Advertisement

Life tables without confidence intervals may mislead

      To the Editors:

      We agree with Unger and Meeks (Unger JB, Meeks GR. Hysterectomy after endometrial ablation. Am J Obstet Gynecol 1996; 175: 1432-7) that further long-term studies are needed to define the role of endometrial ablation for menorrhagia. We take issue with their conclusions and inferences, principally on the ground that these are based on the results obtained from a study of 41 women. They report that 34% of their patients underwent hysterectomy within 5 years of the initial surgery, a higher rate than shown in other studies.
      • O'Connor H
      • Magos A
      Endometrial resection for the treatment of menorrhagia.
      Our study of endometrial resection with use of life tables reported a 20% risk of requiring further surgery by 5 years, and 9% of our patients had hysterectomy, some opting for repeat resection. Furthermore, 98% of repeat surgery occurred in the first 3 years and there was no linear relationship between the rate of further surgery and time.
      There are several explanations why the results of Unger and Meeks are at variance with other published data. There is some evidence that endometrial resection is more effective in long-term relief of menorrhagia than rollerball ablation is. Operator experience can have a major influence on outcome, and it is unclear whether the cases reported were part of the learning curve. Unger and Meeks found no effect of patient age on the risk of treatment failure, in contrast to several publications showing that younger women do less well.
      The small size of the study of Unger and Meeks seems the key to why care must be taken in interpreting their results. They do not give confidence intervals for their life-table analysis. We calculated the 95% confidence intervals for their data, and they show that, because the overall study group is so small, the intervals with respect to the cumulative risk of hysterectomy are very large (Table 1). Their results have to be taken in the context of these figures, which shows that it is impossible to judge from this study the efficacy of endometrial ablation. It is also inappropriate to extrapolate their results over the next 5 to 8 years.
      Table ICalculated 95% confidence intervals for data of Unger and Meeks
      Interval No.Cumulative probability of avoiding hysterectomy (%)Confidence interval (%)
      197.6±5.8
      287.8±26.8
      380.5±38.9
      470.8±51.0
      565.7±54.9
      It is now 16 years since the first report of hysteroscopic endometrial ablation by Goldrath et al.
      • Goldrath MH
      • Fuller TA
      • Segal S
      Laser photovaporization of endometrium for the treatment of menorrhagia.
      Since then, carefully controlled prospective randomized trials with hysterectomy have confirmed the many advantages, both clinical and financial, of hysteroscopic surgery.
      • O'Connor H
      • Broadbent M
      • Magos A
      • McPherson K.
      The Medical Research Council randomised trial of endometrial resection versus hysterectomy in the management of menorrhagia.
      We commend Unger and Meeks on their use of life tables, but their study has insufficient power to put these results into question.

      References

        • O'Connor H
        • Magos A
        Endometrial resection for the treatment of menorrhagia.
        N Eng J Med. 1996; 335: 151-156
        • Goldrath MH
        • Fuller TA
        • Segal S
        Laser photovaporization of endometrium for the treatment of menorrhagia.
        Am J Obstet Gynecol. 1981; 140: 14-19
        • O'Connor H
        • Broadbent M
        • Magos A
        • McPherson K.
        The Medical Research Council randomised trial of endometrial resection versus hysterectomy in the management of menorrhagia.
        Lancet. 1997; 349: 897-901