Hormonal IUD Side Effects (Mirena & Kyleena): Severity by Patient Phenotype

What Hormonal IUDs Are and How They Work in Your Body

Mirena and Kyleena both release levonorgestrel (LNG) directly into the uterine cavity. Mirena releases approximately 20 mcg per day initially, tapering to roughly 10 mcg per day by year five. Kyleena starts at about 17.5 mcg per day and tapers to 7.4 mcg per day by year five. Because the hormone acts locally, systemic LNG levels are far lower than with oral contraceptive pills containing the same progestin.

Serum LNG with Mirena averages 150 to 200 pg/mL, compared to 1,500 to 2,000 pg/mL typical of a 150 mcg LNG oral pill. That local-first pharmacokinetics matters for side-effect prediction. Women who previously struggled with systemic progestin side effects on oral or injectable contraceptives may tolerate an LNG IUD much better. The flip side is that the low serum level means the IUD provides minimal protection against ovarian cysts or acne compared to combined oral contraceptives.

How the Menstrual Cycle Changes After Insertion

The endometrium becomes decidualized and atrophic within weeks of insertion. Follicular development and ovulation usually continue, which is why you may still notice premenstrual symptoms even when bleeding has lightened or stopped. In the Phase III Kyleena trial (Gemzell-Danielsson et al., 2017), 88% of cycles were ovulatory in year one despite dramatic endometrial suppression.

Systemic vs. Local Effects: Why One Woman's Experience Differs From Another's

Serum LNG levels vary two- to threefold between individuals at the same device dose, partly due to differences in cytochrome P450 3A4 activity. Women taking enzyme-inducing medications (rifampicin, some anti-epileptics) may have lower LNG exposure, which could reduce contraceptive efficacy slightly, though the ACOG Practice Bulletin on Contraception (2023) notes the intrauterine mechanism provides a substantial independent barrier. Higher baseline sex-hormone-binding globulin (SHBG), common in women on estrogen therapy, binds more free LNG and further lowers the effective systemic dose.

Side Effects by Severity: A Frequency-Based Map

Most side-effect guides lump everything together. This section separates them by how often they occur and how serious they are, because those two dimensions do not always line up.

Very Common (More Than 10% of Users)

Unscheduled bleeding and spotting is the number-one reason women discontinue in the first year. In the large CHOICE Project cohort (Peipert et al., 2011), bleeding irregularity was cited by 31% of Mirena users as a concern at three months, dropping sharply by month six. Kyleena users report similar patterns, though total bleeding volume tends to be slightly higher with Kyleena versus Mirena because Kyleena's lower dose produces less endometrial suppression.

Cramping and pelvic discomfort peak at insertion and in the first four to eight weeks. Nulliparous women report more severe insertion pain than parous women, a gap confirmed in a 2019 BMJ meta-analysis that found nulliparity was the single strongest predictor of procedural pain scores.

Ovarian follicular cysts develop in up to 12% of Mirena users in any given year, per the Mirena US prescribing information. Most resolve spontaneously within two to three months without intervention.

Common (1 to 10% of Users)

Headache is reported in approximately 7 to 8% of users in registration trials. Whether this is causally related to LNG versus insertion stress is difficult to separate.

Acne and seborrhea occur in 5 to 7% of users. LNG is androgenic relative to other progestins, so women already prone to hormonal acne may notice worsening, particularly in the first three months before steady-state is reached.

Mood changes and depressed mood appear in a large Danish registry study by Skovlund et al. (2016) involving over one million women, which found a relative risk of 1.40 for first antidepressant prescription among progestin-IUD users versus non-hormonal contraception users. The absolute risk increase was small, and the study cannot confirm causality, but it is the most-cited real-world signal on this question.

Decreased libido is listed in Mirena's EU label but not prominently in the US label. The evidence linking LNG IUD specifically to HSDD (hypoactive sexual desire disorder) is limited. Free testosterone may fall modestly due to LNG's weak androgen receptor activity raising SHBG in some users, but this is inconsistent across studies.

Uncommon (0.1 to 1% of Users)

Expulsion occurs in roughly 3 to 6% of users overall, placing this at the boundary of common and uncommon depending on the population. In adolescents under 20, expulsion rates may reach 10 to 15%, per a 2021 ACOG Committee Opinion on Adolescent Long-Acting Reversible Contraception.

Uterine perforation occurs in approximately 1.3 per 1,000 insertions based on the EURAS-IUD observational study. Postpartum insertion before 36 weeks is the strongest risk factor, with a perforation rate approaching 6 per 1,000 in that window.

Rare (Fewer Than 1 in 1,000 Users)

Pelvic inflammatory disease (PID) risk is concentrated in the first 20 days after insertion and is driven by pre-existing infection, not the device itself. Background STI rates in the population matter more than the IUD. After 20 days, PID risk returns to baseline in monogamous women.

Ectopic pregnancy is rare because the overall pregnancy rate is very low (0.1 to 0.5 per 100 woman-years), but if pregnancy does occur while an LNG IUD is in place, approximately 50% are ectopic. The Mirena label advises immediate evaluation if pregnancy is suspected.

