Cystatin C is a 13 kDa protein that is filtered by the glomerulus and completely reabsorbed and degraded by the proximal tubule. Cystatin C has been validated as an alternative marker of glomerular filtration. ^, As the result of its shorter serum half-life, serum cystatin C concentrations change more rapidly than serum creatinine in response to changes in kidney function, allowing changes in serum cystatin C to be detected sooner than changes in serum creatinine following the onset of AKI. Under normal circumstances, urinary cystatin C is virtually undetectable; however, after tubular injury, tubular reabsorption of filtered cystatin is diminished and urinary cystatin C can be detected, raising the possibility of its use as an early marker of tubular injury.

Neutrophil gelatinase–associated lipocalin (NGAL) is a 25 kDa protein whose expression by renal tubular epithelial cells is markedly upregulated after ischemic or nephrotoxic kidney injury. ^, NGAL is believed to enhance the trafficking of iron-siderophore complexes, enhancing the delivery of iron, upregulating hemeoxygenase-1, reducing apoptosis, and increasing the normal proliferation of renal tubule epithelial cells. Urine and plasma NGAL have been evaluated in numerous clinical settings as an early biomarker of tubular injury. Initial studies in children undergoing cardiac surgery demonstrated extremely high sensitivity and specificity for identification of AKI, with an area under the receiver operating characteristic (ROC) curve of >0.99; however, these early results have not been reproduced across other clinical settings.

Kidney injury molecule 1 (KIM-1) is a transmembrane protein whose expression is markedly upregulated in the proximal tubule after tubular injury. The extracellular component of the KIM-1 protein is shed into the urine post tubular injury, permitting its potential use as a marker of tubular damage; however, the time course of peak KIM-1 expression in the urine is later than seen with NGAL. ^, Moreover, studies have demonstrated elevated levels of KIM-1 in non-AKI settings, including CKD and renal cell carcinoma, which reduce its specificity for AKI. ^,

Interleukin 18 (IL-18) is a proinflammatory cytokine whose expression is increased in the kidney after ischemic and nephrotoxic injury. Urinary IL-18 levels have been shown to rise within 6 hours after tubular injury and cardiac surgery, as well as in critically ill patients. ^{, , ,}

Despite its name, liver fatty acid–binding protein (L-FABP) is expressed in the proximal tubule. ^, Elevated L-FABP levels may be detected in the urine within 6 hours of ischemic or nephrotoxic injury, permitting its potential use as a marker of tubular injury. ^, In a meta-analysis of published studies, the sensitivity and specificity of urinary L-FABP for diagnosis of AKI were each approximately 75%.

Tissue inhibitor of metalloproteinase 2 (TIMP-2) and insulin-like growth factor–binding protein 7 (IGFBP7) are expressed in epithelial cells and act in an autocrine and paracrine manner to arrest cell cycle in AKI. In three discovery cohorts comprising 522 patients, these two biomarkers were identified as having the highest discriminant ability among 340 candidate biomarkers of AKI. In a subsequent validation study of 728 patients, this pair of biomarkers had an area under the ROC curve of 0.80, which was significantly better than the performance of other candidate biomarkers including NGAL, KIM-1, IL-18, and L-FABP. A combination test that includes TIMP-2 and IGFBP7 is available for commercial use.

Hyperkalemia is a common and potentially life-threatening complication of AKI. ^, Serum K ⁺ typically rises by 0.5 mmol/L/day in oligoanuric patients and reflects impaired excretion of K ⁺ derived from a patient’s diet, the administration of K ⁺ -containing solutions and drugs administered as potassium salts, as well as the release of K ⁺ from the injured tubular epithelium. Hyperkalemia may be compounded by coexistent metabolic acidosis and/or hyperglycemia or other hyperosmolar states that promote K ⁺ efflux from cells. Hyperkalemia present at the time of diagnosis of AKI or the rapid development of severe hyperkalemia suggests massive tissue destruction, as might be seen with rhabdomyolysis, hemolysis, or tumor lysis. ^{, ,} Hyperuricemia and hyperphosphatemia may accompany hyperkalemia in these settings. Mild hyperkalemia (<6.0 mmol/L) is usually asymptomatic. Higher levels are frequently associated with electrocardiographic abnormalities including peaked T-waves, prolongation of the PR-interval, flattening of P-waves, widening of the QRS complex, and intraventricular conduction defects. These electrocardiographic findings may precede the onset of life-threatening cardiac arrhythmias such as bradycardia, heart block, ventricular tachycardia, ventricular fibrillation, and asystole. In addition, hyperkalemia may induce neuromuscular abnormalities such as paresthesias, hyporeflexia, weakness, ascending flaccid paralysis, and respiratory failure.

Hypokalemia is unusual in AKI but may complicate nonoliguric ATI caused by aminoglycosides, cisplatin, or amphotericin B, presumably because of impaired K ⁺ reabsorption resulting from epithelial cell injury in the thick ascending limb of the loop of Henle. ^, For more discussion on potassium homeostasis, see Chapter 16 .

Normal metabolism of dietary protein yields between 50 and 100 mmol/day of fixed nonvolatile acids (principally sulfuric and phosphoric acid) that are excreted by the kidneys to maintain acid-base homeostasis. Predictably, AKI is commonly complicated by metabolic acidosis, typically with a widening of the serum anion gap due to retention of phosphates, sulfates, and organic anions. Acidosis may be severe (daily fall in plasma HCO ₃ ⁻ >2 mmol/L) when the generation of H ⁺ is increased by additional mechanisms (e.g., diabetic or fasting ketoacidosis; lactic acidosis complicating generalized tissue hypoperfusion, liver disease, or sepsis; metabolism of ethylene glycol). ^{, ,} In contrast, metabolic alkalosis is an infrequent finding but may complicate overly aggressive correction of acidosis with HCO ₃ ⁻ , overzealous use of combination loop and thiazide diuretics, or loss of gastric acid by vomiting or nasogastric aspiration. For further discussion on acid-base homeostasis, see Chapter 15.

Mild to moderate hyperphosphatemia (5–10 mg/dL) is a common consequence of AKI, and hyperphosphatemia may be severe (10 to 20 mg/dL) in highly catabolic patients or when AKI is associated with rapid cell death as in rhabdomyolysis, severe burns, hemolysis, or tumor lysis. Factors that potentially contribute to hypocalcemia include skeletal resistance to the actions of parathyroid hormone, reduced levels of 1, 25-dihydroxyvitamin D, Ca ²⁺ sequestration in injured tissues such as muscle in the setting of rhabdomyolysis, and metastatic deposition of calcium phosphate salts in the setting of severe hyperphosphatemia. Hypocalcemia is usually asymptomatic, possibly because of the counterbalancing effects of acidosis on neuromuscular excitability. However, symptomatic hypocalcemia can occur in patients with rhabdomyolysis or acute pancreatitis or after treatment of acidosis with HCO ₃ ⁻ . Clinical manifestations of hypocalcemia include perioral paresthesias, muscle cramps, seizures, hallucinations, and confusion, as well as prolongation of the QT-interval and nonspecific T-wave changes on electrocardiogram. The Chvostek (contraction of facial muscles on tapping of the jaw over the facial nerve) and Trousseau (carpopedal spasm after occlusion of arterial blood supply to the arm for 3 minutes with a blood pressure cuff) signs are useful indicators of latent tetany in high-risk patients.

Mild asymptomatic hypermagnesemia is common in oliguric AKI and reflects impaired excretion of ingested magnesium (dietary magnesium, magnesium-containing laxatives, or antacids). ^, More significant hypermagnesemia is usually the result of overzealous parenteral magnesium administration, as in the management of AKI associated with preeclampsia. Hypomagnesemia occasionally complicates nonoliguric ATI associated with cisplatin or amphotericin B and, as with hypokalemia, likely reflects injury to the thick ascending limb of loop of Henle, a principal site for Mg ²⁺ reabsorption. ^{, ,} Hypomagnesemia is usually asymptomatic but may occasionally manifest as neuromuscular instability, cramps, seizures, cardiac arrhythmias, or resistant hypokalemia or hypocalcemia. ^,

Uric acid is cleared from blood by glomerular filtration and secretion by proximal tubule cells, and asymptomatic hyperuricemia (10–15 mg/dL) is not atypical in established AKI. Higher levels suggest increased production of uric acid and may point to a diagnosis of acute urate nephropathy. The urinary uric acid-to-creatinine ratio on a random specimen has been proposed as a means to distinguish between hyperuricemia caused by overproduction and impaired excretion. In a small series of patients, this ratio was >1 in 5 patients with acute uric acid nephropathy and was <1 in 27 patients with acute kidney injury due to other etiologies. In a subsequent case series, elevations in the uric acid-to-creatinine ratio to values of >1 were described in other etiologies of AKI, most notably in patients with infections who were markedly hypercatabolic.

Extracellular volume overload is an almost inevitable consequence of diminished salt and water excretion in AKI and may present clinically as mild hypertension, increased jugular venous pressure, pulmonary vascular congestion, pleural effusion, ascites, peripheral edema, increased body weight, and life-threatening pulmonary edema. Hypervolemia may be particularly troublesome in patients receiving multiple IV medications, high volumes of enteral or parenteral nutrition, and/or excessive volumes of maintenance IV fluids. Moderate or severe hypertension is unusual in ATI and should suggest other diagnoses such as hypertensive nephrosclerosis, glomerulonephritis, preeclampsia, renal artery stenosis, and other diseases of the renal vasculature. ^, Excessive water ingestion or administration of hypotonic saline or dextrose solutions can trigger hyponatremia, which, if severe, may cause cerebral edema, seizures, and other neurologic abnormalities. Cardiac complications include arrhythmias and myocardial infarction. Although these events may reflect primary cardiac disease, abnormalities in myocardial contractility and excitability may be triggered or compounded by hypervolemia, acidosis, hyperkalemia, and other metabolic sequelae of AKI. For further discussion on cardiorenal syndromes, see Chapter 42.

Anemia develops rapidly in AKI and is usually multifactorial in origin. Contributing factors include inhibition of erythropoiesis, hemolysis, bleeding, hemodilution, and reduced RBC survival time. Prolongation of the bleeding time is also common, resulting from mild thrombocytopenia, platelet dysfunction, and clotting factor abnormalities (e.g., factor VIII dysfunction).

Malnutrition remains one of the most frustrating and troublesome complications of AKI. The majority of patients have net protein breakdown, which may exceed 200 g/day in catabolic patients. Malnutrition is usually multifactorial in origin and may reflect inability to eat, loss of appetite, and/or inadequate nutritional support; the catabolic nature of the underlying medical disorder (e.g., sepsis, rhabdomyolysis, and trauma); nutrient losses in drainage fluids or dialysate; and increased breakdown and reduced synthesis of muscle protein and increased hepatic gluconeogenesis, probably through the actions of toxins, hormones (e.g., glucagon and parathyroid hormone), or other substances (e.g., proteases) that accumulate in AKI. Nutrition may also be compromised by the high incidence of acute gastrointestinal hemorrhage, which complicates up to 15% of cases of AKI. Mild gastrointestinal bleeding is common (10% to 30%) and usually due to stress ulceration of gastric or small intestinal mucosa. ^,

Infection is the most common and serious complication of AKI, occurring in 50% to 90% of cases and accounting for up to 75% of deaths. ^{, ,} It is unclear whether this high incidence of infection is due to a defect in host immune responses or repeated breaches of mucocutaneous barriers (e.g., intravenous cannulas, mechanical ventilation and bladder catheterization) resulting from therapeutic interventions.

Protracted periods of severe AKI or short intervals of catabolic anuric AKI often lead to development of uremic syndrome. Clinical manifestations of uremic syndrome, in addition to those already listed, include pericarditis; pericardial effusion; cardiac tamponade; gastrointestinal complications such as anorexia, nausea, vomiting, and ileus; and neuropsychiatric disturbances including lethargy, confusion, stupor, coma, agitation, psychosis, asterixis, myoclonus, hyperreflexia, restless leg syndrome, focal neurologic deficit, and/or seizures. The uremic toxin(s) responsible for this syndrome has yet to be defined. Candidate molecules include urea, other products of nitrogen metabolism such as guanidine compounds, products of bacterial metabolism such as aromatic amines and indoles, and other compounds that are inappropriately retained in the circulation in AKI or are underproduced, such as nitric oxide (NO).

