SEARCH

SEARCH BY CITATION

Keywords:

  • Apgar score;
  • caesarean section;
  • perinatal death;
  • Sweden

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References
  14. Journal Club
  15. References

Please cite this paper as: Carlsson Wallin M, Ekström P, Maršál K, Källén K. Apgar score and perinatal death after one previous caesarean delivery. BJOG 2010;117:1088–1097.

Objective  To assess the impact of the indication for a previous caesarean section on the outcome of a subsequent delivery.

Design  Population-based cohort study.

Setting  Sweden.

Population  Women with two deliveries between 1987 and 2007 identified using the Swedish Medical Birth Registry.

Methods  The outcome of 69 133 pregnancies after one caesarean section was compared with the outcome of 487 610 pregnancies following one vaginal delivery. The indication for the first caesarean section was estimated using a new hierarcharchical system based on information from birth records.

Main outcome measures  Perinatal death, low Apgar score (less than seven at 5 minutes).

Results  Infants of women with one previous caesarean section were at increased risk of low Apgar score compared with infants of women with one previous vaginal delivery (OR, 2.0; 95% CI, 1.9–2.1). The risk estimate was reduced when adjustment for maternal and fetal/infant characteristics was made (OR, 1.6; 95% CI, 1.5–1.8). The corresponding crude and adjusted odds ratios for perinatal death were 1.6 (95% CI, 1.4–1.7) and 1.1 (95% CI, 1.0–1.2), respectively. The infant outcome of the delivery after one caesarean section was mainly dependent on the indication for the first-delivery caesarean section and, when no medical indication was present, no increase in risk was detected.

Conclusions  Infants of women with one previous caesarean section were at increased risk of low Apgar score and/or perinatal death compared with infants of women with one previous vaginal delivery. The results suggest that medical conditions, not the previous caesarean section per se, contributed to the increase in risk.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References
  14. Journal Club
  15. References

The increasing frequency of deliveries by caesarean section (CS) has generated animated discussions regarding the benefits and harm of caesarean versus vaginal delivery. Studies have focused on the possible complications, immediate and delayed, that could threaten the mother and her infant in association with a CS.1,2 More recently, these investigations have been followed by several studies on delivery following one CS.3–9 It is obvious that every surgical intervention yields tissue damage. When a CS is performed in the first delivery, risks will be increased in the second pregnancy and delivery for, for example, rupture of the uterus, placenta praevia or placenta accreta. Furthermore, it has been suggested by Smith et al.3 that the risk of unexplained stillbirth in the pregnancy following a first CS is augmented. However, evidence remains to be explored regarding the infant outcome of delivery after one CS.

Sweden has, in concordance with many other countries, experienced an increase in the caesarean delivery rate over the years, from just below 11% in 1990 to 18% in 2007.6,10 Factors contributing to the increase in frequency include increased maternal age at the first delivery, increasing body mass index (BMI) among pregnant women, increasing proportion of elective CS for breech presentation, CS being performed on maternal request, and a trend among obstetricians to intervene more liberally.11 Furthermore, every CS performed increases the likelihood for a CS to be performed without any strict medical indication in the subsequent pregnancy.6,10 Thus, the perinatal outcome of the delivery following a CS is bound to be a matter of great importance in countries with increasing caesarean delivery rates.

Many complications appearing in the first pregnancy or birth are likely to recur in subsequent pregnancies.7 Therefore, it is essential that the underlying medical conditions be controlled for in an analysis of the perinatal outcome of the second delivery. By using retrospective data from the Swedish Medical Birth Registry, the present study focused on the outcome of deliveries preceded by a CS, compared with the outcome of deliveries preceded by a vaginal delivery, in terms of low Apgar score and perinatal death. To our knowledge, the influence of the indication for a CS to be performed at the first delivery on the perinatal outcome of the second delivery has not been investigated thoroughly.6,8 In the present study, a new hierarchical system using objective medical information was developed, and the main condition/indication for the first-delivery CS was assessed. For each indication group, the infant outcome of the subsequent pregnancy was compared with the infant outcome of the pregnancy following one vaginal delivery.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References
  14. Journal Club
  15. References

Women having experienced their first two deliveries between 1987 and 2007 were identified using the Swedish Medical Birth Registry,10 containing medical information on nearly all deliveries in Sweden (coverage about 99%).12 The obstetric history of each woman could be tracked using the unique Swedish personal identification number. Medical diagnoses were reported by the International Classification of Diseases (ICD) codes (1987–1997, ICD-9; 1997 onwards, ICD-10). In Sweden, standardised record forms are used at all antenatal clinics, all delivery units and all paediatric examinations of newborn infants. Copies of these forms are sent to the National Board of Health where they are computerised and linked with Statistics Sweden to obtain, among other things, information on the date of death.

Nearly all pregnant women receive free antenatal care. At the first visit to the antenatal clinic, each woman is interviewed by her midwife, and maternal height, weight and smoking habits are recorded. Thus, information regarding maternal BMI and smoking used in the present study refers to the first trimester of pregnancy. All women are offered a free routine ultrasound examination before 20 weeks of pregnancy, in order to estimate the expected date of delivery, and to check for multiple pregnancy and fetal anomalies.

