8573

Saturday 19 December 1987

INDUCTION OF THERAPEUTIC ABORTION INEARLY PREGNANCY WITH MIFEPRISTONE IN

COMBINATION WITH PROSTAGLANDINPESSARY

MARY W. RODGER DAVID T. BAIRD

Department of Obstetrics and Gynaecology, Centre for ReproductiveBiology, University of Edinburgh, 37 Chalmers Street, Edinburgh

EH3 9EW

Summary Therapeutic abortion was induced in 100women in early pregnancy (less than 56

days’ amenorrhoea) with a combination of the antigestagenmifepristone (RU 486) and a synthetic prostaglandinanalogue, gemeprost. Mifepristone in oral doses of 400-600mg was followed 48 h later by a gemeprost vaginal pessary(0·5-1·0 mg). Bleeding was induced in all women 22-70 hafter the mifepristone dose and although bleeding continuedfor 4-43 days (median 12) the total measured blood-loss wasonly a median of 72·5 ml (range 15-398). Complete abortionoccurred in 95 women. Surgical evacuation of the uterus forminimum debris was required in the remaining 5. Only 10women had diarrhoea or pain that required opioid analgesia.The combination of mifepristone and gemeprost provides asafe and effective alternative to surgical evacuation of theuterus for therapeutic abortion in early pregnancy.

Introduction

IN England and Wales in 1986, 172 286 pregnancies wereterminated by therapeutic abortion.’ In the first 12 weeks ofpregnancy the uterus can be evacuated safely by vacuumaspiration, when the complication rate is related to the

length of gestation. However, surgical termination requiresskilled doctors and nurses and often general anaesthesia.

Before 8 weeks of amenorrhoea spontaneous abortionsare frequently complete. If therapeutic abortion with

prostaglandin analogues is induced in this stage of

pregnancy incomplete abortion is rare with over 90% of

women requiring no surgical intervention.3-5 However, theusefulness of these agents is limited by the high frequency ofgastrointestinal side-effects, which appears unaffected evenby controlled release of prostaglandin per vaginam.6

Mifepristone (RU 486, Roussel), a derivative of theprogestagen norethisterone, is a powerful antiprogesteroneand antiglucocorticoid, and mild antiandrogen. It binds

reversibly at the receptor and has a half-life of about 24 h.7When used to induce abortion, mifepristone has fewerside-effects than prostaglandin analogues but is less

effective, with between 60 and 85% of women abortingcompletely."-10 Preliminary reports suggest that the additionof a prostaglandin to treatment with mifepristone increasesthe efficacy of the antigestagen without major increases inmifepristone side-effects.11,12 We report here our experiencewith mifepristone in combination with a vaginal pessarycontaining gemeprost (May & Baker) in 100 women

presenting for termination of early pregnancy. Gemeprost isa synthetic prostaglandin E, analogue (16, 16-dimethyl-trans-&Dgr;2-PGE1 methylester).

Patients and Methods

Patients

Local family-planning services and general practitioners wereasked to refer women of less than 56 days’ amenorrhoea who hadrequested termination of pregnancy to the gynaecologicaloutpatient department of the Royal Infirmary of Edinburgh, whenit had been established that there were grounds for termination ofpregnancy under the 1967 Abortion Act (Scotland). 100 womenparticipated in this open study (table i). Pregnancy was confirmedby measurement of the serum level of human chorionic

gonadotropin ((3HCG). Gestational age was assessed by an accuratemenstrual history, clinical examination, and a pelvic ultrasoundscan. Women with evidence of multiple pregnancy or spontaneousabortion were excluded from the study as were those with a historyof serious medical disorder and those aged below 17. Local ethicalcommittee approval was granted for the study and written informedconsent was obtained from each participant. All women wereadmitted on the first day of treatment (day 1). No dietaryrestrictions were imposed.