Prerenal AKI is defined as hemodynamically mediated kidney dysfunction that is rapidly reversible following normalization of renal perfusion. Prevention of prerenal AKI from intravascular volume depletion pertains to early recognition and treatment of conditions that involve the loss of extracellular fluid including vomiting, diarrhea, excessive diuresis, and bleeding before hypoperfusion of the kidneys occurs. In patients in whom intravascular volume depletion leads to prerenal AKI, treatment consists of restoration of normal circulating blood volume. The optimal composition of administered fluids in patients with hypovolemic prerenal AKI depends on the source of fluid loss and associated electrolyte and acid-base disturbances. The initial management commonly consists of intravascular volume resuscitation with an isotonic crystalloid solution. Recent data demonstrate that balanced crystalloids reduce major adverse kidney events compared with isotonic saline in hospitalized patients. ^, Red blood cell transfusion should be used for hemorrhagic hypovolemia when there is ongoing bleeding, particularly if the patient is hemodynamically unstable or the blood hemoglobin concentration is dangerously low.

The relative merits of colloid and crystalloid resuscitation fluids in the management of nonhemorrhagic renal, extrarenal, and third-space fluid losses are controversial with advocates for the use of colloids positing that they are more effective at restoring circulating blood volume due to greater retention in the intravascular compartment. However, randomized controlled trials and meta-analyses comparing colloid with crystalloid replacement for resuscitation in critically ill patients have not confirmed this theoretical benefit and have demonstrated an increased need for KRT and other adverse outcomes associated with colloid formulations containing hydroxyethyl starch. In a meta-analysis of 55 trials involving 3504 patients randomly assigned to treatment with albumin or crystalloid, there was no evidence of either improved outcomes, decreased mortality or other complications associated with albumin administration. These results were subsequently confirmed in a nearly 7000-patient multicenter, randomized, controlled trial of fluid resuscitation in hypovolemic medical and surgical ICU patients in which 28-day survival, development of single or multiple organ failure, and duration of hospitalization were similar in both groups. Although specific data on the development of AKI were not described, the need for KRT was similar with saline compared with albumin resuscitation. However, in a post hoc analysis of patients with traumatic brain injury, albumin resuscitation was associated with increased mortality risk. The use of synthetic colloid solutions has been proposed as an alternative to albumin administration; however, hydroxyethyl starch preparations have been associated with an increased risk of AKI. In a multicenter, randomized, controlled trial comparing fluid resuscitation with hydroxyethyl starch with a 3% gelatin solution in 129 patients with sepsis, hydroxyethyl starch was associated with a more than twofold increased risk of AKI. A subsequent meta-analysis confirmed the increased risk of AKI associated with hydroxyethyl starch across 34 studies that included 2604 individuals. In an ensuing randomized controlled trial that included 7000 critically ill patients who were assigned to receive either 6% hydroxyethyl starch or isotonic saline, there was an approximately 20% increased risk of AKI treated with KRT with use of hydroxyethyl starch. On the basis of these data demonstrating no benefit and potential increased risk of AKI, along with the higher costs associated with colloid administration, their routine use for volume resuscitation in hypovolemia and sepsis is not advisable. In particular, hydroxyethyl starch solutions should be used sparingly. If utilized, there should be regular monitoring of kidney function. In such instances, the risk of hyperoncotic renal failure should be minimized by the concomitant use of appropriate crystalloid solutions. ^{, , , ,}

Experimental data have suggested that volume resuscitation with isotonic sodium chloride solutions, which contain supraphysiologic concentrations of chloride, may exacerbate renal vasoconstriction and diminish GFR as compared with isotonic crystalloid solutions with lower chloride content. In healthy patients, magnetic resonance imaging demonstrated that infusion of isotonic saline was associated with reduced renal blood flow velocity and renal cortical tissue perfusion as compared with administration of a reduced-chloride isotonic crystalloid solution. In a subsequent open-label, sequential period study conducted in a single ICU, replacing use of high-chloride IV solutions with fluids containing lower chloride content was associated with a reduction in the incidence of KDIGO stage 3 AKI from 14% to 8.4% and in the use of KRT from 10% to 6.3%. Subsequent prospective randomized clinical trials have yielded divergent results. ^{, , ,} Two of these were large pragmatic, cluster-randomized clinical trials conducted in parallel at the same institution. They compared the administration of isotonic saline with balanced crystalloid solution (i.e., fluid with electrolyte composition that more closely resembles plasma). ^, In the Saline Against Lactated Ringer’s of Plasma-Lyte in the Emergency Department (SALT-ED) trial of >13,000 non–critically ill patients, balanced crystalloid was associated with a decrease in the incidence of 30-day major adverse kidney events (i.e., death, new KRT, persistent kidney impairment defined by a >200% increase in serum creatinine at time of hospital discharge, within 30 days) compared with saline (4.7% vs. 5.6%, P = 0.01), but not with the primary outcome of hospital-free days (days alive after discharge to day 28). The Isotonic Solutions and Major Adverse Renal Events Trial (SMART) that included 15,802 critically ill patients found that compared with isotonic saline, balanced crystalloids were associated with a decrease in 30-day major adverse kidney events (14.3% vs. 15.4%, P = 0.04). In the SALT-ED trial, the benefit in the composite outcome was predominantly due to a lower rate of persistent kidney impairment while in SMART, 30-day mortality predominated. Collectively, these findings support a benefit of balanced crystalloid compared with isotonic sodium chloride, yet the benefits were not homogenous. The greatest benefit was present among the critically ill patients with sepsis and among non–critically ill patients who had a baseline serum creatinine >1.5 mg/dL or hyperchloremia (Cl >110 mmol/L) on initial presentation. Although the balanced fluids were more physiologic regarding their chloride content, they were hypotonic and associated with higher rates of hyponatremia.

The Balanced Solutions in Intensive Care Study (BaSICS) was a multicenter trial in Brazil that randomized 11,052 critically ill patients in a 2 × 2 factorial design to receive either isotonic saline or a balanced electrolyte solution infused rapidly or at a slower rate. ^, In this study, neither electrolyte composition nor rate of fluid administration was associated with mortality at 90 days or with the development of AKI. Similarly, the Plasma-Lyte 148 versus Saline (PLUS) study randomized 5037 critically ill patients in Australia and New Zealand to receive Plasma-Lyte, a balanced electrolyte solution, or isotonic saline as intravenous fluids in the ICU. As in BaSICS, the use of balanced electrolyte solutions was not associated with either a reduction in mortality, AKI, or need for acute KRT as compared with isotonic saline. A subsequent systematic review and meta-analysis that included all four of these studies and included data from 34,450 patients found risk ratios for 90-day mortality of 0.96 (95% confidence interval [CI]: 0.91–1.01), AKI of 0.96 (95% CI: 0.89–1.02), and need for acute KRT of 0.95 (95% CI: 0.81–1.11). Using Bayesian techniques, the authors estimated that the effect of using balanced crystalloid solutions in place of isotonic saline ranged from a 9% relative reduction to a 1% relative increase in the risk of death at 90 days and from an 11% reduction to a 2% relative increase in the risk of developing AKI.

The volume and electrolyte content of urinary and gastrointestinal losses, as well as patients’ serum electrolytes and acid-base status, should be closely monitored to guide adjustments in the composition of the replacement fluids. Although the potassium content in gastric juices tends to be low, concomitant urinary potassium losses may be quite high as the result of metabolic alkalosis.

Management of AKI in the setting of heart failure depends on the clinical setting and etiology of the heart failure. In patients with heart failure in whom AKI has developed in the setting of excessive diuresis, withholding of diuretics and administering cautious volume replacement may be sufficient to restore kidney function. In acute decompensated heart failure (ADHF), AKI may develop despite worsening volume overload; intensification of diuretic therapy is often required for treatment of pulmonary vascular congestion. Although diuretic therapy may exacerbate prerenal AKI, it can also result in improvement in kidney function via several postulated mechanisms: 1) decreasing ventricular distention resulting in a shift from the descending limb to the ascending limb of the Starling curve and improvement in myocardial contractility; 2) decreasing venous congestion ^, ; and 3) diminishing intraabdominal pressure. Additional therapies for ADHF in the setting of AKI include inotropic support, vasodilators for afterload reduction, and mechanical support including intraaortic balloon pumps and ventricular assist devices. The use of invasive hemodynamic monitoring in ADHF has been controversial; although it is often used to guide pharmacologic management, clinical data have not demonstrated improved renal outcomes when management is guided by pulmonary artery catheters. The role of isolated ultrafiltration in ADHF is also controversial. Although negative fluid balance can be achieved more readily using extracorporeal ultrafiltration than with conventional diuretic therapy, studies have not demonstrated differences in kidney function or survival. In the Ultrafiltration versus Intravenous Diuretics for Patients Hospitalized for Acute Decompensated Heart Failure (UNLOAD) trial, hypervolemic patients with heart failure who were randomized to isolated ultrafiltration had more rapid fluid loss and decreased rehospitalizations within 90 days as compared with patients randomized to diuretic therapy, with no differences in kidney function. In contrast, in the subsequent Cardiorenal Rescue Study in Acute Decompensated Heart Failure (CARRESS-HF) trial, ultrafiltration was inferior to diuretic therapy with respect to the bivariate endpoint of change in serum creatinine level and body weight 96 hours after enrollment ( P = 0.003), owing primarily to worsening of kidney function in the ultrafiltration group. On the basis of these data, extracorporeal ultrafiltration cannot be recommended for primary management of patients with decompensated heart failure.

Although volume-responsive prerenal azotemia is common in patients with advanced liver disease, differentiation from HRS and intrinsic AKI may be difficult. While patients with liver failure are typically total body sodium overloaded with peripheral edema and ascites, true hypovolemia or reduced effective systemic arterial blood volume is often an important contributory factor to the development of AKI. The underlying pathophysiology of salt and water retention in cirrhosis involves multiple pathways. Portal hypertension leads directly to ascites formation, while splanchnic and peripheral vasodilatation results in a state of relative arterial underfilling, which activates neurohumoral vasoconstrictors that produce intrarenal vasoconstriction, salt and water retention, and decreased GFR. Volume-responsive AKI may develop in the setting of excessive diuresis, increased gastrointestinal losses (often as the result of therapy for hepatic encephalopathy), rapid drainage of ascites, or spontaneous bacterial peritonitis. Worsening hepatic function is often associated with diuretic resistance and progressive or precipitous worsening of kidney function. It has been postulated that an inadequate increase in cardiac output in response to the fall in peripheral vascular resistance may be central to development of HRS.

Differentiation between volume-responsive prerenal AKI and the HRS is based on the clinical response to volume loading. The optimal fluid for volume expansion in this setting has been controversial. Most recent expert opinion advocated the use of hyperoncotic (20% or 25%) albumin at a dose of 1 g/kg per day. ^, However, there is an absence of rigorous data supporting this regimen as compared with volume expansion with isotonic crystalloid solutions. There are more data regarding the use of albumin infusion to prevent AKI in patients undergoing large-volume (>5 L) paracentesis ^{, ,} and in the treatment of spontaneous bacterial peritonitis. In a randomized controlled trial, patients undergoing paracentesis who received an infusion of approximately 10 g of albumin per liter of drained ascites experienced less activation of the renin-angiotensin system and a significantly lower rate of worsening kidney function than patients who did not receive albumin infusion. In a subsequent study, albumin infusion was superior to administration of either dextran or gelatin solutions in preventing AKI after large-volume paracentesis. Current recommendations are to infuse 6 to 8 g of albumin per liter of ascites drained when paracentesis volume exceeds 5 L. In a randomized controlled trial comparing antibiotics alone with antibiotics plus albumin for the treatment of spontaneous bacterial peritonitis, infusion of 1.5 g/kg of albumin at the initiation of treatment and an additional 1 g/kg on the third day of treatment was associated with reduced rates of both AKI and mortality, although the benefit appears to be restricted to patients in whom the serum creatinine is >1 mg/dL, the BUN is >30 mg/dL, or the total bilirubin is >4 mg/dL.

