Today, falling fertility rates and aging populations are of concern both in the United States1 and Europe.2 The causes of such demographic trends are complicated, and probably both socioeconomic and biological factors play a part.3
At the same time, cesarean delivery rates are rising world-wide, in Norway from 2.0% in 1968 to 15.4% in 2004.4 An association between cesarean delivery and reduced subsequent fertility has been debated. A few authors have claimed that cesarean delivery does not negatively affect fertility,5,6 while most have reported reduced fertility after cesarean delivery compared with vaginal delivery.7–17 Some have concluded this was a physical consequence of the surgery itself,7,8 others that it was mainly the woman's decision to avoid subsequent pregnancies,15,17 and yet others that both these mechanisms were involved.9 A third possible mechanism was pointed out by Hemminki14; all studies suffer from potential selection bias or confounding by indication, because it is not random which women have a cesarean in the first place. In addition, the cause–effect relationship between cesarean delivery and fertility is not straightforward, as women with a history of infertility have a higher risk of cesarean delivery.18,19
The aim of the present study, using data from the population-based Medical Birth Registry of Norway, was first to assess if cesarean delivery was associated with reduced subsequent fertility. If so, we wanted to investigate whether the cesarean per se, or factors related to it, was the most likely explanation. We also wanted to investigate whether the magnitude of an association had changed over time.
MATERIALS AND METHODS
The Medical Birth Registry of Norway was established in 1967 and is based on compulsory notification of all live births and stillbirths in the country from 16 weeks of gestation. A standardized form, comprising demographic data of the parents and newborn, maternal health before and during index pregnancy, interventions and complications during delivery, and the condition of the newborn, is filled in by the midwife or attending physician.20 The Medical Birth Registry of Norway is routinely linked to the Central Population Registry for national identification numbers and dates of deaths. Data on maternal education were obtained from Statistics Norway.
The national identification number was used to link births to their mothers, providing sibship files with the mother as the unit of analyses. We had data on the reproductive history of all 719,544 women who had their first child during 1967–1996, followed up through 2003. Excluding women who died before age 50 and mothers with new partners between pregnancies and multiple pregnancies, 114,127 women had one child, 276,081 had two children, and 206,133 had three or more children. To analyze temporal trends in light of the vast increase in cesarean delivery over this 37-year period, we stratified the material into two periods of time with equal lengths of follow-up; women who had their first child during 1967–1981 were followed up through 1988, and women who had their first child during 1982–1996 were followed up through 2003.
We compared rates of continuation to a subsequent birth (continuation rates) according to the mode of the previous delivery (cesarean compared with vaginal) for all women, for the subgroups of women with preeclampsia or breech delivery, and for women in an obstetric low-risk group (excluding women with pregestational diabetes, preeclampsia, breech presentation, premature rupture of membranes, placenta previa, abruptio placentae, age below 20 years or above 35 years, and children with gestational age below 37 weeks or above 42 weeks, birth weight below 2,500 grams or above 4,500 grams or any congenital anomaly). In all categories, analyses were stratified according to infant survival the first year.
All calculations were performed using SPSS v14.0.2 (SPSS Inc., Chicago, IL). Relative risks (RRs) of continuation with 95% confidence intervals (CIs) were calculated, tested by χ2 tests. The analyses were stratified according to maternal level of education (less than 10 years, 10–12 years and 13 or more years) and maternal age (less than 20 years, less than 25 years, more than 30 years, more than 35 years, above or below median age) to check for confounding and interactions. The statistical significance of interaction terms was tested in logistic regression models (maternal age categorized as less than 25 years, 25–34 years, and 35 or more years), using likelihood ratio tests. As interbirth intervals were not normally distributed, their differences were tested with t tests on a log-transformed scale. Two-sided P values below .05 were considered statistically significant.
The study was based on de-identified data from the Medical Birth Registry of Norway. Such studies are exempt from institutional review board approval in Norway.
