Blood Thinners and a GLP-1: Warfarin, Eliquis, Xarelto With Ozempic or Mounjaro

A GLP‑1 drug and a blood thinner can be taken together. No label forbids it. Warfarin has been tested with semaglutide, dulaglutide and exenatide, and none of them changed warfarin levels or INR in a way that mattered. The gaps are elsewhere. Tirzepatide (Mounjaro, Zepbound) has never been tested with warfarin, and no approved GLP‑1 has a published study with Eliquis, Xarelto, Pradaxa or Savaysa. The risks that are documented come from the side effects and the weight loss: eating less shifts vitamin K, vomiting and diarrhea cost doses and dry out the kidneys, and lower body weight can change the right DOAC dose.

What the GLP-1 labels say about blood thinners

None of the four labels names Eliquis, Xarelto, Pradaxa, Savaysa or clopidogrel.^[1][2][3][4] All four say the same basic thing: the drug slows the stomach, so it may change how other pills are absorbed. Past that, the wording differs.

Monitor patients on oral medications dependent on threshold concentrations for efficacy and those with a narrow therapeutic index (e.g., warfarin) when concomitantly administered with MOUNJARO.

— Mounjaro prescribing information, section 7.2

That sentence is the only place warfarin appears in the Mounjaro or Zepbound label.^[3][4] It does not say “check the INR” in those words, and it is not based on a warfarin study. It is a general caution, and warfarin is the example. The tirzepatide labels also note that the slowing of the stomach is largest after the first dose and fades over time. After a first 5 mg dose, the peak level of acetaminophen fell by about half and came an hour late; total exposure did not change.^[3][4]

Warfarin: four drugs tested, no meaningful change

Four GLP‑1 drugs have been given with a single dose of warfarin in healthy volunteers, and INR was measured.

The semaglutide injection study concluded that no dose adjustment should be needed.^[5] The tablet study found the same for warfarin.^[6] Dulaglutide moved the INR curve by 2%.^[7] Exenatide lowered the INR effect slightly, and the authors judged it no safety concern.^[8] Novo Nordisk ran both semaglutide studies and Lilly the other two.

The limits matter. Every one of these was a single warfarin dose in healthy people, not months of daily warfarin in someone with atrial fibrillation or a valve. A 2026 review of the labels and studies pointed out that no study has tested a narrow-therapeutic-index drug like warfarin at more than one point in a GLP‑1 schedule.^[9] So nobody has measured what happens in the first weeks, or at each dose step, when the stomach slows the most.

What changes the INR in practice: food, and the stomach

The warfarin label says plainly that diet changes INR. It asks patients to eat a normal, balanced diet with a steady amount of vitamin K and to avoid drastic changes in eating habits.^[13] A GLP‑1 is a drastic change in eating habits. Less food usually means less vitamin K, and less vitamin K pushes the INR up. The same label lists diarrhea and “poor nutritional state” as things that can raise the INR.^[13]

Real-world records show this effect, and it is small. In 1,021 warfarin patients, time in the target INR range fell from 64.2% to 62.1% after a GLP‑1 was started. Average INR did not move. Time above the range went up by 1.3 points.^[10] That study is covered in more detail in GLP-1 drugs and your other medications. The authors’ conclusion was closer INR monitoring after a GLP‑1 is started.^[10]

Two published tirzepatide cases show the INR can move either way:

• INR fell. A 66-year-old woman with a mechanical mitral valve started tirzepatide in hospital. Ten days later her INR had dropped sharply, though her diet was strict and stable. It took about three weeks of warfarin increases to get back to target.^[11] The authors suspected the slowed stomach.
• INR rose. A 64-year-old man on warfarin for a chronic clot had an INR of 8.7 shortly after his tirzepatide dose was raised. He also developed a rare autoimmune vasculitis with bleeding into the lungs, and he died. The authors wrote that causation was not clear.^[12]

Two cases prove nothing about how often this happens. They do show why the warfarin label asks for extra INR tests “whenever other medications are initiated, discontinued, or taken irregularly”.^[13] Starting a GLP‑1, raising its dose, missing weeks of it, and stopping it all count.

Eliquis, Xarelto, Pradaxa and Savaysa: not studied together

The DOACs do not use INR and are not affected by vitamin K, so the diet problem that dogs warfarin does not apply. But no approved GLP‑1 has a published interaction study with apixaban, rivaroxaban, dabigatran or edoxaban. A 2024 systematic review of GLP‑1 interaction studies found the usual pattern for the oral drugs that were tested: lower and later peaks, total exposure not changed in a way that mattered. It warned that the studies used healthy volunteers and that people with kidney disease or on narrow-therapeutic-index drugs were not covered.^[14]

Two pieces of evidence touch dabigatran (Pradaxa), and they point different ways:

• A computer model. A 2025 simulation applied GLP‑1-style stomach slowing to a virtual group of 1,000 adults with obesity. It predicted dabigatran exposure would rise by 205%, the largest change of any drug modeled.^[15] This is a prediction. Nobody has measured it in people.
• A real measurement, with the wrong drug. Pfizer tested dabigatran with lotiglipron, an oral GLP‑1 pill it later abandoned over liver problems. Peak dabigatran levels fell by up to 55%, and total exposure did not change significantly.^[16] Lotiglipron is not semaglutide or tirzepatide, and the result may not carry over.

