Factors that confound the prediction of renal medullary oxygenation and risk of acute kidney injury from measurement of bladder urine oxygen tension

Abstract Aim Urinary oxygen tension (uPO 2 ) may provide an estimate of renal medullary PO 2 (mPO 2 ) and thus risk of acute kidney injury (AKI). We assessed the potential for variations in urine flow and arterial PO 2 (aPO 2 ) to confound these estimates. Methods In 28 sheep urine flow, uPO 2 , aPO 2 and mPO 2 were measured during development of septic AKI. In 65 human patients undergoing cardiac surgery requiring cardiopulmonary bypass (CPB) uPO 2 and aPO 2 were measured continuously during CPB, and in a subset of 20 patients, urine flow was estimated every 5 minutes. Results In conscious sheep breathing room air, uPO 2 was more closely correlated with mPO 2 than with aPO 2 or urine flow. The difference between mPO 2 and uPO 2 varied little with urine flow or aPO 2 . In patients, urine flow increased abruptly from 3.42 ± 0.29 mL min −1 to 6.94 ± 0.26 mL min −1 upon commencement of CPB, usually coincident with reduced uPO 2 . During hyperoxic CPB high values of uPO 2 were often observed at low urine flow. Low urinary PO 2 during CPB (<10 mm Hg at any time during CPB) was associated with greater (4.5‐fold) risk of AKI. However, low urine flow during CPB was not significantly associated with risk of AKI. Conclusions uPO 2 provides a robust estimate of mPO 2 , but this relationship is confounded by the simultaneous presence of systemic hyperoxia and low urine flow. Urine flow increases and uPO 2 decreases during CPB. Thus, CPB is probably the best time to use uPO 2 to detect renal medullary hypoxia and risk of post‐operative AKI.