Fertility control: Oral versus self-administered vaginal misoprostol at home before surgical termination of pregnancy: a randomised controlled trial


  • Kevin Sunde Oppegaard,

    Corresponding author
    1. Department of Gynaecology, Women and Children's Division, Ullevål University Hospital, Kirkeveien 166, N-0407 Oslo, Norway
    Search for more papers by this author
  • Erik Qvigstad,

    1. Department of Gynaecology, Women and Children's Division, Ullevål University Hospital, Kirkeveien 166, N-0407 Oslo, Norway
    Search for more papers by this author
  • Britt-Ingjerd Nesheim

    1. Department of Obstetrics, Women and Children's Division, Ullevål University Hospital, Kirkeveien 166, N-0407 Oslo, Norway
    Search for more papers by this author

Dr KS Oppegaard, Department of Gynaecology, Helse Finnmark, Klinikk Hammerfest, Sykehusveien 35, N-9613 Hammerfest, Norway.


Objective  To compare the impact of 400 μg oral versus self-administered vaginal misoprostol at home on pre-operative cervical priming in both primigravid and multigravid women prior to first trimester surgical abortion.

Design  Randomised controlled trial.

Setting  Norwegian University Teaching Hospital.

Sample  Three hundred and thirty-eight women undergoing surgical abortion between 7 and 12 weeks of gestation.

Methods  The women were randomised to either 400 μg of oral misoprostol the evening before or 400-μg of self-administered vaginal misoprostol at home the same day as vacuum aspiration.

Main outcome measures  Pre-operative cervical dilatation, complications and acceptability.

Results  The median cervical dilatation was 6.2 mm (range 0–11 mm) for the women in the 400 μg oral misoprostol and 6.5 mm (range 0–11 mm) in the 400-μg vaginal misoprostol groups. The median pre-operative dilatation was larger in multigravidae (6.4 and 6.7 mm for the oral and vaginal routes, respectively) than in primigravidae (5.8 and 6.0 mm, respectively). In primigravidae, 19% achieved a pre-operative dilatation of ≥7 mm, with no significant difference between oral and vaginal dosage. In multigravidae, 52% achieved a pre-operative dilatation of ≥7 mm with vaginal dosage, compared with 36% with oral dosage (P= 0.03). There was no difference between non-immigrant versus immigrant women in pre-operative cervical dilatation. The 400-μg oral dosage group had a higher risk of bleeding, compared with the group receiving 400-μg vaginal misoprostol [odds ratio (OR) = 10.4; confidence interval (CI) 5.2–20.8]. There was no difference between non-immigrant and immigrant women in acceptability of self-administered vaginal misoprostol; almost all women found this administration route acceptable. Complications were minor and were distributed equally between the two dosage groups.

Conclusions  The vaginal route will result in a satisfactory dilatation in about half of multigravidae but is much less effective in primigravidae. The oral route does not lead to satisfactory dilatation in either group and is associated with a higher occurrence of pre-operative bleeding. Self-administered vaginal misoprostol at home is highly acceptable.


Women in Norway requesting to terminate their pregnancy in the first trimester have the option of either the medical or surgical method before the ninth week of pregnancy.1 All pregnant women are offered free health care, including termination of pregnancy. Surgical methods are used exclusively for termination between 9 and 12 weeks of pregnancy. In 2004, 96% of all terminations were performed within the first trimester, and 61% of first trimester terminations were done surgically.2 Cervical ripening prior to first trimester abortions reduces the incidence of cervical injury.3,4 The ripening can be achieved with prostaglandins.5 Misoprostol is an inexpensive prostaglandin E1 analogue that is used for cervical ripening,6 but it is not approved for any use in pregnancy in any of the countries in which it is registered.7

The aim of this study was to assess whether the 400-μg self-administered vaginal misoprostol was superior to oral administration for pre-operative cervical priming. We were particularly interested in comparing the impact on primigravidae versus multigravidae. Misoprostol for cervical priming was first offered routinely in 1997 to Norwegian women via the oral route, based on published trials demonstrating the efficacy and simplicity of this method.8,9 However, the oral route has previously been shown to not result in a satisfactory pre-operative dilatation in the majority of Norwegian women.10 We wished to examine the acceptability of self-administered vaginal misoprostol at home, and if acceptability was dependent on whether the women were of Scandinavian origin or not.


