Kidney replacement therapy (dialysis) in acute kidney injury in adults: Indications, timing, and dialysis dose

INTRODUCTION — The management of patients with acute kidney injury (AKI) is supportive, with kidney replacement therapy (KRT) indicated in patients with severe kidney injury. Multiple modalities of KRT are available. These include intermittent hemodialysis (IHD); continuous kidney replacement therapies (CKRTs); and hybrid therapies, also known as prolonged intermittent kidney replacement therapies (PIKRTs), such as sustained low-efficiency dialysis (SLED) and extended-duration dialysis (EDD). Despite these varied techniques, mortality in patients with AKI remains high, exceeding 40 to 50 percent in severely ill patients. (See "Kidney and patient outcomes after acute kidney injury in adults".)

The initiation of KRT in patients with AKI prevents uremia and immediate death from the adverse complications of kidney failure. It has been suggested that variations in the timing of initiation, modalities, and/or dosing may affect clinical outcomes, particularly survival, although few studies have directly examined these issues.

The optimal timing, type of modality, and dosing strategy for patients with AKI who require KRT is reviewed here. The different modalities are discussed separately. (See "Continuous kidney replacement therapy in acute kidney injury" and "Prolonged intermittent kidney replacement therapy" and "Use of peritoneal dialysis (PD) for the treatment of acute kidney injury (AKI) in adults" and "Acute hemodialysis prescription".)

URGENT INDICATIONS — Accepted urgent indications for KRT in patients with AKI generally include:

●Fluid overload refractory to diuretic therapy

●Severe hyperkalemia (plasma potassium concentration >6.5 mEq/L) or rapidly rising potassium levels

●Overt manifestations of uremia, such as pericarditis, encephalopathy, or an otherwise unexplained decline in mental status

●Severe metabolic acidosis (pH <7.1) despite medical management, though the benefit of KRT in patients with lactic acidosis is uncertain (see 'Timing of elective initiation' below)

●Certain alcohol and drug intoxications amenable to extracorporeal therapy

The likelihood of requiring KRT is increased in patients with underlying chronic kidney disease (CKD) in proportion to the degree of reduction in glomerular filtration rate (GFR) at baseline. This was illustrated in a study that compared the prehospitalization estimated GFR (eGFR; from the most recent serum creatinine) in 1746 hospitalized patients who developed dialysis-requiring AKI with that of 600,820 hospitalized patients who did not [1].

Compared with patients with an estimated baseline GFR >60 mL/min/1.73 m², the risk of developing AKI requiring dialysis progressively and significantly increased with the severity of underlying CKD. The adjusted odds ratios (ORs) were 1.7, 4.6, and 20.4 for patients with stage 3 (eGFR of 30 to 59 mL/min/1.73 m²), 4 (eGFR 15 to 29 mL/min/1.73 m²), and 5 CKD (eGFR <15 mL/min/1.73 m²), respectively. (See "Overview of the management of chronic kidney disease in adults", section on 'Definition and classification'.)

TIMING OF ELECTIVE INITIATION — Patients who meet one or more of the criteria listed above require urgent initiation of KRT. (See 'Urgent indications' above.)

Even if one of the urgent indications is not satisfied, we electively initiate KRT in patients with AKI that is unlikely to resolve quickly and who, in addition, have one or more of the following:

●Serum potassium >6.0 mEq/L that is unresponsive to aggressive medical management, or >5.5 mEq/L if there is ongoing tissue breakdown (eg, rhabdomyolysis, crush injury, tumor lysis syndrome), or ongoing potassium absorption (eg, due to severe gastrointestinal bleeding). Elective initiation of KRT in patients before the potassium reaches 6.5 mEq/L can help avoid emergency initiation and potentially life-threatening arrhythmias. Serum potassium can rise rapidly in patients with acute kidney injury (AKI) who have ongoing tissue breakdown or potassium absorption.

●Severe metabolic acidosis (pH <7.15) without reversible cause (eg, ketoacidosis) and despite optimal medical management (eg, intravenous sodium bicarbonate therapy as volume status permits). Data supporting a precise pH threshold for initiation of KRT in this setting are lacking; some experts would suggest initiation of KRT at higher pH levels (eg, pH <7.2).

The benefit of KRT in patients with severe metabolic acidosis due to lactic acidosis is controversial, as the rate of clearance that can be provided by KRT is substantially lower than endogenous generation [2]. Although KRT is often employed as supportive therapy performed as a bridge to definitive management of the underlying cause of lactic acidosis (eg, bowel resection for ischemic bowel), there is little evidence of mortality benefit. The exception is the treatment of metformin-associated lactic acidosis, in which KRT reverses the underlying cause.

●Hypervolemic patients who are oliguric or who remain in persistent positive fluid balance despite high doses of loop diuretics (often used in combination with a thiazide or thiazide-like diuretic), particularly if oxygen requirements are increasing. Elective initiation in such patients can help avoid the need for intubation and mechanical ventilation.

We do not use blood urea nitrogen (BUN) values to make decisions regarding timing of initiation of KRT. This is because BUN is often affected by several factors unrelated to kidney function (eg, gastrointestinal bleeding, use of high-dose glucocorticoids). We also do not use any specific time threshold (eg, 72 hours of severe AKI) to initiate KRT in the absence of the urgent or elective indications presented above.

Initiating KRT even earlier in the course of AKI (ie, before the patient develops an urgent indication or one of the elective indications just mentioned) is generally not beneficial. Early initiation of KRT may be harmful, may delay recovery of kidney function, and results in increased health care utilization.

