Cardiac surgery–associated AKI (CSA-AKI) is the second most common cause of AKI in the ICU and is associated with increased mortality. ^, CSA-AKI is defined as AKI occurring 2 to 7 days after surgery and is reported, by KDIGO criteria, to complicate 20% to 30% of cardiac surgeries. The pathophysiology of CSA-AKI is multifactorial including hypoperfusion, inflammation, oxidative stress, atheroembolism, hemolysis, nephrotoxin exposure, and exposure to cardiopulmonary bypass (CPB). ^, The duration of CPB is a significant predictor of AKI, but extensive randomized controlled trials (RCTs) aimed at avoiding or reducing CPB have yielded mixed results. Similarly, RCTs of therapeutics to prevent or treat CSA-AKI have failed to result in benefits. Single-center RCTs ^, have found that perioperative nitric oxide (NO) delivered via the CBP circuit or ventilator decreases the rate of CSA-AKI or chronic kidney disease (CKD), presumably via correction of hemolysis-induced depletion of renal NO. Though promising, the use of NO to prevent CSA-AKI remains investigational and additional RCTs are ongoing (NCT04216927, NCT05757557). ^, The use of biomarkers to prevent CSA-AKI also holds promise. The multicenter PrevAKI study randomized 278 postoperative patients at high risk of CSA-AKI, identified on the basis of elevated urinary tissue inhibitor of metalloproteinases 2 (TIMP-2) ∗ insulin-like growth factor binding protein 7 (IGFBP7), to standard care or the implementation of a KDIGO-based AKI bundle, which includes minimization of nephrotoxic medications, prevention of hyperglycemia, and volume optimization via hemodynamic monitoring. Though the need for KRT and mortality were similar in both groups, the rate of stage 2-3 AKI was significantly reduced.

The use of extracorporeal membrane oxygenation (ECMO) in adults grew exponentially after the 2009 H1N1 influenza outbreak, with overall ECMO usage in adults more than tripling between 2008 and 2014. Notably, 2023 guidelines recommend the use of venovenous (VV-) ECMO for refractory ARDS.

Up to 85% of adults on ECMO develop AKI, with roughly 45% requiring KRT. As in other settings, AKI in patients requiring ECMO is independently associated with mortality. AKI develops more commonly in patients receiving venoarterial- (VA-) ECMO for cardiac failure than those on VV-ECMO for isolated respiratory failure. Observational data suggest that timely KRT may be beneficial in preventing or mitigating volume overload in ECMO patients. ^, Fluid overload in ECMO patients is independently associated with increased mortality , and guidelines suggest targeting euvolemia with diuretics or KRT once feasible to facilitate ECMO weaning. However, experts recommend against preemptive KRT use in patients on ECMO, given a lack of data for benefit of KRT beyond the usual indications.

The pathogenesis of AKI on ECMO is complex. ECMO-specific factors that contribute to AKI include inflammation and disordered coagulation induced by the circuit, pigment injury from hemolysis, and nonpulsatile blood flow in VA-ECMO. Excess circuit-related hemolysis may manifest with pink urine or KRT effluent and can be confirmed by measuring plasma-free hemoglobin levels. ^,

Like ECMO, the use of left ventricular assist devices (LVADs) to treat end-stage heart failure has grown dramatically during the past 2 decades. ^, On average, LVAD placement produces an initial improvement in kidney function, likely reflecting improved renal perfusion in patients with type 2 CRS (i.e., kidney dysfunction due to chronic cardiac dysfunction), followed by a gradual decline back toward baseline. However, kidney function may not improve in those with intrinsic kidney disease. Like patients receiving ECMO, patients managed with LVAD are at elevated risk for AKI throughout their course, and AKI (particularly AKI requiring KRT) is associated with poor outcomes. LVAD-specific causes of AKI include hemolysis, right ventricular failure, and possibly nonpulsatile blood flow. ^,

Hepatorenal syndrome (HRS), particularly HRS-AKI, which was previously referred to as type 1 HRS, is a complication of decompensated cirrhosis associated with high mortality in which the vascular and neurohormonal disturbances of advanced liver disease and portal hypertension result in profound renal hypoperfusion unresponsive to intravenous (IV) albumin or other fluids. In modern cohorts, HRS-AKI accounts for 12% of AKI cases in the setting of cirrhosis, whereas prerenal AKI (44%) and ATN (30%) are significantly more common. Notably, the mortality of HRS and ATN in the setting of cirrhosis is similar at approximately 50% at 90 days. ^,

Developments in the diagnosis of HRS-AKI include adoption of the KDIGO criteria for AKI diagnosis. ^, In addition, though these urinary indices tend to perform poorly in ICU patients in differentiating prerenal AKI from ATN, cutoffs of fractional excretion of sodium (FENa) and urea (FeUrea) have been suggested to differentiate HRS from ATN or prerenal AKI (with FENa <0.2% and FeUrea <28.2% being consistent with HRS-AKI and <0.1% and <21%, respectively, being especially suggestive). Furthermore, though the absence of proteinuria >500 mg/g creatinine remains a traditional diagnostic criterion for HRS, urinary albumin has been shown in studies to be useful in differentiating ATN and HRS, with albuminuria >100 mg/g creatinine being suggestive of ATN. ^, Likewise, studies suggest that the tubular injury biomarker neutrophil gelatinase–associated lipocalin (NGAL) may help differentiate HRS and ATN, with a urinary NGAL level >220 to 250 μg/g creatinine being consistent with ATN. Despite these diagnostic advances, the diagnosis of HRS-AKI in clinical practice remains a tremendous challenge. Though the diagnostic criteria imply that HRS is a diagnosis of exclusion, HRS can coexist with ATN, CRS from cirrhotic cardiomyopathy, abdominal compartment syndrome from tense ascites, or other unrelated causes of AKI or CKD.

