Abstract Background The storage lesion is a multifactorial degradation effect seen in units of banked blood, and results in an increase in the time-constant of oxygen-unloading from red cells (tau). We have previously shown the relevance of this effect in observational data from a recent clinical trial of prolonged normothermic kidney perfusion as a preservation technique ahead of transplantation. We sought direct confirmatory evidence for diffusion-limited oxygen release by stored red cells (RBCs) during normothermic kidney perfusion in a precisely controlled experimental setting. Methods A novel twinned-circuit kidney perfusion system using common, recirculated dialysis was developed, achieving physical separation of RBCs of two conditions whilst allowing non-cellular components of the perfusate to equilibrate throughout. Using human kidneys unsuitable for transplantation, perfusion was alternated between stored and rejuvenated RBCs (treated with PIPA solution) originating from the same donation, to manipulate oxygen-unloading whilst holding all non-RBC parameters constant, without interrupting blood flow. Results Tau improved with PIPA treatment (1.6 to 1.1s). Rejuvenated oxygen-unloading kinetics reversibly and significantly improved the kidney’s oxygen diffusion capacity and increased cortical oxygen tension by 60%. This supported observations from our clinical trial, where increased tau was strongly associated with restricted renal oxygen consumption. Conclusions We have confirmed that oxygen delivery to tissues can become diffusion-limited during normothermic perfusion with stored blood. This effect is reversible by treatment with PIPA solution. This has important implications for ex-vivo normothermic organ perfusion, and potentially for other scenarios such as major hemorrhage, elective surgery with high blood product utilisation, and paediatric transfusion.