HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by LYDON-ROCHELLE, M. Articles by MARTIN, D. P. Search for Related Content PubMed PubMed Citation Articles by LYDON-ROCHELLE, M. Articles by MARTIN, D. P.

Cesarean Delivery and Postpartum Mortality Among Primiparas in Washington State, 1987–1996

From the Department of Health Services, School of Public Health and Community Medicine, Department of Family and Child Nursing, School of Nursing, Department of Epidemiology, School of Public Health and Community Medicine, and Department of Obstetrics and Gynecology, School of Medicine, University of Washington, Seattle, Washington.

Address reprint requests to: Mona Lydon-Rochelle, PhD, MPH University of Washington Center for Women’s Health Research Mailstop 357262 Seattle, WA 98195-7262 E-mail: minot{at}u.washington.edu

	Abstract

Top Abstract Material and Methods Results Discussion References

 
Objective: To examine the association between delivery method and mortality within 6 months of delivery among primiparas.

Methods: We conducted a population-based, retrospective cohort analysis using statewide, maternally linked birth certificate, hospital discharge, and death certificate data. The present cohort was all primiparas who gave birth to live-born infants in civilian hospitals in Washington State from January 1, 1987 through December 31, 1996 (n = 265,471). Odd ratios (OR) and 95% confidence intervals (CI) were calculated for overall mortality, pregnancy-related mortality, and pregnancy-unrelated mortality associated with delivery method.

Results: Thirty-two women (12.1 per 100,000 singleton live births) died within 6 months of delivery of their first child. Eleven of 32 deaths were pregnancy related (4.1 per 100,000 singleton live births, 95% CI 1.6, 6.5), and 21 of the 32 deaths were not pregnancy related (7.9 per 100,000 singleton live births, 95% CI 4.5, 11.3). The pregnancy-related mortality rate was higher among women delivered by cesarean (10.3/100,000) than among women delivered vaginally (2.4/100,000). In logistic regression analyses, women who had cesarean delivery were not at significantly higher risk of death overall after adjustment for maternal age (OR 1.7, 95% CI 0.3, 3.6), pregnancy-related death after adjustment for maternal age and severe preeclampsia (OR 2.2, 95% CI 0.6, 7.9), or pregnancy-unrelated death after adjustment for maternal age and marital status (OR 0.9, 95% CI 0.3, 2.7), relative to women who had vaginal delivery.

Conclusion: Cesarean delivery might be a marker for serious preexisting morbidities associated with increased mortality risk rather than a risk factor for death in and of itself. Data from additional sources such as medical records and autopsy reports are necessary to disentangle preexisting mortality risk from risk associated solely with delivery method.

In 1998, approximately 825,870 women had cesarean deliveries in the United States.¹ Although reduction of the maternal mortality has been identified as a high priority for recent national research and policy agendas, the relationship between cesarean delivery and risk of maternal death is not known.^2,3 Three previous studies in the United States reported that women delivered by cesarean had significantly higher mortality rates than women who had vaginal delivery, whereas two other studies found lower mortality rates among women who had cesarean delivery.^4–8 However, those studies examined mortality rates among women delivering 15–46 years ago, and their conclusions have limited value today because of changes in obstetric practices that are likely to decrease risk, including maternal transport to tertiary care centers and increased use of regional anesthesia.^9,10 We used statewide, maternally linked birth certificate, hospital discharge, and death certificate data to examine the association between method of delivery and mortality within 6 months of delivery among primiparas with singleton live births.

	Material and Methods

Top Abstract Material and Methods Results Discussion References

 
We conducted a population-based, retrospective cohort analysis using data obtained from the Washington State Birth Events Record Database and Washington State death certificates. This file successfully links over 95% of Washington State birth certificates with maternal and infant Comprehensive Hospital Discharge Reporting System records for the delivery hospitalization. In an earlier study of delivery method and maternal postpartum rehospitalization, we included women without selected chronic medical conditions (n = 256,795); details of the sampling design are documented elsewhere.¹¹ In the absence of decisive data from previous observational studies regarding the association between cesarean delivery and maternal mortality we decided to control for the confounding influence of parity, fetal death, and multiple gestations by restricting selection of subjects for our study. Thus, the present study cohort was based on all primiparas who gave birth to live-born infants in civilian hospitals in Washington State from January 1, 1987 through December 31, 1996 (n = 283,908). From this group we excluded women with unknown delivery method (n = 13,401), multiple gestation (n = 4723), infants weighing less than 500 g (n = 209), or infants with unknown birth weight (n = 631). Some women met more than one exclusion criterion, and after all exclusions were made, 265,471 subjects remained for analysis.

