Lithium

Limits on what this shows

• Key negative/null findings: The review is explicit that this is a distinct, more often multiglandular disease pattern rather than sporadic hyperparathyroidism incidentally co-occurring in lithium users, which argues against pure confounding by age or indication alone.
• At-risk subgroups: Longer duration of lithium exposure and older age track with higher prevalence in the reviewed literature; women (as in sporadic hyperparathyroidism generally) are overrepresented in surgical series.
• Counter-evidence on record: None identified that contradicts the direction of effect; the review notes prevalence estimates vary with case-finding intensity (routine calcium screening vs. symptom-triggered testing) across the pooled studies, which affects the precision of the 7.2%/4.8% figures but not their direction.

Regulatory status

FDA label, Section 5.8, names persistent hyperparathyroidism and hypercalcemia as an established long-term risk requiring periodic calcium monitoring.

Population evidence

2026 systematic review (PRISMA), 125 included studies, cohorts/case series/surgical series: hypercalcemia in 7.2% (423/5,853) of lithium-treated patients across studies; confirmed hyperparathyroidism in 4.8% (156/3,239) where PTH was measured; multiglandular parathyroid disease in 25-83% of surgical cases versus 15-20% in sporadic (non-lithium) primary hyperparathyroidism.

Mechanism

Two convergent pathways: (1) lithium produces a rightward shift in the calcium-PTH set point at the calcium-sensing receptor, requiring higher serum calcium to suppress PTH secretion; (2) lithium inhibits GSK-3β, driving Wnt/β-catenin-mediated parathyroid cell proliferation and reduced apoptosis.

Counter-evidence held on record

This is an excess/retention finding, not a deficiency. Calcium or vitamin D supplementation is not an appropriate response to lithium use and could worsen hypercalcemia in a patient who already has lithium-associated parathyroid overactivity. The correct response is periodic serum calcium and PTH monitoring, not calcium repletion.

Limits on what this shows

• Key negative/null findings: Not applicable in the sense of a null study; the finding is a labeled bidirectional mechanism rather than a single-arm deficiency claim.
• At-risk subgroups: Patients starting a low-sodium diet, using diuretics (especially thiazides) or NSAIDs, experiencing protracted sweating, vomiting, or diarrhea, and the elderly with reduced renal reserve.
• Counter-evidence on record: None found contradicting the direction; the main limitation is source count, not conflicting data.

Regulatory status

FDA label, Section 5.3: "Lithium can cause hyponatremia by decreasing sodium reabsorption by the renal tubules, leading to sodium depletion." Section 2/5.1 instructs patients to maintain a normal diet including salt and 2,500-3,000 mL fluid intake, especially during initial stabilization.

Population evidence

Regulatory-label-grade evidence (synthesized clinical trial and post-marketing safety data) rather than a single named cohort study; bounded here to ORANGE because this review could confirm only one independently opened Tier A source (the label itself) rather than two, despite the relationship being long-established in pharmacology references.

Mechanism

Bidirectional and dilutional/hormonal (Step 2 mechanism F): lithium itself promotes natriuresis via the collecting duct; independently, reduced sodium intake or volume depletion increases proximal tubular sodium (and co-transported lithium) reabsorption, raising serum lithium. NSAIDs, ACE inhibitors/ARBs, and diuretics compound the clearance side of this loop.

Counter-evidence held on record

This is not a sodium deficiency to correct with supplementation. The clinically important action is maintaining consistent (not increased) dietary sodium and fluid intake and avoiding abrupt low-sodium diets, dehydration, or new diuretic/NSAID use without checking a lithium level - supplementing sodium or restricting it without medical guidance are both hazards here, not remedies.

Limits on what this shows

• Key negative/null findings: Frank B12-deficiency prevalence did not differ significantly between groups (a nonsignificant trend at one center only); the same study found no significant folate difference between groups (see Fails verification, folate).
• At-risk subgroups: Not defined by dose or duration in the source study; the authors note the clinical significance of the serum difference "is not yet clear."
• Counter-evidence on record: No second independent study was located confirming or contradicting this finding despite a multi-channel search (PubMed, Europe PMC, Semantic Scholar attempted and rate-limited); this is disclosed as a single-study finding, not replicated.

