Has endometrial ablation replaced hysterectomy for the treatment of dysfunctional uterine bleeding? National figures
Correspondence: Dr S. A. Bridgman, Department of Epidemiology, Keele University School of Postgraduate Medicine, Thornbur-row Drive, Hartshill, Stoke-on-Trent, Staffordshire ST4 7QB, UK.
Objectives To describe trends in the use of endometrial ablation and hysterectomy for the treatment of dysfunctional uterine bleeding.
Design Analysis of hospital admissions data.
Setting National Health Service Hospitals in England.
Population Women who underwent a hysterectomy or endometrial ablation for dysfunctional uterine bleeding between 1989 and 1996.
Main outcome measures Annual operation rates and standardised operation ratios for England and for the National Health Service Regions within it, and proportion of operations for dysfunctional uterine bleeding that were endometrial ablations or hysterectomies.
Results There was an initial rise in operation rates for endometrial ablation until 1992/3, since when the rates have fallen. Hysterectomy rates have remained relatively steady since the introduction of endometrial ablation. The total operation rates for dysfunctional uterine bleeding initially increased but have tended to fall since 1992/3. The ratio of hysterectomy to endometrial ablation for dysfunctional uterine bleeding troughed at 3:1 in 1992/3, but by 1995/6 had increased to 4:1.
Conclusions Rather than replacing hysterectomy in the treatment of dysfunctional uterine bleeding, endometrial ablation appears to have added an alternative operative technique. This led to an increase in the total number of operations for this condition, perhaps by lowering the threshold for intervention.
The introduction of endometrial ablation for the treatment of dysfunctional uterine bleeding in the United Kingdom in May 1988 was heralded with great enthusiasm by its proponents1. The procedure is associated with lower morbidity and costs compared with hysterectomy, while it preserves the uterus2. Consequently Magos1, among others2, predicted that techniques aimed at ablating the endometrium would probably replace hysterectomy for the treatment of menorrhagia, while Jourdain et al.3 asserted that endometrial ablation ‘significantly reduces the number of hysterectomies’.
Analysis of 1988 to 1993 routine hospital episode statistics from the Mersey Region of the English National Health Service suggested that endometrial ablations were not replacing hysterectomy, since endometrial ablation rates rose rapidly but there was little change in hysterectomy rates4,5. This paper aims to describe trends in the use of endometrial ablation and hysterectomy in the treatment of dysfunctional uterine bleeding in England from 1989 to 1996.
Data were obtained from the English Department of Health's hospital episode statistics database. This database holds a standard dataset6 for English residents treated in National Health Service hospitals since the 1987–8 financial year. The first year's data was not considered as complete as in later years. In 1988–9 operation codes were coded by Version 3 of the Office of Population Censuses and Surveys Classification of Surgical Operations7, which makes them difficult to compare with later years which were coded using version 48. This study therefore considers the financial years 1989–90 to 1995–6, with the years running from 1 April to 31 March.
Each episode of care has a code for the primary diagnosis (ICD-99 for 1989–90 to 1994–5 and ICD-1010 for 1995–6). There is no single diagnostic code for dysfunctional uterine bleeding. Therefore it was defined in ICD-9 as 626–2 (other dysfunctional uterine bleeding or functional uterine haemorrhage not otherwise specified), 626.8 (other dysfunctional uterine bleeding or functional uterine haemorrhage not otherwise specified) or 627.0 (premenopausal menorrhagia). Dysfunctional uterine bleeding was defined in ICD-10 as N92 (excessive, frequent and irregular menstruation), N93.8 (other specified abnormal uterine and vaginal bleeding—dysfunctional uterine or vaginal bleeding), or N93.9 (abnormal uterine and vaginal bleeding, unspecified).
Each episode of care also has a primary operation code8. The operation codes for hysterectomy are Q07 (abdominal excision of uterus) or Q08 (vaginal excision of uterus). Endometrial ablation was assumed to be coded as Q17 (therapeutic endoscopic operation on uterus) in episodes where the diagnostic code was for dysfunctional uterine bleeding. Episodes with both a primary diagnostic code and operation code defined above were selected from the Hospital Episode Statistics database.