Infants weighing less than two standard deviations (SDs) below the expected birth weight according to gestational age13 were defined as small for gestational age (SGA), whereas infants weighing more than two SDs above the expected birth weight were regarded as large for gestational age (LGA).

In the Swedish Medical Birth Registry, several diagnoses were recorded for the majority of the CSs registered. To avoid each CS appearing in more than one group, a hierarchical system was created. In this system, efforts were made to classify according to the underlying conditions, instead of focusing on conditions appearing during labour (e.g. fetal distress), in an attempt to diminish the subjective impact of diagnoses being recorded after delivery. Information on the medical indication for caesarean delivery was based on the reported ICD codes and well-defined medical information obtainable from registry data (see Table 1).

Table 1.   Hierarchical classification of indications/conditions
Classification groupSpecification
  1. ICD, International Classification of Diseases; SD, standard deviation.

  2. *Exceptions: congenital dislocation of hip/unstable hip, undescended testicle, persistent ductus arteriosus, single umbilical artery and preauricular tags.

 1. Multiple gestationInformation on type of gestation available in Swedish Medical Birth Registry
 2. Preterm birth, <37 weeksInformation on gestational length available in Swedish Medical Birth Registry
 3. Breech or other malpresentationInformation on fetal presentation available in Swedish Medical Birth Registry
 4. Congenital malformationsSignificant congenital malformations*
 5. Rupture of uterusICD9: 665.0–1; ICD10: O71.0–1
 6. Placenta praeviaICD9: 641.0–1; ICD10: O44.0–1
 7. Diabetes mellitus or gestational diabetesICD9: 648.0, 648.8; ICD10: O24.0–4, O24.9
 8. Small for gestational age/poor fetal growth<−2 birth weight SD scores according to a Swedish ultrasound-based weight curve13 or ICD9: 656.5; ICD10: O36.5
 9. Large for gestational age/excessive fetal growth>+2 birth weight SD scores according to a Swedish ultrasound-based weight curve13 or birth weight > 4500 g or ICD9: 656.6; ICD10: O36.6
10. Prolonged pregnancy ≥42 weeksInformation on gestational length available in Swedish Medical Birth Registry
11. Severe conditions complicating   pregnancy/severe maternal diseasePremature separation of placenta (ICD9: 641.2; ICD10: O45) Antepartum haemorrhage (ICD9: 641.3, 641.8–9; ICD10: O46) Pre-existing hypertension, severe pre-eclampsia, eclampsia (ICD9: 642.0–2, 642.5–7; ICD10: O10, O11, O14.0–1, O15.0–1, O15.9) Maternal renal, cardiovascular or thyroid disease (ICD9: 646.2, 648.1, 648.5–6; ICD10: E00–E07, I00–I99) Maternal care for known or suspected fetal abnormality/damage, including fetal–maternal haemorrhage, iso-immunisation and fetal hydrops (ICD9: 655, 656.0–2; ICD10: O35, O36.0–2) Polyhydramniosis or oligohydramniosis (ICD9: 657, 658.0; ICD10: O40, O41.0)
12. Complications during labour and deliveryDystocia, prolonged labour or suspected cephalopelvic disproportion (ICD9: 653, 660, 661, 662; ICD10: O33, O62, O63.0–1, O63.9, O66, O75.0, O75.8–9) Abnormalities of pelvic organs (ICD: 654.0–1, 654.3–9; ICD10: O34.0–1, O34.4–9, O65) Problems associated with amniotic cavity and membranes (ICD9: 658.1–2, 658.4–9, 659.2, 663.5; ICD10: O41.1–9, O69.4, O75.2, O75.6) Problems associated with induction of labour (ICD9: 658.3, 659.0–1; ICD10: O61.0–1, O75.5) Prolapse of cord, or cord entanglement with compression (ICD9: 663.0–2; ICD10: O69.0–2) Other complications of obstetric surgery and procedures (ICD10: O75.4)
13. Fetal distress or fetal death unexplained by conditions mentioned aboveFetal distress, fetal hypoxia, or birth asphyxia (ICD9: 656.3, 768; ICD10: O36.3, O68, P21.0) Intrauterine death (ICD9: 656.4, ICD10: O36.4)
14. Caesarean section (CS) without medical indication mentioned above, or mild conditions not classified elsewhereOther or not specified indication for CS (ICD9: 659.8–9, 669.7; ICD10: O82.8) Mild hypertension or mild pre-eclampsia (ICD9: 642.3–4, 642.9; ICD10: O13, O14.9) Oedema or excessive weight gain in pregnancy, without mention of hypertension (ICD9: 646.1; ICD10: O12) Peripheral neuritis, liver disorders, anaemia, bone and joint disorders of pregnancy (ICD9: 646.4, 646.7, 648.2, 648.7; ICD10: O99.0) Mild infections in pregnancy or labour (ICD9: 646.5–6; ICD10: O23, O75.3) Other conditions predominantly related to pregnancy (ICD9: 646.8–9, 656.7–9, 663.3–4, 663.6, 663.8–9; ICD10: O26, O36.8–9, O69.3, O69.5, O69.8–9)
15. No diagnosis availableNeither abnormal conditions, nor indications for CS could be detected using Swedish Medical Birth Registry data

