Early pregnancy failure is a major public health problem throughout the world. Although approximately 15% of all pregnancies end in spontaneous miscarriage, there are also an estimated 46 million induced abortions annually. Many of these are performed illegally in unsafe situations resulting in approximately 78,000 deaths annually worldwide,1 with the majority of these deaths occurring as a result of septicemia and hemorrhage.2 In addition, many more women suffer long-term morbidity from pelvic infection, uterine perforation, and anemia. In Uganda, many women have abortions initiated by “backstreet abortionists” and then present to a health care facility with an incomplete abortion. Because spontaneous and induced abortions are usually impossible to distinguish, all women in this situation are treated as if illegally induced. For these women, many of whom will be human immunodeficiency virus (HIV) positive, the risk of infection is very high, and expectant management is therefore not considered a safe option.
Ultrasound examination is rarely available for the diagnosis of abortion in Uganda, and instead the diagnosis relies on clinical findings. The current treatment of choice in the first trimester is manual vacuum aspiration.3 This simple technique makes use of a syringe to produce the vacuum for a suction curettage. Like sharp curettage, manual vacuum aspiration is highly effective at uterine evacuation,4 but it causes less pain and bleeding and can be used in an outpatient setting.5
Misoprostol, a medical alternative to manual vacuum aspiration, is an orally active prostaglandin analogue that is inexpensive, easy to administer, and stable at room temperature. Previous studies of misoprostol for incomplete abortion have used a wide range of diagnostic criteria, doses, and routes and report success rates of between 13% and 100%.6 This variation largely occurs as a result of the different ways in which previous studies have used ultrasonography. If ultrasonography is used before treatment, approximately 29% of women will be found to have an empty uterus and will therefore not require treatment.7 Conversely, if ultrasonography is used after treatment to assess success, a thickened endometrium is often seen in the first few days. This finding has been frequently misinterpreted as demonstrating “retained products,” leading to unnecessary further intervention.8 Hence the reported “success rate” in previous trials is closely related to the length of time after treatment at which the uterus is reassessed with an ultrasound examination (Weeks A, Alia G. Ultrasonography may have role in assessing spontaneous miscarriage [letter]. BMJ 2001;323:694.). However, in women with prolonged bleeding following treatment, ultrasonography may have a role in distinguishing between retained products and endometritis, because those with a thin endometrium have been shown to have minimal tissue at evacuation.9 Therefore, the highest success rates (of 100%) are reported in situations where the passage of products before discharge from hospital is not a criterion for success, where endometrial thickness is not measured in clinically stable women following treatment, and where the final assessment of success is made 7 or more days posttreatment.10–12
In the developing world, surgical evacuation or manual vacuum aspiration is the standard treatment, partly because of concerns about the risk of pelvic infection. However, access to these procedures is often hampered by the lack of skilled personnel and facilities. In these settings, misoprostol could save many lives annually by providing a treatment that can be administered at even the most basic rural health center. Misoprostol and manual vacuum aspiration have, however, never previously been compared in a randomized clinical trial. (To identify trials, we searched the Cochrane central register of systematic reviews, controlled trials, and methodology reviews [The Cochrane Library, Issue 4, 2004] and PubMed database [1980–2005] using the key words “misoprostol” and “vacuum aspiration.”) This study was designed to compare the safety, efficacy, and acceptability of misoprostol and manual vacuum aspiration for the treatment of incomplete abortion in a low-income population in Uganda.
MATERIALS AND METHODS
Three hundred seventeen women who presented to Mulago Hospital in Kampala, Uganda, with an incomplete abortion of less than 13 weeks gestation (as estimated by date of last menstrual period and uterine size) were studied in a hospital-based randomized trial. Incomplete abortion was clinically diagnosed, based upon symptoms of bleeding with or without a history of the passage of tissue associated with central abdominal pain and an open cervical os. Although many of the participants will have had inevitable abortions, they were not treated differently within the trial protocol due to the difficulty in distinguishing the 2 entities clinically without ultrasonography. Women presenting with hemorrhage causing hemodynamic changes, any suspicion of an ectopic gestation (forniceal masses or tenderness), severe asthma, signs of severe infection, or known sensitivity to misoprostol were excluded. A patient information leaflet in English was given to each woman, and translated into the local language (Luganda) when necessary by the recruiting nurse/doctor. Ethical approval for this study was obtained from the Makerere University Research and Ethics Committee and the Institutional Review Board of the Population Council.