The study protocol was designed according to the recommendations in the CONSORT statement.11 The study was a randomised, controlled, one-centre study at a central university gynaecological outpatients department. Both women and gynaecologists were aware of the treatment allocated, after the women were randomised (i.e. the study was not blinded). Women with an unwanted pregnancy are, according to Norwegian law, able to phone the department clerking office to book a consultation, without being referred by a general practitioner. This exception to the general rule of referral is so that the patients avoid having their consultation delayed. They may choose either a medical or surgical abortion according to personal preference. Women with a viable intrauterine pregnancy of 7–12 weeks of gestation confirmed by transvaginal ultrasound scan, requesting termination of pregnancy by vacuum aspiration and who had given informed consent, were eligible for study recruitment. The study was carried out from April 1st until August 31st, 2004 (1Fig. 1). The exclusion criteria were as follows: women with a gestation length outwith 7–12 weeks, women who were unable to communicate in Norwegian or English or women with a known allergy to misoprostol.

Figure 1.

Flow diagram.

Two outpatient clerks were primarily responsible for recruiting women. During the six months the study was carried out, approximately three quarters of the total number of women requesting a surgical termination of pregnancy participated in the study. The rest declined the offer, or did not fulfil the inclusion criteria. The study participants were assigned to either the 400-μg self-administered vaginal or 400-μg oral misoprostol groups. Randomisation was performed with permuted blocks using the randomisation plan generator, as described by Dallal.12

Opaque, numbered, sealed envelopes containing slips of paper assigning the study participant to either the self-administered vaginal or oral misoprostol group were prepared. The envelopes were delivered to the outpatient clerks. Hence, those involved in administering the intervention were blinded to the administration method, until the woman had opened the envelope revealing her designated group. The women received two tablets of 200-μg misoprostol together with written instructions, with the oral group instructed to swallow the tablets at 10 p.m. the evening before the operation (currently the standard treatment regime at the hospital). The self-administered vaginal group received written instructions to insert two tablets 200-μg misoprostol vaginally at 6 a.m. the morning of the operation. The women were instructed to insert the tablets as deep as possible vaginally, after voiding urine. Those not included in the study received the current treatment regime.

A gynaecologist at the pre-operative consultation recorded the following data: primigravida (yes/no), age, weight, height and gestational week on the day of the consultation (confirmed by transvaginal ultrasound scan). The woman was recorded as being non-immigrant, if she originated from any of the Nordic countries (Norway, Sweden, Denmark, Finland or Iceland). Any other women were recorded as immigrants.

On admission to the ward, approximately three days after the pre-operative consultation, nurses recorded symptoms and signs: Pre-operative bleeding/spotting was coded yes or no. The woman recorded pain experienced between the intake of misoprostol and the operation on a visual analogue scale score, the scale ranged from 0 (no pain) to 10 (unbearable pain). Women are not routinely issued with prescriptions for analgesics at the pre-operative consultation. The use of off-prescription analgesics before admittance to the ward was noted. Women who had self-administered misoprostol vaginally were asked to rate acceptability of the method on a scale from 1 to 4 (1 = completely acceptable; 2 = fairly acceptable; 3 = fairly unacceptable; 4 = completely unacceptable). The oral administration group were not asked to rate acceptability. Misoprostol tablets do not dissolute completely in the vagina.13,14 The gynaecologist noted if there were tablets visible vaginally before the operation and recorded the findings. The degree of pre-operative cervical dilatation was measured by passing Hegar dilatators through the cervix in ascending order starting with a size of 4 mm. The size of the largest dilatator passed into the cervical os without subjective resistance felt by the operator was recorded as the pre-operative degree of dilatation. If there was resistance with Hegar dilatator size four mm, the result was recorded as <4. If cervical dilatation was judged to be particularly difficult, this was recorded as ‘difficult procedure’. Any pre-operative complications were recorded. No routine follow up consultations were offered to the study participants, but all the patient records were reviewed after 60 days to see if the women had contacted the hospital. Any recorded post-operative complications were noted.

