Predicting Postoperative Acute Kidney Injury through Integration of Genetics and Electronic Health Records

点击翻译按钮获取中文摘要

ABSTRACT Candidate: Dr. Nicholas Douville is a critical care anesthesiologist with board certification in anesthesiology at the University of Michigan. Through completion of the Medical-Scientist Training Program (MSTP) and clinical training in Anesthesiology and Critical Care Medicine, Dr. Douville has developed expertise in bioinformatics and perioperative outcomes research. This proposal builds on Dr. Douville’s expertise, providing protected time for training in bioinformatics, data science, and statistical techniques necessary to drive forward the prediction of patients at risk for postoperative acute kidney injury (poAKI). Environment: The University of Michigan is the coordinating center for the Multicenter Perioperative Outcomes Group (MPOG), an international consortium of over 50 anesthesiology and surgical departments with perioperative information systems. Dr. Sachin Kheterpal, MD, MBA is the primary mentor for Dr. Douville, and is the Director for MPOG and ex-member of the NIH Precision Medicine Initiative Advisory Panel. The proposed research will be completed under the guidance of Dr. Kheterpal, as well as co-mentors Cristen Willer, PhD (genetics) and Michael Heung, MD (nephrology), and Daniel Clauw, MD (general career guidance). Background: Acute Kidney Injury (AKI) occurs after 6-13% of non-cardiac procedures, and is associated with a six-fold increase in postoperative mortality. Numerous metrics for identifying at-risk patients have been developed incorporating preoperative and intraoperative data. Family and linkage studies have demonstrated renal dysfunction to be a heritable trait, however, the specific genetic underpinnings of acute, as opposed to chronic, kidney injury has only recently been explored in the perioperative period. These studies were limited by small sample size, did not consistently identify variants, and failed to utilize advanced genetic analysis, such as polygenic risk scores (PRS). Furthermore, predictive algorithms for poAKI fail to incorporate any genetic data, despite evidence that this may explain a substantial portion of the overall risk. Research: Our goal is to assist perioperative providers in improving patient outcomes through a unified platform that identifies patient attributes that may affect their care and stratifies the risk of key perioperative complications. Our proposed algorithm will combine clinical information (divided into preoperative and intraoperative data) with genetic information to identify patients with greater than baseline risk for developing poAKI. We will validate our methodology using clinical and genetic data from our institutional Michigan Genomics Initiative (MGI), where we have genetic data on over 70,000 individuals who have had surgery at the University of Michigan. We will first develop a polygenic risk score for poAKI (Aim 1). The polygenic risk score (developed in Aim 1) will then be integrated with other variables from the electronic health record (EHR) to provide a comprehensive risk assessment which will be benchmarked against a validated metric (Aim 2).