Sitagliptin and DPP-4 Inhibitors: Independent Evidence & Funding

Table of contents

• Evidence summary
• What sitagliptin is
• How it works
• What it is prescribed for
• What works and what does not
• Benefits by claim
• Risks and all side effects
• All interactions
• Who should avoid sitagliptin
• Dosage and how to take it
• Follow the money: who makes it and who funded the evidence
• Related research
• Frequently asked questions
• Sources and funding notes

Evidence summary

Independent evidence and credibility scorecard

Independence tiers: 1 = regulator or government body; 2 = academic, Cochrane or professional-society team with no declared commercial funding for the work (or only donated supplies); 3 = independent authors with declared industry funding from a company with a stake in the result; 4 = manufacturer-funded or manufacturer-authored. Credibility runs from A (highest) to D.

Our sources cluster in the UK and US (ten of twelve rows involve one or both), so the verdict mainly reflects Anglo-American regulators and research groups. The pivotal outcome trials were all paid for by the companies selling the drugs. The strongest signal is convergence. The publicly funded GRADE trial and the Cochrane reviews agree with the maker-funded trials that gliptins are heart-neutral, rarely cause hypos and lower glucose only modestly. The two harms that matter most, heart failure with saxagliptin and a small rise in pancreatitis, came out of the makers' own trials, against their commercial interest.

What sitagliptin is

Sitagliptin is an oral medicine for adults with type 2 diabetes. It is used alongside diet and exercise to improve blood-sugar control (US label). It was the first of the dipeptidyl peptidase-4 (DPP-4) inhibitors, often called "gliptins". The FDA approved Merck's Januvia as a new molecular entity on 16 October 2006 (Drugs@FDA). The European Commission authorised it in March 2007, with Merck Sharp and Dohme B.V. in the Netherlands as the EU holder (EMA). It also comes combined with metformin (Janumet).

Other members of the class include:

• Linagliptin (Trajenta in the UK, Tradjenta in the US), from Boehringer Ingelheim, first approved in the US in 2011 (Tradjenta label).
• Alogliptin (Vipidia in the UK, Nesina in the US), from Takeda, first approved in the US in 2013 (Nesina label).
• Saxagliptin (Onglyza), tested in the AstraZeneca and Bristol-Myers Squibb-funded SAVOR trial (Scirica 2013). US generics are now on DailyMed from several makers.
• Vildagliptin (Galvus), sold in the UK and Europe but not in the US. It appears in NHS prescribing data (NHSBSA).

Sitagliptin is prescription-only in the UK, US and EU. In England it is the most-dispensed gliptin: about 3.05 million items in 2025/26, ahead of linagliptin (2.57 million) and alogliptin (1.14 million). Saxagliptin and vildagliptin made up fewer than 100,000 items each (our queries of NHSBSA PCA). Almost all English sitagliptin is now dispensed as an unbranded generic.

How it works

After a meal, the gut releases two "incretin" hormones, GLP-1 and GIP. They tell the pancreas to release more insulin and less glucagon, but only when blood sugar is normal or high. The enzyme DPP-4 breaks these hormones down within minutes. Sitagliptin blocks DPP-4, so the hormones last longer and insulin release after meals is boosted (US label). Because the effect depends on glucose, gliptins taken alone rarely push blood sugar too low. That is their main practical advantage over sulfonylureas.

This is a gentler version of what GLP-1 injections such as semaglutide do, which helps explain the small glucose effect, no weight change and few stomach side effects. The mechanism is proven in humans; claims that gliptins preserve insulin-producing beta cells long term are not (Richter 2008).

Pharmacokinetics that matter in practice

Sitagliptin is about 87% absorbed, unaffected by food and has a half-life of about 12.4 hours, which supports once-daily dosing. About 79% leaves the body unchanged in urine, so blood levels rise when the kidneys are weak. That is why the dose is cut at lower kidney function (US label). Linagliptin is the exception in the class: its label gives a single 5 mg dose with no kidney adjustment (Tradjenta label).

What it is prescribed for

Licensed uses

Sitagliptin is licensed only for type 2 diabetes in adults. It must not be used in type 1 diabetes or for diabetic ketoacidosis (UK SmPC). In the UK it can be used alone when metformin is unsuitable, or added to metformin, a sulfonylurea, pioglitazone or insulin. It is not licensed for weight loss, prediabetes, type 1 diabetes or children.