The sample size was calculated on the primary outcomes of cervical dilatation with the assumption of a type 1 error of 0.05 and a power of 0.90. The estimated pre-operative variability (SD) in cervical dilatation was 2.7 mm.15 We aimed at detecting a pre-operative difference in cervical dilatation of 1 mm between the two treatment groups. The minimum number of women in each group was calculated to be 170. We therefore aimed at including a total of 340 women to ensure enough power.

Statistical analyses between the two groups (self-administered oral and self-administered vaginal misoprostol) were tested by either a two-sided Mann–Whitney test, or Irwin–Fisher test for two-by-two tables. Possible influences of a skewed distribution of the other variables were analysed by stepwise linear regression. The data were recorded and analysed using SPSS 11. The effect of misoprostol on bleeding was estimated as ORs with 95% CIs.

The Regional Committee for Medical Research Ethics in Eastern Norway recommended the study protocol. Each participating woman in the study signed an informed consent.


The two treatment groups were comparable regarding basal clinical pre-operative characteristics (1Table 1). The cervical dilatations in the two treatment groups are shown in 2Table 2. In primigravidae, 19% achieved a pre-operative dilatation of ≥7 mm, with no significant difference between oral and vaginal dosage. In multigravidae, 52% achieved a pre-operative dilatation of ≥7 mm with vaginal dosage, compared with 36% with oral dosage (P= 0.03). A linear regression with cervical dilatation as the dependent variable, and treatment group, pregnancy history and ethnicity as independent variables, showed that pregnancy history was the only significant variable (B= 1.3, P < 0.0001). All women randomised to the self-administered vaginal administration group, except for six who withdrew before treatment, adhered to the protocol, and had inserted the tablets correctly. This was confirmed by visual assessment of tablets present vaginally by the gynaecologist pre-operatively.

Table 1.  Demographic characteristics of patients in the two study groups according to dosage route. Values are presented as mean [SD] or n (%)
CharacteristicOral misoprostol route (n= 158)Self-administered vaginal misoprostol route (n= 163)
Age (years)27.7 [6.9]28.1 [6.2]
Gestation length (days)59.4 [7.4]60.0 [8.1]
BMI (kg/m2)23.0 [3.4]22.9 [4.0]
Primigravidae56 (47.5%)62 (52.5%)
Multigravidae102 (49.2%)101 (49.8%)
Non-immigrant119 (51.7%)111 (48.3%)
Table 2.  Intra-operative findings and distribution of cervical dilatation in the two treatment groups. Values are presented as median (range)
 Dosage routeP
 Oral misoprostol (n= 158)Self-administered vaginal misoprostol (n= 163) 
Cervical dilatation (mm) (n= 321)6.2 (0–11)6.5 (0–11)0.3
Cervical dilatation (mm) primigravidae (n= 118)5.8 (0–11)6.0 (0–8)0.7
Cervical dilatation (mm) multigravidae (n= 203)6.4 (0–11)7.1 (0–11)0.06

Of the 163 women who received misoprostol by the vaginal route, 114 (70%) found it completely acceptable, 25 (15%) fairly acceptable, 12 (7%) fairly unacceptable, 3 (2%) completely unacceptable, while 9 (6%) did not answer the question. We found no difference in acceptability between non-immigrant and immigrant women.

