Objective
The purpose of this study was to determine the impact of increasing numbers of cesarean deliveries on maternal morbidity. This study was performed for the 2010 National Institutes of Health Consensus Development Conference on Vaginal Birth After Cesarean: New Insights.
Study Design
We conducted a systematic review and metaanalysis of observational studies.
Results
Twenty-one studies (2,282,922 deliveries) were included. The rate of hysterectomy, blood transfusions, adhesions, and surgical injury all increased with increasing number of cesarean deliveries. The incidence of placenta previa increased from 10/1000 deliveries with 1 previous cesarean delivery to 28/1000 with ≥3 cesarean deliveries. Compared with women with previa and no previous cesarean delivery, women with previa and ≥3 cesarean deliveries had a statistically significant increased risk of accreta (3.3-4% vs 50-67%), hysterectomy (0.7-4% vs 50-67%), and composite maternal morbidity (15% vs 83%; odds ratio, 33.6; 95% confidence interval, 14.6–77.4).
Conclusion
Serious maternal morbidity progressively increased as the number of previous cesarean deliveries increased.
Almost one-third of all births in the United States are by cesarean delivery (CD) and the rate of both primary and repeat cesarean deliveries in the United States continues to rise each year. In the last decade, there has been a trend away from vaginal birth after CD (VBAC), with one-third of US hospitals banning VBAC and >90% of women delivering by repeat cesarean.
Counseling patients with previous CD regarding delivery options typically focuses on the risks of a trial of labor on a scarred uterus and the potential for uterine rupture in this pregnancy rather than the risk of repeat CD in future pregnancies. Many women choose repeat CD or deliver in a facility that does not allow VBAC. After 2 CDs, most women will not be offered VBAC and are destined to deliver by CD for all future pregnancies.
CD is not without risk. Maternal morbidity may include adhesion formation, surgical injury, postoperative infection, hemorrhage/transfusion, hysterectomy, abnormal placentation, and death. It is unclear whether the incidence of adverse outcomes changes with increasing numbers of CDs.
To provide meaningful counseling, it is important for patients and providers to understand not only the risks of trial of labor but also the risks that are associated with multiple CDs. Similarly, from a policy level, it is important to understand the risks for women with multiple CDs to ensure that hospitals and staff have the resources and skilled personnel required to respond. This systematic review was conducted to inform the 2010 National Institutes of Health Consensus Development Conference: Vaginal Birth After Cesarean: New Insights. The objectives were to (1) determine the incidence of adverse maternal outcome with multiple CDs and (2) determine whether the incidence changes with the number of previous CDs.
Materials and Methods
Search strategy
A systematic literature search was conducted in MEDLINE, Database of Abstracts of Reviews of Effectiveness, and the Cochrane Collaboration resources to identify relevant articles from 1980 to September 2009. Search terms used included variations of repeat CD , previous CD , and multiple CD . Additional relevant studies were identified from reference lists of reviews and editorials and by hand-searching key journals and websites, as has been shown to improve study identification in addition to electronic searches.
Study selection
Two investigators reviewed a random set of titles and abstracts to select articles for full text review. When an appropriate level of reliability was reached for selection of studies (kappa of ≥0.60), the remaining titles and abstracts were divided up and reviewed by 1 investigator. Similarly, 2 investigators screened a random set of articles for inclusion. When an appropriate level of reliability was reached, the remaining articles were divided among the investigators for further screening. Settings that were applicable to a United States population were included. Therefore, non-US studies were included if they originated from a developed country. We excluded studies of or with women without a previous CD, nulliparous patients, ≤10 subjects, breech delivery, exclusive focus on preterm delivery, low birthweight, and pregnancies that included twins or abortions. We also excluded studies that were begun or published before the 1980 National Institutes of Health Consensus Conference on VBAC and studies that were limited to patients with particular conditions, such as gestational diabetes mellitus, human immunodeficiency virus, preeclampsia.
Data abstraction
Data were extracted from each study and entered directly into evidence tables by a primary reviewer. A second reviewer verified the accuracy and completeness of the data, which was summarized descriptively.
Study quality assessment
A “best evidence” approach was applied, in which studies with the highest quality and most rigorous design according to predetermined criteria were emphasized. Reviewers rated the quality of each study using criteria that were specific to particular study designs as developed by the US Preventive Services Task Force and the National Health Service Centre for Reviews and Dissemination. Quality rating categories included maintenance of comparable groups, outcome measures that were reliable and valid, outcome assessor blinding, and adjustment for potential confounders. Details of the methods and results for quality assessments are provided in the evidence review. Two reviewers independently quality-rated all studies, with final rating achieved through consensus. Studies rated poor quality were excluded from the analyses.