Computer-generated random numbers were used to allocate women to either manual vacuum aspiration or misoprostol. The allocation was written on cards and placed in consecutively numbered opaque sealed envelopes. When a woman met the criteria for eligibility and had given her written consent, the researcher (a ward nurse trained by the study team) removed the next envelope from the box and assigned the woman to the group indicated by the card in the study envelope. Neither the patients, assessors, nor the data analyzers were blinded to the allocation.
Women allocated to the misoprostol arm were given the drug orally in a dose of 600 g. A speculum examination was performed before discharge to remove any tissue from the cervical os. Regardless of whether products had passed (except in the event of continued heavy bleeding), all women were discharged after 6 hours. They were not routinely given analgesia, although it was available on request. Women allocated to manual vacuum aspiration were transferred to a clinical room where trained staff (doctors or midwives) administered intramuscular pethidine 50 mg and ergometrine 0.2 mg, per standard hospital protocol, and subsequently evacuated the uterus using manual vacuum aspiration.
Women in both groups were observed in the hospital for 4–6 hours following treatment. Before discharge, they were given doxycycline (100 mg/12 hours for 7 days) and metronidazole (400 mg 3 times daily for 5 days) because of the high incidence of septic abortion in Uganda. They also completed a questionnaire regarding adverse effects and maximum pain and blood loss (absent, mild, moderate, or severe). In addition, women were provided with the name and contact information for study research assistants to speak with in the event of complications or if they desired additional information about their treatment. The importance of the follow-up visit was stressed with all women.
A weekly follow-up clinic was set up for the study. Transport costs were provided for women to attend this clinic at between 7 and 14 days following their initial treatment. At the clinic, all women were examined clinically to assess their abortion status and asked a series of questions relating to the treatment's adverse effects and acceptability. Bimanual examinations were performed to ensure that the uterus had returned to its normal size. In the event of any signs of infection (uterine or adnexal tenderness, pyrexia, or discharge), swabs were taken and treatment provided as appropriate. If bleeding persisted or if an enlarged uterus was identified, women were referred for transvaginal ultrasonography. If retained products were seen on the ultrasonogram (A-P diameter > 1.5 cm), a manual vacuum aspiration was conducted.
The primary outcome measure was completeness of evacuation at follow-up using the allocated method alone. Assuming that manual vacuum aspiration is 98% effective,4 188 women per group were needed to detect an 8% reduction in efficacy with misoprostol (α = 0.05, β = 0.2). An additional 10% were sought to account for loss to follow-up. The trial started in August 2001 and was expected to enroll 400 women within 6 months. However, many women with incomplete abortions were ineligible because of advanced gestations or infection. As a result, enrollment was still ongoing in October 2002. At that point the principal investigators in Kampala were posted to jobs outside of the capital, and the study team therefore decided to stop recruitment and analyze the data from the first 317 women, which still had over 80% power to detect the required difference.
Data were entered into EpiInfo 6.0 (U.S. Centers for Disease Control and Prevention, Atlanta, GA) by 2 research assistants in Kampala, and all discrepancies were resolved by referral to the original study forms. Statistical analysis was conducted with SPSS 11 (SPSS Inc, Chicago, IL). Relative risk (RR) and χ2 tests (with use of the 2-tailed Fisher exact test, as appropriate) were used to analyze categorical data, and t tests were used for the analysis of continuous data. Statistical significance was defined as a P < .05 or if the 95% confidence interval of the relative risk excluded unity.
