactivates surfactants and hence worsens pulmonary function15. It is not so obvious for the obstetrician to decide on the optimal timing of delivery. The significantly higher respiratory distress syndrome rate in the group where caesarean delivery was as a result of a maternal reason, compared with a fetal one, suggests that intrauterine fetal stress can have an improving effect on the neonatal pulmonary outcome.

The higher rate of neonatal infections in the control group, compared with the indicated group, was expected. Infections not only play a marked roll by inducing preterm labour, but they also affect the outcome of the infant16. Infection might have a direct impact at the tissue level, harming both the lungs and the brain16. We found that control group infants had a higher proportion of periventricular leukomalacia, although this difference did not reach a level of statistical significance. This could be explained at least partly by the greater amount of infections in that group compared with the indicated one.

Glucose, like oxygen, is of fundamental importance for the developing brain. However, definition of the level of blood glucose that should be considered too low is difficult, partly because the newborn cannot demonstrate by overt symptomatology when the critical low limit has been passed. A recent study demonstrated changes in brain stem auditory-evoked responses in infants who had blood glucose levels < 2.5 mmol/L (45 mg/100 mL)7. In our study we used the same limit, although this value was higher than that used earlier17. Because intrauterine malnutrition was a very common finding in the indicated group, a higher occurrence of hypoglycaemia was expected. According to our routine practice all sick infants and all infants with a birth-weight < 1500 g were given parenteral glucose support from birth until the infant's nutritional needs were being met and the glucose level stabilised. Thus, hypoglycaemia in most infants was only mild and transient without long term sequelae.

Symptomatic chronic lung disease up to the age of one year was more common in infants in the indicated group than in those in the spontaneous delivery group because of the higher rate of respiratory distress syndrome in the neonatal period. Chronic lung disease occurs almost exclusively in preterm infants who receive supplemental oxygen and mechanical ventilation with positive pressure as a treatment of respiratory distress syndrome. Hence, barotrauma and pulmonary oxygen toxicity appear to play an important role in the pathogenesis of chronic lung disease18. We found earlier that infants born after very preterm rupture the membranes had a worse late pulmonary outcome than their counterparts born spontaneously without preceding membrane rupture19. Although mechanisms behind chronic lung disease may be partly different in the infants of these two groups, both indicated preterm infants and those with very early rupture of the membranes appear to form a high risk group as regards the development of chronic lung disease.

The frequency of abnormal neurosensory development at one year did not differ between the groups, irrespective of the high incidence of respiratory problems in the indicated group, and was of the same magnitude as in other studies20. Respiratory problems, even if severe, do not appear to disturb the development of the central nervous system if treatment is adequate.

More infants with intrauterine growth retardation were seen in the indicated group than in those in the spontaneously delivered group. Some catch-up growth appeared to occur, but at one year of corrected age the infants in the indicated group were still smaller than their counterparts. Hence, the infants in the indicated group appeared to have a compromised growth potential compared with that of preterm infants as a whole21.

It can be concluded that indicated pre-maturity consists mainly of hypertensive disorders in the mother, the consequences of which on placental malfunction extend to the fetus causing intrauterine asphyxia. The indicated delivery appears to have a detrimental impact on pulmonary function in the fetus, in whom these problems appear to extend beyond the neonatal period. To find the optimum between intrauterine chronic asphyxia and extreme pre-maturity, however, is a problem which cannot be solved easily.