Children sick enough to be admitted to hospital are often still recovering from infection when discharged and may not be fit for immunisation. Indeed, children may be discharged less than 30 hours after fever subsides, which would preclude immunisation by the criteria of Dr Riley and colleagues. As already occurs in Oxford, and as suggested by Dr Reeve and colleagues, suitably trained health visitors could catch up missed immunisations. The single most important finding in our study was that once parents have decided against a particular vaccine, often following inappropriate advice, it is difficult to persuade them otherwise. Attention to the clear and concise recommendations in the "green book" should ensure that health professionals give appropriate advice from the outset.4 S R M DOBSON A DAVIES S LEDGER
E BIXBY R BOOY
Margaret Pyke Centre, London WIV 5srw
I Riley 1)j, MN1ughal MZ, Roland J. Immunisation state of young children admitted to hospital and effectiveness of a ward based opportunistic immunisation policy. BM7 1991;302:31-3. (5 Januarv.) 2 Reeve H, Chew C, Waller-Wilkinson I. Effectiveness of ward based opportunistic immunisation policy. BMJf 1991;302:412.
(16 February.) 3 Department of Health and Social Security. Immunisation against itzJectious diseases. London: HMSO, 1984. 4 Department of Health. Immunisation against infectious diseases. London: HMSO, 1990.
Waiting lists SIR,-Mr John Appleby's article on John Yates's resignation gave the impression that the College of Health has received Department of Health funding for the Guide to Hospital Waiting Lists.' The college has published this guide for the past six years to help patients and general practitioners locate shorter waiting lists around the country and, in fact, has produced it from its own modest charitable resources. What the department has funded us to do is to set up a computerised waiting list information service for general practitioners and patients in the current year. It is this service that it has said it will probably not fund after 1992. MARIANNE RIGGE
College of Health, London E2 9PL I Appleby J. \Waiting list advisers resign. BMJ 1991;302:432. 23 Februar. )
Stopping the pill SIR,-Dr Ruth Booker' is right to draw attention to the practical problems implied by our article,2 which supports established teaching that the combined (oestrogen containing) pill should be stopped at least four weeks before any surgery with an intrinsic risk of deep venous thrombosis. One answer is to press for the more humane management of long waiting lists. Except in a crisis it is unfair to give only a few days' notice of admission. Many outpatient departments already show, often by keeping an admission diary, the feasibility of giving at least a month's notice for most inpatient surgery, including (in this context) relevant major operations and all surgery to the legs in women. Even more important, in the prevention of unwanted pregnancies on the waiting list, is to remember that the combined pill is not the only really effective reversible contraceptive. In 1985, and again in 1988, one of us recommended a method that is even more effective than the combined oral contraceptive-namely, an injectVOLUME 302
JOHN GUILLEBAUD GILLIAN E ROBINSON
Department of I'aediatrics, University of Oxford, John Radcliffe Hospital, Oxford OX3 9DU
able contraceptive (for example, Depo-Provera).4 One or more injections, each lasting three months, will cover the time on any waiting list. The combined oral contraceptive may be resumed two or more weeks after operation without regard to the time of the last injection. Progestogen-only pills also do not increase the risk of deep venous thrombosis or pulmonary embolism, and can have a failure rate of less than one pregnancy per 100 women years even in those under the age of 35.' But for women whose compliance with the very strict regimen of the progestogen-only pill is in doubt (and for whom an intrauterine device is unsuitable) the injection remains an even better option-provided always there is good counselling backed by a leaflet explaining the difficulty of reversal if side effects occur.
30 MARCH 1991
1 Booker R. Stopping the pill. BMJ7 1991;302:535. (2 March.) 2 Robinson GE, Burren T, Mackie IJ, et al. Changes in haemostasis after stopping the combined oral contraceptive pill: implications for major surgery. BM_J 1991;302:269-71. (2 February.) 3 Guillebaud J. Surgery and the pill. BMJ 1985;291:498-9. 4 Guillebaud J. Should the pill be stopped pre-operatively? BMJ7 1988;296:786-7. 5 Bisset AM, Dingwall-Fordyce I, Hamilton MIK. The efficacy of the progestogen-only pill as a contraceptive method. British J7ournal of Familv Planning 1990;16:88-93.
