Medical abortion causes pain in almost all patients. This pain is often intense: average pain rating scores generally are reported as 6–8 on a 10-point scale.1–3 However, rigorous evidence to guide clinicians and women attempting to control this serious problem is remarkably scant.4
A randomized trial in patients undergoing first-trimester medical abortion found that a single administration of 1,600 mg ibuprofen at the onset of pain was superior to 2,000 mg acetaminophen in reducing pain severity.2 Notably, by the time the patients had ingested the analgesic, they had already developed pain rated as greater than 8 of 10, on average. Thus, this treatment regimen seems suboptimal for maximizing patient comfort.
An alternate strategy is to initiate the analgesic agent before pain onset. This prophylactic approach has intuitive appeal in that, unlike treatment after pain starts, it theoretically could prevent pain altogether. However, studies of oral analgesics given to patients undergoing medical abortion with or shortly before the misoprostol, which for most patients precipitates the pain, have thus far failed to demonstrate significant benefit in reducing pain.1,3,5–7 Notably, these studies all had limitations. None used currently recommended abortifacient regimens,8 and all studied only a single dose of the analgesic agent, which may have been insufficient to produce the desired effect. Furthermore, in all but one study,7 every participant received either an active drug or a placebo, which could itself have salutary effects.
To respond to these deficiencies, we conducted a randomized trial comparing two ibuprofen regimens to control pain in first-trimester abortion with 200 mg mifepristone followed by misoprostol. We chose ibuprofen as the analgesic agent because it is widely available, inexpensive, nonaddictive, and safe. Also, we postulated that its antiprostaglandin activity might be particularly effective for medical abortion pain, which is thought to result largely from prostaglandin-mediated uterine contractions. Our therapeutic regimen allowed participants to take the drug as needed after pain onset; the prophylactic regimen required participants to take the first dose 1 hour before the misoprostol and continue for 2 days regardless of pain. The prophylactic dosing schedule was designed to produce effective analgesic levels beginning before pain onset and lasting through the entire expected duration of pain.
PATIENTS AND METHODS
The trial was conducted at one clinic in New York, New York, and two in Chicago, Illinois, between October 2011 and December 2012. Enrollment at Chicago clinic B was terminated earlier than expected because of slow recruitment and suboptimal follow-up; this decision was made before any examination of study outcome data. The trial was registered before initiation with ClinicalTrials.gov (identifier NCT01457521) and was approved by the Chesapeake institutional review board.
Pregnant patients with a gestational age of 63 days or less who intended to have a medical abortion with mifepristone and misoprostol, who intended to begin the regimen the same day as enrollment into the study, and who did not plan to take any prescription or nonprescription nonsteroidal anti-inflammatory drug for reasons unrelated to the abortion were eligible for enrollment. Patients were disqualified if they had any specified contraindications to ibuprofen. We did not collect data about those who were screened out or who declined to participate. The medical abortion regimen administered to all participants was 200 mg mifepristone orally in the clinic followed by 800 micrograms misoprostol buccally, which the patient was instructed to take 1–2 days later at home.
After obtaining informed consent, site staff interviewed each volunteer to collect eligibility and baseline data. If she was eligible, staff assigned her to either the prophylactic regimen group or the therapeutic regimen group by opening the next unused consecutively numbered opaque envelope containing a random assignment. The one-to-one randomization scheme was stratified by site and used randomly permuted blocks with sizes of eight and 20 generated by computer by the study statistician before the start of enrollment. Staff dispensed at least 12 tablets of 800 mg ibuprofen to each participant and provided verbal and written instructions according to group assignment. Participants in the prophylactic group were told to take the first ibuprofen tablet 1 hour before the misoprostol and to take one additional tablet every 4–6 hours for 48 hours regardless of pain, then as needed. Participants in the therapeutic group were advised to take one ibuprofen tablet every 4–6 hours as needed starting at the onset of pain. All participants were told not to take more than four tablets (3,200 mg) per 24-hour period. Every participant was advised that she could obtain additional free ibuprofen tablets if needed by contacting the clinic, and she also received a prescription for acetaminophen with codeine or hydrocodone for pain not relieved by the study medication. Staff provided each participant with a diary card on which to record the date and time of misoprostol ingestion, her daily maximum pain on a numeric rating scale of 0–10,9,10 date and time of pain onset and first ibuprofen ingestion, number of study ibuprofen tablets taken daily, use of other analgesics, and vaginal bleeding or spotting through her follow-up visit.
