Introduction

Chronic kidney disease (CKD) is a major complication of type 2 diabetes (T2D) and a leading cause of kidney failure worldwide1. Patients with CKD and T2D have a substantially increased risk of cardiovascular disease and progressive kidney dysfunction, despite treatment with renin-angiotensin system (RAS) inhibitors and sodium-glucose cotransporter-2 (SGLT2) inhibitors2,3. However, even with these therapies, a high residual risk of adverse outcomes remains, highlighting the need for additional treatment strategies.

Finerenone, a nonsteroidal mineralocorticoid receptor antagonist (MRA), has demonstrated cardiorenal benefits in patients with CKD and type 2 diabetes in the FIDELIO-DKD and FIGARO-DKD trials4,5. In contrast, spironolactone, a traditional and cost-effective steroidal MRA, has been shown to reduce proteinuria and blood pressure but is limited by a high incidence of hyperkalemia6,7,8,9. The BARACK-D trial evaluated spironolactone in patients with CKD and found limited cardiovascular benefits, with high discontinuation rates due to declines in estimated glomerular filtration rate (eGFR) and hyperkalemia10.

Recently, steroidal agents such as spironolactone and eplerenone, along with the nonsteroidal agent finerenone, have been guideline-recommended for heart failure with reduced ejection fraction (HFrEF)11 and are increasingly used in CKD12, particularly in diabetic nephropathy13. However, direct comparisons of finerenone and spironolactone in real-world CKD and T2D populations remain scarce. Given their distinct pharmacological properties, further evaluation of their comparative effectiveness and safety is warranted in clinical practice. In this work, we conducted a target trial emulation using real-world electronic health records from the TriNetX network to compare the effectiveness and safety of finerenone and spironolactone in patients with CKD and T2D.

Results

Patient characteristics

Among the 18,314 patients who initiated finerenone or spironolactone between June 2021 and September 2024, 1345 were identified as eligible new users of finerenone (mean [SD] age, 68.5 [10.6] years; 799 [59.4%] male; 585 [43.5%] White), and 16,969 were eligible new users of spironolactone (mean [SD] age, 72.3 [11.1] years; 7435 [43.8%] male; 11,029 [65.0%] White). After propensity score matching (PSM), 1132 individuals were included in each group for outcome analyses (Table 1). The study cohort construction and exclusion criteria are shown in Fig. 1.

Fig. 1: Study cohort construction and exclusion criteria.
figure 1

*CKD was defined by two eGFR values < 60 mL/min/1.73 m² (using the Modification of Diet in Renal Disease Study [MDRD] formula), separated by at least 90 days. †Strong CYP3A4 inhibitors, for example ketoconazole, clarithromycin, etc. CKD chronic kidney disease, CYP3A4 cytochrome P450 3A4, ESKD end stage kidney disease, ICH intracranial hemorrhage, MRA mineralocorticoid receptor antagonist, PSM propensity score matching.

Full size image
Table 1 Baseline characteristics of patients with chronic kidney disease and type 2 diabetes initiating finerenone or spironolactone, before and after propensity score matching
Full size table

Prior to matching, the finerenone group had a higher prevalence of hyperuricemia compared with the spironolactone group. Conversely, the spironolactone group had a higher prevalence of anemia, cardiovascular and kidney comorbidities, other organ comorbidities, systemic disorders, sleep apnea, mood disorders, and lifestyle-related health hazards. After PSM, baseline characteristics were well balanced between the two groups.

Comparison of finerenone and spironolactone on outcome of interests

After a follow-up of 1.3 years (IQR, 0.8–1.5), treatment with finerenone was associated with a significantly lower risk of all-cause mortality compared with spironolactone (aHR, 0.31; 95% CI, 0.21–0.45; P < 0.001). A total of 35 patients (3.1%) in the finerenone group and 112 patients (9.9%) in the spironolactone group died. This corresponded to an absolute risk reduction (ARR) of 7% (95% CI, 5–9) and a number needed to treat (NNT) of 15 (95% CI, 11–21) (Table 2). The intention-to-treat survival probability for each treatment group is presented in Fig. 2. 112 patients (9.9%) in the finerenone group and 150 patients (13.3%) in the spironolactone group experienced MACE. Finerenone was associated with a significantly lower risk of MACE than spironolactone (aHR, 0.74; 95% CI, 0.58–0.94; P = 0.013), with an ARR of 3% (95% CI, 1%–6%) and a NNT of 29 (95% CI, 17–143) (Table 2). This association remained consistent across individual components, including acute myocardial infarction, cardiac arrest/ cardiogenic shock, and alternative MACE definitions (Supplementary Table 5). MAKE occurred in 46 patients (4.1%) in the finerenone group and 96 patients (8.5%) in the spironolactone group, with a lower hazard ratio in the finerenone group (aHR, 0.47; 95% CI, 0.33–0.67; P < 0.001). This corresponded to an ARR of 4% (95% CI, 2%–6%) and an NNT of 23 (95% CI, 16–42) (Table 2). The E-values for MACE (2.06), MAKE (3.68), and mortality (5.91) suggest that the observed associations are robust to unmeasured confounding (Table 2).