Three hundred seventeen women were recruited to the trial (Fig. 1). One woman in the misoprostol group and 7 women in the manual vacuum aspiration group were mistakenly given the wrong treatment. Their outcomes were included in the analysis under the “intention-to-treat” rule. Also included in the analysis were 6 additional women in the manual vacuum aspiration group for whom manual vacuum aspiration was not possible: for 5 women, the amount of retained products was too great for manual vacuum aspiration (the clinical estimate of gestation at trial entry had been underestimated), and for one woman, the os had closed by the time of manual vacuum aspiration, making the procedure impossible. There were 5 other women in the manual vacuum aspiration group for whom no outcome data were collected: 3 women discharged themselves before treatment, and 2 women were incorrectly excluded by the recruiter after randomization but before treatment—one because she did not fit the entry criteria and one because no manual vacuum aspiration kit was available.
Participant characteristics are reported in Table 1. Clinical characteristics at presentation were similar between the 2 groups, with women having an average reported gestational age of just over 9 weeks. Bleeding ranged from mild, for about two thirds of women, to moderate for another third. Few women enrolled in the study had severe bleeding at the first visit.
When the women who returned for follow-up at 1–2 weeks were analyzed, 96.3% of women assigned to misoprostol and 91.5% of women assigned to manual vacuum aspiration had a completed abortion following use of their allocated treatment alone (Table 2). Four women in the misoprostol group and one in the manual vacuum aspiration group required an additional re-evacuation of the uterus after the initial treatment. Five women randomized to the manual vacuum aspiration group did not have the allocated treatment because, at the time of evacuation, the uterus was clinically found to be over 13 weeks in size. Each was classified as a “treatment failure” and surgically evacuated using sharp curettage. Approximately 40% of participants were lost to follow-up. Significantly more women in the manual vacuum aspiration group than in the misoprostol group did not return for the follow-up appointment.
In the 6 hours following treatment, bleeding was the most common symptom reported by women in the 2 study groups (misoprostol 95.6%, manual vacuum aspiration 98.0%, Table 3).
When women were asked to rate the maximum pain they experienced, most reported that the maximum pain was mild or moderate (Table 3). Women in the manual vacuum aspiration group rated their pain significantly higher than those receiving misoprostol, despite the fact that all women undergoing manual vacuum aspiration received a 50-mg pethidine injection at the time of the operation. For maximum bleeding following treatment, the opposite was found, with women in the misoprostol group rating their bleeding significantly greater than those receiving manual vacuum aspiration.
The rate of complications was higher in the manual vacuum aspiration group (Table 2). In that group, 5 women experienced bleeding from cervical trauma during treatment (although no women needed more than a single cervical suture), and 3 women had pelvic infection at follow-up, requiring additional antibiotics. In the misoprostol group, only one woman experienced a complication. She had a laparotomy for a suspected ectopic pregnancy when she returned for her follow-up visit. A pelvic abscess was found, and she made a slow recovery on antibiotic treatment.
Table 4 shows women's reports of overall satisfaction when assessed at their follow-up visits. Regardless of the treatment they received, over 90% of women indicated that they were either “very satisfied” or “satisfied” with their experience, would choose the same method again, and would recommend it to a friend. Each woman was asked to name up to 2 best and worst features of her treatment. Women receiving misoprostol were more likely to report that the treatment had no worst feature (75.2% versus 39.4%, P ≤ .001), and women receiving manual vacuum aspiration were more likely to name pain as the treatment's worst feature (56.3% versus 17.8%, P ≤ .001). In contrast, women receiving misoprostol were more likely to report that having “less pain” was a best feature of the method (36.5% versus 20.0%, P = .02). Just over half of women in both groups indicated that the effectiveness of the treatment was its best feature.