Maternal mortality was defined as one that occurred for any reason after delivery and within 6 months of delivery. All death certificate records registered with the State of Washington from January 1, 1987 through June 30, 1997 were searched and then linked to the file using a sequential deterministic method (details available on request). We chose to study deaths within 6 months of delivery because we speculated that maternal mortality has a potential relationship to delivery method extending through that time. The study was approved by the Human Subjects Review Committee at the University of Washington, Seattle, and the Human Research Review Board at the Washington State Department of Health, Olympia.

Deliveries were classified as cesarean if "cesarean" was checked on the birth certificate or any of the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM)¹² cesarean delivery procedural or diagnostic codes were used on hospital discharge data. Given these criteria, there were 56,776 (21.4%) women delivered by cesarean and 208,695 (78.6%) women delivered vaginally available for analysis. The definition of pregnancy-related mortality for the study was that developed by ACOG and the Centers for Disease Control and Prevention (CDC): death resulting from complications of pregnancy itself, a chain of events initiated by pregnancy, or aggravation of an unrelated event by the physiologic effects of pregnancy.^13,14 For each identified maternal death, maternal age, ICD-9-CM discharge and procedure codes for the delivery hospitalization, delivery-to-death interval, death location, underlying cause of death, and multiple causes of death were reviewed independently by one obstetrician-gynecologist and two maternal-fetal medicine specialists who were masked to the subjects’ method of delivery. Deaths were then classified as pregnancy related or pregnancy unrelated, with 100% agreement among reviewers.

To examine the risk of mortality overall and pregnancy-related mortality associated with cesarean delivery, we used logistic regression to calculate odds ratios (ORs) and test-based 95% confidence intervals (CIs), using women who had vaginal delivery as the reference group. The following variables, reported at the time of delivery, were examined as possible confounders between delivery method and mortality: maternal age, maternal race or ethnicity, marital status, maternal smoking during pregnancy, initiation of prenatal care, payer, hospital level, infant’s estimated gestational age, infant’s birth weight, maternal medical conditions, pregnancy complications, and medical induction of labor. All potential confounding variables were considered for adjustment if their inclusion into the logistic model changed the ORs for mortality associated with method of delivery by 10% or more. We formulated models separately for the overall mortality, pregnancy-related, and pregnancy-unrelated mortality analyses. Maternal age met the criterion for overall mortality, maternal age and marital status for pregnancy-related mortality, and maternal age and preeclampsia for pregnancy-unrelated mortality.

In a subanalysis, we examined the frequency of pregnancy-related deaths among only the 173,231 women (65.2%) with term gestations and spontaneous labors and without medical or obstetric complications, because pregnancy-related mortality could be attributable to antenatal morbidity rather than delivery method.

	Results

Top Abstract Material and Methods Results Discussion References

 
Thirty-two women (12.1 per 100,000 singleton live births) died within 6 months of delivering their first child. Several maternal and newborn characteristics differed by mortality status (Table 1). Compared with women who survived, those who died were more likely to be teenaged, unmarried, or white, and to smoke during pregnancy, be covered by Medicaid, and deliver at a level III hospital. Women who died were more likely than survivors to have had cardiac, renal, mental health, or severe preeclampsia complications and to have had medically indicated induction of labor during the birth hospitalization. The infants of women who died were more likely to have had low birth weight (LBW, 500–2499 g) or be preterm (estimated gestational age under 36 weeks).

