Perioperative Acute Kidney Injury: Diagnosis, Prediction, Prevention, and Treatment

Common definitions of AKI rely on changes in serum creatinine and absolute urine output to provide an easy to interpret, objective, and validated approach useful for epidemiologic studies; however, each component of common AKI metrics has limitations.⁸  Notably, urine output is significantly influenced by hypovolemia and diuretics, decreasing specificity, while serum creatinine only begins to increase after 50% of functional nephrons are lost, and these changes may not manifest for 48 to 72 h after injury, leading to low sensitivity in previously healthy kidneys.⁷  Furthermore, serum creatinine can be impacted by nonrenal factors including age, muscle mass, nutritional status, liver function, and gastrointestinal elimination, further limiting diagnostic accuracy,^9,10  and not all patients have a recorded baseline serum creatinine. Despite these limitations, serum creatinine remains the standard biomarker for AKI in perioperative practice.¹⁰ 

The KDIGO criteria, currently used to define AKI,⁸  update two previous consensus definitions, (1) Risk, Injury, Failure, Loss, End-stage Renal Disease (RIFLE)¹¹  and (2) AKI-Network (AKIN),¹²  to capture patients with smaller creatinine increases, allowing earlier disease detection.¹³ 

Currently, the perioperative medical community has largely adopted the KDIGO consensus criteria to define postoperative AKI within 7 days of an operative intervention. The KDIGO consensus criteria, published in 2012, included a few modifications from previous consensus definitions. Notably, the surveillance window for the relative changes in serum creatinine was extended to 7 days (the time window for the absolute increase of 0.3 mg/dl or greater rise in serum creatinine remains at 48 h). Additionally, the requirement for stage 3 AKI that serum creatinine increase by 0.5 mg/dl or greater in those with serum creatinine 4.0 mg/dl or greater was removed, and estimated glomerular filtration rate–based criteria for pediatric patients were added (table 1).⁸ 

When KDIGO AKI (stage 1 or greater) persists for more than 7 days after the inciting event, it is termed acute kidney disease.¹⁴  When the disease process persists beyond 90 days, it transitions from acute kidney disease to chronic kidney disease.

AKI represents a clinical syndrome, rather than a single disease, with most cases being multifactorial. Commonly implicated mechanisms include (1) oxido-inflammatory stress (for example, hyperoxia or glyco-oxidative injury), (2) renal hypoperfusion (for example, hypotension or anemia), (3) endogenous or exogenous nephrotoxins, and (4) iatrogenic causes (fig. 1). Venous congestion has also been associated with postoperative AKI through multiple potential mechanisms including (1) hypoperfusion, (2) inflammation, (3) oxidative stress, (4) endothelial activation, and (5) sympathetic activation.¹⁵  Administration of nephrotoxic drugs and the need for intravenous contrast also place surgical patients at risk for acute interstitial nephritis and acute tubular damage.¹⁶  Prerenal etiologies, like true volume depletion and volume dysregulation secondary to congestive heart failure, and acute tubular injury from a myriad of insults including nephrotoxic drugs and intravenous contrast, are common mechanisms in the postoperative setting. The surgical environment also introduces the possibility of iatrogenic injury. Examples include ureteral damage in an intraabdominal surgery and emboli from vascular procedures.

The incidence of AKI varies with socioeconomic status, healthcare setting, and population risk profile.¹⁷  AKI in the surgical population is usually reported separately for (1) cardiac surgery and (2) noncardiac surgery populations. The pooled incidence of any stage of cardiac surgery–associated AKI in a meta-analysis of 320,086 cardiac surgery patients worldwide was 22.3% (95% CI, 19.8 to 25.1).¹⁸  Of the patients, 13.6% were classified as stage 1, 3.8% were stage 2, 2.7% were stage 3 (without renal replacement therapy), and 2.3% were stage 3 (requiring renal replacement therapy).¹⁸ 

AKI is also a frequent complication after thoracic surgery. The incidence varies significantly based upon the extent of the procedure: (1) sublobar resection, 3.8% (2.0 to 6.2%); (2) lobectomy, 6.7% (4.1 to 9.9%); (3) bilobectomy or pneumonectomy, 12.1% (8.1 to 16.6%); and (4) esophagectomy, 10.5% (5.6 to 16.7%).¹⁹  Risk factors for AKI after thoracic surgery are advanced age, male sex, higher body mass index, higher American Society of Anesthesiologists (Schaumburg, Illinois) Physical Status classification, hypertension, diabetes, long-term angiotensin-converting enzyme inhibitor or angiotensin receptor blockers use, hypoalbuminemia, lower pulmonary function, extent or duration of surgery, intraoperative colloid administration, and higher estimated blood loss.¹⁹ 

The incidence of AKI after noncardiac or nonthoracic surgery varies substantially based upon the type of procedure. Data from 37,345 noncardiac surgeries demonstrated a total AKI incidence of 4.7% (KDIGO stage 1, 2.9%; stage 2, 0.7%; and stage 3, 1.1%).⁵  A smaller, retrospective study of 1,869 noncardiac surgical procedures found 128 (6.8%) cases of AKI, divided between 5.4% KDIGO stage 1 and 1.4% KDIGO stage 2 or 3.¹ 

Patients undergoing major vascular surgery have also been shown to have an increased risk for postoperative AKI due to a combination of patient and surgical risk factors. Although the incidence across all vascular procedures is estimated to be as high as 49%,³  the incidence differs dramatically based on specific procedure. For example, procedures on the lower extremities have a relatively low incidence of AKI (4% for infrainguinal bypass; 19% for endovascular revascularization of critical limb ischemia).³  This contrasts with much higher incidence in aneurysm repairs, with some reports as high as 75% for emergent, open repair of a ruptured aneurysm.²⁰ 

Retrospective studies place the incidence in noncardiothoracic, nonvascular surgery between 6 and 13%, depending on the diagnostic criteria used.^21–23  Trauma patients are at increased risk for AKI due to a variety of overlapping mechanisms, including renal hypoperfusion (due to hemorrhagic shock), direct renal injury, rhabdomyolysis, abdominal compartment syndrome, and nephrotoxic exposure.²⁴  Independent risk factors for the development of AKI after trauma include (1) older age, (2) blood transfusion, (3) comorbidities like diabetes or chronic kidney disease, (4) systolic blood pressure upon admission, and (5) lactic acidosis.²⁴  The incidence of AKI in the trauma population ranges from 12% in the general population to 20% in the population admitted to the intensive care unit (ICU).²⁵ 

The incidence of AKI varies dramatically within intraabdominal surgery from 0.2% after appendectomy to 3.5% in exploratory laparotomies.²⁶  The overall incidence is 1.1%. More than half of liver transplant recipients experience postoperative AKI,²⁷  with a recent study demonstrating a significant proportion of stage 2 (24.9%) or stage 3 (17.7%) AKI.²⁸ 

Risk factors for postoperative AKI can generally be divided as (1) patient-related, (2) surgery-related, (3) anesthetic-related, and (4) postoperative risk factors.² 

Although many predictive models have been published, we present five validated models for predicting postoperative AKI (compared in table 2).^29–31  The validated models largely focus on patient factors; however, the two models developed by Kheterpal et al.^29,30  also incorporate surgical factors: (1) emergent surgery and (2) high-risk (or alternatively, intraperitoneal) surgery.