Definitive therapy of HRS requires restoration of hepatic function, usually achieved through liver transplantation. ^, The role of peritoneovenous shunting (e.g., LeVeen and Denver shunts) in HRS has been inadequately studied. In a subset of 33 patients with HRS included in a randomized trial comparing peritoneovenous shunts with medical therapy, shunting was not associated with improved survival. These data need to be interpreted with caution due to the small sample size and because data on improvement in kidney function were not reported. In addition, because of poor long-term patency rates and high rates of complications, particularly encephalopathy, the use of peritoneovenous shunts has largely been supplanted by transjugular portosystemic shunts (TIPS). TIPS has been demonstrated to provide better control of ascites than sequential paracentesis, and in one series lower rates of HRS, albeit with a higher risk of encephalopathy. In a small case series, TIPS was reported to be effective as primary therapy for HRS but has not been evaluated in a randomized trial. Pharmacologic therapy with vasoconstrictors, when combined with albumin infusion, has been associated with improvement in kidney function in patients with HRS. ^, Agents that have shown benefit include norepinephrine, the combination of octreotide and midodrine, ^, and the V ₁ -vasopressin receptor agonist terlipressin, ^, although only terlipressin has been evaluated in randomized controlled trials. In the Reversal of Hepatorenal Syndrome Type 1 with Terlipressin (REVERSE) trial, similar rates of confirmed reversal of HRS were found with terlipressin and placebo (19.6% vs. 13.1%, P = 0.22), although the mean decrease in serum creatinine was more pronounced with terlipressin (1.1 mg/dL vs. 0.6 mg/dL, P = 0.001). In the Study to Confirm Efficacy and Safety of Terlipressin in Hepatorenal Syndrome Type 1 (CONFIRM), verified reversal of HRS was observed in 32% of patients in the terlipressin group and 17% in the placebo group. However, mortality was higher among terlipressin-treated patients (51% vs. 45%), fewer terlipressin-treated patients received liver transplants (23% among terlipressin-treated patients vs. 29% among placebo-treated patients), and terlipressin was associated with more adverse events, particularly respiratory failure, with death due to respiratory disorders in 11% of patients treated with terlipressin as compared with 2% of patients in the placebo arm. In a subsequent meta-analysis that included 974 patients, the risk ratio for reversal of HRS was 2.1 (95% CI: 1.5–2.9) with a 40% reduction in the need for KRT within 30 days but no benefit in 90-day survival. In a meta-analysis of studies comparing terlipressin to norepinephrine, both agents were associated with similar efficacy in reversal of HRS but adverse events were more common with terlipressin. On the basis of these data, terlipressin has been approved for use in the United States. However, given the risk of respiratory failure, it should not be used in patients with hypoxia, it should not be used in patients with grade 3 acute-on-chronic liver failure, and given a low probability of benefit, it should not be used in patients whose serum creatinine is >5 mg/dL.

Acute kidney injury can result from elevations in intraabdominal pressure, resulting in a clinical presentation with similar features to prerenal AKI. The abdominal compartment syndrome is defined by an intraabdominal pressure ≥20 mm Hg associated with dysfunction of one or more organ systems. However, intraabdominal pressures lower than 20 mm Hg may be associated with abdominal compartment syndrome, while values higher than this threshold do not universally lead to abdominal compartment syndrome. Abdominal compartment syndrome typically develops in critically ill patients, most commonly in the setting of trauma with abdominal hemorrhage, abdominal surgery, massive fluid resuscitation, liver transplantation, and gastrointestinal conditions including peritonitis and pancreatitis. Mechanisms underlying the development of AKI in abdominal compartment syndrome are believed to involve renal vein compression and constriction of the renal artery from sympathetic and renin-angiotensin system activation and reduced cardiac output. Oliguria, which can lead to anuria, often develops and, as is true for other forms of AKI associated with impaired renal perfusion, urine sodium concentration is commonly reduced.

The diagnosis of abdominal compartment syndrome, which should be suspected in patients with acute abdominal distension and/or rapidly accumulating ascites or abdominal trauma, can be made by simple transduction of bladder pressure. ^{, ,} Treatment is prompt abdominal decompression; if ascites is present, decompression may be achieved by performing large-volume paracentesis and in patients with severe ileus or colonic distension, bowel decompression may be sufficient, yet surgical laparotomy is often required for definitive therapy.

Strategies to prevent intrinsic AKI vary on the basis of the specific etiology of kidney injury. Optimization of cardiovascular function and restoration of intravascular volume status are key interventions to minimize the risk that prerenal AKI evolves into ischemic ATI. There is compelling evidence that aggressive intravascular volume expansion reduces the incidence of ATI after major surgery or trauma, burns, and cholera. ^{, , ,} AKI due to sepsis is common and is associated with mortality rates as high as 80%. ^{, ,} The role of early goal-directed therapy (EGDT) using resuscitation to defined hemodynamic targets (MAP >65 mm Hg, CVP 10–12 mm Hg, urine output >0.5 mL/kg per hour, ScvO ₂ >70%) using a combination of crystalloid solutions, red cell transfusion, and vasopressors guided by invasive hemodynamic monitoring in improving overall outcomes and decreasing the risk of AKI has been controversial. In a seminal single-center randomized controlled trial, EGDT resulted in a significant reduction in overall organ dysfunction and mortality in patients presenting with severe sepsis or septic shock, although specific data on the incidence of AKI were not reported. However, EGDT was not associated with a reduction in dialysis-requiring AKI in the Protocolized Care for Early Septic Shock (ProCESS), the Australasian Resuscitation in Sepsis Evaluation (ARISE), or the Protocolized Management in Sepsis (PRoMISe) trials or in a patient-level meta-analysis that combined data from these trials ( Fig. 28.1 ). Although the benefits of EGDT were not confirmed in these three later trials, early recognition of sepsis, prompt initiation of antibiotic therapy, and rapid volume resuscitation and hemodynamic stabilization improve outcomes and are likely to minimize the risk of AKI. The role of maintenance of normoglycemia in critically ill patients in minimizing the risk of AKI has also been controversial. Two single-center randomized controlled trials that used intensive insulin management to maintain blood glucose levels of 80 to 110 mg/dL as compared with conventional management maintaining the glucose concentration between 180 and 220 mg/dL each resulted in decreased rates of AKI, defined on either the basis of change in serum creatinine or the need for KRT. However, the benefits of tight glycemic control were not confirmed in the Normoglycemia in Intensive Care Evaluation–Survival Using Glucose Algorithm Regulation (NICE-SUGAR) trial, a 6104-patient multicenter trial that compared intensive therapy to achieve a target glucose of approximately 80 to 110 mg/dL with more conventional therapy designed to maintain the blood glucose <180 mg/dL. In the NICE-SUGAR trial, intensive glycemic control was associated with an increased risk of hypoglycemia, an increased mortality risk (27.5% vs. 24.9%, P = 0.02), and no reduction in the need for KRT.

One-year survival comparing early goal-directed therapy (EGDT) to usual care in patients with sepsis Kaplan-Meier survival curves comparing EGDT to usual care in patients with sepsis from pooled patient-level data from the ProCESS, ARISE, and ProMISe trials.

Kidney replacement therapy was required in 11.0% of patients who were randomized to EGDT as compared with 10.6% of patients randomized to usual care (OR: 1.02; 95% CI: 0.81–1.28; P = 0.88).

From The PRISIM Investigators. Goal-directed therapy for septic shock—a patient-level meta-analysis. N Engl J Med. 2017;376:2223–2234.

In surgical patients, avoidance of hypotension has been associated with a decreased risk of AKI. In a retrospective analysis of more than 33,000 patients who underwent noncardiac surgery, episodes of intraoperative hypotension with a mean arterial blood pressure <55 mm Hg were associated with a marked increase in the probability of AKI. The adjusted odds of AKI with intraoperative hypotension increased with duration of hypotension, from an odds ratio of 1.2 with <10 minutes of hypotension, to 1.3 with 10 to 20 minutes of hypotension and 1.5 with more than 20 minutes of hypotension. Volume overload is typical after surgical procedures and small trials suggested better patient outcomes following abdominal surgery using a restrictive fluid management strategy. However, a large randomized controlled trial comparing restrictive and liberal fluid management strategies in patients undergoing major abdominal surgery at increased risk of complications found no benefit with regard to the primary outcome of 1-year disability-free survival but found a higher rate of acute kidney injury associated with the restrictive fluid strategy (8.6% vs. 5.0%, P < 0.001).

Intravascular volume depletion has been identified as a risk factor for ATI resulting from iodinated contrast, rhabdomyolysis, hemolysis, cisplatin, amphotericin B, multiple myeloma, aminoglycosides, and other nephrotoxins; crystal-associated AKI related to acyclovir and acute urate nephropathy; and AKI stemming from hypercalcemia. ^{, , , , , , ,} Restoration of adequate intravascular volume prevents the development of experimental and human ATI in many of these clinical settings. Avoidance of potentially nephrotoxic medications or insults in high-risk patients and settings is also important to reduce the risk for ATI. Specifically, among patients with advanced cardiac and/or liver disease in whom renal perfusion may be diminished, the use of selective or nonselective NSAIDs that inhibit the production of vasodilatory prostaglandins may exacerbate intrarenal vasoconstriction and precipitate AKI. Diuretics, NSAIDs (including selective cyclo-oxygenase-II [COX-II] inhibitors), ACE inhibitors, angiotensin receptor blockers, and other inhibitors of the renin-angiotensin-aldosterone system (RAAS) should be used with caution in patients with suspected absolute or effective intravascular volume depletion or in patients with renovascular disease as these agents may convert reversible prerenal AKI to ischemic ATI. The combined use of agents that block the RAAS, diuretics, and NSAIDs has been identified as a risk factor for AKI, particularly among patients with heart failure, liver failure, or other conditions with reduced baseline renal perfusion. ^,

Careful monitoring of circulating drug levels appears to reduce the incidence of AKI associated with aminoglycoside antibiotics and calcineurin inhibitors. The observation that the antimicrobial efficacy of aminoglycosides persists in tissues even after the drug has been cleared from the circulation (postantibiotic killing) has led to the use of once-daily dosing with these agents. Dosing regimens that provide higher peak drug levels but less frequent administration appear to provide comparable antimicrobial activity and less nephrotoxicity than older conventional dosing regimens. Nephrotoxicity of drugs may also be reduced through changes in formulation. For example, the use of lipid-encapsulated formulations of amphotericin B may decrease the risk of amphotericin-induced AKI.

Contrast-associated AKI (CA-AKI) has historically been defined by and commonly manifests clinically as small absolute (>0.5 mg/dL) and/or relative (>25%) increases in serum creatinine that occur within 2 to 4 days post iodinated contrast administration. While severe AKI is relatively uncommon due to iodinated contrast alone, the incidence of CA-AKI defined by relatively minor increments in serum creatinine has been shown in past studies to occur in up to 15% or more of high-risk patients. A series of recent observational retrospective studies have questioned the true incidence of CA-AKI and challenged the concept that iodinated contrast is nephrotoxic, particularly among individuals receiving intravenous (IV) contrast media for computed tomography. A meta-analysis of 13 nonrandomized studies that included 25,950 patients found a similar risk of AKI in patients who received IV iodinated contrast compared with patients who underwent radiographic procedures without intravascular contrast (relative risk [RR] = 0.79; 95% CI: 0.62–1.02). Similarly, an observational study of more than 29 million patients hospitalizations found that the risk for AKI among patients who received intravascular iodinated contrast during their hospitalization was comparable with the risk observed among patients who had not received contrast (RR 0.93; 95% CI 0.88–0.97). While the results of these retrospective analyses are adjusted for medical comorbidity and other risk factors for renal injury, such adjustment cannot fully account for all factors that influence the decision to administer contrast, rendering it likely that patients in these studies who did not receive contrast may have been at higher risk for AKI than patients who received contrast.