Terlipressin has been validated as an effective treatment for HRS-AKI in the Study to Confirm Efficacy and Safety of Terlipressin in Hepatorenal Syndrome (HRS) Type 1 (CONFIRM). Though CONFIRM demonstrated a significant increase in HRS reversal (29% vs. 16% in placebo arm), the terlipressin arm also experienced an increased risk of adverse effects including respiratory failure (10% with terlipressin vs. 3% in the placebo arm) and death due to respiratory failure (11% vs. 2%, respectively). This is thought to result from excessive albumin use and the regional hemodynamic effects of terlipressin including intense systemic vasoconstriction, pulmonary venoconstriction, lack of pulmonary arterial vasoconstriction, and absent inotropic effect. The positive primary outcome of the CONFIRM trial resulted in FDA approval of terlipressin in 2022 with a black box warning about the risk of fatal respiratory failure. Contraindications to terlipressin include any hypoxemia or need for supplemental oxygen, acute-on-chronic liver failure (ACLF) grade 3, and serum creatinine >5.0 mg/dL, as these patients are at increased risk of respiratory failure or treatment nonresponse. ^{, , ,} Norepinephrine is recommended as an alternative to terlipressin, as it appears to have similar efficacy in achieving HRS reversal, but it requires ICU admission in most centers. ^, Octreotide and midodrine are not recommended because they are rarely effective and may cause harm by delaying effective therapy. ^{, ,}

The results of CONFIRM and a second large RCT published in 2021 examining albumin use in hospitalized patients with cirrhosis highlight the risk of harmful volume overload from excessive albumin administration in patients with liver disease. Similarly, the association between intraabdominal hypertension (IAH) and AKI in cirrhosis, as well as the potential for improved kidney function through cautious therapeutic paracentesis, is gaining recognition. Therefore while albumin remains the initial treatment for most cirrhosis patients with AKI, it is important to first assess their volume status thoughtfully. Furthermore, given data that delay in implementation of effective vasoconstrictor therapy increases the risk of treatment failure, ^{, ,} those who felt likely to have HRS may benefit from initiation of vasoconstrictors before failing 48 hours of albumin therapy. ^,

Acute liver failure (ALF), previously known as fulminant hepatic failure, is defined by 1. hepatic encephalopathy, 2. coagulopathy (international normalized ratio [INR] ≥1.5), 3. onset of illness <26 weeks, and 4. no evidence of cirrhosis. ^, The distinction between ALF and chronic liver disease is vitally important because some of the complications and management options apply specifically to ALF. Common causes of ALF include toxic, viral, and ischemic liver injury, with acetaminophen toxicity alone accounting for about 50% in modern cohorts. Though some cases will spontaneously resolve, the mortality of ALF is high with supportive care alone. Complications of ALF overlap with decompensated cirrhosis and include hepatic encephalopathy AKI, coagulopathy and bleeding, hypoglycemia, infection and sepsis, and multiorgan failure. However, in contrast to cirrhosis, in which clinically significant cerebral edema is characteristically absent, ^, severe hepatic encephalopathy from ALF is frequently accompanied by cerebral edema with progressive intracranial hypertension and risk of lethal brainstem herniation. Cerebral edema is present in up to 35% of patients with ALF and grade 3 hepatic encephalopathy and 75% of patients with grade 4 hepatic encephalopathy. First-line treatment for confirmed or suspected intracranial hypertension is hyperosmolar therapy with mannitol, though it requires monitoring of serum osmolality and its use may be limited by renal dysfunction in those not on KRT. ^, Similarly, hypertonic saline to achieve serum sodium of 145 to 155 mEq/L is recommended in patients with or at high risk of cerebral edema. ^,

AKI in the setting of ALF may be due to renal hypoperfusion (e.g., sepsis, bleeding, and HRS) or intrinsic injury (e.g., ischemic ATN, direct tubular toxicity of acetaminophen or Amanita ) and is associated with increased risk of cerebral edema and overall poor prognosis. For a variety of reasons, patients with ALF may require KRT relatively early in the course of their AKI. Typically, continuous renal replacement therapy (CRRT) is preferred to intermittent hemodialysis (IHD) to minimize the risk of hypotension and increase in intracranial pressure (ICP), which can combine to severely reduce cerebral perfusion pressure. CRRT also offers more flexibility to manage volume status, clear ammonia, and precisely manage hyponatremia, which frequently accompanies ALF and may aggravate cerebral edema. ^, In contrast to chronic liver disease, KRT does appear to have a role in specifically treating hyperammonemia in ALF in the absence of AKI, particularly when ammonia is >250 μg/L, with higher-than-standard doses of CRRT typically required to effectively clear ammonia. ^, Beyond the risk of hypotension and increased ICP, a preference for CRRT over IHD in ALF is supported by limited observational and trial data. ^,