The first three groups consisted of mutually exclusive, well-defined conditions (multiple gestation, preterm singleton birth and term singleton breech birth). The fourth group consisted of major congenital malformations. The following main group comprised all term singleton cephalic births without major malformations. This group was further divided using medical diagnoses combined with information on birth weight and gestational length. Groups 5 and 6 (rupture of the uterus and placenta praevia, respectively) consisted of severe conditions/diagnoses that were imperative indications for CS. Furthermore, the diagnosis ‘rupture of the uterus’ was of special interest for further planned studies, and was highlighted in the hierarchy instead of being combined with other severe labour complications. Groups 7–10 (diabetes or gestational diabetes, SGA, LGA or macrosomia, and pregnancy ≥42 weeks, respectively) were designed to show the impact of fetal growth independent of maternal diabetes, and the impact of prolonged pregnancy independent of fetal size, respectively. Group 11 (severe conditions complicating pregnancy/severe maternal disease) consisted of conditions for which CS is likely to be considered. Group 12 consisted of conditions arising during labour and delivery. Group 13 consisted of cases of fetal distress or fetal death, not explained by any of the conditions mentioned above. Group 14 comprised CSs for which medical information was available, but reflected mild conditions which were seldom judged to be true indications for CS. It also included ICD codes for ‘other or not specified indication for caesarean section’. Group 15 consisted of CSs for which the underlying indication could not be assessed – the records lacked ICD codes, and did not provide any other medical information implying an increased risk for CS.

The perinatal outcome of women who had experienced one previous CS was compared with that of women with one previous vaginal delivery. The outcomes studied were an Apgar score below seven at 5 minutes and perinatal death (stillbirth from 28 completed weeks of pregnancy or early neonatal death at 0–6 days).

Statistical methods

Statistical analyses were performed using Gauss (Gauss™; Aptec Systems Inc., Maple Valley, WA, USA; http://www.aptech.com). Odds ratios with 95% confidence intervals for low Apgar score or perinatal death were obtained using logistic regression analyses. For each indication, for the first caesarean delivery, the odd ratios for low Apgar score and perinatal death were computed using three different models: (a) crude odds ratio; (b) odds ratio adjusted for year of birth and maternal characteristics; and (c) odds ratio adjusted for year of birth, maternal characteristics and fetal/infant parameters. Year of birth was entered as a linear continuous variable in the logistic models. The ‘maternal characteristics’ were maternal age (continuous), smoking (semi-continuous; 1, no; 2, 1–9 cigarettes per day; 3, 10 or more cigarettes per day), height (cm, continuous) and BMI (kg/m2, continuous). The ‘fetal/infant’ parameters were restricted to ‘hard facts’, not dependent on delivery mode or hospital level: multiple birth (yes/no), preterm birth <37 weeks (yes/no), breech presentation (yes/no) and birth weight SD scores, according to a Swedish ultrasound-based growth curve13 (continuous, second grade model). Adjustment was performed for all possible confounders valid for the second delivery.

Nonparametric tests (Mann–Whitney U-test) were performed in order to compare the background characteristics of the study groups (women with one previous CS versus one previous vaginal delivery).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References
  14. Journal Club
  15. References

The demographic characteristics of women with one previous delivery are shown in Table 2. Compared with women with one previous vaginal delivery, women who had experienced one previous CS were more likely to give birth towards the end of the study period (P < 0.001), to be older (P < 0.001), to be shorter (P < 0.001), to have a higher BMI (P < 0.001) and to be delivered preterm (before 37 completed weeks of pregnancy) (P < 0.001). No apparent differences between the groups could be detected with regard to the interval between deliveries, maternal country of birth or maternal smoking. The infants of women who had experienced one previous CS were more likely to weigh less than two SDs below the expected birth weight according to gestational age (P < 0.001), or more than two SDs over the expected weight (P < 0.001), when compared with infants of women with one previous vaginal delivery. The gender distribution was similar between the groups.

Table 2.   Demographic characteristics. All women with at least two deliveries in Sweden between 1987 and 2007; data from the first and second deliveries as specified
 Caesarean first deliveryVaginal first delivery
N = 69 133 (%)*= 487 610 (%)*
  1. *Percentages based on records with obtainable information.

  2. **The pregnancy leading to the second delivery.

  3. ***Deviation in standard deviation (SD) from the expected birth weight according to Swedish growth standard curves.13