Data synthesis
Metaanalyses were conducted, when appropriate, to generate a meaningful combined estimate to summarize rates with STATA software (version 10.1; StataCorp, College Station, TX). A random effects model was used to combine the studies while incorporating variations among studies. Statistical heterogeneity was assessed with the standard Q-test and the I 2 statistic (the proportion of variation in study estimates because of heterogeneity rather than sampling error). Forest plots were presented to graphically summarize the study results and the pooled results.
Results
We identified 3134 citations and reviewed 963 articles for inclusion, of which 203 articles met inclusion and were quality rated and 21 articles provided information on the association of maternal morbidity with multiple CDs ( Figure 1 ). For multiple CDs, 69 citations were identified, and 39 full-text articles were quality rated. Eleven studies met inclusion criteria and were rated good or fair quality. Individual topics were informed in the following manner ( Table 1 ): hysterectomy (7 studies), hemorrhage (3 studies), adhesions (3 studies), surgical injury (2 studies), perioperative infection (4 studies), and wound complications (2 studies). To determine the incidence and outcomes of abnormal placentation (including placental abruption, previa, and accreta) after previous CD, 82 full-text articles were reviewed. Nineteen articles met inclusion criteria and consisted of 8 good- or fair-quality cohort studies, 7 fair-quality case control studies, and 4 good- or fair-quality case series. Each study provided evidence for ≥1 of the sections on abruption (5 studies), previa (16 studies), and accreta (5 studies).
| Study | Design/years/location | Aim of study | Population | Exclusion criteria | Criteria for diagnosis | Relevant outcomes evaluated |
|---|---|---|---|---|---|---|
| Bodelon et al, 2009 | Case-control/1987-2006/United States | To identify factors associated with peripartum hysterectomy | Cases: women undergoing peripartum hysterectomy within 30 days after delivery/control subjects: women without peripartum hysterectomy | Not reported | ICD-9 codes | Maternal risk factors for peripartum hysterectomy that included placenta abnormalities and delivery method |
| Gilliam et al, 2002 | Case-control/1986-1989/United States | To estimate the relationship between previous CD and previa | Cases: multiparous women with previa/control subjects: multiparous women with spontaneous vaginal deliveries | Multiple gestation, no previa with CD | Documentation on perinatal abstract form and entered in perinatal registry. | Incidence of previous CD and previa |
| Grobman et al, 2007 | Cohort/1999-2002/ United States | To estimate the association between the number of previous CD and pregnancy outcomes in women with previa | Cases: women with previa and singleton gestation | Antepartum stillbirth, unknown number of previous CDs | Documentation in intrapartum record of “placenta previa” | Maternal morbidity that included placenta accreta, hysterectomy, and composite maternal morbidity |
| Hemminki et al, 2005 | Cohort/1987-1998/Finland | To investigate the effects of mode of delivery on problems with subsequent births | Women with >1 delivery during the study period, comparison by mode of delivery of first birth | Women with missing information (904 births) | ICD-9 codes | Prelabor hemorrhage, placental problems at birth, mode of delivery |
| Hershkowitz et al, 1995 | Cohort/1985-1992/Israel | To determine whether multiple previous CDs is associated with higher frequency of previa vs 1 previous CD | Cases: multiparous women with previa/control subjects: multiparous women with no previa | Nulliparous, <2 consecutive deliveries at same institution | Placenta attachment totally or mostly low uterine segment diagnosed by ultrasound evaluation or in labor | Characteristics of women with previa compared with women without previa |
| Juntunen et al, 2004 | Case-control/1982-2002/Finland | To evaluate outcomes in CD that is repeated several times | Cases: women with ≥4 CDs/control subjects: women with 1-3 CDs | Not reported | Number of previous hysterectomies | Maternal morbidity that included adhesions, blood loss >1000 g, placenta previa, abruptio placentae, and major perioperative complication |
| Knight et al, 2008 | Case-control/2005-2006/United Kingdom | To investigate the incidence of peripartum hysterectomy | Cases: women with peripartum hysterectomy/control subects: no hysterectomy | Not reported | Women undergoing a hysterectomy in the same clinical episode as delivery of a fetus or infant | Risk and indication of hysterectomy that was associated with previous CD |
| Laughon et al, 2005 | Case-control/2000-2003/United States | To determine whether increased risk of previa at delivery with previous CD results from an increased risk of abnormal implantation or lower likelihood of resolution | Cases: singleton pregnancies with previa on second trimester ultrasound evaluation/control subjects: singleton pregnancies with no previa | Low-lying placentas | Review of computerized database records | Incidence of placenta previa with previous CD |
