Prevention of Cardiac Surgery–Associated Acute Kidney Injury by Implementing the KDIGO Guidelines in High-Risk Patients Identified by Biomarkers

The PrevAKI-Multicenter Randomized Controlled Trial

Alexander Zarbock, MD; Mira Küllmar, MD; Marlies Ostermann, MD; Gianluca Lucchese, MD; Kamran Baig, MD; Armando Cennamo, MD; Ronak Rajani, MD; Stuart McCorkell, MD; Christian Arndt, MD; Hinnerk Wulf, MD; Marc Irqsusi, MD; Fabrizio Monaco, MD; Ambra Licia Di Prima, MD; Mercedes García Alvarez, MD; Stefano Italiano, MD; Jordi Miralles Bagan, MD; Gudrun Kunst, MD; Shrijit Nair, MD; Camilla L'Acqua, MD; Eric Hoste, MD; Wim Vandenberghe, MD; Patrick M. Honore, MD; John A. Kellum, MD; Lui G. Forni, MD; Philippe Grieshaber, MD; Christina Massoth, MD; Raphael Weiss, MD; Joachim Gerss, PhD; Carola Wempe, PhD; Melanie Meersch, MD


Anesth Analg. 2021;133(2):292-302. 

In This Article


Study Design and Ethics

We conducted a randomized controlled, multicenter, multinational, parallel-group trial in 12 centers across Europe. Institutional review board approval was obtained from the Research Ethics Committee of the Chamber of Physicians Westfalen-Lippe and the Westfalian Wilhelms University Muenster (2017-291-f-S) and the corresponding institutional review boards of all participating centers. The trial was registered before patient enrolment at (NCT03244514, principal investigator: A. Zarbock, date of registration: September 8, 2017). The study was conducted in accordance with the Declaration of Helsinki (version Fortaleza, 2013). Written informed consent was obtained from all participating patients according to local requirements and legislation.


Eligible patients included adults undergoing cardiac surgery involving the use of CPB. High risk for AKI was defined as having urinary [TIMP-2]•[IGFBP7] ≥ 0.3 (Nephrocheck Test, Biomérieux, France) 4 hours after CPB.[7] Patients were excluded if any of the following criteria were present: preexisting AKI (≥KDIGO stage 1 defined by creatinine as well as urine output), need for cardiac assist devices (extracorporeal membrane oxygenation, ventricular assist devices, intraaortic balloon pump), pregnancy or breastfeeding, known glomerulonephritis/interstitial nephritis/vasculitis, CKD with an estimated glomerular filtration rate (eGFR) <20 mL/min/1.73 m2, chronic dialysis dependency, prior kidney transplant within the last 12 months, and participation in another interventional trial within the last 3 months. In addition, patients with any type of relationship with the investigator or used by the sponsor/investigator and patients held in an institution by legal or official order were excluded.[14]

Randomization and Masking

Eligible patients were randomized in a 1:1 ratio to 1 of the 2 treatment arms. Randomization was performed centrally using a web-based randomization system provided by the European Society of Intensive Care Medicine (ESICM). Randomization lists were stratified by center and generated using block randomization with randomly varying block sizes.

Blinding of the intervention was not possible. Study intervention was performed by clinical staff involved in anesthesia and perioperative care and supervised by member of the research team on a daily basis. End point assessment was performed by a blinded researcher not involved in providing anesthesia and perioperative care. Patients remained blinded to allocation as they were unaware of the study-related procedures.


Patients scheduled for cardiac surgery were approached by a member of the research team before surgery and invited to give informed consent for participation in the study. The surgical procedure and perioperative care were performed according to the standard of care at each center. With informed consent in place, urinary [TIMP-2]•[IGFBP7] was measured 4 hours after CPB. Patients with urinary [TIMP-2]•[IGFBP7] ≥0.3 were eligible to be randomized. Patients assigned to the control group were managed according to the standard of care at each center, including the specification to keep mean arterial pressure (MAP) >65 mm Hg and central venous pressure (CVP) between 8 and 10 mm Hg. Patients assigned to the intervention group received treatment according to the KDIGO bundle consisting of the following measures: avoidance of nephrotoxic agents, discontinuation of angiotensin-converting enzyme inhibitor (ACEi), and angiotensin-II receptor blocker (ARB) during the first 48 hours after surgery, avoidance of hydroxyethyl starch (HES), gelatine, and chloride-rich solutions (including 0.9% saline), close monitoring of serum creatinine (every 12 hours), fluid balance and urine output (hourly), avoidance of hyperglycemia in the first 72 hours after surgery (defined as blood glucose levels >150 mg/dL for >3 hours), consideration of alternatives to radiocontrast agents, close hemodynamic monitoring by using a functional hemodynamic monitoring with an optimization of the volume status, and hemodynamic parameters according to a prespecified algorithm (Supplemental Digital Content, Figure S1,


The primary end point was the compliance rate to the KDIGO bundle defined as proportion of patients who were treated according to the trial protocol (bundle fulfilled at all times). Adherence to the bundle was achieved if all elements of the bundle were met (Supplemental Digital Content, Table S1, Secondary end points were the occurrence and severity of AKI within 72 hours (as defined by the KDIGO criteria), free days of vasoactive medications and mechanical ventilation through day 28, renal recovery (defined as serum creatinine <0.5 mg/dL higher than baseline)[13,15] at days 30, 60, and 90, all-cause mortality at days 30, 60, and 90, length of stay in intensive care unit (ICU) and hospital, use of renal replacement therapy (RRT) at days 30, 60, and 90, and major adverse kidney events consisting of mortality, dialysis dependency or persistent renal dysfunction at day 90 (MAKE90).

Clinical variables were extracted from medical records. Adherence with the hemodynamic protocol was documented every 3 hours. Initiation of RRT was at the discretion of the ICU clinicians. Specific criteria for initiation of RRT were not included in the protocol.

Statistical Analysis

Descriptive statistical analyses were performed reporting frequency of categorical variables as well as location and scale statistics of quantitative variables (mean and standard deviation or median and quartiles). Categorical variables were compared between the randomized groups using χ 2 test, and continuous variables were compared using Student t test or Wilcoxon test, as appropriate. The primary statistical analysis of the primary outcome adherence to the trial protocol provided confirmatory statistical evidence on a 5% significance level. Secondary end points were planned a priori. All other statistical analyses were considered exploratory. P values were regarded as noticeable ("significant") in case P ≤ .05 without adjustment for multiplicity. Statistical analyses were performed using SAS software (Version 9.4 for Windows, SAS Institute Inc, Cary, NC) and IBM SPSS Statistics 26 for Windows (IBM Corporation, Somers, NY).

Sample Size

In the primary statistical analysis, the compliance rate of the study patients is estimated. As the compliance rate could not be quantified in advance, we pursued a worst-case approach and assumed a compliance rate of 50%. In this case, the 95% confidence interval (CI) has maximal width. Based on this assumption, with a sample size of n = 140 patients per group, the calculated 95% CI according to Clopper-Pearson ranges from 41% to 59%. Therefore, in case of any observed compliance rate apart from the worst-case scenario, the 95% CI is never wider than 59 − 41 = 18. This corresponds to an estimation of the compliance rate with a precision of ±9%.