Unintended pregnancies and contraceptive use SIR,-Ms Anne Fleissig comments that the combined contraceptive pill has a lowered efficacy in some women after a gastrointestinal upset or while they are taking antibiotics,' but the reference she gives contains no data supporting this suggestion.2
The evidence in support of lowered efficacy is largely anecdotal. In patients who have gastrointestinal upset vomiting within one to two hours after tablet ingestion may be associated with contraceptive failure. Both components of the pill, however, are absorbed rapidly and completely,' and we have shown that in patients with an ileostomy the absorption of contraceptive steroids is complete.4 Diarrhoea would need to be very severe to result in reduced absorption. The evidence for an interaction of broad spectrum antibiotics with oral contraceptives is based on a few individual case reports, but all the systematic scientific studies have failed to show any evidence of this interaction. In our review of the Committee on Safety of Medicines data there were 63 pregnancies reported over 16 years.' Even allowing for underreporting of adverse reactions this is still a small number of pregnancies in an estimated 2 to 2 5 million women taking oral contraceptives each year. In addition, some of the reported failures of contraception implicated cotrimoxazole, which is known to enhance the efficacy of oral contraceptives.6 Our group and others have failed to show any systematic interaction with antibiotics such as ampicillin, tetracycline, and erythromycin.7 We recently completed two studies with temafloxacin7 and clarithromycin' and again found no evidence of reduced contraceptive efficacy. If anything, clarithromycin enhanced the efficacy of the coadministered oral contraceptive. If an interaction exists between antibiotics and oral contraceptives it is likely to be extremely rare. Thus, to give the impression that the pill has reduced efficacy in women (in general) taking antibiotics is misleading. The decision of the pharmaceutical industry to issue a warning about potential failure of oral contraceptives is understandable given the considerable media interest arising from Ms Fleissig's paper, but the decision has been taken on medico-
legal rather than scientific grounds. The use of references in scientific papers that have no academic basis and merely perpetuate inaccuracies must be avoided. M ORME D J BACK University of Liverpool, Liverpool L69 3BX 1 Fleissig A. Unintended pregnancies and the use of contraception: changes from 1984 to 1989. BMJ7 1991;302:147. (19 January.) 2 Griffiths M. Contraceptive practices and contraceptive failures among women requesting termination of pregnancy. British Journal ofFamily Planning 1990;16:16-8. 3 Orme M, Back DJ, Breckenridge AM. Clinical pharmacology of oral contraceptive steroids. Clinical Pharmacokinetics 1983;8: 95-136. 4 Grimmer SFM, Back DJ, Orme MLE, Cowie A, Gilmore 1, Tjia J. The bioavailability of ethinyloestradiol and levonorgestrel in patients with an ileostomy. Contraception 1986;33: 51-9. 5 Back DJ, Grimmer SFM, Orme MLE, Proudlove C, Mann RD, Breckenridge AM. Evaluation of Committee on Safety of Medicines yellow card reports on oral contraceptive drug interactions with anticonvulsants and antibiotics. B J Clin Pharmacol 1988;25:527-32. 6 Grimmer SFM, Allen WL, Back DJ, Breckenridge AM, Orme MLE, Tjia JF. The effect of cotrimoxazole on oral contraceptive steroids in women. Contraception 1983;28:53-9. 7 Back DJ, Orme M. Pharmacokinetic drug interactions with oral contraceptives. Clinical Pharmacokinetics 1990;18:472-84. 8 Back DJ, Tjia J, Martin C, Millar E, Salmon P, Orme M. The interaction between clarithromycin and oral contraceptive steroids. ournal ofPharmaceutical Medicine (in press). 9 Back DJ, Tjia J, Martin C et al. The lack of interaction between temafloxacin and combined oral contraceptive steroids. Contraception (in press).
AUTHOR'S REPLY,-In commenting on the recent increase in unintended pregnancies, which seemed to be associated with use of the contraceptive pill, I referred to Griffiths's paper' primarily to illustrate the importance of giving women enough information to make contraception effective. If interference from antibiotics or gastrointestinal upset is rare as Professor Orme and Dr Back suggest this detail is clearly less important than published recommendations to general practitioners and family planning doctors say.2 I am glad they have raised this pharmacological issue, which I was not in a position to do. ANNE FLEISSIG
Institute for Social Studies in Medical Care, London NW3 2SB I Griffiths M. Contraceptive practices and contraceptive failures among women requesting termination of pregnancy. British
Journal ofFamily Planning 1990;16:16-8. 2 Department of Health, Scottish Home and Health Department, and Welsh Office. Handbook of contraceptive practice (1990 edition). London: DoH, 1990.
SIR,-The incidence of unintended pregnancies was recently reported by Ms Anne Fleissig as increasing from 27% to 31%.' She recognised the difficulty of postal contact with women who had had a live birth six months earlier (76% response rate). It was therefore impossible to calculate the total number of unintended pregnancies as she had no information about non-responders, women who miscarried, and those who had had a termination. Some idea of the total number of pregnancies that are unplanned may be inferred from a prospective survey of pregnancies done in the town of Alton during 1989 as part of a two year cohort study in a population of 21000. The diagnosis of pregnancy generally occurred at about the sixth week, and in 1989 there were 312 new pregnancies seen by 13 general practitioners. The outcome was ascertained in 303 (97%). The pregnancy had been planned by only 60% of the women at the time of conception. We divided the unplanned ones into wanted (22%), uncertain (5%), and unwanted (13%) pregnancies. Of the 41 women with unwanted pregnancies, 30 had a termination carried out. There were 238 live births, and at the first postnatal visit the midwife noted whether the pregnancy was said to have been planned or not.