Probably not, on the best evidence available. Across 48 placebo-controlled trials with 94,245 participants, the odds ratio for kidney cancer was 1.12 (95% CI 0.78 to 1.60) [1]. In people, that is between 5 fewer and 13 more cases per 10,000 treated. The reviewers graded it as moderate-certainty evidence of little or no effect. The worry came from a matched cohort that put the hazard ratio at 1.38 (95% CI 0.99 to 1.93) [2]. Its own authors called that marginally nonsignificant.
Where the kidney signal came from
The cohort drew on 2014 to 2024 health records and matched 43,317 GLP-1 users to 43,315 eligible non-users [2]. Everyone had obesity or overweight and no prior cancer. Mean age was 52.4, and 68.2% were women. Fourteen cancers were tracked.
Overall, the picture favored the drug. The fourteen cancers ran at 13.6 per 1,000 person-years in users against 16.4 in non-users, a hazard ratio of 0.83 (95% CI 0.76 to 0.91). Endometrial cancer came in at 0.75, ovarian at 0.53 and meningioma at 0.69. Kidney cancer was the one estimate above 1. Its interval starts at 0.99, so it does not exclude no effect. The authors asked for longer follow-up rather than calling it a harm. Thyroid was among the fourteen, with no separate figure given; the thyroid evidence is set out in the thyroid cancer guide.
What the randomized trials add
Trials remove the question of who gets prescribed a drug, which a cohort cannot fully do. The meta-analysis pooled 48 placebo-controlled trials searched to August 2025 [1]. For kidney cancer the odds ratio was 1.12, with an interval from 0.78 to 1.60. The same moderate-certainty grade applied to thyroid, pancreatic and breast cancer. Results held in trials of semaglutide or tirzepatide and across dose and duration subgroups.
Two comparisons against SGLT2 drugs, pointing opposite ways
A target-trial emulation matched 294,664 adults with type 2 diabetes starting an SGLT2 inhibitor against the same number starting a GLP-1 [3]. Over a mean of 44.2 months, renal cell carcinoma ran at 1.07 per 1,000 patient-years on SGLT2 drugs and 1.26 on GLP-1 drugs. The reported hazard ratio for SGLT2 use was 0.85 (95% CI 0.79 to 0.92). Urothelial cancer showed the same pattern.
That is a comparison between two treatments, not against no treatment. It cannot say whether GLP-1 drugs raise kidney cancer or SGLT2 drugs lower it. The same trade-off appears in the kidney stone comparisons, where every study ranks one diabetes drug against another.
A second emulation looked only at adults with type 2 diabetes after bariatric surgery [4]. Over five years, starting a GLP-1 was linked to lower kidney cancer risk than metformin, insulin, SGLT2 inhibitors or sulfonylureas. The risk differences ran from 0.18 to 0.22 percentage points. Against SGLT2 drugs it was 0.19 points lower, across 8,613 matched pairs. That is the opposite direction to the first comparison, in a different population. How these drugs work after surgery is covered in the post-bariatric trials.
How to read four answers at once
The randomized evidence carries the most weight and sits closest to no effect. The cohort signal is real enough to watch and too uncertain to act on. The two SGLT2 comparisons disagree, and neither uses a placebo. Pancreatic cancer shows a similar pattern of comparator-dependent answers in the pancreatic cancer guide.
A personal or family history of kidney cancer is a reason to raise the question with a prescriber. It is not, on this evidence, a reason the drug was shown to be unsafe. Whether an online intake asks about it is one of the questions sellers leave open.