The apparent reduction in risk of kidney toxicity from iodinated contrast over time likely reflects advancement in technology. Over the past 25 years, there has been considerable progress in developing less nephrotoxic contrast agents. The use of lower-osmolal contrast agents in place of the older and more nephrotoxic high-osmolal agents resulted in a decreased incidence of CA-AKI. ^, Data regarding the added benefit associated with the iso-osmolal radiocontrast agent iodixanol have been less consistent and may reflect heterogeneity in the risk of CA-AKI associated with specific lower-osmolal agents. In addition, imaging technology has improved with use of higher-resolution computed tomography scanners and improved imaging technology for angiography, allowing reduction in the volume of administered contrast media, further reducing the risk of toxicity.

On the basis of these data, the American College of Radiology and National Kidney Foundation have issued consensus statements on the use of intravenous iodinated contrast media. On the basis of an extensive review of published data, they estimated the risk of AKI following contrast administration to be similar to that of the general population for individuals with an eGFR of 45 mL/min/1.73 m ² or higher, to be 0% to 2% for individuals with an eGFR of 30 to 44 mL/min/1.73 m ² and to be 0% to 17% for individuals with an eGFR <30 mL/min/1.73 m ² and recommend prophylaxis for prevention of AKI in individuals undergoing imaging with intravenous contrast who have an eGFR <30 mL/min/1.73 m ² and are not undergoing maintenance dialysis and in high-risk patients with an eGFR of 30 to 44 mL/min/1.73 m ² . Similar guidance was developed for the management of patients with kidney disease requiring cardiovascular catheterization or intervention in a workshop jointly sponsored by the National Kidney Foundation and Society of Cardiovascular Angiography and Interventions (SCAI).

Although is prudent to consider intravascular iodinated contrast as potentially nephrotoxic and to implement evidence-based preventive care in patients at risk for CA-AKI, excessive concern regarding the risk of CA-AKI should not prevent performance of clinically indicated diagnostic and therapeutic procedures. Two decades ago, Chertow and colleagues observed that elderly individuals with CKD who presented with acute myocardial infarction were less likely to undergo coronary angiography than propensity-matched individuals without CKD, a finding that they described as “renalism.” They further estimated that the failure to perform clinically indicated angiography was associated with an increase in mortality risk. In an analysis of U.S. veterans hospitalized with acute coronary syndrome between 2013 and 2017, individuals with CKD were again less likely to undergo angiography despite matching for clinical indication for angiography based on the Global Registry of Acute Coronary Events (GRACE) 2.0 risk score, with nonperformance of angiography associated with an adjusted risk ratio for all-cause mortality at 6 months of 1.4 (95% CI: 1.3–1.5). Thus while CA-AKI is associated with increased morbidity and mortality, failure to perform clinically indicated diagnostic and therapeutic interventions because of excessive concern regarding the risk of AKI also has untoward consequences.

Interventions to ameliorate risk of contrast-associated acute kidney injury Periprocedural intravenous crystalloid

Periprocedural Intravenous Crystalloid

Given the known timing of administration of iodinated contrast, prevention of CA-AKI has been the focus of numerous clinical trials ( Table 28.15 ). Past studies have demonstrated that the administration of IV fluids to high-risk patients before and after exposure to intravascular iodinated contrast diminished the risk for contrast-associated AKI (CA-AKI), although the optimal regimen for fluid administration is unknown. ^{, ,} In a small clinical trial that was stopped early due to safety concerns, periprocedural IV isotonic saline was associated with a markedly lower rate of AKI after contrast exposure than oral fluid administration. In a larger randomized trial, isotonic IV saline significantly reduced the incidence of CA-AKI following coronary angiography compared with IV half-normal saline with a particular benefit noted in diabetic patients and those receiving large volumes of contrast. One study challenged the principle that IV fluid administration reduces the risk for CA-AKI. This noninferiority clinical trial randomized 660 patients with baseline CKD undergoing procedures with iodinated contrast to receive IV isotonic saline or no IV fluids and found a similar rate of CA-AKI (2.7% with saline vs. 2.6% without fluid), concluding that withholding IV fluid was not inferior to administering IV saline with regard to the prevention of small increments in serum creatinine. A subsequent study randomized 523 patients with stage 3 CKD undergoing contrast-enhanced computed tomography to no preprocedural volume administration or preprocedural isotonic sodium bicarbonate. Overall, AKI developed in 2.1% of participants, 2.7% in the group receiving no preprocedural fluids, and 1.5% in participants administered preprocedural intravenous sodium bicarbonate. Interpretation of both trials is limited by relatively low sample size and inclusion of a predominantly low-risk population who did not meet criteria for periprocedural prophylactic intervention proposed by the American College of Radiology–National Kidney Foundation consensus recommendations.

Table 28.15

Effectiveness of Preventive Interventions for Contrast-Associated Acute Kidney Injury

No benefit	Unclear benefit	Beneficial
Loop diuretics a	Statins	Isotonic intravenous fluids
Mannitol a	Natriuretic peptides	Low/iso-osmolal contrast media
Dopamine a	Theophylline/aminophylline
Fenoldopam a	Ascorbic acid
Hemodialysis a	Hemofiltration
N-acetylcysteine Sodium bicarbonate b

Intravenous Sodium Bicarbonate.

A number of small clinical trials compared the effects of isotonic sodium bicarbonate compared with isotonic saline for prevention of contrast-associated AKI. These studies were generally underpowered and yielded conflicting results. Subsequent meta-analyses concluded that there was an overall benefit associated with bicarbonate administration with regard to AKI defined by small changes in serum creatinine, although there was no demonstrable benefit with regard to the need for dialysis, leading several clinical practice guidelines to recommend the administration of either IV isotonic sodium chloride or sodium bicarbonate to high-risk patients receiving iodinated contrast. ^{, ,} However, the Prevention of Serious Adverse Events Following Angiography (PRESERVE) trial provided more definitive findings regarding the comparative effects of sodium bicarbonate and sodium chloride for the prevention of CA-AKI. PRESERVE was a multinational randomized clinical trial that used a 2 × 2 factorial design to compare IV isotonic sodium bicarbonate with IV isotonic sodium chloride and oral N-acetylcysteine with placebo for the prevention of serious adverse outcomes and CA-AKI following angiographic procedures in 4993 patients with CKD. IV sodium bicarbonate did not reduce the incidence of a primary outcome composed of 90-day death, need for dialysis, or persistent decline in kidney function (OR = 0.93; 95% CI: 0.72–1.22). Similarly, bicarbonate did not reduce the incidence of CA-AKI assessed 3 to 5 days post angiography (OR = 1.16; 95% CI: 0.96–1.41). While the optimal rate and duration of IV saline administration is not known, it is reasonable to administer isotonic sodium chloride at a rate of 1 mL/kg per hour for 6 to 12 hours before and 6 to 12 hours after a contrast-enhanced procedure in at-risk hospitalized patients. For at-risk outpatients, an alternative regimen of 3 mL/kg over 1 hour before the procedure followed by 6 mL/kg administered over 2 to 6 hours following the procedure may be more feasible.

N-acetylcysteine

N-acetylcysteine (NAC) is an antioxidant with vasodilatory properties that was postulated to potentially prevent CA-AKI based on its capacity to scavenge ROS, reduce the depletion of glutathione, and stimulate the production of vasodilatory mediators including nitric oxide. ^, Clinical trials of oral and IV NAC have yielded conflicting findings. Although initially used at a dose of 600 mg twice daily, subsequent studies suggested greater efficacy with higher doses of up to 1200 mg twice daily. ^, In the Acetylcysteine for Contrast-Induced Nephropathy (ACT) trial, 2308 patients were randomized to receive 1200 mg of NAC or placebo twice daily beginning before the procedure and continuing for three doses post procedure. No differences were observed in the incidence of CA-AKI at 48 to 96 hours post contrast administration or in the incidence of death or need for dialysis within 30 days. However, the overall study population had relatively well-preserved kidney function, with a median serum creatinine of 1.1 mg/dL, and less than 16% of patients had a baseline serum creatinine >1.5 mg/dL. In the aforementioned PRESERVE trial, the administration of NAC in a dose of 1200 mg by mouth twice daily for 5 days beginning just before angiography was not associated with a reduction in 90-day death, need for dialysis, or persistent impairment in kidney function (OR = 1.02; 95% CI: 0.78–1.33) or in a reduction in CA-AKI (OR = 1.06; 95% CI: 0.87–1.28). On the basis of these findings, NAC should not be used to reduce the risk of CA-AKI.

Other Agents.

Trials of other pharmacologic interventions including furosemide, dopamine, fenoldopam, calcium channel blockers, and mannitol have failed to demonstrate significant benefit and, in some cases, have been associated with an increased risk of CA-AKI ( Table 28.15 ). Studies on the benefit of natriuretic peptides, aminophylline, theophylline, and ascorbic acid have also yielded conflicting results. Given the absence of convincing data on the efficacy of these interventions, as well as potential safety concerns with the use of natriuretic peptides, aminophylline, and theophylline in patients with cardiovascular disease, their routine use is not recommended. There have been multiple trials and meta-analyses that investigated statins for the prevention of CA-AKI. While many, albeit not all, demonstrated a benefit of statins, particularly in high dose, with respect to the development of CA-AKI, the effect of this class of medication on more serious outcomes including need for dialysis and progressive CKD remains unclear. Furthermore, in the majority of patients requiring angiographic procedures, there are likely other indications for statin therapy. Kidney replacement therapies for the prevention of CA-AKI have been largely ineffective, and in some instances the use of “prophylactic” hemodialysis has been associated with harm. The interpretation of studies of hemofiltration for prevention of CA-AKI is confounded by their use of change in serum creatinine as an endpoint because hemofiltration lowers serum creatinine concentration. ^, Given the risks associated with intravenous line placement and the kidney replacement procedures themselves, along with lack of definitive benefit, use of dialysis or hemofiltration to prevent CA-AKI is not currently recommended ^, ( Fig. 28.2 ).

Algorithm for mitigation of risk of contrast-associated acute kidney injury (CA-AKI).

COX-2, Cyclo-oxygenase-2; NSAIDs, nonsteroidal antiinflammatory drugs.

Allopurinol (10 mg/kg/day in three divided doses, or 100 mg/m ² every 8 hours, max 800 mg per day) is useful for limiting uric acid generation in patients at high risk for acute urate nephropathy; however, AKI can develop despite the use of allopurinol, probably through the toxic actions of hypoxanthine crystals on tubule function. ^{, , , , ,} In the setting of high rates of uric acid generation such as tumor lysis syndrome, the use of recombinant urate oxidase (rasburicase, 0.05–0.2 mg/kg) may be more effective. Rasburicase catalyzes the degradation of uric acid to allantoin and has been shown to be effective both as prophylaxis and treatment for acute uric acid–mediated tumor lysis syndrome and to prevent the development of AKI due to tumor lysis syndrome associated hyperuricemia. ^{, ,} In oligoanuric patients, prophylactic hemodialysis may be used to acutely lower uric acid levels.

Amifostine, an organic thiophosphate, has been demonstrated to ameliorate cisplatin nephrotoxicity in patients with solid organ or hematologic malignancies. N-acetylcysteine limits acetaminophen-induced renal injury if given within 24 hours of ingestion, and dimercaprol, a chelating agent, may prevent heavy metal nephrotoxicity. ^, Ethanol inhibits ethylene glycol metabolism to oxalic acid and other toxic metabolites, but its use has been largely replaced by fomepizole, an inhibitor of alcohol dehydrogenase that decreases production of ethylene glycol metabolites and prevents the development of AKI.