Finally, postoperative inpatient or ICU factors have been implicated in the development of late postoperative AKI, occurring more than 48 h after surgery. These factors commonly include but are not limited to sepsis, mechanical ventilation, acute lung injury, blood transfusion, fluid balance, urinary obstruction, and medications (diuretics, vasopressors, and nonsteroidal anti-inflammatory drugs [NSAIDs]).^2,34 

The relationship between age and sex on the development of AKI remains unresolved,³⁵  with preclinical^36,37  and clinical³⁸  evidence suggesting female sex to be protective against AKI, while the KDIGO guidelines report female sex as a risk factor for AKI.⁸  The discrepancy is hypothesized to result from age-related changes in female-associated hormones such as estrogen and progesterone leading to studies showing postoperative AKI risk being performed in older, postmenopausal populations.³⁵  After adjusting for other risk factors, Privratsky et al. demonstrated that younger females had the lowest odds of postoperative AKI, which increased with age.³⁵  Furthermore, preoperative hormone replacement therapy was not shown to modify the risk for postoperative AKI. Males had a greater risk for postoperative AKI, but that difference gradually narrowed as the patient aged, with no difference noted in the 86- to 90-yr-old age stratification.

The kidney, specifically, the renal medulla, is susceptible to ischemia and hypoxia during periods of hypoperfusion, most commonly presenting as systemic arterial hypotension.³⁹  Whereas direct causal relationships between hypotension and AKI are yet to be definitively established, it is well-described that even brief periods of mean arterial pressure (MAP) less than 65 mmHg during noncardiac surgery are associated with myocardial injury, AKI, and death.^39,40  In this context, the PeriOperative Quality Initiative consensus-building (POQI-3) consensus statement recommends maintaining an intraoperative MAP greater than 65 mmHg to potentially reduce the risk of postoperative AKI, with higher targets in patients with pre-existing hypertension.²  Conversely, the PeriOperative Quality Initiative consensus review did not find sufficient evidence in noncardiac surgery to recommend a hypertensive blood pressure threshold above which blood pressure–lowering therapies should be initiated, but did determine that systolic pressures greater than 140 mmHg in cardiac surgery were associated with 30-day mortality. Guidelines are further hampered by the lack of standardized definitions for intraoperative hypotension.

A 2009 meta-analysis found that perioperative hemodynamic optimization with fluids and inotropes reduced postoperative renal injury.⁴¹  This study was limited by significant heterogeneity in terms of (1) goals of optimization, (2) modality of optimization, (3) definition of AKI, (4) surgical type, and (5) risk stratification. Notably, the fluid-only subgroup (no inotropes) showed no reduction in kidney injury; however, it was not possible to ascertain whether fluid and inotropes work synergistically, or whether inotropes conferred the overall protective effect.

In the Intraoperative Norepinephrine to Control Arterial Pressure (INPRESS) trial, Futier et al. found that individualized treatment (i.e., maintaining systolic blood pressure within 10% of baseline value using a norepinephrine infusion) resulted in lower rates of postoperative organ dysfunction (a composite metric that included renal failure) after noncardiac surgery when compared to standard management (i.e., maintaining systolic blood pressure greater than 80 mmHg or within 40% of baseline value using ephedrine boluses).⁴²  These findings have not been reproduced and contrast with another randomized trial showing no difference in a composite outcome consisting of major adverse cardiovascular events (MACE), AKI, and all-cause mortality between populations with target MAP greater than 75 mmHg compared to 60 mmHg.⁴³ 

The PeriOperative ISchemic Evaluation-3 (POISE-3) trial was an international randomized controlled trial to compare a hypotension-avoidance versus a hypertension-avoidance strategy on major vascular complications after noncardiac surgery.^44,45 Hypotension-avoidance (target MAP, 80 mmHg or greater; antihypertensives for systolic blood pressure greater than 130 mmHg) and hypertension-avoidance (target MAP, 60 mmHg or greater; antihypertensives continued) strategies resulted in a similar incidence of major vascular complications.⁴⁴  As the primary outcome in POISE-3 did not include AKI, further research (including subanalysis of POISE-3 data)⁴⁶  is necessary to identify hemodynamic targets and modifications associated with renal protective effect.⁴⁵ 

The ongoing, tight perioperative blood pressure management to reduce serious cardiovascular, renal, and cognitive complications (GUARDIAN trial) compares tight pressure management compared to routine pressure management on a composite of major perfusion-related complications (myocardial injury, stroke, nonfatal cardiac arrest, stage 2 to 3 AKI, sepsis, and death) in patients undergoing major noncardiac surgery.⁴⁷  Preliminary results demonstrate that tight blood pressure management with norepinephrine infusion decreases intraoperative hypotension, but delaying antihypertensive medications has little effect on postoperative blood pressure.⁴⁷ 

A retrospective multicenter study found that the impact of intraoperative hypotension on postoperative AKI varies based upon the patient’s risk quartile.³²  Specifically, they assessed the relationship between duration of intraoperative hypotension (at absolute, MAP less than 50 mmHg, 50 to 54 mmHg, 55 to 59 mmHg, and 60 to 64 mmHg; and relative hypotension thresholds, MAP more than 40% below preinduction baseline, 30 to 40%, and 20 to 30%) and KDIGO stage 1, 2, or 3 AKI. They found 12,431 (9.0%) cases of postoperative AKI among 138,021 major noncardiac surgical procedures reviewed.³²  Major risk factors for postoperative AKI were anemia, preoperative estimated glomerular filtration rate, surgery type, American Society of Anesthesiologists Physical Status, and anesthesia duration (table 2). As risk increased, less severe intraoperative hypotension was associated with postoperative AKI. Among the lowest risk quartile, no hypotension range was associated with postoperative AKI. In patients with medium risk (quartile 2), only severe intraoperative hypotension (MAP less than 50 mmHg) was associated with increased postoperative AKI risk. Whereas, in patients with high (quartile 3) or highest (quartile 4) risk, even mild intraoperative hypotension (55 to 59 mmHg) was associated with postoperative AKI.

