Objective
The objective of the study was to determine the optimal procedure for midtrimester uterine evacuation.
Study Design
This was a retrospective cohort study of women undergoing midtrimester uterine evacuation by prostaglandin induction or dilation and evacuation (D&E). Primary outcome was composite complication, defined as any of the following: infection, need for additional surgery, unexpected admission or readmission, serious maternal morbidity, and/or maternal death.
Results
Two hundred twenty patients met inclusion criteria: 94 D&E and 126 induction. D&E was associated with less composite complications (15% vs 28%, P = .02), which persisted in adjusted analysis (adjusted odds ratio, 0.38; 95% confidence interval, 0.15–0.99; P = .05). Women in the induction group had higher rates of retained placenta requiring curettage (22% vs 2%, P = .01), whereas cervical injury was more common in the D&E group (5% vs 0%, P = .01). Median length of stay was significantly shorter in the D&E group (5.7 hours vs 28.4 hours, P < .001).
Conclusion
Midtrimester D&E is associated with fewer complications than prostaglandin induction.
Midtrimester uterine evacuation is a process in which the contents of the uterus are removed during the second trimester of pregnancy. It can be performed electively, for fetal or pregnancy complications, or for maternal medical indications. Techniques for uterine evacuation include intrauterine prostaglandin instillation, prostaglandin induction, and dilation and evacuation (D&E). Older studies suggested that D&E is the safest of these methods. Changes in maternal characteristics, physician experience, abortion techniques, indications for prenatal screening for fetal anomalies, and increasing number of abortions for fetal anomalies over the last decade warrant reevaluation of the relative safety of these techniques.
Two contemporary studies have found that D&E continues to be safer than induction for second-trimester uterine evacuation. Their retrospective/nonrandomized design, limited information regarding technique (such as use of ultrasound guidance), variation in clinical protocols between and within the 2 studies, and discrepancy in complication rates limits the certainty and generalizability of their conclusions.
The purpose of our study therefore was to compare the relative safety of ultrasound-guided D&E and prostaglandin induction for midtrimester uterine evacuation in a single institution with consistent clinical protocols.
Materials and Methods
We performed a retrospective cohort study of all women undergoing a planned midtrimester uterine evacuation procedure (13 0/7 weeks to 23 6/7 weeks) at our institution, Lehigh Valley Health Network (Allentown, PA) from January. 2005 to June 2010. This time period was chosen because a change in providers at the beginning of the time period resulted in a more standardized approach to offering and performing midtrimester uterine evacuation procedures.
Women undergoing D&E by an experienced provider under continuous ultrasound guidance were compared with those undergoing a prostaglandin induction during the study period. Patients were identified by procedure (Current Procedural Terminology) and diagnosis ( International Classification of Diseases , ninth revision) codes. To identify women who may have been missed by our initial search criteria, we also reviewed Report of Induced Termination of Pregnancy forms (completion of these forms for each every elective termination are required by the Pennsylvania Department of Health) as well as delivery logs to identify any inductions for fetal demise.
Exclusion criteria included those in active labor and/or advanced cervical dilation during the initial evaluation, preexisting chorioamnionitis, and pregnancies outside the gestational age range of interest. Women in active labor and/or advanced cervical dilation were excluded because they did not require cervical ripening/dilation, an aspect of midtrimester termination that poses potential risk. Women with preexisting chorioamnionitis were excluded because the increased maternal risk required significant deviation from the clinical protocols of those undergoing planned procedures. Medical records were reviewed for maternal demographics (age, race/ethnicity, insurance type, parity), potential comorbidities (maternal medical conditions, previous uterine surgery, invasive testing), and procedure indications, details, and complications.
The primary outcome was occurrence of a composite complication, defined as any one of the following events: infection requiring the use of antibiotics for chorioamnionitis, endometritis or sepsis; blood transfusion; need for additional surgery (including cervical repair, uterine perforation repair, dilation and curettage [D&C], exploratory laparotomy or diagnostic laparoscopy); unexpected admission or readmission to the hospital for procedure-related complications within 6 weeks of the procedure; serious maternal morbidity; and/or maternal death. D&C included immediate and delayed procedures. Cervical repair included suturing of the cervix for laceration repair and/or to achieve hemostasis (in cases of bleeding without clear laceration such as for a puncture site from the tenaculum).