Fig. 2: Cumulative incidence of study outcomes in patients with chronic kidney disease and type 2 diabetes.
figure 2

A MACE. B MAKE. C All-cause mortality. Kaplan–Meier curves show estimated survival/event-free probabilities for each treatment group (centre), with shaded areas indicating 95% confidence intervals (error bands). The number at risk and the number of events at each time point are shown below each panel. Sample sizes for both finerenone and spironolactone groups are n = 1132, with each patient representing an independent biological replicate. P values were calculated using a two-sided log-rank test. Exact p values are provided where available; otherwise, p values are reported as <0.001 when below this threshold. No adjustments were made for multiple comparisons. Source data are provided as a Source data file. MACE major adverse cardiac events, MAKE major adverse kidney events.

Full size image
Table 2 Comparison of primary outcomes between finerenone and spironolactone: intention-to-treat analysis
Full size table

Secondary outcome

Patients receiving spironolactone experienced significantly more episodes of hyperkalemia. Hyperkalemia (potassium ≥5.5 mmol/L) occurred in 26.4% of patients receiving spironolactone and 17.2% of those receiving finerenone (P < 0.001). Hyperkalemia (potassium ≥6.0 mmol/L) was observed in 10.2% of the spironolactone group and 5.9% of the finerenone group (P < 0.001). Severe hyperkalemia with potassium ≥6.5 mmol/L occurred in 3.9% of the spironolactone group and 2.1% of the finerenone group (P = 0.014) (Supplementary Table 6).

Subgroup and sensitivity analyses

The association between finerenone use and a lower risk of the primary outcomes remained consistent across subgroups stratified by age, sex, HbA1c level (≥7% vs <7%), kidney function (eGFR ≥45, 30–44, and <30 mL/min/1.73 m²), proteinuria (UPCR ≥ 300 vs <300 mg/g), heart failure status, concurrent SGLT2 inhibitor use, RAS inhibitor use, and enrollment year. Notably, the association with lower mortality was more pronounced among patients receiving RAS inhibitors (p for interaction = 0.048), and the association with lower MAKE risk was stronger among those enrolled in 2024 or later (p for interaction <0.001). No significant effect modification was observed in other subgroups (p for interaction >0.05 for all) (Fig. 3).

Fig. 3: Adjusted hazard ratios for primary outcomes across prespecified subgroups.
figure 3

Points represent the estimated hazard ratio (measure of centre) for each subgroup, and horizontal lines represent the corresponding 95% confidence intervals (error bars). Sample sizes for both finerenone and spironolactone groups are n = 1132, with each patient representing an independent biological replicate. P values for interaction were derived from likelihood‑ratio χ² tests comparing Cox models with versus without the treatment × subgroup interaction terms in the PS‑matched cohort. Exact p values are provided where available; otherwise, p values are reported as <0.001 when below this threshold. No adjustments were made for multiple comparisons. Source data are provided as a Source data file. eGFR estimated glomerular filtration rate, HbA1c glycated hemoglobin, HF heart failure, MACE major adverse cardiac events, MAKE major adverse kidney events, RAS renin-angiotensin system, SGLT2 sodium-glucose cotransporter-2, UPCR urine protein and creatinine ratio.

Full size image

Treatment persistence was modest overall, with 49.0% of patients remaining on finerenone and 55.5% on spironolactone at 6 months, declining to 33.8% and 28.3% at 12 months, respectively. In landmark analyses restricted to patients with continued treatment at 3 and 6 months, finerenone was consistently associated with lower risks of MACE, MAKE, and all-cause mortality compared with spironolactone (Supplementary Table 7). Additional analyses performed before PSM, censoring events within the first 30 days of treatment initiation, restricting the cohort to those with documented drug doses (10 mg for finerenone, 25 mg for spironolactone), excluding patients who switched to other MRAs, and limited to patients with complete laboratory data also supported the robustness of our findings (Supplementary Tables 8 and 9).