Another large, multicenter retrospective study by Chiu et al. found an increased risk of AKI (despite decreased duration of hypotension) when liberal use of vasopressors was employed at the expense of fluid resuscitation.⁴⁸ 

Although preoperative anemia has consistently been identified as a risk factor for postoperative AKI,^32,49–51  studies have not demonstrated that correcting anemia (through either transfusion or transfusion-sparing strategies) decreases the risk for AKI.⁵¹  Specifically, a randomized controlled trial of patients undergoing elective cardiac surgery showed that neither intravenous ferric carboxymaltose nor oral iron (administered for 3 to 8 weeks before surgery) improved (1) baseline hemoglobin, (2) intraoperative transfusion requirement, or (3) risk for postoperative AKI.⁵²  Furthermore, a large meta-analysis of cardiac surgery patients found that patients with preoperative anemia were greater than three times more likely to develop postoperative AKI than those in the nonanemic group, but failed to show an association between perioperative transfusion and mortality.⁵¹ 

In addition to not improving outcomes, transfusion may actually be associated with an increased risk for postoperative AKI. A retrospective study of more than 33,000 cardiac surgery patients demonstrated transfusion of packed red blood cells was associated with both mortality and postoperative renal failure.⁵³  The association between transfusion and postoperative morbidity may be due to a sicker patient population in the transfusion group that persists despite risk adjustment and propensity score matching.^53–55  Alternatively, components of blood products (i.e., myeloid-related protein 14) have been linked to direct renal injury in animal models.⁵⁶ 

Within the noncardiac surgery⁵⁷  and critically ill⁵⁸  population, evidence supports a restrictive approach to blood transfusion. The optimal transfusion strategy for the cardiac surgery population remains unresolved, with the Transfusion Requirements After Cardiac Surgery trial showing a restrictive transfusion strategy (hematocrit less than 24%) to be noninferior to a liberal transfusion strategy (hematocrit greater than 30%),⁵⁹  while the Transfusion Indication Threshold Reduction trial showed a higher 90-day mortality in the restrictive transfusion group (hemoglobin less than 75 g/l) compared to the liberal threshold group (hemoglobin less than 90 g/l).⁵⁹  More recently, the Transfusion Requirements in Cardiac Surgery III trial demonstrated a restrictive transfusion strategy was noninferior to a liberal transfusion strategy with respect to the composite outcome of death from any cause, myocardial infarction, stroke, or new-onset renal failure with dialysis, with less blood transfused.⁶⁰  The 2023 American Association of Blood Banks Guidelines recommend (1) a restrictive transfusion strategy (defined as less than 7 g/dl) for hospitalized adult patients who are hemodynamically stable, (2) a transfusion threshold of 7.5 g/dl in cardiac surgery, and (3) 8 g/dl in orthopedic surgery or for patients with pre-existing cardiovascular disease.⁶¹ 

Saline (0.9% sodium chloride; “normal saline”) has historically been the most commonly used intravenous crystalloid for intraoperative resuscitation and maintenance fluid therapy⁶² ; however, physiologically balanced crystalloids, like lactated Ringer’s solution, more closely approximate the physiologic levels of sodium, potassium, and chloride found in plasma.⁶³  Because normal saline has a superphysiologic concentration of chloride (154 mM, compared to 94 to 111 mM in plasma),⁶⁴  hyperchloremic metabolic acidosis is a known consequence of normal saline administration, but whether this is of any clinical consequence is unresolved.^65,66 

A recent meta-analysis of randomized controlled studies on the use of balanced versus normal saline in the perioperative period demonstrated (1) uncertain effect on postoperative mortality (low certainty), (2) uncertain effect on need for renal replacement therapy (low certainty), and (3) improved postoperative acid–base status (moderate certainty).⁶⁶  Further studies are necessary in specific surgical populations. For example, the Saline versus Lactated Ringer’s Solution: The Saline or Lactated Ringer’s (SOLAR) trial of elective colorectal and orthopedic surgery patients found no difference in composite postoperative complications or AKI between patients receiving perioperative saline or lactated Ringer’s solution.⁶⁷  However, balanced crystalloid improved postoperative electrolyte (lower chloride and higher bicarbonate) and acid–base (higher pH) status; the clinical implications of these changes remains unresolved.⁶⁶ 

This builds upon evidence from the critically ill population, which showed a high probability that using balanced crystalloids reduces AKI compared to saline,⁶⁸  although evidence from key randomized trials has been conflicting.^64,69  Specifically, the Balanced Solutions in Intensive Care Study (BaSICS) did not show a difference in 90-day mortality among critically ill patients bolused with saline compared to balanced solutions.⁶⁹  This contrasts with the earlier Isotonic Solutions and Major Adverse Renal Events Trial (SMART), which showed balanced crystalloid use resulted in lower rate of death, new renal replacement, or persistent renal dysfunction.⁶⁴  Among noncritically ill emergency department patients in the Saline against Lactated Ringer’s or Plasma-Lyte in the Emergency Department (SALT-ED) trial, there was no difference in hospital-free days (primary outcome), but there was lower incidence of major adverse kidney events within 30 days (composite of death, new renal replacement therapy, and persistent renal dysfunction).⁷⁰ 

Whether colloids have a role in surgery has been debated and studied.⁷¹  Colloids, specifically hydroxyethyl starch solutions, initially held promise as a way to minimize overall volume of fluids administered, but have been shown to be nephrotoxic in the intensive care setting.^72–74  While the nephrotoxic effect was not replicated in surgical populations,^71,75,76  strong evidence of coagulopathy, bleeding, and transfusion⁷⁷  prompted the U.S. Food and Drug Administration (Silver Spring, Maryland) to issue a black box warning in 2013 and the European Medicines Agency (Amsterdam, The Netherlands) to suspend marketing authorization in 2018.⁷⁸  Widespread use of the synthetic colloid hydroxyethyl starch waned behind mounting evidence of potential clinical harm⁷⁹ ; however, whether this risk persists with human-derived albumin (the primary colloid replacement for hydroxyethyl starch) remained unresolved.⁸⁰ 

Albumin is routinely utilized in cardiac surgery to prime the cardiopulmonary bypass circuit and for volume replacement, although evidence supporting such practice is limited or conflicting.⁸¹  The randomized, double-blind, single-center Albumin in Cardiac Surgery (ALBICS) trial showed no difference in major adverse events (a composite including AKI) between cardiac surgery patients with albumin or lactated Ringer’s solution use for cardiopulmonary bypass priming fluid and perioperative volume replacement,⁸²  suggesting that routine use of albumin did not provide any benefit for patients undergoing cardiac surgery with cardiopulmonary bypass.⁸⁰  A recent meta-analysis confirmed this finding, leading to the conclusion that albumin should be used restrictively in cardiac surgery.⁸¹ 

Data on the perioperative use of albumin is almost nonexistent outside of cardiac surgery. The Choosing Wisely in Anesthesiology Initiative stated that routine administration of colloid for volume resuscitation “without appropriate indication” is one of the top five “low-value” activities in the field of anesthesiology⁸³ ; however, it did not give guidance on when albumin might be indicated. Ongoing research seeks to better define populations that may derive benefit. Albumin may have a limited role in specific surgical populations⁸⁰  including hepatorenal syndrome,^84–86  spontaneous bacterial peritonitis,^87,88  or after paracentesis in liver failure patients.^89–91 

The major goals of intraoperative fluid resuscitations are optimizing intravascular volume to increase cardiac output and perfusion pressure in order to improve renal blood flow and glomerular function.⁹²  Excessive fluid has also been shown to be associated with AKI⁹³ ; however, the cause-and-effect relationship between the two remains unclear.^84,94  A further complexity is that positive fluid balance can be caused by a combination of increased administration, which is potentially modifiable, and decreased urine output, which could only be addressed through renal replacement therapy.⁹⁴  Finally, the definitions of “liberal,” “restrictive,” and “traditional” fluid administration have evolved and can overlap across trials, requiring careful review of specific interventions when interpreting study results.