Participants returned to the clinic according to the site's standard medical abortion protocol. At each follow-up visit, study staff collected the participant's diary and interviewed her without reference to the diary about her experience with pain during and after the abortion. Staff then reviewed the diary with the participant and updated it as needed. Staff questioned her about specific expected ibuprofen side effects (upset stomach or heartburn, nausea or vomiting, dizziness, and rash) and about any other adverse events and also recorded information about the abortion outcome. Further follow-up was arranged if necessary until the abortion was complete, all abortion pain had ceased, and all adverse events were resolved. If a participant was unable to come to the clinic for follow-up, study staff contacted her by telephone to collect key outcome information.
We calculated the sample size to test the null hypothesis of no average difference in the highest maximum pain score between groups. We aimed to enroll 250 participants into the study, which would provide 86% power to detect a mean difference between groups of 1.3 points on an 11-point numeric rating scale of pain severity (the minimum clinically significant difference in pain scores according to prior research11) assuming a standard deviation of 3 points and that 20% of participants would fail to follow up.
Analyses included all randomized participants who provided any follow-up data. Each participant was maintained in her randomly assigned group. The primary pain analysis compared the average maximum pain scores in the two groups as recorded on the diaries or as recalled at the follow-up visit if diary data were missing. We did not impute missing outcome data. We used nonparametric Mantel-Haenszel mean score tests (equivalent to nonparametric Wilcoxon rank-sum tests for ordinal variables),12 controlling for site, for all between-group comparisons. All pain variables were analyzed as ordinal. In addition, we used linear regression to estimate the effect of treatment on maximum pain score, to explore associations of this outcome with baseline characteristics, and to explore possible effect modification by these characteristics. We conducted all tests at the two-sided .05 significance level with no adjustment for multiple comparisons. All analyses used SAS 9.2.
All 250 participants who received random assignment met all admission criteria. Follow-up data were collected from 228 participants, approximately 90% of participants in both randomized groups (Fig. 1). Those who were lost were not notably different in any identified characteristic from those who completed the study except that they were more likely to have been enrolled at Chicago clinic B. In the prophylactic and therapeutic groups, respectively, 93% and 88% of participants with follow-up data submitted diaries containing pain scores; primary outcome data from the remaining participants were obtained by recall.
Most of the 250 enrolled participants were given 12 ibuprofen tablets at admission, but 18 in the prophylactic group and five in the therapeutic group requested and received more than 12 (specifically, they received 20 or 22 tablets) before leaving the clinic. No participant received additional ibuprofen tablets after admission.
The 228 participants in the analysis population had a median age of 25 years, and most had received some education after high school (Table 1). More than half rated the intensity of their usual menstrual cramps as 5 or less on a scale of 0–10. Only 18% of participants had previously had a medical abortion. Half expected to have pain with the upcoming abortion rated at least 7 on a scale of 0–10.
Most participants were reportedly compliant with their assigned analgesic regimen (Table 2). In the therapeutic group, a large majority of participants waited until the onset of pain before their first dose of ibuprofen. In the prophylactic group, almost all took the first ibuprofen tablet before the misoprostol; 89% took it at least 1 hour before. In addition, most participants in the prophylactic group took at least five ibuprofen tablets, which we estimated was the minimum number that could be construed to fulfill the instruction to take the pills four times a day for 2 days. The median number of ibuprofen tablets taken in the prophylactic and therapeutic groups, respectively, were nine and four (P<.001). In each group, more participants took exactly 12 tablets than took either 11 or 13 (Fig. 2). Of the 22 participants who received more than 12 tablets at admission, only six, all in the prophylactic group, actually took more than 12 tablets. Two participants, both in the prophylactic group, ever took more than five ibuprofen tablets on a single date. (Each took six on one date.) The average of the mean number of tablets taken per day between the first and last tablets was 2.1 in the prophylactic group and 1.4 in the therapeutic group (P<.001). More than 80% of participants in each group were compliant with all instructions relevant to their group.
Our analysis found no evidence of a difference in maximum pain score in the two groups or in any other measures of pain severity, duration, or acceptability (Table 3, Fig. 3). The mean maximum pain scores in the prophylactic and therapeutic groups were 7.1 (standard deviation 2.5) and 7.3 (standard deviation 2.2), respectively; the estimated mean difference, adjusted for site, was −0.17 (95% confidence limits –0.78 to 0.44). In both groups combined, at least 47% (107 of 228) of participants took analgesics other than the study ibuprofen; among participants who used fewer ibuprofen tablets than they received, this proportion was at least 41% (74 of 181).