Negative outcome analysis

In the negative outcome analysis of overall cancer risk, patients with a history of cancer at treatment initiation were excluded. No significant difference in the risk of de novo cancer was observed between the two treatment groups (Supplementary Table 6).

Discussion

In this real-world cohort study emulating a randomized clinical trial, finerenone use in patients with CKD and T2D was associated with a lower risk of all-cause mortality, MACE, and MAKE compared with spironolactone, with a median follow-up of 1.3 years (IQR, 0.8–1.5). Notably, the combination of finerenone with RAS inhibitors was associated with a significant reduction in the risk of MACE. Furthermore, clinical use of finerenone after 2024 was linked to a decreased risk of MAKE. Additionally, in this new-user intention-to-treat study, treatment with finerenone was associated with a lower incidence of hyperkalemia compared with spironolactone.

To our knowledge, this is the first study to head-to-head compare the effectiveness of finerenone versus spironolactone in a real-world cohort of patients with CKD and T2D using a target trial emulation approach. Previous randomized trials, such as FIGARO-DKD and FIDELIO-DKD, have demonstrated the cardiovascular and kidney benefits of finerenone in this population, with the pooled FIDELITY analysis confirming its ability to delay CKD progression4,5,14. Despite differences in baseline characteristics and follow-up duration, the 9.9% MACE incidence and 3.1% mortality we observed over 1.3 years are broadly consistent with those reported in the FIDELIO-DKD trial, reinforcing the external validity of our real-world findings. In contrast, spironolactone, commonly used for CKD-related hypertension and proteinuria, may be limited by a higher incidence of hyperkalemia. The BARACK-D trial did not identify significant cardiovascular benefits associated with spironolactone use in patients with CKD, likely due to its frequent discontinuation owing to safety concerns, even at low doses10. Furthermore, low-dose spironolactone users did not demonstrate a significant benefit in slowing renal function decline over the study period10. In a post hoc analysis of the FIDELITY-TRH and AMBER studies, finerenone was associated with a lower risk of hyperkalemia and treatment discontinuation compared with spironolactone in patients with treatment-resistant hypertension and CKD15.

Both steroidal and non-steroidal MRAs mitigate endothelial dysfunction, oxidative stress, and albuminuria via aldosterone blockade8,16,17,18. Spironolactone has been shown to reduce left ventricular (LV) mass, improve arterial stiffness, and enhance LV function in early CKD19,20. However, spironolactone use is often limited by early discontinuation, with nearly two-thirds stopping treatment within 6 months in the BARACK-D trial10, compared to lower discontinuation rates (~27–29%) over full follow-up in finerenone trials4,5. In our real-world study, treatment persistence between 6 and 12 months was higher in the finerenone group (33.8%) compared to spironolactone (28.3%). This pattern may reflect better long-term tolerability in patients with finerenone.

Hyperkalemia remains a key limitation of MRAs therapy in CKD management. Discontinuation or down-titration of MRAs, due to hyperkalemia, is associated with significantly increased risks of adverse outcomes in patients with CKD and heart failure. The BIOSTAT-CHF and Swedish HF registries previously found that the hyperkalemia itself was not directly linked to adverse outcomes—instead, it marked suboptimal therapy that led to worse prognosis21,22. In our study, finerenone was associated with a lower risk of hyperkalemia compared with spironolactone, which aligns with prior research15. Before matching, mean serum potassium was marginally higher in participants treated with finerenone than in those given spironolactone, a disparity that likely reflects clinicians’ preference for finerenone in patients with hyperkalemia liability. This imbalance was eliminated after matching, and finerenone still conferred a lower incidence of treatment-emergent hyperkalemia than spironolactone. Our specificity analyses yielded consistent findings, with finerenone being associated with greater benefit than spironolactone, particularly in the lower risk of cardiac arrest. This observation aligns with the lower incidence of hyperkalemia-related sudden cardiac death observed in the finerenone group. These findings support the potential role of finerenone as an alternative to spironolactone in the management of patients with CKD and type 2 diabetes, particularly among those at increased risk of potassium dysregulation.