The Restrictive versus Liberal Fluid Therapy in Major Abdominal Surgery (RELIEF) trial demonstrated that restrictive fluid resuscitation (goal, net zero fluid balance) protocol had a significantly higher rate of AKI when compared to the group receiving liberal fluid resuscitation (goal, reflective of traditional practice).⁹⁵  This contrasts with the results from a randomized trial showing that restrictive fluid resuscitation (in combination with pre-emptive norepinephrine infusion) reduced postoperative complications (which included both renal failure and transient creatinine increase) and hospitalization after radical cystectomy and urinary diversion.⁹⁶  In light of this clinical evidence, providers need to recognize that both hypovolemia and hypervolemia may be associated with AKI and adopt personalized management guided by physiologic endpoints.⁹⁴  Assessment of euvolemia can be guided by judicious fluid challenges with frequent assessment of fluid responsiveness, hemodynamic homeostasis, and clinical signs of hypoperfusion. Intraoperative positional changes can function like a passive leg raise maneuvers, potentially revealing fluid responsiveness. Furthermore, assessment should prioritize dynamic markers (stroke volume, pulse pressure variation) rather static metrics (central venous pressure, pulmonary artery occlusion pressure)⁹² —which have been shown to be an unreliable predictor of fluid responsiveness.⁹⁷ 

A variety of factors in the perioperative environment, including (1) positive pressure ventilation, (2) intravenous fluids, (3) transfusion, and (4) cardiopulmonary bypass, impact venous congestion and central venous pressure.¹⁵  While optimal volume status remains a complex topic, venous congestion has been associated with cardiac surgery–associated AKI in patients from the Statin AKI Cardiac Surgery trial¹⁵  and the Aortic To RAdial Pressure (ATRAP) Gradient Study.⁹⁸  Furthermore, venous congestion was found to more accurately predict AKI than arterial hypotension in the latter population,⁹⁸  although prospective studies are necessary to assess whether treatment strategies to alleviate venous congestion alter the development of postoperative AKI.

Dysglycemia (hypoglycemia, hyperglycemia, and glycemic variability) is hypothesized to be associated with postoperative AKI through a variety of mechanisms including oxidative stress, increased inflammation, and endothelial dysfunction.⁹⁹  Diabetes is a well-established risk factor for AKI,³⁰  but more recent research has found a link between poor glycemic control in the perioperative period and postoperative AKI,^100,101  raising a question whether careful management of blood sugar during the perioperative period has a renal protective effect. Studies assessing the renoprotective effect of intensive glycemic control have found conflicting results, with some studies demonstrating a renoprotective effect^102–104  and others showing no advantage (or even a higher risk of death and stroke) to intensive (target, 81 to 108 mg/dl) compared to conventional (target, less than 180 mg/l) glycemic control.^105,106  Based upon these studies, the POQI-3 recommends treating hyperglycemia to a target blood glucose less than 180 mg/dl to reduce the risk of postoperative AKI.² 

The α₂-adrenoreceptor agonist dexmedetomidine has been associated with lower risk of AKI after cardiac surgery¹⁰⁷  and delayed graft function after renal transplantation.¹⁰⁷  The mechanism of this potential renoprotective effect has not been fully resolved but is hypothesized to result from a combination of intrinsic anti-inflammatory properties, sympatholytic activity leading to reduction of renal vasoconstriction, and smooth muscle relaxation.¹⁰⁸  Larger multicenter trials are needed to confirm this renoprotective effect and to resolve the ideal dexmedetomidine dose and timing.¹⁰⁸ 

Angiotensin-converting enzyme inhibitors and angiotensin receptor blockers have been hypothesized to increase the risk for AKI through a variety of mechanisms, including systemic hypotension, renal artery constriction, and interstitial nephritis.¹⁰⁹  Recommendations and practice for the management of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers in the perioperative period are conflicting. The American College of Cardiology (Bethesda, Maryland) and American Heart Association (Dallas, Texas) recommending continuing during noncardiac surgery,¹¹⁰  while anesthesia groups routinely hold on the day of surgery due to concerns of intraoperative hypotension.¹¹¹  Subanalysis of the Vascular events In noncardiac Surgery patIents cOhort evaluatioN (VISION) cohort found that withholding angiotensin-converting enzyme inhibitors or angiotensin receptor blockers before major noncardiac surgery was associated with a lower risk of death and postoperative vascular events, although this study was not designed to assess renal outcomes.¹¹²  This contrasts with an earlier meta-analysis that did not demonstrate a difference in mortality, morbidity, and complications between patients who continued or held their angiotensin-converting enzyme inhibitors or angiotensin receptor blockers in the perioperative period.¹¹³  In the largest prospective study to date in major elective noncardiac surgery, conducted by the Student Audit and Research in Surgery (STARSurg) Collaborative, withholding angiotensin-converting enzyme inhibitors or angiotensin receptor blockers did not protect against the development of postoperative AKI.¹⁰⁹  The POQI-3 consensus statement recommends restarting angiotensin-converting enzyme inhibitors or angiotensin receptor blocker therapy within the first 48 h postoperatively in patients, as long as the patient is hemodynamically stable and not exhibiting signs of postoperative AKI.² 

Additionally, angiotensin-converting enzyme inhibitors or angiotensin receptor blockers can lead to a “functional” AKI (defined as an increase in serum creatinine). The prospective Translational Research Investigating Biomarker Endpoints for Acute Kidney Injury (TRIBE-AKI) study demonstrated that the graded increase in functional AKI (continued angiotensin-converting enzyme inhibitors or angiotensin receptor blocker greater than held angiotensin-converting enzyme inhibitors or angiotensin receptor blocker greater than no angiotensin-converting enzyme inhibitors or angiotensin receptor blocker use) was not accompanied by “structural” AKI (no difference in AKI risk between the three exposure groups, as assessed by four different urinary biomarkers for AKI).¹¹⁴  The results of a study can be dramatically altered based upon how AKI is defined (serum creatinine changes vs. biomarkers), and this discrepancy may be especially pronounced in patients on angiotensin-converting enzyme inhibitors or angiotensin receptor blockers. This study demonstrates that a rise in serum creatinine is not always accompanied by tissue damage (and tissue damage does not always result in rising serum creatinine, i.e., subclinical AKI), highlighting one of the major limitations of serum creatinine as a biomarker.