Twenty-one deaths were classified as pregnancy unrelated (7.9 per 100,000 singleton live births, 95% CI 4.5, 11.3), and all of them occurred after discharge from the initial birth hospitalization. Four pregnancy-unrelated deaths occurred among women who had cesarean delivery, and all were due to injuries (two suicide, one poisoning, and one homicide). Seventeen pregnancy-unrelated deaths occurred among women who had vaginal delivery, including ten due to injuries (three suicide, three motor vehicle accident, two homicide, and two injury), three due to central nervous system complications, two due to cancer, and one each due to cardiac complications and unknown causes. The pregnancy-unrelated mortality rate was lower among women who had cesarean delivery (6.9/100,000) than among women who had vaginal delivery (8.2/100,000).

In a logistic regression model adjusted for maternal age, women who had cesarean delivery were not at significantly higher risk of overall mortality relative to women who had vaginal delivery (OR 1.7, 95% CI 0.3, 3.6) (Table 2). Similarly, women who had cesarean delivery were not at significantly higher risk of pregnancy-related death relative to women who had vaginal delivery, after adjustment for maternal age and severe preeclampsia (OR 2.2, 95% CI 0.6, 7.9). Women with cesarean delivery also were not at significantly higher risk of pregnancy-unrelated death, after adjustment for maternal age and marital status (OR 0.9, 95% CI 0.3, 2.7). Finally, in a subanalysis among women who were without apparent preexisting morbidity, only two pregnancy-related deaths remained (one among women with cesarean delivery and one among women with vaginal delivery).

	Discussion

Top Abstract Material and Methods Results Discussion References

Previous studies concerning the relationship between cesarean delivery and risk of maternal mortality have produced findings that are inconsistent. Two previous statewide studies of 13 deaths in Rhode Island and 30 deaths in Georgia, and a United States population-based study of 691 maternal deaths reported higher mortality rates among women who had cesarean delivery.^4–6 The Rhode Island study of vital records data from 1965 through 1973 were used to calculate a cesarean-delivery mortality rate of 30.9 per 100,000, which was significantly higher than the vaginal-delivery mortality rate of 2.7 per 100,000.⁴ In Georgia, a cohort analysis using statewide, maternally linked birth and death certificate data based on women who gave birth to live-born infants in 1975 and 1976 found a cesarean-delivery mortality rate of 59.3 per 100,000, which was significantly higher than the vaginal-delivery mortality rate of 9.7 per 100,000.⁶ The United States population-based study used hospital discharge data from 1970, 1974, and 1978 and reported that the 1978 in-hospital mortality rate was significantly higher among women with cesarean delivery (40.9/100,000) than among women with vaginal delivery (9.8/100,000).⁵

In contrast, a Massachusetts statewide study examined 174 reported maternal deaths from 1954 through 1985 and found that women who had cesarean delivery were not at significantly higher risk of death directly due to cesarean delivery relative to women who had vaginal delivery.⁷ Also, a hospital-based study in Massachusetts, using medical records data of 68,645 women who gave birth between 1968 and 1978, found six deaths and reported that women who had cesarean deliveries were not at increased risk of in-hospital death compared with women who delivered vaginally; in fact, no maternal deaths were reported among women delivered by cesarean.⁸

Each of those studies used different definitions of maternal mortality, and none controlled for the confounding influence of parity, maternal age, or obstetric and medical complications, factors that could account for the different maternal mortality risks associated with method of delivery. In our study cohort, obstetric and medical complication rates were higher among women who had cesarean delivery than among women who had vaginal delivery. Women who died were more likely than those who survived to have had preexisting morbidities and preterm or induced labor, suggesting that their health might have been compromised antenatally. Thus, cesarean delivery might be a marker for serious preexisting morbidities that increase mortality risk rather than a risk factor for death in and of itself. The diminution of risk after we adjusted for severe preeclampsia supports this hypothesis. It also is possible that residual confounding and spurious mortality risk elevation remain in our results, as we did not have complete information on all relevant maternal antenatal complications.