TABLE 1-PATIE?v’TS’ CHARACTERISTICS

1416

Treatment

Group I (n = 20) received 150 mg mifepristone orally each day for4 days (17 of these patients have been reported briefly12). Groups II(n = 30), III (n - 30), and IV (n = 20) received a single oral dose ofmifepristone, 400 mg, 500 mg, or 600 mg, respectively, between0900 and 1200 h. In group I half a 1 mg gemeprost vaginal pessarywas inserted into the posterior fomix of the vagina 48 h after the firstdose of mifepristone; the other half was inserted 6 h later. The first 6women in group II received a whole 1 mg pessary 48 h after the

mifepristone dose. The remaining 74 women were given onehalf-pessary 48 h after mifepristone administration; the rest of thepessary was inserted 3 h later only if there are been no uterine pain orbleeding.

Prophylactic medication was not prescribed. Analgesic drugswere given as required: oral paracetamol or dihydrocodeine for mildto moderate pain, intramuscular pethidine (100 mg) or diamorphine(5 mg) for severe pain. Intramuscular cyclizine was given forvomiting.

Assessments

Samples of peripheral blood were collected at recruitment formeasurement of the concentration of haemoglobin, urea,

electrolytes, cortisol, and HCG and for liver function tests. Bloodwas also taken for oestradiol and progesterone assay from all women

except those in group I. Each woman recorded symptoms in a diaryfrom the day before treatment started. Some women were asked tocollect all soiled sanitary protection once treatment had started sothat blood-loss could be estimated.Women in group I had their temperature, pulse rate, and

blood-pressure measured 4 hourly for 12 h after drug adminis-tration. The first 10 women in group II had the same observations

hourly for 10 h. Because side-effects were few in these 10 women,these observations were made at 1 and 2 h after mifepristone in theremaining women.The women remained in hospital for at least 3 h after gemeprost

administration during which time hourly recordings of

temperature, pulse rate, and blood-pressure were made. Patientswere discharged once uterine pain and vaginal bleeding hadlessened.

-

Follow-up Visits

The women were reviewed 1, 2, and 4 weeks after treatment anddischarged from follow-up after the onset of the next menstrualperiod. At each visit uterine size and the cervix were assessed. Anendocervical swab was taken for culture from women who

complained of offensive vaginal discharge. At the first visit bloodwas collected for estimation of the concentration of haemoglobin,urea, electrolytes, cortisol, and HCG and for liver function tests.Blood was taken for oestradiol and progesterone assay in groupsII-IV. At the other visits blood was collected for estimation of HCGconcentration in groups II-IV, and urine was collected in group Ifor an immunological pregnancy test.13

Abortion was considered complete if by 4 weeks vaginal bleedinghad ceased, the cervix was closed, the uterus was normal sized, andeither the serum level of HCG was less than 200 mIU/ml or animmunological pregnancy test in urine was negative. 13 of thewomen in group I, 11 in group II, 28 in group III, and 20 in groupIV collected all their soiled sanitary protection.

Hormone Assays

Progesterone and oestradiol were measured byradioimmunoassay. 14,15 HCG was assayed with commercial

reagents (Serono), the lower limit of sensitivity was 3 mIU/mlplasma. Cortisol was measured by radioinirnunoassay.11

Measurement of Blood-loss

Blood-loss was calculated from used sanitary towels and tamponsby a modification of Hallberg and Nilsson’s methodP Soiledsanitary wear was placed in a molar solution of sodium hydroxidefor 48 h. A sample of liquid was filtered and its optical density

TABLE II-OUTCOME OF TREATMENT*

measured at 550 nm. Blood-loss was calculated by comparison withthe optical density of 1 ml of peripheral blood placed in molarsodium hydroxide for 24 h.

AnalysisThe U test, the t test for independent samples, and the X2 test

were used. Data are presented as median (range) or mean (standarderror).

