Objective
We sought to examine outcomes of mifepristone and misoprostol for early pregnancy failure (EPF) treatment in a nonresearch setting.
Study Design
A protocol was developed for physicians to use mifepristone 200 mg orally and misoprostol 800 μg vaginally for EPF. Success rates were analyzed and an adjusted multivariable regression was used to identify factors predictive of success.
Results
Treatment success occurred in 99 (80%; 95% confidence interval, 72–87%) of 123 patients after mifepristone and a single dose of misoprostol and 102 (83%; 95% confidence interval, 75–89%) patients overall. The odds of successful medical treatment were increased in women with a diagnosis of intrauterine embryonic/fetal demise (odds ratio, 3.80) and decreased in women who made additional emergency department visits (odds ratio, 0.12).
Conclusion
Patients and clinicians may be more likely to intervene surgically with an EPF when a strict study protocol is not being followed.
Early pregnancy failure (EPF) is a common complication, occurring in approximately 15-20% of pregnancies. One in 4 women will experience a pregnancy loss over the course of her life. EPF can be further defined by the subtypes of pregnancy abnormality including intrauterine embryonic/fetal demise, anembryonic pregnancy, inevitable abortion, and incomplete abortion diagnosed at <12 weeks’ gestation.
Use of medical management with misoprostol, a prostaglandin E 1 analog, has become increasingly popular since the mid-1990s. Studies of misoprostol for treatment of EPF are difficult to compare because of the different criteria used to identify pregnancy failure, varying drug doses, delivery methods, criteria for success, and time interval between medication and outcome assessment. Accordingly, success rates using misoprostol vary significantly. The results of the first large, multicenter, randomized trial comparing surgical management to medical management with misoprostol were published by Zhang et al in 2005, reporting successful uterine evacuation using 1 or 2 doses of misoprostol 800 μg vaginally in 86% of women.
However, studies of elective medical abortion using a combination of mifepristone, a progesterone antagonist, with misoprostol demonstrate a higher success rate than misoprostol alone, exceeding 95% in women up to 9 weeks’ gestation. Additionally, multiple studies have shown efficacy of ≥90% >9 weeks’ gestation. Oddly, when used for EPF, the reported overall success rates with mifepristone and misoprostol are much lower, with wide discrepancies because of varying dosing regimens and differing criteria for success, ranging primarily from 67-84%. Only 2 studies have shown efficacy at ≥90%. Schreiber et al reported a 90% first-dose success rate in a small pilot study using 200 mg mifepristone and 800 μg misoprostol per vagina, which increased to 93% after a second dose of misoprostol. Kushwah and Singh reported 92% efficacy of 200 mg mifepristone followed by 600 μg misoprostol sublingually; however, medications were administered on an inpatient basis and up to 3 additional doses of misoprostol were given.
Importantly, all of the trials in the literature examine outcomes based on prospective research protocols. With the understanding that study protocols create situations that likely overestimate real-life outcomes, we sought to examine the results of such treatment by nonresearch providers in standard patient care settings. Unlike study protocols that require strict timing of study follow-up and treatment, real-life treatment could potentially include women having more input into how to proceed if initial treatment is unsuccessful and permit providers to recommend surgical intervention without needing to meet specified study criteria. We used an existing database created for a quality assurance investigation of clinical use of mifepristone and misoprostol for medical treatment of EPF to examine outcomes.
Materials and Methods
We reviewed a deidentified database created for quality assurance of an innovative treatment program using mifepristone and misoprostol for EPF at Magee-Womens Hospital of the University of Pittsburgh Medical Center. The institutional review board (IRB) at the University of Pittsburgh granted exempt approval for this analysis. The database included 28 women from an IRB-approved clinical trial, conducted from May 2006 through August 2007, and 128 women treated under the innovative treatment program from October 2007 through September 2009. The clinical trial was prospectively investigating outcomes with this treatment in a clinical setting. The 2 protocols were identical except that consent during the clinical trial had to be obtained by a study investigator whereas, for the innovative treatment program, the program-specific consent form could be provided by any clinician initiating treatment.
A nonviable pregnancy was defined by one of the following ultrasound criteria: (1) an embryonic pole ≥5 mm without cardiac activity ; (2) a gestational sac with mean sac diameter >16 mm and no embryonic pole ; (3) no documented growth on ultrasound examination of a confirmed pregnancy defined as growth in mean sac diameter of <3 mm over 5 days or <4 mm over 7 days ; or (4) an increase in serum β-human chorionic gonadotropin <15% over a 2-day period with a yolk sac present.