In first births in our study population, cesarean delivery rates increased from an average of 5.2% during 1967–1981 to 12.9 % during 1982–1996. The stillbirth rate was 11.9 per 1,000, and the infant mortality rate 9.4 per 1,000 during 1967–1981, while the corresponding rates were 7.0 per 1,000 and 4.9 per 1,000 during 1982–1996. During 1982–1996, the median interbirth interval, that is, the interval between two successive births, was 34.8 months after a vaginal delivery and 35.9 months after a cesarean delivery (Table 1), a statistically significant difference (P<.001). If the newborn died, the interval was 16.2 months after vaginal delivery and 18.0 months after cesarean delivery (P=.001).
If the infant survived the first year of life, the rate of continuation to a subsequent birth (continuation rate) was consistently lower after a cesarean delivery than after a vaginal delivery (Fig. 1A). The absolute difference was greatest in the earliest period, 18%, and gradually diminished to stabilize at around 9% from 1982–1986 onward. If the neonate was stillborn or died during the first year of life, continuation was initially lower after cesarean delivery, but the absolute difference was only 7% during 1967–1971. The difference gradually diminished, and from 1982–1986 onward, the continuation rate was the same regardless of mode of delivery (Fig. 1B). Thus, two major points are illustrated: First, the association between cesarean delivery and reduced fertility was weaker (or nonexistent) if the neonate was stillborn or died, and second, the effect of cesarean delivery on continuation rates decreased over time. This was observed in all deliveries, in subgroups of women with preeclampsia or breech delivery, in the obstetric low-risk group (Table 2) and when we analyzed continuation from the second to the third birth (Table 3).
Figure 2 shows the risk of cesarean delivery and continuation rate to a subsequent birth by previous mode of delivery. For instance, the average risk of cesarean delivery in nulliparous women during our study period was 9%. Among women with a first cesarean delivery, 70% continued to a second delivery, and their risk of a repeat cesarean was 52%. Among women with a vaginal first delivery, 82% continued to a second delivery, and their risk of this being cesarean was 4%.
Continuation to a third pregnancy was related to the most recent mode of delivery (Fig. 2). Among the women who had two children, 28% continued to a third birth after two cesarean deliveries, while 27% continued after a vaginal delivery followed by a cesarean delivery. Among the women who first had a cesarean and then a vaginal delivery, 37% continued to a third birth, which was close to the continuation rate after two vaginal deliveries, namely 40%. Also, we observed that the a priori probability of a woman having three or more children was 32% (0.82×0.40) if her first two deliveries were vaginal compared with 19% (0.70×0.28) if her first two deliveries were cesarean.
In the strata of surviving infants, we observed an increasing effect of cesarean delivery on continuation rates with increasing maternal age. This interaction was statistically significant (P<.05). For instance, for women who had their first child by cesarean delivery before 25 years of age during 1967–1981, the RR of continuation when compared with a vaginal birth was 0.92 (95% CI 0.91–0.93), increasing to 0.96 (0.95–0.97) during 1982–1996. For women who had their first child by cesarean delivery at age 35 years or older, the RR of continuation was 0.68 (0.62–0.76) during 1967–1981, and 0.72 (0.68–0.77) during 1982–1996. If the neonate was stillborn or died, there was no interaction between cesarean delivery and maternal age, and continuation rates were independent of mode of delivery during 1982–1996 for mothers of all ages.
Stratification of analyses on maternal educational levels did not alter the results.
The association between mode of delivery and continuation to a subsequent birth was dependant on whether the infant survived the first year of life or not. If the infant survived, a lower rate of continuation to a subsequent birth (continuation rate) was observed after cesarean delivery. If the infant was stillborn or died, continuation was independent of mode of delivery from 1982 onward. This was found in all deliveries, deliveries in an obstetric low-risk group, in subgroups of women with preeclampsia or breech presentation, for continuation from first to second birth, and for continuation from second to third birth. The findings suggest that the reduced fertility after cesarean delivery cannot be completely explained by the indication or any physical consequence of the cesarean. Rather, it is probably, to a large degree, voluntary.
Important strengths of our study include the large, population-based cohort of women and the long period of follow-up. Even though perinatal and infant deaths are rare in Norway, the size of the cohort allowed us to stratify on infant survival, which proved to be of great importance. Also, the long period of follow-up allowed us to study continuation from second to third birth, which few authors have done before.