This study shows that 600 μg oral misoprostol is as effective as manual vacuum aspiration for the treatment of incomplete abortion and, in so doing, suggests that the medical management of this condition may be feasible and successful in less-developed countries. The high success rate observed in the misoprostol group is consistent with that reported in similar European studies.13,14
Rates of complications and reported adverse effects were infrequent in both treatment groups. Pelvic infection was observed in one woman receiving misoprostol and 3 receiving manual vacuum aspiration, and each of these infections was successfully treated with antibiotics. This low rate probably reflects the use of routine antibiotics for all women. Compared with their manual vacuum aspiration counterparts, women receiving misoprostol reported significantly more bleeding but less pain at the time of the procedure. Irrespective of the method they received, women were generally satisfied with their treatment.
Pragmatic studies conducted in areas of high maternal mortality are crucial if the medical community is to reduce maternal mortality. Although clinical studies are difficult to complete in these settings, it is precisely such studies that must be undertaken, despite the problems of protocol violations and loss to follow-up. All randomized trials of incomplete abortion treatments thus far have used ultrasonography for diagnosis, yet their results may be of limited relevance to areas where ultrasound facilities are not readily available.
For low-resource settings, this study suggests that misoprostol may have a number of advantages over manual vacuum aspiration. Firstly, misoprostol appears to be a much more flexible treatment. For manual vacuum aspiration, a definite diagnosis of both abortion status and gestation needs to be made. The same may not be true of misoprostol. A review of the known failures in this study demonstrates that, while all 4 failures in the misoprostol group were attributable to failure of the drug, 5 of the 7 failures in the manual vacuum aspiration group were due to inaccurate clinical diagnosis, preventing the use of manual vacuum aspiration. In this pragmatic trial, clinical examination, rather than ultrasonography, was used to diagnose an incomplete abortion, as is the usual practice in Uganda. The problems of this practice are seen in the 5 women allocated to manual vacuum aspiration who were found to have retained second-trimester placentas at the time of treatment, and one woman who was found to have a closed cervix after a complete abortion. It is likely that a similar number of women from the misoprostol group were incorrectly diagnosed. However, the misoprostol users, unlike those receiving manual vacuum aspiration, were not re-examined before treatment, and so this remained undiscovered. The failure to correctly diagnose gestational age on initial clinical examination is a reality of clinical practice in many parts of the world. Indeed, the fact that there was no excess of failures in the misoprostol group suggests that 600 μg oral misoprostol may be equally effective at slightly later gestations. This study suggests that both methods can be safely offered to women without unnecessary recourse to ultrasound examination, which is expensive and dependent on skilled providers.
The second benefit of misoprostol is its ease of use. Manual vacuum aspiration requires a specific piece of equipment along with a trained operator. Even in a training center such as Mulago Hospital, manual vacuum aspiration kits or trained surgical providers may be unavailable (as was the case for 2 women in the manual vacuum aspiration arm). Misoprostol, however, is readily available at low cost (approximately $1.50 for a 200-μg tablet) in Kampala pharmacies and could be safely administered by midlevel providers. Access to misoprostol treatment, therefore, may be greater than that of manual vacuum aspiration.
The major weakness of this trial is the low rate of attendance at the follow-up visit. The low rate of follow up was recognized early in the study, and numerous attempts were made to improve it. Despite these efforts, which included a review of the quality of counseling offered to women before and just after treatment and an increase in the monetary sum provided to cover their transportation costs for additional hospital visits, nearly 40% of the study participants did not return to the hospital for follow-up care. The low rate of return is not that surprising, given that induced abortion is illegal and highly stigmatized in Uganda. Indeed, to receive any formal treatment for an abortion (whether induced or spontaneous) can be seen as shameful, and this may have inhibited women from attending their follow-up visits, irrespective of whether they had an manual vacuum aspiration or misoprostol. The discomfort of having an manual vacuum aspiration may also have discouraged women in that group from reattending, through fear of needing to undergo a further surgical technique. In the misoprostol group, the ease of simply swallowing 3 tablets may have led to less anxiety; hence, the finding that they were significantly more likely to return for follow-up.
