Increasing access to safe abortion can decrease worldwide maternal morbidity and mortality.1 Medical abortion with mifepristone and misoprostol is a safe and effective method of pregnancy termination.2 Despite the potential for medical abortion to increase access to safe abortion, less than 25% of eligible abortions in the United States were medical in 2008.3
Women cite length of time of treatment and number of required office visits as reasons for not choosing medical abortion.4,5 In addition, women who find transvaginal ultrasonography invasive or have difficulty returning to the clinic may view ultrasound follow-up as an obstacle to medical abortion. Studies have shown that ultrasonography is useful for confirming expulsion of the gestational sac and that endometrial thickness should not be used as a determinant of completion of medical abortion.6,7 However, in practice, when medical abortion follow-up ultrasonographies are inconclusive or show a large volume of heterogeneous material in the uterus, additional interventions may be recommended despite true completion of abortion and minimal symptoms.
Alternatives to ultrasound follow-up include serial serum human chorionic gonadotropin (hCG) measurements, a standardized assessment of women's symptoms combined with low-sensitivity urine pregnancy testing, and telephone consultation with either low or high-sensitivity urine pregnancy testing.8–11 An 80% drop in serum hCG values 1 week after abortion is highly correlated with complete abortion and is an objective measure as opposed to ultrasound examination.12–16 Retrospective studies have reported the successful implementation of serum hCG follow-up in the United States; however, no randomized controlled studies have compared the two methods of follow-up.17,18 The objective of this study was to estimate whether women who follow-up with serum hCG testing have fewer unplanned follow-up visits and interventions than women who follow-up with ultrasonography.
MATERIALS AND METHODS
This randomized controlled trial of 376 women was conducted from May 2009 to April 2010 at the Planned Parenthood League of Massachusetts in Boston, Massachusetts, after receiving approval from the Partners institutional review board. Study participants were healthy women aged 18 years or older desiring medical abortion, less than or equal to 63 days of gestation by ultrasonography, and able to comply with the medical abortion regimen and follow-up. All participants needed to be proficient in English, have a working telephone, and be open to phone contact and agree to complete study surveys and follow-up. Exclusion criteria included inability to document intrauterine pregnancy, concern for ectopic pregnancy, clinical instability or signs of pelvic infection, medical contraindications to medical abortion, and situations in which additional phone calls or contact by research staff might jeopardize patient safety (ie, domestic violence). All women who met recruitment criteria were invited to participate until our sample size requirement was fulfilled.
Once study participants were clinically cleared to proceed with a medical abortion and gave informed consent to participate, they were assigned a participant number and randomized in a 1:1 ratio to either medical abortion follow-up with serial serum hCG testing or medical abortion follow-up with ultrasonography. The random allocation sequence was computer-generated by a person who was not affiliated with the study and followed randomly permuted blocks with block sizes of four and six. The allocation sequence was concealed by the use of sealed, sequentially numbered opaque envelopes. At the time of randomization, research staff opened the next envelope, which contained the assigned follow-up method. The assigned follow-up method was coded as arm 1 or 2 when entered into a database by research staff for analysis. Although participants, clinicians administering the abortion, and research staff implementing the study were aware of assigned follow-up method, investigators conducting analyses remained blinded to assigned follow-up method until data analysis was complete.
Once a participant was randomized, a survey was administered verbally before her abortion procedure to collect basic demographic information. All women had a fingerstick to determine hemoglobin and Rh status. Women in the serum hCG follow-up arm also had a blood draw at this time to establish a baseline serum hCG value. Women who underwent medical abortion swallowed 200 mg mifepristone under direct observation and self-administered 800 micrograms buccal misoprostol 24–48 hours later per routine Planned Parenthood League of Massachusetts protocol. All women were given prescriptions for pain medication and emergency care precautions.