Regulatory status

Not addressed in the FDA label.

Population evidence

Cervantes, Ghadirian & Vida, Biological Psychiatry 1999: cross-sectional comparison at two mood-disorder clinics (Allan Memorial Institute; Montreal General Hospital), lithium-treated vs. non-lithium patients. Mean serum B12 approximately 20% lower in the lithium group at each center; statistically significant at each site and combined (two-tailed p =.017,.021, and.0009).

Mechanism

Not established. Lithium is not metabolized, does not chelate B12, and has no documented effect on intrinsic factor or ileal cobalamin transport; the authors describe the mechanism as unclear.

Counter-evidence held on record

This supports periodic B12 monitoring in long-term lithium users, not a standing supplementation recommendation. No dose is specified here; B12 has no established tolerable upper intake level, but any repletion decision belongs with a prescriber, informed by an actual measured level.

Limits on what this shows

• Key negative/null findings: The interaction is specific to pharmacological/supplemental iodine doses (e.g., potassium iodide, high-dose kelp); ordinary dietary iodine intake is not documented as posing the same risk.
• At-risk subgroups: Patients taking iodine-containing supplements, kelp/seaweed supplements, or iodinated contrast/amiodarone concurrently; pre-existing subclinical hypothyroidism.
• Counter-evidence on record: Paradoxically, some lithium users develop hyperthyroidism (Graves' disease, toxic multinodular goiter, silent thyroiditis) rather than hypothyroidism, per the FDA label - the endocrine effect is not uniformly suppressive in every patient.

Regulatory status

FDA label, Sections 5.7 and 7.1: lithium "is concentrated within the thyroid and can inhibit thyroid synthesis and release," with monitoring required before treatment, at three months, and every 6-12 months thereafter; iodide preparations are listed as a named interaction ("concomitant use may produce hypothyroidism"), example given potassium iodide.

Population evidence

Linus Pauling Institute (Oregon State University) drug-nutrient interaction reference independently confirms: "concomitant use of lithium and pharmacological doses of potassium iodide may increase the risk of hypothyroidism due to inhibition of thyroid hormone synthesis (i.e., Wolff-Chaikoff effect)," recommending thyroid-status monitoring.

Mechanism

Direction check applied (Hard Rule 4): this is iodine intake affecting the drug's known adverse endocrine effect, not lithium depleting iodine status. Both agents independently suppress thyroid hormone synthesis; combined exposure is additive at pharmacological iodine doses.

Counter-evidence held on record

Not a call to increase iodine intake. The actionable guidance runs the opposite direction: avoid high-dose iodine or kelp supplements while on lithium unless a prescriber directs otherwise, and maintain routine TSH/free T4 monitoring regardless.

Limits on what this shows

• Key negative/null findings: Both controlled, drug-specific, chronic-treatment studies located (Dunner 1975, Makara-Studzińska 2013) found no significant magnesium disturbance, directly bounding the label's unquantified hypermagnesemia listing. Nielsen J, Acta Psychiatrica Scandinavica 1964 (40:2, 190-196, DOI 10.1111/j.1600-0447.1964.tb05745.x, PMID 14235255) is the specific earlier report Dunner et al. contrast their null finding against, but this 1964 paper predates the routine indexing of abstracts and its full findings could not be verified during this review (see Sources); it is disclosed by bibliographic record only and not used as evidentiary support here.
• At-risk subgroups: Not defined by the label or by either controlled study. The acute-versus-chronic contrast Dunner et al. draw suggests early treatment, if the phenomenon is real at all, rather than long-term maintenance therapy.
• Counter-evidence on record: Two independent controlled chronic-treatment studies found no significant magnesium effect, which is why this is graded orange (a real, drug-specific, regulatory-labeled finding) rather than red (magnitude and reproducibility are not established, and the best controlled data lean toward no effect at steady state).

Regulatory status

FDA label, Section 6 (Adverse Reactions), lists "hypermagnesemia" among miscellaneous adverse reactions reported with lithium carbonate, alongside leukocytosis and transient hyperglycemia. No frequency, magnitude, dose, or exposure-duration data are given.