Currently, Uganda has among the highest abortion-related morbidity and mortality rates in the world, and government hospitals such as Mulago Hospital in Kampala manage the majority of postabortion complications.1 In low-resource settings such as Uganda, inexpensive and easy-to-use treatments are badly needed, and the introduction of misoprostol could greatly benefit both women and providers. Postabortion care based on misoprostol treatment requires minimal technical skill and is ideal for rural health centers with limited facilities. Provided that clinicians can correctly assess whether the cervical os is open or not, they can safely administer misoprostol. Manual vacuum aspiration requires more equipment and more advanced staff and is therefore better suited to larger health centers for use as second-line therapy.
Future studies should address the appropriate application of misoprostol-based postabortion care outside of tertiary level facilities, as well as the development of home follow-up via simple symptom checklists to assist women in determining whether they need to return to a hospital after treatment. If we can develop practical training programs and treatment protocols for use in rural areas, then the use of misoprostol could potentially lead to a major reduction in abortion-related maternal morbidity and mortality.
1. World Health Organization. Unsafe abortion: global and regional estimates of incidence of mortality due to unsafe abortion with a listing of available country data. 3rd ed. Geneva: World Health Organization; 1997.
2. Uganda Bureau of Statistics (UBOS) and ORC Macro. Uganda Demographic and Health Survey 2000–2001. Calverton (MD): UBOS and ORC Macro; 2001.
3. Fraser R. Abortion. In: Lawson JB, Harrison KA, Bergstrom S, editors. Maternity care in developing countries. London (UK): RCOG Press; 2001. p. 289.
4. Greenslade FC, Leonard AH, Benson J, Winkler J, Henderson VL. Manual vacuum aspiration: a summary of clinical and programmatic experience worldwide. Carrboro (NC): IPAS; 1993.
5. Forna F, Gälmezoglu AM. Surgical procedures to evacuate incomplete abortion (Cochrane Review). In: The Cochrane Library, Issue 1. Chichester (UK): John Wiley & Sons; 2004.
6. Creinin MD, Schwartz JL, Guido RS, Pymar HC. Early pregnancy failure: current management concepts. Obstet Gynecol Survey 2001;56:105–113.
7. Chung T, Leung P, Cheung LP, Haines C, Chang AM. A medical approach to management of spontaneous abortion using misoprostol. Acta Obstet Gynecol Scand 1997;76:248–51.
8. Creinin MD, Harwood B, Guido RS, Fox MC, Zhang J. Endometrial thickness after misoprostol use for early pregnancy failure. Int J Gynecol Obstet 2004;86:22–6.
9. Chung TK, Cheung LP, Sahota DS, Haines CJ, Chang AM. Evaluation of the accuracy of transvaginal sonography for the assessment of retained products of conception after spontaneous abortion. Gynecol Obstet Invest 1998;45:190–3.
10. Hinshaw K. Medical uterine evacuations in the management of first trimester miscarriage. In: Prendiville W, Scott JR, editors. Clinical management of early pregnancy. London (UK): Arnold; 1999. p. 79–90
11. Ngai SW, Chan YM, Tang OS, Ho PC. Vaginal misoprostol as medical treatment for first trimester spontaneous miscarriage. Hum Reprod 2001;16:1493–6.
12. Bagratee JS, Khullar V, Regan L, Moodley J, Kagoro H. A randomized controlled trial comparing medical and expectant management of first trimester miscarriage. Hum Reprod 2004;19:266–71.
13. Henshaw RC, Cooper K, el-Refaey H, Smith NC, Templeton AA. Medical management of miscarriage: non-surgical uterine evacuation of incomplete and inevitable spontaneous abortion [published erratum appears in BMJ 1993;306:1303]. BMJ 1993;306:894–5.
14. Sahin HG, Sahin HA, Kocer M. Randomized outpatient clinical trial of medical evacuation and surgical curettage in incomplete miscarriage [published erratum appears in Eur J Contracept Reprod Health Care 2002;7:iv]. Eur J Contracept Reprod Health Care 2001;6:141–4.