NSAIDs have been used to effectively mitigate postoperative pain while decreasing the opioid requirement.¹¹⁵  NSAIDs can adversely affect renal via prostaglandin mediated afferent arteriolar dilation, which causes reduced glomerular perfusion,¹¹⁶  leading to acute interstitial nephritis. Drug-associated acute interstitial nephritis presents a cell-mediated immune response, characterized by acute loss of kidney function, interstitial infiltrates, edema, and tubulitis on biopsy (with relative sparing of glomerulus and vasculature).¹¹⁷  Because of this risk, the Medicines and Healthcare Products Regulatory Agency (London, United Kingdom) has recommend that NSAIDs be avoided in hypovolemic patients¹¹⁸ ; however, best practices in low-risk patients remain unresolved.²  A 2018 meta-analysis found that although NSAIDs led to a slight increase in serum creatinine, this was of unclear clinical significance, and no reliable conclusions could be drawn regarding the association between NSAIDs and renal replacement therapy, hospital length of stay, or death.¹¹⁶  A secondary analysis of data from the Restrictive versus Liberal Fluid Therapy in Major Abdominal Surgery (RELIEF) trial found that intraoperative NSAID or cyclooxygenase-2 inhibitor use was significantly associated with increased postoperative AKI, supporting the need for a randomized trial to better define the risk–benefit balance of these perioperative analgesic adjuncts.¹¹⁹  Emerging proteomic research sheds light on the molecular pathways, specifically differences in prostaglandin synthesis, driving NSAID-attributed AKI in the perioperative space.¹²⁰ 

Nephrotoxic medication exposure represents one of the most common etiologies of AKI in hospitalized patients.¹²¹  Antimicrobials (antivirals, antibiotics, and antifungals) have been implicated in impaired renal function through mechanisms ranging from acute interstitial nephritis to acute tubular necrosis.¹²²  A variety of strategies to mitigate this risk has been proposed including (1) adjusting the dose based upon creatinine clearance, (2) avoiding concomitant nephrotoxins, and (3) regular assessment of antibiotic necessity and duration.¹²²  Implementation of an electronic health record screening and clinical decision support reduced the rate of AKI by 64% among noncritically ill children.¹²¹ 

Exposure to iodinated contrast media has been associated with acute injury in numerous historical studies,¹²³  leading many clinicians to avoid contrast in patients with reduced renal function.¹²⁴  Studies showing the association between contrast and AKI often lacked a control group of patients not exposed to contrast,¹²³  whereas more recent well-controlled observational studies suggest a much lower association between contrast exposure and AKI.^125–127  The American College of Radiology (Reston, Virginia) clarified the implied causal relationship between contrast media and AKI through the adoption of two new terms: (1) contrast-induced AKI and (2) contrast-associated AKI.¹²⁸  Contrast-associated AKI is defined as any AKI occurring within 48 h after the administration of contrast media, while contrast-induced AKI is the subset of cases that can be causally linked to the contrast exposure.¹²⁴  While the incidence of contrast-induced AKI is markedly lower than that of contrast-associated AKI,¹²⁴  the actual risk remains unresolved, with some studies showing no elevated risk for contrast-induced AKI regardless of baseline renal function¹²⁶  and others showing risk only at severely reduced renal function.¹²⁷  The primary risk factor for contrast-associated AKI remains baseline estimated glomerular filtration rate, with diabetes mellitus, nephrotoxic exposure, hypotension, and hypovolemia conferring additional risk.¹²⁴  From a pragmatic perspective, given these uncertainties, renal risk alone should not contraindicate truly necessary imaging.

In addition to the previously described patient-specific risk factors, procedure-specific factors impact the development of contrast-associated AKI. For example, percutaneous coronary interventions (particularly in ST-Elevation Myocardial Infarction [STEMI] patients) are particularly susceptible to AKI due to contrast exposure, reduced perfusion secondary to transient hypotension or reduced cardiac output, and high baseline patient risk due to comorbidities.¹²⁹  Intraarterial contrast places the patient at higher risk than intravenous contrast administration, which has improved with decreasing osmolar concentration of intravenous contrast.¹³⁰  The volume and osmolarity of the of contrast media represent two modifiable risk factors.^129,131 

The KDIGO guidelines make a number of recommendations for minimizing AKI, including (1) avoidance of nephrotoxic agents, (2) discontinuation of angiotensin-converting enzyme-inhibitors and angiotensin receptor blockers for 48 h after surgery, (3) close monitoring of serum creatinine and urine output, (4) avoidance of hyperglycemia, (5) avoidance of radiocontrast, (6) hemodynamic monitoring with a pulse index continuous cardiac output catheter, and (7) algorithm-based optimization of volume and hemodynamic status.^8,132  Meersch et al. found that a therapeutic bundle derived from these KDIGO recommendations significantly reduced the rate of cardiac surgery–associated AKI (55.1% vs. 71.7%; absolute risk reduction, 16.6; 95% CI, 5 to 27.9%; P = 0.004) within 72 h compared to standard care (MAP greater than 65 mmHg and central venous pressure between 8 and 10 mmHg).¹³²  This contrasts with a more recent study from Shen et al. demonstrating that an enhanced recovery after surgery protocol consisting of balanced fluid management, flexible use of vasoactive drugs, and multimodal analgesia containing NSAIDs was safe but did not impact incidence of postoperative AKI.¹³³  The bundled interventions limit speculation on the contribution of any single measure.

Because advances in the treatment of postoperative AKI have been limited,^6,7  management in the perioperative period has traditionally focused on prevention and diagnosis. Once postoperative AKI develops, supportive management (i.e., preventing ongoing renal insult through removal of nephrotoxic agents and optimized renal perfusion through fluids and vasopressors) is the mainstay of therapy.¹³⁴ The 2012 KDIGO guidelines for the management of AKI recommend a stage-based treatment approach.^7,135  For stage 1 AKI or patients at high risk for AKI, primary management goals include identifying the likely cause of AKI and avoiding secondary insults where possible (such as administration of NSAIDs or other nephrotoxic medications when alternatives are available).⁷  In addition, hemodynamic optimization to ensure adequate kidney perfusion is an important goal, although the guidelines do not provide specific hemodynamic targets. Two single-center randomized controlled trials (one in cardiac surgery and one in noncardiac surgery patients) demonstrated significant reductions in postoperative AKI with a bundled care approach including target MAP 65 mmHg or greater.^136,137  Notably, a meta-analysis suggested that protocolized approaches to hemodynamic optimization are associated with reduced risk of AKI overall despite differences in specific targets.⁴¹ 

In patients who progress to stage 2 AKI, adjusting drug dosages to account for glomerular filtration and careful attention to volume input or output become important. Stage 3 AKI is associated with acid–base imbalance, altered electrolyte levels, and accumulation of uremic toxins.⁷  Decisions on the need for intensive care and initiation of kidney replacement therapy require expertise in nephrology or critical care medicine and should be considered early in disease progression.¹³⁵ 

Phase of care, monitoring, and follow-up can vary greatly depending on the severity of disease. The KDIGO clinical practice guidelines recommend that all patients diagnosed with AKI are evaluated at 3 months to screen for new or progressive chronic kidney disease,^135,138  while the Acute Disease Quality Initiative and PeriOperative Quality Initiative recommend all postoperative patients with AKI or acute kidney disease have a kidney health assessment within 30 days of hospital discharge.²  As morbidity and mortality increase with severity of AKI (and in patients without renal recovery), these factors should inform the frequency and nature of follow-up.²  Management of AKI during the period of inpatient hospitalization provides an opportunity to prevent or delay progression of AKI to chronic kidney disease.¹³⁸  Patients requiring acute renal replacement therapy (or necessitate continuous renal replacement therapy due to hemodynamic instability) should be managed in the ICU.¹³⁹  Furthermore, a variety of critical conditions associated with AKI including shock, sepsis, acid–base imbalance, and respiratory failure secondary to fluid overload require intensive care. The KDIGO management guidelines recommend considering ICU admission beginning at stage 2 AKI.¹³⁵ 

