Skip to content
GLP Loss
← Research
Safety

GLP-1s and Sulfonylureas: The Interaction That Reaches the Hospital

Hypoglycemia below 54 mg/dL ran 13.8%, 9.9% and 12.8% across three tirzepatide doses — flat. Split the same trial by whether a sulfonylurea was present and it runs 10.3% against 2.1%. The variable is the tablet, not the injection.

Owen Castellanos9 min read
The rate tracks the other drugGlucose below 54 mg/dL in the tirzepatide label trialsWith a sulfonylurea, by tirzepatide dose5 mg: 13.8% — 10 mg: 9.9% — 15 mg: 12.8%Same weight-loss trial, split by companion drugWith a sulfonylurea 10.3%; without one 2.1%Sulfonylureas against low-risk comparatorsHypoglycemia odds ratio 5.24; 31 head-to-head trialsTitrating the injection more slowly does not address this.The labeled instruction is to reduce the tablet, which a different prescriber wrote.

Glipizide, glimepiride and glyburide are among the oldest and cheapest diabetes drugs still in wide use. They force the pancreas to release insulin whether or not glucose is elevated, which is precisely why they work and precisely why they cause hypoglycemia. Put one alongside a drug that also lowers glucose and reduces food intake, and the result is the interaction in this class most likely to put somebody in an ambulance — not because the GLP-1 is dangerous, but because the tablet already was.

The instruction names a phrase patients do not use

The semaglutide label devotes a warnings section to hypoglycemia with concomitant use of insulin secretagogues or insulin, and states that the risk may be lowered by a reduction in the dose of sulfonylurea or other concomitantly administered insulin secretagogue. Its drug interactions section adds that when initiating treatment, prescribers should consider reducing the dose of the secretagogue.[1] The tirzepatide labels carry equivalent language.[2]

The phrase doing the work there is insulin secretagogue. A person taking a small white tablet for diabetes will answer no to a question about insulin, truthfully, and will not recognize the category their medicine belongs to. The labeled instruction is written in pharmacology; the intake form has to be written in brand names.

The rate does not follow the dose being titrated

In the tirzepatide diabetes program, hypoglycemia at a glucose level below 54 mg/dL among patients also taking a sulfonylurea occurred in 13.8%, 9.9% and 12.8% of patients on 5 mg, 10 mg and 15 mg respectively, with severe hypoglycemia in 0.5%, 0% and 0.6%.[2] Across a threefold range of the drug being escalated, the rate is flat, and the middle dose is the lowest of the three.

The comparison that does move is the companion drug. In the weight-management trial of the same molecule in people with type 2 diabetes, hypoglycemia below 54 mg/dL was reported in 4.2% of treated patients against 1.3% on placebo — but within the treated arm, those taking a sulfonylurea ran 10.3% against 2.1% for those who were not.[3] Same drug, same trial, same doses; one variable separates a 2% rate from a 10% one.

This is why slowing down the escalation schedule is the wrong lever here, even though it is the lever most cash subscriptions actually offer. Nausea and vomiting track the injection dose; secretagogue hypoglycemia tracks the tablet.

What the semaglutide numbers add

The semaglutide label reports its own sulfonylurea subgroup separately. Severe hypoglycemia occurred in 0.8% and 1.2% of patients when 0.5 mg and 1 mg were coadministered with a sulfonylurea, documented symptomatic hypoglycemia in 17.3% and 24.4%, and severe or blood-glucose-confirmed symptomatic hypoglycemia in 6.5% and 10.4%.[1] Severe hypoglycemia, in that label's definition, is an episode requiring the assistance of another person — not a number on a meter but an event with a witness.

Those percentages sit alongside a monotherapy row where severe hypoglycemia was 0% at every dose. The distance between those two rows is the entire subject of this page.

The swallowed version of the same molecule was tested in a mixed background population and lands in between. In a 26-week trial of the semaglutide tablet added to metformin and/or a sulfonylurea, or to basal insulin, in patients with moderate renal impairment, severe hypoglycemia was 0% in both the placebo arm of 161 and the 14 mg arm of 163, while plasma glucose below 54 mg/dL was recorded in 3% and 6% respectively.[9] A doubled rate of a laboratory threshold with no severe events on either side is a different finding from the sulfonylurea-only subgroup, and the reason is that the background regimens were pooled.

The pharmacology explains why the stack is uneven

The semaglutide label states the mechanism in one line: the drug stimulates insulin release in the presence of elevated blood glucose concentrations.[1] That glucose dependence is why these drugs produce almost no hypoglycemia on their own — as the glucose falls, the stimulus to secrete insulin falls with it.

