Cesarean deliveries, the most common inpatient surgical procedure performed in the United States,1 accounted for 32.8% of births in 2012.2 Because fewer than 10% of women deliver vaginally after a cesarean delivery,3 virtually all cesarean deliveries will be followed by a repeat cesarean delivery. Cesarean deliveries increase the risk for placenta accreta, a dangerous obstetric complication that has undergone a 13-fold rise in incidence since the early 1900s4 and increases in frequency with higher numbers of cesarean deliveries.5
Given the risks of multiple cesareans deliveries, women may undertake future childbearing differentially based on how they deliver. Two publications indirectly support this hypothesis by demonstrating that women who had a cesarean, as compared with a vaginal, delivery were 11% less likely to have a subsequent birth6 and that this was driven by maternal choice.7,8 However, what remains unclear is whether a pattern exists, similar to that found between cesarean deliveries and accreta, between increasing numbers of cesarean deliveries and reduced childbearing and whether this relationship is consistent for all women. For example, women from lower socioeconomic backgrounds, who are more likely to have limited health literacy9,10 and fewer pre- and postnatal encounters,11,12 may have less knowledge of their risks, which could mitigate the observed association between mode of delivery and continued childbearing.
Our aims were to use the National Survey of Family Growth to determine whether increasing numbers of cesarean, as compared with vaginal, deliveries are associated with reduced childbearing and whether a socioeconomic difference exists in that relationship.
MATERIALS AND METHODS
We used cross-sectional data collected during the 2006–2010 National Survey of Family Growth to perform an analysis of women and their childbearing trajectory. The National Survey of Family Growth is a periodic, in-person survey conducted by the U.S. Centers for Disease Control and Prevention's National Center for Health Statistics. The purpose is to provide nationally representative data on topics related to reproductive health. The 2006–2010 survey was the seventh time it has been conducted since 1973. Participants are selected from a national probability sample of men and women, aged 15–44 years, and the overall response rate for the 2006–2010 cycle was 77%. Interviews are conducted in both English and Spanish. Further details about the sampling methodology can be found in previous reports.13 Because the National Survey of Family Growth is a publicly available, deidentified data set, the Northwestern University institutional review board does not consider its use to be human subjects research. Therefore, board approval was not necessary for this study.
At the time of the National Survey of Family Growth interview, information is collected and inputted into two separate data files: 1) the respondent file, where the individual respondent is the unit of analysis; and 2) the pregnancy file, where each pregnancy a respondent discloses is the unit of analysis. By merging these two files using the common identifier of the respondent identification number, we were able to obtain data on multiple pregnancies for a single woman and determine her number of births and mode of delivery for each birth. Women were included in the study if their first birth was a singleton live-born neonate. Those with death of their first child were excluded. We assessed their childbearing trajectory through four births because 95% of the childbearing population in the National Survey of Family Growth had fewer than five deliveries.
Our analyses were adjusted for several potential confounders, including age at prior delivery, race or ethnicity, household income, educational attainment, and intention of prior birth.14 We used age at prior delivery as a proxy for potentially different follow-up time, controlling for any age differences by mode of delivery. Consistent with prior National Survey of Family Growth publications,15 we categorized race and ethnicity as non-Hispanic white, non-Hispanic black, Hispanic, and non-Hispanic other or multiple races. We also classified pregnancy intention as intended, unwanted, and mistimed per prior National Survey of Family Growth Publications.15 Although unwanted and mistimed together comprise the larger category of unintended, we wanted to separate them to present how each varies with increasing parity. Education was dichotomized into greater or less than a high school diploma or successful completion of the general educational development examination. Family income, adjusted for household size, was classified as greater than or less than 100% of the federal poverty level for the year before the respondent's National Survey of Family Growth interview.
We determined the proportion of women who had a second, third, or fourth birth and whether they had subsequent births as a function of the mode of their previous delivery. We estimated differences in the likelihood of second births as a function of a first vaginal or cesarean delivery. In determining the likelihood of third births, we compared women with two prior vaginal deliveries with those with either one or two prior cesarean deliveries. Because our primary research question pertained to the number of cesarean deliveries rather than sequence, we grouped women with vaginal births after cesarean delivery into the same group as women who had a cesarean delivery after a vaginal delivery. Finally, to determine the likelihood of fourth births, we again compared women according to the number of previous cesarean deliveries.