Exploratory regression models indicated significant, independent associations of maximum pain score with younger age, greater usual menstrual cramps severity, and higher expected pain from the upcoming abortion but not with any other baseline characteristic (Table 1), including number of prior pregnancies or gestational age. No benefit of the prophylactic regimen was apparent in any subgroup defined by these or any other characteristic or in participants who received only the standard 12 ibuprofen tablets at admission, in those who were fully compliant with the assigned regimen, or in those who did or did not take analgesics other than the study ibuprofen. The prophylactic regimen was associated with significantly more pain than the therapeutic regimen among the 21 participants enrolled at Chicago clinic B (estimated mean difference 2.6, 95% confidence limits 0.6 to 4.6).
No serious adverse events occurred during the study. Vaginal bleeding or spotting occurred over a mean of 10.3 days in the prophylactic group and 10.8 days in the therapeutic group (P=.8; ranges 0–19 days and 4–26 days in the two groups, respectively). Significantly fewer participants reported nausea or vomiting or reported dizziness in the prophylactic group than in the therapeutic group, but we found no evidence of a difference between groups in the incidence of other expected ibuprofen side effects (Table 4). One unexpected adverse event was judged by the site investigator as possibly related to ibuprofen: a case of mouth tingling and numbness in the prophylactic group. Group assignment was not associated with the likelihood of having surgery or additional medical treatment to complete the abortion.
Our trial found no evidence that pretreatment with high-dose ibuprofen followed by around-the-clock administration offered any advantage over ibuprofen as needed in reducing pain in first-trimester medical abortion. This conclusion applied to every measure of pain that we examined, including maximum pain score, duration of pain, average daily pain, recalled maximum pain, qualitative pain description, acceptability of pain, and use of alternative analgesic agents. It further applied to every population subgroup that we evaluated, particularly those defined by characteristics shown in our study or others13,14 to be associated with amount of pain such as age, prior pregnancies, usual dysmenorrhea, gestational age, and expected abortion pain.
We do not believe that this disappointing result can be explained by flaws in our study. Follow-up in our trial was high: we collected primary outcome data from more than 90% of enrolled participants. Compliance with the assigned regimen was reasonably high in both groups, and indeed when we excluded noncompliant participants, the conclusion of the analysis was unchanged. Although the quality of paper diary data is notoriously questionable,15,16 the results of our primary analysis, which was largely based on diary entries, were consistent with the findings regarding both quantitative and qualitative descriptions of pain as recalled at follow-up, which may in any case be a better long-term indicator of the overall experience.17 Dispensing more than the standard number of 12 ibuprofen tablets preferentially to the prophylactic group at admission could have caused bias, but the effect presumably would have been to enhance pain relief in that group; it is unlikely to account for our negative findings.
Because the trial was not masked, site staff who counseled participants at enrollment could unintentionally have influenced their subsequent pain perceptions or their pain reports. However, before the trial, staff tended to expect that the prophylactic regimen would be superior, suggesting that any health care provider-induced bias would likely have favored that regimen. Moreover, in actual clinical practice, health care providers and patients would clearly be fully aware of the prescribed analgesic regimen, and thus any effect of that awareness on treatment success would not have been bias but rather an integral component of the treatment's mechanism. Like in any open-label trial, differential loss to follow-up between randomized groups was also a risk. Given our high retention rates, however, we do not suspect that loss caused notable bias.
In retrospect, perhaps our negative results should not be surprising. As noted previously, prior studies of the prophylactic approach to analgesia in medical abortion were not encouraging.1,3,5–7 Similarly, randomized trials of prophylactic ibuprofen for prevention of cramping associated with gynecologic procedures other than abortion, in particular intrauterine device insertion18 or colposcopy19, have failed to show benefit. Prophylactic ibuprofen for dysmenorrhea (starting treatment at or before the onset of menses) is widely recommended but to our knowledge has never been rigorously compared with therapeutic use. Moreover, the concerted effort over many years to demonstrate the use of prophylactic and around-the-clock analgesia for other types of pain, most notably postsurgical pain, has mostly been inconclusive.20
Our findings provide no strong guidance for clinical practice. Medical standards generally advocate using the lowest effective dose of any drug; this principle would argue against the prophylactic regimen, which entailed a higher total ibuprofen dose than the therapeutic regimen with no efficacy advantage. On the other hand, our study found no indication that the prophylactic regimen caused more side effects, and the higher ibuprofen exposure did not appear to increase the risk of abortion failure.
