Frequently Asked Questions

Yes. In this real-world cohort of adults with bipolar disorder and chronic kidney disease, GLP-1 receptor agonist use was associated with a lower observed risk of kidney replacement therapy than DPP-4 inhibitor use, with an adjusted hazard ratio of 0.56 (95% CI, 0.50–0.64; P<.001). In a secondary active-comparator analysis, GLP-1 receptor agonist use was also associated with lower observed kidney replacement therapy risk than RAAS inhibitor use (aHR 0.66, 95% CI 0.55–0.78; P<.001).

After propensity score matching against DPP-4 inhibitor users, 5-year event-free survival for kidney replacement therapy was 87.93% with GLP-1 receptor agonists versus 84.63% with DPP-4 inhibitors, an absolute risk reduction of 3.30% (2b9=15.40; P<.001). In the matched comparison with RAAS inhibitors, 5-year kidney replacement therapy-free survival was 94.68% with GLP-1 receptor agonists versus 89.77% with RAAS inhibitors (log-rank 2b9=17.66; P<.001), corresponding to a 4.91 percentage point absolute reduction and a number needed to treat of 21.

The answer is mixed. In the primary Cox model, GLP-1 receptor agonist use was associated with lower observed health care utilization than DPP-4 inhibitor use, with an adjusted hazard ratio of 0.74 (95% CI, 0.71–0.78; P<.001), and it was also associated with lower observed risk than RAAS inhibitor use (aHR 0.62, 95% CI 0.58–0.65; P<.001).

However, in propensity-matched Kaplan–Meier analyses, health care utilization-free survival did not differ significantly between groups. Versus DPP-4 inhibitors, 5-year event-free survival was 58.06% with GLP-1 receptor agonists versus 60.64% with DPP-4 inhibitors (2b9=0.018; P=.893), and versus RAAS inhibitors it was 60.62% versus 59.24% (log-rank 2b9=2.09; P=.148).

Yes. The observed reduction in kidney replacement therapy risk remained significant across multiple psychotropic subgroups, including antipsychotics (aHR 0.53, 95% CI 0.47–0.61; Q=.001), lithium (aHR 0.41, 95% CI 0.24–0.71; Q=.001), valproate (aHR 0.65, 95% CI 0.52–0.83; Q=.001), and lamotrigine (aHR 0.57, 95% CI 0.44–0.75; Q=.001). The authors note that this suggests the association was not contingent on a specific psychotropic background, although they also caution that known GLP-1 receptor agonist–associated lithium toxicity risks still warrant prospective safety monitoring.

The kidney replacement therapy association was generally consistent across sex, race, and obesity status. Reported adjusted hazard ratios were 0.62 in males and 0.53 in females, 0.58 in White individuals and 0.51 in Black individuals, and 0.58 in those with obesity and 0.41 in those without obesity.

Among Hispanic individuals, the point estimate was directionally similar but did not reach statistical significance (aHR 0.66, 95% CI 0.41–1.05; Q=.08), whereas the association remained significant in non-Hispanic individuals (aHR 0.56, 95% CI 0.48–0.64; Q<.001).

In the agent-level sensitivity analyses, both semaglutide and tirzepatide were associated with lower observed kidney replacement therapy risk. Semaglutide had an adjusted hazard ratio of 0.56 (95% CI, 0.47–0.67; P<.001), and tirzepatide had an adjusted hazard ratio of 0.40 (95% CI, 0.25–0.62; P<.001).

Both agents were also associated with lower observed health care utilization risk: semaglutide aHR 0.59 (95% CI, 0.55–0.63; P<.001) and tirzepatide aHR 0.57 (95% CI, 0.49–0.66; P<.001).

This was a retrospective cohort study using TriNetX electronic health record data from approximately 131 million adults across 117 health care systems, with mean follow-up of 4.0 years and outcomes assessed over 5 years. The study included adults aged 18 years or older with bipolar disorder and stage 1 to 3 chronic kidney disease who were new users of either GLP-1 receptor agonists or DPP-4 inhibitors; a secondary analysis compared GLP-1 receptor agonists with RAAS inhibitors.

The new-user active-comparator design matters because it was used to reduce immortal-time bias and limit confounding that can occur when treated patients are compared with nonusers. The first qualifying prescription was the index event in both groups, exposures were mutually exclusive, and propensity score matching produced balanced cohorts with all standardized mean differences below 0.1 after matching.

Kidney replacement therapy was defined using diagnosis and procedure codes that captured end-stage kidney disease, dialysis dependence, kidney transplant status, renal transplant complications, and hemodialysis procedures. Specifically, the outcome included ICD-10 codes Z91.15, Z94.0, Z99.2, N18.6, T86.10, T86.11, T86.12, and T86.19, plus CPT codes 90935 and 90937 for hemodialysis.

Health care utilization was defined as any emergency department visit or hospital inpatient/observation care encounter during the 5-year follow-up period, identified using procedural billing codes in the electronic health record.

The findings are observational and hypothesis-generating, not proof of causality. The authors state that residual confounding cannot be excluded despite propensity score matching and covariate adjustment, especially from indication, disease severity, lifestyle factors, psychiatric illness severity, treatment engagement, and social determinants of health that were not fully captured in TriNetX.

Other limitations included possible coding inaccuracies, incomplete chronic kidney disease staging, lack of laboratory confirmation such as serial eGFR or albuminuria, possible misclassification of incident versus prevalent kidney replacement therapy, inability to model death as a competing risk, lack of dose and adherence data, and changing GLP-1 receptor agonist prescribing patterns across the 2009–2025 study period. The authors conclude that prospective replication and randomized trials are needed before these data can support treatment recommendations.

No firm treatment recommendation can be made from this study alone. The authors conclude that GLP-1 receptor agonist use was associated with lower observed kidney replacement therapy risk in adults with bipolar disorder and chronic kidney disease, but they also state that the findings do not establish causality or provide a basis for treatment recommendations.

The article suggests these results may support closer collaboration between psychiatry, nephrology, and endocrinology and may be relevant when patients are already candidates for GLP-1 receptor agonists, but prospective studies and randomized clinical trials are needed to define clinical applicability and risk-benefit more clearly.