Women who were immigrants experienced more pain, particularly with the oral misoprostol administration route (3Table 3). In a stepwise linear regression with pain as the dependent variable and dosage route, primigravidae or multigravidae, and ethnicity as independent variables, dosage route (B= 1.7, P < 0.001) and ethnicity (B= 0.9, P= 0.006) were statistically significant.

Table 3.  Pain in the two treatment groups. Pain was measured with a visual analogue scale score, scale ranges from 0 (no pain) to 10 (unbearable pain). Values are presented as median (range)
 Dosage routeP
Oral misoprostolSelf-administered vaginal misoprostol 
Pain2.9 (0–10)1.2 (0–10)<0.001
Non-immigrant2.3 (0–10)1.0 (0–10)<0.001
Immigrant4.0 (0–10)1.7 (0–9)<0.001

There was more bleeding with oral misoprostol (4Table 4). The odds ratio for pre-operative bleeding was 10.4 (95% CI 5.2–20.8) in the oral misoprostol group, as compared with the self-administered vaginal administration group. There was no difference in the occurrence of bleeding between non-immigrant (23.9%) and immigrant women (26.4%). No woman spontaneously aborted during the study. There were a total of 15 (4.4%) complications. Ten women were re-admitted with retained products of conception (six in the oral administration group) and had to undergo an extra curettage. Three women (one in the oral group) were treated with antibiotics for post-operative endometritis, but they could be managed as outpatients. One woman in the vaginal administration group sustained a false passage through the cervix during cervical dilatation but did not require any further treatment following her suction termination of pregnancy. There was one failed abortion (in the vaginal dosage group). No further complications were reported. Twelve of the cervical dilatations (six in each dosage group) were judged to be particularly difficult, eight of which were in primigravidae.

Table 4.  Occurrence of bleeding in the two treatment groups. Values are presented as n (%). The odds ratio of pre-operative bleeding was 10.4 in the oral group, compared with the vaginal group
 Dosage routeP
Oral misoprostolSelf-administered vaginal misoprostol 
Bleeding68 (43.0)11 (6.7)<0.001
Primigravidae25 (44.6)3 (4.8) 
Multigravidae43 (42.2)8 (7.9) 


A pre-operative cervical dilatation of 7 mm has, in previous randomised controlled trials, been classified as a ‘satisfactory’ achievement for the pre-operative priming agent.16 Our study suggests that 400-μg misoprostol administered via the vaginal route will give a cervical dilatation ≥7 mm in about half of multigravidae but is much less effective in primigravidae. The oral route does not lead to satisfactory dilatation in either group. Occurrence of bleeding was much more common in women taking oral misoprostol, compared with women using misoprostol vaginally. Women of Scandinavian origin recorded significantly lower pre-operative pain scores on a visual analogue scale score, compared with immigrant women. There were no differences between non-immigrant and immigrant women in pre-operative cervical dilatation, acceptability or occurrence of bleeding. Self-administered vaginal misoprostol at home was considered highly acceptable, regardless of ethnic group.

The main strength of this study is its large sample size. It was designed to be able to detect a difference between the two dosage groups of misoprostol of 1-mm cervical dilatation. The study was designed and conducted strictly in adherence with the CONSORT criteria.11 Furthermore, the study included all women requesting termination of pregnancy and was not confined to groups perceived to be at high risk of cervical injury.4 The dropout rate was low. Because of its size, it was also possible to estimate the rate of more rare complications. To our knowledge, this is the first study on cervical priming to compare different ethnic populations.

The main weakness is that the study was not blinded, and that both the patients and doctors involved in the data collection on cervical priming were aware of the treatment given after randomisation. In addition, there were 15 specialists involved in the data collection on cervical ripening. All the doctors involved had more than five years operating experience, and the majority were senior consultants, with more than 20 years experience. Although the doctors were individually instructed in assessing pre-operative cervical dilatation, it is difficult to be certain that every doctor's assessment was valid, reliable and unbiased but there is no reason to suspect differential misclassification.

