Special Feature
Postmenopausal Hormone Therapy and Endometrial Cancer
By Leon Speroff, MD
In the february 15, 1997, issue of Lancet, epidemiologists from Seattle reported a case-control study of endometrial cancer and the postmenopausal sequential, cyclic use of estrogen-progestin therapy (Beresford SAA, et al. Lancet 1997;349:458-461). The surprising conclusion was that the use of combined estrogen-progestin regimens for five or more years was associated with an increased relative risk of endometrial cancer, even with 10-21 days of added progestin per month. What are we to make of this? Have we been wrong in believing that 10 or more days of progestin per month added to an estrogen regimen will protect against endometrial cancer?
First, let’s look closely at this study. It is a population-based, case-control study using cases identified from a regional cancer registry in the state of Washington. Out of a total of 1154 eligible cases, 832 completed interviews, and ultimately a total of 394 cases were matched with 788 controls. As is already well-known, the use of unopposed estrogen for at least six months was associated (after adjustments for age and body mass index) with a four fold increased risk of endometrial cancer (relative risk [RR], 4.0; confidence interval [CI], 3.1-5.1). These cases accounted for the majority of cases of endometrial cancer, 324 of the 394 total (82%). After discontinuing unopposed estrogen treatment, the risk of endometrial cancer lingers for several years. Thus, it is not surprising that those women (76 cases) who had used unopposed estrogen for at least six months and then used combined, sequential estrogen-progestin therapy also had a statistically significant increased risk of endometrial cancer (RR, 2.7; CI, 1.9-4.0). Among the women who used only a sequential, combination regimen (67 cases), the relative risk of 1.4 did not achieve statistical significance (CI, 1.0-1.9).
There were 25 cases of endometrial cancer in women who used a progestin for less than 10 days per month, and the relative risk of 3.1 was statistically significant. However, the 25 cases of endometrial cancer in women who used a progestin for 10-21 days per month (RR, 1.3) did not achieve statistical significance. This would suggest that an adequate duration of exposure would protect against endometrial cancer. But most concerning was the analysis based upon duration of current use. (See Table 1.)
Table 1
Cases of Endometrial Cancer Among Current Users of Progestin
Current users of progestin for less than 10 days per month:
| Duration of progestin use | Cases | Controls | Relative Risk |
| 6 months to 5 years | 11 | 13 | 2.2 (0.9-5.2) |
| 5 or more years | 14 | 9 | 4.8 (2.0-11.4) |
Current users of progestin for 10-21 days per month:
| Duration of progestin use | Cases | Controls | Relative Risk |
| 6 months to 5 years | 12 | 48 | 0.7 (0.4- 1.4) |
| 5 or more years | 12 | 15 | 2.7 (1.2-6.0) |
First notice that statistical significance was achieved only with five or more years of use in each category. But also notice the relatively small number of cases and controls in these groups, reflected in the wider confidence intervals in contrast to the overall results noted earlier. Thus, we clinicians are left with deciding how real these conclusions are based on small numbers with wide confidence intervals that reflect the imprecision of small numbers. Indeed, where the case numbers were respectable, the increased risk with a combination regimen did not achieve statistical significance.
Our concern is due to the fact that combined estrogen-progestin therapy is relatively recent (in epidemiological terms) and we have based our convictions on the results with surrogate end points. We are comfortable that standard regimens of combination therapy are effective in preventing endometrial hyperplasia (as demonstrated in the PEPI trial). However, surrogate end points (in this case, endometrial hyperplasia) are never a completely satisfactory substitute for actual clinical events. Furthermore, the endometrial hyperplasia data are based upon relatively short-term trials, and we do not have the comfort of long-duration documentation.
This current report and the Upsalla, Sweden, report (Persson I, et al. BMJ 1989;298:147) dispel the hope that combined therapy would be associated with a risk of endometrial cancer that would even be below that of women who never used hormones. Nevertheless, previous studies have not indicated an increased relative risk, but rather neither an increased nor a decreased risk (see Brinton and Hoover, Obstet Gynecol 1993;81:265-271, for review). Only time will tell whether the current report is pointing in the right direction; however, the data are not overwhelming in their strength and numbers.
The effect of the daily, continuous combination regimen has thus far been identical to the sequential regimen on the rate of the surrogate end point, hyperplasia (i.e., equally effective). About six months ago, I heard a report in Europe that documented 15 and more years of current use of a daily combination that used 2 mg estradiol valerate, a higher estrogen dose than we use, and after more than 7 million woman-years of use, there were only seven cases of endometrial cancer, four of which had hyperplasia before beginning treatment.
Certainly, combined estrogen-progestin treatment is effective in reducing the risk of endometrial cancer associated with unopposed estrogen. And, it is certain that not all cases of endometrial cancer can be prevented with combination therapy, but the risk of endometrial cancer with combination regimens must be very, very small, if it is at all elevated. In terms of the size of risk, this new report is actually reassuring. It does remind us, however, that endometrial surveillance and appropriate diagnostic intervention are on-going obligations for clinicians who are proponents of the use of postmenopausal hormone therapy.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content