All study participants were scheduled to return to Planned Parenthood League of Massachusetts for routine follow-up 1–2 weeks after taking mifepristone. Women who were in the ultrasound follow-up arm were scheduled to return for an ultrasonography and visit with a clinician; and women in the serum hCG arm were scheduled to return for a serum hCG blood draw, which was sent to Quest Laboratories for analysis. Women in the serum hCG follow-up arm were informed of their serum hCG results by phone. Medical abortion was determined to be “complete” in the serum hCG follow-up arm if there was at least an 80% drop in preabortion to postabortion serum hCG values. Women who had an inadequate drop in serum hCG, who had symptoms concerning for ongoing pregnancy, or retained products of conception or infection were scheduled for a follow-up visit with a clinician.
At the routine follow-up visit, research staff administered a verbal postabortion survey to all participants to inquire about additional care, visits, and interventions as well as patient satisfaction with the assigned follow-up method. All research participants were again contacted by phone 4–6 weeks after their initial medical abortion visit for a second postabortion survey. This second postabortion survey captured any additional follow-up that was required and any delayed complications of the medical abortion. On completion of all surveys, medical records were reviewed for evidence of additional interventions and visits.
If women did not keep their scheduled follow-up ultrasound or serum hCG appointment, clinic staff made three attempts to contact them within 1 week to reschedule their appointment per Planned Parenthood League of Massachusetts clinical protocol. To minimize loss to follow-up, research staff collected multiple methods for contacting participants and made three additional attempts to schedule their follow-up appointment.
The primary outcome was the frequency of unplanned interventions or visits beyond the one routinely scheduled follow-up ultrasonography or follow-up serum hCG draw. The composite binary outcome of unplanned interventions or visits included: additional clinic or emergency room visits, additional ultrasonographies or blood draws, repeat dosing of misoprostol, and surgical evacuation. The secondary outcome was patient satisfaction with the assigned follow-up method.
It was estimated that 10% of those in the transvaginal ultrasound arm would have an unplanned intervention or visit compared with 2% in the hCG arm.12 For a comparison of these proportions with a two-sided significance level of .05, an approximate sample size of 149 people in each arm was estimated to achieve a power of approximately 0.80. A 20% loss to follow-up rate was estimated based on previous studies at Planned Parenthood League of Massachusetts, and the total study population was calculated, resulting in 376 participants. Our primary outcome was analyzed through intention-to-treat analysis using SAS 9.2 statistical software. Binary and ordinal outcomes were compared using Fisher’s exact tests and χ2 as appropriate. Numeric outcomes were compared using t tests or Wilcoxon rank-sum tests as appropriate. Logistic regression was performed to model outcomes with any identified confounders. Variables with an unadjusted P value of ≤.05 were considered for inclusion.
Figure 1 shows study participant flow. Of the 1,555 women who were screened for participation in the study, 192 (12%) were ineligible and 987 (63%) declined participation. Of those who were ineligible, 86 were not proficient in English (45%), 38 changed to surgical abortion (20%), 35 were minors (18%), 16 were not candidates for medical abortion (8%) and four were not reachable by phone (2%). A total of 376 participants was enrolled in the study and randomized to serum hCG or ultrasound follow-up. Of the 188 participants randomized to serum hCG follow-up, eight participants were withdrawn because they were found to be ineligible for a medical abortion (n=3), they had poor venous access that prohibited blood draws (n=3), they were not study candidates as a result of an inconclusive intrauterine pregnancy on ultrasonography (n=1), or they decided they preferred ultrasound follow-up (n=1). Of the 188 participants randomized to ultrasonography, one woman was withdrawn because she switched to a surgical abortion. Two serum hCG participants crossed over to the ultrasound group and none crossed over from ultrasound to the serum hCG group. By the 2-week medical abortion follow-up visit, 21 of 180 participants (12%) in the serum hCG arm were lost to follow-up compared with 36 of 187 women (19%) in the ultrasound arm (P=0.04). By the 4-week time point, a total of 38 of 180 participants (21%) were lost to follow-up in the serum hCG arm compared with 52 of 187 (28%) in the ultrasound arm (P=.14).
