Increased Hematocrit During Sodium-Glucose Cotransporter-2 Inhibitor Therapy

To the Editor

Treatment with sodium-glucose cotransporter-2 (SGLT-2) leads to increased hematocrit. Other than hemoconcentration, this phenomenon could be attributed to enhanced erythropoiesis, as indicated by a rise in plasma erythropoietin (EPO) and reticulocytosis [1]. SGLT-2 inhibitors lead to improved renal cortical oxygenation, reflecting reduced transport activity in proximal tubules [2]. In their recent article in JOCMR, Sano et al proposed that improved survival of cortical peritubular interstitial cells with improved cortical oxygenation is the cause of increased EPO with SGLT-2 inhibition [3]. While this hypothesis is interesting, it is not evidence-based, and we would like to propose a more likely alternative explanation, in line with the current understanding of EPO regulation [4], namely intensified hypoxia at the renal cortico-medullary junction.

EPO is a hypoxia-triggered gene, up-regulated by hypoxia-inducible factors (HIFs). HIFs are heterodimers consisting of α and β sub-units, that upon attachment to hypoxia response elements along nuclear DNA strands induce trans-activation of numerous HIF-dependent genes, including EPO [4]. Under normoxic conditions, cytoplasmic HIF-α subunits undergo proteasomal degradation, permeated by specific HIF-prolyl hydroxylases. These enzymes are inhibited under hypoxic conditions, leading to cytoplasmic accumulation of HIF-α subunits. Consequently, HIF-α subunits undergo nuclear translocation and binding with HIF-β subunits, promoting HIF-mediated gene responses [4].

Renal parenchymal oxygenation declines in the diabetic kidney under experimental settings, reaching 10 mm Hg at the cortico-medullary junction [5]. We have previously reported that such intensified hypoxia in the diabetic rat kidney leads to stabilization of α subunits of HIF-2 isoforms in peritubular interstitial cells within the deep cortex and the outer medulla, leading to nuclear immunostaining of HIF-2 [6] and that these cells are the origin of EPO upon HIF-2 stabilization [7]. Furthermore, HIF-prolyl hydroxylases have recently been shown to trigger EPO and erythropoiesis in phase 2 clinical trials [8].

While cortical oxygenation improves with SGLT-2 inhibition, medullary oxygenation substantially declines, both in diabetic and in intact rats, conceivably due to increased solute delivery to distal tubular segments, with enhanced regional oxygen expenditure for tubular transport [2]. Therefore, intensified hypoxia at the deep cortex and outer medulla induced by SGLT-2 inhibition is likely the cause of enhanced EPO transcription and consequent reticulocytosis.

Another observation in line with this narrative is the reversal of post-renal transplantation erythrocytosis with angiotensin II inhibition [9], conceivably reflecting attenuation of renal parenchymal hypoxia by blocking the renin-angiotensin-aldosterone axis (RAAS).

Taken together, we conclude that increased EPO levels and erythropoiesis following SGLT-2 inhibition likely result from enhanced hypoxia at the renal cortico-medullary junction, rather than through the amelioration of cortical oxygenation, as suggested by Sano et al [3].