Descriptive statistics are reported as nationally weighted proportions based on survey weights obtained from the National Survey of Family Growth. Statistical testing was performed with the complex survey module of Stata 12. For likelihood of second, third, and fourth births, bivariate associations were obtained with the Pearson χ2 test of proportions. We then used a modified Poisson regression to obtain the likelihood of a second, third, and fourth birth as a function of previous mode(s) of delivery, controlling for relevant characteristics. This technique provides the adjusted incidence rate ratio (the closest approximation of relative risk) with 95% confidence intervals for the binary outcome of subsequent birth.16 To assess the simultaneous influence of sociodemographic characteristics on both route of delivery and the likelihood of subsequent births, we tested the significance of interaction terms between race or ethnicity, income, and educational attainment and number of prior cesarean deliveries in the models.
There were 10 respondents from the original study sample whose births were mistakenly recorded as live births (eight women with first births, two with second births, and one with a third birth). We excluded them from the analyses because the outcomes linked to these pregnancies were miscarriage or ectopic, abortion, or stillbirth. We also removed respondents from our intention analysis that responded “didn't know” or “didn't care,” approximately 1% of the sample per birth. No imputation of missing data was performed.
Of the 12,279 women surveyed by the National Survey of Family Growth, 7,538 (61%) reported a pregnancy and 6,683 (54%) reported at least one live birth (Fig. 1) between 1974 and 2010. After exclusions, our final study sample consisted of 6,526 (98%) of the childbearing population. This corresponds to a weighted population of just more than 33.6 million women at the 2008 midpoint of the survey time period, which is consistent with previous reports from the National Survey of Family Growth.14
Table 1 presents the descriptive characteristics of our study sample by parity. Cesarean deliveries comprised the minority of all births, and, as parity increased, so did the proportion of mothers who are nonwhite, nonadolescent, lower income, and less educated. By the fourth birth, the proportion of women who characterized the pregnancy as unintended predominantly reported it as unwanted.
Figure 2 describes the birth trajectory of the study population from their first birth through their fourth birth. Percentages represent the weighted proportion of women who had a subsequent birth, stratified by prior mode of delivery, using their previous mode of delivery as the denominator. For ease of interpretation and because 83% of women who had four children either had three prior cesarean deliveries or three prior vaginal deliveries, we only present fourth births for women who sequentially delivered by the same method.
When subsequent childbearing was calculated as a proportion of women who had a prior delivery, regardless of delivery mode, 70% had a second delivery, 45% had a third delivery, and 36% had a fourth delivery. On the other hand, when all women with a first birth were considered, only 30% had a third birth, 11% had a fourth birth, and 69% had only vaginal deliveries, whereas 17%, 9%, 4%, and 1% had one, two, three, and four cesarean deliveries, respectively (data not depicted). Of the women who had a primary cesarean delivery, 14% had a successful vaginal birth after cesarean delivery.
When subsequent childbearing was calculated, taking mode of delivery into account, the association between mode of delivery and subsequent childbearing materialized. Women who had cesarean deliveries were less likely to have a subsequent delivery regardless of birth order. This pattern intensified when women with the same consecutive mode of delivery were compared. For example, 66% of women who had two successive cesareans deliveries, as compared with 52% of women with two successive vaginal deliveries, did not have a third child and 84% of women after three cesarean deliveries as compared with 60% after three vaginal deliveries discontinued childbearing.
Table 2 presents the likelihood of a second birth based on respondents' characteristics. The bivariate results suggest that women who had a vaginal first delivery, lower income and educational attainment, and younger age at first delivery were more likely to have a second birth. However, in the multivariable regression, income and educational attainment were no longer significant. Black race, first delivery by cesarean, and having an unintended first birth were significantly associated with a lower likelihood of a second birth, whereas younger age at first birth was significantly associated with a higher likelihood of a second birth.
Table 3 presents the analysis of the women with a second birth who had a third. Bivariate results revealed a dose–response relationship between number of prior cesarean deliveries and continued childbearing. Specifically, the proportion of women who had a third birth significantly decreased as the number of cesarean deliveries increased. However, in multivariable regression, the likelihood of a third birth was no longer significantly different between women who had two vaginal deliveries and those who had one cesarean delivery, including women who underwent a vaginal birth after cesarean delivery. Hispanic ethnicity, lower income, and younger age at second delivery were also significantly associated with having a third birth and an unwanted second birth was significantly associated with a lower likelihood of a third birth.
By the third birth, a history of at least two cesarean deliveries was significantly associated with a lower likelihood of fourth birth, even after adjustment for confounding variables (Table 4). Women with three births who had undergone two or three cesarean deliveries were 37% and 59% less likely, respectively, to have a fourth birth than women who had all three births vaginally. Women who were nonwhite, low income, and younger at their third birth were also significantly more likely to have a fourth birth, whereas an unwanted third birth was significantly associated with a lower likelihood of a fourth birth regardless of mode of prior deliveries.