Optimal cervical priming for surgical termination of pregnancy in the shortest time interval, with a minimum of side effects, is desirable. Randomised controlled studies have shown different results as to which timing intervals and which dosage is optimal, and whether the oral or the vaginal administration route is most effective.15,17–23 Few of the studies include information on the impact of previous pregnancies. A consensus has emerged that vaginal administration is superior to oral administration, with 400-μg misoprostol administered vaginally 3–4 hours before surgery providing optimal dilatation with minimal side effects.5,17,18,24–27 We have previously shown that the degree of cervical dilatation was lower after 400 μg oral misoprostol in women requesting termination at our hospital,10 compared with all other studies where 400 μg oral misoprostol was given.8,9,14,20,21,24,27 In addition, this dosage and administration route was associated with a high occurrence of light pre-operative bleeding in the majority of patients. In one of these studies,8 even placebo resulted in a larger dilatation than oral misoprostol in our experience for multigravidae, but not for primigravidae. We speculated that this might be due to population differences. In our current study, 91 women (28%) were registered of being of non-Scandinavian origin. Previous registration of country of origin in women seeking termination of pregnancy at our hospital between 1999 and 2003 included first and second generation immigrants from four continents.28 The majority of these immigrants come from non-Western countries traditionally associated with high fertility rates,29 such as the Asian subcontinent, the Middle East, Central and South America and Africa. We found no difference in cervical dilatation due to population differences in our study.

The vagina is considered an ideal site for drug administration.30 It has been suggested that the vaginal route may not be acceptable to many women due to social reasons,31 and nursing staff responsible for vaginal application of misoprostol prefer the oral or sublingual route.25,32,33 There has recently been interest in comparing sublingual with vaginal administration,32,34,35 and sublingual with oral administration.36 Sublingual misoprostol is an alternative to nurse-administered vaginal administration but is associated with more pre-operative side effects than vaginal administration.32,35

Studies have shown that it is safe, feasible and effective for misoprostol to be self-administered vaginally at a TOP unit 2–4 hours pre-operatively for cervical priming prior to manual vacuum aspiration in South Africa.16 One study has also suggested that supervised self-administration of vaginal misoprostol in the clinic is feasible in the UK,37 and that this regimen not only demedicalises the problem but also decreases the workload for the medical staff. This study did not perform qualitative research to find out women's views regarding satisfaction, however. To our knowledge, there have been no previous studies assessing self-administration of vaginal misoprostol at home prior to surgical termination of pregnancy. However, successful self-administration of vaginal misoprostol at home and in hospital prior to medical termination of pregnancy has been reported in the literature.38–42 It has previously been suggested that women of different cultural and ethnic backgrounds might have a completely different view regarding self-administration,37 although our study does not support this view.

It has been estimated that there are a large number of women at risk of not receiving a cervical priming agent pre-operatively in England and Wales.43 Self-administered vaginal misoprostol at home will prove particularly beneficial in units where the need for quick patient turnover means that the recommended priming interval of 3–4 hours after nurse-administered misoprostol is not practical. For the small minority who do not find vaginal administration acceptable, the alternative oral or sublingual route should be offered.

If one were to recommend treatment according to pregnancy history, we would recommend that primigravidae receive a dosage greater than 400 μg self-administered vaginal misoprostol at home 3–4 hours pre-operatively, while multigravidae receive 400 μg by the same route and time interval. We would also suggest that further studies to assess the impact of 600 μg misoprostol of self-administered misoprostol on both primigravidae and multigravidae prior to surgical termination of pregnancy.


This study would not have been possible without the support and enthusiasm of the health care personnel at the Department of Gynaecology at Ullevål University Hospital, who recruited the patients and ensured that the data collection was processed smoothly and competently.

Conflict of interest

None reported.