Update on Perioperative Acute Kidney Injury

Alexander Zarbock, MD; Jay L. Koyner, MD; Eric A. J. Hoste, MD; John A. Kellum, MD


Anesth Analg. 2018;127(5):1236-1245. 

In This Article

Oliguria in the Perioperative Setting

Oliguria by itself may be an appropriate response in patients with intravascular volume depletion or hypovolemia. In these patients, adequate volume expansion with the appropriate crystalloid (nonhyperchloremic) or colloid should reverse the process. However, oliguria can also be maladaptive in the settings of congestive heart failure, end-stage liver disease with ascites, and acute tubular necrosis/AKI with volume overload. In these settings, the kidney is not able to provide the appropriate natriuresis perhaps due to renal congestion in the setting of increased right-sided vascular pressures or activity of antidiuretic hormone (ADH). Determining the underlying source of oliguria can perhaps lead to improved volume status and eventually improved kidney function.

In prior investigations, isolated urine output–based AKI has been shown to be more common but associated with less severe adverse outcomes than isolated SCr-based AKI.[50–53] In a retrospective cohort study looking at adults without end-stage renal disease who underwent major noncardiac surgery, 4229 subjects had sufficient creatinine and urine output (UOP) measurements for analysis.[50] Adding UOP to the SCr-based AKI increased the incidence of AKI from 8% to 60%. In adjusted analyses, stage 3 UOP-based AKI was associated with a 2.84 (1.41–5.70) increased odds of 30-day mortality compared to those with no AKI, but this was less compared to those with stage 3 SCr-based AKI (5.00 [1.65–15.17]) and those with combined UOP- and SCr-based stage 3 (7.85 [2.76–22.33]). However, Mizota et al[54] published a single-center study demonstrating that in 320 living donors for liver transplantation, 12% had SCr-based AKI, 22% had UOP changes, and 28% had AKI by both criteria. They demonstrated significantly longer ICU and hospital stays in those with isolated UOP-based AKI compared to SCr-based AKI. These studies highlight that oliguria is common in the perioperative period and is associated with adverse outcomes.

Oliguria is usually thought to be due to intravascular volume depletion or prolonged systemic hypoperfusion, both of which are believed to lead to reduced renal perfusion and result in decreased filtered load and urine production. However, oliguria can also occur in the setting of ADH release as well as in the presence of increased aldosterone secretion.[55,56] In the perioperative setting, ADH release can come from a variety of nonrenal stimuli, including pain, nausea, and type of surgery (neurological or otherwise).[56,57] Regardless of its source, oliguria in the ICU has been associated with increased mortality.[51,52,58] In a systematic review and meta-analysis, Prowle and coworkers[59] investigated the incidence of perioperative AKI following major abdominal surgery. While they did not separate UOP-based AKI from SCr-based AKI, they demonstrated that AKI occurred in 13.4% of patients (from 8 studies and 82.514 patients) and was associated with a 12.6-fold (95% CI, 6.8–23.4) increased relative risk of short-term perioperative mortality.[59]

Intraoperative Fluid Balance and UOP

The impact of fluid management on AKI and UOP in the perioperative setting has been investigated. The prevailing thought is that volume expansion could lead to improved hemodynamics and renal perfusion and consequently more UOP and perhaps less AKI. Historically, in 2 separate randomized controlled trials investigating fluid-restrictive versus fluid-liberal intravenous fluid strategies, evidence showed that fluid-liberal strategies led to longer hospital stays and more perioperative complications.[60,61] However, these studies neither demonstrated a difference in AKI rates nor reported UOP-based outcomes.

More recently, 2 small randomized controlled trials demonstrated no difference in intraoperative UOP in patients randomized to liberal (8–10 mL/kg/h) versus restrictive (2–4 mL/kg/h) fluid strategies.[62,63] Both these separate studies in 107 patients undergoing laparoscopic bariatric surgeries and 102 patients undergoing video-assisted thoracoscopic surgery showed no difference in intraoperative UOP and postoperative AKI rates in patients receiving lactated ringers.[62,63] Some have hypothesized that the lack of impact on UOP and AKI outcomes are due to altered distribution and elimination of intravenous fluids in the setting of laparoscopic procedures. These alterations lead to rapid fluid administration (eg, as a part of fluid-liberal strategies) and may preferentially be less responsive to diuretics as well as lead to increased peripheral edema.[64]