Overall, most participants were white (56%), single (83%), nulliparous (63%), and had completed high school (96%). Average participant age was 26±6 years and mean gestational age was 46±6 days. Patient demographics are presented by arm in Table 1. The only significant difference between arms was employment status with 77% (n=145) of the women in the serum hCG arm reporting current employment and 66% (n=124) of the women in the ultrasound follow-up arm reporting current employment (P=.02).
By the 2-week medical abortion follow-up visit, a total of 23 participants reported an unplanned intervention or visit. As shown in Table 2, 8.2% (13/159) of these participants were in the serum hCG follow-up arm and 6.6% (10/151) were in the ultrasound follow-up arm (relative risk [RR] 1.23, 95% confidence interval [CI] 0.56–2.73, P=.60). There continued to be no significant difference between the two groups after logistic regression was performed to compare these percentages (odds ratio [OR] 1.33, 95% CI 0.60–2.94, P=.51). A breakdown of the type of unplanned events is reported in Table 2. There was no significant difference in type of unplanned interventions or visits between the two arms.
By 4-weeks postabortion, a total of 32 participants reported an unplanned intervention or visit. As shown in Table 3, by 4 weeks postabortion, similar percentages of women in each arm had an unplanned intervention or visit, 12.7% (18/142) in the serum hCG follow-up arm compared with 10.4% (14/135) in the ultrasound arm (RR 1.22, 95% CI 0.63–2.36, P=.55). There continued to be no significant difference between the two groups after logistic regression was performed to compare these percentages (OR 1.28, 95% CI 0.66–2.48, P=.49). There was no significant difference in type of additional follow-up and intervention between the two arms (Table 3). By 4 weeks postabortion, 4.4% (6/135) in the ultrasound arm and 1.4% (2/142) in the serum hCG arm had undergone surgical evacuation (RR 0.32, 95% CI 0.07–1.54, P=.16). One ongoing pregnancy was diagnosed in the hCG follow-up arm and this patient underwent a dilation and curettage procedure. Pain (n=3), bleeding (n=4), and persistent tissue in the uterus (n=4) were the other indications for dilation and curettage. There were no adverse events in either arm.
The secondary outcome, satisfaction with assigned follow-up method, was collected at the 2-week medical abortion follow-up visit and analyzed in an intent-to-treat manner. Eighty-eight percent (140/159) of women in the serum hCG arm and 95% (144/151) of women in the ultrasound follow-up arm reported that they were satisfied or very satisfied with their assigned follow-up method (P=.25). Women in the serum hCG arm were asked what they liked about follow-up with serum hCG, and the most common responses were that the method was quick, easy, did not require disrobing, and was less invasive than ultrasound follow-up. When the same women were asked what they did not like about serum hCG follow-up, the most common complaints were that the method was painful, involved a needlestick and sometimes multiple attempts at obtaining blood, and that immediate results were not available. Women in the ultrasound arm were asked what they liked about follow-up with ultrasonography, and the most common responses were that the follow-up method was quick, easy, painless, visual, and provided immediate results. When the same women were asked what they did not like about ultrasound follow-up, they reported that the ultrasonography was uncomfortable and invasive. Overall, when asked whether they were satisfied with the abortion procedure, 95% (145/153) of women in the serum hCG arm reported that they were satisfied or very satisfied compared with 87% (129/149) in the ultrasound follow-up arm (P=.02).
Finally, at the 2-week medical abortion follow-up, women were asked whether they would have been interested in using a home pregnancy test for medical abortion follow-up. Of the 310 women who responded, 66% (n=205) reported that they would have been interested in a home urine test and 32% (n=99) stated that they would not have been interested. Of note, 72% (115/159) of the participants in the serum hCG follow-up group reported interest in the home urine pregnancy test for follow-up compared with 60% (90/151) of the participants in the ultrasound follow-up group (P=.02).