Diuretics may be used to manage volume overload secondary to AKI and increase potassium excretion.⁸  Administration of loop diuretics has been shown to reverse oligo-anuria, revealing patients with less severe renal injury or adequate renal reserve,¹⁴⁰  but has not been shown to prevent AKI, need for renal replacement therapy, or overall mortality.^141,142  Judicious fluid administration combined with loop diuretics (i.e., “matched hydration”) may increase tubular flow of filtrate without volume depletion and associated renal hypoperfusion^143–145 ; however, further randomized control trials are necessary before this strategy can be broadly recommended in the prevention of postoperative AKI.¹⁴¹ 

Despite initially promising studies,^146,147  neither preoperative aspirin nor clonidine reduced the risk of postoperative AKI in the POISE-2 randomized trial of noncardiac surgery patients.²³  Dexamethasone potentially attenuates the systemic inflammatory response to cardiopulmonary bypass and reducing inflammatory mediators, but the Dexamethasone in Cardiac Surgery (DECS) trial and the Steroids in Cardiac Surgery (SIRS) trial failed to support prophylactic use of steroids during cardiac surgery.^148,149  Finally, a meta-analysis found no evidence that N-acetylcysteine improved renal outcomes after major surgery¹⁵⁰ ; therefore, the KDIGO management guidelines recommend against using oral or intravenous N-acetylcysteine for prevention of postsurgical AKI.¹³⁵ 

The Acute Disease Quality Initiative and PeriOperative Quality Initiative consensus report recommends determining and resolving the underlying cause of postoperative AKI.²  Mainstays of management include removal or avoidance of nephrotoxic exposure and hemodynamic optimization. Renal perfusion can be maintained through a combination of fluids and vasopressors, with protocol-based management to prevent worsening AKI in high-risk perioperative patients. Diuretics should not be used in the treatment of AKI, except to manage volume overload.²  A number of agents including dopamine, fenoldopam, atrial natriuretic peptide, insulin-like growth factor-1, and diuretics have failed to improve the course of AKI and do not have a role in the management of AKI.^2,8 

Renal replacement therapy is used only if there is a specific indication, like severe acidosis, hyperkalemia, uremia, or inability to manage volume status with diuretics.^2,8  The Standard versus Accelerated Initiation of Renal-Replacement Therapy in Acute Kidney Injury (STARRT-AKI) trial showed no evidence of benefit and higher risk for dependence at 90 days when renal replacement therapy was initiated early for postoperative AKI compared to conservative management.¹⁵¹  The optimal timing and indications for renal replacement therapy remain unresolved and should be carefully weighed on an individual basis.¹  Absolute indications for renal replacement therapy include (1) symptoms or signs of uremia (i.e., pericarditis, bleeding, or encephalopathy), (2) pulmonary edema resistant to diuretics, (3) refractory hyperkalemia (greater than 6.5 mM, rapidly increasing, or associated with cardiac arrhythmias), (4) metabolic acidosis (pH less than 7.2), or (5) intoxication with dialysable drug or toxin.^152,153  Relative indications are (1) organ dysfunction impacted by either fluid overload or from the AKI; (2) administration of large volumes of fluid (i.e., blood transfusion); (3) solute burden such as from tumor lysis syndrome, rhabdomyolysis, or hemolysis; and (4) oligo-anuria due to severe AKI with low probability of rapid renal recovery.¹⁵² 

The incidence of AKI after cardiac surgery ranges from 15 to 30%.^18,154  Even patients who fully recovered renal function after developing AKI had a significantly higher mortality than those not developing AKI.¹⁵⁵  The increased incidence of AKI in cardiac surgery patients can be attributed to both the unique stress profile of the procedure (e.g., cardiopulmonary bypass impairing autoregulatory mechanisms and inducing and inflammatory response, hemodilution, and atheroembolic events associated with bypass cannulation and the surgical technique) and the risk profile of the patient population. A variety of risk prediction models for AKI after cardiac surgery have been developed including the Cleveland Clinic Score,³³  the Mehta Score,¹⁵⁶  and the Simplified Renal Index.¹⁵⁷  Intraoperative risk factors include cardiopulmonary bypass, blood transfusion, proteinuria, urgent surgery, reoperation, and preoperative intra-aortic balloon pump.^49,158,159  Patients undergoing valvular surgery have a higher incidence of cardiac surgery–associated AKI than those undergoing coronary artery bypass graft surgery.^49,160  Strategies for prevention and mitigation of cardiac surgery–associated AKI have been comprehensively reviewed,^155,161  and include implementation of the “KDIGO bundle of care,”¹³²  goal-directed oxygen delivery while on cardiopulmonary bypass,^162,163  and a restrictive transfusion threshold.¹⁶⁴ 

The Vasopressin versus Norepinephrine in Patients with Vasoplegic Shock after Cardiac Surgery (VANCS) trial found that patients receiving vasopressin as a first-line agent had lower incidence of moderate to severe AKI than those receiving norepinephrine,¹⁶⁵  and dobutamine reduced cardiac surgery–associated AKI when implemented as a part of a “bundle of care” approach.^132,166  Further research is necessary on the optimal and mitigation strategies and hemodynamic targets.

The Society of Thoracic Surgeons (Chicago, Illinois) definition for postoperative acute renal failure essentially only includes stage 3 (or dialysis-requiring) AKI.¹⁶⁷  While other studies apply the KDIGO criteria after cardiac surgery,^168–170  Society of Thoracic Surgeons database reporting notably focuses on this acute renal failure definition.