However, the results of these smaller studies were not substantiated in a recent large prospective multicenter international trial looking at 3000 patients who had an increased risk for complications undergoing major abdominal surgery and who were randomized to receive either a fluid-liberal or fluid-restrictive strategy in the first 24 hours after surgery (Restrictive Versus Liberal Fluid Therapy for Major Abdominal Surgery trial).[39] A total of 1490 patients in the restrictive arm received a median (interquartile range) of 3.7 (2.9–4.9) L compared to 6.1 (5.0–7.4) L in the liberal group, and this led to increased rates of intraoperative oliguria and lower intraoperative UOP in the restrictive arm (250 [144–440] vs 350 [200–600]; P < .001). In addition to impacting perioperative UOP, there was an increased rate of KDIGO SCr-based AKI in the restrictive group (26.4% vs 19.0% in the liberal group; P < .001). More specifically, patients in the restrictive group were twice as likely to develop stage 2 or 3 AKI compared to those in the liberal arm (odds ratio [OR], 2.02; 95% CI, 1.43–2.85; P < .001). Finally, 0.9% of the restrictive group required RRT compared to 0.3% of the liberal group (hazard ratio, 3.27; 95% CI, 1.01–13.8; P = .048).[39] These emerging data point to the potential dangers of volume restriction on perioperative kidney function and may help guide future clinicians in terms of striking the balance of finding the optimal volume status.

Prior to this large-scale investigation, there had only been a handful of retrospective cohort studies that published data on the interplay between intraoperative UOP and postoperative AKI.[65,66] Mizota et al[66] published a single-center, retrospective cohort study of 3560 patients who underwent major abdominal surgery and investigated several intraoperative oliguria thresholds to determine their association with postoperative AKI. After excluding all patients who received intraoperative diuretics (furosemide, human atrial natriuretic peptide, or mannitol), they demonstrated that patients with UOP between 0.3 and 0.5 mL/kg/h did not have an increased risk of postoperative AKI (adjusted OR, 1.37; 95% CI, 0.88–2.13; P = .160). However, an intraoperative UOP < 0.3 mL/kg/h (which occurred in 11.3% of the cohort) was independently associated with a 2.65-fold (1.77- to 3.97-fold) increase odds of postoperative AKI (P < .001).[66] This effect was present across a variety of subgroups, including those who did and did not receive laparoscopic surgeries and those with and without 10 mL/kg of intraoperative blood loss and across several types of surgeries. While these data/novel cutoffs need to be further validated (eg, in other cohorts like the aforementioned Restrictive Versus Liberal Fluid Therapy for Major Abdominal Surgery trial), it remains to be seen if 0.3 mL/kg/h is a better threshold for the prediction of postoperative AKI and other adverse outcomes. However, it is line with emerging data that demonstrate that the intensive monitoring of UOP in hospitalized patients (defined as no gaps in urine output data for >3 hours) leads to improved AKI detection, decreased incidence of fluid overload, and improved survival in those who develop AKI.[51] Thus, we anticipate future studies to attempt to balance the benefits of a postoperative fluid-liberal strategy with the risk of volume overload and to determine the optimal resuscitation strategy to prevent postoperative oliguria and the resultant AKI.

Furosemide Stress Test

The majority of this prior work has focused on fluid balance and monitoring urine output during surgeries to predict AKI. More recently, several studies have investigated the kidney's UOP response to a protocoled dose of furosemide to predict adverse outcomes in the setting of early AKI (furosemide stress test).[67–70] Furosemide is an ideal agent for interrogating renal function because it requires several aspects of the nephron to be intact. On delivery to the kidney through systemic circulation, furosemide is transported across from the basolateral side of the proximal tubule through the tubular cells and into the urinary space through the organic anion transporters.[71] After that it requires a functioning thick ascending limb of the loop of Henle, where it acts on the sodium, potassium 2 chloride channel, resulting in increased sodium excretion and increased urine output. Finally, it requires an intact distal nephron and collecting ducts to maintain the diuresis.

Initially, Chawla et al[67] demonstrated that the 2-hour cumulative UOP following 1 or 1.5 mg/kg of intravenous furosemide in patients in euvolemic or hypervolemic ICU with stage 1 or 2 AKI provided an area under the curve (AUC; SE) of 0.87 (0.09) for the prediction of progression to stage 3 AKI. A cutoff of <200 mL of urine in the first 2 hours following the furosemide challenge provided 87% sensitivity and 84% specificity for the progression to stage 3. While the finding of this study has been replicated in other similar mixed-medical patients in the surgical ICU, the authors of 2 other studies used a modified retrospective version of this concept to investigate the ability of furosemide responsiveness to predict AKI outcomes.[69,72] In a retrospective investigation of infants undergoing cardiac surgery, they demonstrated that even after correcting for the impact of fluid balance on SCr that decreased UOP following a furosemide challenge predicted postoperative AKI with an AUC of 0.74 at 2 hours but 0.77 at 6 hours.[72] This 6-hour time point is similar to findings published by McMahon et al[70] who demonstrated the 6-hour UOP following a single dose of 100 mg of intravenous furosemide provided an AUC of 0.85 for the development of delayed graft function in adults undergoing deceased donor kidney transplantation. Thus, while there is a wealth of data accumulating about the prognostic benefits of the furosemide stress test, it still requires further prospective investigation across several types of surgical patients.