Though controversy persists given inherent limitations in observational data, vancomycin has been increasingly recognized as a contributor to AKI in critically ill patients. The nephrotoxicity of vancomycin was initially attributed to the impurities of early formulations, referred to as “Mississippi mud.” Though current purified formulations are clearly less toxic than prior preparations, increasing rates of methicillin-resistant Staphylococcus aureus (MRSA) infections and the concomitant increase in vancomycin use have led to increasingly frequent reports of vancomycin-associated AKI (VA-AKI) with modern formulations. In meta-analyses, the reported rates of VA-AKI ranged from 5% to 43% of vancomycin-treated patients with a relative risk of AKI with vancomycin of 2.45 compared with other antimicrobials (mostly linezolid). ^, VA-AKI has a mean onset of 4 to 17 days after drug initiation and approximately 75% of cases improve or resolve by hospital discharge. ^,

Vancomycin undergoes no metabolism and is entirely cleared unchanged by the kidneys. The mechanism of VA-AKI appears to be toxic acute tubular injury caused by oxidative injury and obstructive tubular casts. ^{, ,} The casts are composed of noncrystal nanospheric vancomycin aggregates embedded in uromodulin, as demonstrated by immunohistochemistry of human kidney biopsy and urinary sediment specimens, and reproduced in mouse models. ^, Risk factors include both patient factors—baseline kidney function, obesity, severity of illness, concomitant nephrotoxin exposure (e.g., aminoglycosides), and age—and drug factors like higher target trough levels (i.e., >15 mg/L, as used for severe MRSA infections); higher total daily dose (i.e., >4 g per day); intermittent dosing (rather than continuous infusions); and longer duration of therapy (i.e., >7 days). ^{, , ,}

An additional reported risk factor for VA-AKI is concomitant exposure to piperacillin-tazobactam (P-TZ), which alone is not generally considered nephrotoxic (except for rare cases of AIN). Though not found in all studies, a signal for increased rates of AKI with this combination compared with vancomycin alone has been repeatedly detected in the literature. However, with some exceptions, the majority of the studies tying the combination of P-TZ and vancomycin to AKI have failed to detect a difference in need for KRT or mortality, outcomes strongly linked to AKI. Moreover, the mechanism of this vancomycin-specific additive nephrotoxicity of P-TZ is unknown and multiple attempts to replicate this synergistic nephrotoxicity in animal models and in vitro studies have failed. One proposed explanation for these contradictory findings is that P-TZ, which is excreted via organic anion transporters, inhibits tubular secretion of creatinine, leading to increased rates of creatinine-defined pseudo-AKI without any associated tubulopathy or morbidity. Support for this theory comes from a 2022 biomarker study of nearly 200 ICU patients treated with vancomycin and P-TZ or vancomycin and cefepime. The P-TZ-vancomycin combination was associated with higher rates of creatinine-defined AKI, but there was no association with AKI as defined by cystatin C, need for KRT, or mortality. Subsequently, the 2023 Antibiotic Choice on Renal Outcomes (ACORN) study randomized more than 2500 adults hospitalized for acute infection to cefepime or P-TZ and, though 77% of the participants were also on vancomycin at enrollment, the rates of AKI or death were similar in both arms while the cefepime arm had increased rates of delirium or coma. Considering that kidney dysfunction is a major risk factor for cefepime neurotoxicity, these results suggest that cefepime should not be used preferentially to PT-Z in combination with vancomycin or in kidney dysfunction. However, as observational data suggest that synergistic nephrotoxicity is more likely with longer courses, ^, caution may be warranted with prolonged courses of PT-Z and vancomycin.

Emerging data suggest that IV iodinated contrast, as given with computed tomography (CT), is less nephrotoxic than previously thought. Although a few exceptions exist, ^, more than a dozen studies and meta-analyses analyzing thousands of patients across a variety of settings have been published since 2013, suggesting that modern low- or iso-osmolar IV contrast exposure has little to no nephrotoxicity. Many of these analyses focused on patients at higher risk of contrast-associated AKI (CA-AKI) due to underlying CKD. A 2023 multicenter analysis of >14,000 patient encounters, using weighting by propensity scores and entropy balancing to mitigate confounding and biases, found no increased risk of persistent kidney dysfunction or need for KRT in patients with established AKI before IV contrast exposure. While the risk of CA-AKI and relevant AKI outcomes after IV contrast administration is far lower than previously thought, the risk may not be zero, and measures to reduce the risk, such as minimization of contrast exposure and provision of IV fluids when appropriate, should still be pursued.

Importantly, nearly all of these studies questioning the nephrotoxicity of iodinated contrast analyzed IV rather than intraarterial contrast exposure. Both modern and historical data support the notion that intraarterial contrast exposure is more nephrotoxic, though many have speculated that this differential risk relates primarily to procedural aspects of angiography, such as the risk of subclinical atheroembolism, rather than enhanced toxicity related to the route of contrast administration. Regardless of the mechanism, the risk of AKI after intraarterial contrast exposure needs to be carefully considered in at-risk patients, and implementation science to mitigate the risk of AKI after cardiac catheterization has shown promise. ^,

Observational studies examining coronary angiography in advanced CKD patients with acute coronary syndromes reveal that although CKD patients face a higher risk of AKI or death compared to non-CKD patients, those with CKD who undergo angiography have lower mortality rates than CKD patients who do not receive indicated angiography. Though these observational studies are subject to residual confounding and biases, they suggest that withholding indicated diagnostic tests or interventions in patients with renal dysfunction due to excessive concern about AKI may inadvertently lead to worse outcomes, a concept termed “renalism.” Though data demonstrating that “renalism” exists in ICU populations are lacking, many experts ^{, , , ,} and 2020 consensus guidelines recommend performing studies using IV contrast regardless of kidney function when needed for definitive diagnosis or treatment of life-threatening conditions (e.g., sepsis or suspected pulmonary embolism).

