Background
Preterm birth (PTB) is the number one cause of perinatal mortality. Prior surgery on the cervix is associated with an increased risk of PTB. History of uterine evacuation, by either induced termination of pregnancy (I-TOP) or spontaneous abortion (SAB), which involve mechanical and/or osmotic dilatation of the cervix, has been associated with an increased risk of PTB in some studies but not in others.
Objective
The objective of the study was to evaluate the risk of PTB among women with a history of uterine evacuation for I-TOP or SAB.
Data Sources
Electronic databases (MEDLINE, Scopus, ClinicalTrials.gov , EMBASE, and Sciencedirect) were searched from their inception until January 2015 with no limit for language.
Study Eligibility Criteria
We included all studies of women with prior uterine evacuation for either I-TOP or SAB, compared with a control group without a history of uterine evacuation, which reported data about the subsequent pregnancy.
Study Appraisal and Synthesis Methods
The primary outcome was the incidence of PTB < 37 weeks. Secondary outcomes were incidence of low birthweight (LBW) and small for gestational age (SGA). We planned to assess the primary and the secondary outcomes in the overall population as well as in studies on I-TOP and SAB separately. The pooled results were reported as odds ratio (OR) with 95% confidence interval (CI).
Results
We included 36 studies in this metaanalysis (1,047,683 women). Thirty-one studies reported data about prior uterine evacuation for I-TOP, whereas 5 studies reported data for SAB. In the overall population, women with a history of uterine evacuation for either I-TOP or SAB had a significantly higher risk of PTB (5.7% vs 5.0%; OR, 1.44, 95% CI, 1.09–1.90), LBW (7.3% vs 5.9%; OR, 1.41, 95% CI, 1.22–1.62), and SGA (10.2% vs 9.0%; OR, 1.19, 95% CI, 1.01–1.42) compared with controls. Of the 31 studies on I-TOP, 28 included 913,297 women with a history of surgical I-TOP, whereas 3 included 10,253 women with a prior medical I-TOP. Women with a prior surgical I-TOP had a significantly higher risk of PTB (5.4% vs 4.4%; OR, 1.52, 95% CI, 1.08–2.16), LBW (7.3% vs 5.9%; OR, 1.41, 95% CI, 1.22–1.62), and SGA (10.2% vs 9.0%; OR, 1.19, 95% CI, 1.01–1.42) compared with controls. Women with a prior medical I-TOP had a similar risk of PTB compared with those who did not have a history of I-TOP (28.2% vs 29.5%; OR, 1.50, 95% CI, 1.00–2.25). Five studies, including 124,133 women, reported data about a subsequent pregnancy in women with a prior SAB. In all of the included studies, the SAB was surgically managed. Women with a prior surgical SAB had a higher risk of PTB compared with those who did not have a history of SAB (9.4% vs 8.6%; OR, 1.19, 95% CI, 1.03–1.37).
Conclusion
Prior surgical uterine evacuation for either I-TOP or SAB is an independent risk factor for PTB. These data warrant caution in the use of surgical uterine evacuation and should encourage safer surgical techniques as well as medical methods.
Preterm birth (PTB) is the number one cause of perinatal mortality in many countries, including the United States. Defining risk factors for prediction of PTB is an important goal for several reasons. First, identifying women at risk allows initiation of risk-specific treatment. Second, it may define a population useful for studying particular interventions. Finally, it may provide important insights into the mechanisms leading to PTB.
Prior surgery on the cervix, such as cone biopsy and loop electrosurgical excision procedure, is associated with an increased risk of spontaneous PTB. A history of uterine evacuation, by either induced termination of pregnancy (I-TOP) or treatment of spontaneous abortion (SAB) by suction dilation and curettage or by dilation and evacuation (D&E), which may involve mechanical and/or osmotic dilatation of the cervix, has been associated with an increased risk of PTB in some studies but not in others.
