Acute kidney injury risk with SGLT2 inhibitors versus sulfonylureas by baseline cardiorenal status in type 2 diabetes: A nationwide cohort study

Technical Summary: Acute Kidney Injury Risk with SGLT2 Inhibitors versus Sulfonylureas by Baseline Cardiorenal Status

Problem Statement
While Sodium–glucose cotransporter-2 inhibitors (SGLT2i) are established for reducing major adverse cardiovascular and renal outcomes in type 2 diabetes mellitus (T2DM), evidence regarding their short-term renal safety—specifically the risk of acute kidney injury (AKI) and acute interstitial nephritis (AIN)—remains limited in Asian populations undergoing routine second-line therapy. Previous randomized controlled trials (RCTs) often enrolled select populations and were not primarily designed to assess short-term AKI or biopsy-confirmed AIN in standard care settings. Furthermore, observational data on the comparative risk of SGLT2i versus sulfonylureas (SUs) as metformin-based second-line therapy is heterogeneous, with varying estimates depending on the comparator and clinical setting. This study addresses the gap in real-world evidence concerning the 180-day AKI risk of initiating SGLT2i compared to SUs, specifically examining how this risk varies by baseline cardiorenal status (heart failure and chronic kidney disease).

Methodology
The study employed a nationwide, retrospective cohort design using Korean National Health Insurance Service (NHIS) claims data from January 1, 2013, to December 31, 2022.

• Study Population: Adults (≥18 years) with T2DM who newly initiated either an SGLT2i or an SU as add-on therapy to metformin between 2016 and 2022. A 3-year washout period ensured new-user status. Individuals with prior AKI, advanced CKD, dialysis dependence, or renal transplantation were excluded.
• Matching Strategy: To enhance comparability, the study utilized a two-step matching process: exact matching on the interval from T2DM diagnosis to treatment initiation, followed by 1:1 propensity score matching (PSM) using greedy nearest-neighbor matching (caliper 0.2 SD). This resulted in a final matched cohort of 251,664 individuals (125,832 per group).
• Exposure: Initiation of SGLT2i (class and ingredient-specific: dapagliflozin, empagliflozin, ipragliflozin, ertugliflozin) versus SU. Cumulative exposure was measured via cumulative defined daily dose (cDDD).
• Outcomes: The primary outcome was incident AKI within 180 days, identified via ICD-10 code N17. The secondary outcome was biopsy-confirmed acute interstitial nephritis (AIN).
• Statistical Analysis: Stratified Cox proportional hazards models were used to estimate adjusted hazard ratios (aHRs), with matched pairs as the stratification variable. Models were adjusted for covariates showing residual imbalance (e.g., heart failure, dyslipidemia, specific medications). Subgroup analyses were conducted by heart failure (HF) and CKD status. Sensitivity analyses included inverse probability of treatment weighting (IPTW), a 7-day lag period, restricted follow-up, and self-controlled risk interval (SCRI) analyses.

Key Contributions and Results

• Primary Finding: In the propensity-matched cohort, SGLT2i initiation was associated with a significantly lower 180-day risk of AKI compared to SU initiation (aHR: 0.63; 95% CI: 0.55–0.72). The incidence rate was 5.90 per 1,000 person-years for SGLT2i versus 8.95 for SUs.
• Effect Modification by Cardiorenal Status:
  • Heart Failure (HF): The protective association persisted in patients with HF (aHR: 0.55) and those without HF (aHR: 0.64), with no significant interaction by HF status.
  • Chronic Kidney Disease (CKD): A significant interaction was observed based on baseline CKD. The lower AKI risk was pronounced in patients without pre-existing CKD (aHR: 0.50). However, in patients with pre-existing CKD, the association was attenuated and not statistically significant (aHR: 0.94; 95% CI: 0.77–1.15).
• Ingredient-Specific Analysis: Estimates for individual SGLT2i agents (dapagliflozin, empagliflozin, ipragliflozin, ertugliflozin) were directionally consistent with the class-level finding, though estimates for less frequently used agents were less precise.
• Secondary Outcomes: Biopsy-confirmed AIN was rare, yielding imprecise estimates that were directionally consistent with the primary AKI findings.
• Robustness: Sensitivity analyses (IPTW, lag periods, SCRI) supported the robustness of the primary findings, suggesting the association was not driven solely by early events or specific analytic approaches.

Significance and Claims
The authors claim that this large East Asian real-world cohort provides critical evidence that SGLT2i initiation as metformin-based second-line therapy is associated with a lower short-term risk of AKI compared to SUs. A central contribution of the study is the clarification that baseline cardiorenal status modifies this risk. Specifically, the study underscores that HF and CKD should not be treated as a single cardiorenal phenotype when interpreting early renal risk; the safety benefit regarding AKI appears preserved in HF but is attenuated in pre-existing CKD.

The paper concludes that these findings support the comparative short-term renal safety of SGLT2i in routine clinical practice while highlighting the importance of distinguishing between HF and CKD when assessing early renal risk. The results suggest that baseline cardiorenal status may inform early renal risk assessment and monitoring strategies for patients initiating second-line therapy. The authors maintain a modest stance, noting that while the findings support short-term safety, residual confounding cannot be excluded, and the attenuation of benefit in CKD patients warrants cautious interpretation and further investigation.