Furthermore, the observed association between finerenone and reduced mortality and cardiorenal outcomes may be partly attributed to its distinct pharmacological properties beyond treatment persistence and lower hyperkalemia risk14. Finerenone has distinct pharmacological properties, including inhibition of pro-inflammatory and pro-fibrotic pathways via suppression of serum/glucocorticoid-regulated kinase 1 (Sgk1) activation and prevention of cofactor recruitment to mineralocorticoid receptors23,24,25. Finerenone shows superior antifibrotic effects in the heart and kidneys, slowing CKD progression in patients with diabetes15. These pharmacological distinctions provide a potential explanation for our findings and highlight finerenone’s role as a more targeted therapeutic option in CKD and T2D23.

The low NNT for mortality (15) from our study underscores the potential benefit even over short-term follow-up in this high-risk population. Furthermore, finerenone appears to act synergistically with RAS inhibitors, providing a more comprehensive strategy for managing cardiorenal complications in T2D. A meta-analysis demonstrated that finerenone, when combined with RAS inhibitors, improved cardiovascular and kidney outcomes and reduced the risk of hyperkalemia compared with than traditional MRAs such as spironolactone26. Clinically, this dual blockade has been shown to enhance cardiorenal outcomes: the FIDELIO-DKD and FIGARO-DKD trials reported that the addition of finerenone to optimized RAS blockade significantly reduced rates of MACE, particularly hospitalizations for heart failure and cardiovascular mortality4,5,14.

This study has several limitations. First, as a retrospective analysis, residual confounding cannot be fully excluded despite the rigorous adjustments made using propensity score matching and the target trial framework. However, E-value analysis suggests that an unmeasured confounder would be unlikely to account for the observed associations. Indication bias remains a limitation, as finerenone was more frequently prescribed to patients with lower eGFR and higher potassium levels, though we adjusted for complications and cardiorenal risk factors associated with diabetes as proxy of disease severity, which likely mitigated potential bias. However, the use of a target trial emulation approach strengthens the validity of causal inferences by systematically mitigating key sources of bias, including immortal time bias, selection bias, and depletion of susceptible bias. Second, the reliance on EHRs and standardized coding systems, including the International Classification of Diseases, 10th Revision (ICD-10), may introduce ascertainment bias due to potential misclassification or incomplete capture of events. While negative control outcome was used to assess potential systemic bias, overdiagnosis, underdiagnosis, or misclassification cannot be fully excluded. Third, the relatively short follow-up period of 1.3 years, reflecting the recent approval of finerenone, limits our ability to draw definitive conclusions on long-term efficacy and safety, which are critical in managing chronic diseases like CKD. Fourth, treatment adherence over time could not be fully assessed; exposure was defined at baseline, and dynamic changes such as discontinuation or switching were not captured. However, treatment persistence and landmark sensitivity analyses supported the robustness of our findings among sustained users, although residual misclassification was possible. Additionally, the reasons for treatment discontinuation, such as hyperkalemia or other adverse events, were not captured in the database, limiting further insight into discontinuation patterns. Fifth, the predominance of spironolactone 25 mg and finerenone 10 mg prescriptions limited our ability to evaluate dose–response effects. A dose-restricted sensitivity analysis confined to these common doses corroborated the primary findings. Moreover, a notable portion of patients had missing or unrecorded MRA dosing data, further restricting detailed dose stratification. Sixth, missing laboratory data, especially proteinuria (UPCR), is a common limitation of EHR studies. High UPCR missingness may underestimate proteinuria prevalence and affect generalizability. Finally, CKD was defined based on two eGFR measurements at least 90 days apart, which may have inadvertently included patients with transient declines due to acute kidney injury. However, this algorithm has been previously validated for CKD stage 3 or greater27.

Ongoing clinical trials, such as FIND-CKD (NCT05047263) and CAPTIVATE (NCT06058585), will further elucidate the efficacy of finerenone in a broader CKD population, including individuals without diabetes. Additionally, the CONFIDENCE trial (NCT05254002) will evaluate the potential benefits of combining MRAs with SGLT2 inhibitors, particularly in managing potassium homeostasis and improving long-term clinical outcomes in patients with CKD. As finerenone moves toward broader clinical adoption, its higher cost relative to spironolactone warrants careful consideration. Although modeling studies suggest that finerenone may be cost-effective in CKD and T2D28,29,30,31,32, cost-effectiveness data for spironolactone remain limited10. Rigorous cost-utility analyses in resource-limited settings are therefore required before widespread adoption.