This randomized trial compares the effectiveness of serum hCG and ultrasound follow-up and suggests that they are similar. We required participants to return to the clinic for their serum hCG testing to standardize laboratory follow-up and create a uniform study population to focus purely on follow-up method and its relation to unnecessary interventions. We found that medical abortion follow-up with in-clinic serum hCG did not result in fewer unplanned interventions and visits than follow-up with ultrasonography. Routine follow-up for medical abortion is usually scheduled 1–2 weeks after mifepristone administration to assess abortion completion and diagnose any complications in a timely manner. Our study did not show a statistically significant difference in unplanned interventions and visits between the hCG and ultrasound follow-up arms by the clinically relevant 2-week follow-up visit. We extended data collection to include a 4-week phone call to capture possible delayed complications and interventions. Although we did not note a difference in unplanned interventions and visits between arms by 4-weeks postabortion, we did not have adequate statistical power to detect a difference at this time point as a result of loss to follow-up. Overall, the number of unplanned interventions and visits were low and similar between groups at both the 2- and 4- week postabortion time points.
Of note, by 4 weeks postabortion, 4.4% of the participants in the ultrasound follow-up arm had undergone surgical evacuation compared with 1.4% of the participants in the serum hCG follow-up arm (RR 0.32, 95% CI 0.07–1.54, P=.16). Our study was not designed to detect small differences in rates of suction curettage. A much larger study would be required to evaluate whether ultrasound follow-up may be more likely than serum testing to identify findings such as a thickened endometrial stripe containing heterogenous material that result in treatment by suction curettage.
In our study, women were more likely to complete the 2-week follow-up visit if they were returning to the clinic for serum hCG testing compared with ultrasound follow-up. This may speak to the relative ease of blood-based follow-up even in a setting where it required return to the clinic. Serum hCG follow-up does not require a visit with a clinician, can be quicker, and may be considered less invasive by some women. A major advantage of serum follow-up is that it offers women the option of following up at more convenient locations. Interestingly, Horning et al showed that in clinical practice where women are able to choose whether to return to the clinic for ultrasonography or have serum testing in the community, the population of women who chose not to return to the clinic were more likely to live far from the clinic, be unemployed, and have had prior pregnancies and abortions. In this setting, community-based serum testing was actually associated with lower rates of follow-up than return to clinic testing.18 Societal factors and inherent personality characteristics can certainly influence the type and location of medical abortion follow-up a woman chooses and whether she completes follow-up. Some women may prefer returning to the clinic for their follow-up, be it serum or ultrasonography, for psychological reasons because returning to the clinic may formalize the follow-up process and provide the opportunity to speak with a clinician. By offering patients a variety of medical abortion follow-up options, we can help increase access to medical abortion and improve the acceptability of the service.
Many women have strong preferences in terms of their follow-up method and consider factors such as the relative discomfort of transvaginal ultrasonography compared with needlesticks, the need to return to the abortion clinic, desire for visual confirmation of complete abortion, desire for clinician interaction, and cost. In our study, the majority of participants in both groups was satisfied with their assigned method of follow-up. Women who did not like the serum hCG follow-up primarily felt that the blood draw was painful and did not provide immediate results, whereas women who did not like the ultrasound follow-up found the test to be invasive and uncomfortable. In addition to follow-up with serum hCG, other options that have been explored include use of a semiquantitative urine pregnancy test,19 in-clinic follow-up with standardized questions and a low sensitivity urine pregnancy test,8 telephone consultation and a low-sensitivity home pregnancy test administered at 2 weeks postabortion,11 or telephone consultation at 1 week postabortion followed by a high-sensitivity home urine pregnancy test administered at 1 month postabortion.9 Interestingly, we found that significantly more women who were assigned to the serum hCG group were interested in a home urine pregnancy test follow-up than women assigned to the ultrasound follow-up group. This may indicate that women in the ultrasound group felt reassured by visual confirmation of complete abortion or their interaction with a clinician and having had that experience, they are less interested in a home-based follow-up method.
Our study is an addition to the small body of literature that demonstrates the effectiveness and acceptability of serum hCG follow-up and adds that neither method of follow-up resulted in many unnecessary interventions. Women often have preferences for medical abortion follow-up based on personal circumstances, comfort with the medical system, and convenience. By increasing the options that women have for medical abortion follow-up, we can remove barriers to access and provide patient-centered care. Given similar effectiveness between the follow-up methods studied, patient preference, time, and cost should factor into the decision of which method of follow-up to use.