Diagnostic criteria relying on serum creatinine can overestimate kidney function in cirrhotic patients, and bilirubin can decrease the accuracy of creatinine assays.¹⁷¹  A joint consensus statement from the International Club of Ascites (ICA) and the Acute Disease Quality Initiative recommends defining AKI in patients with cirrhosis using the KDIGO criteria, while using the lowest serum creatinine value up to 12 months before.¹⁷²  In addition to AKI meeting the KDIGO definition, a diagnosis of hepatorenal syndrome, AKI requires (1) cirrhosis with ascites, (2) absence of improvement in serum creatinine and/or urine output within 24 h after adequate volume resuscitation, and (3) absence of an alternative explanation for the primary cause of AKI.^171,172  ICA modifications to KDIGO have shown to accurately predict 30-day mortality, length of hospital stay, and organ failure.¹⁷³  However, because automated algorithms calculate and surgical registries curate AKI outcomes using KDIGO standard criteria, the unmodified KDIGO definition has been employed in studies in the liver transplant population without issue.²⁸ 

Delayed graft function, typically defined as dialysis requirement within 1 week of transplant, is a manifestation of AKI attributable to the transplant process.¹⁷⁴  Delayed graft function occurs in about one of five transplants, with incidence in deceased donor transplants as high as 40%,¹⁷⁵  and is associated with clinical outcomes including acute rejection,¹⁷⁶  poor graft survival,¹⁷⁴  and increased mortality.^177,178  Furthermore, delayed graft function exacts a significant economic burden, increasing transplant cost by more than $18,000, length of hospitalization by 6 additional days, and ICU stay by 2 days.^175,179,180  Despite improved understanding of the pathophysiology leading to delayed graft function, including the molecular mechanism of ischemia–reperfusion injury and allograft immunogenicity,¹⁷⁷  the incidence of delayed graft function has continued to rise (presumably due to increased cold ischemia time resulting from changes in the Kidney Allocation System).¹⁷⁵ 

Traditional risk factors¹⁸¹  for delayed graft function can be divided into (1) donor factors (age, sex, body mass index, etc.), (2) recipient factors (age, race, sex, comorbidities, etc.), (3) graft factors (cold ischemia time, human leukocyte antigens or ABO incompatibility, deceased donor compared to living donor, and donation after cardiac death compared to donation after brain death), and (4) surgical factors (right nephrectomy, laparoscopic approach for donor nephrectomy, etc.).^182–184  Improved risk stratification informs management decisions (for example, invasive monitoring in high-risk patients) and organ allocation; however, characterized risk factors are mostly nonmodifiable. The intraoperative and early ICU periods remain an understudied source of outcome variation and represent a window where early intervention could modify the clinical trajectory.

Despite standardization of AKI criteria, diagnosis remains dependent on changes in serum creatinine and urine output.¹⁸⁵  A major limitation of serum creatinine as an AKI biomarker is that it does not increase until substantial parenchymal injury has occurred. This time “lag” prevents recognition of renal injury during the time window, when early intervention can alter the clinical course.¹⁸⁶  Furthermore, functional biomarkers like serum creatinine are altered by a variety of overlapping processes, independent of renal injury (like fluid overload, malnutrition, and muscle wasting).^186,187  Significant effort has been devoted on identifying and validating novel biomarkers for AKI in the perioperative period that may reflect the specific process of nephron injury and enable earlier detection of AKI.¹⁸⁵ 

Biomarkers can be categorized into multiple overlapping classification schema: (1) blood versus urine, (2) anatomic location (glomerulus, proximal tubule, loop of Henle, and distal tubule), and (3) functional mechanism (injury, inflammation, and repair or fibrosis).¹⁸⁵  Injury biomarkers are released from injured kidney cells (analogous to troponin release from injured myocardial cells during an infarction), potentially providing a more sensitive and specific marker of AKI than blood urea nitrogen and serum creatinine.¹⁸⁸  The injury biomarkers, neutrophil gelatinase-associated lipocalin, kidney injury molecule-1, and liver fatty acid–binding protein, are early markers of renal tubular injury and potentially identify AKI during the “blind window” before rise in serum creatinine or BUN.¹⁸⁸ Inflammatory biomarkers, interleukin-18, tumor necrosis factor receptor-1 (TNFR-1), tumor necrosis factor receptor-2 (TNFR-2), and osteopontin, are activated and recruited as an early response to kidney injury, and hold promise (in combination with other biomarkers) of predicting prognosis and progression of postoperative AKI.^189–191  Additionally, specific inflammatory markers, tumor necrosis factor-α (TNF-α) and interleukin-9, can be used to differentiate acute interstitial nephritis from other causes of AKI.¹⁹²  After injury and inflammation, a tightly regulated reparative process occurs. Biomarkers associated with repair (epidermal growth factor and chitinase 3-like 1) and fibrosis (transforming growth factor-β [TGF-β]), monocyte chemoattractant protein 1, and type III procollagen peptide can be leveraged to predict renal recovery and interstitial fibrosis or tubular atrophy.¹⁹³ 

Injury biomarkers rise at different rates, with each having a characteristic injury response curve (analogous to the sequential rise of myoglobin, troponin I, creatine phosphokinase, and lactate dehydrogenase after myocardial injury).¹⁹⁴  For example, urinary neutrophil gelatinase-associated lipocalin begins to rise within 2 h and peaks within 6 h after surgery.^195–197  This compares to kidney injury molecule-1, which typically rises within 12 h, peaking around 24 h.^197,198  A comparison of the time course for novel AKI biomarkers can be found in table 3.

Combining multiple biomarkers enables improved risk prediction (quantified by discrimination and net reclassification improvement) earlier in the disease course, when interventions are more likely to impact disease trajectory.¹⁹⁵  Nephrocheck (bioMerieux, France), a test that combines urinary tissue inhibitor of metalloproteinases 2 and insulin like growth factor binding protein 7, was approved by the Food and Drug Administration as the first biomarker for AKI in 2014. The sensitivity and specificity of all biomarkers depend on (1) the temporal relationship between renal injury and when the sample is drawn, (2) the diagnostic threshold selected, and (3) the AKI population (critically ill, cardiac surgery, etc.). Plasma biomarkers demonstrate better discrimination for early postoperative AKI (compared to urine biomarkers) even when normalized to urinary creatinine and injury biomarkers had better predictive ability than either stress or inflammatory biomarkers.^199,200  Across all populations and AKI etiologies, serum and urine neutrophil gelatinase-associated lipocalin had the best diagnostic accuracy (table 3); however, performance was more limited in surgical populations (with neutrophil gelatinase-associated lipocalin/creatine performing best in this population).²⁰⁰ 

Continued development, validation, and integration of biomarker panels will enable earlier, more accurate diagnosis of AKI.¹⁹²  More recently, an independent dose–response association was seen between the level of urinary biomarkers (interleukin-18, neutrophil gelatinase-associated lipocalin, kidney injury molecule-1, liver fatty acid–binding protein, and albumin) and duration of AKI in the TRIBE-AKI cardiac surgery cohort.²¹⁵  Urinary epidermal growth factor and monocyte chemoattractant protein 1 predict which AKI patients are likely to develop chronic kidney disease after cardiac surgery.²¹⁶  In addition to predicting AKI progression and AKI–to–chronic kidney disease transition, novel biomarkers can impact the management of AKI¹⁸⁵ —for example, (1) guiding diuresis in patients with acute decompensated heart failure²¹⁷  and (2) triggering implementation of the KDIGO care bundle after major abdominal surgery.¹³⁷  Ongoing research will define meaningful subgroups using a combination of genomic data, transcriptomics, and biomarkers.²¹⁸  Furthermore, advanced analytical techniques will integrate molecular signatures from multiple biomarkers, establish unbiased predictive cutoff values, and provide insight into the pathology and progression of renal disease.¹⁸⁵ 

