The Role of Monitoring the Partial Pressure of Oxygen in Urine for Early Detection of Acute Kidney Injury

Salt Lake Community College Westpointe Campus, Salt Lake City, UT

Acute kidney injury (AKI) is defined as an abrupt decrease in renal function due to structural injury. AKI has a large effect on critical care patients due to a relatively high occurrence and association with poor outcomes. Currently, the diagnostic tools that we have, serum creatinine and urine output, are delayed indicators of the presence of injury and simply capture the drop in renal function. However, there is evidence that renal hypoxia plays a role in the development of most forms of AKI. Furthermore, there is a correlation between the partial pressure of oxygen in urine (PuO₂) and the renal tissue oxygen levels. Thus, the purpose of this work was to establish noninvasive PuO₂ as an early indicator of AKI and as a tool to guide interventions. To achieve this goal, we focused our work on three specific aims. In Aim 1, we tested the feasibility of measuring PuO₂ outside the body with a novel device. Based on data from a large clinical study, we found that patients who developed AKI after cardiac surgery had lower intraoperative PuO₂ compared to patients who did not develop AKI. We also determined that reliable PuO₂ measurements could be acquired at much more frequent intervals than current monitoring methods. In Aim 2, we built upon the work of Aim 1 and incorporated the PuO₂ data into a prediction algorithm. We used a forward feature selection process to identify the optimal preoperative patient features. We also found the optimal PuO₂ based feature using receiver operator characteristic curves. Next, we used machine learning techniques to train and test the performance of a preoperative-only model and a model that included the same preoperative features plus the PuO₂ feature. We found that the model that included the PuO₂ feature was a significantly better predictor than the preoperative-only model. Lastly, in aim 3, we explored the relationship between critical care interventions and PuO₂. This study was built upon a porcine hemorrhagic shock model. We found that when targeting a mean arterial pressure of 65 mmHg following shock, more resuscitation support led to improved PuO₂ independent of changes in mean arterial pressure. Future studies are needed to expound upon these relationships. Ultimately, this work was the beginning of developing a noninvasive PuO₂ monitor and related algorithm that could potentially improve outcomes in cardiac surgery and, more generally, in resource-limited or austere settings such as during space travel.