How does Mirena work?

The contraceptive and therapeutic effects of Mirena are mainly derived from the local effects of LNG in the uterus¹.

Effect on cervical mucus and sperm function

Barbosa et al.² suggested that the cervical mucus volume was reduced by Mirena in some users, while Jonsson et al.³ reported an increase in the weight of cervical mucus, thus inhibiting the passage of sperm. A recent study by Lewis et al.⁴ showed that the mid-cycle cervical mucus of Mirena users is of poor quality and prevents endocervical sperm transport in vitro.

It has also been postulated that the migration of sperm through the uterine and fallopian tubal fluid is inhibited^5,6.

Effect on endometrium

The high levonorgestrel (LNG) concentrations in the endometrium down-regulate endometrial estrogen and progesterone receptors, making the endometrium insensitive to circulating estradiol (thereby suppressing endometrial growth)^7,8.

After only a couple of months of Mirena use, the glands of the endometrium atrophy, the stroma becomes swollen and decidual, the mucosa thins and the epithelium becomes inactive – described by Perino et al.⁹ as ‘cylindrico-cubic, monostratified and without mitosis’. Vascular changes include a thickening of arterial walls, suppression of the spiral arterioles and capillary thrombosis¹⁰. An inflammatory reaction characterized by an increase in neutrophils, lymphocytes, plasma cells and macrophages has been observed^11,10 and focal stromal necrosis may also occur^7,10. The endometrial changes are uniform within 3 cycles after placement of the system¹⁰ and no further histological development takes place over the long term⁷.

The endometrial morphological features associated with Mirena use are summarized in Table 1.

Table 1: Endometrial morphological features associated with Mirena use¹¹

Morphological feature	Proportion of cases with feature
Decidualization of stroma	96%
Atrophy of glands	87%
Stromal inflammatory cell infiltrate	79%
Stromal inflammatory cell infiltrate including plasma cells	27%
Surface papillary formations	51%
Stromal myxoid change	39%
Stromal hemosiderin disposition	32%
Glandular metaplasia	9%
Stromal necrosis	7%
Reactive atypia in surface glands	4%
Stromal calcification	1%

The initial changes in the endometrium caused by Mirena may be associated with irregular bleeding or spotting, particularly in the first few months of treatment. Although the mechanism underlying irregular bleeding or spotting with hormonal therapy is not well understood, it may involve matrix metalloproteinases and their tissue inhibitors, and/or a hormonal influence upon endometrial sex steroid receptors, or ligand availability in the endometrium^12,13.

With Mirena, once the endometrial effects are established, the bleeding becomes less in quantity than usual, or may cease altogether. Mirena may also induce the release of the binding protein for endometrial insulin-like growth factor 1, thus inhibiting its growth-promoting effect¹⁴. This may explain the protective effect of Mirena on fibroid formation reported by Sivin & Stern¹⁵. An alternative explanation may be that the thin estrogen-resistant endometrium will not produce enough growth factors to stimulate the growth of fibroids. The morphological changes in the endometrium revert to ‘normal’, and menstruation has been reported as early as the first month following removal of Mirena¹⁶.

After an initial increase in spotting during the first few months of use, Mirena brings about a highly significant reduction in days of bleeding and/or spotting (Figure 1)^17,16. In the study by Nilsson et al.¹⁶ a significantly greater proportion of women in the Mirena group than in the Cu-IUD (Nova T) group reported an improvement in dysmenorrhea (35% vs. 9%; p<0.005).

Figure 1: Mean number of days of a) spotting and b) bleeding during the initial 12 months’ use of Mirena and Nova T¹⁷.

In a study comparing the effects of Mirena and the Cu-IUD (Nova T) 3 months after placement, the volume of menstrual bleeding decreased by 60% in users of Mirena, whereas it increased by 90% in users of the Cu-IUD. Mean menstrual blood loss following placement of Mirena also differed from that after placement of the Cu-IUD (reduction of >75% with Mirena vs. increase of >50% with Cu-IUD; p<0.001) in women with normal pre-placement menstruation (Figure 2)¹⁸.

Figure 2: Change in menstrual blood loss with Mirena and a Cu-IUD during 12 months of use.

The number of bleeding days in women with normal blood loss is reduced by Mirena¹⁹. Approximately 20% of women using the system for contraception experience amenorrhea (defined as the absence of bleeding or spotting within the last 90 days) within 1 year. The blood hemoglobin levels and serum ferritin levels also respond positively to the marked reduction in blood loss brought about by the system²⁰.

During a second consecutive period of Mirena use, a higher proportion of women experience amenorrhea, compared with the first 5-year period (Figure 3)²¹. Many women become free of bleeding soon after the second Mirena placement, and the initial period of spotting that occurred after the first Mirena placement does not typically recur²¹.

Figure 3: Bleeding pattern with continuous long-term use of Mirena.

Ovarian function

During the first year of use, some women experience suppression of ovarian function. Thereafter, most cycles are ovulatory²² and the incidence of ovulatory cycles with Mirena and with the Cu-IUD is the same (85%)¹⁸. The effect of LNG on ovarian function depends on plasma LNG levels and there are marked inter-individual differences in the plasma levels achieved^22,23. In general, the anovulatory cycles (5–15% of treatment cycles) correlate with higher levels of LNG^18,24.

For complete suppression of ovulation, a daily intrauterine release of more than 50 µg of LNG is required²³. With Mirena, only 20 µg/day LNG is released. Determination of plasma estradiol and progesterone (P) levels indicates that women using Mirena generally have normal ovulatory cycles^25,24. Other types of ovarian function that may be exhibited by women using Mirena include: anovulation with some inhibition of estradiol production; anovulation with high follicular activity; and ovulation with an inadequate luteal phase²⁴.

There is no reduction in E2 levels during the use of Mirena. Figure 4 shows the mean plasma E2 and LNG concentrations in menstruating and amenorrheic women²³. Menstrual bleeding does not itself reflect ovarian function among women using Mirena: not only are average progesterone levels the same among those with regular, scanty bleeds as those with oligomenorrhea, but the levels of E2 and the incidence of ovulation are similar for the two groups²².

Figure 4: Mean plasma estradiol and LNG concentrations in menstruating and amenorrheic women using Mirena for 1 or 2 years²³.

The gradual reduction of endometrial thickness and conversion of the functional endometrium to a rest stage resistant to estrogen stimulation is visible in the gradual
reduction in menstrual blood loss during the first few months. Oligomenorrhea develops, despite normal ovarian function. No difference in the incidence of ovulation is found between menstruating and amenorrheic women²².