Observational and animal data ^, suggest that 0.9% saline, when compared with balanced salt solutions such as lactated Ringer (LR) or Plasma-Lyte (PL, Baxter), may increase the risk of AKI, need for KRT, and mortality. The pathophysiology underlying the nephrotoxicity of saline is thought to be related to the supraphysiologic chloride content and the resultant hyperchloremic acidosis that saline induces. Proposed mechanisms include reduced glomerular filtration rate via activation of the tubuloglomerular feedback triggered by increased chloride delivery to the macula densa, vasoconstriction caused by chloride-induced thromboxane release, and increased inflammatory cytokine expression induced by acidosis. Imaging studies on human volunteers demonstrate that saline infusion reduces RBF.

Since 2010, a series of large clinical trials have analyzed crystalloid choice in the ICU, starting with a 2012 before-and-after study and a 2015 cluster crossover RCT, which yielded conflicting results. ^, In 2018, two large, single-center, pragmatic, unblinded, multiple-crossover cluster trials were published analyzing the use of LR or PL vs. saline, one in patients admitted from the ED to a non-ICU bed, the Saline against Lactated Ringer or Plasma-Lyte in the Emergency Department (SALT-ED) trial ( n >13,000), and another in ICU patients, the Isotonic Solutions and Major Adverse Renal Events Trial (SMART; n >15,000). Both studies demonstrated that balanced solutions resulted in an approximate 1% reduction in the composite endpoint of major adverse kidney events (a composite of death, need for KRT, or persistent doubling of creatinine above baseline) by 30 days or discharge.

These two single-center studies were followed in 2021 and 2022 by two large multicenter RCTs. First, the Balanced Solutions in Intensive Care Study (BaSICS) in Brazil randomized >11,000 individual patients in 75 ICUs to the blinded use of PL or saline whenever possible throughout the entire ICU course and found no difference in 90-day mortality. There was also no difference in subgroups with sepsis or AKI, but the risk of death was significantly increased with LR in the subgroup with traumatic brain injury (TBI). Though methodologically rigorous, BaSICS was criticized for being underpowered despite the large sample size due to lower-than-predicted mortality, relatively low acuity of the subjects, administration of nonstudy fluids before-and-after randomization, and relatively low total volumes of fluids administered, which resulted in statistically significant but modest differences in serum chloride levels (e.g., 2–3 mEq/L higher in the saline arm at 48 hours). ^, The subsequent Plasma-Lyte 148 versUs Saline (PLUS) study, carried out in 53 ICUs in Australia and New Zealand, was similar in design, with the major exceptions being that the study excluded TBI and was aborted after enrollment of 5037 of the planned 8800 patients due to the COVID-19 pandemic. Otherwise, though PLUS also had some contamination by nonstudy fluid, the severity of acute illness and volumes of study fluid administered were somewhat higher than in the SMART or BaSICS trials. Like BaSICS, the primary outcome of 90-day mortality was not significantly different in the PL and saline arms. Soon after the publication of PLUS, a Bayesian meta-analysis of 13 RCTs including >35,000 ICU patients (of which >30,000 come from SMART, BaSICS, and PLUS) was published and concluded that the risk ratio for 90-day mortality with balanced crystalloids versus saline was 0.96 (95% confidence interval [CI], 0.91–1.01; I ² =12.1%), with similar effects on risk of AKI and need for KRT. Large multicenter RCTs of PL versus saline in patients undergoing cadaveric renal transplantation ( n = 808) and in pediatric septic shock ( n = 708) have demonstrated benefits in kidney function with PL. ^,

Though somewhat more expensive, there are no data (outside of TBI) to suggest harm from balanced solutions, and, as such, the use of balanced solutions seems reasonable in most ICU settings. Alternatively, one can reasonably interpret the BaSICS and PLUS trials as evidence that saline is largely safe in the ICU. Importantly, saline will continue to be the fluid of choice in patients with TBI or hypovolemic hypochloremic metabolic alkalosis.

There are no compelling data to support the routine use of colloids over crystalloids as primary resuscitation fluid in ICU patients given a series of negative large RCTs and the increased expense of colloids. Two of these, the Saline versus Albumin Fluid Evaluation (SAFE) and Albumin Italian Outcome Sepsis (ALBIOS) trials, have shown albumin to be a safe alternative to crystalloids in ICU patients. Both trials suggested a possible benefit of lower mortality with albumin in subgroups with sepsis, via a nonsignificant trend in SAFE and a significant post hoc subgroup analysis in ALBIOS. ^, Likewise, most (but not all ) meta-analyses have concluded that albumin use is associated with reduced mortality in sepsis. Additional RCTs evaluating the role of albumin in the treatment of sepsis, such as albumin as endothelial rescue therapy, are ongoing (NCT03869385) or planned including the ALBIOSS-BALANCED trial (NCT03654001).

