Cowett, Allison A. MD*; Cohen, Leeber S. MD*; Lichtenberg, E Steve MD, MPH*†; Stika, Catherine S. MD*
The endometrial thickness in the nonpregnant uterus can be reliably measured by transvaginal ultrasonography.1,2 The endometrial stripe is visualized in the sagittal plane, and the distance between the anterior and posterior walls at the interface between the endometrium and the myometrium is measured.3,4 A range of measurements of the normal endometrial stripe has been described for each phase of the menstrual cycle, where differences in endometrial thickness reflect the changing balance of hormonal influences.5
In the only published study to examine ultrasound findings after medical abortion, the mean endometrial thickness in patients treated with tamoxifen and misoprostol was 11.2 mm 1 week after the procedure, with 85% of values measuring 16 mm or less.6 Identifying a normal range for endometrial thickness and appearance after medical termination with mifepristone and misoprostol, a regimen now widely available in the United States, will guide clinicians in the interpretation of postprocedure ultrasound studies and may reduce the need for additional doses of medication, surgical intervention, or both. We performed this retrospective review of a cohort of women who underwent medical abortion at the estimated gestational age of 63 days or less to determine the ultrasound parameters associated with the need for additional doses of medication or surgical intervention after mifepristone and misoprostol pregnancy termination.
MATERIALS AND METHODS
The study protocol was approved by the Northwestern University Office for the Protection of Research Subjects Institutional Review Board before the initiation of chart review. The study was performed on a convenience sample. Charts from a cohort of patients who underwent medical termination at an urban family planning clinic from January 2001 through January 2002 were reviewed. A patient was included in the study if she underwent medical termination of pregnancy according to the protocol outlined below, she returned for follow-up evaluation, and her chart contained a baseline and postprocedure vaginal ultrasound examination.
All patients were managed similarly by using an evidence-based modification of the U.S. Food and Drug Administration (FDA)–approved protocol previously shown to be 96% effective at estimated gestational age of 63 days or less.7 Initial evaluation included a history, physical examination, urine pregnancy test, hematocrit, and ultrasonography to confirm the presence of an intrauterine pregnancy at 63 days or less. Patients received mifepristone (Mifeprex; Danco Laboratories, New York, NY) 200 mg in a single oral dose under direct observation. Patients who experienced emesis within 1 hour of mifepristone administration received a second dose. Patients were then given 4 misoprostol 200-μg tablets to be self-administrated intravaginally 24–72 hours after the mifepristone. Patients were instructed to return for follow-up within 7–10 days. At this follow-up visit, each patient was questioned about passage of pregnancy tissue, and a vaginal ultrasonography was performed. If the gestational sac was present, the patient was given an additional misoprostol dose. Absence of an intrauterine gestational sac confirmed a completed medical termination. If additional medications were required, patients were followed up for another 7–10 days. Surgical termination was considered if there was continued vaginal bleeding at 4 weeks or incomplete or unsuccessful termination after 3 doses of misoprostol. Patients with an incomplete medical termination had the option to undergo surgical termination at any time during the follow-up period. The global fee for a medical abortion covered all medications, Rh-immune globulin when indicated, unlimited ultrasonography, and procedural interventions.
The following data were extracted from the patient’s chart: date of birth, gravidity, parity, height, weight, last menstrual period, estimated gestational age based on vaginal ultrasound examination, date of mifepristone administration, and date of follow-up evaluation and ultrasonography. The postprocedure ultrasonogram was evaluated for endometrial thickness and the presence of a fluid interface, complex echoes, or a persistent gestational sac. The number of misoprostol doses, necessity of a dilation and curettage (D&C), hospital admission, and any other complications were recorded.
Successful medical termination was defined as an abortion completed after a single dose of mifepristone and misoprostol without complications, supplemental surgical procedures, or additional medications. Because the FDA recommends that patients return for a follow-up visit at approximately 14 days, patients whose first postprocedure ultrasonography occurred more than 17 days after initiation of therapy were excluded from the study.
Statistical analysis was performed with the SPSS 10.1 software package (SPSS Inc, Chicago, IL). For continuous variables, data are expressed as mean ± standard deviation (range). Height and weight data were converted to body surface area. The associations between endometrial stripe thickness and estimated gestational age at the time of the abortion, as well as posttreatment day of the ultrasound examination, were determined with the Pearson correlation coefficient. Ultrasound data were also grouped by week postprocedure, and one-way analysis of variance (ANOVA) was performed for endometrial thickness. When the ANOVA F test was significant at P < .05, post hoc analysis was performed by using the Scheffé method. Preliminary univariate analyses were performed for comparisons between women who successfully aborted after a single dose of medication and those who required additional therapy. The Student t test was used for continuous variables. A χ2 or Fisher exact test was performed for categorical variables. Multivariable logistic regression with forward stepwise selection was then performed to estimate the independent effect of the baseline ultrasound predictors. All analyses were 2-tailed, with a significance level of P < .05.