Year of delivery (second delivery)
1987–199212 494 (18.1)108 278 (22.2)
1993–199719 236 (27.8)152 178 (31.2)
1998–200219 461 (28.2)128 477 (26.3)
2003–200717 942 (26.0)98 677 (20.2)
Maternal age (years) (second delivery)
<20188 (0.3)2389 (0.5)
20–246846 (9.9)72 785 (14.9)
25–2922 243 (32.2)191 180 (39.2)
30–3426 081 (37.7)163 253 (33.5)
35–3911 761 (17.0)51 538 (10.6)
≥402014 (2.9)6465 (1.3)
Interval between deliveries (years)
0–118 910 (27.4)136 009 (27.9)
2–545 171 (65.3)315 928 (64.8)
6–94213 (6.1)28 860 (5.9)
≥10839 (1.2)6813 (1.4)
Maternal country of birth
Sweden60 399 (87.4)428 585 (87.9)
Other8734 (12.6)59 025 (12.1)
Maternal height (cm) (second pregnancy**)
Not known8397 (12.1)61 284 (12.6)
<1553195 (5.3)9191 (2.2)
155–16425 185 (41.5)142 315 (33.4)
165–17428 328 (46.6)232 278 (54.5)
≥1754028 (6.6)42 542 (10.0)
Maternal body mass index (BMI) (kg/m2) (second pregnancy**)
Not known17 931 (25.9)135 313 (27.8)
<204128 (8.1)40 833 (11.6)
20–2425 931 (50.6)200 701 (57.0)
25–2914 120 (27.6)81 046 (23.0)
30–344909 (9.6)22 012 (6.2)
35–391567 (3.1)5896 (1.7)
≥40547 (1.1)1809 (0.5)
Maternal smoking (second pregnancy**)
Not known4436 (6.4)26 897 (5.5)
No56 521 (87.4)399 432 (86.7)
<10 cigarettes per day5504 (8.5)41 002 (8.9)
≥102672 (4.1)20 279 (4.4)
Infant gender (second delivery)
Male35 692 (51.6)251 000 (51.5)
Female33 441 (48.4)236 610 (48.5)
Gestational length (completed weeks) (second delivery)
<28183 (0.3)756 (0.2)
28–31483 (0.7)1786 (0.4)
32–363586 (5.2)16 950 (3.5)
37–4160 220 (87.1)439 674 (90.2)
≥424661 (6.7)28 444 (5.8)
Birth weight deviation*** (second delivery)
<−3SD289 (0.4)1020 (0.2)
−3 to −2.1SD1600 (2.3)6049 (1.2)
−2 to −1.1SD7849 (11.4)46 933 (9.6)
−1 to −0.1SD21 048 (30.4)158 033 (32.4)
0–1SD22 961 (33.2)178 004 (36.5)
1.1–2SD10 839 (15.7)76 158 (15.6)
2.1–3SD3256 (4.7)16 691 (3.4)
>3SD1291 (1.9)4722 (1.0)

In the cohort of women who gave birth at least twice (in the current study, only data from the first two deliveries were included), the overall CS rate at the first delivery was 12.4%. The CS rate was 47.6% among women with multiple births, 29.0% among women with singleton preterm births, 69.0% in term singleton breech presentation and 9.2% in deliveries with term singleton cephalic presentation. The corresponding CS rate in the second delivery was 11.2% overall, 40.6% in multiple births, 34.0% in preterm singletons, 69.1% in term singleton breech presentation and 8.8% in term singleton cephalic presentation.

The overall caesarean delivery rate among women having undergone a CS at their first delivery was 50% (increasing gradually from 46% in 1987–1992 to 55% in 2003–2007). The corresponding percentage among women with one previous vaginal delivery was 5.7%.

In Table 3, the delivery indications/conditions of all women experiencing a CS at their first delivery are shown. The recurrence of indications/conditions in the second delivery, whether vaginal or by CS, is then shown, first using all the diagnoses given for each woman at the second delivery and, second, for comparison, using the hierarchical system. In order to highlight the degree of recurrence, Table 3 also shows the crude odds ratios for second delivery conditions/diagnoses among women with the same conditions/diagnoses in both deliveries versus women having a different condition/diagnosis at the second delivery from the indication for the first-delivery CS.

Table 3.   Delivery indications/conditions of all women experiencing a caesarean section (CS) at their first delivery, and recurrence of indications/conditions in the second delivery, whether vaginal or by CS, first using all the conditions/diagnoses of each woman at the second delivery, and then using the hierarchical system
 Proportion with this CS indication at 1st delivery and repeated condition/diagnosis at 2nd deliveryProportion without this CS indication at 1st delivery but with this condition/diagnosis at 2nd deliveryOR for recurrence: this indication at 1st delivery versus another indication at 1st delivery
n/N (%)n/N (%)OR (95% CI)
  1. CI, confidence interval; IUGR, intrauterine growth restriction; OR, odds ratio; SD, standard deviation.

  2. *Small for gestational age: birth weight below the mean − 2SDs of the Swedish growth standard curve.13

  3. **Large for gestational age: birth weight above the mean + 2SDs of the Swedish growth standard curve.13