Remote ischemic preconditioning (RIPC) has been investigated as a potential intervention for the prevention of AKI. RIPC involves the implementation of brief episode(s) of ischemia/reperfusion of distant tissue (e.g, with sequential brief inflation and deflation of a blood pressure tourniquet on a limb that is hypothesized to enhance the kidneys’ resistance to a subsequent more prolonged period of ischemia by means of hormonal mediators, hormonal and neuronal signaling pathways, and antiinflammatory and molecular mediators ( Fig. 28.3 ). Various trials and meta-analyses have examined the benefit of RIPC in the setting of cardiac surgery. A trial that randomized 240 high-risk patients undergoing cardiac surgery to RIPC or sham RIPC demonstrated a lower rate of AKI, defined by KDIGO criteria, with RIPC (37.5% vs. 52.5%, P = 0.02), with no effect on secondary endpoints including myocardial infarction, stroke, or death. A considerably larger trial that enrolled 1612 patients undergoing cardiac surgery found no difference in the rate of AKI (a secondary end point) between RIPC and sham conditioning groups or in the incidence of the combined primary end point of death from cardiovascular causes, nonfatal myocardial infarction, coronary revascularization, or stroke within 12 months. A series of meta-analyses of studies that included patients undergoing cardiac and, in some instances, vascular surgery found lower rates of AKI with RIPC compared with control but failed to demonstrate a benefit on need for KRT or death. On the basis of these cumulative data, the use of RIPC for the prevention of adverse outcomes following cardiac and vascular procedures is not routinely recommended.

Postulated mechanisms for remote ischemic preconditioning.

Postulated mechanisms for renal protective effects of remote ischemic preconditioning. AP-1, Activator protein-1; cGMP, cyclic guanosine monophosphate; CGRP, calcitonin gene–related peptide; COX2, cyclooxygenase 2; HIF-1α, hypoxia-inducible factor 1α; HSP, heat shock protein; iNOS, inducible nitric oxide synthase; JAK, Janus kinase; MEK, MAPK kinase; mPTP, mitochondrial permeability transition pore; Nrf2, nuclear factor (erythroid-derived 2)-like 2; STAT1/3, signal transducer and activator of transcription.

From Gassnov N, Nia AM, Caglayan E, Er F. Remote ischemic preconditioning and renoprotection: from myth to a novel therapeutic option? J Am Soc Nephrol . 2014;25(2):216–224.

Historically, low-dose dopamine (“renal-dose” dopamine; <2 mg/kg/min) was widely advocated for the management of oliguric AKI. In experimental studies of animals and healthy human volunteers, low-dose dopamine increased renal blood flow and to a lesser extent GFR. However, low-dose dopamine has not been demonstrated to prevent or alter the course of ischemic or nephrotoxic ATI in prospective clinical trials. This absence of clinical benefit may relate to differences in the hemodynamic response to low-dose dopamine in patients with renal disease as compared to healthy individuals. In contrast to the reduction in renal resistive index associated with low-dose dopamine in critically ill patients without kidney disease, dopamine infusion is associated with an increase in renal resistance in patients with AKI. Moreover, dopamine, even at low doses, is potentially toxic in critically ill patients and can induce tachyarrhythmias, myocardial ischemia, as well as extravasation necrosis. Thus the routine administration of low-dose dopamine to ameliorate or reverse the course of AKI is not justified on the basis of the balance of experimental and clinical evidence. ^,

Fenoldopam is a selective postsynaptic dopamine agonist that acts on D1 receptors and mediates more potent renal vasodilatation and natriuresis than dopamine. However, fenoldopam is a potent antihypertensive agent and causes hypotension by decreasing peripheral vascular resistance. Several small studies suggested that fenoldopam could reduce the incidence of AKI in high-risk clinical situations ^, ; however, a subsequent larger randomized trial comparing fenoldopam to standard hydration in patients undergoing invasive angiographic procedures found no benefit in regard to decreasing the incidence of contrast-associated AKI. In another large RCT, fenoldopam administration failed to reduce mortality or the need for KRT in ICU patients with early ATI. Therefore there is currently no clinical role for fenoldopam in the prevention or treatment of AKI.

ANP is a 28–amino acid polypeptide synthesized in cardiac atrial muscle. ^, ANP augments GFR by triggering afferent arteriolar vasodilatation and constriction of the efferent arteriole. ^, In addition, ANP inhibits sodium transport and lowers oxygen requirements in several nephron segments. ^, Synthetic analogs of ANP showed promise in the management of ATI in the laboratory setting; however, these benefits in animal models of AKI have failed to translate into clinical benefit in humans. A large multicenter, prospective, randomized placebo-controlled trial of anaritide, a synthetic analog of ANP, in patients with ATI failed to show clinically significant improvement in dialysis-free survival or overall mortality, although there was an improvement in dialysis-free survival in oliguric patients. This benefit in oliguric patients was not confirmed in a subsequent prospective study. It has been suggested that the absence of benefit may be related to both the relatively late initiation of therapy and the effect of ANP on systemic blood pressure. In a subsequent pilot study, low-dose recombinant ANP administration in high-risk cardiac surgery patients was associated with a reduction in the requirement for postoperative KRT. Until these results are confirmed in a larger, multicenter trial, the use of ANP in this setting cannot be recommended. Trials of ANP for the prevention of contrast-associated AKI have generated mixed results. ^, Ularitide (urodilantin) is a natriuretic pro-ANP fragment produced within the kidney. In a small randomized trial, ularitide did not reduce the need for dialysis in patients with AKI. A recent meta-analysis of ANP for the treatment of AKI concluded that the paucity of high-quality studies precluded a determination of the effects of this therapy.

High-dose IV diuretics are commonly prescribed to increase urine output in patients with oliguric AKI. Although this strategy assists in volume management and minimizes the risk of progressive volume overload, there is no evidence that diuretic therapy alters the natural history of AKI or improves mortality or dialysis-free survival. In a retrospective analysis, diuretic therapy was associated with an increased risk of death and nonrecovery of renal function. These risks were restricted, however, to patients who did not respond to diuretic administration with increased urine volume; in diuretic-responsive patients, outcomes were similar to untreated patients. In a prospective randomized trial, high-dose IV furosemide augmented urine output but did not alter the outcome of established AKI. In a post-hoc analysis of data from the Fluid and Catheter Treatment Trial, a positive fluid balance after AKI in patients with acute lung injury was strongly associated with increased mortality while diuretic therapy was associated with improved 60-day patient survival. Given the risks of loop diuretics in AKI, including irreversible ototoxicity and exacerbation of prerenal AKI, these agents should be used solely to facilitate the management of extracellular volume overload (vide infra). Of note, a single administration of furosemide in a dose of 1.0 t 1.5 mg/kg has been shown to help characterize the risk for progressive AKI. Specifically, a urine volume of <200 mL in the 2 hours following furosemide administration demonstrated sensitivity and specificity for progression to AKIN stage 3 of 87.1% and 84.1%, respectively.

The osmotic diuretic mannitol, which also has renal vasodilatory and oxygen-free radical scavenging properties, has been investigated as a preventive treatment for AKI. ^, No adequate data exist to support the routine administration of mannitol to oliguric patients. Moreover, when administered to severely oliguric or anuric patients, mannitol may trigger expansion of intravascular volume and pulmonary edema, as well as severe hyponatremia due to an osmotic shift of water from the intracellular to the intravascular space. ^,

The management of acute vasculitis involving the kidney and acute glomerular disease is covered in detail in Chapters 31 to 34. AKI caused by acute glomerulonephritis or vasculitis may respond to corticosteroids, alkylating agents, rituximab, and plasmapheresis depending on the primary etiology of the disease. Plasma exchange is useful in the treatment of sporadic TTP and possibly sporadic HUS in adults. ^, The role of plasmapheresis in drug-induced thrombotic microangiopathies is less certain, and removal of the offending agent is the most important initial therapeutic maneuver. ^{, ,} Postdiarrheal HUS in children is usually managed conservatively as evidence suggests that early antibiotic therapy may actually promote the development of HUS. Treatment with eculizumab, a humanized monoclonal antibody that prevents cleavage of complement component C5 into C5a and C5b, inhibiting terminal complement activation, may be considered in patients with nondiarrheal (complement mediated) HUS unresponsive to plasma exchange. Hypertension and AKI associated with scleroderma may be exquisitely sensitive to treatment with ACE inhibitors. Management of vasculitis is discussed in detail in Chapter 31.

Early studies suggested that plasmapheresis may be of benefit in AKI due to myeloma cast nephropathy. ^{, ,} Clearance of circulating light chains with concomitant chemotherapy to decrease the rate of production had been postulated to reverse renal injury in patients with circulating light chains, heavy Bence Jones proteinuria, and AKI. A subsequent randomized controlled trial compared plasma exchange and standard chemotherapy with chemotherapy alone. Although the study did not demonstrate improvement with plasma exchange with regard to a composite outcome of death, dialysis dependence, or GFR <30 mL/min at 6 months, the study was inadequately powered to definitively exclude a clinical benefit and there was a trend toward improved outcomes with plasmapheresis. More recently, it has been suggested that the use of dialysis membranes that are permeable to light chains and other proteins with molecular weights lower than albumin (high cutoff membranes) may be an effective therapeutic strategy in patients with AKI due to light chain cast nephropathy; however, data from clinical trials have yielded inconclusive results. Thus primary management should focus on prompt initiation of highly effective chemotherapy. Other contributors to AKI in multiple myeloma, such as hypercalcemia, should also be promptly treated. Multiple myeloma is discussed in more detail in Chapter 35 .

Most cases of acute interstitial nephritis (AIN) are due to an allergic response to a medication. The initial therapeutic step in AIN is discontinuation of the offending medication or treatment of the probable inciting factor if not drug induced. Data on the efficacy of corticosteroids derive from small observational studies, which have yielded highly discordant results. While some studies suggest that early use of corticosteroids (i.e., before significant renal damage and within 7–14 days of discontinuation of the offending medication) may be beneficial, other studies demonstrate no clear evidence of efficacy. There have been no large, prospective randomized clinical trials investigating the role of corticosteroids in the treatment of AIN. As corticosteroids are associated with a series of potentially serious side effects, their use should be considered on a case-by-case basis. If corticosteroid therapy is being considered and no patient-related contraindications exist, one potential regimen used in a recent study involves the IV administration of methylprednisolone (250–500 mg/day) for 3 to 4 days, followed by oral prednisone at a dose of 1 mg/kg/day tapered over 8 to 12 weeks. However, there are no data supporting the superiority of this specific approach over others. Mycophenolate mofetil has also been investigated as a therapeutic agent for AIN. In a study of eight patients with AIN, six experienced improvement and two experienced stabilization in kidney function with mycophenolate mofetil therapy. While this small case series suggests a possible role for mycophenolate mofetil in the treatment of AIN, additional data are needed to confirm its safety and efficacy for this indication. Tubulointerstitial diseases are discussed further in Chapter 37 .

After correction of intravascular volume deficits, salt and water intake should be adjusted to match ongoing losses (urinary, gastrointestinal, drainage sites, insensible losses). Extracellular volume overload can usually be managed by restriction of salt and water intake and judicious use of diuretics. High doses of loop diuretics (e.g., the equivalent of 200 mg of furosemide administered as an IV bolus infusion or 20 mg per hour as a continuous infusion) or combination therapy with both thiazide and loop diuretics may be required. If adequate diuresis cannot be attained, further use of diuretics should be discontinued to minimize the risk of complications such as ototoxicity. Fluid administration should be closely monitored to avoid progressive volume overload. Although there is a strong association between progressive fluid overload and mortality risk in patients with AKI, a causal relationship has not been definitively established because volume overload may also be a surrogate for other determinants of mortality such as hemodynamic instability and capillary leak. Fluid conservative management has, however, been demonstrated to result in improved outcomes in critically ill patients with respiratory failure. Ultrafiltration or dialysis may be required for volume management when conservative measures fail.

