Hysterectomy: will it pay the bills in 2007?
BMJ 1997; 314 doi: https://doi.org/10.1136/bmj.314.7075.160 (Published 18 January 1997) Cite this as: BMJ 1997;314:160Treatment of choice for cancer, but a choice of treatment for menorrhagia
- R J Lilford, Professor of health services researcha
- a Department of Public Health, Faculty of Medicine, The University of Birmingham, Edgbaston, Birmingham B15 2TT
To study the indications for hysterectomy is to study the interface between medicine and society. In California barely half of all women will carry their uterus to the grave,1 whereas a gynaecologist in Saudi Arabia may do no more than one hysterectomy a year and, as often as not, this will be a lifesaving operation for catastrophic obstetric haemorrhage. In Britain hysterectomy rates are somewhere between these two extremes. To understand the variations, do not gaze endlessly at histological specimens but examine the societies from which they originate.
Perceived abnormal bleeding accounts for 70% of hysterectomies in pre-menopausal British women, and in most cases of “menorrhagia,” menstrual blood loss is within the “normal” range.2 Much of the variation in hysterectomy rates is therefore attributable to the psychosocial factors that influence demand.3
Provider factors are also important. Women general practitioners are less likely than their male counterparts to refer women with menstrual symptoms for a specialist opinion;4 and hysterectomy, like cholecystectomy and tonsillectomy, varies considerably in frequency from surgeon to surgeon.5 This is not to say that gynaecologists exploit women for personal gain or take some sort of covert delight in the procedure; doctors' wives, after all, undergo hysterectomy as often as controls matched for social class.6 Yet hysterectomy rates are not only variable but labile: a public education campaign in the Italian speaking cantons of Switzerland resulted in a sharp fall in hysterectomy rates compared with the control cantons (French and German speaking).7
It is tempting to conclude that a lot of unnecessary surgery is going on and that we should campaign against it. But what are the effects of hysterectomy? Although in the short term women who have had a hysterectomy score worse than non-surgical controls on measures of wellbeing, if the same women are followed prospectively their scores improve.8 An authoritative study failed to confirm fears that hysterectomy increases the risk of urinary incontinence.9 And while hysterectomy is a major operation with serious morbidity and, very rarely, mortality,10 it may also save lives by reducing the risk of uterine cancer. The relative risks of operative mortality and death from cancer are such that hysterectomy is the safer option.11 Risk of cancer is reduced still further by removing the ovaries, although this increases the risk of ischaemic heart disease unless hormone replacement therapy is taken. However, oestrogen replacement does not correct the loss of libido that follows oophorectomy, perhaps as a result of androgen deficiency.8 Preliminary evidence that hysterectomy may predispose to ovarian failure, even when the ovaries are conserved, is a subject of current research.
How should hysterectomy be performed? Vaginal hysterectomy is associated with fewer complications than abdominal hysterectomy.12 However, this has not been confirmed by a recent systematic review,13 and many surgeons feel uncomfortable with the vaginal route in the absence of prolapse. Laparoscopically assisted vaginal hysterectomy is a new technique which is currently under evaluation in a large trial and a nationwide observational study, both funded through the NHS R and D programme.
Should the cervix be removed during abdominal or laparoscopic hysterectomy? Surgical morbidity is often the result of removal of the cervix, which lies close to the ureters and bladder and from which it must be carefully dissected with a consequent risk of vault haematoma or urological injury.14 This, along with a suspicion that the presence of the cervix may enhance orgasm, has led to calls for “sub-total” hysterectomy. Removal of the cervix was strongly advocated in the first half of the century, in part because of the risk of cancer. However, the residual risk of this disease in women with a history of regular negative cervical smears is sufficiently low to be traded off against lower complication rates.14
What about other surgical alternatives? Various techniques to remove the endometrium while leaving the remainder of the uterus in situ have been investigated.15 Although these procedures result in lower morbidity and shorter hospital stay than hysterectomy, up to 30% of patients will eventually lose their uterus. Endometrial surgery provides additional choice, but overall rates of surgical treatment have not declined and may have risen.16
About one woman in seven will decline hysterectomy if she can be shown to have blood loss within the normal range.17 For these and other women, medical treatments are appropriate. However, referral is almost always followed by surgery,18 and medical treatments may merely delay a surgical “solution.” The new progesterone-coated intrauterine device proposes to revolutionise non-surgical management. Recently licensed in this country, the levonorgestrel-bearing device substantially reduces menstrual flow (and seems to be much more effective than oral medical treatment) while also providing effective contraception and reducing the risk of sexually transmitted diseases.19 Whether this will remove hysterectomy from its pre-eminent place in the repertoire of gynaecological treatments remains to be seen. Although the first hysterectomy was carried out in 1822, 20 it has become a mainstay of gynaecological practice. It is quite probable that the operation has “peaked” and will now decline in incidence. There is no “correct” hysterectomy rate, but “correct” practice is to make explicit the trade offs between this operation and an increasing number of alternatives.
I thank Professors Angela Coulter, Klim McPherson, James Drife, and Nicholas Johnson for help with this article.
Footnotes
I thank Professors Angela Coulter, Klim McPherson, James Drife, and Nicholas Johnson for help with this article.