Some studies have also postulated that the method of uterine evacuation may influence the association (or not) with PTB. Moreover, with recent increases in the use of medications (misoprostol and mifepristone), it would be important to assess outcomes in subsequent pregnancies after medical termination of pregnancy as the element of cervical trauma is minimized with these techniques.
The aim of this metaanalysis was to evaluate the risk of PTB among women with a history of uterine evacuation for either I-TOP or SAB.
Materials and Methods
Search strategy
Electronic databases (ie, MEDLINE, Scopus, ClinicalTrials.gov , EMBASE, Sciencedirect) were searched from their inception until January 2015 with no limit for language. Search terms used were the following key words: low birthweight, premature birth, preterm birth, small for gestational age, miscarriage, pregnancy, premature, newborn, uterine evacuation, abortion, induced abortion, spontaneous abortion, termination of pregnancy, curettage, first trimester, second trimester, mifepristone, misoprostol, laminaria, subsequent, and dilatation and evacuation; dilation and curettage; spontaneous preterm birth.
In addition, the reference lists of all identified articles were examined to identify studies not captured by electronic searches. The electronic search and the eligibility of the studies were independently assessed by the authors (G.S. and V.B.). Differences were resolved by discussion.
Study selection
We included all studies of women with prior uterine evacuation for either I-TOP or SAB, compared with a control group without prior uterine evacuation, which reported data about the subsequent pregnancy. We excluded studies without a control group (eg, case series) as well as studies about stillbirth.
I-TOP was defined as an intervention to voluntarily terminate a pregnancy (ie, induced abortion) by either surgical or medical means so it does not result in a live birth. SAB was defined as spontaneous intrauterine pregnancy loss prior to 20 weeks. Surgical uterine evacuation (for either I-TOP or SAB) was defined as a procedure using surgical instruments, either D&E or vacuum aspiration (VA), to remove the fetus and placenta from the uterus.
D&E was defined as a procedure that includes mechanical cervical dilatation (usually by using uterine dilators of increasing diameter to stretch the cervix) followed by the removal of uterine contents using a combination of suction and instruments (eg, sharp curette, ring clamp, or forceps). VA was defined as evacuation of the uterine contents using an electric vacuum aspirator or manual vacuum aspirator. Medical uterine evacuation (for either I-TOP or SAB) was defined as a nonsurgical uterine evacuation in which pharmaceutical drugs are used to empty the uterus. Therefore, we had the potential for several sensitivity analyses according to the type of abortion ( Table 1 ).
Intervention group | Control group |
---|---|
Overall analysis | |
Prior uterine evacuation (I-TOP and SAB) | No prior uterine evacuation |
Planned sensitivity analyses in women with prior uterine evacuation for I-TOP | |
Prior I-TOP | No prior I-TOP |
Prior surgical (either D&E or VA) I-TOP | No prior I-TOP |
Prior surgical I-TOP by D&E | No prior I-TOP |
Prior surgical I-TOP by VA | No prior I-TOP |
Prior surgical I-TOP by D&E | Prior surgical I-TOP by VA a |
Prior medical I-TOP | No prior I-TOP |
Planned sensitivity analyses in women with prior uterine evacuation for SAB | |
Prior SAB | No prior SAB |
Prior surgical (either D&E or VA) SAB | No prior SAB |
Prior surgical SAB by VA | No prior SAB |
Prior surgical SAB by D&E | No prior SAB |
Prior surgical SAB by D&E | Prior surgical SAB by VA a |
Prior medical SAB | No prior SAB |
Planned sensitivity analyses comparing I-TOP with SAB | |
Prior uterine evacuation for I-TOP | Prior uterine evacuation for SAB a |
Planned subgroup analyses in study on I-TOP and in study on SAB, separately | |
Only 1 prior uterine evacuation | No prior uterine evacuation |
More than 1 prior uterine evacuation | No prior uterine evacuation |
More than 1 prior uterine evacuation | Only 1 prior uterine evacuation a |
Prior uterine evacuation in singletons | No prior uterine evacuation in singletons |
Prior uterine evacuation in multiple gestations | No prior uterine evacuation in multiple gestations |
Prior uterine evacuation in cohort studies | No prior uterine evacuation in cohort studies |
Prior uterine evacuation in case-control studies | No prior uterine evacuation in case-control studies |
According to gestational age at uterine evacuation | — |
a Because none of the included studies evaluated this outcome, we used an indirect comparison metaanalysis to assess this outcome.
