Shannon, Caitlin S. MPH*; Winikoff, Beverly MD, MPH*; Hausknecht, Richard MD†; Schaff, Eric MD‡; Blumenthal, Paul D. MD§; Oyer, Deborah MD¶; Sankey, Heather MD∥; Wolff, Jessica NP∥; Goldberg, Rachel MPH*
In 2000, the United States Food and Drug Administration (FDA) approved a mifepristone (Mifeprex, Danco Laboratories, LLC, New York, NY) and misoprostol regimen for termination of pregnancies up to 49 days from the first day of the woman's last menstrual period (LMP). This approved regimen consists of oral administration of 600 mg mifepristone followed in 48 hours by the oral administration of 400 μg misoprostol, with both drugs taken in the provider's office and a follow-up visit around day 14.
A growing body of literature has shown a comparable level of efficacy with a lower dose of 200 mg mifepristone, combined with varying routes and doses of misoprostol.1–9 For example, recent studies have reported successes of a regimen of 200 mg mifepristone followed in 24 to 48 hours by 800 μg vaginal misoprostol in 90 to 98% of cases (Shannon C, Wiebe E, Jacot F, Guilbert E, Sheldon WR, Winikoff B. Regimens of misoprostol with mifepristone for early medical abortion, submitted article).6–8 Several studies also suggest that regimens combining 200 mg mifepristone and 400 μg or 600 μg oral misoprostol administered 24 to 48 hours after mifepristone are highly effective, with success rates from 91% to 96% (Shannon et al, submitted).3–5 Many of the trials of alternative regimens have also demonstrated the safety and efficacy of a regimen that offers women the choice of administering misoprostol at home (Shannon et al, submitted).4,8–9
Practice has come to reflect these evidence-based alternative regimens. For example, the Planned Parenthood Federation of America allows its affiliates to use 200 mg of mifepristone, with additional patient consent, and likewise to use home administration of misoprostol. Similarly, the National Abortion Federation advises providers in its protocol that “mifepristone 200 mg is as effective as 600 mg in a regimen with 400 μg misoprostol orally when used up to 49 days,” and that home administration of misoprostol has been found to be safe, effective, and acceptable.10 Individual providers are not limited to the regimens set forth in FDA-approved labeling but are frequently guided by accepted medical standards when determining whether to use drugs in evidence-based regimens.
This article reports on a prospective, open-label, multicenter trial of 200 mg mifepristone, followed in 48 hours by home administration of a single oral dose of 400 μg misoprostol. The rationale for this trial was to produce sufficient information to support re-labeling of mifepristone in 2 respects: lowering the recommended dose to 200 mg and incorporating administration of misoprostol at home. Doing so would serve to bring labeling in line with current practice and literature. In addition, it would remove barriers to use inherent in the current labeling, such as cost and number of office visits.
PATIENTS AND METHODS
The study was conducted in 5 centers in the United States. Each site had institutional review board approval. The sites represented several health care environments, including a Planned Parenthood affiliate, university hospitals, and private practices. Each of the centers was capable of performing surgical abortions and had experience performing mifepristone medical abortions.
Study participants were women who requested medical abortion for termination of pregnancy, were at least 18 years of age, had a positive urine pregnancy test, an intrauterine pregnancy dated at less than 50 days since LMP, were in generally good health, and had none of the conditions that are contraindicated in the U.S. labeling for mifepristone. (Administration of mifepristone and misoprostol for the termination of pregnancy is contraindicated in patients with any 1 of the following conditions: confirmed or suspected ectopic pregnancy or undiagnosed adnexal mass; intrauterine device in place; chronic adrenal failure; concurrent long-term corticosteroid therapy; history of allergy to mifepristone, misoprostol, or other prostaglandin; hemorrhagic disorders or concurrent anticoagulant therapy; or inherited porphyrias.) Other exclusion criteria were threatened abortion (as evidenced by vaginal bleeding), inability or unwillingness to return to the study center for a follow-up visit, and an unwillingness to complete a diary of adverse effects or be followed up at home in case of failure to return to the clinic at the scheduled date. Each participant gave written informed consent for her participation in the study and was advised to undergo surgical pregnancy termination if the medical method failed.
At the initial clinic visit on day 1, counseling, a complete medical history and physical examination, and gestational age determination were performed. Gestational age was assessed through at least 1 of the following 4 methods: menstrual history (36% of cases), bimanual examination (32%), transvaginal ultrasound (97%), and β-hCG levels (2%). Subjects who met all enrollment criteria and who gave written informed consent ingested 1 200 mg mifepristone tablet. Each woman was given 2 tablets of 200 μg misoprostol to take home and instructions to take those tablets orally in 48 hours. They were also given medication for pain according to clinic practice. Before leaving the clinic, subjects received a diary card on which to record each day any vaginal bleeding (heavier than, equal to, or less than menses), expulsion of conceptus, pain or cramps, vomiting, nausea, weakness, fever or chills, headache, other adverse effects, or use of concomitant medications between day 1 and day 15. They also received information describing symptoms requiring emergency treatment (such as very heavy bleeding, fever, and severe pain), and a telephone number to call to reach an on-call provider 24 hours per day. Finally, patients were counseled on use of contraception.