Finally, when interaction terms were added to the models, there was no effect modification by sociodemographic characteristics on the relationship between number of cesarean deliveries and the likelihood of subsequent delivery after first and second births. However, for women with three births, household income significantly interacted with having had two or three prior cesarean deliveries and the likelihood of a fourth birth. We found that lower income women were significantly more likely to have a fourth birth after undergoing two or three cesarean deliveries than women with higher incomes (adjusted incidence rate ratio 2.50, 95% confidence interval [CI], 1.23–5.05 and adjusted incidence rate ratio 2.39, 95% CI 1.01–5.65, respectively).
This analysis not only confirms prior reports of decreased childbearing after cesarean delivery6,7,17 and other covariates associated with parity,14 but it also reveals a dose–response relationship between number of prior cesarean deliveries and the likelihood of subsequent childbirth that intensifies with increasing parity. For example, for women with two births, those who had two cesarean deliveries were 27% less likely to have a third birth than women with two vaginal deliveries. Among women who had three births, the likelihood of a fourth birth was 37% and 59% lower for women who had two and three cesarean deliveries, respectively, as compared with women with three vaginal deliveries (P<.05). For both groups of women, there was no difference in likelihood of subsequent birth between women who had one cesarean delivery and all vaginal deliveries. Our model with interaction terms revealed that household income also functioned as a moderator in this relationship for women with three children such that low-income women who had two or three cesarean deliveries were significantly more likely to have a fourth birth than their higher income counterparts.
Although the National Survey of Family Growth data provided an answer to our primary research question, they do not allow certainty with regard to the reason for this result. Despite earlier studies' suggesting an “infertility” phenomenon resulting from the procedure itself,17,18 more recent literature suggests a volitional etiology.7,8 Tollanes et al7 assessed postcesarean delivery childbearing in women who gave birth in Norway between 1967 and 2003. Their cohort, which was divided into two groups based on whether there was a stillbirth or neonatal death within 1 year of delivery, demonstrated a much larger reduction in subsequent childbearing among the women whose neonate survived and had undergone a cesarean delivery as compared with a vaginal delivery. Based on these results, the authors concluded that the observed reduction in childbearing after cesarean delivery was by maternal choice rather than an outcome of the surgical procedure. Likewise, a 2002–2003 survey of Scottish women who failed to have a second birth within 5 years of their first delivery revealed no significant difference in self-reported infertility between the women who had a primary cesarean, as compared with a primary vaginal, delivery.8 Our data support the volitional nature of the decrease in childbearing related to prior cesarean delivery, because it seems unlikely that women of higher socioeconomic status become infertile more frequently, specifically in relation to cesarean delivery, than their lower socioeconomic status counterparts.
Socioeconomic status long has been considered an important social determinant of health. In the field of reproductive health, women from lower socioeconomic backgrounds have higher rates of unintended pregnancy and births, abortions, teen pregnancies, and contraceptive failures.19,20 Each of these disparities engender substantial mental, physical, and economic challenges for women, their families, and society.21–24 Our data now introduce another difference—the higher probability of multiple cesarean deliveries—to that list. The exact reason that women of lower socioeconomic status have a dissimilar childbearing pattern after cesarean deliveries than higher income women cannot be known from these data. For example, they may have less access to contraception or they may desire a greater number of children despite mode of delivery and potential subsequent morbidity. It is also possible that they are not as cognizant, as a result of lower health literacy or less counseling, of the attendant higher risk of obstetric complications, such as accreta and hysterectomy. Further research can help to elucidate the underlying dynamics for this difference.
There are several limitations to and potential biases in this study. We do not have sufficient information on the indication for cesarean delivery or significant comorbidities such as preeclampsia. Also, because these are self-reported and retrospective data, we can neither eliminate the possibility of recall bias nor establish a causal relationship between cesarean delivery and reduced childbearing. It is unlikely, however, that many women would incorrectly recall whether they had a cesarean delivery. Because our results are derived from nationally representative data, external validity with regard to the U.S. childbearing population should be present.
In conclusion, our data indicate that women who have cesarean deliveries are less likely to continue childbearing, especially as their number of cesarean deliveries increases, and that income moderates this relationship. These findings emphasize that desired family size and the risks associated with multiple cesarean deliveries should be part of the comprehensive conversation that takes place during informed consent for nonemergent cesarean deliveries, specifically elective cases. Moreover, they underscore the need to investigate why the association between cesarean delivery and fewer subsequent deliveries exists and what factors are driving the income-based difference in postcesarean childbearing.