Lack of data on the exact indication of cesarean delivery represents a weakness, as confounding by indication is possibly present. In other words, the indication for the cesarean delivery, and not the cesarean per se, might end a woman's reproductive career. To reduce this problem, we stratified the analyses on two important indications (preeclampsia and breech presentation) as well as an obstetric low-risk group, which did not alter the results. However, possible residual confounding by indication cannot be completely excluded in our study.
Most authors agree that cesarean delivery is associated with reduced subsequent fertility.7–17 These authors have usually excluded stillbirths and early deaths, and their results agree with our findings on families with surviving infants. Albrechtsen et al17 stratified analyses on perinatal outcome of the first child and concluded that the mother's wish to continue to a subsequent birth was influenced by cesarean delivery, per se, and not just the circumstances leading up to it.
A few authors have concluded that cesarean delivery is not associated with reduced subsequent fertility. The cohort of Tower et al6 consisted of 1,152 primiparous women, among whom half had been delivered by cesarean, and the follow-up period was 5 years. This small cohort with such limited follow-up time, and the fact that the overall continuation rate to a second birth was only about 60%, which is much lower than in our cohort, make our studies hard to compare. Smith et al5 found a statistically significant association between primary cesarean delivery in the early 1980s and reduced probability of a second birth but concluded that this was due to residual confounding.
The number of pregnancies and children a woman has during her reproductive career is influenced by a complex set of determinants. Having twins for instance, reduces the likelihood of a new pregnancy.21 Hence, we excluded multiple pregnancies. A change of partner might influence the probability of another pregnancy, thus we only analyzed families where both parents were the same for all the studied pregnancies. Chronic illness, serious complications during pregnancy or having a child with a congenital anomaly might influence a woman's wish to have further children. We therefore selected an obstetric low-risk group, from which such women were excluded.
The loss of a child may cause a desire for a new pregnancy to replace the loss.22 Previous studies have therefore excluded stillbirths and early losses. However, by stratifying on infant loss instead, we gained new insight: cesarean delivery was more strongly associated with reduced continuation to a subsequent pregnancy if the child survived; if the child died, continuation was independent of mode of delivery from 1982 onwards. This suggests that the association between cesarean delivery and reduced fertility is not completely explained by confounding by indication, nor does a cesarean delivery on average leave a woman physically less able to have more children. Rather, we conclude that a cesarean delivery makes a woman more likely to voluntarily end her reproductive career. However, should she experience the trauma of losing a child, the wish to replace that loss is stronger than the negative influence of the cesarean delivery.
We observed a temporal trend in which cesarean delivery was associated with a less-negative influence on continuation rates during 1982–1996 than during 1967–1981. The primary cesarean delivery rate increased from 5% on average in the early period to 13% in the last; thus, the indications for cesarean delivery have become broader over these decades. The women delivered by cesarean in the early period were more likely to have serious conditions that per se might reduce the probability of further pregnancies; thus, our results were probably partly confounded by indication. In recent years, many women who have cesarean deliveries would probably have had vaginal deliveries at an earlier point in time, yet reduced continuation is still observed after cesarean delivery, although not to the same extent. In addition, the fact that the results are consistent in groups of both high and low risk of cesarean delivery implies that the cesarean delivery in itself, not just its indication, has a negative influence on continuation rates.
Both the risk of cesarean delivery and the risk of infertility increase with maternal age. Part of the stronger effect of cesarean delivery on fertility among the oldest women could thus be an involuntary age-related reduced fertility, especially since the interbirth interval is slightly longer after a cesarean delivery than after a vaginal delivery. However, our results showed reduced fertility after cesarean delivery in all age groups, also among the youngest mothers, implying that confounding by maternal age does not alone explain our results.