Nonhierarchical system
Multiple gestation46/1674 (2.7)932/67 459 (1.4)2.0 (1.5–2.7)
Preterm birth, <37 weeks1826/10 494 (17.4)2426/58 639 (4.1)4.9 (4.6–5.2)
Breech or other malpresentation, ≥37 weeks1225/13 962 (8.8)2104/55 171 (3.8)2.4 (2.3–2.6)
Congenital malformations106/2592 (4.1)1611/66 541 (2.4)1.7 (1.4–2.1)
Rupture of uterus0/33 (0.0)552/69 100 (0.8)
Placenta praevia22/742 (3.0)298/68 391 (0.4)7.0 (4.8–10.2)
Diabetes mellitus/gestational diabetes944/1373 (68.8)907/67 760 (1.3)162.2 (152.0–173.1)
SGA*/suspected IUGR1016/7006 (14.5)1328/62 127 (2.1)7.8 (7.2–8.4)
LGA**/suspected LGA/>4500 g1760/4606 (38.2)3991/64 527 (6.2)9.4 (8.9–9.9)
Prolonged pregnancy, ≥42 weeks1285/8871 (14.5)3376/60 262 (5.6)2.9 (2.7–3.0)
Severe pregnancy complications/severe maternal disease612/6420 (9.5)1240/62 713 (2.0)5.2 (4.8–5.7)
Complications during labour/delivery15 691/38 500 (40.8)7754/30 633 (25.3)2.0 (2.0–2.1)
Fetal distress/death2966/18 858 (15.7)4787/50 275 (9.5)1.8 (1.7–1.9)
Hierarchical system
Multiple gestation46/1674 (2.7)932/67 459 (1.4)2.0 (1.5–2.7)
Preterm birth, <37 weeks1677/9584 (17.5)2212/59 549 (3.7)5.5 (5.2–5.8)
Breech or other malpresentation, ≥37 weeks834/11 408 (7.3)1730/57 725 (3.0)2.6 (2.4–2.8)
Congenital malformations50/1412 (3.5)1320/67 721 (1.9)1.8 (1.4–2.4)
Rupture of uterus0/24 (0.0)513/69 109 (0.7)
Placenta praevia4/325 (1.2)125/68 808 (0.2)6.8 (2.9–16.2)
Diabetes mellitus/gestational diabetes468/896 (52.2)934/68 237 (1.4)78.8 (72.7–85.3)
SGA/suspected IUGR269/2871 (9.4)1103/66 262 (1.7)6.1 (5.4–6.9)
LGA/suspected LGA/>4500 g1116/3474 (32.1)3514/65 659 (5.4)8.4 (7.8–9.0)
Prolonged pregnancy, ≥42 weeks819/6658 (12.3)2862/62 475 (4.6)2.9 (2.7–3.2)
Severe pregnancy complications/severe maternal disease91/1805 (5.0)806/67 328 (1.2)4.4 (3.6–5.4)
Complications during labour/delivery5803/17 516 (33.1)9363/51 617 (18.1)2.2 (2.2–2.3)
Fetal distress/death506/5701 (8.9)2681/63 432 (4.2)2.2 (2.0–2.4)

For the hierarchical system, as well as for the nonhierarchical system, strong associations between the conditions/diagnoses of the first and second delivery were found. In addition to the self-evident correlation for diabetes at the first and second delivery, strong associations with odds ratios above 2.0 (according to both the hierarchical and nonhierarchical systems) were found for all conditions/diagnoses, except for congenital malformations. The strongest correlation between first and second delivery conditions/diagnoses was seen for LGA/excessive fetal growth, SGA/intrauterine growth restriction (IUGR) and placenta praevia.

Table 4 shows the risk of an Apgar score below seven at 5 minutes among second-delivery infants of women having experienced one previous CS or one previous vaginal delivery. Infants of women having undergone a CS at the first delivery (regardless of indication) were twice as likely to have a low Apgar score as were infants of women who had experienced a previous vaginal delivery. When the odds ratio was adjusted for year of birth and maternal characteristics (age, BMI, height and smoking habits), the odds ratio for a low Apgar score – previous CS versus previous vaginal delivery – was marginally lowered. The corresponding odds ratio was further reduced, but still strongly significant, when also adjusting for fetal/infant parameters (multiple birth, preterm birth, breech presentation and birth weight SD score).

Table 4.   Association between the mode of the first delivery and Apgar score below seven at 5 minutes at the second delivery. Crude and adjusted odds ratios as specified
 Apgar score at second deliveryCrude ORAdjusted OR*Adjusted OR**
<7 n (%)≥7OR (95% CI)OR (95% CI)OR (95% CI)
  1. CI, confidence interval; IUGR, intrauterine growth restriction; LGA, large for gestational age; OR, odds ratio; SGA, small for gestational age.

  2. *Adjusted for year of birth and maternal characteristics (maternal age, smoking, height and body mass index), second delivery.

  3. **Adjusted for year of birth, maternal characteristics and fetal/infant parameters (multiple birth, preterm birth, breech presentation and birth weight standard deviation score), second delivery.