Where guidelines place it

NICE rewrote its type 2 diabetes medicines advice in February 2026. Most adults now start on modified-release metformin plus an SGLT-2 inhibitor. People with atherosclerotic cardiovascular disease also get subcutaneous semaglutide, for its proven heart and kidney benefits (NICE NG28, initial medicines). Gliptins are not a first-line choice for most people. They come in at three points:

• Next add-on. For adults with no relevant other conditions who need more glucose lowering, NICE says to "offer to add a DPP-4 inhibitor". A sulfonylurea, pioglitazone or insulin comes next only if that is not suitable or does not work (NICE NG28, further medication).
• Heart failure. NICE makes a DPP-4 inhibitor the first add-on choice for people with heart failure, judging the class "relatively safe". It adds that all gliptins have similar efficacy and services should use the one with the "lowest acquisition cost" (NICE NG28). NICE does not, in the text we read, single out saxagliptin or alogliptin. Their US labels carry specific heart-failure warnings (Nesina label).
• Severe kidney disease and frailty. With an eGFR of 20 to under 30, NICE advises dapagliflozin or empagliflozin plus a DPP-4 inhibitor; below 20, a DPP-4 inhibitor may be considered alone. For frail adults who cannot safely take an SGLT-2 inhibitor, a gliptin is an option (NICE NG28).

NICE frames the gliptin option for people with early-onset type 2 diabetes as being for those not taking a GLP-1 drug or tirzepatide, used only when those drugs are not appropriate (NICE NG28).

What works and what does not

Benefits by claim

Lowering blood sugar (HbA1c)

HbA1c reflects average blood sugar over about three months. The 2008 Cochrane review pooled 25 trials lasting 12 to 52 weeks. Sitagliptin lowered HbA1c by about 0.7 percentage points against placebo, and vildagliptin by about 0.6. Limited head-to-head data showed "no improved metabolic control" compared with older tablets (Richter 2008). The review declared no conflicts. In practice, 0.7 points might take someone from 8.0% to 7.3%. That is meaningful, but less than many other options.

The best independent test is GRADE, run by the US National Institute of Diabetes and Digestive and Kidney Diseases. It randomised 5,047 people already on metformin to one of four add-ons and followed them for a mean of 5 years. All four lowered HbA1c. Sitagliptin was the least durable. People lost their HbA1c target at 38.1 per 100 person-years, against 26.1 for liraglutide, 26.5 for insulin glargine and 30.4 for the sulfonylurea glimepiride. The gap was larger in people who started with higher HbA1c (GRADE 2022). Merck donated the sitagliptin but did not fund or run the trial (full text).

In TECOS, where doctors could freely add other drugs, the HbA1c gap with placebo was only 0.29 points (Green 2015, funded by Merck).

Heart safety and heart failure

After 2008, US and European regulators required new diabetes drugs to prove they did not raise heart risk. Each gliptin maker therefore ran a large placebo-controlled trial. All four showed the drug did not raise heart attack, stroke or cardiovascular death:

• TECOS (sitagliptin, Merck-funded): 14,671 people with heart disease, median 3 years. Major events 11.4% vs 11.6% (HR 0.98, 0.88 to 1.09). Heart-failure admissions identical (HR 1.00, 0.83 to 1.20) (Green 2015).
• SAVOR-TIMI 53 (saxagliptin, AstraZeneca and Bristol-Myers Squibb-funded): 16,492 people, median 2.1 years. Major events 7.3% vs 7.2% (HR 1.00). More people on saxagliptin were admitted for heart failure: 3.5% vs 2.8% (HR 1.27, 1.07 to 1.51) (Scirica 2013).
• EXAMINE (alogliptin, Takeda-funded): 5,380 people 15–90 days after a heart attack or unstable angina. Major events 11.3% vs 11.8% (HR 0.96) (White 2013). Takeda's US label reports heart-failure admissions of 3.9% vs 3.3% (Nesina label).
• CARMELINA (linagliptin, Boehringer Ingelheim and Eli Lilly-sponsored): 6,979 people at high heart and kidney risk, median 2.2 years. Major events 12.4% vs 12.1% (HR 1.02, 0.89 to 1.17) (Rosenstock 2019).