While the genetic contribution of chronic kidney disease has been well characterized,²¹⁹  the genetics of AKI remains poorly understood. Most studies on the genetic underpinnings of AKI have employed a candidate gene approach and failed to identify contributory variants consistently.²²⁰  A genome-wide association study of AKI in a general hospitalized population (mixture of different AKI risk factors including surgery, sepsis, and nephrotoxic medications) failed to identify any variants at the level of genome-wide significance (P < 5 × 10^-8), but did report two novel variants adjacent to the dispatched RND transporter family member 1 (DISP1) and Toll-like receptor 5 (TLR5) genes at a lower suggestive threshold.²²¹  Genome-wide association studies focused on postoperative AKI²²²  and cardiac surgery–associated AKI^223–226  have found very few significant associations, and these associations were not consistently replicated across studies. Additionally, a polygenic score that summarizes the genetic burden across the whole genome into a single number that can be used in disease risk modeling was not associated with cardiac surgery–associated AKI.²²⁵  These may be because cardiac surgery–associated AKI is only minimally influenced by baseline genetics, and clinical risk factors play a more influential role in the development and pathogenesis.²²⁵  The expansion of biobanks will enable greater statistical power on a larger, more diverse population.²²⁵  Future studies may also identify subpopulations where genetics may exert disproportionate influence (for example, stratified by patient risk or stage of AKI).

Although preservation of renal perfusion is fundamental to AKI prevention and management,⁷  the optimal strategy to ensure this remains unresolved. Ongoing research will provide insight into optimal (1) choice of vasopressors or inotropes, (2) fluid management, and (3) hemodynamic or oxygenation goals for targeted management,^7,155  and will likely require a personalized or precision-medicine approach. In addition to understanding interventions to optimize renal perfusion, more research on the temporal relationship between each of these exposures and renal injury is necessary. For example, intraoperative hypotension has traditionally been evaluated over the entire course of the procedure.³⁹  Future studies that employ a more nuanced definition for hypotension could overcome such limitations while providing insight into the specific mechanism and timing between perfusion and AKI. Furthermore, the risk–benefit balance between nephrotoxic exposures, like NSAIDs and intravascular contrast, remains an area of active research. The incidence of contrast-associated AKI and effective interventions in high-risk patients remain areas of controversy and active research.¹⁵⁵  Advances in analytical and predictive algorithms, including machine-learning and clinical decision support, will enable more accurate prediction and stratification of patients at risk for AKI.

Many studies have focused on the serum creatinine–based diagnostic criteria, while less attention has been paid to prognostic implications and management of oliguria.^2,227  Despite key limitations, serum creatinine remains the standard biomarker for AKI in perioperative practice.¹⁰  Advances in biomarkers may enable improved sensitivity, earlier detection, prognostic insight, and more precise classification. The management of AKI is largely supportive (prevent ongoing renal insult and optimize renal perfusion),¹³⁴  with the optimal timing and indications for renal replacement therapy unresolved.¹  Proposed treatments for AKI, including dopamine, fenoldopam, aspirin,²³  steroids,¹⁴⁹  and ischemic preconditioning,²²⁸  have not been found to alter the course of disease; however, prevention and treatment modalities remain key areas for future research.

Perioperative AKI has been consistently associated with long-term morbidity,¹  mortality,²  and healthcare costs,³  even among patients with complete renal recovery.⁴  Treatment options for postoperative AKI are limited,^6,7  so perioperative management has traditionally focused on prevention and diagnosis. The KDIGO criteria, currently used to define AKI,⁸  rely on changes in serum creatinine and absolute urine output to provide an easy-to-interpret, objective, and validated approach useful for epidemiologic studies; however, this definition has notable limitations, including a susceptibility to be influenced by nonrenal factors and a time lag between nephron injury and rise in serum creatinine.⁸  AKI represents a clinical syndrome, rather than a single disease, with commonly implicated mechanisms including (1) oxido-inflammatory stress, (2) renal hypoperfusion, (3) endogenous or exogenous nephrotoxins, and (4) iatrogenic causes. Risk factors for postoperative AKI are divided into (1) patient-related, (2) surgery-related, (3) anesthetic-related, and (4) ICU-related risk factors. Recent studies have attempted to integrate anesthetic factors into risk assessment.³²  Notably, characterizing hypotension is complicated by nonstandardized definitions and baseline risk. A variety of strategies for prevention and mitigation of postoperative AKI have been proposed and include preservation of renal perfusion, avoidance of nephrotoxic exposure, implementation of bundles of care, goal-directed oxygen delivery, and a restrictive transfusion threshold. Ongoing research will provide insight into optimal choice of vasopressors or inotropes, fluid management, and hemodynamic or oxygenation goals for targeted management, and will likely require a personalized or precision-medicine approach. Continued development, validation, and integration of biomarker panels will enable earlier, more accurate diagnosis of AKI.

All work and partial funding are attributed to the Department of Anesthesiology, University of Michigan Medical School (Ann Arbor, Michigan) and the Department of Internal Medicine, Division of Nephrology, University of Michigan Medical School. Dr. Douville received support from the National Institute of Diabetes and Digestive and Kidney Diseases (Bethesda, Maryland; 1 K08 DK131346) and a Foundation for Anesthesia Education and Research (Schaumburg, Illinois) Mentored Research Training Grant. Dr. Mathis received grant funding support from the National Heart, Lung, and Blood Institute (Bethesda, Maryland; K01-HL141701) and the National Institute of Diabetes and Digestive and Kidney Diseases (R01-DK133226).

Dr. Mathis received funding paid to the University of Michigan (Ann Arbor, Michigan) from Chiesi USA (Cary, North Carolina) for work unrelated to the scope of the research presented in this manuscript. Dr. Heung has received payments from legal firms for expert testimony, CardioSounds LLC Ann Arbor, Michigan) as a paid consultant, grant funding from Astute Medical Inc. (Canton, Ohio), grant funding to the University of Michigan from Spectral Diagnostics (Toronto, Ontario), and Wolters-Kluwer Inc. for consulting fees as a senior editor for Lexicomp drug dosing guidelines. Dr. Schaub has received funding through the RPC2 consortium (Astra Zeneca [Cambridge, United Kingdom], Genetech [San Francisco, California], NovoNordisk [Bagsvaerd, Denmark], Eli Lilly [Indianapolis, Indiana]). Dr. Kretzler has received funding from the Chan Zuckerberg Initiative, Janssen (Beerse, Belgium; as an employee of the University of Michigan), Eli Lilly, Gilead (Foster City, California), Boehringer-Ingelheim (Ridgefield, Connecticut), Moderna (Cambridge, Massachusetts), the European Union Innovative Medicine Initiative, Certa (Saratoga, California), Chinook Tulsa, Oklahoma), Angion (San Francisco, California), RenalytixAI (New York, New York), Travere (San Diego, California), Regeneron, IONIS (Carlsbad, California), Maze Therapeutics (San Francisco, California), Roche-Genentech, Sanofi (Paris, France), and Dimerix (Victoria, Australia; as an employee of the University of Michigan). Additionally, Dr. Kretzler holds a patent, PCT/EP2014/073413: Biomarkers and Methods for Progression Prediction for Chronic Kidney Disease. The other authors declare no competing interests.