We observed that cesarean delivery was more strongly associated with reduced fertility if the child survived than if it was stillborn or died. This suggests the reduced fertility was probably to a large degree voluntary and not related to the indication, nor to any physical consequence, of the cesarean delivery. Cesarean delivery rates are still increasing in many developed countries. In a country such as the United States, where the cesarean delivery rate in 2004 was almost twice as high as in Norway (29.1%23 compared with 15.4%4), cesarean delivery could have a substantial overall effect on reducing fertility.
1. Hamilton BE, Ventura SJ. Fertility and abortion rates in the United States, 1960-2002. Int J Androl 2006;29:34–45.
2. Grant J, Hoorens S, Sivadasan S, Loo MVH, Davanzo J, Hale L, et al. Trends in European fertility: should Europe try to increase its fertility rate or just manage the consequences? Int J Androl 2006;29:17–24.
3. Skakkebaek EN, Jørgensen N, Main KM, Rajpert-De Meyts E, Leffers H, Andersson AM, et al. Is human fecundity declining? Int J Androl 2006;29:2–11.
5. Smith SGC, Wood AM, Pell JP, Dobbie R. First cesarean birth and subsequent fertility. Fertil Steril 2006;85:90–5.
6. Tower CL, Strachan BK, Baker PN. Long-term implications of caesarean section. J Obstet Gynaecol 2000;20:365–7.
7. Hemminki E, Graubard BI, Hoffman HJ, Mosher WD, Fetterly K. Cesarean section and subsequent fertility: results from the 1982 National Survey of Family Growth. Fertil Steril 1985;43:520–8.
8. LaSala AP, Berkeley AS. Primary cesarean section and subsequent fertility. Am J Obstet Gynecol 1987;157:379–83.
9. Jolly J, Walker J, Bhabra K. Subsequent obstetric performance related to primary mode of delivery. BJOG 1999;106:227–32.
10. Hall MH, Campbell DM, Fraser C, Lemon J. Mode of delivery and future fertility. BJOG 1989;96:1297–303.
11. Mollison J, Porter M, Campbell D, Bhattacharya S. Primary mode of delivery and subsequent pregnancy. BJOG 2005;112:1061–5.
12. Zdeb MS, Therriault GD, Logrillo VM. Frequency, spacing, and outcome of pregnancies subsequent to primary cesarean childbirth. Am J Obstet Gynecol 1984;150:205–12.
13. Bahl R, Strachan B, Murphy DJ. Outcome of subsequent pregnancy three years after previous operative delivery in the second stage of labour: cohort study. BMJ 2004;328:311.
14. Hemminki E. Impact of caesarean section on future pregnancy: a review of cohort studies. Paediatr Perinat Epidemiol 1996;10:366–79.
15. Bhattacharya S, Porter M, Harrild K, Naji A, Mollison J, van Teijlingen E, et al. Absence of conception after caesarean section: voluntary or involuntary? BJOG 2006;113:268–75.
16. Hemminki E. Pregnancy and birth after cesarean section: a survey based on the Swedish birth register. Birth 1987;14:12–7.
17. Albrechtsen S, Rasmussen S, Dalaker K, Irgens LM. Reproductive career after breech presentation: subsequent pregnancy rates, interpregnancy interval, and recurrence. Obstet Gynecol 1998;92:345–50.
18. Pandian Z, Bhattacharya S, Templeton A. Review of unexplained infertility and obstetric outcome: a 10 year review. Hum Reprod 2001;16:2593–7.
19. Murphy DJ, Stirrat GM, Heron J. The relationship between Caesarean section and subfertility in a population-based sample of 14 541 pregnancies. Hum Reprod 2002;17:1914–7.
20. Irgens L. The Medical Birth Registry of Norway: epidemiological research and surveillance throughout 30 years. Acta Obstet Gynecol Scand 2000;79:435–9.
21. Bjerkedal T, Erickson JD. Association of birth outcome with subsequent fertility. Am J Obstet Gynecol 1983;147:399–404.
22. Skjaerven R, Wilcox AJ, Lie RT, Irgens LM. Selective fertility and the distortion of perinatal mortality. Am J Epidemiol 1988;128:1352–63.
23. Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F, Kirmeyer S. Births: final data for 2004. Natl Vital Stat Rep 2006;1–101.
Figure. No caption available.