Primary and secondary outcomes were planned a priori. The primary outcome was the incidence of PTB (ie, preterm delivery < 37 weeks). Secondary outcomes were neonatal outcomes including incidence of low birthweight (birthweight < 2500 g) and of small for gestational age (birthweight < 10th percentile for gestational age).
We planned to assess the primary and the secondary outcomes in the overall population as well as in studies on I-TOP and SAB, separately. We also planned several subgroup analyses according to the number of prior uterine evacuation; the number of fetuses in the index pregnancy; the gestational age at abortion; or the type of the study (either cohort or case-control study) ( Table 1 ). We assessed these subgroup analyses for only the primary outcome (ie, incidence of PTB) in both surgical and medical I-TOP and SAB, separately and not in the overall combined data ( Table 1 ).
Data extraction and risk of bias assessment
Data abstraction was completed by 2 independent investigators (G.S. and V.B.). Each investigator independently abstracted data from each study separately. Data from each eligible study were extracted without modification of original data onto custom-made data collection forms. Differences were resolved by consensus. Information of confounders adjusted and adjusted risk estimates were collected when available. When possible, all authors were contacted for missing data.
Reviewers (G.S. and V.B.) independently assessed the risk of bias of the included studies via the Methodological Index for Non-Randomized Studies. Seven domains related to risk of bias were assessed in each study: (1) aim (ie, clearly stated aim), (2) rate (ie, inclusion of consecutive patients and response rate), (3) data (ie, prospective collection of data), (4) bias (ie, unbiased assessment of study endpoints), (5) time (ie, follow-up time appropriate), (6) loss (ie, loss to follow-up), and (7) size (ie, calculation of the study size). Review authors’ judgments were categorized as low risk, high risk, or unclear risk of bias. Discrepancies were resolved by discussion.
Data analysis
The data analysis was completed independently by two authors (G.S. and V.B.) using Review Manager 5.3 (Copenhagen, Denmark: The Nordic Cochrane Centre, Cochrane Collaboration, 2014). Discrepancies were resolved by discussion.
Heterogeneity across studies was assessed using the Higgins I 2 test. In case of statistically significant heterogeneity (I 2 > 0%), the random-effects model of DerSimonian and Laird was used; otherwise, in case of no inconsistency in risk estimates (ie, I 2 = 0%), a fixed-effect model was managed. The pooled results were reported as odds ratio (OR) with 95% confidence interval (CI).
For the outcomes not directly assessed by any of the included studies, an indirect comparison metaanalysis was performed ( Table 1 ). In the indirect comparison metaanalyses, data were combined in a 2-stage approach in which outcomes were analyzed in their original study and then summary statistics combined using standard summary data metaanalysis techniques to give an overall measure of effect (summary relative risk with 95% CI).
For studies that reported both unadjusted and adjusted risk for confounders statistically proven, we performed metaanalyses using a generic inverse variance method to obtain the adjusted risk estimate (aOR) of the primary outcome (ie, incidence of PTB). We assessed the aOR only for the primary outcome (ie, incidence of PTB) in studies on both surgical and medical I-TOP and SAB, separately.
Before data extraction, the review was registered with the PROSPERO International Prospective Register of Systematic Reviews (registration number CRD42015026482). Therefore, all the analyses and the outcomes were planned a priori before the data extraction.