Results

The effectiveness of the four treatment regimens wassimilar (table 11). Of the 74 women who received half agemeprost pessary, only 10 (14%) required the second half.There were no on-going pregnancies and 95 of the womenaborted completely. Only 5 women required surgicalintervention. In group I, a woman of 48 days’ amenorrhoeaunderwent curettage 6 weeks after treatment because of

persistent vaginal spotting and an ultrasound scan

suggestive of minimum uterine debris. Her HCG levels haddecreased and an immunological pregnancy test in urine wasnegative. Histological examination of the scant curettingsrevealed necrotic trophoblastic tissue but no fetal parts. InGroup II, a woman of 50 days’ amenorrhoea was taken totheatre because of brisk vaginal bleeding 5 h after theinsertion of half a gemeprost pessary. Under anaesthesia the

products of conception were removed from the cervical oswith a rapid decrease in blood-loss. She was not shocked anddid not require transfusion. 3 women in group III requiredcurettage. 1 woman of 53 days’ amenorrhoea underwentsurgical evacuation of the uterus 5 weeks after initial

treatment; the rest of her history was identical to that of thewoman in group I. Another woman of 50 days’ amenorrhoeaunderwent curettage 17 days after treatment because herHCG levelled out despite a large initial decrease. Anultrasound scan confirmed a small area of retained productsof conception. The third woman in this group whounderwent curettage was 48 days from her last menstrualperiod when treated. Her HCG level had fallen

appropriately for 2 weeks but at her final visit was found tohave increased. Ultrasound revealed minimum uterinedebris. Histological examination of the curettings from these3 patients revealed trophoblastic tissue but no fetal parts.

Because there was no significant difference between thefour groups in the onset of bleeding and pain, requirementfor analgesia, side-effects, duration of bleeding, measuredblood-loss, the time until the next menstrual period, and all

TABLE III-ONSET OF BLEEDING AND BLOOD-LOSS

*Before vs after, p = 0 50 (U test).

1417

TABLE IV—XUMBER OF PATIEXTS WITH GASTROIXTESTIXALSIDE-EFFECTS

Significant difference between the days&mdash;*=2221, p<0001 and

tx2 = 8-00, p < 0-05.

TABLE V&mdash;HORMONE CONCEXTRATIOXS BEFORE AXD A-TER

TREATMENT (n = 80)

Significant difference compared with before treatment&mdash;*p<001 andt<0001.

laboratory results, data were pooled for analysis. Bleedingwas induced in all 100 women (table III). Bleeding started36 h after mifepristone administration and continued for 12days; the median measured blood-loss of 72 women was 72 5ml. The haemoglobin concentration was similar before andafter treatment. No woman required blood transfusion.The 94 women who had pain were first aware of pelvic

discomfort 46-6 h (19 5-70) after the start of treatment. Theproducts of conception were identified during the treatmentof 57 women at 4-2 h (2-3-9-2) after insertion of the

gemeprost pessary. No patient required analgesia during thefirst 48 h of treatment. After insertion of the pessary 44women received an oral analgesic drug and 9 womenreceived an intramuscular opioid. 47 women required noanalgesia.There was no significant difference in the frequency of

nausea before and during treatment (table iv). There was,however, a significant increase in the incidence of vomitingand of diarrhoea. 30 women vomited after the pessary wasinserted compared with 13 the day before treatment, and 10women had diarrhoea compared with 3 before treatment. 40women had pelvic cramps for 2 (1-5) days after treatment.No women had clinical evidence of pelvic infection. 7women received antibiotics when pathogenic organisms(p-haemolytic streptococcus, Trichomonas vaginalis,gonococcus, Bacteraides) were isolated on culture of anendocervical swab. The interval until the next menstrual

period was 31 (14-48) days from treatment.Liver function tests and cortisol levels were similar before

and after treatment. Levels of HCG decreased significantlyover the week after treatment although HCG was stilldetectable in plasma 2 and 4 weeks after treatment (table v).Oestradiol and progesterone also decreased significantlyafter treatment. There was no significant difference in theresults between women who required evacuation and thosewho did not.