Women who had cesarean delivery had pregnancy-related deaths due to eclampsia, puerperal infection, hemorrhage, renal failure, and Moyamoya disease. We previously reported that cesarean delivery was associated with significantly higher risks of maternal rehospitalization for uterine infection, obstetric surgical wound complications, and cardiopulmonary and thromboembolic conditions among primiparas without prior identified high-risk medical conditions.¹¹ Because these complications occur more frequently among women who have cesarean delivery than among woman who have vaginal delivery, the two deaths from infection and hemorrhage among women who had cesarean delivery might have been related to delivery method. It is unlikely, however, that the deaths from severe preeclampsia or arteriovenous malformation were caused by cesarean delivery, thus the potential effect of delivery method on mortality risk among primiparous women in Washington State is small.

The use of linked birth certificate, hospital discharge, and death certificate data to investigate mortality after childbirth has advantages. Linked datasets allowed us to study an entire cohort of primiparous women who gave birth to live singletons during a 10-year period. Because our record linkage process provided the delivery-to-death interval and unique identifiers, we could identify more deaths than would be detected by death certificate review alone, because ICD-9 codes 630–676 (complications of pregnancy, childbirth, and puerperium) for the underlying cause of death were assigned on only 20% of death certificates. Conversely, data derived from vital statistics and administrative records might be limited by inaccuracies or incompleteness. In this study we have some assurance that there was minimal misclassification of method of delivery, as previous research found that 99.8% of cesarean deliveries are correctly classified using Washington State linked birth certificate-hospital discharge files.¹⁵ Despite 100% agreement among reviewers, we had limited ability to classify accurately the pregnancy relatedness of all deaths. In the absence of medical records, autopsy reports, and family interviews, death from events such as motor vehicle accidents, which theoretically could be influenced by use of pain relievers that impair function after cesarean delivery, might have been misclassified as unrelated to pregnancy. Another concern about misclassification was that it is possible that women who died of suicide were more likely than women who survived to have had a pregnancy-related complication such as postpartum depression. However, the absence of mental health-related puerperium diagnoses and presence of nonpuerperium mental health diagnoses (ie, psychosis or poisoning by psychotropic drugs) listed on hospital discharge summaries or death certificates among women who died of suicide contradict that interpretation.

Because this study was part of a larger study that examined short-, medium-, and long-term maternal health outcomes associated with delivery method, we restricted selection of study subjects to primiparas to avoid the confounding influence of parity.¹¹ Despite performing power calculations, which indicated that the number of deliveries in Washington State over a 10-year period would be adequate to detect a difference in mortality rate by delivery method, the number of deaths was much lower than expected and our risk estimates had very wide confidence intervals. In addition, there was extremely low statistical power to detect an association between cesarean delivery and pregnancy-related death when adjusting for confounding variables; however, adjustment demonstrated the effect that serious preexisting morbidity had on estimated cesarean delivery–related risk. Thus, our results should be interpreted cautiously.

The Safe Motherhood Monitoring and Prevention Research Act of 1999 authorized the United States Secretary of Health and Human Services, acting through the CDC, to establish and implement local, state, and national monitoring and surveillance programs to identify and promote the research investigation of maternal mortality. Our findings indicate that data from additional sources such as medical records and autopsy reports are necessary to disentangle preexisting mortality risk from risk associated solely with delivery method. Further studies to determine the reasons for all pregnancy-associated deaths after childbirth in Washington State would aid in determining where to direct prevention efforts.

	Footnotes

 
Dr. Lydon-Rochelle was supported in part by grants from the Agency for Health Care Policy and Research (5 T32 HS00034) and the National Institute of Health-National Institute of Nursing Research (1 P30 NR0400 & T32 NR07039).

We acknowledge the data linkage programming support of William O’Brien. We are grateful to George J. Gilson, MD, University of New Mexico, and Susan E. Gerber, MD, MPH, Northwestern University Medical School for assistance with classification of maternal deaths.

PII S0029-7844(00)01119-4

Received July 10, 2000. Received in revised form September 18, 2000. Accepted October 12, 2000.

	References

Top Abstract Material and Methods Results Discussion References

 
1. Ventura SJ, Martin JA, Curtin SC, Matthews TJ, Park MM. Births: Final data for 1998. National Vital Stat Rep 2000;48(3):1–100.