First delivery vaginal (reference)3938 (0.8)483 6721.01.01.0
First delivery by caesarean section (CS), regardless of indication1108 (1.6)68 0252.0 (1.9–2.1)1.8 (1.7–1.9)1.6 (1.5–1.8)
Conditions/diagnoses, first-delivery CS (hierarchical order)
Multiple gestation25 (1.5)16491.9 (1.3–2.8)1.7 (1.1–2.5)1.6 (1.0–2.3)
Preterm birth, <37 weeks179 (1.9)94052.3 (2.0–2.7)2.2 (1.9–2.5)1.1 (0.9–1.3)
Breech or other malpresentation, ≥37 weeks172 (1.5)11 2361.9 (1.6–2.2)1.8 (1.5–2.1)1.7 (1.4–1.9)
Congenital malformations28 (2.0)13842.5 (1.7–3.6)2.2 (1.5–3.1)2.2 (1.5–3.2)
Term singleton cephalic
All704 (1.6)44 3512.0 (1.8–2.1)1.7 (1.6–1.9)1.8 (1.6–1.9)
Rupture of uterus0 (0.0)240 (NS)
Placenta praevia3 (0.9)3221.2 (0.4–3.6)1.1 (0.4–3.3)0.9 (0.3–3.0)
Diabetes mellitus/gestational diabetes21 (2.3)8752.9 (1.9–4.5)2.4 (1.5–3.6)1.8 (1.1–2.8)
SGA/suspected IUGR64 (2.2)28072.8 (2.2–3.6)2.4 (1.9–3.1)2.1 (1.6–2.7)
LGA/suspected LGA/>4500 g44 (1.3)34301.6 (1.2–2.1)1.4 (1.0–1.9)1.6 (1.2–2.2)
Prolonged pregnancy, ≥42 weeks100 (1.5)65581.9 (1.5–2.3)1.6 (1.3–2.0)1.9 (1.5–2.3)
Severe pregnancy complications/severe maternal disease41 (2.3)17642.9 (2.1–3.9)2.6 (1.9–3.6)2.4 (1.8–3.4)
Complications during labour/delivery260 (1.5)17 2561.9 (1.6–2.1)1.6 (1.4–1.8)1.7 (1.5–1.9)
Fetal distress/death unexplained by conditions mentioned above107 (1.9)55942.3 (1.9–2.9)2.1 (1.7–2.6)2.2 (1.8–2.6)
CS without indication mentioned above, or mild conditions not classified elsewhere34 (1.0)35191.2 (0.8–1.7)1.1 (0.8–1.5)1.0 (0.7–1.5)
No diagnosis available30 (1.3)22021.7 (1.2–2.4)1.6 (1.1–2.3)1.5 (1.0–2.1)

When investigating the relationship between previous CS and risk of perinatal death in the second pregnancy/delivery (Table 5), it was apparent that the adjustment for maternal as well as fetal/infant characteristics had a substantial influence on the estimated odds ratio. The crude odds ratio for perinatal death, regardless of the indication for the first CS, was highly statistically significant, and was only marginally reduced after adjustment for maternal characteristics. When fetal/infant variables were also adjusted for, no statistically significant association could be detected between previous caesarean delivery and perinatal death in the second pregnancy/delivery.

Table 5.   Association between the mode of the first delivery and perinatal death at the second delivery. Crude and adjusted odds ratios as specified
 Perinatal outcome at second deliveryCrude OR Adjusted OR* Adjusted OR**
Perinatal death, n (%)Survival, 7 days
OR (95% CI)OR (95% CI)OR (95% CI)
  1. CI, confidence interval; IUGR, intrauterine growth restriction; LGA, large for gestational age; NS, not significant; OR, odds ratio; SGA, small for gestational age.

  2. *Adjusted for year of birth and maternal characteristics (maternal age, smoking, height and body mass index), second delivery.

  3. **Adjusted for year of birth, maternal characteristics and fetal/infant parameters (multiple birth, preterm birth, breech presentation and birth weight standard deviation score), second delivery.

  4. ***P < 0.05.

First delivery vaginal (reference)1965 (0.4)485 6451.0 (–)1.0 (–)1.0 (–)
First delivery by caesarean section (CS), regardless of indication433 (0.6)68 7001.6 (1.4–1.7)1.5 (1.3–1.6)1.1 (0.97–1.2)
Conditions/diagnoses, first-delivery CS (hierarchical order)
Multiple gestation15 (0.9)16592.2 (1.3–3.7)2.1 (1.3–3.5)1.9 (1.1–3.3)
Preterm birth, <37 weeks110 (1.1)94742.9 (2.4–3.5)2.7 (2.2–3.3)0.7 (0.6–0.9)
Breech or other malpresentation, ≥37 weeks55 (0.5)11 3531.2 (0.9–1.6)1.2 (0.9–1.6)1.1 (0.8–1.4)
Congenital malformations11 (0.8)14011.9 (1.1–3.5)1.8 (0.97–3.2)1.8 (0.95–3.3)
Term singleton cephalic
All242 (0.5)44 8131.3 (1.2–1.5)1.2 (1.1–1.4)1.3 (1.1–1.5)
Rupture of uterus0 (0.0)240 (NS)
Placenta praevia1 (0.3)3240.8 (0.1–5.4)0.7 (0.1–5.3)0.7 (0.1–5.1)
Diabetes mellitus/gestational diabetes8 (0.9)8882.2 (1.1–4.5)2.0 (0.97–3.9)1.6 (0.8–3.4)
SGA/suspected IUGR35 (1.2)28363.1 (2.2–4.3)2.7 (2.0–3.9)1.4 (1.0–2.1)
LGA/suspected LGA/>4500 g12 (0.3)34620.9 (0.5–1.5)0.8 (0.4–1.4)1.4 (0.8–2.5)
Prolonged pregnancy, ≥42 weeks34 (0.5)66241.3 (0.9–1.8)1.2 (0.8–1.6)1.6 (1.1–2.3)
Severe pregnancy complications/severe maternal disease14 (0.8)17911.9 (1.1–3.3)1.7 (1.0–3.0)***1.4 (0.8–2.4)
Complications during labour/delivery76 (0.4)17 4401.1 (0.9–1.4)1.0 (0.8–1.3)1.2 (0.9–1.5)
Fetal distress/death unexplained by conditions mentioned above34 (0.6)56671.5 (1.1–2.1)1.4 (1.0–2.0) NS1.4 (1.0–1.9) NS
CS without indication mentioned above, or mild conditions not classified elsewhere14 (0.4)35391.0 (0.6–1.7)1.0 (0.6–1.7)0.8 (0.5–1.4)
No diagnosis available14 (0.6)22181.6 (0.9–2.6)1.4 (0.8–2.3)1.3 (0.7–2.2)