Hyponatremia associated with a fall in effective serum osmolality can usually be corrected by restriction of water intake. Conversely, hypernatremia is treated by administration of water, hypotonic saline solutions, or hypotonic dextrose-containing solutions (the latter are effectively hypotonic because dextrose is rapidly metabolized).

Mild hyperkalemia (<5.5 mmol/L) should be managed initially by restriction of dietary potassium intake and the discontinuation of potassium supplements and potassium-sparing diuretics. More severe degrees of hyperkalemia (5.5–6.5 mmol/L) can usually be controlled by combining these measures with the administration of exchange resins to enhance gastrointestinal potassium losses. While sodium polystyrene sulfonate has been widely used for decades, concerns have been raised regarding its safety, particularly when administered in 70% sorbitol, due to reports of bowel necrosis. ^, Newer exchange resins including sodium zirconium cyclosilicate and patiromer are also effective at decreasing serum potassium concentration, although patiromer is not labeled for acute management of hyperkalemia. Loop diuretics can also increase potassium excretion in diuretic-responsive patients. Emergency measures need to be employed in patients with more severe hyperkalemia and in patients with electrocardiographic manifestations of hyperkalemia. In patients with severe hyperkalemia with concomitant electrocardiographic manifestations, the IV administration of calcium will antagonize the cardiac and neuromuscular effects of hyperkalemia and is a valuable emergency temporizing measure, allowing time for the additional measures described later to be implemented. IV calcium must be used with caution, however, if there is concomitant severe hyperphosphatemia or evidence of digitalis toxicity. IV insulin (10–20 U of regular insulin) promotes potassium entry into cells and lowers extracellular potassium concentration within 15 to 30 minutes, with an effect that lasts for several hours. ^, Concomitant administration of IV dextrose (25–50 g over 30 to 60 minutes) is required to prevent hypoglycemia in patients who do not have hyperglycemia. β-adrenergic agonists, such as inhaled albuterol (10–20 mg by nebulizer), also promote rapid potassium uptake into the intracellular compartment. Although sodium bicarbonate also stimulates potassium uptake into the intracellular compartment, this effect is not sufficiently rapid to be clinically useful for the emergent management of hyperkalemia. Emergent dialysis is indicated if hyperkalemia is resistant to these measures.

Treatment of metabolic acidosis depends on the clinical setting and etiology. As a general rule, metabolic acidosis does not require emergent treatment unless the serum HCO ₃ ^– concentration falls below 15 mmol/L or the pH is lower than 7.15 to 7.20. In patients with AKI in whom metabolic acidosis is due to the underlying renal failure, more severe acidosis can be corrected by either oral or IV bicarbonate administration. Initial rates of replacement should be based on estimates of HCO ₃ ^– deficit and adjusted thereafter according to serum levels. In patients with underlying lactic acidosis, the role of bicarbonate therapy is controversial and the primary focus of therapy should be on correction of the underlying cause. Patients treated with IV bicarbonate need to be monitored for complications of therapy including metabolic alkalosis, hypocalcemia, hypokalemia, hypernatremia, and volume overload.

Hypocalcemia does not usually require treatment unless it is severe or symptomatic, as may occur in patients with rhabdomyolysis or pancreatitis, or after administration of bicarbonate. Hyperphosphatemia can often be controlled by restricting dietary phosphate intake and the use of oral phosphate binders (e.g., aluminum hydroxide, calcium salts, sevelamer carbonate, and lanthanum carbonate). Caution should be employed when using aluminum-containing phosphate binders as prolonged use may result in aluminum intoxication, which can contribute to osteomalacia; short-term use is rarely associated with bone disease, and the feared neurologic complications of aluminum intoxication are restricted to patients with inadvertent parenteral exposure. Hypermagnesemia can be prevented through avoidance of magnesium-containing medications, such as antacids, and limiting magnesium content of parenteral nutrition. Hyperuricemia is usually mild in AKI (<15 mg/dL) and does not require specific intervention. Severe hyperuricemia secondary to cell lysis may be managed by blocking xanthine oxidase with allopurinol or by enhancing degradation with recombinant uricase as previously described.

Patients with AKI are clinically heterogeneous and individualized nutritional management is required, especially in critically ill patients on KRT in whom protein catabolic rates can exceed 1.5 g/kg body weight/day. ^{, , , , , ,} The objective of nutritional management in AKI is to provide sufficient calories to preserve lean body mass, avoid starvation ketoacidosis, and promote healing and tissue repair while minimizing production of nitrogenous waste. If the duration of impaired kidney function is likely to be short and the patient is not extremely catabolic and does not require KRT, then dietary protein should be approximately 0.8 to 1.0 g/kg body weight/day. Protein intake should not be restricted in patients in whom AKI is likely to be prolonged, are hypercatabolic, or are receiving KRT. Protein intake in these patients should generally be 1.0 to 1.5 g/kg body weight/day. ^{, , ,} There is no evidence of improved outcomes with protein intake higher than 1.7 g/kg body weight/day, even in extremely hypercatabolic patients with a meta-analysis of data from clinical trials in critically ill patients demonstrating no overall benefit to higher protein delivery with increased mortality in the subgroup of individuals with AKI (risk ratio 1.4; 95% CI: 1.14–1.8). ^, Total caloric intake should generally be 20 to 30 kcal/kg body weight/day and should not exceed 35 kcal/kg per day. ^{, , , ,} Benefits of vigorous parenteral hyperalimentation have not been consistently demonstrated; enteral nutrition support is preferred as it avoids the morbidity associated with parenteral nutrition while providing support to intestinal function. Water-soluble vitamins and trace elements should be supplemented in patients receiving KRT. ^,

Severe anemia is generally managed with blood transfusion. Transfusion is usually not required for patients with a hemoglobin concentration above 7 g/dL. Whether there is a role for erythropoiesis stimulating agents in AKI has not been definitively determined. Patients with AKI or other acute illness are relatively resistant to the effects of these agents, and their onset of action is delayed. In randomized controlled trials in critically ill patients, recombinant human erythropoietin decreased transfusion requirement but had no effect on other outcomes. ^, Uremic bleeding usually responds to desmopressin, correction of anemia, estrogens, or dialysis.

Doses of drugs that are excreted by the kidney must be adjusted for impaired kidney function and the use of KRT. Whenever possible, pharmacokinetic monitoring should be employed to ensure appropriate drug dosing, especially for agents with narrow therapeutic windows (see Chapter 56). In addition to careful monitoring for toxicity of agents that are normally excreted by the kidney, careful attention must be paid to dosing of antibiotics and other drugs removed by KRT to assure that therapeutic drug levels are achieved, particularly in patients receiving augmented intensity of KRT. Drug dosing in kidney disease is further discussed in Chapter 56.

KRT is the generic term for the multiple modalities of dialysis and hemofiltration employed in the management of kidney failure. Although kidney transplantation is also a form of KRT for end-stage renal disease, transplantation does not play a role in the management of AKI. KRT facilitates the management of patients with AKI, allowing correction of acid-base and electrolyte disturbances, amelioration of volume overload, and removal of byproducts of nitrogen metabolism (so-called “uremic solutes”). Although KRT can forestall or reverse the life-threatening complications of uremia associated with severe and prolonged AKI, it does not hasten and can potentially delay the recovery of kidney function in patients with AKI, and it can be associated with potentially life-threatening complications. Despite more than 60 years of research and clinical experience, ^, numerous questions regarding the optimal application of KRT in AKI remain. ^,

In clinical practice there are wide variations in the timing of initiation of KRT for patients with AKI. Widely accepted indications for initiation of KRT include volume overload unresponsive to diuretic therapy, severe metabolic acidosis or hyperkalemia despite appropriate medical therapy, and overt uremic manifestations including encephalopathy, pericarditis, or uremic bleeding diathesis ( Table 28.17 ). However, even these specific indications are subject to substantial clinical interpretation, and in many patients, KRT is initiated in the absence of these specific indications in response to a clinical course marked by progressive azotemia or sustained oliguria. The correlation between the BUN concentration and onset of uremic symptoms is relatively weak, although the longer the duration and the greater the severity of azotemia, the more likely it is that overt symptoms will develop. Observational series and small clinical trials dating from the 1950s through the 1980s suggested that initiating KRT when the BUN concentration approached 90 to 100 mg/dL was associated with improved survival as compared with more delayed initiation of therapy. More recent observational studies have suggested that initiation of KRT at even less severe degrees of azotemia may further improve survival. These studies need to be interpreted with caution, as the outcomes associated with earlier initiation of KRT may reflect differences related to the reasons for initiation of therapy (e.g., volume overload or hyperkalemia vs. progressive azotemia) rather than a benefit due to the earlier therapy per se. In addition, these observational series only included patients in whom KRT was actually initiated rather than the broader population of patients with AKI, including patients who either recovered kidney function or died without receiving KRT.

Numerous prospective clinical trials have evaluated the timing of initiation of KRT in AKI. In a small, randomized, controlled trial of critically ill patients randomized to early, high-volume hemofiltration, early low-volume hemofiltration, or late low-volume hemofiltration, there was no benefit associated with earlier initiation of treatment. In a subsequent trial comparing earlier to later initiation of dialysis in patients with community-acquired AKI, mortality was lower in patients initiated on dialysis later than in the group of patients started earlier with no difference in recovery of kidney function between groups. This latter trial needs to be interpreted with caution, however, as almost half of the patients admitted with community-acquired AKI were excluded due to urgent need for dialysis.

Four larger, randomized clinical trials have evaluated the timing of KRT among critically ill patients with AKI ( Table 28.18 ) . The Effect of Early versus Delayed Initiation of Renal Replacement Therapy on Mortality in Critically Ill Patients with Acute Kidney Injury (ELAIN) trial randomized 231 patients with AKI (based on KDIGO stage 2 and plasma NGAL level >150 ng/mL) at a single surgical ICU to early or delayed initiation of KRT and demonstrated reduced 90-day mortality (hazard ratio 0.66; 95% CI: 0.45–0.97) with earlier initiation of KRT. Patients in the early KRT group also demonstrated decreased duration of KRT (9 days vs. 25 days, P = 0.04) and reduced hospital length of stay (51 days vs. 82 days, P < 0.001). However, the separation in timing of initiation of KRT between early and late groups was less than 1 day, raising questions about the therapeutic mechanisms underlying the observed marked reductions in mortality, duration of KRT, and hospital length of stay and suggesting that unrecognized differences between the two treatment groups may have contributed to the surprisingly large effect size.

The Artificial Kidney Initiation in Kidney Injury (AKIKI) trial was a multicenter trial that randomized 620 patients with KDIGO stage 3 AKI who required mechanical ventilation and/or catecholamine support to early or delayed initiation of KRT. KRT was initiated as soon as possible in patients randomized to the early arm but was withheld unless hyperkalemia, metabolic acidosis, or volume overload resulting in pulmonary edema were present or oliguria persisted for more than 72 hours or the BUN exceeded 112 mg/dL. There was no difference in 60-day mortality between the groups (48.5% in early strategy group vs. 49.7% in delayed strategy group, P = 0.79). Notably, nearly half of patients assigned to the delayed strategy group never required initiation of KRT (see Fig. 28.4 ).

Timing of initiation of kidney replacement therapy (KRT) and survival in the Acute Kidney Initiation in Kidney Injury (AKIKI) Trial comparing early and delayed strategies for initiation of KRT.

Timing of initiation of KRT (A) and Kaplan-Meier probability of survival (B) in the Acute Kidney Initiation in Kidney Injury (AKIKI) trial. Most (98%) of the patients in the early-treatment group initiated KRT at a median of 4.3 hours after reaching stage 3 AKI as compared with 51% of patients who initiated KRT at a median of 57 hours. Sixty-day mortality was 48.5% in the early-treatment group versus 49.7% in the delayed-treatment group (HR: 1.03; 95% CI: 0.81–1.29; P = 0.84).

