Levonorgestrel-releasing IUD

Limits on what this shows

• Key negative/null findings: Khokha et al. 1987 (PMID 3103689), a separate levonorgestrel-specific rat hepatic study, found one mitochondrial lipid-synthesis enzyme (glycerol-3-phosphate acyltransferase) unchanged with levonorgestrel treatment, a genuine null finding for that specific mitochondrial endpoint, though at systemic oral/parenteral dosing, not IUD-equivalent exposure.
• At-risk subgroups: Not established in humans, and not established at IUD-equivalent systemic exposure in any species; both supporting studies used substantially higher systemic dosing than the IUD delivers.
• Counter-evidence on record: The Khokha et al. null finding above; more fundamentally, the IUD's own labeled pharmacokinetics (peak serum 175-300 pg/mL versus an oral/parenteral rat toxicology dose) argue against translatability, though no study has directly tested mitochondrial endpoints at IUD-equivalent exposure to confirm the mechanism is absent, which is why this is graded yellow (plausibility not ruled out) rather than white.

Regulatory status

Not addressed in any of the three labels reviewed.

Population evidence

Kretzschmar et al. 1989 (PMID 2500261, Chemico-Biological Interactions 70:157-166) is a Tier B rat hepatic mechanistic study using systemically administered levonorgestrel (reported in this database's prior review of the same paper as 1 and 10 mg/kg body weight), finding mitochondrial ultrastructural lesions and preferential mitochondrial glutathione depletion at the higher dose. Title, authors, and journal were reconfirmed via the OpenAlex works API this session; the full abstract text and dose-response figures were not independently re-verified against primary text this session (abstract_inverted_index unavailable via OpenAlex; individual PubMed and Europe PMC pages were rate-limited on every attempt), so the specific dose figures are carried forward from this database's existing verified record of the same source rather than freshly re-confirmed.

Mechanism

The mechanism proposed (mitochondrial glutathione depletion at high systemic/hepatic exposure) requires levonorgestrel to reach a substantial hepatic concentration. The IUD's own label data show peak serum levonorgestrel of roughly 175-300 pg/mL across all three devices reviewed, far below sustained oral or injectable progestin exposure and orders of magnitude below any plausible human-equivalent extrapolation of a rat 10 mg/kg systemic dose. This route/dose gap is larger for the IUD than it is for oral levonorgestrel, where the same rat data were also judged only weakly translatable.

Counter-evidence held on record

This is high-dose rat hepatic toxicology data using a systemic route the IUD does not replicate at anything like the same exposure. Do not present this as evidence that the levonorgestrel IUD damages mitochondria; the exposure gap makes this one of the weakest yellow-tier findings in this dossier. No supplement or monitoring action is indicated.

Limits on what this shows

• Key negative/null findings: The copper-IUD arm changing in the same direction as the levonorgestrel-IUD arm is itself the key limiting finding: it argues against attributing the effect specifically to levonorgestrel rather than to IUD placement generally.
• At-risk subgroups: Not established; the study did not stratify by baseline risk factors for oxidative stress.
• Counter-evidence on record: The shared-direction change in the copper-IUD comparator (above) is the primary counter-evidence against a levonorgestrel-specific attribution, which is why this entry is capped at yellow rather than orange despite resting on a real, drug-specific, Tier A/B comparative study.

Regulatory status

Not addressed in any label reviewed.

Population evidence

Buyuk et al. 2021 (PMID 33594922, Gynecological Endocrinology, DOI 10.1080/09513590.2021.1878132), a Tier A/B human comparative study (n=30 per group), measured thiol-disulfide homeostasis before insertion and at 6 months in women using a T380A copper IUD versus a levonorgestrel IUD. Abstract retrieved via the OpenAlex works API this session: native thiol was similar between groups at baseline (293.1±43.5 copper vs 290.4±42.4 mcmol/L levonorgestrel); by 6 months, both groups showed significant increases in disulfide/native-thiol and disulfide/total-thiol ratios, with corresponding decreases in native-thiol/total-thiol ratio, indicating a shift toward oxidative stress in both device types.

Mechanism

No hormone-specific mechanism is named; the authors' own framing attributes the change to both device types alike, consistent with a local uterine foreign-body/inflammatory response rather than a levonorgestrel pharmacologic effect. Meric Turgut et al. 2025 (PMID 41012385), a non-mammalian zebrafish toxicology study, provides only Tier B, environmentally-dosed, non-human evidence of levonorgestrel-associated oxidative damage, of limited relevance to human intrauterine exposure.

Counter-evidence held on record

Do not present this as "the levonorgestrel IUD increases oxidative stress" in isolation; the same study found the same direction of change with a copper IUD, so the more accurate framing is that IUD placement generally, not the hormone specifically, is associated with this marker shift at 6 months. No supplement or antioxidant recommendation is indicated by this evidence.