2. United States Department of Health and Human Services. Tracking Healthy People 2010. Washington, DC: United States Government Printing Office. November 2000.

3. Maternal mortality—United States, 1982–1996. JAMA 1998;280: 1042–3.[Free Full Text]

4. Evrard JR, Gold EM. Cesarean section and maternal mortality in Rhode Island. Incidence and risk factors, 1965–1975. Obstet Gynecol 1977;50:594–7.[Abstract/Free Full Text]

5. Petitti DB, Cefalo RC, Shapiro S, Whalley P. In-hospital maternal mortality in the United States: Time trends and relation to method of delivery. Obstet Gynecol 1982;59:6–12.[Abstract/Free Full Text]

6. Rubin GL, Peterson HB, Rochat RW, McCarthy BJ, Terry JS. Maternal death after cesarean section in Georgia. Am J Obstet Gynecol 1981;139:681–5.[Medline]

7. Sachs BP, Yeh J, Acker D, Driscoll S, Brown DA, Jewett JF. Cesarean section-related maternal mortality in Massachusetts, 1954–1985. Obstet Gynecol 1988;71:385–8.[Medline]

8. Frigoletto FD Jr, Ryan KJ, Phillippe M. Maternal mortality rate associated with cesarean section: An appraisal. Am J Obstet Gynecol 1980;136:969–70.[Medline]

9. Powell SL, Holt VL, Hickok DE, Easterling T, Connell FA. Recent changes in delivery site of low-birth-weight infants in Washington: Impact on birth weight-specific mortality. Am J Obstet Gynecol 1995;173:1585–92.[Medline]

10. Hawkins JL, Koonin LM, Palmer SK, Gibbs CP. Anesthesia-related deaths during obstetric delivery in the United States, 1979–1990. Anesthesiology 1997;86:277–84.[Medline]

11. Lydon-Rochelle M, Holt VL, Martin DP, Easterling T. The association between method of delivery and maternal rehospitalization. JAMA 2000;283:2411–6.[Abstract/Free Full Text]

12. International classification of diseases, 9th revision: Clinical modification. Vol. I–III. Salt Lake City, Utah: Medicode, 1998.

13. Koonin LM, MacKay AP, Berg CJ, Atrash HK, Smith JC. Pregnancy-related mortality surveillance—United States, 1987–1990. MMWR CDC Surveill Summ 1997;46:17–36.[Medline]

14. Atrash HK, Alexander S, Berg CJ. Maternal mortality in developed countries: Not just a concern of the past. Obstet Gynecol 1995;86: 700–5.[Abstract]

15. Parrish KM, Holt VL, Connell FA, Williams B, LoGerfo JP. Variations in the accuracy of obstetric procedures and diagnoses on birth records in Washington State, 1989. Am J Epidemiol 1993;138: 119–27.[Abstract/Free Full Text]

This article has been cited by other articles:

				C. Deneux-Tharaux, E. Carmona, M.-H. Bouvier-Colle, and G. Breart Postpartum Maternal Mortality and Cesarean Delivery. Obstet. Gynecol., September 1, 2006; 108(3): 541 - 548. [Abstract] [Full Text] [PDF]

				L. Carbillon Vaginal Versus Cesarean Delivery for Breech Presentation in California: A Population-Based Study Obstet. Gynecol., May 1, 2004; 103(5): 1003 - 1004. [Full Text] [PDF]

				V. M. Allen, C. M. O'Connell, R. M. Liston, and T. F. Baskett Maternal Morbidity Associated With Cesarean Delivery Without Labor Compared With Spontaneous Onset of Labor at Term Obstet. Gynecol., September 1, 2003; 102(3): 477 - 482. [Abstract] [Full Text] [PDF]

				Z. A. Brown, A. Wald, and L. Corey Cesarean Delivery and Risk of Herpes Simplex Virus Infection--Reply JAMA, May 7, 2003; 289(17): 2208 - 2209. [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by LYDON-ROCHELLE, M. Articles by MARTIN, D. P. Search for Related Content PubMed PubMed Citation Articles by LYDON-ROCHELLE, M. Articles by MARTIN, D. P.