Know where the question sits in the trial
This report was a post hoc kidney analysis of SURPASS-4, an open-label randomized trial involving 2,002 adults with type 2 diabetes and high cardiovascular risk. Tirzepatide was compared with insulin glargine. Median treatment duration was 85 weeks, and the kidney analysis combined the tirzepatide groups. [1]
Post hoc means the analysis was developed after the original study framework rather than serving as the main trial question. It can provide useful evidence, especially when grounded in randomized treatment assignment, but should be labeled clearly. Readers need that context before treating a secondary paper like a dedicated outcomes trial.
What the kidney measurements showed
Estimated filtration declined by an average 1.4 versus 3.6 mL/min/1.73 m² per year with tirzepatide and insulin glargine, respectively. A combined kidney endpoint also favored tirzepatide, with a hazard ratio of 0.58. That composite included new macroalbuminuria as well as filtration decline, kidney failure and kidney-related death. [1]
The components prevent the headline from being reduced to fewer cases of kidney failure. A combined result can be influenced by more frequently occurring laboratory outcomes. The exact endpoint matters as much as the relative estimate when judging what a study contributes to a clinical decision.
Why this is different from FLOW
Our FLOW analysis covers a trial designed around kidney outcomes in a defined chronic-kidney-disease population. The designs, medicines and comparators differ. Putting their hazard ratios in a table does not establish which drug is better for kidneys.
A post hoc analysis also raises questions about multiplicity, selection of outcomes and the need for confirmation. Those are reasons to qualify the inference, not to pretend the observed findings do not exist. A useful summary states both the direction of the result and the limitations of how the question was examined.
The comparator was an actively titrated diabetes treatment, not placebo. That makes the clinical contrast distinct from a no-active-drug comparison. Keep the comparator named even in a short summary so that a favorable kidney estimate does not appear to describe an unspecified alternative.
A sensible question for the clinician
Ask how much weight the clinician gives this additional analysis compared with trials specifically designed for your diagnosis. That opens a discussion of the whole evidence base rather than asking one paper to settle a decision it was not built to answer.
Use our research-reading guide to identify the main endpoint and comparator when opening a study. If several clinicians manage diabetes, cardiovascular risk and kidney monitoring, clarify who integrates their recommendations. This paper supports an informed discussion of a kidney signal; it is not an instruction to change insulin or other treatment independently.