Side-Effect Severity by Patient Phenotype

This is where one-size-fits-all side-effect lists fail you. Your baseline hormonal environment, reproductive history, and underlying conditions shape which effects you are most likely to experience and how severely.

Reproductive-Age Women With No Underlying Conditions

For most healthy women aged 18 to 40, the hormonal IUD is one of the most effective and best-tolerated long-acting contraceptive methods, with a first-year typical-use failure rate of 0.1 to 0.2%. Bleeding change is the dominant concern. Women who enter with regular, moderate cycles adapt more smoothly than women with heavy periods (who often improve dramatically) or women with very light periods who find any irregularity distressing.

Women With PCOS

PCOS involves androgen excess, insulin resistance, and often heavy or irregular bleeding. The LNG IUD can reduce menorrhagia in PCOS, which is a genuine benefit. Acne and hirsutism are less likely to improve with an LNG IUD than with combined oral contraceptives, because the IUD does not suppress ovarian androgen production or raise SHBG enough to reliably reduce free testosterone. A 2020 systematic review in Fertility & Sterility concluded that LNG IUDs reduce bleeding in PCOS but provide no consistent benefit for androgen-mediated symptoms.

Women with PCOS who are also insulin-resistant should know that LNG has weak glucocorticoid-like metabolic effects at higher doses; at IUD doses the clinical impact on insulin sensitivity appears negligible based on available data, though this has not been studied in large RCTs focused specifically on PCOS and LNG IUD.

Women With Endometriosis

The LNG IUD is endorsed by the European Society of Human Reproduction and Embryology (ESHRE) as a first-line medical management option for endometriosis pain and dysmenorrhea. In a Cochrane review (Abou-Setta et al., 2013), post-surgical Mirena insertion significantly reduced endometriosis recurrence rates and pain scores versus surgery alone.

Side effects in women with endometriosis follow similar patterns to the general population, but this group may find the initial cramping period more tolerable given their baseline pain experience. The reduction in dysmenorrhea typically outweighs the nuisance of early spotting for this phenotype.

Women With Uterine Fibroids

Fibroids that distort the uterine cavity are a contraindication, because they increase expulsion risk and the device may not sit correctly. Women with subserosal or intramural fibroids that do not distort the cavity can often use an LNG IUD safely. Heavy bleeding due to fibroids may improve, although fibroid-related heavy bleeding responds less completely to the IUD than adenomyosis-related bleeding.

Nulliparous and Adolescent Women

The American Academy of Pediatrics and ACOG both recommend LARCs, including hormonal IUDs, as first-line contraception for adolescents. The main phenotype-specific concerns are insertion pain (higher without prior vaginal delivery) and expulsion risk (roughly double the adult parous rate). A smaller-framed uterus, measured by ultrasound at under 6 cm in length, is associated with both higher expulsion and higher perforation rates.

Mood effects deserve particular attention in adolescents, who have higher baseline rates of depression and anxiety. The Skovlund registry data showed the largest relative risk increases in the youngest age groups, though absolute numbers remained small.

Perimenopausal Women (Ages 40 to 55)

This is one of the most clinically underserved use cases. Mirena is FDA-approved as the progestogen component of the FDA-approved Bijuva (estradiol/progesterone) and off-label alongside transdermal or patch estrogen for menopausal hormone therapy (MHT). A 52 mg LNG IUD provides endometrial protection for women using systemic estrogen, eliminating the need for a separate oral progestogen. The British Menopause Society endorses this approach, with the IUD offering local progestogen delivery and fewer systemic progestogen side effects (bloating, mood changes) than oral options.

Perimenopausal women may experience more variable bleeding patterns because their own ovarian estrogen is fluctuating. The IUD does not control anovulatory bleeding driven by estrogen surges. Women who expect amenorrhea may find cycle unpredictability frustrating in this transition.

The IUD also provides contraception, which matters: spontaneous ovulation can still occur in perimenopause, and pregnancy after 45 carries substantially elevated risks.

Postpartum and Breastfeeding Women

Postpartum insertion timing affects both side effects and risk profile. Immediate postpartum insertion (within 10 minutes of placenta delivery) carries a higher expulsion rate (approximately 15 to 25%) versus interval insertion at six or more weeks. Perforation risk at four to six weeks postpartum is also elevated compared with non-postpartum insertion.

Breastfeeding women may notice an earlier return of menses when the IUD is the sole contraceptive method, compared to women using lactational amenorrhea alone. LNG transfer into breast milk is measurable but low. A WHO Medical Eligibility Criteria for Contraceptive Use review rates LNG IUD as Category 2 (benefits generally outweigh risks) in breastfeeding women from birth to six weeks postpartum, and Category 1 (no restriction) after six weeks.

Pregnancy, Lactation, and Contraception Requirements

This section is required reading if you are pregnant, planning pregnancy, or postpartum.

During established pregnancy: Contraindicated. If you become pregnant while an LNG IUD is in place, the device should be removed as soon as possible if the strings are visible. Leaving it in place is associated with increased risks of spontaneous abortion, preterm birth, and chorioamnionitis. Because approximately 50% of in-situ IUD pregnancies are ectopic, an ultrasound to confirm intrauterine location is the first step.

