Most of the interaction questions on this site are about a tablet arriving late. This one is not. Lithium is an element, not a molecule assembled by a liver, and its concentration in blood is governed by how much water and sodium a body is carrying. The list of things that push a lithium level upward reads like the list of things a GLP-1 receptor agonist does in its first weeks: eating less, drinking less, vomiting, diarrhea. The drugs have never been studied together. The label of the older one already predicts the collision.
How little room there is
The extended-release lithium label opens with a boxed warning stating that toxicity is closely related to serum lithium levels and can occur at doses close to therapeutic levels, and that facilities for prompt and accurate serum determinations should be available before treatment begins.[1] The numbers behind that sentence are the whole problem: toxic concentrations of 1.5 mEq/L or above sit against a therapeutic range of 0.8 to 1.2, and the same section adds that some patients show toxic signs at concentrations considered to be within the therapeutic range.[1]
Three tenths of a milliequivalent is the entire margin, and what fills it is not drug metabolism. The label’s clinical pharmacology section explains why: lithium is distributed through total body water, excreted almost entirely by the kidney in proportion to its plasma concentration, and it decreases sodium reabsorption by the renal tubules in a way that can itself lead to sodium depletion. The instruction that follows is dietary — maintain a normal diet including salt, and an adequate fluid intake of 2,500 to 3,500 mL.[2] A drug that changes how much a person eats and drinks is operating directly on that instruction.
The label’s own list of what raises a level
The lithium label does not leave the mechanism to inference. Under Lithium Toxicity it prints a bulleted list of what increases the risk, and two of the bullets are volume depletion or dehydration and changes in electrolyte concentrations, especially sodium and potassium.[1] A further paragraph is more specific still: decreased tolerance to lithium has been reported to follow protracted sweating or diarrhea, and where that happens the label calls for supplemental fluid and salt under medical supervision.[2]
Now read the other label. Semaglutide carries a warning headed Acute Kidney Injury Due to Volume Depletion, which records postmarketing reports of acute kidney injury, in some cases requiring hemodialysis, and states that the majority of reported events occurred in patients who experienced gastrointestinal adverse reactions leading to dehydration such as nausea, vomiting, or diarrhea.[3] Tirzepatide carries the same warning in nearly the same words.[4] Neither manufacturer was writing about lithium. They were describing the exact physiological event the lithium label names as a risk factor, in the population most likely to experience it, during the weeks the dose is being escalated. The class-wide version of that risk is the subject of the dehydration article.
Six labels, no mention
Given all that, the reasonable expectation is that somebody wrote it down. Nobody did. Searched as documents this week, the prescribing information for both tirzepatide products, both semaglutide injections, the semaglutide tablet and liraglutide contains the word “lithium” a combined total of zero times. There is no interaction entry, no monitoring note, and no mention in any warnings section.
Nor is there a study. A 2026 case report states the position plainly: no pharmacokinetic studies have been published evaluating this potential drug-drug interaction.[5] That is a stronger claim than it looks. Manufacturers ran interaction trials against warfarin, digoxin, metformin, atorvastatin, lisinopril, rosuvastatin, furosemide and a combined oral contraceptive — the whole apparatus described in the oral medications article. Lithium was not on the list, because these were diabetes programs and lithium is a psychiatric drug.
What the four published reports describe
The human evidence for this interaction is four papers, and it is worth being exact about their size. A 2025 case series from a psychiatric pharmacy group reports three patients on stable lithium regimens who started semaglutide. In two, levels rose significantly and produced toxicity despite stable renal function and no change in concurrent medications. In the third, the lithium dose was reduced before the semaglutide started, and toxicity was avoided — but levels still came in higher than expected.[6]
The fourth report is the most specific, because the patient changed only one variable. A 23-year-old man with schizoaffective disorder had been titrated to lithium 1,800 mg nightly while already taking semaglutide, with a level of 0.9 mEq/L. Four days after his first dose of tirzepatide he reported early symptoms of toxicity, which progressed after the second dose; a level drawn fifteen hours after a dose came back at 1.7 mEq/L, with no change in renal function, hydration status or other medications. Intravenous fluids were enough to manage it, and a lithium dose reduction prevented recurrence. The authors scored it on the Drug Interaction Probability Scale as a probable interaction.[5] A fourth publication, a 2025 letter in the same psychopharmacology journal, reports semaglutide-mediated lithium toxicity under that title.[7]
Four reports is not an incidence. It is a signal that something happens to some people, in a pairing common enough that both drugs are prescribed to overlapping populations at scale. Psychiatric medication-related weight gain is one of the more common reasons a person on lithium is offered a GLP-1 in the first place, which is a reason to expect the pairing to grow rather than to stay rare.
Nobody agrees on which mechanism it is
Three explanations are on the table and no study has separated them. The case series lists altered kidney function, dehydration from reduced oral intake, vomiting or diarrhea, and delayed gastric emptying as candidate contributors.[6] The tirzepatide case adds a fourth and more surprising one: that tirzepatide may enhance the absorption of lithium, which runs against the intuition that a slower stomach means less drug.[5]
That inversion should not be surprising to anyone who has read the rest of this class’s interaction data, where the significant changes in total exposure have generally been increases rather than decreases — the pattern described for thyroid replacement. But lithium is not absorbed like a tablet of thyroxine, and an extended- release lithium formulation is designed specifically to blunt its own peak. A stomach that holds it longer is a plausible way to defeat that design. Plausible is as far as any source here goes.
The largest dataset points the opposite way
One finding has to be reported alongside the case reports or the picture is false. A 2026 propensity-matched cohort drawn from a multinational electronic health record network examined adults with type 2 diabetes who had lithium exposure, and compared those who started a GLP-1 receptor agonist against those who did not. Across 452 matched pairs over 24 months, renal events occurred in 23.5% of the GLP-1 group against 44.7% of the comparator, a risk ratio of 0.53 (95% CI, 0.30 to 0.91) and a hazard ratio of 0.45 (0.23 to 0.85).[8] A companion analysis of spontaneous adverse event reports found lithium associated with elevated reporting of every renal outcome examined, while semaglutide’s associations ran inverse and none of them met the threshold to count as a signal.[9]
Both things can be true, because they are measured on different clocks. Lithium’s long-term renal risk is a chronic kidney story unfolding over years, and it is exactly the kind of story this drug class has improved elsewhere — the subject of the kidney outcomes article. An acute level climbing from 0.9 to 1.7 in four days is a different event with a different cause. A drug can plausibly protect a kidney over two years and disturb a level over two weeks, and neither of those findings cancels the other.
The gap this leaves for a cash-pay buyer
Every author cited above asks for the same thing, and it is not a dose change. The case series calls for baseline renal function, hydration status and lithium levels to be assessed before semaglutide is started and for levels to be monitored more frequently during therapy.[6] The tirzepatide report asks for increased monitoring when one of these drugs is started, when a dose changes, and when a patient switches from one to another.[5] None of that is possible unless somebody knows both prescriptions exist.
That is the structural problem rather than a pharmacological one. A lithium level is ordered by the clinician managing the bipolar disorder; the GLP-1 is often bought from a telehealth seller who has no contact with that clinician and whose intake form may or may not ask. What those forms cover, and how much they vary, is set out in the intake article. A compounded vial also arrives with no FDA-reviewed prescribing information at all, because compounded semaglutide and tirzepatide are not approved by the FDA and are not reviewed by the agency for safety, effectiveness or quality before dispensing — so even the general warning about volume depletion is not in the package. Nothing on this page is medical advice or a reason to change any dose; how claims here are established is set out in the methodology.