Tables 4 and 5 focus on the perinatal outcome by indication for the first-delivery CS. It is evident that, among infants of women having experienced a CS at their first delivery, the outcome of the second pregnancy/delivery varied considerably with the indication for the first-delivery CS. For low Apgar score, as well as for perinatal death, the P value for homogeneity of the odd ratios across indications strata was <10−6. Crude odd ratios for low Apgar score at the second delivery were increased significantly for all classes of medical indications for the previous CS, except for placenta praevia. In contrast, infants of women who had a previous CS without mention of indication, or a previous CS indicated by minor complications, did not have a significantly increased risk for low Apgar score, compared with infants of women who had experienced a previous vaginal delivery. Adjustment for the year of birth and maternal characteristics only marginally lowered the odds ratios for low Apgar score among the respective conditions/indications.

The impact of adjustment for fetal/infant parameters varied among the subgroups of indications for the previous CS; for example, the odds ratio for a low Apgar score at the second delivery was not increased significantly for infants of women who had a previous CS before 37 weeks of pregnancy when fetal parameters were considered. Among women who had diabetes as the indication for the first-delivery CS (but no indication for CS higher in the hierarchy), the odds ratio for low Apgar score at the second delivery was substantially lowered, but still significant when fetal parameters were adjusted for. For the other classes of indications, adjustment for fetal/infant parameters only marginally altered the magnitude of the association between a previous CS and the risk of a low Apgar score at the subsequent delivery.

For several indication classes, a significant association between delivery mode at the first delivery and perinatal death at the second delivery was found. For women without a strict medical indication for CS performed at the first delivery, no association between this first delivery by CS and perinatal death at the second delivery could be detected.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References
  14. Journal Club
  15. References

An infant being born at the second delivery to a woman having had a CS performed at the first delivery was found to have an increased risk for a low Apgar score and/or perinatal death compared with an infant born to a woman having experienced a previous vaginal delivery. The significance of a low Apgar score has been reported by others.14 In order to examine to what extent the indication for the first CS affected the perinatal outcome of the second delivery, a hierarchical system of caesarean delivery indications and pregnancy/delivery conditions was developed. Following this hierarchy, it was revealed that, for most classes of indications, there was a substantial risk for recurrence in the second pregnancy/delivery. It became evident that the conditions/indications present at the first delivery must be considered when evaluating the impact of the mode of the first delivery on the perinatal outcome of the second delivery.

The results suggested that the indication for the first caesarean delivery had a major influence on the risk for a low Apgar score or perinatal death at the second delivery. Compared with the perinatal outcome following the first vaginal delivery, the risk for a low Apgar score was more than doubled at deliveries following a first CS with a main indication of congenital malformations, suspect intrauterine growth restriction, severe pregnancy complications or unexplained fetal distress during labour. When the first delivery by CS was performed without a strict medical indication, no increased risk for a low Apgar score or perinatal death at the second delivery was found. For both outcomes, the risk estimates were close to unity. However, in spite of the large numbers, the upper confidence limits for the risk estimates for low Apgar scores and perinatal death were 1.5 and 1.4, respectively. Thus, even if unlikely, a weak association between a previous CS and an adverse outcome in the subsequent delivery could not be ruled out.

In addition to the impact of the indication for the first-delivery CS on the outcome of the second delivery, it was shown that maternal baseline characteristics had a substantial influence on the risk estimates. In concordance with the report by Taylor et al.,9 women having undergone a previous caesarean delivery were older than women having experienced a previous vaginal delivery. In the present study, women having experienced a CS at their first delivery also had a higher BMI and smoked more than women giving birth vaginally.

Fetal/infant parameters were also found to have an important influence on the outcome of the second delivery. After adjusting the outcome results for year of birth and maternal characteristics, the risk was still increased for both low Apgar score and perinatal death. However, after adjusting for certain fetal/infant parameters (multiple birth, preterm birth, breech presentation and birth weight SD score), only the risk for an Apgar score below seven at 5 minutes remained significantly increased.

Despite the growing rates of caesarean delivery, there are few published studies that have focused on the perinatal outcome of deliveries following one previous CS. The available literature does not address all of the subgroups analysed in this study. Richter et al.5 found a similar odds ratio for perinatal death at the second delivery (CS at first delivery versus vaginal first delivery; 1.39 from German data 1993–1995), but they did not study the outcome by indication for the CS at first delivery. After having performed a systematic review of the safety of vaginal birth after caesarean delivery, Guise et al.15 stressed that the identification of high-risk and low-risk groups of women and settings for morbidity remains a key research priority.