The metaanalysis was reported following the Preferred Reporting Item for Systematic Reviews and Meta-Analyses statement.
Results
Study selection and study characteristics
We included 36 studies in this metaanalysis (1,047,683 women). The flow of study identification is shown in Figure 1 . Risk of publication bias was assessed by visual inspection of funnel plot; the symmetric plot suggested no publication bias ( Figure 2 ). Publication bias, assessed using Begg’s and Egger’s tests, showed no significant bias ( P = .87 and P = .71, respectively).
Thirty-one studies reported data about prior uterine evacuation for I-TOP, whereas 5 studies reported data regarding prior uterine evacuation for SAB ( Tables 2 , 3 , and 4 ).
Study | Study location | Type of study | Number of included women | Method of abortion | GA at abortion | Confounders adjusted | Primary outcome |
---|---|---|---|---|---|---|---|
Pantelakis et al, 1973 | Greece | Case-control | 4779 | Surgical | N/A | None | PTB |
Papaevangelou et al, 1973 | Greece | Retrospective cohort | 3467 | Surgical | N/A | None | PTB |
Daling and Emmanuel, 1975 | Taiwan | Retrospective cohort | 1516 | Surgical | N/A | None | PTB |
Daling and Emmanuel, 1977 | United States | Retrospective cohort | 553 | D&E | N/A | None | PTB |
Van der Slikke and Treffers, 1978 | The Netherlands | Case-control | 3432 | Surgical | N/A | None | GA at delivery |
World Health Organization, 1979 | Europe | Retrospective cohort | 3352 | Surgical | N/A | None | GA at delivery |
Obel, 1979 | Denmark | Prospective cohort | 497 | Surgical | N/A | None | Placental complications |
Mandelin and Karjalainen, 1979 | Finland | Prospective cohort | 696 | Surgical | N/A | None | Birthweight |
Meirik et al, 1982 | Sweden | Prospective cohort | 1442 | Vacuum | < 13 wks | Marital status, smoking | Birthweight |
Linn et al, 1983 | United States | Retrospective cohort | 9823 | Surgical | N/A | Age, ethnicity, smoking, economic status, parity | Birthweight |
Meirik et al, 1983 | Sweden | Retrospective cohort | 1292 | Vacuum | < 13 wks | Marital status, smoking | PTB |
Meirik et al, 1984 | Sweden | Prospective cohort | 269 | Prostaglandins followed by D&E | < 13 wks | Parity | PTB |
Park et al, 1984 | Korea | Case-control | 681 | Surgical | None | N/A | PTB |
Frank et al, 1985 | United Kingdom | Prospective cohort | 1545 | Surgical | < 22 wks | Age, marital status, gestational age at entry | LBW |
Pickering and Forbes, 1985 | United Kingdom | Population-based cohort study | 7000 | Surgical | N/A | Maternal age, height, sex of infant, marital status, social class | PTB |
Lekea-Karanika et al, 1990 | Greece | Case-control | 4391 | Surgical | N/A | Race, smoking | PTB |
Martius et al, 1998 | Germany | Population-based case-control study | 106,124 | Surgical | N/A | Gravidity, uterine surgery, type of work, urinary tract infection | PTB |
Zhou et al, 1999 | Denmark | Retrospective cohort | 64,125 | Surgical | <14 wks | Maternal age | PTB |
Henriet and Kaminski, 2001 | French | Population-based cohort study | 12,336 | Surgical | < 22 wks | Maternal age, parity, education, smoking | SGA |
Fox-Helias and Blondel, 2000 | French | Case-control | 17,411 | D&E | N/A | None | PTB |
Che et al, 2001 | China | Retrospective cohort | 2707 | Vacuum | N/A | Parental age, occupation, education, maternal BMI | PTB |
El-Bastawissi et al, 2003 | United States | Case-control | 654 | Surgical | N/A | Maternal age, race, smoking, parity | PTB |
Ancel et al, 2004 | Europe | Case-control | 7721 | Surgical | N/A | Maternal age, marital status, social class, smoking, parity | PTB |