On the basis of a statistically significant increase in mortality rate in the subgroup with TBI in SAFE, albumin use should be avoided in TBI patients, for which saline is considered the fluid of choice. The use of artificial colloids, such as hydroxyethyl starch, gelatins, and dextrans, should be avoided in ICU patients due to an increased risk of AKI and a lack of evidence for their use. ^{, ,}

Metabolic acidosis is common in the ICU, present in 6% to 42% of patients, and, when persistent, is associated with a mortality of >50%. ^, Using IV sodium bicarbonate to correct metabolic acidosis in the ICU, though commonly done, remains controversial. Despite the theoretical benefits of improved hemodynamic responsiveness to catechol vasopressors, animal models, ^, human observational data, ^, and small human trials suggest little to no benefit by attempting to correct acidemia with supplemental IV bicarbonate, with a possible signal in observational data for improved outcomes in patients with sepsis complicated by AKI. Potential harms of exogenous sodium bicarbonate administration include intracellular acidosis, cerebral spinal fluid acidosis, impaired oxygen offloading in tissues, stimulation of lactic acidosis, disinhibition of glycolysis, hypercarbia and/or depression of respiratory drive, hypokalemia, hypocalcemia, overshoot alkalosis, volume overload, and (when given in hypertonic solutions) hypernatremia. ^,

The highest-quality data addressing the use of IV bicarbonate in the ICU available thus far come from the Sodium Bicarbonate to Treat Severe Acidosis in the Critically Ill (BICAR-ICU) trial. The study randomized nearly 400 patients in 26 ICUs in France with severe metabolic or mixed acidosis (pH ≤7.20, arterial pCO ₂ ≤45 mm Hg, serum bicarbonate ≤20 mmol/L) to unblinded treatment with 4.2% (500 mmol/L) sodium bicarbonate or standard care. Patients with pure respiratory acidosis, ketoacidosis, prior alkali therapy or KRT, stage IV CKD, and proven digestive or urinary tract loss of sodium bicarbonate were excluded, with the latter two criteria functionally excluding most non–anion gap metabolic acidosis. The primary outcome was the composite of death at 28 days or ≥1 organ failure at 7 days and was not statistically significant. However, patients in the subgroup with stage 2 or 3 AKI ( n = 182) had significant decreases in the primary outcome and both of its components, with a 28-day mortality of 46% vs. 63% ( P = 0.017) favoring sodium bicarbonate administration. In addition, there was a reduction in the need for KRT in the sodium bicarbonate group (35% vs. 52%). Though this subgroup analysis was prespecified and adjusted for multiple comparisons, the results remain hypothesis-generating. The findings of BICAR-ICU support the cautious use of sodium bicarbonate in critically ill patients with severe acidosis and AKI and the general avoidance of liberal sodium bicarbonate use in ICU patients without AKI. Larger studies are ongoing to better inform the use of sodium bicarbonate, specifically in ICU patients with AKI including the BICAR-ICU-2 trial (NCT04010630) and the U.K. Multicentre Evaluation of Sodium Bicarbonate in Acute Kidney Injury in Critical Care (MOSAICC) trial (ISRCTN14027629).

IHD sessions are typically prescribed three times weekly, lasting 3 to 4 hours each, with blood flow rates of 300 to 500 mL/min and dialysate flow rates of 500 to 800 mL/min ( Table 64.3 ). Ultrafiltration removes excess fluid, and solute clearance mainly relies on diffusion, with some larger solutes cleared via convection due to increased membrane porosity and internal filtration and backfiltration. The frequency and duration of dialysis depend on patient metabolic control, volume status, and hemodynamic stability. Initial sessions typically use lower blood and dialysate flows for patients susceptible to dialysis disequilibrium.

IHD is preferred for rapid correction of electrolyte imbalances and drug intoxications due to its rapid solute removal. However, IHD can lead to intradialytic hypotension in 10% to 70% of acute treatments. Proposed interventions for improving hemodynamic stability are mostly extrapolated from patients on maintenance IHD. These interventions include decreasing the ultrafiltration rate, lowering the blood flow rate, sodium modeling, cooling the dialysate, increasing the dialysate calcium concentration, using isolated ultrafiltration, using hypertonic infusions, extending the duration of dialysis or increasing the frequency of treatments in order to slow the ultrafiltration rate, and using midodrine. In general, these interventions lack robust evidence, especially in critically ill patients with AKI. Additionally, the rapid shifts of intracellular fluid and solutes during IHD can result in increased ICP and cerebral edema in patients with acute brain injury.

CRRT offers an alternative for managing hemodynamically unstable patients with AKI intolerant to IHD. CRRT uses diffusion, convection, or a combination to remove solutes and fluids gradually. This method theoretically allows for better hemodynamic tolerance compared with IHD by providing slower solute and fluid removal per unit time. CRRT is generally performed 24 hours a day with blood flow rates typically between 100 and 300 mL/min ( Table 64.3 ). Advantages of CRRT include better hemodynamic tolerance, continuous solute clearance, and improved volume control ( Fig. 64.3 ). Drawbacks include vascular access challenges, filter clotting, increased anticoagulation needs, reduced patient mobility, and higher costs and ICU nurse workload compared with IHD. Although rare, dialysis disequilibrium has been reported with CRRT. ^,

Superior volume control with CRRT versus intermittent hemodialysis (IHD).

(A) Fluid accumulation over time in patients on CRRT and on IHD. Data from the PICARD (Program to Improve Care in Acute Renal Disease) study. (B) Twenty-four-hour total fluid balance (I/Os; milliliters) during the first 3 days of kidney replacement therapy in patients on IHD and CVVHD therapy. Shown are median values with interquartile range (box borders) and extreme values (whiskers). CRRT, Continuous renal replacement therapy; CVVHD, continuous venovenous hemodialysis; IHD, intermittent hemodialysis.

A reproduced with permission from Bouchard J, Soroko SB, Chertow GM, et al. Fluid accumulation, survival and recovery of kidney function in critically ill patients with acute kidney injury. Kidney Int. 2009;76[4]:422–427. B reproduced with permission from Augustine JJ, Sandy D, Seifert TH, Paganini EP. A randomized controlled trial comparing intermittent with continuous dialysis in patients with ARF. Am J Kidney Dis. 2004;44[6]:1000–1007.