Pooling these and similar trials, the 2021 Cochrane review found high-certainty evidence of no effect on cardiovascular death (OR 1.00), heart attack (0.97), stroke (1.00) or death from any cause (1.03). It found moderate-certainty evidence of no overall effect on heart-failure admissions (OR 0.99, 0.80 to 1.23) (Kanie 2021). So the heart-failure problem seems to belong to saxagliptin (and possibly alogliptin) rather than the whole class. The current Januvia label still advises weighing risks in people prone to heart failure, citing "two other members" of the class (US label).

GRADE adds a small independent signal: "any cardiovascular disease" ran at 2.0 per 100 person-years on sitagliptin vs 1.4 on liraglutide (sitagliptin vs the other three combined, HR 1.2, 95% CI 1.0 to 1.5), with no difference in major events, heart-failure admissions or deaths (GRADE outcomes 2022). We treat this as a weak signal, not proof of harm.

Lower risk of death compared with newer drugs?

No. A British Heart Foundation-supported network meta-analysis of 236 trials (176,310 people) linked SGLT-2 inhibitors (HR 0.80) and GLP-1 agonists (HR 0.88) with lower all-cause death than control, but not DPP-4 inhibitors (HR 1.02, 95% credible interval 0.94 to 1.11) (Zheng 2018). A 453-trial analysis funded by the European Foundation for the Study of Diabetes, with an unrestricted AstraZeneca grant, reached the same broad conclusion for people at higher cardiovascular risk (Tsapas 2020). This is why guidelines moved gliptins down the list.

Kidney protection

CARMELINA was designed partly to test kidney benefit from linagliptin and found none. The combined kidney outcome (kidney death, end-stage kidney disease or a sustained 40% fall in eGFR) occurred in 9.4% vs 8.8% (HR 1.04, 0.89 to 1.22) (Rosenstock 2019). GRADE found no difference in kidney outcomes between the four drugs (GRADE outcomes). The real kidney advantage of gliptins is practical: they can be used at doses adjusted for very low eGFR, where many other tablets cannot.

Hypoglycaemia and weight

Gliptins rarely cause hypos on their own. In GRADE, severe hypoglycaemia occurred in 0.7% of people on sitagliptin over 5 years. The figures were 1.0% for liraglutide, 1.3% for glargine and 2.2% for glimepiride (GRADE). Cochrane found no clear rise against placebo (OR 1.11, 0.95 to 1.29) (Kanie 2021). The risk rises when a gliptin is added to a sulfonylurea or insulin, and the labels advise a lower dose of those drugs (UK SmPC). Weight does not rise, though people on placebo lost slightly more (Richter 2008).

Risks and all side effects

Common side effects (licensing trials)

In placebo-controlled trials, the effects seen in at least 5% of people on sitagliptin, and more often than on placebo, were nasopharyngitis (cold-like symptoms), upper respiratory infection and headache (US label). In monotherapy trials nasopharyngitis affected 5.2% vs 3.3% on placebo. The UK SmPC lists headache and hypoglycaemia (when combined with other drugs) as common. Dizziness, constipation and itching are listed as uncommon. Osteoarthritis and pain in the limbs occurred slightly more often (UK SmPC). The 2008 Cochrane review found all-cause infections increased significantly with sitagliptin (Richter 2008).

Serious and rare risks

Other serious reactions on the labels:

• Allergic reactions: anaphylaxis, angioedema and exfoliative skin reactions including Stevens-Johnson syndrome, with onset reported within the first 3 months and sometimes after the first dose (US label).
• Kidney injury: worsening kidney function, including acute kidney failure sometimes needing dialysis, reported after marketing (US label).
• Others: interstitial lung disease, vomiting, cutaneous vasculitis and low platelets (rare), all from post-marketing reports (UK SmPC).
• Liver (alogliptin): post-marketing reports of fatal and non-fatal liver failure, although some reports lacked detail to prove cause (Nesina label).

Pancreatic cancer

TECOS found no significant difference in pancreatic cancer (P=0.32) (Green 2015), and EXAMINE reported similar cancer rates (White 2013). Trials of 2–3 years cannot rule out a slow-developing risk, but none has shown one.

Overdose

Healthy volunteers took single doses of up to 800 mg in trials, with only minimal QT changes. Sitagliptin is modestly removed by dialysis, about 13.5% of a dose in a 3–4 hour session. Treatment is supportive, with ECG monitoring (UK SmPC).