Moreau et al, 2005 | French | Case-control | 2561 | Surgical | N/A | None | PTB |
Raatikainen et al, 2006 | Finland | Retrospective cohort | 26,967 | Vacuum | < 14 wks | Maternal age, weight, marital status, education, smoking, alcohol consumption, parity, uterine surgery | N/A |
Bhattacharya et al, 2012 | Scotland | Population-based cohort study | 577,510 | Vacuum | N/A | Maternal age, weight, smoking | PTB |
McCarthy et al, 2013 | Multicenter | Prospective cohort | 4812 | D&E | N/A | Maternal age, weight, smoking | PTB |
Woolner et al, 2014 | Scotland | Population-based cohort study | 45,631 | D&E | N/A | Smoking, social class | PTB |
Study | Study location | Type of study | Number of included women | Method of abortion | GA at abortion | Confounders adjusted | Primary outcome |
---|---|---|---|---|---|---|---|
Zhu et al, 2009 | China | Prospective cohort | 9363 | 200 mg mifepristone | < 14 wks | None | Placental complications |
Mirmilstein et al, 2009 | Australia | Prospective cohort | 154 | 400 μg misoprostol | 14-24 wks | None | PTB |
Winer et al, 2009 | France | Prospective cohort | 736 | 200 mg mifepristone followed by 400 μg misoprostol | < 22 wks | None | PTB |
Study | Study location | Type of study | Number of included women | Method of abortion | GA at abortion | Confounders adjusted | Primary outcome |
---|---|---|---|---|---|---|---|
Doyle 2000 | Taiwan | Case-control | 12,273 | Surgical | N/A | None | PTB |
Nguyen et al, 2004 | Vietnam | Case-control | 1709 | Surgical | N/A | None | PTB |
Smith et al, 2006 | Scotland | Population-based case-control | 84,391 | Surgical | N/A | None | PTB |
Selo-Ojeme and Tewari, 2006 | United Kingdom | Case-control | 206 | Surgical | N/A | None | PTB |
Freak-Poli et al, 2009 | Australia | Population-based case-control study | 25,554 | Surgical | < 20 wks | None | N/A |
The quality of the studies included in our metaanalysis was assessed by the Methodological Index for Non-Randomized Studies’ tool for assessing the risk of bias ( Figure 3 ). Nine of the included studies were retrospective cohorts, whereas 9 were prospective cohorts; 11 were case-control studies ; 7 were large, high-quality population-based studies. The majority had a low risk of bias in the aim and the time.
Synthesis of results
Uterine evacuation for induced termination of pregnancy or spontaneous abortion: combined data
In the overall population, women with a history of uterine evacuation for either I-TOP or SAB had a significantly higher risk of PTB (5.7% vs 5.0%; OR, 1.44, 95% CI, 1.09–1.90; Figure 4 A; 34 studies, 1,031,320 women), low birthweight (7.3% vs 5.9%; OR, 1.41, 95% CI, 1.22–1.62; Figure 4 B; 11 studies, 675,197 women), and small for gestational age (10.2% vs 9.0%; OR, 1.19, 95% CI, 1.01–1.42; Figure 4 C; 3 studies, 43,411 women) compared with controls (ie, women without a history of uterine evacuation).
Induced termination of pregnancy
Of the 31 studies reporting data regarding I-TOP, 28 included 913,297 women with a history of surgical I-TOP, whereas 3 included 10,253 women with a prior medical I-TOP ( Tables 2 and 3 ). Women with a history of uterine evacuation for I-TOP had a significantly higher risk of PTB (5.5% vs 4.4%; OR, 1.52, 95% CI, 1.09–2.13; Figure 5 A, 29 studies, 907,187 women), low birthweight (7.3% vs 5.9%; OR, 1.41, 95% CI, 1.22–1.62; Figure 5 B; 11 studies, 675,197 women), and small for gestational age (10.2% vs 9.0%; OR, 1.19, 95% CI, 1.01–1.42; Figure 5 C; 3 studies, 43,411 women) compared with controls (ie, women without history of uterine evacuation for I-TOP).
