EARLIER use of sodium-glucose cotransporter-2 (SGLT-2) inhibitors following severe acute kidney injury (AKI) may be associated with better kidney outcomes in adults with Type 2 diabetes, new research suggests.
Patients prescribed an SGLT-2 inhibitor within 30 days of leaving hospital had a 38% lower relative risk of adverse kidney events compared with those starting treatment between 31 and 90 days after discharge.
SGLT-2 inhibitors are established treatments for reducing kidney and cardiovascular complications in people with Type 2 diabetes, chronic kidney disease, and heart failure. However, the optimal time to initiate treatment following severe AKI remains uncertain.
Clinicians may delay SGLT-2 inhibitors following AKI because of concerns including fluid balance, changes in kidney function, recurrent kidney injury, and diabetic ketoacidosis.
Researchers therefore investigated outcomes associated with earlier versus later treatment in adults with Type 2 diabetes recovering from dialysis-requiring AKI.
The cohort study used electronic health records from the TriNetX Global Collaborative Network between 2015 and 2024. Eligible patients had required dialysis during an AKI episode but subsequently discontinued dialysis after hospital discharge.
Amongst 129,179 eligible patients, 4,406 started an SGLT-2 inhibitor within 90 days of discharge. Of these, 2,268 began treatment within 30 days, while 2,138 started between day 31 and 90.
Researchers used propensity score matching to account for differences between the groups, resulting in 1,912 matched patients in each treatment group.
Adverse kidney events occurred in 3.9% of patients who started SGLT-2 inhibitors within 30 days, compared with 6.2% of those who started treatment later.
After adjustment, earlier treatment was associated with a 38% lower relative risk of adverse kidney events.
The composite kidney outcome included restarting dialysis, developing end-stage kidney disease, or kidney function declining to an estimated glomerular filtration rate (eGFR) below 15 mL/min/1.73 m².
Importantly, the association appeared to be driven primarily by preservation of kidney function rather than avoidance of further dialysis. Severe decline in kidney function occurred in 3.4% of the early-treatment group compared with 5.3% of the later-treatment group, while there was no significant difference in restarting dialysis.
No significant differences were observed in all-cause mortality or major adverse cardiovascular events between the two groups.
One reason clinicians may hesitate to prescribe SGLT-2 inhibitors shortly after severe AKI is concern about potential adverse effects.
However, during the 6-month safety assessment, researchers found no significant differences between early and later treatment in recorded urinary tract infections, diabetic ketoacidosis, hypoglycaemia, volume depletion, heart failure, diabetic retinopathy, or genital fungal infections.
Sensitivity analyses using alternative definitions of early treatment, including 15 and 45 day thresholds, produced consistent kidney findings.
Nevertheless, the researchers stressed that 30 days should not be considered a definitive biological cut-off for treatment. Rather, the findings suggest that the early period following hospital discharge warrants further investigation as a potential window for initiating kidney-protective therapy.
As an observational study, the research cannot establish that earlier SGLT-2 inhibitor treatment directly caused the improved kidney outcomes. Unmeasured differences between patients, including their clinical readiness to start treatment, could have influenced both prescribing decisions and outcomes.
The authors conclude that prospective randomised trials are needed to determine the optimal timing, efficacy, and safety of SGLT-2 inhibitor initiation following severe AKI.
Chou Y-T et al. Early Sodium-Glucose Cotransporter-2 Inhibitor Use After Acute Kidney Injury in Type 2 Diabetes. JAMA Netw Open. 2026;9(8).
Featured image: NanSan on AdobeStock
Each article is made available under the terms of the Creative Commons Attribution-Non Commercial 4.0 License.