Pregnancy category: The FDA's pre-2015 letter system classified LNG IUD as Category X for use during an established pregnancy. Under the current PLLR labeling, the Mirena and Kyleena prescribing information state that the device is contraindicated in known or suspected pregnancy.

Fertility return: Ovulation typically resumes within the first cycle after IUD removal. The LNG IUD does not reduce long-term fertility, unlike some injectables. Women trying to conceive should have the device removed when they are ready to attempt pregnancy; there is no recommended waiting period.

Breastfeeding: LNG is present in breast milk at low concentrations. Studies have not identified developmental effects in exposed infants. WHO and CDC Medical Eligibility Criteria consider the LNG IUD compatible with breastfeeding after six weeks postpartum. Before six weeks, the theoretical concern is neonatal LNG exposure during the period of greatest hepatic immaturity, though real-world evidence of harm is absent.

If you are trying to conceive: Remove the IUD when you are ready. No waiting period is needed. Fertility returns to your baseline, not to a contraceptive-suppressed baseline, because the IUD does not suppress ovulation in most cycles.

Who This Is Right For, and Who Should Think Carefully

No contraceptive suits every woman. Here is a life-stage and condition-based summary.

Well-matched phenotypes:

• Women with heavy menstrual bleeding or adenomyosis seeking bleeding reduction
• Women with endometriosis using it as maintenance therapy after surgery
• Perimenopausal women who want contraception plus the progestogen arm of MHT in one device
• Women who cannot use estrogen-containing contraceptives (migraine with aura, elevated clotting risk, cardiovascular risk factors)
• Adolescents who want a low-maintenance, highly effective contraceptive and are counseled on insertion discomfort

Phenotypes that warrant extra discussion:

• Women with a history of depression or mood disorders. The Skovlund signal is not definitive, but mood monitoring is warranted.
• Women with androgenic skin conditions (hormonal acne, hirsutism) who may not see the skin benefit they would get from combined oral contraceptives.
• Women with cavity-distorting fibroids. MRI or sonohysterography before insertion is advisable.
• Women who are 4 to 12 weeks postpartum. Perforation risk is highest in this window; insertion at 6 weeks or later (once uterine involution is complete) reduces that risk.
• Women on enzyme-inducing medications. Efficacy data for this combination is thin, and backup or alternative contraception may be discussed.

Absolute contraindications (from the Mirena prescribing information):

• Known or suspected pregnancy
• Unexplained uterine bleeding
• Uterine anomaly or fibroid distorting the cavity
• Acute or recurrent PID
• Postpartum endometritis or septic abortion in the past three months
• Known or suspected uterine or cervical malignancy
• Liver disease or liver tumor
• Hypersensitivity to any component

Adverse Event Signals From Post-Market Surveillance (FAERS)

The FDA's Adverse Event Reporting System (FAERS) captures voluntary reports, which overrepresent serious events and underrepresent common mild ones. The most frequently reported FAERS categories for levonorgestrel IUDs (2010 to 2022) include device expulsion, device malposition, uterine perforation, and mood-related events, consistent with trial-era data. A 2021 analysis of FAERS reports for intrauterine devices found that the signal-to-noise ratio for depression and anxiety was higher for LNG IUDs than for copper IUDs, mirroring the Skovlund registry finding. FAERS data cannot establish causation or incidence rates.

One pattern in FAERS that deserves attention: reports of persistent side effects after removal, sometimes called "post-IUD syndrome" colloquially online. This term has no formal clinical definition, and no peer-reviewed trial has characterized a distinct post-removal syndrome. Hormonal IUDs do not suppress the hypothalamic-pituitary-ovarian axis the way injectable or implant contraceptives do, so protracted post-removal hormonal disruption is physiologically less plausible than with depot medroxyprogesterone acetate. Women who have symptoms after removal should be evaluated for other causes (autoimmune conditions, thyroid dysfunction, perimenopause) rather than assuming device causation.

Evidence Gaps: What We Don't Know Yet

Women have been chronically underrepresented in contraceptive research as anything other than the subject of pregnancy prevention trials. Several gaps are specific to sex-specific physiology and deserve acknowledgment.

The mood signal from the Skovlund study is real and worth taking seriously, but the original authors themselves noted that confounding by indication is difficult to exclude, since women with mood symptoms may seek hormonal contraception at different rates than women without.

Libido and sexual function data for LNG IUD specifically (not combined oral contraceptives, which dominate the literature) are sparse. Most trials did not measure sexual function outcomes at all, reflecting a historical blind spot in contraceptive research.

The interaction between LNG IUD and PCOS insulin resistance has not been studied in a powered RCT. Extrapolations from oral LNG doses should not be applied directly to IUD doses.

As WomanRx clinical reviewer Rachel Goldberg, MD, puts it: "The missing data on mood and libido is not reassuring silence. It is a gap. Women deserve trials that measure what matters to them, not just pregnancy rates. Until we have those trials, my job is to monitor each patient individually and take her symptom reports seriously rather than dismissing them because we lack a population-level signal."