All interactions

• Sulfonylureas (gliclazide, glimepiride) and insulin: the most important interaction. Adding a gliptin raises hypo risk, and a lower sulfonylurea or insulin dose may be needed (UK SmPC; US label).
• Digoxin: sitagliptin raised digoxin exposure by about 11% (peak 18%). No dose change is advised, but people at risk of digoxin toxicity should be monitored (UK SmPC).
• Strong enzyme inhibitors (for example ketoconazole, itraconazole, ritonavir, clarithromycin): sitagliptin is mostly cleared by the kidneys. The SmPC notes that CYP3A4 metabolism may matter more in severe kidney disease, so strong inhibitors could raise levels there (UK SmPC).
• Rifampicin and other strong inducers (linagliptin): rifampicin lowers linagliptin levels. The label "strongly" recommends alternatives with strong P-gp or CYP3A4 inducers (Tradjenta label).
• Metformin, warfarin, simvastatin, oral contraceptives, glibenclamide, rosiglitazone: no meaningful change in their levels in clinical studies (UK SmPC).
• GLP-1 drugs (semaglutide, liraglutide, dulaglutide) and tirzepatide: these work on the same hormone pathway. NICE positions gliptins for people not taking a GLP-1 drug or tirzepatide, rather than alongside them (NICE NG28).
• Alcohol: no direct drug interaction is listed. The NHS says sitagliptin may not suit people with alcohol-use disorder because it can raise pancreatitis risk, and advises that it is best not to drink alcohol while taking it because alcohol can affect blood sugar (NHS).

Who should avoid sitagliptin

Contraindications

Anyone who has had a serious allergic reaction to sitagliptin, such as anaphylaxis or angioedema, should not take it (US label). It is not for type 1 diabetes or diabetic ketoacidosis (UK SmPC).

Use with care

• Past pancreatitis: not studied. The UK SmPC asks for caution (UK SmPC).
• Heart failure or high risk of it: saxagliptin and alogliptin carry specific warnings. For sitagliptin the label asks prescribers to weigh risks (Nesina label; US label). NICE still regards the class as a suitable add-on in heart failure (NICE NG28).
• Bullous pemphigoid or other blistering skin disease: the class is linked to it, especially in older adults (Lee 2020).

Children and adolescents

Three 20-week trials in 410 young people aged 10 to 17 found sitagliptin no better than placebo (US label). The UK SmPC says it should not be used in this age group because of "insufficient efficacy" (UK SmPC).

Older adults

No dose change is needed for age alone, but kidney function should be checked more often (US label). The low hypo risk suits frailty (NICE NG28), but older people carry most of the pemphigoid risk (Lee 2020).

Pregnancy and breastfeeding

Human data are lacking. The UK SmPC says Januvia should not be used during pregnancy or breastfeeding (UK SmPC). The US label says available data are insufficient to judge birth-defect risk and stresses the dangers of poorly controlled diabetes in pregnancy (US label). The NHS says doctors will usually suggest a different medicine (NHS).

Liver and kidney disease

Sitagliptin has not been studied in severe liver impairment, and the SmPC advises care. Because it is cleared by the kidneys, liver disease is not expected to change its levels much (UK SmPC). Alogliptin carries liver-failure reports (Nesina label). In kidney disease the dose must be lowered (see below), and kidney function should be checked before starting and regularly after.

Dosage and how to take it

These are label doses, not personal advice. Your prescriber sets the dose based on your kidney function and other medicines.

How to take it

Sitagliptin can be taken with or without food, once a day (UK SmPC). The Nesina label says not to double the next dose after a missed one and not to split tablets (Nesina label). Gliptins are often prescribed as combination tablets with metformin (Janumet and others), which add metformin's own cautions.

Driving

The SmPC says sitagliptin has "no or negligible influence" on driving. It notes dizziness and sleepiness have been reported, and hypos are possible with a sulfonylurea or insulin (UK SmPC).

Product quality

UK generic sitagliptin comes from many suppliers (NHSBSA), and Merck's patent settlements allowed US generics from May 2026 (Merck 10-K). Zydus's separately approved Zituvio (sitagliptin free base, made in India) has its own FDA label (Zituvio label). We could not verify reports of nitrosamine impurities in some batches from a primary source.

Follow the money: who makes it and who funded the evidence

Who made it, and who sells it now

Sitagliptin was developed by Merck & Co. of Rahway, New Jersey, USA, known as MSD outside North America (US label). The UK licence holder is Merck Sharp & Dohme (UK) Limited in London (UK SmPC). Januvia and Janumet earned Merck $2.544 billion worldwide in 2025, up 12% on 2024 ($2.268 billion) but down from $3.366 billion in 2023. Merck put the 2025 rise down to higher US net pricing, partly offset by generic competition abroad and falling US volumes (Merck 2025 Form 10-K; Merck Q4 2025 results).

