This retrospective cohort study included patients with type 2 diabetic kidney disease (T2DKD) treated with finerenone at Xiamen Humanity Hospital Endocrinology Department (January 2023–May 2025). Diagnosis followed 2025 ADA criteria: urinary albumin-to-creatinine ratio (UACR) ≥30 mg/g or 24 h urinary albumin ≥30 mg or eGFR <60 mL/min/1.73 m² sustained ≥3 months after excluding other causes (Epstein, 2008).

Inclusion required: ADA-defined T2DKD; baseline eGFR ≥60 mL/min/1.73 m²; finerenone treatment ≥1 month (stable dose 10/20 mg/day adjusted by eGFR); complete pre-treatment (≤1 month) and post-treatment (1–3 months) lab records.

Exclusion criteria: active/treated malignancy (≤1 year); rheumatic disease (e.g., SLE, RA); acute gout/uric acid-lowering drug use (≤3 months); Child-Pugh ≥ B liver dysfunction; pregnancy/lactation; baseline serum potassium ≥5.0 mmol/L or finerenone contraindications; potassium/uric acid-modifying drug use (≤3 months).

The patient selection process is summarized in Supplementary Figure S1.

Baseline data were collected within 1 month prior to finerenone initiation, including demographic characteristics (age, sex), metabolic profiles (fasting plasma glucose, HbA1c, triglycerides, serum uric acid, urine pH), renal parameters (urine albumin-to-creatinine ratio, serum creatinine, estimated glomerular filtration rate calculated via CKD-EPI equation), serum potassium levels, and concurrent medication use (GLP-1 receptor agonists, SGLT2 inhibitors, and initial finerenone dose of 10 mg/day). Follow-up assessments at 1–3 months post-treatment involved repeated measurements of all baseline parameters, with specific focus on changes in urine albumin-to-creatinine ratio, estimated glomerular filtration rate, and serum uric acid levels, as well as monitoring for treatment-emergent hyperkalemia (defined as serum potassium ≥5.5 mmol/L). All laboratory tests were conducted under standardized fasting conditions.

Hyperuricemia was defined as a baseline SUA level >420 μmol/L for men or >360 μmol/L for women.

The primary endpoint was the absolute change in SUA from baseline to follow-up. Subgroup analyses compared this change between patients with and without baseline hyperuricemia. Secondary endpoints included: (1) the relative change in UACR; and (2) the absolute change in eGFR (Fidelity Study Group, 2024a; Fidelity Study Group, 2024b).

Safety assessments focused on the incidence of treatment-emergent hyperkalemia (defined as serum potassium ≥5.5 mmol/L on at least one consecutive measurement after treatment initiation).

The finerenone dosing strategy adhered to the prescribing information: an initial dose of 10 mg once daily for patients with serum potassium ≤5.0 mmol/L and eGFR ≥25 mL/min/1.73 m². The dose could be escalated to 20 mg once daily after 4 weeks if serum potassium was ≤4.8 mmol/L and eGFR had not declined by >30% from baseline. The dose was maintained at 10 mg if serum potassium was between 4.8 and 5.5 mmol/L or if eGFR decline exceeded 30%. Finerenone was discontinued if serum potassium exceeded 5.5 mmol/L and could be re-initiated at 10 mg once daily once it fell to ≤5.0 mmol/L (Halimi et al., 2025).

A total of 124 patients diagnosed with type 2 diabetic kidney disease (DKD) were included in this retrospective analysis. The baseline demographic, clinical, and laboratory characteristics are summarized in Table 1. The mean age was 55.8 ± 11.6 years, with males comprising 64.5% of the cohort. A significant proportion of patients were on concurrent medications, including SGLT2 inhibitors (65.3%) and GLP-1 receptor agonists (38.7%). At baseline, 44.4% (n = 55) of patients met the criteria for hyperuricemia.

The impact of finerenone on key clinical parameters was assessed using a linear mixed model (LMM) for repeated measures data, with results presented in Table 4. Serum Uric Acid (SUA): LMM analysis demonstrated a significant reduction in SUA levels after treatment, with an adjusted mean difference of −47.9 μmol/L (95% CI: −63.5 to −32.3; p < 0.001). Urinary Albumin-to-Creatinine Ratio (UACR): UACR decreased significantly, with a mean reduction of −148.2 mg/g (95% CI: −202.6 to −93.9; p < 0.001),corresponding to a 39.4% decline from baseline. Estimated Glomerular Filtration Rate (eGFR): A small but statistically significant decline in eGFR was observed (mean change: −2.7 mL/min/1.73 m²; 95% CI: −5.0 to −0.4; p = 0.019). The individual trajectories and paired changes for UACR, SUA, and eGFR in each patient are visually presented in Figure 1. Paired t-test analysis further verified the significant changes in these key parameters, with detailed results shown in Table 2.

The reduction in SUA was heterogeneous across predefined subgroups (Table 3). Hyperuricemia Status: The SUA-lowering effect was markedly more pronounced in patients with baseline hyperuricemia (−88.6 μmol/L; 95% CI: −104.7 to −72.5) compared to those without (−16.6 μmol/L; 95% CI: −29.8 to −3.4), with a statistically significant interaction (p for interaction = 0.003). Concomitant Medications: The magnitude of SUA reduction was not significantly modified by concomitant use of SGLT2 inhibitors (Heerspink et al., 2019) (p for interaction = 0.750) or GLP-1 receptor agonists (p for interaction = 0.680).