CRRT modalities for solute removal are continuous venovenous hemofiltration (CVVH), continuous venovenous hemodialysis (CVVHD), and continuous venovenous hemodiafiltration (CVVHDF). CVVH relies on convection without dialysate, using a hemofiltration solution (often referred to as “replacement fluid” based on historical practices) infused before the filter (prefilter) or after the filter (postfilter) to restore volume and electrolytes. When using postfilter hemofiltration solution, the filtration fraction should be kept <0.25 to reduce the likelihood of clotting of the hemofilter. This can be done by increasing the blood flow rate or introducing some hemofiltration solution before the filter, although the latter will dilute the concentration of solutes entering the hemofilter, leading to a decrease in solute clearance. ^, CVVHD primarily uses diffusion with slower dialysate flow rates, resulting in near-complete saturation of the dialysate. Ultrafiltration is primarily used for volume control. CVVHDF combines convection and diffusion, utilizing both dialysate and replacement fluid.

Despite enhanced middle-molecule clearance in convective therapies, meta-analyses comparing hemofiltration to hemodialysis for AKI found no significant differences in patient survival. Although high-volume hemofiltration (HVHF) with ultrafiltration rates >50 mL/kg/h has been used to enhance cytokine clearance in AKI patients with sepsis, controlled studies have not consistently demonstrated better outcomes. ^{, , ,} With no clear advantage of one CRRT modality over another, choice should be based on clinician preference and expertise, with attention to circuit survival.

Prolonged intermittent renal replacement therapy (PIRRT), an alternative to CRRT for critically ill patients with AKI, offers a hybrid approach. These techniques mitigate the hemodynamic impact of IHD by extending the duration of therapy and adapting conventional hemodialysis machines to operate at lower blood pump speeds (100–300 mL/min) and dialysate flow rates (100–300 mL/min) ( Table 64.3 ). Alternatively, CRRT equipment can be adapted for PIRRT, with sessions lasting 6 to 18 hours. PIRRT practices vary substantially in technology, prescription, and anticoagulation. ^, Diffusion-based PIRRT is described by terms like “sustained low-efficiency (daily) dialysis” (SLED or SLEDD), “extended daily dialysis” (EDD), “slow continuous dialysis” (SCD), and “go slow dialysis.” When combining convection and diffusion, these therapies are known as low-efficiency daily diafiltration (SLEDD-f) and sustained hemodiafiltration (S-HDF). “Accelerated venovenous hemofiltration” is used for exclusively convective therapy.

PIRRT combines advantages of CRRT and IHD, offering improved hemodynamic stability by gradually removing solutes and fluids over an extended timeframe. It eliminates the need for 24-hour therapy and costly CRRT equipment and solutions. PIRRT provides high solute clearances similar to IHD, allowing for uninterrupted diagnostic and therapeutic procedures. ^, It can also be administered overnight, promoting patient mobility during the day. Studies confirmed hemodynamic control comparable with CRRT. ^, In some medical institutions, PIRRT is employed to transition patients from CRRT to conventional IHD when their hemodynamic stability improves. ^,

PD leverages the peritoneum as a semipermeable membrane for solute removal by diffusion. A catheter is used to instill a dialysate solution into the peritoneal cavity, where it dwells for a specified duration, allowing solutes to diffuse from the blood into the dialysate before drainage. The dialysate contains varying high-glucose concentrations to establish an osmotic gradient for fluid removal. Acute PD can be conducted intermittently or continuously and either manually or through an automated cycler. PD techniques for AKI encompass high-volume PD, acute intermittent peritoneal dialysis (AIPD), continuous equilibrated peritoneal dialysis (CEPD), tidal peritoneal dialysis (TPD), and continuous flow peritoneal dialysis (CFPD).

High-volume PD employs an automated cycler to deliver 18 to 22 exchanges of 2 L every 24 hours, offering the highest small-solute clearance and ultrafiltration rate among PD methods. ^, AIPD employs brief dwell times with dialysate flows of 1 to 4 L/hour for 12 to 24 hours every other day. CEPD resembles continuous ambulatory PD for chronic outpatient PD but without patient ambulation. ^, In TPD, a portion of dialysate remains in the peritoneal cavity, facilitating rapid exchanges through a cycler, and improving solute clearance. CFPD involves simultaneous inflow and outflow of dialysate through two routes, maintaining a fixed intraperitoneal volume, with high dialysate flow rates and urea clearances of 30 to 50 mL/min. ^,

PD offers advantages like technical simplicity, hemodynamic stability, reduced dialysis disequilibrium risk, no need for anticoagulation or vascular access, and cost-effectiveness. ^, Gradual solute and fluid removal makes it safer for patients at risk of intracranial hypertension. Drawbacks include specialized catheter training, catheter placement complications, peritonitis risk, potential solute inadequacy, unpredictable ultrafiltration, hyperglycemia, protein loss, and respiratory compromise due to intraabdominal pressure. ^, PD is contraindicated for individuals with recent abdominal surgery, abdominal drains, diaphragmatic disease, or ileus. ^,