Discussion

Previous studies of mifepristone, when used alone toterminate early pregnancy, have shown an unacceptablyhigh rate of incomplete abortion and ongoing pregnancy. 89Although gemeprost effectively induces early abortion, thedoses required are associated with a high incidence ofside-effects such as pain that requires opioid analgesia and

vomiting and diarrhoea: 18 This study has confirmed ourprelirninan- report12 that the addition of a small dose ofprostaglandin to treatment with mifepristone increases thefrequency of complete abortion. In the present study 95women aborted completely after treatment with a

combination of mifepristone and gemeprost. There were noongoing pregnancies after treatment. Although 5 womenunderwent curettage, in retrospect the first 3 probably didnot require surgical intervention. As we gained experiencewe became more confident in disregarding minimumcontinued vaginal bleeding.No serious complications were encountered during

treatment. Heavy bleeding, attributed to insufficient

endogenous prostaglandin production, during treatmentwith mifepristone alone has been reported.8.9 However,none of the women in our study required transfusion andhaemoglobin concentration did not decrease after treatment.The measured blood-loss was similar to that expectedduring a heavy period" and to that after vacuum aspirationand medical termination with gemeprost alone, although theduration of bleeding was found to be slightly longer thanwith these methods.4.l8The occurrence of pain that required opioid analgesia was

low compared with that when prostaglandin analogues areused alone.4.20 Only 9 of our patients required intramuscularpethidine or diamorphine and almost half the patientsneeded no analgesic drug at all. Nausea occurred in morethan half the patients before and during treatment.

Although 30 women vomited after administration of

gemepost, 13 had done so the day before treatment. Nauseaand vomiting are common symptoms in early pregnancy,and diarrhoea, a more specifically prostaglandin-relatedside-effect, occurred in only 10 women. That the occurrenceof gastrointestinal side-effects was less frequent than duringtreatment with gemeprost alone is a reflection of the smallerdose of prostaglandin.There is evidence that mifepristone sensitises the uterus

to the oxytocic action of prostaglandins by releasingendogenous prostaglandins from the decidua. 1121 Clinicallythis would appear to be the case with mild pelvic painstarting before exogenous prostaglandin administration anda high abortion rate after an otherwise subtherapeutic doseof gemeprost. This combination of mifepristone andgemeprost exploits the advantages of both compounds: alow frequency of side-effects and a high frequency ofcomplete abortion.

There was no difference in the efficacy of the fourtreatment regimens and the dose of prostaglandin requiredto achieve complete abortion was not greater when the doseof mifepristone was lower. Kovacs et all studied 37 women,of less than 42 days’ amenorrhoea, after 25, 50, or 100 mgmifepristone twice daily for 4 days, and found that theoccurrence of complete abortion was not dose-related.Couzinet et al10 treated 100 women within 10 days of theirmissed period and found no difference in the effectiveness of400, 600, or 800 mg mifepristone. Our study confirms thesefindings and also demonstrates the effectiveness of single asopposed to sequential administration of mifepristone.The comparison of medical termination of pregnancy

with vacuum aspiration is not straightforward. The 5%incidence of incomplete abortion after treatment with

mifepristone and gemeprost must be weighed against theneed for a skilled operator to do vacuum aspiration and,arguably, the need for and therefore risk of a generalanaesthetic.4.22 The occurrence of incomplete abortion aftermedical termination of pregnancy, however, makes careful

1418

follow-up a necessity. Undoubtedly experience leads to alower rate of surgical intervention, as we found during ourstudy.

Acceptability to the patient is important if medicaltermination of early pregnancy is to be used as an alternativeto surgical abortion. The acceptability of medicaltermination has been assessed.23 Women who found this

type of abortion unacceptable were usually those who hadhad pain and other side-effects. The use of mifepristone incombination with gemeprost is an effective and safe meansof medical induction of therapeutic abortion. The lowfrequency of severe pain and gastrointestinal side-effectssuggest that this is an acceptable means of termination andcould provide an alternative to surgical abortion. Thecombination would have particular application in thosecountries where skilled medical and surgical experience arein short supply.The help of Sister A. Michie, Sister H. Hillier, and Sister D. Will in the

management of patients is appreciated. We thank Margaret Harper for typingthe manuscript and Dr Ian Roberts of Roussel for supplying mifepristone.We also thank local general practitioners and Brook Advisory and FamilyPlanning Centres for their co-operation.

