Tirzepatide lowered liver fat more than insulin degludec in SURPASS-3 MRI substudy

Adults with type 2 diabetes and a high fatty liver risk score lost significantly more liver fat over a year on tirzepatide than on insulin degludec, according to an MRI substudy of the phase 3 SURPASS-3 trial ^[1]. Pooled results for the tirzepatide 10 mg and 15 mg doses showed an 8.09 percentage-point drop in liver fat content, against 3.38 points on insulin degludec, an estimated difference of 4.71 percentage points in tirzepatide's favor ^[1].

The substudy enrolled 296 of 502 people screened, or 59%, at 45 medical research centers and hospitals in eight countries, including the United States ^[1]. Participants were assessed for eligibility between April 1 and Nov 15, 2019 ^[1]. They were adults with type 2 diabetes, an HbA1c between 7.0% and 10.5%, a BMI of at least 25 kg/m2, stable weight, no previous insulin treatment, and at least three months on metformin alone or with an SGLT2 inhibitor ^[1]. Everyone in the substudy also had a fatty liver index of at least 60 ^[1]. They were randomly assigned to weekly tirzepatide at 5 mg (71 people), 10 mg (79 people) or 15 mg (72 people), or to once-daily titrated insulin degludec (74 people) ^[1].

Liver fat was measured with MRI proton density fat fraction ^[1]. The average at the start was 15.71% ^[1]. The primary result was the change at week 52 for the combined 10 mg and 15 mg groups versus insulin degludec ^[1]. The difference was -4.71 percentage points (95% CI -6.72 to -2.70; p<0.0001) ^[1]. The study also reported reductions in visceral and abdominal subcutaneous fat, and said the liver fat reduction was correlated with baseline liver fat (rho = -0.71) and with reductions in visceral fat (rho = 0.29), subcutaneous fat (rho = 0.33) and body weight (rho = 0.34), all at p of 0.0006 or less ^[1].

The trial was randomized, open-label and parallel-group, meaning participants and investigators knew which treatment was assigned ^[1]. It was funded by Eli Lilly and Company, and four of the listed authors are employees and shareholders of the company ^[1]. The paper, published in 2022, describes tirzepatide as a dual GIP and GLP-1 receptor agonist that was under development for type 2 diabetes at the time ^[1].

Why it matters for patients

Fatty liver is common in people with type 2 diabetes, and this substudy was restricted to those whose fatty liver index was at least 60 ^[1]. Participants started with a mean liver fat content of 15.71% ^[1]. Both treatment groups ended the year with less liver fat, but the tirzepatide groups ended with a larger reduction ^[1]. Because the insulin group also fell 3.38 points, the 4.71-point figure compares two active treatments rather than comparing treatment with no treatment ^[1].

Liver fat on MRI is an imaging measurement, not a symptom or a clinical event. The substudy did not test whether lowering liver fat this way prevents liver inflammation, scarring or liver-related illness; that has not been reported ^[1]. The correlations with weight loss mean the study cannot say whether tirzepatide affected the liver directly or mainly through weight loss ^[1].

The results cover a specific group: adults with type 2 diabetes and this risk score, most of them on metformin. People without diabetes, or with lower fatty liver index scores, were not studied, so the findings may not describe them ^[1]. The pooled primary analysis covered only the 10 mg and 15 mg doses; results for the 5 mg dose are not part of that comparison ^[1]. The comparator was insulin degludec, so the substudy did not compare tirzepatide with semaglutide or with a placebo ^[1].

What has not been reported

Whether these MRI changes lead to fewer liver problems over years, or whether they persist after treatment stops. Participants were followed to week 52, and the substudy is listed as complete ^[1].