According to the 2012 KDIGO AKI guidelines, the preferred site for temporary dialysis catheter placement is the right internal jugular (IJ) vein, followed by the femoral vein, followed by the left IJ, with the subclavian veins as a last resort. The catheter should be inserted with the tip in a large vein at an appropriate depth: 15 to 20 cm for right IJ catheters and 20 to 24 cm for left IJ catheters with a tip at the cavoatrial junction and at least 20 to 25 cm for femoral catheters with a tip in the inferior vena cava. ^, Ultrasound guidance and strict aseptic technique are recommended to minimize complications. While there is consensus favoring the right IJ, some argue for left IJ over femoral placement, especially in obese patients at higher risk of femoral catheter complications. In an RCT comparing femoral and jugular catheters, infection rates were similar overall but higher in the femoral group for patients with a BMI >28.4 kg/m ² . In a post hoc analysis, right IJ placement had the lowest catheter malfunction rate, followed by femoral, with highest in the left IJ, though the malfunction of the left IJ catheters in this study may have been due to inadequate depth of insertion (16 cm). ^, Observational studies suggest lower infection and thrombosis risks with tunneled catheters, along with better blood flow rates. ^, Therefore considering a tunneled dialysis catheter over a nontunneled one should be considered for prolonged KRT requirements, but further research is needed before making it a standard practice for AKI in the ICU. ^,

KRT dosing in AKI is determined by Kt/V _urea and frequency in IHD and PIRRT, as well as effluent flow rate (mL/kg/h) in CRRT. Two large RCTs, the VA/NIH ATN study and the RENAL study, evaluated KRT dosing in AKI. In the ATN trial, intensive KRT included high-dose CVVHDF or PIRRT for unstable patients and IHD for stable patients. In the intensive approach, patients received CVVHDF with an effluent rate of 35 mL/kg/h or six PIRRT/IHD sessions per week, with each session targeting a Kt/V _urea of 1.2 to 1.4. The less intensive approach used CVVHDF with an effluent rate of 20 mL/kg/h or three PIRRT/IHD treatments per week with the same Kt/V _urea target. Notably, the intensive KRT strategy did not reduce 60-day mortality or improve kidney recovery. The RENAL trial compared CRRT at doses of 40 and 25 mL/kg/h and revealed no survival advantage with higher CRRT intensity at 90 days. Multiple meta-analyses also showed no survival benefit with increased KRT intensity. Studies of high-volume CRRT with doses up to 70 to 85 mL/kg/h have similarly shown no benefit. ^,

On the basis of these studies, KDIGO 2012 AKI guidelines recommend a CRRT-delivered dose of 20 to 25 mL/kg/h and a Kt/V _urea of ≥1.2 per session for PIRRT or IHD. ^, In IHD, continued patient assessment is needed to ascertain whether additional treatments are required to maintain electrolyte and acid-base balance and control volume status. For CRRT, doses of 25 to 30 mL/kg/h can be considered to account for therapy interruptions. Temporary use of higher doses may be required for specific conditions including hyperkalemia, severe acidosis, or hyperammonemia, but dose should be adjusted back to standard ranges once solute control is achieved. CRRT dosing should remain adaptable and continuously adjusted to meet the dynamic needs of critically ill patients.

High-intensity KRT can have adverse consequences by removing beneficial solutes, such as essential nutrients, antiinflammatory cytokines, and antibiotics. ^, Both the ATN and RENAL studies observed increased rates of hypophosphatemia with higher KRT intensity. ^, In the ATN study, higher-intensity KRT was associated with lower rates of successful extubation and fewer ventilator-free days, especially in patients with lower serum phosphate levels. Higher KRT intensity also increased the risk of a ≥50% decline in urine output (HR = 1.29; 95% CI, 1.1–1.51). Patients treated only with IHD in the ATN study had more KRT-free days with lower-intensity KRT. A meta-analysis of six trials involving 1926 subjects found that higher-intensity KRT was associated with longer KRT duration and a higher likelihood of KRT dependence at day 28 (HR = 1.15, 95% CI 1–1.33). Additionally, high-intensity KRT is costlier due to expenses for disposable items and nursing labor. ^, The threshold for harmful underdosing in KRT intensity is uncertain. Some experts suggest a potential mortality risk below 20 mL/kg/h, but observational data from Japan suggest that doses as low as 15 mL/kg/h are safe.

KDIGO 2012 guidelines for AKI recommend using regional citrate anticoagulation (RCA) over heparin in CRRT unless contraindicated. Citrate chelates ionized calcium (iCa) in the extracorporeal circuit, and optimal anticoagulation is achieved when hemofilter iCa is <0.4 mmol/L. A portion of the calcium-citrate complex is lost across the hemofilter while the rest enters the systemic circulation, where the liver metabolizes citrate to bicarbonate and calcium is released into the circulation. Calcium is infused into the patient to replace the calcium lost across the hemofilter.

Proactive management of RCA is essential to mitigate potential side effects including metabolic alkalosis and acidosis, hypocalcemia and hypercalcemia, and hypomagnesemia. Patients with severe shock liver and lactic acidosis may not effectively metabolize citrate. ^{, ,} Citrate accumulation is characterized by low systemic iCa, elevated total calcium, total calcium to systemic iCa ratio >2.5, anion gap metabolic acidosis, and escalating calcium infusion requirements ( Table 64.4 ). ^, Nonetheless, RCA has been safely used in advanced liver disease and perioperative liver transplant patients.

Multiple RCTs including the multicenter Effect of Regional Citrate Anticoagulation versus Systemic Heparin Anticoagulation During Continuous KRT on Dialysis Filter Life Span and Mortality Among Critically Ill Patients With Acute Kidney Injury (RICH) trial support RCA’s superiority over systemic heparin in extending filter life and reducing bleeding risk. In the RICH trial, exploratory secondary outcomes demonstrated a potential association between citrate and an increased risk of infection. ^, However, a post hoc analysis found that filter lifespan exceeding 48 h, but not anticoagulation choice, was associated with higher infection rates. Further research is needed to clarify this observation.