Correspondence should be addressed to D. T. B.

REFERENCES

1. Office of Population Censuses and Surveys. Abortion Statistics. London: HM

Stationery Office, 1986: 1.2. Grimes DA, Cates W. Complications from legally induced abortions: a review. Obstet

Gynecol Survey 1979; 34: 177-91.3. Bygdeman M, Breune K, Christensen N, et al. A comparison of two stable

prostaglandin E analogues for termination of early pregnancy and for cervicaldilatation Contraception 1980; 22: 471-483.

4. Smith SK, Baird DT. The use of 16,16-dimethyl-trans-&Dgr;2-PGE1 methyl ester (ONO802) vaginal suppositories for the termination of early pregnancy: a comparativestudy. Br J Obstet Gynaecol 1980; 87: 712-17.

5. Csapo AI, Peskin EG, Pulkkinen M, et al. "Menstrual induction" with sulprostone.Prostaglandins 1982; 24: 657-65.

6. Cameron IT, Baird DT. A controlled release form of 16,16-dimethyl-trans-&Dgr;2-PGE1methyl ester for early abortion. Contraception 1986; 33: 121-25

7. Swahn ML, Cekan S, Wang G, et al Pharmacokinetic and clinical studies of RU 486and human fertility regulation. In: Baulieu EE, Segal SJ, eds. The antiprogestinsteroid RU 486 and human fertility control. New York: Plenum Press, 1985:249-58.

8. Herrmann W, Wyss R, Riondel A, et al. The effect of an antiprogesterone steroid inwomen: Interruption of the menstrual cycle and of early pregnancy. C R Acad SciParis 1982, 294: 933-38.

9. Kovacs L, Sas M, Resch BA, et al. Termination of very early pregnancy by RU 486: anantiprogestational compound Contraception 1984; 29: 399-410.

10. Couzinet B, Strat NL, Ulmann A, Baulieu EE, Schaison G. Termination of earlypregnancy by the progesterone antagonist RU 486 (mifepristone). N Engl J Med1986; 315: 1565-70.

11. Bygdeman M, Swahn M-L. Progesterone receptor blockage: Effect on uterine

contractility and early pregnancy. Contraception 1985; 32: 45-51.12. Cameron IT, Michie AF, Baird DT. Therapeutic abortion in early pregnancy with

antigestagen RU 486 alone or in combination with prostaglandin analogue(gemeprost). Contraception 1986; 35: 459-67.

13. Hobson BM. Pregnancy diagnosis using Pregnosticon haemaglutination inhibitiontest. Br J Obstet Gynaecol 1968; 75: 718-23.

14. Scaramuzzi RJ, Corker CS, Young G, Baird DT. Production of antisera to steroidhormones in sheep. In Cameron EHD, Hillier SG, Gnffiths K, eds Steroid

immunoassay. 5th Tenovus Workshop. Cardiff: Alpha Omega Alpha Publishing,1975:111-22.

15. Backstrom CT, McNeilly AS, Leask RL, Baird DT. Pulsatile secretion of LH, FSH,prolactin and oestradiol during the human menstrual cycle. Clin Endocrinol 1982;17: 29-42.

16. Gray SM, Seth J, Beckett CJ. Comparison of separation methods m the I125

radioimmunoassay of plasma cortisol. Ann Clin Biochem 1983; 20: 312-17.17. Hallberg L, Nilsson L Determination of menstrual blood loss. Scand J Clin Lab Invest

1964, 16: 244-48.18. Cameron IT, Baird DT. Early pregnancy termination: a comparison between vacuum

aspiration and medical abortion using prostaglandin (16,16,-dimethyl-trans-&Dgr;2-PGE1 methyl ester) or the antigestagen RU 486. Br J Obstet Gynaecol (in press).