Six hundred eighty-four medical terminations were performed during the 13-month study period, and 572 patients returned for a postprocedure ultrasound examination. Of these, 525 patients returned at 17 days or less after initiation of therapy and were included in the analysis. All patients received 1 oral dose of mifepristone 200 mg. One patient aborted the pregnancy before self-administration of misoprostol; all others received at least 1 intravaginal misoprostol 800-μg dose. Patient characteristics for all patients and those who had successful versus failed medical terminations are presented in Table 1. There was a statistically significant difference in estimated gestational age between the success and the failure groups (45.3 ± 7.26 days and 49.2 ± 7.51 days, respectively; P = .02).
Of the 525 women who returned within 17 days of their initial treatment, 437 (83.2%) had an ultrasound examination with a measurable endometrial stripe. The observed mean endometrial thickness in these women was 4.10 ± 1.80 mm (range 0.67–13.4 mm). Ultrasonograms could be evaluated for the presence of complex echoes, fluid interface, and persistent gestational sac in 516 (98.3%) subjects. Complex echoes were observed in 91 (17.6%) women. Fluid interface was observed in 35 (6.8%) women, and 6 (1.2%) women had a persistent gestational sac. Only 9 women had 2 of these findings, and 1 woman had all 3 present on ultrasound examination.
Overall, 96.4% of returning patients had a successful medical termination. Nineteen of the 525 (3.6%) were considered failed medical abortions because they required either more than 1 misoprostol dose, a surgical procedure, or both to complete the abortion process. In 14 of these patients, a second misoprostol dose produced successful abortion of the residual pregnancy tissue or clot. One patient received 3 doses of misoprostol. A second received 4 doses of misoprostol and underwent D&C for persistent bleeding. Three other D&C procedures were performed because of heavy vaginal bleeding, and 1 woman received a blood transfusion. No other complications occurred, and there were no hospitalizations.
The endometrial stripe was evaluated in relation to both estimated gestational age and timing of the follow-up ultrasonography. There was no significant correlation between the gestational age of the pregnancy at the time of abortion and the postprocedure endometrial thickness. As depicted in Figure 1, the interval from initiation of therapy to the follow-up ultrasonography was inversely correlated to endometrial thickness (r = –0.22; P < .001). When grouped by weekly intervals, the mean endometrial thickness was 4.39 ± 1.77 mm, 3.82 ± 1.76 mm, and 3.26 ± 1.85 mm in patients returning by day 7 (n = 255), between days 8 and 14 (n = 238), and between days 15 and 17 (n = 32), respectively. By using ANOVA and the Scheffé post hoc analysis, a statistically significant difference was observed in endometrial thickness between women examined during the first postprocedure week and those examined in both the second (P = .003) and third (P = .008) weeks. Although the endometrial thickness appeared to further decrease between days 15 and 17, this difference was not statistically significant.
Endometrial stripe thickness was greater in the patients who required additional medication or surgical intervention than in the patients who underwent complete abortion with a single course of medication (Table 2). However, given the wide range of endometrial thicknesses in both groups, this finding is not likely to be clinically useful. Figure 2 illustrates the receiver operating characteristic curve for cutoff points of endometrial thickness measurements ranging from 2 mm or less to 8 mm or less, demonstrating that there is no acceptable measurement at which false-positive and false-negative rates are minimized. In the 4 women who required a D&C, the initial postmedication ultrasound measurement of the endometrial stripe was not different from that in women who successfully aborted after a single course of medication (4.91 ± 1.62 mm versus 4.06 ± 1.80 mm, respectively; P = .34). Both the presence of a fluid interface and complex echoes on postprocedure ultrasound were significantly associated with the need for further clinical intervention (Table 3). A persistent gestational sac was apparent in 6 (31.6%) of the 19 patients in the failure group (P < .001).
Multivariable logistic regression was used to analyze the following ultrasound parameters: endometrial stripe, complex echoes, and fluid interface. Persistent gestational sac was not included in the model because it was uniformly predictive of failure. Only endometrial stripe thickness remained in the model as a significant positive predictor of an unsuccessful medical termination.
The safety and efficacy of medical termination of early pregnancy with mifepristone and misoprostol has been previously demonstrated in multiple studies.8,9 Postprocedure ultrasonography is used by the majority of providers to document the absence of a gestational sac and confirm a completed medical abortion. Despite the widespread use of ultrasonography, normal parameters for the endometrial appearance after medical abortion have not been described.