Anticoagulation-free KRT sessions are feasible with IHD and PIRRT due to the higher blood flow and shorter durations. CRRT without anticoagulation may require more circuit changes due to clotting, but strategies such as higher blood flow rates and prefilter replacement solution may mitigate clotting risk. In one RCT, different blood flow rates (150 vs. 250 mL/min) in CRRT had no significant impact on filter clotting, possibly due to higher flow rates triggering more pressure alarms and reductions or stoppages in blood pump speed, with consequent increased clotting risk.

Achieving euvolemia is crucial in critical illness, especially due to the harms of fluid overload. ^{, ,} Ultrafiltration and rapid osmotic shifts during IHD can cause intradialytic hypotension from inadequate plasma refill and impaired hemodynamic compensatory mechanisms. Additionally, IHD can induce hypotension due to “myocardial stunning,” unrelated to ultrafiltration, observed in both patients on maintenance IHD and those with AKI undergoing IHD. ^, Surprisingly, myocardial stunning was also seen in one study of patients with AKI on CRRT, suggesting microcirculatory factors might play a role.

Data on prescribing net ultrafiltration (UF _NET ) in CRRT are limited. UF _NET is the net volume of fluid removed per hour adjusted for patient body weight (mL/kg/h). In a secondary analysis of the RENAL trial, UF _NET rates above 1.75 mL/kg/h were associated with lower survival and a higher risk of hypophosphatemia. Accumulating observational data suggest a nonlinear relationship between UF _NET and mortality, with the lowest mortality at UF _NET rates of 1.0 to 1.75 mL/kg/h ( Fig. 64.4 ). ^, Iterative assessment of patient’s fluid balance (desired vs. achieved) and tolerance to fluid removal is key for risk stratification. ^, Prospective studies are needed to validate the proactive targeting of UF _NET within the 1.0 to 1.75 mL/kg/h range and/or other methods to achieve personalized fluid balance goals.

Net ultrafiltration rate is independently associated with mortality in patients treated with continuous renal replacement therapy (CRRT) in a nonlinear fashion.

The relationship between mortality and net ultrafiltration (UF _NET ) in patients in the intensive care unit with acute kidney injury (AKI) is J shaped. In a post hoc analysis of >1400 patients from the Randomized Evaluation of Normal versus Augmented Level Replacement Therapy (RENAL) study comparing high versus low dose of CRRT for AKI in the intensive care unit, low UF _NET rates <1.01 mL/kg per hour, and high UF _NET rates >1.75 mL/kg per hour were associated with higher risk-adjusted 90-day mortality compared with UF _NET rates in a middle range of 1.01–1.75 mL/kg per hour. It has been proposed that these effects are mediated by harms of organ edema in those treated with low UF _NET rates and by organ ischemia in those treated with high UF _NET rates, but no prospective trial data yet exist to demonstrate that targeting a moderate rate of UF _NET improves outcomes.

Data from Murugan R, Kerti SJ, Chang CH, et al. Association of net ultrafiltration rate with mortality among critically ill adults with acute kidney injury receiving continuous venovenous hemodiafiltration: a secondary analysis of the randomized evaluation of normal vs augmented level [RENAL] of renal replacement therapy trial. JAMA Netw Open. 2019;2[6]:e195418. Reprinted with permission from Murugan R, Bellomo R, Palevsky PM, Kellum JA. Ultrafiltration in critically ill patients treated with kidney replacement therapy. Nat Rev Nephrol . 2021;17[4]:262–276.

Specific drug-dosing considerations in critically ill patients with AKI include attention to drugs with long half-lives and those that rapidly achieve therapeutic concentrations. Importantly, loading doses do not typically need to be adjusted in patients with AKI or those requiring KRT. In contrast, maintenance doses require adjustments in patients with AKI or those requiring KRT. When KRT is initiated, dosing strategies must further account for the extracorporeal clearance of medications. Drug characteristics (e.g., volume of distribution, molecular weight, and extent of protein binding), KRT prescriptions (e.g., blood flow, session duration, proportion of convective and diffusive clearances), and dialyzer features (e.g., porosity or flux, surface area or efficiency) are important determinants of medication clearance. In some cases, KRT may have a more subtle impact on drug clearance by enhancing extrarenal drug metabolism, perhaps through the removal of uremic toxins. For recipients of IHD, medications should be administered after dialysis or a postdialysis supplemental dose should be considered. With the predominant use of high-flux filters, which enhance drug clearance, dosing recommendations that were derived for use with low-flux filters typically need to be increased by 25% to 50%.

The advent of PIRRT poses a further challenge because extracorporeal clearance is augmented by this modality as compared with IHD. This may result in increased drug removal and inadequate blood concentrations of vital medications including antibiotics. To date, drug-dosing guidance with PIRRT is available for only a few agents. As a general rule, drug doses should probably be higher than those administered to patients receiving IHD and, where feasible, their use should be accompanied by careful drug level monitoring. A “top-up” dose following the conclusion of a PIRRT session is advisable for agents that are likely to be dialyzable.

In patients receiving CRRT, total effluent dose and the relative breakdown of convective and diffusive therapy will affect drug removal and subsequent dosing changes. CRRT at usual doses, assuming no interruptions to therapy, begins to approximate endogenous kidney function (∼30 to 60 mL/min), and more frequent drug dosing is generally required than with IHD or PIRRT. For a more detailed discussion of this topic, the reader is referred to Chapter 56 .