In this real-world cohort study emulating a randomized clinical trial, use of finerenone in patients with chronic kidney disease (CKD) and type 2 diabetes (T2D) was associated with a significantly lower risk of all-cause mortality, as well as cardiovascular and kidney events, compared with spironolactone. Finerenone use was also linked to a reduced incidence of hyperkalemia, a common safety concern with MRA therapy. These findings suggest that finerenone may offer improved clinical outcomes and a more favorable safety profile in this high-risk population. However, confirmation through prospective, long-term randomized clinical trials is needed to validate these real-world observations and inform optimal treatment strategies in patients with CKD and T2D.

Methods

The study adhered to the ethical principles of the Declaration of Helsinki33 and followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline. Ethical approval for this study was obtained from the Institutional Review Board of Chi Mei Hospital (approval number: 11210- E01). In addition, all participating healthcare organizations contributing data to the TriNetX Research Network had obtained institutional review board (IRB) or ethics committee approval to share de-identified data. The use of de-identified, aggregated data was deemed exempt from informed consent by the Western Institutional Review Board. This exemption is based on the TriNetX platform’s data de-identification process, which complies with the Health Insurance Portability and Accountability Act (HIPAA) Privacy Rule and General Data Protection Regulation (GDPR) standards. The data are formally certified by a qualified expert as de-identified, contain no protected health information (PHI), and are presented only as aggregated summaries, thus, the research is considered non-human subjects research and exempt from informed consent requirements.

Data source

This study used data from the TriNetX platform, which aggregates electronic health records (EHRs) from the Global Collaborative Network, comprising 146 healthcare organizations (HCOs). These HCOs span 21 countries across the Americas, Europe/Middle East/Africa (EMEA), and Asia-Pacific (APAC) regions, including the United States, United Kingdom, Germany, France, Israel, Japan, Taiwan, and Australia. The available EHR data include patient demographics (including sex as recorded in the individual’s EHR), diagnoses, medications, procedures, laboratory tests, genomics, visits, and details related to socioeconomic factors and lifestyle. Race and ethnicity are recorded separately in TriNetX, consistent with clinical documentation standards in the United States and other participating regions. This network encompasses both insured and uninsured patients from a range of clinical settings, including hospitals, primary care units, and specialty clinics34,35,36,37,38,39,40,41,42.

Target trial specification and emulation

The target trial emulation framework was used to design and analyze this observational study, replicating an RCT structure using observational data43,44. This framework has been widely applied in clinical research, particularly in studies on CKD45,46. To emulate randomization, the finerenone and spironolactone groups were propensity-score matched for these covariates47. Details of the target trial specification, including eligibility criteria, treatment strategies, outcomes, and analysis approach, are provided in Supplementary Table 1 and Supplementary Fig. 1.

Eligibility criteria

The target trial emulation included adults aged ≥18 years with CKD and T2D who had medical encounters between July 2021 and September 2024. This study period was selected as finerenone was first approved by the US FDA in July 2021. Incident CKD was defined as two eGFR values < 60 mL/min/1.73 m², measured at least 90 days apart, using the Modification of Diet in Renal Disease Study (MDRD) formula. Patients were grouped based on the first prescription (new users) of either finerenone or spironolactone, marking the baseline or index event. Eligible patients had no history of either MRAs use in the preceding 6 months before index event. Exclusion criteria included prior eGFR values < 15 mL/min/1.73 m², end-stage renal disease (ESRD), or recent events such as acute coronary syndromes, stroke, cardiac arrest, cardiogenic shock, or ever dialysis within 60 days of the index prescription. Other exclusions included medical contraindications (e.g., adrenal insufficiency such as Addisonian crisis or history of strong CYP3A4 inhibitors) and safety concerns (e.g., hyperkalemia, defined as serum potassium ≥5.5 mmol/L, as per the safety warnings for either finerenone or spironolactone). Eligibility criteria and baseline covariates were evaluated during the baseline period, defined as the one-year period prior to the index event. (Supplementary Table 2 and Supplementary Fig. 1).

Treatment strategies

Two treatment strategies were compared: initiation of finerenone or spironolactone at baseline (index event). Treatment initiation was defined as the first prescription of the respective medication (new-user design), following an intention-to-treat approach, with no adjustments for medication adherence, switches, or addition of other MRAs.