In our study, we note a statistically significant difference in the mean endometrial thickness in women who underwent a completed medical termination with a single course of medication versus those who required further medical or surgical intervention. However, the range of endometrial thicknesses in both groups was wide, with a significant overlap between the groups. Therefore, it is not possible to establish an endometrial thickness measurement above which additional medication or surgical intervention is universally indicated. As demonstrated by the receiver operating characteristic curve in Figure 2, choosing a low endometrial thickness cutoff point for clinical intervention would lead to the unnecessary treatment of a significant number of patients. Increasing the cutoff would decrease the false-positive rate and improve the specificity of the test measurement. However, this improved specificity would be at the expense of sensitivity, and fewer patients requiring further therapy would be identified. Although the presence of complex echoes and fluid interface were associated with termination failure on univariate analysis, these parameters fell out of the multivariable logistic regression model and should not be used to guide additional therapy in isolation from the overall clinical picture. The only ultrasound parameter that consistently predicted failed therapy was the persistence of a gestational sac. However, in our study, a gestational sac was present in only one third of patients who required further medical or surgical intervention. Given these findings, we recommend that clinicians intervene not only when a persistent gestational sac is seen on a postprocedure ultrasonogram, but also on the basis of compelling clinical indications.
There are several limitations to our study. One hundred fifty-nine (23.2%) of the 684 women treated during the study period failed to return for evaluation within 17 days of the procedure and were not included in the analysis. Ultrasound findings in the women who failed to return may have differed significantly from those included in our study. Because the ultrasound records were evaluated retrospectively, the endometrial stripe could not be measured in a second cohort of patients (88 of 525, 16.8%). Also related to our study’s retrospective nature, treating physicians were not blinded to ultrasound findings and could have been influenced by ultrasound results in their decision to provide further treatment.
Our mean endometrial thickness of approximately 4 mm was significantly smaller than that previously reported after other medical termination regimens and surgical abortion.6,10 This difference may be due to the abortifacient properties inherent in using mifepristone and misoprostol, as well as their doses, timing, and mode of administration.
Furthermore, we used a strict definition of failure that included patients requiring either surgical intervention or additional doses of medication for the completion of the procedure rather than surgical intervention alone. Although this is not the most common definition of failure, we used it because we sought to identify ultrasound parameters that would predict the need for both types of intervention. And finally, given the 96.4% effectiveness of the medical abortion protocol, the number of ultrasonograms obtained for patients who failed was small and may not have been adequate to make significant conclusions.
This study describes the ultrasound appearance of the endometrium in a cohort of patients after medical termination of early pregnancy with mifepristone and misoprostol. We conclude that clinicians should not base the decision to intervene on endometrial thickness measurements. Rather, providers of medical abortion should treat patients based on the presence of a persistent gestational sac on postprocedure ultrasonography or, in the absence of a gestational sac, compelling clinical signs and symptoms.
1. Delisle M, Villeneuve M, Boulvain M. Measurement of endometrial thickness with transvaginal ultrasonography: is it reproducible? J Ultrasound Med 1998;17:481–4.
2. Spandorfer SD, Arrendondo-Soberon F, Loret de Mola JR. Reliability of intraobserver and interobserver sonographic endometrial stripe thickness measurements. Fertil Steril 1998;70:152–4.
3. Ylöstalo PR. Ultrasonography of endometrium. Ann Med 1990;22:105–8.
4. Bourne T, Hamberger L, Hahlin M, Granberg S. Ultrasound in gynecology: endometrium. Int J Gynaecol Obstet 1997;56:115–27.
5. Fleischer AC, Mendelson EB, Böhm-Vélez M, Entman SS. Transvaginal and transabdominal sonography of the endometrium. Semin Ultrasound CT MR 1988;9:81–101.
6. Harwood B, Meckstroth KR, Mishell DR, Jain JK. Serum beta-human chorionic gonadotropin levels and endometrial thickness after medical abortion. Contraception 2001;63:255–6.
7. Schaff EA, Fielding SL, Eisenger SH, Stadalius LS, Fuller L. Low-dose mifepristone followed by vaginal misoprostol at 48 hours for abortion up to 63 days. Contraception 2000;61:41–6.
8. Newhall EP, Winikoff B. Abortion with mifepristone and misoprostol: regimens, efficacy, acceptability and future directions. Am J Obstet Gynecol 2000;183 suppl:S44–53.
9. Kruse B, Poppema S, Creinin MD, Paul M. Management of side effects and complications in medical abortion. Am J Obstet Gynecol 2000;183 suppl:S65–75.
10. Ben-Hava I, Aschkenazi S, Orvieto R, Perri T, Shalev J, Dicker D, et al. Spectrum of normal intrauterine cavity sonographic findings after first-trimester abortion. J Ultrasound Med 2001;20:1277–81.