Peer-reviewed Zepbound cost-offset numbers published in Diabetes, Obesity and Metabolism

A new peer-reviewed study found that older adults who kept taking Zepbound (tirzepatide) had lower health care costs, outside the price of the drug itself, than similar adults who did not take a GLP-1 medicine. The study appeared in the journal Diabetes, Obesity and Metabolism and was funded by Eli Lilly, the drug's maker ^[1][2].

Researchers compared 15,843 tirzepatide users over age 55 with 15,843 matched adults who had not started any GLP-1 or GIP/GLP-1 medicine, using insurance claims data ^[1][2]. In the main analysis, monthly costs for people on tirzepatide moved from $1,061 to $955 over 18 months. Costs for the untreated group rose from $1,031 to $1,244 over the same period ^[1]. That gap worked out to $145 per person per month between 6 and 12 months, and $319 per person per month between 12 and 18 months, both statistically significant differences ^[1].

A second method of analyzing the same data, called pairwise comparison, found even larger gaps: $181 per person per month at 6 to 12 months and $607 per person per month at 12 to 18 months ^[1][2]. The study's authors treated the smaller, primary estimates as more reliable, because so few patients remained in follow-up by 18 months that the pairwise method could be skewed by which patients stayed longest ^[1].

Much of the savings tracked with fewer hospital stays and emergency visits. People on tirzepatide had lower combined rates of inpatient admissions and ER visits at every checkpoint, with the gap widening over time ^[1]. Routine outpatient and office visits were not significantly different between the groups, though people on tirzepatide had numerically more of them, which researchers said could reflect more routine care rather than more illness ^[1].

Importantly, the study excluded the cost of tirzepatide itself, because claims data do not show the drug's real, negotiated price ^[1][2]. That means the findings show potential savings elsewhere in the health system, not whether tirzepatide is a net cost or a net saving once its own price is included ^[1][2].

Why it matters for patients

For people over 55 already on tirzepatide, this study suggests that staying on treatment was linked to lower spending on hospital stays and ER visits, at least among those who kept taking the drug and were captured correctly in claims data. It does not answer whether the total cost of care, including what patients or Medicare pay for tirzepatide, goes down or up. It also only describes people who stayed on treatment; anyone who stopped was removed from the analysis at that point, so the results say nothing about what happens to costs after stopping ^[1].

The timing matters for Medicare beneficiaries specifically, because CMS is running a GLP-1 Bridge demonstration in which eligible beneficiaries pay $245 for a 30-day supply with a $50 copay, leaving Medicare a net cost of $195 per month ^[1]. The study's authors noted the $319 monthly offset seen at 12 to 18 months has implications for that $195 figure, though this is not the same as saying tirzepatide pays for itself ^[1].

What happens next

CMS is expected to use the GLP-1 Bridge program, which started July 1, 2026, to help decide whether GLP-1 drugs should become a permanent part of Medicare coverage ^[1]. A separate July 2026 analysis from the Institute for Clinical and Economic Review found tirzepatide's cost per quality-adjusted life-year was the lowest among GLP-1 drugs it reviewed, but warned that treating even 1% of eligible patients would exceed the group's roughly $880 million annual budget threshold ^[1]. Whether the cost patterns in this new study hold up depends heavily on how long patients stay on treatment, since the study's authors and the ICER report both note that real-world persistence with GLP-1 drugs tends to be lower than in clinical trials, and weight regain can occur within 1.5 years of stopping ^[1].