Prespecified outcomes

The primary outcomes were major adverse cardiovascular events (MACE), major adverse kidney events (MAKE), and all-cause mortality. MACE was defined as acute coronary syndromes, nonfatal stroke, hemorrhagic stroke, cardiac arrest, cardiogenic shock. To complement our primary definition, we additionally evaluated alternative MACE definitions based on narrower myocardial infarction criteria. MAKE was defined as progression to end-stage kidney disease (ESKD) or initiation of dialysis. The secondary outcome was hyperkalemia, assessed at thresholds of ≥5.5 mEq/L. Each patient was followed from the index event until the occurrence of an outcome of interest, loss to follow-up, death, administrative censoring (March 14, 2025), or a maximum follow-up period of 1.5 years, whichever occurred first (Supplementary Table 3).

Covariates

The predefined covariates, selected based on clinical knowledge and previous evidence, were measured within 1 year before the index event to balance treatment group differences. These included sociodemographic factors—age, race, sex, and socioeconomic status—as documented in the electronic health record; laboratory and vital sign measurements (glycated hemoglobin, eGFR, blood pressure, total cholesterol, low-density lipoprotein cholesterol, body mass index, and potassium); medications (insulin, metformin, glucagon-like peptide 1 receptor agonists, sodium-glucose cotransporter 2 (SGLT2) inhibitors, renin-angiotensin system (RAS) inhibitors, β-blockers, calcium channel blockers, aspirin, anticoagulants, and hydroxymethylglutaryl-CoA (HMG-CoA) reductase inhibitors); comorbidities (ischemic heart disease, heart failure, hypertension, cerebrovascular disease, peripheral vascular disease, atrial fibrillation and flutter, acute kidney injury, anemia, chronic obstructive pulmonary disease, liver disease, systemic connective tissue disorders, neoplasms, hyperuricemia, sleep apnea, depressive episodes, and anxiety disorders); diabetic complications (ophthalmic, neurologic, and circulatory); and lifestyle factors (nicotine dependence and alcohol-related disorders) (Supplementary Table 4).

Statistical analysis

To minimize confounding and emulate the randomization process, one-to-one propensity score matching (PSM) was performed using logistic regression with greedy nearest-neighbor matching and a caliper width of 0.1 pooled standardized differences42. Adequate balance between the matched groups was considered achieved when the standardized difference was less than 0.1, indicating minimal differences48. To address missing laboratory data (e.g., BMI, HbA1c, SBP, lipids, and UPCR), we included a distinct “no measurement” category for each variable in the propensity score model.

For the primary analysis, hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using Cox proportional hazards models under an intention-to-treat framework. Cumulative incidence curves were generated using the Kaplan-Meier method, and differences were evaluated with the log-rank test. Absolute risk reduction (ARR) was calculated as the difference in event rates between groups, and number needed to treat (NNT) was derived as its reciprocal (NNT = 1/ARR). Confidence intervals for ARR and NNT were estimated using standard binomial methods. The incidence of hyperkalemia was assessed using odds ratios (ORs) with corresponding p values, which were derived from the chi-square test. To assess robustness to unmeasured confounding, we calculated E-values, with higher values indicating greater resistance to bias49.

Predefined subgroup analyses were conducted in separate PSM cohorts stratified by clinically relevant baseline characteristics, including age, sex, glycated hemoglobin level, baseline eGFR, proteinuria, heart failure, and use of SGLT2 inhibitors and RAS inhibitors, and enrollment year, to examine potential effect modification. To assess treatment persistence, we calculated the proportion of patients with ongoing prescriptions at 6 and 12 months after initiation. As a sensitivity analysis, we performed landmark analyses at these timepoints, including only patients who were event-free and remained on treatment to examine the associations with subsequent outcomes. Additional analyses included conducting analyses before propensity score matching (PSM), excluding events within 30 days of treatment initiation to reduce misclassification bias, and limiting dose-specific analyses to participants with recorded drug doses. To minimize bias from treatment switching, patients who transitioned to the alternative medication class were excluded. We also restricted analyses to patients with complete laboratory data to assess the impact of missingness. A negative control outcome analysis was performed by examining the association between treatments and overall cancer incidence, for which no association was expected50,51,52,53. To avoid bias associated with the interpretation of composite endpoints, we further conducted specificity analyses on the individual components of the outcomes of interest.

All analyses were conducted using the TriNetX platform (TriNetX LLC, Cambridge, MA, USA) and R software (version 4.4.1; The R Foundation for Statistical Computing, Vienna, Austria). A two-sided p value < 0.05 was considered statistically significant. Data were collected and analyzed from March 14, 2025, to April 30, 2025.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.