Tirzepatide vs SGLT2 Drugs in MASLD

A large US health-record study found lower recorded rates of death, hospitalization, major cardiovascular events and major liver outcomes among people with metabolic dysfunction-associated steatotic liver disease (MASLD) who started tirzepatide rather than an SGLT2 inhibitor. After matching, the analysis included 43,743 people in each group. It was retrospective and treatment was not assigned at random, so the results show associations rather than proof that tirzepatide caused better outcomes^[1].

Khalil and colleagues published the multicenter cohort in Hepatology International on 3 July 2026. At one year, the reported hazard ratio for all-cause mortality was 0.328 for tirzepatide versus SGLT2 inhibitors, with a 95% confidence interval from 0.272 to 0.394. The corresponding event counts were 144 and 533^[1]. That is a large difference, but a hazard ratio from a matched database cohort is not a randomized estimate of efficacy.

What the study compared

The researchers used the TriNetX US Collaborative Network to identify new users of tirzepatide or an SGLT2 inhibitor who had MASLD and at least one metabolic comorbidity. They matched the groups one to one with a caliper of 0.1 across more than 50 measured covariates. The paper describes a multicenter, retrospective, propensity score-matched cohort, not a clinical trial^[1].

The prespecified outcomes covered several organ systems. They included all-cause mortality, all-cause hospitalization, major adverse cardiovascular events (MACE), major adverse liver outcomes (MALO) and major adverse kidney events (MAKE). The authors used Kaplan-Meier estimates and Cox proportional hazards models at one and three years. These methods compare the timing of recorded events, but they do not recreate random assignment.

Propensity matching can balance diagnoses, medicines and other characteristics that appear in the data and were included in the model. It cannot balance a factor that was not measured, was coded poorly or was omitted. A clinician may choose tirzepatide or an SGLT2 inhibitor for different reasons, including glycaemic goals, kidney function, heart failure, body weight, insurance access and prior treatment. Those choices can remain linked to later outcomes after matching.

The one-year associations were large

At one year, 3,033 tirzepatide initiators and 7,143 SGLT2 inhibitor initiators had an all-cause hospitalization. The reported hazard ratio was 0.460 (95% CI 0.441 to 0.480). MACE occurred in 1,064 versus 2,376 people, corresponding to a hazard ratio of 0.506 (95% CI 0.471 to 0.544)^[1]. These estimates describe the matched cohorts as recorded in TriNetX.

For the liver composite, the study recorded 149 MALO events with tirzepatide and 465 with SGLT2 inhibitors. The hazard ratio was 0.379 (95% CI 0.315 to 0.456). Hepatic failure and ascites also numerically favored tirzepatide, while varices and hepatic encephalopathy did not differ significantly^[1]. MAKE also showed no significant difference, which argues against treating the findings as a uniform advantage across every endpoint.

The mortality result deserves particular restraint. The one-year cumulative incidences were 0.48% and 1.38%, and the event counts were modest relative to 43,743 people per group. A difference this large can reflect a real treatment association, but it can also magnify treatment selection, incomplete outcome capture, differences in continued drug use or other residual confounding. The observational design cannot tell how much each explanation contributed.

Three-year results pointed the same way

The three-year analysis reported 39,838 people per group. Mortality events numbered 208 with tirzepatide and 963 with SGLT2 inhibitors, for a hazard ratio of 0.374 (95% CI 0.321 to 0.436). Hospitalization had a hazard ratio of 0.528, MACE 0.545 and MALO 0.478^[1]. The direction therefore persisted across both follow-up windows.

Longer follow-up does not remove the original treatment-selection problem. Electronic health records can also miss events that occur outside participating systems, and routine-care records do not guarantee that a prescription was filled or continued. The publicly accessible abstract does not report how switching, stopping treatment or loss to follow-up affected these comparisons. Readers should not convert the hazard ratios into a personal forecast.

Randomized liver evidence answers a different question

Tirzepatide has also been tested in randomized liver research, but that evidence does not validate this head-to-head outcome comparison. In a phase 2 trial, 190 people with biopsy-confirmed metabolic dysfunction-associated steatohepatitis (MASH) and stage F2 or F3 fibrosis were randomly assigned to tirzepatide or placebo for 52 weeks. MASH resolution without worsening fibrosis occurred in 44% to 62% of the tirzepatide groups and 10% of the placebo group^[2].

The phase 2 trial used biopsy endpoints and a placebo comparator. The 2026 TriNetX study used clinical events and an active SGLT2 inhibitor comparator. One cannot substitute for the other. The randomized trial supports a liver-histology signal in a narrower MASH population, while the database study raises a hypothesis about comparative outcomes in routine care. Our tirzepatide research overview keeps those evidence types separate.

What this study does not establish

The analysis does not establish that tirzepatide is superior to SGLT2 inhibitors for treating MASLD. It also does not show that patients should switch classes. SGLT2 inhibitors and tirzepatide may be selected for overlapping but different clinical priorities, and the study compared classes as used in records rather than assigning a single standardized regimen.

A randomized head-to-head trial would better address causality by balancing measured and unmeasured baseline factors through allocation. It would also need clear definitions for treatment changes, outcome adjudication and follow-up. Until such evidence exists, the accurate summary is narrower: tirzepatide initiation was associated with fewer recorded events than SGLT2 inhibitor initiation in this matched cohort, and the size of that association requires prospective testing.

Medical disclaimer: This article is for educational and journalistic purposes only and does not constitute medical advice. Peptides discussed may be classified as prescription medicines or research chemicals depending on your jurisdiction. Always consult a qualified healthcare professional before using any peptide product. PeptideMethods.com does not sell, distribute, or facilitate the sale of any peptide product.