**5. Rationale for prevention using levonorgestrel-releasing intrauterine system**

The first beneficial results with frameless FibroPlant® (Contrel Research, Ghent, Belgium) LNG-IUS in women with non-atypical and atypical endometrial hyperplasia were published in 2003. [45] Later, several reports followed using the framed Femilis® (Contrel Research) LNG-IUS and included two cases of endometrial carcinoma, all successfully managed. [46, 47] No failures occurred during follow-up for many years. The numbers were small but indicative of the extraordinary impact of LNG to suppress the hyperplastic and atypical endometrium.

The risk of developing cancer in women with atypical endometrial hyperplasia untreated for 20 years is between 8.6 and 42.5%. [48] In contrast, less than 5% of women with nonatypical endometrial hyperplasia will experience progression to carcinoma. The lower risk for women with nonatypical than atypical hyperplasia can assist in decision making for nonsurgical management of endometrial hyperplasia. The higher risks with atypical hyperplasia progressing to carcinoma warrant consideration of appropriate approaches as concurrent carcinoma among patients with atypical endometrial hyperplasia can range from 17 to 52% across studies. [49]

The standard treatment for endometrial hyperplasia with atypia or early stage (pT1a) adeno‐ carcinoma of the endometrium is staging hysterectomy with bilateral oophorectomy. It is rather clear that for many young women who wish to preserve their fertility, such a decision is unacceptable. Studies using LNG-IUS releasing 20 µg of LNG/day indicate that successful treatment is possible especially when such potent progestins as LNG are administered. Intrauterine route of delivery appears much more effective than oral administration. [50, 51, 52] Successful treatment of early endometrial carcinoma has been reported with a 65 µg/day progesterone-releasing IUS, followed up to 36 months, yet, results of biopsies were negative only in 7 of 11 at 6 months and 6 of 8 at 12 months. [8] A significant reduction of the ERs and PRs expression observed during treatment with the LNG-IUS appears to be a marker for the strong antiproliferative effect of the hormone at the cellular level. Furthermore, as the treat‐ ment is continuous, compliance is not an issue.

The most common source of endogenous unopposed estrogen is chronic anovulation which can be accompanied by either continued ovarian secretion of estradiol or conversion of circulating androstenedione and testosterone to estrone and estradiol, respectively, by aromatase in the adipocytes. Chronic anovulation is a feature of PCOS and is also a condition occurring frequently in the perimenopause. In all these women, whether they have a hyper‐ plastic condition due to exogenous or endogenous excess estrogen, direct delivery of LNG to the target cells of the endometrium using an intrauterine device will cause substantial histologic changes. The result is usually a very uniform suppression of the endometrium, regardless of the duration of treatment, the histologic picture being independent of the distance from the delivery system. [53] Intrauterine progestin delivery, particularly LNG, is therefore probably the most effective intervention in preventing endometrial proliferation because of the uniform suppression of the endometrium throughout the whole thickness of the mucosa caused by the high tissue concentrations of the locally applied hormone. With oral progestin therapy, a higher rate of residual hyperplasia, including complex and atypical hyperplasia, has been observed. [54] It has been demonstrated that the duration of the progestin adminis‐ tration is more important than the daily dose for prevention of endometrial hyperplasia. Again, continuous intrauterine progestin delivery seems optimal. [55]

Over the past 5 to 10 years many reports on the management of atypical endometrial hyper‐ plasia and early cancer of the endometrium with LNG-IUS have been published. [56-63] Remissions were frequent, but also failures occurred. Some consider that at present there is no sufficient evidence because most studies were underpowered or contained a high proportion of patients lost to follow-up. [64] A meta-analysis and systematic review of the literature concludes that the available evidence suggests that treatment with oral or intrauterine progestin are of comparable effectiveness. The risk of progression during treatment is small, but longer follow-up is required. [65] As recently highlighted by Scarselli et al., who followed patients for over 15 years, LNG-IUS represents an option to treat hyperplasia and prevent endometrial carcinoma in at least 85% of patients. [66] Consistent with that study, in order to support stable regression over long periods of time, we recommend continued preventive LNG-IUS use on a life-long basis using a suitable LNG-IUS as the uterine cavity is often very small. Annual assessment of the endometrial thickness by ultrasound is probably sufficient unless vaginal bleeding occurs.