(From Gaudry S, et al. Initiation strategies for renal-replacement therapy in the intensive care unit. N Engl J Med . 2016;375:122–133.

The Initiation of Dialysis Early versus Delayed in the Intensive Care Unit (IDEAL-ICU) trial also failed to demonstrate a benefit to earlier initiation of KRT in 488 patients with sepsis-associated AKI. In the IDEAL-ICU trial, patients were enrolled if they met the criteria for RIFLE-F, AKI, had sepsis, and did not have an emergent indication for KRT. In the early-treatment arm, patients initiated KRT within 12 hours of eligibility while in the delayed-treatment arm, KRT was initiated if a specific indication for KRT developed or if there was no recovery of kidney function after 48 hours. In the early-treatment arm, mortality at 90 days was 58% as compared with 54% in the delayed-treatment arm ( P = 0.38). Ninety-seven percent of patients in the early-treatment arm received KRT as compared with only 62% of patients in the delayed-treatment arm. Of those in the delayed arm who did not receive KRT, 75% had spontaneous recovery of kidney function while 23% died before specified criteria for initiation of KRT.

The Standard versus Accelerated Initiation of Renal Replacement Therapy in Acute Kidney Injury (STARRT-AKI) trial was a multinational trial that enrolled 3019 critically ill patients with KDIGO stage 2 or 3 AKI at 168 hospitals in 15 countries. Patients were eligible if they did not have advanced CKD, a treatable form of AKI, or an urgent indication for KRT, and their treating physicians had equipoise regarding the immediate need to start or defer initiation of KRT. Patients randomized to early initiation had KRT started as soon as possible and within 12 hours of meeting full eligibility while patients randomized to the standard (delayed) initiation arm had KRT withheld unless specific criteria were present (serum potassium ≥6.0 mmol/L, pH ≤7.20, serum bicarbonate ≤12 mmol/L, or the presence of respiratory failure due to volume overload) or persistent AKI for 72 hours after randomization. Ninety-seven percent of patients randomized to the early initiation arm received KRT with a median time from full eligibility to start of KRT of 6.1 hours (IQR: 3.9–8.8 hours) as compared with 62% of individuals randomized to the standard (delayed) arm, with a median time to initiation of KRT of 31.1 hours (IQR 19.0–17.8 hours). All-cause mortality at 90 days was not different between the two treatment arms (43.9% in the early arm vs. 43.7% in the standard arm) ( Fig. 28.5 ); however, dialysis dependence at day 90 among surviving patients was higher in the early initiation arm (10.4% vs. 6.0%; HR 1.74, 95% CI: 1.24–2.43).

Survival with standard (delayed) versus accelerated initiation of kidney replacement therapy (KRT) in the Standard versus Accelerated Initiation of Renal Replacement Therapy in the Acute Kidney Injury (STARRT-AKI) Trial.

Kaplan-Meier plot of survival probability among 2927 evaluable patients in the STARRT-AKI Trial, which compared strategies of early and delayed KRT in critically ill patients with AKI. Death at 90 days occurred in 643 patients (43.9%) in the accelerated KRT arm as compared with 639 patients (43.7%) in the standard (delayed) KRT arm ( P = 0.92). RRT, Renal replacement therapy.

From STARRT-AKI Investigators, Australian and New Zealand Intensive Care Society Clinical Trials Group, United Kingdom Critical Care Research Group, et al. Timing of initiation of renal-replacement therapy in acute kidney injury. N Engl J Med . 2020;383:240–251.

A meta-analysis that pooled data from these four large trials and several smaller studies found no benefit associated with earlier initiation of KRT, with a log odds ratio of–0.04 (95% CI:–0.16–0.07). There was also a trend to greater dialysis dependence associated with earlier initiation of KRT, although this did not achieve statistical significance (log odds ratio 0.24; 95% CI:–0.03–0.51).

In a follow-up to their initial trial, the AKIKI investigators evaluated a strategy of even later initiation of KRT. Two-hundred seventy-eight critically ill patients with AKI who had oliguria for more than 72 hours or a BUN of more than 112 mg/dL without other indications for initiation of KRT were randomized to begin KRT immediately or wait until objective criteria (hyperkalemia, acidosis, or pulmonary edema) were present or the BUN exceeded 140 mg/dL. Greater delay in initiation of KRT did not result in further reduction in the number of KRT-free days but was associated with an increased hazard ratio for death (1.65; 95% CI: 1.09–2.50; P = 0.018) in a prespecified multivariable analysis. Thus the authors concluded that while early initiation of KRT was not beneficial, excessive delay was associated with an increased risk of preventable death.

Although volume overload unresponsive to diuretic therapy is a widely accepted indication for initiation of KRT, wide variations in the degree of volume overload at initiation of therapy exist. ^{, ,} Observational studies have demonstrated a strong association between the degree of volume overload and mortality risk, leading to the suggestion that KRT should be initiated early, before the development of progressive volume overload. ^, The association between volume overload and mortality risk does not, however, establish a causal relationship; disease processes that contribute to the development of volume overload may independently contribute to mortality risk in these patients. In a newer, single-center, prospective observational study of 820 critically ill patients treated with continuous KRT (CKRT), mortality was associated with greater positive fluid balance at time of CKRT initiation. However, after adjusting for covariates including age, SOFA score, and vasopressor dose, cumulative fluid balance at time of CKRT initiation was no longer associated with ICU or hospital mortality (OR 1.01 per 1000 mL fluid accumulation; 95% CI: 0.97–1.05). Similarly, in a secondary analysis of the STARRT-AKI trial, earlier initiation of KRT attenuated cumulative fluid balance; however, among patients with greater fluid accumulation at the time of randomization, earlier KRT initiation did not have an effect on all-cause mortality.

The KDOQI Clinical Practice Guidelines for Acute Kidney Injury, which predate the ELAIN, AKIKI, IDEAL-ICU, and STARRT-AKI trials, do not make strong recommendations on the timing of initiation of KRT. They suggest that KRT be “ … initiated emergently when life-threatening changes in fluid, electrolyte, and acid-base balance exist” and further suggest that “ … the broader clinical context, the presence of conditions that can be modified by KRT, and trends of laboratory tests—rather than single BUN and creatinine thresholds alone—[be considered] when making the decision to start KRT.” The accumulated data since publication of these guidelines demonstrate that early initiation of KRT before the presence of objective clinical indications is not associated with improved clinical outcomes; however, undue delay may be associated with increased mortality.

KRT should be discontinued when kidney function recovers or because continued provision of dialytic support is no longer consistent with the patient’s overall goals of care. Recovery of kidney function is usually heralded by increased urine volume. Although no specific threshold of urine output correlates with sufficient recovery of kidney function, it is unlikely that a urine output of less than roughly 1 L/ day is sufficient to sustain dialysis independence. Although diuretics may increase daily urine volume, there is no evidence that diuretic therapy promotes recovery of kidney function. Improved solute clearance is manifested by spontaneous fall in blood urea and creatinine concentrations or a persistent downward trend in predialysis values. The role of creatinine clearance measurement to assess recovery of kidney function is uncertain, with a paucity of data to define specific thresholds for recovery of kidney function. In the Acute Renal Failure Trial Network (ATN) Study, KRT was continued if measured creatinine clearance on a 6-hour timed urine collection was <12 mL/min; KRT was stopped if the clearance was >20 mL/min; and the decision was left to the discretion of the clinician if the creatinine clearance was between 12 and 20 mL/min. ^, While specific criteria for discontinuation of KRT cannot be provided, patients should be carefully monitored for evidence of recovery.

Multiple modalities of KRT are available for the management of patients with AKI including conventional intermittent hemodialysis (IHD), peritoneal dialysis (PD), multiple forms of continuous KRT (CKRT), and prolonged intermittent kidney replacement therapies (PIKRTs), such as sustained low-efficiency dialysis (SLED; also known as extended duration dialysis). Detailed descriptions of the technical aspects of these modalities are provided in Chapters 62, 63, and 64 (CKRT and PIKRT). Objective data to guide the selection of modality for individual patients are limited, and the choice of modality is often guided by the resources of the health care institution and technical expertise of the physicians and nursing staff. The KDIGO Clinical Practice Guideline for Acute Kidney Injury suggests that for the majority of patients, the available modalities of KRT are complementary, with the caveats that CKRT and PIKRT be used in hemodynamically unstable patients and CKRT be used for patients with acute brain injury or other causes of increased intracranial pressure or generalized brain edema.

Intermittent hemodialysis

Intermittent hemodialysis has been the mainstay of KRT in AKI for 7 decades. Patients typically undergo dialysis treatments for 3 to 5 hours on a daily or thrice-weekly, alternate-day schedule depending on catabolic demands, electrolyte disturbances, and volume status. Just as with the timing of initiation of dialysis in AKI, the most appropriate dosing strategy for IHD in patients with AKI has been the subject of considerable investigation. The dose of IHD may be adjusted by altering the intensity of each individual dialysis session, usually quantified as the product of urea clearance and dialysis duration normalized to volume of distribution of urea (Kt/V _urea ), or by changing the frequency of the dialysis sessions. In an observational study, Paganini and colleagues demonstrated a survival benefit in patients with intermediate severity of illness scores when the delivered Kt/V _urea was >1.0 per treatment as compared with a delivered Kt/V _urea <1.0 per treatment. However, there have been no prospective clinical trials evaluating the relationship between the delivered Kt/V _urea and outcomes when dialysis is provided on a constant treatment schedule. Schiffl and colleagues reported on a prospective trial of 160 patients with AKI assigned in an alternating fashion to alternate-day or daily intermittent hemodialysis. The more frequent treatment schedule was associated with a reduction in mortality at 14 days after the last dialysis session from 46% in the alternate-day dialysis arm to 28% in the daily treatment arm ( P = 0.01). The duration of renal failure declined from 16 ± 6 days to 9 ± 2 days ( P = 0.001). This study has been criticized, however, because the delivered dose of therapy per session was low in both treatment arms (Kt/V _urea <0.95), resulting in a high rate of symptoms in the alternate-day dialysis arm that may have been related to overtly inadequate dialysis. The impact of frequency of IHD was also evaluated in the ATN study. In the ATN study, 1124 critically ill patients were randomized to an intensive or less-intensive strategy for the management of KRT. When patients were hemodynamically stable, they received IHD, and when hemodynamically unstable they received CKRT or SLED, regardless of treatment arm. Patients randomized to the less-intensive treatment strategy received IHD on a thrice-weekly (alternate-day except Sunday) schedule while patients randomized to the intensive arm received six-times IHD per week (daily except Sunday). The 60-day all-cause mortality was 53.6% in the intensive treatment arm compared with 51.5% in the less-intensive arm ( P = 0.47) ( Fig. 28.6 ). The mean delivered Kt/V _urea was 1.3 per treatment after the first IHD session. Although the study was not designed to evaluate outcomes by individual modality of KRT, there were no differences in mortality between groups when evaluated on the basis of percentage of time treated using IHD. According to these results, it does not appear that there is further benefit to routinely increasing the frequency of IHD treatments beyond three times per week as long as the delivered Kt/V _urea is at least 1.2 per treatment. More frequent treatments may be necessary if the target dose per treatment cannot be achieved (e.g., in hypercatabolic patients, in patients with severe hyperkalemia or metabolic acidosis, and for issues related to volume management). The KDIGO Clinical Practice Guideline for Acute Kidney Injury recommends delivering a Kt/V _urea of 3.9 per week when using IHD in AKI, calculating the weekly Kt/V _urea as the arithmetic sum of the delivered dose per treatment. It should be recognized, however, that this approach for calculating an equivalent weekly Kt/V _urea is not consistent with urea kinetic principles and that rigorous data for the appropriate dose of therapy when treatments are delivered more frequently than three times per week are not available.

Sixty-day mortality with intensive versus less-intensive kidney replacement therapy in the Acute Renal Failure Trial Network (ATN) study.