Surgically induced termination of pregnancy
Table 2 shows the characteristics of the included studies on surgical I-TOP. A total of 913,297 women from 28 studies with at least 1 prior surgical I-TOP were included. Seventeen studies reported information on confounders and adjusted risk estimates. Ten studies included only singleton gestations. The vast majority (27 of the 28) stratified data for number of prior I-TOP, whereas 1 did not report informative data about it.
Most of the studies had incidence of PTB as the primary outcome. Regarding the method of abortion, 5 studies defined the procedure as only VA, 5 studies defined the procedure as only D&E, whereas the others used both methods. One study reported the use of prostaglandins followed by D&E.
Women with a prior surgical I-TOP had a significantly higher risk of PTB (5.4% vs 4.4%; OR, 1.52, 95% CI, 1.08–2.16; Figures 6 A; 27 studies, 906,297 women), low birthweight (7.3% vs 5.9%; OR, 1.41, 95% CI, 1.22–1.62; Figure 6 B; 11 studies, 675,197 women), and small for gestational age (10.2% vs 9.0%; OR, 1.19, 95% CI, 1.01–1.42; Figure 6 C; 3 studies, 43,411 women) compared with controls (ie, women without a history of uterine evacuation for I-TOP).
The risk of PTB was still significantly higher after adjusting for confounders statistically proven, including marital status, smoking, age, ethnicity, economic status, parity, maternal height, race, social class, gestational age at entry, gravidity, parity, parental age, education, body mass index, uterine surgery, type of work, alcohol consumption, urinary tract infection, and sex of the infant (aOR, 1.25, 95% CI, 1.13–1.38; Figure 7 ; 16 studies, 874,080 women).
Subgroup analysis: method of abortion
Comparing the women with a prior surgical I-TOP with those who did not, both VA (3.6% vs 3.1%; OR, 1.20, 95% CI, 1.16–1.24; Figure 8 ; 5 studies, 609,912 women) and D&E (5.5% vs 4.3%; OR, 1.39, 95% CI, 1.08–1.80; Figure 9 ; 5 studies, 68,679 women) were associated with an increased risk of PTB. Moreover, by using an indirect comparison metaanalysis, we found that women who received D&E had a significantly higher risk of PTB compared with those who received VA (5.5% vs 3.6%; OR, 1.54, 95% CI, 1.38–1.73).
Subgroup analysis: number of prior I-TOP
Women with only 1 prior surgical I-TOP had a significantly higher risk of PTB compared with those who did not have any prior I-TOP (5.1% vs 4.4%; OR, 1.53, 95% CI, 1.02–2.31; Figure 10 A; 23 studies, 875,356 women). Women with more than 1 prior surgical I-TOP had a significantly higher risk of PTB compared with those without any prior I-TOP (23.4% vs 8.6%; OR, 1.98, 95% CI, 1.46–2.68; Figure 10 B; 9 studies, 165,085 women). Moreover, by using an indirect comparison metaanalysis, we found that women with more than 1 prior surgical I-TOP had a significantly higher risk of PTB compared with those who had only 1 prior surgical I-TOP (23.4% vs 5.1%; OR, 5.65, 95% CI, 5.10–6.25).
Subgroup analysis: number of fetuses
In a subgroup analysis of studies in which only singleton gestations in the index pregnancy were enrolled, women with a history of surgical I-TOP had a significantly higher risk of PTB compared with controls (9.6% vs 6.6%; OR, 1.45, 95% CI, 1.27–1.65; Figure 11 ; 10 studies, 152,668 women). No separate data about multiple gestations were reported in any studies.