The US patent on the sitagliptin compound expired in January 2023. Settlements over a later patent on the specific salt form kept US exclusivity until May 2026 for Januvia and Janumet, and July 2026 for Janumet XR. Merck told investors it expected to lose "nearly all U.S. sales" of Januvia and Janumet after that (Merck 10-K). Before then, Januvia was one of the first ten drugs chosen for Medicare price negotiation. Its negotiated price for a 30-day supply in 2026 is $113, against a 2023 list price of $527 (79% lower). Medicare Part D spent about $4.09 billion gross on it in 2023 for 843,000 people (CMS fact sheet).

In England the brand has nearly vanished. Our queries of NHS dispensing data found sitagliptin's net ingredient cost fell from about £78.2 million (2.20 million items) in 2022/23 to about £13.5 million (3.05 million items) in 2025/26. That works out at roughly £35.55 per item before generics and £4.44 after. MSD-branded packs were about 11,000 items in 2025/26, under 0.4% of the total. In contrast, linagliptin, still sold under Boehringer Ingelheim's brand, cost about £76.4 million for 2.57 million items, roughly £29.72 per item. Alogliptin cost about £31.7 million, around £27.92 per item (NHSBSA PCA). NICE's advice to use the gliptin with the "lowest acquisition cost" therefore now points to generic sitagliptin in England (NICE NG28).

The other gliptins have their own owners. Linagliptin is licensed from Boehringer Ingelheim International GmbH of Ingelheim, Germany (Tradjenta label). Alogliptin is a trademark of Takeda Pharmaceutical Company Limited (Nesina label). Saxagliptin's outcome trial was funded by AstraZeneca and Bristol-Myers Squibb (registry).

Who profits either way? Before May 2026, Merck and the makers of branded gliptins gained from wider use. SGLT-2 and GLP-1 makers gain when guidelines rank those classes higher. Health systems gain from cheap generic sitagliptin being seen as acceptable.

Who funded the evidence

Every large outcome trial was paid for by the maker of the drug it tested: Merck for TECOS (with Duke and Oxford academic partners), AstraZeneca and Bristol-Myers Squibb for SAVOR, Takeda for EXAMINE, and Boehringer Ingelheim with Eli Lilly for CARMELINA and CAROLINA (TECOS, SAVOR, EXAMINE, CARMELINA registries; CAROLINA). In CARMELINA, Boehringer staff helped design, analyse and write the report, while the academic steering committee decided on publication (full text). The main independent counterweights are the NIH-funded GRADE trial, where Merck only donated drug (registry), the Cochrane reviews and the British Heart Foundation-supported Zheng analysis. These maker-funded trials were large, double-blind and registered, and some reported results against their sponsors' interests. But they were designed to rule out harm (a non-inferiority margin of 1.3), not to find benefit.

Documented integrity and regulatory events

We found no regulator ban or withdrawal of sitagliptin. The main regulatory events are label changes: class warnings on pancreatitis, severe joint pain and bullous pemphigoid, plus heart-failure warnings for saxagliptin and alogliptin after SAVOR and EXAMINE (Nesina label). The US price was set through Medicare negotiation, and exclusivity was extended to May 2026 through patent settlements, as Merck disclosed (Merck 10-K).

Related research

For wider, evidence-graded context, see these Pure City Research guides:

• Type 2 diabetes: prevention and management evidence
• Diabetes and heart disease
• Diabetes supplements: benefits, limits and safety
• Berberine: glucose evidence and safety
• Heart failure: treatment evidence and safety
• Heart failure with reduced ejection fraction
• Heart failure with preserved ejection fraction
• Chronic kidney disease and the heart
• Acute pancreatitis
• Chronic pancreatitis

Frequently asked questions

Sources and funding notes

What we could not verify: the FDA's 2015 joint-pain safety communication and its 2016 heart-failure communication for saxagliptin and alogliptin no longer loaded on fda.gov when we checked. We therefore relied on the current labels and do not quote their case counts. We could not access the sources-of-support section of the 2021 Cochrane review, the funding statements of Tkáč 2017 and Yang 2021, or CARMELINA's heart-failure results (full text blocked by rate limits). We used abstracts and registry records instead. We could not verify reports of nitrosamine impurities in sitagliptin from a primary source. We did not find a UK-specific date for generic sitagliptin launch, or Merck's sitagliptin sales by country. NHSBSA figures are our own queries of the open Prescription Cost Analysis data: community dispensing in England only, items not patients, and net ingredient cost before discounts. We could not confirm the ownership structures of Boehringer Ingelheim or Takeda from a primary source, so we report only what the labels state.