1. World Health Organization (WHO). Unsafe abortion: global and regional estimates of incidence of unsafe abortion and associated mortality in 2003. Geneva (Switzerland): WHO; 2007.
2. Jones RK, Henshaw SK. Mifepristone for early medical abortion: experiences in France, Great Britain and Sweden. Perspect Sex Reprod Health 2002;34:154–61.
3. Jones RK, Kooistra K. Abortion incidence and access to services in the United States, 2008. Perspect Sex Reprod Health 2011;43:41–50.
4. Wiebe ER. Choosing between surgical abortions and medical abortions induced with methotrexate and misoprostol. Contraception 1997;55:67–71.
5. Newhall EP, Winikoff B. Abortion with mifepristone and misoprostol: regimens, efficacy, acceptability and future directions. Am J Obstet Gynecol 2000;183:S44–53.
6. Cowett AA, Cohen LS, Lichtenberg ES, Stika CS. Ultrasound evaluation of the endometrium after medical termination of pregnancy. Obstet Gynecol 2004;103:871–5.
7. Reeves MF, Fox MC, Lohr PA, Creinin MD. Endometrial thickness following medical abortion is not predictive of subsequent surgical intervention. Ultrasound Obstet Gynecol 2009;34:104–9.
8. Clark W, Bracken H, Tanenhaus J, Schweikert S, Lichtenberg ES, Winikoff B. Alternatives to a routine follow-up visit for early medical abortion. Obstet Gynecol 2010;115:264–72.
9. Perriera LK, Reeves MF, Chen BA, Hohmann HL, Hayes J, Creinin MD. Feasibility of telephone follow-up after medical abortion. Contraception 2010;81:143–9.
10. Grossman D, Grindlay K. Alternatives to ultrasound for follow-up after medication abortion: a systematic review. Contraception 2011;83:504–10.
11. Cameron ST, Glasier A, Dewart H, Johnstone A, Burnside A. Telephone follow-up and self-performed urine pregnancy testing after early medical abortion: a service evaluation. Contraception 2012;86:67–73.
12. Fiala C, Safar P, Bygdeman M, Gemzell-Danielsson K. Verifying the effectiveness of medical abortion; ultrasound versus hCG testing. Eur J Obstet Gynecol Reprod Biol 2003;109:190–5.
13. Parashar P, Iversen OE, Midbøe G, Myking O, Bjørge L. Medical abortion in the first trimester: the use of serum hCG and endometrial thickness as markers of completeness. Eur J Contracept Reprod Health Care 2007;12:366–71.
14. Chou SY, Chen CY, Chiang HK, Chow PK, Wang CC, Hsu CS. Monitoring medical abortion using mifepristone/misoprostol combination with ultrasonogram and serum human chorionic gonadotropin. Taiwan J Obstet Gynecol 2006;45:48–52.
15. Walker K, Schaff E, Fielding S, Fuller L. Monitoring serum chorionic gonadotropin levels after mifepristone abortion. Contraception 2001;64:271–3.
16. Clark WH, Gold M, Grossman D, Winikoff B. Can mifepristone medical abortion be simplified? A review of the evidence and questions for future research. Contraception 2007;75:245–50.
17. Clark W, Panton T, Hann L, Gold M. Medication abortion employing routine sequential measurements of serum hCG and sonography only when indicated. Contraception 2007;75:131–5.
18. Horning EL, Chen BA, Meyn LA, Creinin MD. Comparison of medical abortion follow-up with serum human chorionic gonadotropin testing and in-office assessment. Contraception 2012;85:402–7.
19. Grossman D, Berdichevsky K, Larrea F, Beltran J. Accuracy of a semi-quantitative urine pregnancy test compared to serum beta-hCG measurement: a possible screening tool for ongoing pregnancy after medication abortion. Contraception 2007;76:101–4.