Mifepristone and misoprostol could be prescribed by any obstetrician/gynecologist affiliated with Magee-Womens Hospital for women who met one of the following criteria: (1) a nonviable pregnancy with an embryonic pole with a crown-rump length of ≤40 mm; (2) a nonviable pregnancy without an embryonic pole and with a mean sac diameter ≤45 mm; (3) an incomplete abortion with expulsion of some products of conception with or without active bleeding from the cervical os and an endometrial lining >30 mm and the uterine size <13 weeks by pelvic examination; or (4) an inevitable abortion: an intrauterine gestational sac on ultrasound examination and an open cervical os on digital examination with active vaginal bleeding with a mean sac diameter between 16-45 mm. In addition, patients needed to be hemodynamically stable, willing and able to sign the informed consent, and willing to comply with the treatment protocol and visit schedule.
Patients were ineligible for medical management if they had any of the following: (1) orthostatic hypotension; (2) contraindication to mifepristone such as chronic corticosteroid administration or adrenal disease; (3) contraindication to misoprostol such as glaucoma, mitral stenosis, sickle cell anemia, poorly controlled seizure disorder, or known allergy to prostaglandin; (4) clinical indication requiring karyotyping of products of conception; (5) prior surgical or medical attempted uterine evacuation, either self-imposed or physician provided; (6) known or suspected extrauterine pregnancy; (7) evidence of ovarian hyperstimulation syndrome; (8) known or suspected pelvic infection; (9) hemoglobin <9.5 g/dL; (10) known clotting defect or receiving anticoagulants; (11) cardiovascular disease; (12) current breast-feeding; (13) pregnancy with an intrauterine device in situ; (14) current use of any experimental drug; or (15) suspected or confirmed endometrial arteriovenous malformation.
After consent, measurements of hemoglobin, height, weight, blood pressure, and pulse were performed. Blood type was assessed and Rh-immune globulin administered to Rh-negative patients. A physician in the office or clinic setting or in the emergency department (ED) could initiate treatment; an obstetrics and gynecology faculty or resident provided all treatment in the ED. A prescription for 200 mg of mifepristone and 800 μg of misoprostol was completed by the physician and faxed to the hospital pharmacy. The physician provided all counseling and prescription for pain medication, and arranged follow-up. In addition, the physician provided standardized written information that reviewed how to properly take the medications, expectations for pain and bleeding, instructions for recognizing and managing problems, and telephone numbers for the treating physician and the hospital ED.
The patient was to swallow the mifepristone immediately after it was dispensed by the pharmacist and to use the misoprostol vaginally 24 hours later.
A follow-up transvaginal ultrasound examination was to be performed 1 week later either in the physician’s office or the hospital ultrasound department, and to discuss further management if needed. Physicians were instructed per the protocol that presence of a gestational sac or an endometrial thickness of >30 mm indicated further management was necessary. Options could include suction aspiration, a second dose of vaginal misoprostol, or expectant management. Those women who chose additional misoprostol or expectant management were to be followed up with weekly endovaginal ultrasound examinations until the gestational sac had expelled.
Success was determined by a diagnosis of complete abortion made at a follow-up visit by the treating physician, who was to use ultrasound criteria of absence of a gestational sac and an endometrial lining <30 mm. Failure was automatically assigned to any woman having a surgical intervention. Records for all University of Pittsburgh Medical Center hospitals were searched at the conclusion of data collection to determine whether any unknown surgery related to the pregnancy had been performed.
Data analysis was performed on the modified intent-to-treat (MITT) population, which consisted of all women who started the medical treatment and had at least 1 follow-up interaction ( Figure ). Physicians were defined in groups as gynecologic specialties faculty (academic gynecology faculty), general obstetrics and gynecology faculty (including all academic faculty and private faculty who were not members of the gynecologic specialties faculty), and residents. The resident cases were all supervised by gynecologic specialties faculty. Statistical analysis was performed using statistical software (SPSS, version 17.0; SPSS Inc, Chicago, IL). Associations of demographic, medical, and clinical factors with treatment success were evaluated using Student t test, χ 2 analysis, and Fisher’s exact test, where appropriate. Multivariable logistic regression was performed to identify factors that were independently associated with treatment success. Variables with a univariate P value < .2 were considered for inclusion in the regression analysis. The final multivariable model was adjusted for treating physician and enrollment location. Adjusted P values were calculated based on the Wald χ 2 test.