Age-standardised rates of hysterectomy, endometrial ablation and hysterectomy plus endometrial ablation for dysfunctional uterine bleeding in women aged 25–54 years were calculated for England using five year age bands. The 1991/92 England population estimate was chosen as standard as this is the estimate closest to the 1991 census. Indirectly age-standardised operation ratios for hysterectomy, endometrial ablation and hysterectomy plus endometrial ablation for dysfunctional uterine bleeding were calculated for 1989/90 to 1995/96 for England and the National Health Service Regions using the 1991/2 operation rates for England as the standard. National and Regional population estimates used in the computation of rates were obtained from the Office for National Statistics11. Confidence intervals for standardised operation ratios were calculated by the method of Gardner and Altman12.
Following the introduction of endometrial ablation in 1988 there was an increase in the operation rates for dysfunctional uterine bleeding until 1992–3 (Table 1). During this period, the rates of hysterectomy dipped shortly after the introduction of endometrial ablation, before stabilising at a slightly lower rate than that prior to the introduction of the new technique, and then dipped again in the final year of observation. In contrast, the rates of endometrial ablation increased rapidly from its introduction until 1992–3, after which they have tended to fall (Table 1). The total operation rates for dysfunctional uterine bleeding initially increased to peak in 1992/3, since when they have tended to fall (Table 1).
Table 1. Age-standardised rates. per 1000 women aged 25–54 years, of hysterectomy, endometrial ablation, and hysterectomy plus endometrial ablation, for dysfunctional uterine bleeding (DUB) and ysterectomy:endometrial ablation ratios in National Health Service hospitals in England in 1989/1990 to 1995/1996. The standard population estimate is England 1991/1992.
Initially the ratio of hysterectomy to endometrial ablation for dysfunctional uterine bleeding fell rapidly, reaching a trough of 3:1 in 1992–3. Since then, the ratio has increased to 4:1 (Table 1).
The Regional analyses are not presented in detail here, except that of Oxford Region where endometrial ablation was first introduced into the UK. Oxford had the highest standardised operation ratio in England throughout the study period (Table 2). It also reached its peak rate earlier than in England as a whole. The hysterectomy rate in Oxford on average tended to be little different than that for England as a whole, although there were some considerable year to year fluctuations (Table 2).
Table 2. Indirectly standardised operation ratios (SORs) and 95% confidence intervals for dysfunctional uterine bleeding (DUB) for the Oxford Region of the English National Health Service, using the age-specific (five year age band) operation rates for England in 1991/1992 as standard.
The predicted substantial drop in hysterectomy rates after the introduction of endometrial ablation was not observed1, although there was a small early drop in rates. Overall, the new procedure was initially accompanied by increased total operation rates for dysfunctional uterine bleeding. This suggests a lower threshold for surgical intervention. After a peak in 1992–3, the rate of endometrial ablation has declined slightly.
There are several validity issues. The change in the diagnostic coding from ICD-99 Up to 1994–5 to ICD-1010 in 1995–6 could theoretically have had an effect on the recording of hospital episodes. The data only include operations carried out in the National Health Service, and whether there was any shift of work to and from the private sector is unknown. As patients can wait 18 months or more for their operations, some fluctuation of numbers between years is possible depending on time between the prescription of an operation and its implementation.
As numbers represent episodes of care and not individual women, they may include re-operations. Post endometrial ablation surgery rates (re-ablation and hysterectomy) have been published and range from 18% in the first year13 to 22% in the first three years14. Thus it is possible that the decline in the number of women receiving endometrial ablations since 1992–3 is steeper than first appears from the data.
It is likely that the rise in the number of endometrial ablations is accounted for by initial enthusiasm, as would be expected for a new procedure with high expectations. However, there are a number of possible reasons for the subsequent decline. One explanation is the realisation that the results were not as good as was first thought.