Limits on what this shows

• Key negative/null findings: Perello-Capo et al. 2023 (PMID 37487868), a 3-year randomized phase 4 trial of the 13.5 mg device (Kyleena-strength) versus a Nova T copper IUD, found no significant hemoglobin or hematocrit difference between groups, indicating the hematologic benefit is dose- and/or baseline-bleeding-dependent, not a universal levonorgestrel-IUD effect. A single case report, Vieira da Motta and Vieira da Motta 2013 (PMID 23843714), documents abnormally elevated serum ferritin and iron in one LNG-IUS user despite normal hemoglobin, an "iron overload syndrome" framing; this is a rare, single-patient counter-signal recorded as a genuine at-risk-subgroup flag, not a reason to discount the population-level finding.
• At-risk subgroups: Benefit is specific to women with baseline heavy menstrual bleeding using a 52 mg device (Mirena, Liletta); not established for the 19.5 mg (Kyleena) or 13.5 mg (Skyla/Jaydess-class) lower-dose systems. Conversely, individuals with a predisposition to iron overload (e.g., hereditary hemochromatosis) may warrant ferritin monitoring given the Vieira da Motta case report, since a device that raises ferritin in the general HMB population could theoretically compound an existing overload risk.
• Counter-evidence on record: See Key negative/null findings above. Every comparative trial located used an active comparator (oral medroxyprogesterone acetate, or a copper IUD), not placebo or no treatment, so part of the apparent benefit reflects the comparator's own bleeding profile rather than an absolute untreated baseline; this active-comparator limitation is the main reason this entry is capped at orange rather than red.

Regulatory status

Mirena and Liletta are both FDA-indicated for treatment of heavy menstrual bleeding. Mirena's labeled trial (n=79 Mirena vs n=81 medroxyprogesterone acetate, 6 cycles, baseline MBL ≥80 mL by alkaline hematin method) found roughly 9 of 10 Mirena-treated subjects had blood loss reduced by more than half, meeting a <80 mL and ≥50%-reduction success definition; Liletta's labeled trial (n=105) reported 80% treatment success and a 96% median MBL reduction by study end. Neither label reports hemoglobin/ferritin figures directly in the sections reviewed.

Population evidence

Imperato et al. 2002 (PMID 12063443), an 8-year prospective study (n=223: 38 regularly-menstruating controls, 185 IUD users), found the LNG-IUS group had significantly reduced menstrual loss and significantly increased serum ferritin within 6 months of insertion, while the copper-IUD comparison group did not improve and the authors concluded copper-releasing devices may worsen iron status via copper release, a contrast, not a confound, with the LNG-IUS finding. Kaunitz et al. 2012 (PMID 22959906, Contraception 86:452-457), titled by the authors "Levonorgestrel-releasing intrauterine system for heavy menstrual bleeding improves hemoglobin and ferritin levels," is a drug-specific Tier A source; its full numeric results could not be retrieved this session (ScienceDirect returned an empty/script-gated page and repeated PubMed/Europe PMC/Semantic Scholar rate-limiting), so this dossier cites its confirmed title, design, and journal indexing rather than asserting unretrieved effect sizes. Lowe and Prata 2013 (PMID 23122687), a systematic review stratified specifically to LNG-IUS versus copper-IUD hemoglobin and ferritin, is Tier A by design; full abstract text was likewise not retrievable this session (see Sources). Xiao et al. 2003 (PMID 12749438, Fertility and Sterility 79:963-969) and Rybo, Andersson, and Odlind 1993 (PMID 8489751) are additional levonorgestrel-IUS-specific sources indexed as measuring menstrual blood loss with hemoglobin and/or ferritin before and after treatment.

Mechanism

Reduced menstrual blood loss lowers cumulative iron loss; not a direct pharmacologic effect on iron absorption, transport, or storage. This is the drug's own intended, labeled therapeutic effect for the heavy-menstrual-bleeding indication, so it is written here as a downstream hematologic consequence of that effect (per Hard Rule 2), not as an independent nutrient-sparing action.

Counter-evidence held on record

This is a therapeutic, protective effect tied to a specific FDA-labeled indication (heavy menstrual bleeding) and to the higher-dose devices, not a general reason to expect iron benefits from any levonorgestrel-IUD use, and not a signal to add iron supplementation. It is also not evidence that the lower-dose devices (Kyleena, Skyla) provide the same benefit. Women switching from heavy bleeding to a 52 mg levonorgestrel IUD who were previously on iron supplementation may need that reassessed by their prescriber, not increased; individuals with a personal or family history of iron overload should have ferritin monitored rather than assumed to need iron.

Limits on what this shows

• Key negative/null findings: Perello-Capo et al. 2023 (PMID 37487868): no significant hemoglobin or hematocrit difference between the 13.5 mg LNG-IUD and a copper IUD over 3 years, again indicating a dose-bounded effect.
• At-risk subgroups: Relevant only to heavy-menstrual-bleeding patients on a 52 mg device; no plausible oxygen-delivery effect proposed or measured for the lower-dose devices.
• Counter-evidence on record: Same active-comparator and dose-boundedness limitations as the Iron entry.

Regulatory status

Same Mirena/Liletta heavy-menstrual-bleeding indication and trial data as the Iron entry. Neither label reports specific hemoglobin change figures in the sections reviewed; hemoglobin is addressed only via the secondary literature below.

Population evidence

Imperato et al. 2002 (PMID 12063443) measured hemoglobin and hematocrit alongside ferritin in its 8-year, 223-woman prospective cohort (see Iron entry for full detail). Lowe and Prata 2013 (PMID 23122687), stratified to LNG-IUS, reviews hemoglobin specifically alongside ferritin; full abstract not retrievable this session (see Sources). Rybo, Andersson, and Odlind 1993 (PMID 8489751) is indexed as reporting that LNG-IUD reduced menstrual blood loss substantially, resulting in increased hemoglobin levels.

Mechanism

Indirect: less menstrual blood loss preserves red-cell mass and hemoglobin; this is a downstream consequence of the drug's approved bleeding-reduction effect (Hard Rule 2), not a direct action on erythropoiesis or arterial oxygenation.

Counter-evidence held on record

This entry shares its underlying evidence and mechanism with the Iron entry above (see Overlap map). It describes a bounded, indication-specific benefit, not a general "boosts oxygen delivery" claim, and carries no supplementation implication in either direction.