On day 3, women took 400 μg misoprostol orally at home. For 2 patients, complete abortion was confirmed by a study investigator on or before Day 3, and the patients were therefore advised not to administer misoprostol. On day 15, they returned for their second scheduled clinic visit and underwent clinical and gynecologic examinations. To determine the status of the termination, transvaginal ultrasound examination, bimanual examination, and β-hCG were carried out at the investigator's discretion. Patients with ongoing pregnancy (based on fetal cardiac activity or fetal growth consistent with a gestational age 2 weeks greater than at day 1) had vacuum aspirations. Patients with incomplete abortions, defined as pregnancy termination with partial expulsion of the products of conception, were given the option of either waiting 1 week for the process to complete or having a vacuum aspiration. Patients with complete abortions were discharged from the clinic's care. Bleeding did not have to have stopped for a patient to be discharged from the clinic's care.
For patients who chose to wait an additional week, another visit was scheduled, at which time a determination of the patient's abortion status was made. Women with incomplete abortions had vacuum aspirations and were discharged from the study immediately following the procedure.
Immediately before discharge from the study, all patients were given a short exit interview, including questions on the acceptability of the procedure, best and worst features of the procedure, and the experience with taking misoprostol at home. The study protocol was approved by the Institutional Review Boards of the Population Council, Johns Hopkins Bayview Medical Center, the University of Rochester, and the Western Institutional Review Board.
Complete medical abortion with no surgery before discharge from the study and no additional misoprostol administration was classified as treatment success. All outcomes that resulted in a surgical intervention were classified as treatment failures. Reasons for surgery were coded as 1) ongoing (continuing) pregnancy, 2) incomplete abortion, 3) medically necessary (ie, for bleeding or pain that warranted surgery), or 4) patient request. The duration and severity of bleeding and adverse effects and the incidence of all adverse events were evaluated. Medical treatments administered for any adverse event were analyzed. The acceptability of the procedure was also evaluated, based on closed and open-ended questions asked of the participants during the exit interview.
The overall sample size calculation was based on the number of patients needed to assure an efficacy of between 85% and 95%. The study was not intended to be a comparative trial. However, compared to the U.S. registration trial reported by Spitz et al11 of the regimen currently approved by the FDA, which demonstrated a 92.1% efficacy in termination of pregnancies in women with up to 49 days amenorrhea (N = 827), it would be possible to detect a 6% or greater difference from that efficacy (92.1 ± 6%) with a power of greater than 90% at the 95% confidence level (α = 0.05).
All statistical tests were performed as 2-tailed tests, and statistical significance was defined as having a P < .05 for the differences in relevant proportions or means. Comparison of proportions was done with χ2 tests, with Yates' correction used where necessary. Backward stepwise logistic regression analysis was used to evaluate the relation between success and various baseline patient characteristics; the significance level required for a variable to stay in the model was .051. All statistical analysis was performed using commercially available software (Statistical Package for the Social Sciences, SPSS Inc., Chicago, IL).
From January 2001 through May 2003, 376 patients were enrolled in the study. All of these patients received mifepristone at visit 1. Ten patients failed to return for their second visit; however, evidence suggesting a successful outcome was available for 1 of these women. This patient contacted the clinic several months later seeking a second medical abortion, indicating that the termination she had obtained as a study participant was complete. The remaining 9 patients were considered lost to follow-up and excluded from efficacy analyses with lack of evidence of either failure or success. An additional twelve patients were excluded from efficacy analyses because, in violation of the protocol, they received an additional dose of misoprostol at the follow-up visit. Another patient was excluded because the gestational age of her pregnancy exceeded 49 days since LMP. These 22 exclusions left a total of 354 women in the final sample.
In 2 of the study sites, women were initially counseled to take their pain medication (ibuprofen) and misoprostol simultaneously. Midway through the trial, an interim review of cases revealed an unusual number of ongoing pregnancies in these sites. Study staff were unable to find pharmacologic evidence that the practice of simultaneous administration of ibuprofen and misoprostol would increase the risk of ongoing pregnancy, but acknowledged that, theoretically, such a practice could reduce the absorption of misoprostol. Sites were instructed to stop advising simultaneous use of misoprostol and ibuprofen. However, because the practice was not a protocol violation, the subjects were retained in the efficacy analysis. Separate analysis of ongoing pregnancies in these sites is detailed in the section below.