1. Januvia (sitagliptin) US prescribing information, DailyMed — FDA-approved label; written by Merck Sharp & Dohme LLC.
2. Januvia film-coated tablets: Summary of Product Characteristics (emc; revised 12 October 2023) — UK product information; Merck Sharp & Dohme (UK) Ltd, MHRA-authorised.
3. Drugs@FDA: NDA 021995 (Januvia) — US regulator record; FDA (US government).
4. EMA: Januvia EPAR — EU regulator record; EMA (EU agency, fee- and EU-funded).
5. Nesina (alogliptin) US prescribing information — FDA-approved label; Takeda.
6. Tradjenta (linagliptin) US prescribing information — FDA-approved label; Boehringer Ingelheim.
7. Zituvio (sitagliptin) US prescribing information — FDA-approved label; Zydus Lifesciences (India).
8. NICE NG28: Type 2 diabetes in adults, further medication (updated 18 February 2026) and initial medicines — UK guideline; NICE (public body).
9. NHS: Sitagliptin (reviewed 19 August 2026) — NHS England patient information.
10. NHSBSA Prescription Cost Analysis, 2022/23 to 2025/26 — official statistics; NHS Business Services Authority.
11. CMS: Negotiated prices for initial price applicability year 2026 (fact sheet) — US government.
12. Richter B et al. DPP-4 inhibitors for type 2 diabetes mellitus. Cochrane Database Syst Rev 2008 — systematic review; conflicts "None known"; funding statement not accessed.
13. Kanie T et al. DPP-4 inhibitors, GLP-1 receptor agonists and SGLT-2 inhibitors for people with cardiovascular disease. Cochrane 2021 (PMC) — network meta-analysis; Cochrane; author conflicts as described above.
14. GRADE Study Research Group. Glycemic outcomes. NEJM 2022 and microvascular and cardiovascular outcomes (full text) — RCT; US NIDDK/NIH, donated drugs from six companies.
15. Green JB et al. TECOS. NEJM 2015 — RCT; funded by Merck Sharp & Dohme.
16. Scirica BM et al. SAVOR-TIMI 53. NEJM 2013 — RCT; funded by AstraZeneca and Bristol-Myers Squibb.
17. White WB et al. EXAMINE. NEJM 2013 — RCT; funded by Takeda Development Center Americas.
18. Rosenstock J et al. CARMELINA. JAMA 2019 (full text) — RCT; sponsored by Boehringer Ingelheim and Eli Lilly.
19. Rosenstock J et al. CAROLINA. JAMA 2019 — RCT; sponsored by Boehringer Ingelheim and Eli Lilly.
20. Zheng SL et al. SGLT-2, GLP-1 and DPP-4 and all-cause mortality. JAMA 2018 (full text) — network meta-analysis; British Heart Foundation fellowship.
21. Tsapas A et al. Comparative effectiveness of glucose-lowering drugs. Ann Intern Med 2020 — network meta-analysis; EFSD with an unrestricted AstraZeneca grant.
22. Tkáč I, Raz I. Gliptins and acute pancreatitis. Diabetes Care 2017 — pooled trial analysis; funding not stated in abstract.
23. Lee H et al. Bullous pemphigoid with DPP-4 inhibitors vs sulfonylureas. JAMA Dermatol 2020 (full text) — cohort study; Brigham and Women's Hospital and US National Institute on Aging.
24. Yang W et al. DPP-4 inhibitors and bullous pemphigoid: meta-analysis of RCTs. Diabetes Metab Res Rev 2021 — meta-analysis; funding not accessed.
25. Trial registries: TECOS, SAVOR-TIMI 53, EXAMINE, CARMELINA, GRADE — US National Library of Medicine.
26. Merck & Co. Form 10-K for 2025 and Q4 and full-year 2025 results — company filings.

Educational research reviewed 7 October 2026. This article is not medical advice; talk to a qualified clinician or pharmacist before starting, stopping or combining treatments.