A composite complication was chosen as our primary outcome because we believed that it was clinically useful for counseling patients regarding overall procedural risk. We also evaluated serious complications (a subset of of composite complications) because these complications are associated with increased short- and long-term morbidity and, as such, are more likely to influence provider and patient decision making regarding choice of procedure. Serious complications were defined as any one of the following: blood transfusion, need for major additional surgery (laparotomy or laparoscopy), serious maternal morbidity, and/or maternal death. Our secondary outcome was length of stay.
Lehigh Valley Health Network serves as a referral center for patients with suspected fetal anomalies and/or pregnancy complications from the surrounding community and, as such, most women planning a second-trimester uterine evacuation procedure in our institution are seen by the Maternal-Fetal Medicine (MFM) team, which consists of perinatologists, genetic counselors, and nurse practitioners, all experienced in counseling patients regarding pregnancy options. Women in the second-trimester planning a uterine evacuation procedure are offered both D&E and induction. Fetuses with ultrasound findings suspicious for aneuploidy are encouraged to have invasive testing (amniocentesis or chorionic villus sampling [CVS]) as part of their evaluation. Women who choose to terminate without having had genetic testing are offered genetic testing at the time of cardioplegia or the evacuation procedure. In cases in which an autopsy may be helpful in making a diagnosis, patients are encouraged to consider induction. Some providers routinely recommend D&E at 13-14 weeks because of provider perception that D&E is easier than induction at earlier gestational ages.
In our institution, D&E is routinely performed as an outpatient procedure in the operating room under continuous ultrasound guidance by experienced providers. Cervical ripening with laminaria is used for all patients undergoing D&E greater than 16 weeks’ gestation and for select patients prior to 16 weeks. Induction is performed as an inpatient procedure with vaginal misoprostol or dinoprostone. Misoprostol is used most often as the first-line agent. The regimen is at the discretion of the provider, typically misoprostol 200-400 μm every 3-4 hours per vagina (using 100 or 200 μg tablets) or dinoprostone suppository 20 mg per vagina every 3 hours. Oxytocin augmentation is occasionally used. The provider may switch agents if the patient does not appear to be responding to the initial agent, usually after multiple doses. In general, induction agents are administered until delivery of both the fetus and the placenta.
Cervical ripening prior to induction with laminaria or a Foley bulb prior to or on the day of induction is also occasionally performed at the discretion of the provider. All pregnancies 20 weeks or longer have injection of intracardiac potassium chloride (KCl) to induce fetal demise prior to the procedure, regardless of procedure type. This is performed 1-2 days prior to D&E and 0-2 days before induction.
Management of the third stage of labor is at the discretion of the provider. An attempt is made to deliver the placenta at the time of delivery (within 20-30 minutes). If the placenta does not deliver during that time, and there is no evidence of significant bleeding, women are generally allowed as many as 4-6 hours to deliver the placenta with continued administration of the induction agent during that time period. If the placenta does not deliver by the end of that time period or if there is significant bleeding at any time, a D&C is performed. Obstetrics and gynecology residents actively participate in the care of both the induction and the D&E patients under the direct supervision of attending physicians experienced in these procedures.
Student t test, χ 2 , Mann Whitney U , and Fisher’s exact test were used to compare characteristics of the women by uterine evacuation method. Risk ratios were calculated to identify predictors of composite complications. Multivariate logistic regression was used to calculate adjusted odds ratios, controlling for potential confounders as identified in unadjusted analysis and those that have been identified in the literature as being associated with complications. This study was reviewed and approved by the Lehigh Valley Health Network Institutional Review Board.
Results
Within the study period, 245 patients were identified as having undergone a midtrimester uterine evacuation; of these, 25 were subsequently excluded because they did not meet inclusion criteria (13 were not within the appropriate gestational age range, 8 had chorioamnionitis on presentation, 3 presented with advanced cervical dilation and prolapsing membranes, and 1 did not actually have an evacuation procedure), for a final cohort of 220 patients. Of those 220 patients, 94 had a midtrimester D&E and 126 had a midtrimester prostaglandin induction.
Baseline characteristics of both groups are outlined in Table 1 . Women undergoing D&E were older, more likely to be white and to have private insurance, and less likely to be Hispanic. Patients undergoing D&E had an earlier median gestational age at the time of the procedure, were more likely to have undergone invasive diagnostic testing, and were more likely to have had cervical ripening prior to the procedure. Patients less than 16 weeks were more likely to have a D&E and those over 20 weeks were more likely to have an induction.