19. Hallberg L, Hogdahl AM, Nilsson L, Rybo G. Menstrual blood loss: a populationstudy. Acta Obstet Gynecol Scand 1966, 45: 320-51.

20 Mackenzie IZ, Embury MP, Davies AJ, Guillebaud J. Very early abortion byprostaglandins. Lancet 1978; i: 1223-26.

21. Kelly RW, Healy DL, Cameron MJ, et al. The stimulation of prostaglandinproduction by two antiprogesterone steroids m human endoemetnal cells. J ClinEndocrinol Metab 1986; 62:1116-23.

22. Goldthorp WO. Ten minute abortions. Br Med J 1977; ii: 562-64.23. Rosen AS, Nystedt L, Bygdeman M, Lindstrom V. Acceptability of a non-surgical

method to terminate very early pregnancy m comparison to vacuum aspiration.Contraception 1979; 19: 107-17.

CONTROLLED TRIAL OF PREDNISOLONE AS

ADJUVANT IN TREATMENT OFTUBERCULOUS CONSTRICTIVEPERICARDITIS IN TRANSKEI

J. I. G. STRANGD. G. GIBSONA. J. NUNN

H. H. S. KAKAZAD. J. GIRLING

W.Fox

Umtata Hospital, Transkei, Southern Africa; Department ofCardiology, Brompton Hospital, London; and MRC Cardiothoracic

Epidemiology Group, Brompton Hospital

Summary In Transkei, 143 patients with activetuberculous constrictive pericarditis

without significant pericardial effusion all received the samedaily 6-month antituberculosis regimen of streptomycin,isoniazid, rifampicin, and pyrazinamide for 14 weeksfollowed by isoniazid and rifampicin. They were randomlyallocated to receive in addition either prednisolone orplacebo for the first 11 weeks; the comparison wasdouble-blind throughout treatment and follow-up. In the114 patients assessable up to 24 months, improvement wassignificantly more rapid in the prednisolone group, as shownby the rate of fall in the mean pulse rate and the rate at whichjugular venous pressure and level of physical activity becamenormal. During follow-up, 2 (4%) of the 53 prednisoloneand 7 (11%) of the 61 placebo patients died from

pericarditis, and 11 (21%) and 18 (30%), respectively,required pericardiectomy. By 24 months 50 (94%)prednisolone and 52 (85%) placebo patients had a

favourable status. 3 patients (1 prednisolone, 2 placebo)were normally active but were classified as not havingachieved a favourable status. It is recommended that, in theabsence of a specific contraindication, antituberculosis

chemotherapy should be initially supplemented by steroids.

Introduction

BEFORE the advent of antituberculosis chemotherapy,tuberculous pericarditis was usually fatal, either in the acutestage or as a result of subsequent constriction. Effectivechemotherapy and, if indicated, pericardiectomy forconstriction greatly improves the prognosis, but there is stillappreciable mortality.2--6 In Transkei and the surroundingregions, tuberculous pericarditis is one of the commonestcauses of the clinical picture of "congestive heart failure",’ /and has indeed been referred to as "Transkei heart".8Non-randomised comparisons, in which the clinician incharge decided whether or not to prescribe corticosteroids.inaddition to antituberculosis chemotherapy, suggested thatwhile steroids may have a role in the treatment oftuberculous pericardial effusion, they do not reduce the riskof constriction or its progression. 5,6,9 We have investigatedthe usefulness of steroids as an adjuvant in the treatment ofactive tuberculous constrictive pericarditis in a randomisedplacebo-controlled double-blind trial with a 2-year follow-up.

Patients and MethodsPatients

Patients from throughout Transkei (which has a population ofabout 3 million) aged 5 years or more with active tuberculousconstrictive pericarditis were eligible for study if they had receivedno previous antituberculosis chemotherapy, or such therapy for lessthan 2 weeks during the previous year. The diagnostic criteria"included symptoms of reduced physical activity and breathlessness,and the finding of an increased jugular venous pressure (JVP),