Population evidence

Dunner DL, et al., Acta Psychiatrica Scandinavica 1975 (51:2, 104-109, DOI 10.1111/j.1600-0447.1975.tb00219.x): 79 outpatients with bipolar or unipolar primary affective disorder on chronic lithium carbonate versus 19 placebo controls. Plasma and erythrocyte magnesium showed only minor, non-statistically-significant differences by treatment, diagnosis, or sex. The authors explicitly noted this contrasts with earlier reports of magnesium increases during acute lithium therapy, and concluded steady-state magnesium is maintained during prolonged treatment. Makara-Studzińska et al. 2013 (already cited in this dossier under the Ungraded section for a different mineral context; Annals of Agricultural and Environmental Medicine, 37 hospitalized lithium-treated bipolar inpatients, 8-week observation) likewise found no significant relationship between plasma lithium level and magnesium status.

Mechanism

Not established for this drug specifically in the sources opened. A plausible renal-tubular mechanism is proposed by analogy to lithium's well-documented effect on tubular sodium handling (see Sodium and water balance, above), but no source directly measured or confirmed a magnesium-specific renal transport effect.

Counter-evidence held on record

This is an excess-risk listing, not a deficiency to correct. Magnesium supplementation is not supported by this entry; nor is reducing dietary magnesium, given the null findings in both controlled chronic-treatment studies. No action beyond the routine electrolyte monitoring lithium already requires is indicated.

Limits on what this shows

• Key negative/null findings: The review is explicit that the inositol-depletion hypothesis "remains controversial and unsupported by some findings"; effects are brain-region-specific (documented in hypothalamus/cortex, not shown as a systemic or serum-level depletion) and are drawn almost entirely from animal models rather than confirmed human peripheral deficiency.
• At-risk subgroups: Not established in humans; this is a CNS tissue-level finding, not a population-risk finding.
• Counter-evidence on record: No systemic (serum) inositol deficiency has been demonstrated in lithium-treated patients; the effect as documented is localized to brain tissue in animal models.

Regulatory status

Not addressed in the FDA label.

Population evidence

Yu & Greenberg, Future Neurology 2016 (review, DOI 10.2217/fnl-2016-0003): rat cerebral cortex myo-inositol fell 30% by 6 hours after lithium injection, persisting to 24 hours; other animal studies show 22-25% decreases in lithium-treated mouse brain inositol (33-37% in inositol-transporter-knockout mice). Human evidence is limited to magnetic resonance spectroscopy studies showing altered (not consistently directional) brain myo-inositol in living bipolar patients.

Mechanism

Direct IMPase inhibition reducing inositol recycling in the phosphatidylinositol signaling cycle, proposed since the 1980s as (part of) lithium's mood-stabilizing mechanism of action.

Counter-evidence held on record

This is not a deficiency to supplement against. If brain inositol reduction is part of how lithium works therapeutically, inositol supplementation is a theoretical mechanism-antagonism risk, not a repletion opportunity - there is no dosing recommendation here in either direction, and any inositol supplement use in a lithium patient should be discussed with the prescriber first.

Limits on what this shows

• Key negative/null findings: The authors explicitly report "no significant relationship... between lithium/valproate dosing and trace element concentrations" - i.e., the study's own data argue against lithium as the cause (confounding by indication, per Hard Rule 6, is not excluded here; if anything it is supported).
• At-risk subgroups: Not defined by lithium exposure; applies to bipolar disorder as a diagnosis in this dataset.
• Counter-evidence on record: The explicit absence of a dose-relationship is itself the strongest counter-evidence against drug attribution.

Regulatory status

Not addressed in the FDA label.

Population evidence

Rog et al., Frontiers in Molecular Neuroscience 2024 (17:1441575): 62 bipolar disorder patients (treated with lithium and/or valproic acid within therapeutic range) vs. 40 healthy controls. Serum zinc lower in BD patients (6,186.85 vs. 6,604.85 µg/L, p = 0.014); selenium and arsenic also lower.

Mechanism

Not drug-specific; proposed as disease-associated oxidative-stress-linked trace-element handling rather than a lithium pharmacological effect.

Counter-evidence held on record

Framed as a bipolar-disorder-population context flag worth monitoring, not a lithium-caused deficiency. No dosing recommendation is made; zinc's tolerable upper intake level (40 mg/day, adults, NIH ODS) is noted for reference only, not as a suggested dose.