1. National Center for Health Statistics. National Hospital Discharge Survey: 2010 table, procedures by selected patient characteristics—number by procedure category and age. 2010; Available at: http://www.cdc.gov/nchs/fastats/insurg.htm
. Retrieved March 17, 2014.
2. Hamilton BE, Martin JA, Ventura SJ. Births: preliminary data for 2012. Natl Vital Stat Rep 2013;62:1–20.
3. Spong CY, Berghella V, Wenstrom KD, Mercer BM, Saade GR. Preventing the first cesarean delivery: summary of a joint Eunice Kennedy Shriver National Institute of Child Health and Human Development, Society for Maternal-Fetal Medicine, and American College of Obstetricians and Gynecologists Workshop. Obstet Gynecol 2012;120:1181–93.
4. Wu S, Kocherginsky M, Hibbard JU. Abnormal placentation: twenty-year analysis. Am J Obstet Gynecol 2005;192:1458–61.
5. Silver RM, Landon MB, Rouse DJ, Leveno KJ, Spong CY, Thom EA, et al.. Maternal morbidity associated with multiple repeat cesarean deliveries. Obstet Gynecol 2006;107:1226–32.
6. Gurol-Urganci I, Bou-Antoun S, Lim CP, Cromwell DA, Mahmood TA, Templeton A, et al.. Impact of Caesarean section on subsequent fertility: a systematic review and meta-analysis. Hum Reprod 2013;28:1943–52.
7. Tollanes MC, Melve KK, Irgens LM, Skjaerven R. Reduced fertility after cesarean delivery: a maternal choice. Obstet Gynecol 2007;110:1256–63.
8. 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.
9. Yee LM, Simon MA. The role of health literacy and numeracy in contraceptive decision-making for urban Chicago women. J Community Health 2014;39:394–9.
10. Gazmararian JA, Parker RM, Baker DW. Reading skills and family planning knowledge and practices in a low-income managed-care population. Obstet Gynecol 1999;93:239–44.
11. Feijen-de Jong EI, Jansen DE, Baarveld F, van der Schans CP, Schellevis FG, Reijneveld SA. Determinants of late and/or inadequate use of prenatal healthcare in high-income countries: a systematic review. Eur J Public Health 2012;22:904–13.
12. Centers for Disease Control and Prevention (CDC). Postpartum care visits—11 states and New York City, 2004. MMWR Morb Mortal Wkly Rep 2007;56:1312–6.
13. Lepkowski JM, Mosher WD, Davis KE, Groves RM, Van Hoewyk J. The 2006–2010 National Survey of Family Growth: sample design and analysis of a continuous survey. Vital Health Stat 2 2010;1–36.
14. Martinez G, Daniels K, Chandra A. Fertility of men and women aged 15-44 years in the United States: national survey of family growth, 2006–2010. Natl Health Stat Report 2012;1–28.
15. Mosher WD, Jones J, Abma JC. Intended and unintended births in the United States: 1982–2010. National Health Statistics Reports. Hyattsville (MD): National Center for Health Statistics; 2012.
16. Barros AJ, Hirakata VN. Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models that directly estimate the prevalence ratio. BMC Med Res Methodol 2003;3:21.
17. 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.
18. LaSala AP, Berkeley AS. Primary cesarean section and subsequent fertility. Am J Obstet Gynecol 1987;157:379–83.
19. Dehlendorf C, Rodriguez MI, Levy K, Borrero S, Steinauer J. Disparities in family planning. Am J Obstet Gynecol 2010;202:214–20.
20. Dehlendorf C, Harris LH, Weitz TA. Disparities in abortion rates: a public health approach. Am J Public Health 2013;103:1772–9.
21. Pinzon JL, Jones VF. Care of adolescent parents and their children. Pediatrics 2012;130:e1743–56.
22. Shah PS, Balkhair T, Ohlsson A, Beyene J, Scott F, Frick C. Intention to become pregnant and low birth weight and preterm birth: a systematic review. Matern Child Health J 2011;15:205–16.
23. Monea E, Thomas A. Unintended pregnancy and taxpayer spending. Perspect Sex Reprod Health 2011;43:88–93.
24. Williams SL, Parisi SM, Hess R, Schwarz EB. Associations between recent contraceptive use and quality of life among women. Contraception 2012;85:282–7.