The more important finding of our trial is that clearly neither regimen was optimal. In both study groups, more than half the patients had maximum pain scores of 8 or higher on a scale of 0–10, and more than half had pain for 5 days or longer. Although most patients were satisfied overall with the abortion, 43% reported that the pain experience was at least somewhat unacceptable. Better pain management in medical abortion is urgently needed. The relatively high proportion of participants in both groups who took exactly 12 ibuprofen tablets suggests that some of our participants ran out of tablets; perhaps giving more than 12 initially would have been helpful. However, 41% of participants who had a sufficient supply of ibuprofen tablets nevertheless still also felt a need to use other analgesics. Because an ibuprofen regimen involving a higher ibuprofen dose or more frequent dosing than we used in our groups is unlikely to be widely accepted, future research should focus on other drugs or possibly multimodal approaches.
1. Fiala C, Swahn ML, Stephansson O, Gemzell-Danielsson K. The effect of non-steroidal anti-inflammatory drugs on medical abortion with mifepristone and misoprostol at 13–22 weeks gestation. Hum Reprod 2005;20:3072–7.
2. Livshits A, Machtinger R, David LB, Spira M, Moshe-Zahav A, Seidman DS. Ibuprofen and paracetamol for pain relief during medical abortion: a double-blind randomized controlled study. Fertil Steril 2009;91:1877–80.
3. Wiebe E. Pain control in medical abortion. Int J Gynaecol Obstet 2001;74:275–80.
4. Jackson E, Kapp N. Pain control in first-trimester and second-trimester medical termination of pregnancy: a systematic review. Contraception 2011;83:116–26.
5. Avraham S, Gat I, Duvdevani NR, Haas J, Frenkel Y, Seidman DS. Pre-emptive effect of ibuprofen versus placebo on pain relief and success rates of medical abortion: a double-blind, randomized, controlled study. Fertil Steril 2012;97:612–5.
6. Weber B, Fontan JE. Acetaminophen as a pain enhancer during voluntary interruption of pregnancy with mifepristone and sulprostone. Eur J Clin Pharmacol 1990;39:609.
7. Jain JK, Harwood B, Meckstroth KR, Mishell DR. Early pregnancy termination with vaginal misoprostol combined with loperamide and acetaminophen prophylaxis. Contraception 2001;63:217–21.
8. World Health Organization. Safe abortion: technical and policy guidance for health systems. 2nd ed. Geneva (Switzerland): World Health Organization; 2012.
9. Summers S. Evidence-based practice part 2: reliability and validity of selected acute pain instruments. J Perianesth Nurs 2001;16:35–40.
10. Flaherty SA. Pain measurement tools for clinical practice and research. AANA J 1996;64:133–40.
11. Bernstein SL, Bijur PE, Gallagher EJ. Relationship between intensity and relief in patients with acute severe pain. Am J Emerg Med 2006;24:162–6.
12. Stokes ME, Davis CS, Koch GG. Categorical data analysis for the SAS system. 2nd ed. Cary (NC): SAS Institute Inc; 2000.
13. Suhonen S, Tikka M, Kivinen S, Kauppila T. Pain during medical abortion: predicting factors from gynecologic history and medical staff evaluation of severity. Contraception 2011;83:357–61.
14. Penney G. Treatment of pain during medical abortion. Contraception 2006;74:45–7.
15. Mishell DR Jr, Guillebaud J, Westhoff C, Nelson AL, Kaunitz AM, Trussell J, et al.. Recommendations for standardization of data collection and analysis of bleeding in combined hormone contraceptive trials. Contraception 2007;75:11–5.
16. Stone AA, Shiffman S, Schwartz JE, Broderick JE, Hufford MR. Patient non-compliance with paper diaries. BMJ 2002;324:1193–4.
17. Kahneman D. Thinking fast and slow. New York (NY): Farrar, Straus, and Giroux; 2011.
18. Gemzell-Danielsson K, Mansour D, Fiala C, Kaunitz AM, Bahamondes L. Management of pain associated with the insertion of intrauterine contraceptives. Hum Reprod Update 2013 May 12 [epub ahead of print].
19. Church L, Oliver L, Dobie S, Madigan D, Ellsworth A. Analgesia for colposcopy: double-masked, randomized comparison of ibuprofen and benzocaine gel. Obstet Gynecol 2001;97:5–10.
20. Dahl JB, Kehlet H. Preventive analgesia. Curr Opin Anaesthesiol 2011;24:331–8.