| Lydon-Rochelle et al, 2001 | Cohort/1987-1996/United States | To assess the association between first birth CD and second birth placental abruption and previa | Cases: women with second singleton with previous CD/control subjects: women with second singleton and previous vaginal | First birth abruption or previa | Classification on birth certificate or ICD-9 code for “placental abruption” or “placenta previa” | Second-birth maternal complications that included abruption, previa, postpartum hemorrhage, hysterectomy, and infection |
| Lynch et al, 2003 | Case series/1990-1999/Ireland | To investigate the incidence of maternal morbidity after CD in women with at least 2 previous CDs | Cases: women with CD after ≥2 previous CDs | Not reported | Women with elective repeat CD solely because of ≥2 previous CDs | Maternal morbidity that included previa, adhesions, hysterectomy, surgical injury, venous thromboembolism, and wound problems |
| Macones et al, 2005 | Cohort/1996-2000/United States | To compare clinical outcomes in women with previous CD | Cases: women with 2 previous CDs | Classic scar | ICD-9 code “previous cesarean delivery, delivered” | Maternal morbidity that included major operative injuries, transfusion, and postpartum fever |
| Miller et al, 1997 | Case series/1985-1994/United States | To define clinical risk factors that are associated with placenta previa-accreta | Cases: placenta accreta confirmed by histologic examination | Not reported | Placenta accreta confirmed by histology on hysterectomy specimens | Clinical risk factors associated with placenta accreta |
| Nisenblat et al, 2006 | Cohort/2000-2005/Israel | To assess maternal complications after multiple CDs | Cases: women with ≥3 CDs/control subjects: second CD | Trial of labor after CD | Women scheduled for planned repeat CD | Maternal morbidity that included excessive blood loss, dense adhesions, hysterectomy, surgical injury, and placental abnormalities |
| Odibo et al, 2007 | Retrospective cohort/ 1996-2000/United States | To evaluate risk factors that are associated with placenta previa and abruption in women with previous CDs | Cases: women with previous CD and previa or abruption/control subjects: women with previous CD without previa or abruption | Not reported | ICD-9 codes for ”previous cesarean delivery, delivered,” sonographic evidence of placenta covering os in third trimester, separation of the placenta before delivery as reported by physicians | Placenta previa and abruption |
| Olive et al, 2005 | Case series/1998-2002/Australia | To determine risk factors and maternal morbidity for women with placenta previa | Cases: women with placenta previa/control subjects: women without placenta previa | Vaginal birth or CD for failure to progress | Women with placenta previa delivered by CD at >26 wk gestation | Maternal morbidity that included severe postpartum hemorrhage, hysterectomy, intensive care unit admission, or composite maternal morbidity |
| Phelan et al, 1987 | Cohort/1982-1984/United States | To evaluate the risks that are associated with trial of labor | Cases: women with 1-2 previous CDs who attempted a trial of labor/control subjects: women with repeat CDs | Classic scar, multiple gestation, malpresentation | Patients who accepted trial of labor | Maternal morbidity that included hysterectomy and abnormal placentation |
| Rouse et al, 2006 | Cohort/1999-2000/United States | To evaluate risks for blood transfusion in women with CD | Cases: women with CD and blood transfusion | Not reported | Transfusion of packed red blood cells intraoperatively or postoperatively before hospital discharge | Characteristics of women with CD who received a blood transfusion |
| Silver et al, 2006 | Cohort/1999-2002/United States | To estimate the magnitude of increased maternal morbidity that is associated with increasing numbers of CDs | Cases: women who had a CD without labor | Women in labor | Women undergoing CD without labor | Maternal morbidity that included placenta accreta, previa, hysterectomy, surgical injury, and composite morbidity |
| Taylor et al, 1994 | Case-control/1984-1987/United States | To investigate the occurrence of placenta previa after CD | Cases: women with previa/control subjects: women without previa | Nonwhite, nulliparous, missing data | Vital records check-box of previous CD and placenta previa | Incidence of placenta previa with previous CD |
| Wu et al, 2005 | Case-control/1982-2002/United States | To determine whether the rate of abnormal placentation is increasing in conjunction with CD rate | Cases: abnormal placentation/control subjects: normal placentation, matched by year of delivery | Myomatous uteri, malignancy | Histopathologic diagnosis; difficult manual piecemeal removal; heavy continued bleeding from implantation site after CD | Incidence of abnormal placentation by number of previous CDs |
| Zelop et al, 1993 | Case series/1983-1991/United States | To evaluate the clinical indications and incidence of emergency peripartum hysterectomy | Cases: women with emergency peripartum hysterectomy | Elective peripartum hysterectomies | Peripartum hysterectomy performed during the same hospitalization as the delivery | Clinical indications and incidence of peripartum hysterectomy |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