The DOAC labels themselves show which drugs are sensitive to the things a GLP‑1 changes:

Sources for the table: the Eliquis, Xarelto, Pradaxa, Savaysa and warfarin labels.^{[17][18][19][20][13]}

Xarelto and eating less

Xarelto is the DOAC most tied to meals. Its label says the 15 mg and 20 mg tablets should be taken with food, and for atrial fibrillation, with the evening meal. Taken fasting, the 20 mg dose is about 66% absorbed; with food, total exposure rises 39%.^[18] The label also says absorption depends on where in the gut the tablet is released.^[18] Someone on a GLP‑1 who skips the evening meal, or eats a few bites, is not getting the condition the dose was set for. Nobody has measured what a slowed stomach does to Xarelto. The meal rule is on the label either way.

Eliquis, Savaysa and losing weight

Weight is written into two DOAC dosing rules. For atrial fibrillation, Eliquis drops from 5 mg to 2.5 mg twice daily in people with at least two of three traits, one of which is weighing 60 kg (about 132 lb) or less.^[17] For treating clots, Savaysa drops from 60 mg to 30 mg at 60 kg or less.^[20] Savaysa’s label also sets dosing by a kidney formula, Cockcroft-Gault, that has body weight in it, so losing weight lowers the number it gives even when the kidneys have not changed.^[20] Most people on a GLP‑1 will not cross 60 kg. Those who started near it can, and the dose was chosen for the old weight.

Vomiting, diarrhea and missed doses

Stomach side effects are common. On Wegovy 2.4 mg, 24% of people vomited against 6% on placebo, and 30% had diarrhea against 16%.^[2] On Zepbound, vomiting ran 8% to 13% by dose against 2%, and diarrhea 19% to 23% against 8%.^[4] For someone on a blood thinner, that creates two problems.

The first is the dose itself. The adult labels for Eliquis and Savaysa say a missed dose should be taken the same day, and not doubled.^[17][20] Pradaxa’s label says to skip the missed dose if it cannot be taken at least six hours before the next one.^[19] Xarelto’s rules depend on the dose schedule.^[18] None of the adult labels says what to do when a tablet comes back up. That is a question for a pharmacist, not a guess. Each DOAC label carries a boxed warning that stopping early raises the risk of clots or strokes;^{[17][18][19][20]} a run of lost doses works the same way.

The second is the kidneys. The Wegovy and Zepbound labels report kidney injury, sometimes needing dialysis, mostly in people dehydrated by nausea, vomiting or diarrhea, and ask for kidney function to be checked in that situation, especially when starting or raising the dose.^[2][4] Dabigatran leans on the kidneys more than any other DOAC: 80% of it is cleared there.^[19] Pradaxa’s label asks for kidney function to be checked more often when it might be falling, and to stop the drug if the kidneys fail acutely.^[19] Xarelto exposure rose 44% to 64% in people with impaired kidneys.^[18] More on GLP‑1 drugs and kidney numbers is in GLP-1 drugs and your kidneys, and on managing the diarrhea in GLP-1 diarrhea: what helps.

Clopidogrel (Plavix) and other antiplatelets

No study was found that measured clopidogrel levels or platelet blocking in people taking a GLP‑1. The closest data on the Plavix label is about food, which also slows absorption: taken with breakfast, the active form peaked 57% lower, total exposure was unchanged, and platelet blocking fell by less than 9%.^[21] That suggests a slower stomach mostly blunts the peak. It is an inference, not a test.

GLP‑1 drugs do not thin the blood on their own in any meaningful way. A 16-person pilot trial of liraglutide found a small, short-lived drop in platelet clumping in the first week that faded after that, and the authors called its significance unknown.^[22] The Plavix label warns that stopping clopidogrel raises the risk of heart events and says to restart it as soon as possible after surgery.^[21]

Before endoscopy or surgery: two drugs to plan around

A procedure is where the two drugs collide on the calendar. The GLP‑1 labels warn about food left in the stomach under anesthesia and ask patients to tell the care team before any planned procedure; they also say the data are too thin to say whether pausing the drug helps.^[2][3] The blood thinner labels give their own stop times. These are label wording, not a plan; the person doing the procedure sets the actual timing.

Sources: the Eliquis, Xarelto, Pradaxa, Savaysa, warfarin and Plavix labels.^{[17][18][19][20][13][21]} Pradaxa’s timing depends on kidney function, which a bout of GLP‑1 dehydration can change. The evidence on holding the GLP‑1 itself, including a randomized trial, is in GLP-1 drugs before endoscopy and surgery.

What to ask

1. On warfarin: when should my INR be checked? Ask for a plan that covers the first weeks, each dose increase, and stopping.
2. On Xarelto: what if I can’t eat a meal with it? The 15 mg and 20 mg tablets are meant to go down with food.
3. On Eliquis or Savaysa: does my dose still fit my weight? Both labels use 60 kg as a cutoff in some uses.
4. On Pradaxa: how will my kidneys be watched? Especially after a stretch of vomiting or diarrhea.
5. What do I do if I vomit a dose? Get the answer from a pharmacist before it happens.
6. Who coordinates a procedure? Make sure the doctor doing it knows about both drugs.

For how a GLP‑1 interacts with other common prescriptions, see levothyroxine and a GLP-1 and antidepressants and a GLP-1.