Author's reply: Detailed Causality between Sodium-glucose Cotransporter 2 Inhibitor and Tubulointerstitial Nephritis

The Authors Reply Thank you for the valuable comments regarding our recent paper (1). We completely agree that we should avoid unneccessary termination of sodium-glucose cotransporter 2 (SGLT2) inhibitors, since the initial drop in glomerular filtration rate (GFR) likely reflects the possible mechanism by which the drug acts. SGLT2 inhibitors improve glomerular hyperfiltration based on tubule-glomerular feedback, which induce an acute reduction in the GFR, and thus has a renal protective effect (2). However, we have to question whether the cause of a drop in the GFR can be attributed solely to SGLT2 inhibitors and not to other causes of acute kidney injury. It might be impossible to diagnose tubulointerstitial nephritis (TIN) based on the initial drop in the GFR except by a renal biopsy, so we reported our case.

We performed a renal biopsy in this case because our patient had several characteristics that were considered different from those of a typical initial drop. First, her acute decline in the renal function was greater than what is usually seen with an initial drop, as the eGFR decreased by 40% from baseline, and the serum creatinine level worsened 1.6-fold (0.89 to 1.43 mg/dL) before termination of the SGLT2 inhibitor. It was reported that an acute eGFR reduction ≥10% was observed in 45% of SGLT2 inhibitor-treated participants, but a >30% reduction was observed in only 0.5% of participants, making it a rare event (3). Second, the decrease in the renal function was accompanied by proteinuria and an elevation of tubular markers. SGLT2 inhibitors may reduce albuminuria by improving glomerular hyperfiltration. In our case, the abnormal elevation of tubular markers decreased to approximately 10% after termination of the SGLT2 inhibitor, and after 3 months, the levels were normalized with the disappearance of proteinuria after steroid use. In addition, uric acid levels spontaneously normalized before intervention, and there were no findings suggestive of Fanconi's syndrome.

Regarding the renal pathology, we were interested in the location of the TIN, but it was difficult to identify specific patterns or specific areas of high inflammation. TIN is generally diffuse, but cell infiltration was predominant around the tubule rather than around the glomerulus. Although clear tubulitis was not found, a renal biopsy revealed TIN and the absence of osmotic injury, such as tubular epithelial cell vacuolation, renal damage known to be associated with the use of SGLT2 inhibitors, as well as poor findings of diabetic changes, which might have induced proteinuria.

The authors state that they have no Conflict of Interest (COI).