A sulfonylurea has no such brake. It closes a potassium channel on the beta cell and insulin follows regardless of what the glucose is doing, which is why the same tablet that works at a blood sugar of 250 keeps working at 55. Combining a glucose-dependent drug with a glucose-independent one does not produce two equal contributions to the risk. It produces one drug that lowers the starting point and a second that keeps pushing after the floor is reached.

The tablet was the higher-risk drug before the injection arrived

A systematic review of 31 randomized head-to-head trials, covering 26,204 patients, compared sulfonylureas against comparators with low hypoglycemic potential — metformin, DPP-4 inhibitors, SGLT2 inhibitors and GLP-1 receptor agonists. Sulfonylureas carried an odds ratio of 5.24 (95% CI, 4.20 to 6.55) for hypoglycemia, 1.32 (95% CI, 1.07 to 1.61) for major adverse cardiovascular events, 1.67 (95% CI, 1.17 to 2.38) for myocardial infarction, and 1.32 (95% CI, 1.00 to 1.75) for all-cause mortality.[4] The review also reported differences within the class, with an increased mortality risk seen for glipizide but not the other molecules.

That is a baseline, not an interaction. It means the person who arrives at a weight-loss intake already on a sulfonylurea was, before any of this, in the group a five-fold hypoglycemia odds ratio describes.

The subgroup where the difference between drug classes is largest

A cohort study of Medicare beneficiaries over 65 with type 2 diabetes compared initiators of newer second-line agents and stratified by what they were already taking. Overall, SGLT2 inhibitors carried a hazard ratio of 0.90 (95% CI, 0.82 to 0.98) for severe hypoglycemia requiring an emergency or inpatient visit against GLP-1 receptor agonists — a small difference. Against DPP-4 inhibitors the overall hazard ratio was 0.75 (95% CI, 0.68 to 0.83).[5]

Then the stratification. Among patients already using a sulfonylurea, that SGLT2-versus-DPP-4 comparison widened to a hazard ratio of 0.57 (95% CI, 0.49 to 0.65) with a rate difference of 6.80 fewer events per 1,000 person-years, while among patients not on one the association was near null. The authors report the same pattern for the GLP-1 comparison.[5] The choice between second-line drugs barely matters for hypoglycemia until a secretagogue is in the background, at which point it matters a great deal — a stratification consistent with the age-specific picture in the literature on adults over 65.

Where these events actually end up

Severe hypoglycemia is one of the few adverse events in this field that reliably produces a hospital record. National surveillance of emergency hospitalizations in United States adults aged 65 and over estimated 99,628 such admissions per year for adverse drug events, of which four medications or classes accounted for 67.0%: oral hypoglycemic agents were implicated in 10.7%, insulins in 13.9%.[6] A later survey of emergency department visits across all ages found anticoagulants, antibiotics and diabetes agents implicated in 46.9% of visits for adverse drug events, with insulin and four oral diabetes agents among the 15 most commonly implicated drugs in older adults.[7]

Post-marketing reports specific to this drug class point the same way without being able to give a rate. A disproportionality analysis of the FDA Adverse Event Reporting System identified 1,164 cases of hypoglycemia associated with GLP-1 receptor agonists, of which 56.08% involved hospitalization and 3.53% were fatal, with a median time to onset of 5 days (interquartile range 0 to 67.75).[8] Spontaneous reports cannot establish incidence — the denominator is unknown and reporting is voluntary — but the onset figure is informative on its own. The median case appeared in the first week, not after months of escalation.

What the labels leave to somebody else

As with the insulin half of the same sentence, covered in the insulin article, the labels instruct a reduction without naming a size. Unlike insulin, a sulfonylurea has no units to halve and no titration log; it is one tablet strength, refilled on autopilot, frequently by a primary care practice that has no idea a weight-loss subscription has started.

Three facts have to meet for the labeled instruction to be executed: the sulfonylurea has to be named on the intake, somebody has to recognize the brand as a secretagogue, and a prescriber somewhere has to change its dose. A questionnaire asking whether the buyer takes insulin will return a correct no from a person taking glimepiride 4 mg twice a day. The wider pattern of what these forms do and do not capture is set out in the telehealth intake article, and the standing exclusion list in who should not take these drugs.

A note on the specific products this market sells. Compounded semaglutide and tirzepatide have not been approved by the FDA, and the agency does not evaluate their safety, efficacy or quality before a pharmacy dispenses them. The glucose-lowering effect that stacks with a sulfonylurea is a property of the molecule and does not care which pharmacy filled the vial — but the warnings section that names the interaction belongs to the branded labels, and the compounded product ships without one, a distinction explored in what that status actually means.