In concordance with the results of the present study, Hemminki et al.7 and Daltveit et al.6 found that women with a previous caesarean delivery were more likely than women with a previous vaginal delivery to suffer from miscellaneous complications during their next pregnancy. Hemminki et al.7 reported poorer infant outcome of deliveries after the first CS, even when women with persistent problems were excluded, and concluded that their findings were unlikely to be explained entirely by the CS indications. No results were presented of the perinatal outcome with regard to the indication for the previous caesarean delivery. Daltveit et al.6 concluded that an increased risk of complications after a caesarean delivery might be caused by CS itself, but could also be a result of confounding by indication. The authors stated that, in analyses of outcomes after previous CS, indications for the first CS should be considered. However, no such results were reported.

A limitation of this study was that the conditions and diagnoses presented in the medical records were lacking for some patients. Different methods of diagnosis and different indications might have been used in different regions of Sweden. We tried to diminish this problem by using as many conditions as possible determined by ‘hard facts’ in the medical records, rather than using the diagnoses/indications given by obstetricians. The hierarchical system was constructed to support this strategy, with the ‘self-given’ conditions higher up the hierarchy.

Another fact that should be kept in mind is that, in the present study, we chose to compare women with one previous CS with women who had experienced one vaginal delivery. This design might have disfavoured the results of the CS group, as a previous vaginal delivery is well known to facilitate a second delivery.16,17 Furthermore, the putative harmful maternal effects of a first delivery by CS were not evaluated in this study.

A strength of this study is that it is population based, involving a large number of women/births from the whole country. As a low Apgar score and, especially, perinatal death are rare events in an industrialised country, large amounts of data are required to detect possible differences. As every delivery unit in Sweden sends its statistical data to the Swedish Medical Birth Registry, it was possible to analyse the results from a mixture of obstetric departments.

In the present study, the CS rate after a first delivery by CS was about 50%. Similar rates have been reported from Finland7 and Norway,6 but substantially different results have been reported recently from the USA, where the frequency of repeated CS was nearly 90% in 2006.15

The determination of the influence of the mode of the second delivery on the perinatal outcome was beyond the scope of the current study. Probably, the choice of delivery mode at the second delivery has a substantial influence on the perinatal outcome. In future studies of deliveries following one previous CS, we intend to identify women belonging to high- and low-risk groups, considering the indication for the first-delivery CS. Our results could then be used to develop evidence-based guidelines for obstetricians counselling women and their partners about the preferable mode for the second delivery. This should, in turn, provide a possibility to influence the caesarean delivery rates17,18 without jeopardising infant health.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References
  14. Journal Club
  15. References

Infants of women having experienced one previous caesarean delivery had almost double the risk for a low Apgar score and perinatal death compared with infants of women with one previous vaginal delivery. The association could be explained, to some extent, by the underlying maternal and fetal baseline characteristics, as adjustments for these factors considerably lowered the overall risk estimates. However, the magnitude of the association varied significantly with the indication for the first-delivery CS. The results suggest that medical conditions, not the previous CS per se, contributed to the increase in risk.

Contribution to authorship

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References
  14. Journal Club
  15. References

MCW performed the analyses and contributed to the writing of the manuscript, PE contributed to the writing of the manuscript, KM contributed to the writing of the manuscript and KK performed the analyses and contributed to the writing of the manuscript.

Details of ethics approval

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References
  14. Journal Club
  15. References

This study was approved by the Regional Research Ethics Committee in Lund, 16 August 2007, reference number 309/2007.

Funding

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References
  14. Journal Club
  15. References

Funding was provided by the Evy and Gunnar Sandberg Foundation, the Birgit and Sven Håkan Olsson Foundation, and grants to researchers in the Public Health Service from the Swedish government.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References
  14. Journal Club
  15. References

We acknowledge the Evy and Gunnar Sandberg Foundation, the Birgit and Sven Håkan Olsson Foundation, and grants to researchers in the Public Health Service from the Swedish government.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References
  14. Journal Club
  15. References

Journal Club

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References
  14. Journal Club
  15. References

Discussion points

  • 1
     Background: Describe the rate of caesarean section (CS) in your area/country and compare with worldwide rates and trends over time.

Discuss the evidence from previous studies1,2 with regard to outcome in subsequent pregnancies for women with previous CS. How is this current study different in its stated aims and approach?

  • 2
     Methods: Discuss the value of the Apgar score as a proxy marker for short- and long-term neonatal outcome; discuss its advantages and disadvantages compared with perinatal death.

Describe the hierarchical system used by the authors to classify the indication for CS and comment on its usefulness and validity.

  • 3
     Results and implications: Describe the odds of recurrence for the maternal and fetal conditions recorded in this study.

Discuss whether, according to this study, CS per se or the indication for CS is a more important risk factor for adverse outcome in subsequent pregnancies, with reference to both statistical significance and effect size.

Will the findings of this study change the way you counsel women who are about to have their first CS, depending on the indication? Will they change your plan of care for women having a second pregnancy after a CS?

D Siassakos Southmead Hospital, Bristol, UK Email jsiasakos@gmail.com

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References
  14. Journal Club
  15. References