Another possible explanation for the subsequent decline in endometrial ablation rates is the replacement of endometrial ablation with Mirena which is an intrauterine progestogen contraceptive device, introduced in the United Kingdom in May 1995, and has a side-effect of a reduction of menses15. It is licensed for menorrhagia as an alternative to hysterectomy in several European countries16, although not yet licensed for this use in the UK. Lahteenmaki et al.16 in Finland reported a high quality randomised controlled trial of a levonorgestrel releasing intrauterine system as an alternative to hysterectomy for dysfunctional uterine bleeding for women already on a surgical waiting list. They studied 56 patients. At six months 64% (95% CI 44 to 81) of study women cancelled their hysterectomies compared with 14% (4–33%) in the control group. At a minimum of 23 months follow up 13/27 women continued with the levonorgestrel intrauterine system. In the UK Khan et al.17 advocated that those with dysfunctional uterine bleeding should be given the option to make an informed choice about the use of the levonorgestrel intrauterine system for menorrhagia. A recent review by the Royal College of Obstetricians and Gynaecologists has recommended that a progestogen releasing intrauterine device is an effective treatment for reducing heavy menstrual blood loss and should be considered as an alternative to surgical treatment18. The epidemiology of the use of Mirena for dysfunctional uterine bleeding in the UK is unknown, as is its effect on rates of hysterectomy and endometrial ablation. The introduction of Mirena was too recent to have much impact on the trends described in this study, although it is possible that there may be an effect on the 1995/96 operative rates. Further epidemiological studies are needed to ascertain its effect on the rates of surgery.
That endometrial ablation was commonly used years before the publication of randomised controlled trials should be a matter of concern for evidence-based doctors and their patients. Subsequently, however, there has been good scientific evidence to support its use. Dwyer et al.19 published the first trial study of endometrial ablation compared with hysterectomy for dysfunctional uterine bleeding in 1993. Subsequently, there have been other randomised trials comparing two modes of ablation13, ablation and hysterectomy14, and ablation and medical management20. These suggest that endometrial ablation may have a useful role in gynaecology, and a review by the Royal College of Obstetrics and Gynaecology concluded that these procedures are an effective treatment for menorrhagia18. A rapid rise in the rate of endometrial ablation was observed before these trials were reported (Table 1).
The belief of many gynaecologists that endometrial ablation would substantially reduce the need for hysterectomy, has not materialised. That there was little change in hysterectomy rates is perhaps surprising as endometrial ablation appears to avoid the need for hysterectomy in 80–85% of women treated1,13. It seems likely that the new less invasive technique reduced the threshold for intervention, as may have happened with the introduction of minimally invasive cholecystectomy21. It was predicted that if endometrial ablation successfully replaced hysterectomy millions of pounds may be saved for the National Health Service5. Economic evaluations alongside randomised trials indicate that, even allowing for the risk of treatment failure, endometrial ablation has a lower cost per patient than abdominal hysterectomy22,23. If endometrial ablation had been used only on women who would otherwise have undergone abdominal hysterectomy, there would have been a saving in NHS costs. However, this paper suggests that the net cost effect on the National Health Service of the introduction of endometrial ablation is more complex, and there does not appear to have been a reduction in the overall cost of surgical treatment for menorrhagia.
In conclusion, the fact that endometrial ablation has not replaced hysterectomy should perhaps have been expected, as it cannot relieve uterine pain or remove entirely the risk and worry of conception2. Recent evidence suggests that it is an effective addition to the surgical armamentarium available to surgeons, although its exact role remains unclear18. This study shows that the simple introduction of a new technology does not necessarily lead to more rational use of treatments. The introduction of Mirena might also be predicted to reduce surgery, but this hypothesis needs to be tested empirically, and equally the most effective use and indications need to be established. Further as many women think they have normal periods when they are heavy and vice versa24, the role of any medical or surgical treatment interventions against no intervention is still an important area for further research.
The authors would like to thank Dr M. Sculpher (Senior Research Fellow, Centre for Health Economics, University of York), Professor P. Croft (Professor of Epidemiology at the University of Keele) the editors and peer reviewers of the British Journal of Obstetrics and Gynaecology and the Statistics Division 2 (Hospital Episode Statistics) at the Department of Health for their advice.