Frequently asked

Which diabetes pills are sulfonylureas?
The ones still commonly prescribed in the United States are glipizide, glimepiride and glyburide, sold under names including Glucotrol, Amaryl, DiaBeta and Glynase. They are inexpensive generics, often prescribed years before a weight-loss drug enters the picture. The labels for semaglutide and tirzepatide call them insulin secretagogues, which is not a phrase most people recognize as describing their own tablet.
Does a lower GLP-1 dose reduce this risk?
The label data suggest not. In the tirzepatide diabetes program, hypoglycemia below 54 mg/dL among patients on a concomitant sulfonylurea ran 13.8%, 9.9% and 12.8% at 5 mg, 10 mg and 15 mg — flat across a threefold dose range, with the middle dose lowest. The rate that moves is the one split by whether a sulfonylurea is present at all: 10.3% against 2.1% within the same weight-loss trial.
What do the labels tell prescribers to do?
Both the semaglutide and tirzepatide labels instruct prescribers to consider reducing the dose of a concomitantly administered insulin secretagogue when initiating treatment, and to monitor blood glucose. Neither states how large that reduction should be. The instruction is addressed to whoever controls the sulfonylurea prescription, which at a cash-pay weight-loss subscription is usually a different clinician entirely.
How serious does sulfonylurea hypoglycemia get?
Severe hypoglycemia is defined on these labels as an episode requiring the assistance of another person, and it was reported in 0.8% and 1.2% of patients taking semaglutide 0.5 mg and 1 mg with a sulfonylurea. At population scale, oral hypoglycemic agents were implicated in 10.7% of an estimated 99,628 annual emergency hospitalizations for adverse drug events among United States adults aged 65 and over.
When in treatment would a problem show up?
Earlier than most people expect. A disproportionality analysis of 1,164 GLP-1-associated hypoglycemia reports in the FDA Adverse Event Reporting System found a median time to onset of 5 days, with an interquartile range of 0 to 67.75 days. Spontaneous reports cannot give an incidence rate, but they place the typical case in the first week rather than at the top of a titration schedule.

Sources

  1. [1] Novo Nordisk Pharmaceutical Industries, LP (2026). OZEMPIC (semaglutide) injection, solution — Warnings and Precautions 5.5 and Adverse Reactions 6.1: Hypoglycemia with a sulfonylurea DailyMed, U.S. National Library of Medicine. Source
  2. [2] Eli Lilly and Company (2026). MOUNJARO (tirzepatide) injection, solution — Warnings and Precautions 5.3, Adverse Reactions 6.1 and Drug Interactions 7.1 DailyMed, U.S. National Library of Medicine. Source
  3. [3] Eli Lilly and Company (2026). ZEPBOUND (tirzepatide) injection — Adverse Reactions 6.1: Hypoglycemia in patients with type 2 diabetes mellitus DailyMed, U.S. National Library of Medicine. Source
  4. [4] Volke V, Katus U, Johannson A, et al. (2022). Systematic review and meta-analysis of head-to-head trials comparing sulfonylureas and low hypoglycaemic risk antidiabetic drugs. BMC Endocr Disord. PMID 36261824
  5. [5] Htoo PT, Paik JM, Alt E, et al. (2023). Risk of Severe Hypoglycemia With Newer Second-line Glucose-lowering Medications in Older Adults With Type 2 Diabetes Stratified by Known Indicators of Hypoglycemia Risk. J Gerontol A Biol Sci Med Sci. PMID 36866496
  6. [6] Budnitz DS, Lovegrove MC, Shehab N, Richards CL (2011). Emergency hospitalizations for adverse drug events in older Americans. N Engl J Med. PMID 22111719
  7. [7] Shehab N, Lovegrove MC, Geller AI, et al. (2016). US Emergency Department Visits for Outpatient Adverse Drug Events, 2013-2014. JAMA. PMID 27893129
  8. [8] Zhao Z, Tang Y, Hu Y, et al. (2021). Hypoglycemia following the use of glucagon-like peptide-1 receptor agonists: a real-world analysis of post-marketing surveillance data. Ann Transl Med. PMID 34734034
  9. [9] Novo Nordisk Pharmaceutical Industries, LP (2026). RYBELSUS (semaglutide) tablets; OZEMPIC (semaglutide) tablets — Warnings and Precautions 5.4 and Adverse Reactions 6.1: Hypoglycemia DailyMed, U.S. National Library of Medicine. Source

More in Safety