Kaplan-Meier plot of mortality in 1124 critically ill patients with acute kidney injury randomized to a strategy of more intensive kidney replacement therapy (6×-per week intermittent hemodialysis or continuous venovenous hemodiafiltration at 35 mL/kg per hour) versus less-intensive kidney replacement therapy (3×-per week intermittent hemodialysis or continuous venovenous hemodiafiltration at 20 mL/kg per hour). At 60 days, mortality was 53.6% in the more intensive arm versus 51.5% in the less-intensive arm (OR: 1.09; 95% CI: 0.86–1.40; P = 0.47).

From The VA/NIH Acute Renal Failure Trial Network. Intensity of renal support in critically ill patients with acute kidney injury. N Engl J Med . 2008; 359:7–20.

The selection of IHD dialyzer membrane may also impact clinical outcomes. Exposure to cellulosic membranes results in accentuated leukocyte and complement activation and delayed recovery of kidney function in experimental models of AKI as compared with exposure to more biocompatible synthetic membranes. ^, Clinical trials comparing dialysis membranes have yielded conflicting results. Although some studies demonstrated delayed recovery of kidney function with cellulosic membranes, other studies observed no difference between cellulosic and other synthetic membranes thought to be more biocompatible. When these data have been aggregated in systematic reviews, a benefit of the synthetic membranes is not convincingly demonstrated. ^, While the effect of membrane type on humoral and cellular activation may still influence recovery of kidney function in AKI, the clinical importance of this issue has diminished as the cost differential between synthetic and cellulosic membranes has narrowed and the use of unsubstituted cellulosic membranes has decreased.

The major complications associated with IHD are related to the need to access the vasculature, the need for anticoagulation to maintain patency of the extracorporeal circuit, and intradialytic hypotension primarily resulting from shifts in solute and volume. ^{, ,} Many of these issues, particularly the need for vascular access and anticoagulation, are similar for CKRT and SLED.

Vascular access is usually obtained through insertion of a double-lumen catheter into a large-caliber central (internal jugular or subclavian) or femoral vein. The major complications associated with vascular access include vascular and organ trauma during insertion, bleeding, catheter malfunction and thrombosis, and infection. Although femoral catheters are generally associated with an increased risk of infection compared with catheters in the subclavian or internal jugular veins, an increased risk of infection was observed only when femoral-vein catheters were used in patients with a high BMI in a randomized controlled trial involving patients undergoing acute KRT. Prompt transition to tunneled hemodialysis catheters (or placement of tunneled hemodialysis catheters in advance of initiation) has been proposed as a means of decreasing the risk of infection in patients undergoing acute dialysis. ^, However, this strategy has not been rigorously evaluated in prospective clinical trials.

Anticoagulation is used to help maintain patency of the extracorporeal dialysis circuit in IHD, as well as CKRT and SLED. ^, The most commonly used anticoagulant for dialysis is unfractionated heparin, with multiple protocols used to attain sufficient anticoagulation of the dialysis circuit while minimizing systemic effects. ^, Regional heparinization, in which heparin is infused proximal to the dialyzer and protamine is infused into the return line to reverse heparin’s effect, can be used but has generally been supplanted by low-dose heparin protocols. Low-molecular-weight heparins (LMWH) may be used as an alternative to unfractionated heparin, but the benefits of this approach are unclear as LMWH is not associated with enhanced efficacy, drug half-life is variably prolonged with impaired kidney function, and monitoring of the anticoagulant effect is more difficult. In patients with heparin-induced thrombocytopenia (HIT), heparin administration is contraindicated. Alternative anticoagulation strategies include regional citrate ^, ; the serine protease inhibitor nafamostat ; the direct thrombin inhibitors hirudin, leprudin, and argatroban ; and, rarely, the prostenoids epoprostenol and iloprost. ^, In many patients, particularly those with underlying coagulopathy or thrombocytopenia, and in patients with active hemorrhage or recent postoperative status, acute KRT can be provided in the absence of anticoagulation. ^{, ,}

Intradialytic hypotension is common in patients undergoing acute IHD. ^{, , , ,} Episodes of hypotension may impair solute clearance and the efficiency of dialysis and can further compromise renal perfusion and delay recovery of kidney function. ^, Intradialytic hypotension is typically triggered by intercompartmental fluid shifts or excessive fluid removal, leading to decreased intravascular volume, and may be exacerbated by altered vascular responsiveness related to the underlying acute process. ^, Hypotension may be particularly problematic in critically ill patients in whom sepsis, cardiac dysfunction, hypoalbuminemia, malnutrition, or large third-space losses may accompany the development of AKI. Prevention of intradialytic hypotension requires careful assessment of intravascular volume; prescription of realistic ultrafiltration targets; extension of treatment time so as to minimize the ultrafiltration rate; increasing the dialysate sodium concentration; and decreasing the dialysate temperature. ^, It is noteworthy that intradialytic hypotension can develop even among patients in whom no ultrafiltration is prescribed; the reason(s) for hypotension are not entirely clear, although many have attributed hemodynamic instability to the rapid exchange of solutes induced by high flux/high-efficiency hemodialysis with extracellular to intracellular shifting of body water.

Although there is a tendency to reduce the extracorporeal blood flow in patients prone to hypotension, there is little evidence that this provides any benefit. Reducing blood flow decreased the volume of the extracorporeal circuit in the past when parallel plate and coil dialyzers were used; however, there is little change in the volume of the extracorporeal circuit in response to changes in blood flow when hollow fiber dialyzers are used. Reducing blood flow may, however, result in reduction of the delivered dose of dialysis.

Continuous kidney replacement therapy

Continuous kidney replacement therapies represent a spectrum of treatment modalities. In their initial description, the continuous therapies were provided using arteriovenous extracorporeal circuits. While this approach provided technical simplicity, blood flow was dependent on the gradient between mean arterial and central venous pressure and there was an increased risk of complications from prolonged arterial cannulation. As a result, the continuous arteriovenous therapies have largely been supplanted by pump-driven, venovenous CKRT. The modalities of venovenous CKRT vary predominantly based on their mechanism of solute removal. With continuous venovenous hemofiltration (CVVH), solute transport occurs by convection; with continuous venovenous hemodialysis (CVVHD) by diffusion; and with continuous venovenous hemodiafiltration (CVVHDF) by a combination of the two. ^, Although, at the same level of urea clearance, convective therapies provide enhanced clearance of higher molecular weight solutes compared with diffusive therapies, no clear clinical benefit has been demonstrated for CVVH or CVVHDF as compared with CVVHD.

The clearance of urea and other small solutes during CKRT is generally proportional to the total effluent flow rate (the sum of ultrafiltrate and dialysate flow rates), ^{, ,} and dose of therapy is usually expressed as the effluent volume indexed to body weight. This approach to estimating solute clearance is based on the assumption of near-complete solute equilibration between blood and effluent and may overestimate the actual solute clearance. ^, Several single-center randomized controlled trials demonstrated an improvement in survival when doses of CVVH were increased from 20 to 25 mL/kg per hour to doses in excess of 35 to 45 mL/kg per hour. ^, However, other small studies did not find a similar benefit. ^, Two large multicenter, randomized controlled trials also did not find a survival benefit associated with more intensive CKRT. ^, In the previously described ATN study, 1124 patients were randomized to two intensities of KRT. In both treatment arms, patients received IHD when hemodynamically stable and CVVHDF or SLED when hemodynamically unstable. In the less-intensive arm, CVVHDF was provided at an effluent flow rate of 20 mL/kg per hour, and in the more intensive arm at 35 mL/kg per hour. Sixty-day all-cause mortality was 51.5% in the less-intensive arm and 53.6% in the more intensive arm ( P = 0.47) ( Fig. 28.6 ). In the Randomized Evaluation of Normal versus Augmented Level (RENAL) Replacement Therapy study, 1508 patients were randomized to CVVHDF at either 25 mL/kg per hour or 40 mL/kg per hour. Ninety-day all-cause mortality was 44.7% in both treatment arms ( P = 0.99) ( Fig. 28.7 ). On the basis of these data, the KDIGO Clinical Practice Guideline for Acute Kidney Injury recommends delivering an effluent volume during CKRT of 20 to 25 mL/kg per hour, recognizing that a slightly higher dose may need to be prescribed in order to achieve the target delivered dose in order to compensate for interruptions in treatment. Although both the ATN and RENAL studies demonstrated an absence of benefit with more intensive dosing of KRT, neither were designed to establish the minimal adequate dose of CKRT. Observational data from Japan have demonstrated that mortality does not increase until the delivered dose is <13 to 5 mL/kg per hour, suggesting that the KDIGO target of 20 to 25 mL/kg per hour may be higher than necessary.

Ninety-day mortality with intensive versus less-intensive continuous venovenous hemodiafiltration (CVVHDF) in the Randomized Evaluation of Normal versus Augmented Level (RENAL) Replacement Therapy Study Kaplan-Meier plot of mortality in 1508 critically ill patients with acute kidney injury randomized to CVVHDF at 35 mL/kg per hour versus 20 mL/kg per hour).

At 90 days, mortality was 44.7% in both treatment groups (OR: 1.00; 95% CI: 0.81–1.23; P = 0.99).

From The RENAL Replacement Therapy Study Investigators. Intensity of continuous renal-replacement therapy in critically ill patients. N Engl J Med. 2009;361:1627–1638.

Given the improved hemodynamic tolerance of CKRT as compared with IHD, particularly in patients with underlying hemodynamic instability, it has been postulated that CKRT would be associated with improved clinical outcomes. Five randomized controlled trials comparing outcomes with CKRT and IHD have been published. In a multicenter, randomized controlled trial of 166 patients with AKI, Mehta and colleagues observed ICU and hospital mortality rates of 59.5% and 65.5%, respectively, in patients randomized to CKRT as compared with 41.5% and 47.6%, respectively, in patients randomized to IHD ( P < 0.02). As the result of an imbalance in randomization, patients in the CKRT arm had higher severity of illness as measured by APACHE III score and a higher rate of liver failure. Adjusting for the imbalanced randomization in a post hoc analysis, the investigators found no difference in mortality attributable to modality of KRT. In another single-center randomized trial ( n = 80), Augustine and colleagues reported more effective fluid removal and greater hemodynamic stability associated with CVVHD as compared with IHD but no difference in survival. Similarly, in another single-center RCT from Switzerland, Uehlinger and colleagues observed no difference in survival in 70 patients randomized to continuous venovenous hemodiafiltration (CVVHDF) as compared with 55 patients assigned to IHD. In the Hemodiafe study, a multicenter RCT conducted in 21 ICUs in France, Vinsonneau and colleagues reported 60-day survival rates of 31.5% in 184 patients randomized to IHD as compared with 32.6% in 175 patients randomized to CVVHDF ( P = 0.98) ( Fig. 28.8 ). Similarly, Lins and colleagues observed hospital mortality rates of 62.5% in 144 patients randomized to IHD and 58.1% in 172 patients randomized to CKRT ( P = 0.43). Multiple meta-analyses have concluded that there is no difference in survival among patients by KRT. Although several studies have suggested that CKRT is associated with improved rates of recovery of kidney function in surviving patients as compared with IHD, ^, all of these studies are confounded by higher mortality rates in the CKRT group. When analyzed across studies in which there were no differences in mortality, rates of recovery of kidney function do not appear to be affected by modality of KRT. ^{, , ,}

60-Day Survival with Intermittent Hemodialysis (IHD) versus Continuous Venovenous Hemodiafiltration (CVVHDF) in the Hemodiafe Study.

Kaplan-Meier plot of survival among 359 critically ill patients with acute kidney injury randomized to intermittent hemodialysis versus continuous venovenous hemodiafiltration. At 60 days survival was 31.5% among patients randomized to IHD versus 32.6% among patients randomized to CVVHDF ( P = 0.98).

From Vinsonneau C, Camus C, Combes A, et al. Continuous venovenous haemodiafiltration versus intermittent haemodialysis for acute renal failure in patients with multiple-organ dysfunction syndrome: a multicentre randomised trial. Lancet 2006;368:379–385.