Of the 354 women in the final efficacy analysis, misoprostol administration was confirmed for 352 (99%). Misoprostol administration occurred at home in all cases.
Table 1 presents baseline and demographic variables of the patients included in the efficacy analysis. The majority of patients (58.8%, n = 208) had gestations of between 43 and 49 days. The mean gestational age was 42.6 days. Most women had a high school education or higher, and almost two-thirds had at least some university education (64.9%, n = 226). The mean age of the study participants was 27.5 years, and 76.8% (n = 272) were not married. More than one half of the women had had a previous abortion (54.7%), and more than three-quarters had had a previous pregnancy (75.9%).
As displayed in Table 2, of the 354 patients included in the efficacy analysis, 91.5% (n = 324) had a successful termination of pregnancy. Logistic regression analysis indicated that successful termination of pregnancy was not significantly related to age, gravidity, number of previous abortions, or clinic site (data not shown).
A total of 8 patients (2.3%, 95% confidence interval [CI] = 1.1–4.4%) had incomplete abortions at the study's end, and 13 patients (3.6%, 95% CI = 2.0–5.3%) had ongoing pregnancies; all 21 had surgical completions. In addition, 8 patients (2.3%, 95% CI = 1.1–4.4%) had undergone medically indicated surgical interventions by study end, and 1 patient (0.3%, 95% CI = 0.0–1.6%) had a surgical completion at her request.
Three patients who were judged to have successful outcomes at the study end underwent a surgical procedure for bleeding more than 3 weeks after study drug administration. Due to this extended period, it was not possible to confirm that the surgery was because of the mifepristone-misoprostol regimen, but the events were treated conservatively, as if they had occurred during the study period, because they might potentially have been related to the procedure. Two of these patients were classified as study failures, and 1 was removed from the efficacy analysis because she had been administered an additional dose of misoprostol. Had these cases been left in the tabulation of successes, the overall study success rate would have been 92.1%.
In the 2 study sites where patients were initially counseled to take ibuprofen and misoprostol simultaneously, the rate of ongoing pregnancy was significantly higher than that in the remaining 3 sites (P = .04). The combined rate of ongoing pregnancy in these 2 sites was 7.1% (8/113); the rate was a full 10.2% before the ceasing of the practice. In contrast, the combined ongoing rate across the remaining study sites was 2.0% (5/241).
Of the 376 study participants, 39 made 1 or more unscheduled visits to the study centers. For 44% of these women (n = 17), no additional intervention was necessary. Reasons for unscheduled visits included heavy bleeding (28.2%; n = 11), cramping (12.8%; n = 5), anxiety (17.9%; n = 7), to request a surgical abortion (7.7%; n = 3), and the woman's belief that the abortion was incomplete (10.3%; n = 4).
There was 1 serious adverse event in the study. A patient was hospitalized for 4 days due to pelvic pain. The physician involved initially suspected infection due to complaints of pain and tenderness, but the pathology report did not support sepsis or uterine infection. After further investigation, a diagnosis of pelvic pain of unknown origin was recorded.
One patient required fluid replacement as a result of intractable vomiting. She underwent a manual vacuum aspiration at the study clinic. In addition, 9 patients were treated with oral antibiotics. Antibiotic use was not a reflection of infection in the current study: only 1 of the 9 patients treated with antibiotics was diagnosed with an infection (mild endometritis), based on tenderness elicited at physical examination. An additional patient was given antibiotics for a presumed, but not confirmed, infection. The remaining antibiotics were given prophylactically in conjunction with surgical procedures.
Bleeding is an expected component of mifepristone medical abortion. On average, patients experienced 2.0 days of heavy bleeding (eg, bleeding heavier than menses), 5.2 days of bleeding similar to menses, and 3.8 days of spotting (data not shown). (Because patients could mark more than 1 type of bleeding on a given day on their home study cards, these numbers do not sum to an overall mean bleeding duration.)
Four patients had a surgical intervention for heavy bleeding, and 3 had a surgical intervention for prolonged bleeding. The mean study days at surgery were 7 days for heavy bleeding (range 3 to 13 days) and 40.3 days (range 21 to 53 days) for prolonged bleeding.
As can be seen in Table 3, the most common adverse effects reported by patients were pain, nausea, weakness, headache, and dizziness. Most women who reported symptoms did not experience them before taking misoprostol. Almost all of the patients returning their home study cards (n = 338) reported having experienced at least 1 day of pain or cramps (93.2%), with 39.3% (133/338) reporting experiencing them before taking misoprostol. The median duration of pain or cramps was 3 days. Two-thirds of patients (66.6%) reported having experienced at least 1 day of nausea, with 42% (142/338) experiencing it before taking misoprostol. The median duration of nausea was 2 days.