The unadjusted correlations between the magnitude of SUA reduction (ΔSUA) and other continuous clinical variables are displayed in Figure 2.

To identify independent predictors, candidate variables were first screened using LASSO regression with 10-fold cross-validation (Figure 3), where the optimal lambda (λ) was selected at the point of minimum mean squared error. The regression coefficients of all variables screened by LASSO regression are presented in Table 4, and the optimal lambda (λ) was selected at the point of minimum mean squared error. This process identified 14 candidate variables for inclusion in the final multivariable linear regression model (results detailed in Table 5 and visualized in Figure 4).

The multivariable model (Table 5) demonstrated that baseline serum uric acid was the strongest independent predictor of ΔSUA (β = 0.58 per 1 μmol/L increase, p < 0.001). HbA1c also showed a weak positive association (β = 7.48, p = 0.040). The model found no significant difference in the magnitude of SUA reduction when comparing the 20 mg dose to the 10 mg reference dose (β = −7.40, p = 0.599). Similarly, SGLT2 inhibitor use was not an independent predictor of ΔSUA in this adjusted model (β = −5.06, p = 0.733). Consistently, in a linear mixed model additionally adjusted for SGLT2i and GLP-1 use, the time effect of finerenone on SUA reduction remained significant (β = −47.8, 95% CI: −63.4 to −32.2, p < 0.001), and neither concomitant medication was a significant independent predictor.

Finerenone was generally well-tolerated. Treatment-emergent hyperkalemia (serum potassium ≥5.5 mmol/L) was observed in one patient (0.8%). The condition resolved after temporary drug discontinuation and dietary adjustment, and the patient was successfully re-challenged with finerenone at 10 mg daily without recurrence.

The reduction in SUA observed in our cohort differs from findings in other studies. We propose that this discrepancy is primarily attributable to the distinct characteristics of our cohort, notably the high prevalence and severity of baseline hyperuricemia (Inker et al., 2012; Jin et al., 2023; Johnson et al., 2003). The mean baseline SUA in our cohort (400.0 μmol/L) was substantially higher than typically reported in RCT populations (Kolkhof and Bärfacker, 2017). This explanation is strongly supported by our subgroup analysis, which showed that the significant SUA reduction was almost entirely driven by patients with baseline hyperuricemia (mean decrease of 88.6 μmol/L). This indicates that any potential urate-modulating effect of finerenone may be most discernible in individuals with pre-existing elevated SUA (Kolkhof et al., 2015). Alternative explanations for the observed reduction must be acknowledged, including regression to the mean—a phenomenon likely in this short-term study of patients selected for high baseline SUA—and the absence of a control group. The small but significant eGFR decline (−2.7 mL/min/1.73 m²) is consistent with the acute hemodynamic effect expected with MRAs and mirrors findings from RCTs. (Levey et al., 2006; Liu et al., 2022). That this decline coincided with, rather than attenuated, the SUA reduction hints at a mechanism potentially independent of pure glomerular hemodynamics, possibly involving improved tubular handling of urate secondary to reduced renal inflammation.

A key ancillary finding is that the magnitude of SUA reduction appeared independent of concomitant use of SGLT2 inhibitors or GLP-1 receptor agonists (Liu et al., 2015). Subgroup and interaction analyses in both linear mixed and multivariable regression models found no significant effect modification or independent predictive value for these agents (Liu, 2024). This suggests that the observed association between finerenone and lower SUA is not merely a confounded effect of these commonly co-prescribed, urate-friendly medications. However, the possibility of residual confounding or unmeasured synergistic effects cannot be excluded.

The primary mechanism of finerenone—blocking mineralocorticoid receptor-driven inflammation and fibrosis—is well-established and aligns with the robust 39.4% reduction in UACR we observed (Naj et al., 2022; Navaneethan et al., 2025). We hypothesize that this attenuation of renal inflammation may secondarily improve the local environment for uric acid excretion, a pathway supported by preclinical links between inflammation and tubular urate transport (Perakakis et al., 2023; Perkovic et al., 2019). Regarding clinical relevance, it is important to note that this study was not designed to and cannot establish whether the observed short-term SUA reduction translates into additional long-term renal or cardiovascular protection beyond finerenone’s proven benefits. Our findings are hypothesis-generating (Pitt et al., 2021). They highlight a subgroup of patients—those with DKD and hyperuricemia—in whom finerenone might offer a particularly compelling therapeutic profile, potentially addressing two interrelated risk factors simultaneously (Sánchez-Lozada et al., 2005; Tentolouris et al., 2025). Future prospective studies are needed to confirm the durability of this effect and its impact on hard endpoints.

Our study has important limitations that necessitate cautious interpretation. First and foremost, the lack of a control group precludes definitive causal attribution of the SUA reduction to finerenone; observed changes could reflect regression to the mean or other unmeasured temporal factors (Vaduganathan et al., 2025). Second, the retrospective, single-center design may introduce selection bias and limits generalizability. Third, the short follow-up (1–3 months) only captures initial biochemical responses; long-term effects on SUA dynamics remain unknown (Wada et al., 2025; Wang et al., 2025). Fourth, we lacked data on important lifestyle confounders (e.g., diet, alcohol intake) and biomarkers of inflammation or urate transport, which limits mechanistic insight (Zhu et al., 2011). Finally, the sample size, while adequate for detecting the main effect, provides limited power for some subgroup and multivariable analyses. Similar limitations have also been reported in other real-world studies on finerenone for diabetic kidney disease (Zhang et al., 2025).