Overall, the acceptability of the regimen was very high, with 63% of patients reporting that it was very satisfactory and an additional 23% reporting that it was satisfactory. Although there was some variation in the rates by study clinic (range 5.7% to 12.5%), dissatisfaction with the procedure was more likely related to whether the woman had a successful medical abortion: almost one half of all women (44.0%) who had failed medical abortions reported dissatisfaction, compared with only 3.1% of women with successful procedures.
During their exit interview, women were asked to describe the 2 best features and the 2 worst features of the procedure. As displayed in Table 4 4% of women said that they could think of no positive attributes of the method; more than double that proportion (8.6%) said that they could not think of any negative attributes. Twenty-three percent of women cited being able to be in the privacy of their own homes as 1 of the top 2 best features of the method. Other positive attributes cited commonly as 1 of the top 2 best features of the method were that it requires no surgery, is noninvasive, and involves taking pills (16.4%); it is fast, simple, and convenient (16.4%); and it results in little (if any) or less pain (10.3%). The most commonly cited worst features were the amount and length of pain or cramping (29.5%) and the amount and length of bleeding or clots (16.9%). In addition, 9% of women cited nausea or vomiting as 1 of the top 2 worst features, and 6.9% mentioned other adverse effects.
This multicenter trial tested a regimen of oral mifepristone and oral misoprostol similar to the regimen currently approved by the FDA except for a lowering of the oral mifepristone dose from 600 mg to 200 mg, and patient self-administration at home of the 400 μg oral misoprostol. The results confirm that this is a safe and effective method of early pregnancy termination.
As can be seen in Table 5, the conservatively calculated 91.5% rate of complete abortion in this study is not statistically different from that found in the U.S. registration trial of 827 women with amenorrhea of 49 days or less (92.1%, P = .722).11 The present study's success rate is also consistent with the results of other trials of the 200 mg mifepristone and 400 μg oral misoprostol regimen, which had success rates for various gestational age groups ranging from 91% to 96%.3,4,9 It also falls within the range of success rates reported in studies of 200 mg mifepristone and 800 μg misoprostol administered vaginally, which vary from 90% to 98%.1,6–8
Although the aggregate failure rate in the present study (5.9%) does not differ significantly from that of the Spitz et al11 study (5.7%, P = .867), the distribution of incomplete abortion and ongoing pregnancies rates is different. The rate of incomplete abortion in the current study was significantly lower than that of the previous trial (2.3% compared with 4.7%, P = .041). In contrast, the rate of ongoing pregnancy in the present study was 3.6%, a rate significantly different from the ongoing pregnancy rate of 1.0% in the Spitz study (P = .001). Although the cause for this apparent trade-off between ongoing pregnancies and incomplete abortions is unknown, several independent factors can be hypothesized. First, classification of study failures may have differed between the 2 trials due to changes in diagnostic criteria between providers over time or increased precision in ultrasonography in the years between the studies. Second, as described earlier, in the 2 study sites where patients were counseled to take ibuprofen and misoprostol simultaneously, the rate of ongoing pregnancy was significantly higher than that in the remaining 3 sites (7.1% compared with 2.0%; P = .04). The 2.0% ongoing rate across the study sites that did not ever counsel patients in this manner is not significantly different from the ongoing rate in the Spitz study (P = .30), and is also consistent with other studies of the 200 mg mifepristone and 400 μg oral misoprostol regimen, in which the ongoing rate has ranged from 1.1% to 2.8%.3,4,9
Consistent with other studies, almost all patients in this study reported some combination of bleeding, pain, and gastrointestinal complaints. However, the severity of those complaints resulted in only 1 hospitalization, no blood transfusions, and no emergency room visits. Nausea and vomiting were only severe enough to warrant intervention (intravenously fluids) in 1 patient. There were no significant differences in the frequency or severity of adverse events between this study and the earlier U.S. trial, despite the difference in dose and in home administration.
The regimen tested in the current study was not only effective and associated with few severe adverse effects or adverse events, but it was also highly acceptable, as illustrated by the 92% satisfaction rate among study successes. In addition, almost one half (41%) of women for whom the regimen did not work still reported that they were satisfied or very satisfied with the method. The most commonly cited positive feature of the method was its home administration: 23% of women named being at home or the privacy of the method as 1 of the top 2 best features of the method
In conclusion, the regimen of 200 mg mifepristone followed in 48 hours by home administration of 400 μg oral misoprostol is safe, effective, and acceptable for women seeking medical abortions of pregnancies of up to 49 days duration when compared with the regimen currently approved by the FDA. It reduces the number of office visits by one third (from 3 to 2), and saves two thirds of medication costs, increasing the feasibility of this method's use for both providers and patients.