Multiple trials have compared strategies of early KRT initiation (in the absence of any indications mentioned above) with delayed KRT initiation (once indications have developed) [3-11]. The best data come from a large multicenter randomized trial and a previously published meta-analysis that synthesized findings from older, smaller trials:

●In the Standard versus Accelerated Initiation of Renal-Replacement Therapy in Acute Kidney Injury (STARRT-AKI) trial, 3019 critically ill patients with severe AKI (but who did not yet have an indication for KRT) were randomly assigned to an early strategy (initiation of KRT as soon as possible and within 12 hours of identification) or delayed strategy (initiation of KRT once indications developed) [3]. At the time of randomization, the mean serum creatinine was 3.5 mg/dL (309 micromol/L), the mean serum potassium was 4.5 mEq/L, the mean serum bicarbonate was 19.6 mEq/L, the mean urine output was approximately 460 mL/24 hours, and 45 percent of patients were oliguric or anuric. Mechanical ventilation and vasopressor support were required in 77 and 70 percent of patients, respectively.

KRT was initiated in 97 percent of patients randomized to the early strategy as compared with 62 percent of patients assigned to the delayed strategy. There was no difference in mortality at 90 days (44 percent in both groups); however, patients assigned to the early strategy were more likely to remain KRT-dependent at 90 days (10.4 versus 6.0 percent) and to require rehospitalization (21 versus 17 percent). In addition, adverse events were more common with the early strategy (23 versus 17 percent). Earlier initiation of KRT was not associated with improved outcomes in a secondary analysis stratified by severity of volume overload [12].

●In a meta-analysis of nine studies and 1879 patients, early KRT initiation did not decrease mortality at 28 (43 versus 44 percent), 60 (51 percent in both groups), or 90 days (56 versus 55 percent) [4]. All patients in the early initiation group received KRT, compared with 58 percent in the delayed initiation group. There were no differences in the risk of hyperkalemia, severe cardiac arrhythmias, or severe bleeding events. At hospital discharge, there were no between-group differences in the proportion of patients with KRT dependence, in serum creatinine among those who were not KRT dependent, or in ventilator or vasopressor use. There was no heterogeneity across the included trials and most trials were high quality.

Even though most studies favor delayed KRT initiation, there are likely to be limits to how long KRT can be safely delayed. However, that exact point in time beyond which the benefits of KRT are lost is not known. In the Artificial Kidney Initiation for AKI 2 (AKIKI-2) trial, 278 critically ill patients with severe AKI who did not have urgent indications for KRT were monitored for development of either oliguria of 72 hours duration or an elevation in BUN to between 112 and 140 mg/dL [13]. Upon meeting one of these criteria, the patients were randomly assigned to either initiate KRT or defer KRT until an urgent indication developed or the BUN exceeded 140 mg/dL. KRT was ultimately initiated in 79 percent of those assigned to have dialysis deferred. There was no difference in the number of days alive and free of KRT between randomization and day 28. At both 28 and 60 days, the mortality was higher in the group who had KRT deferred (45 versus 38 percent at day 28 and 55 versus 44 percent at day 60). However, these differences were not statistically significant. Although KRT dependence was slightly lower at day 28 in the group that had KRT deferred (11 versus 16 percent), the difference had abated at day 60.

To place the findings of the AKIKI-2 trial in context with other similar trials, such as the STARRT-AKI trial, it is important to note that most other trials did not use specific BUN values for decision-making regarding initiation of KRT [3,13]. By design, the AKIKI-2 trial allowed a longer period of non-KRT management of severe AKI, as evidenced by higher BUN values at the time of KRT initiation (mean BUN 89 versus 123 mg/dL among those who had KRT deferred). By comparison, the BUN values at KRT initiation among patients in the STARRT-AKI trial were lower (mean BUN 63 and 85 mg/dL in the early and delayed groups, respectively).

Considering the totality of data across multiple trials, in the absence of indications, accelerated initiation of KRT is not associated with clinical benefit, results in greater utilization of health care resources, and may be associated with an increased risk for delayed recovery of kidney function. However, there does appear to be a poorly defined threshold (that likely varies between patients) beyond which delaying KRT, in favor of spontaneous recovery of kidney function, may be harmful.

OPTIMAL MODALITY — A large number of modalities are available for KRT. These include intermittent hemodialysis (IHD), peritoneal dialysis, continuous kidney replacement therapy (CKRT), and hybrid therapies such as sustained low-efficiency hemodialysis (SLED).

Data do not support the superiority of any particular mode of KRT in patients with AKI. In the majority of patients, selection of modality should therefore be based upon local expertise and availability of staff and equipment. However, in selected patients, other factors may prevail. As an example, in patients with acute brain injury or fulminant hepatic failure, continuous therapy may be associated with better preservation of cerebral perfusion. However, the costs associated with CKRT may be greater than with other modalities of KRT.

Continuous kidney replacement therapies versus intermittent hemodialysis — CKRT represents a family of modalities that provides continuous support for severely ill patients with AKI. These include continuous hemofiltration, hemodialysis, and hemodiafiltration, which involve both convective and diffusive therapies. Although superior clearance of middle- and larger-molecular-weight molecules is associated with convective therapies (hemofiltration) compared with diffusive therapies (hemodialysis), there are no studies clearly showing improved clinical outcomes compared with the type of solute transport.

Survival and recovery of kidney function are similar with both CKRT and IHD, and the Kidney Disease: Improving Global Outcomes (KDIGO) Clinical Practice Guidelines for AKI suggest using intermittent and CKRT as complementary therapies in patients with AKI [14].

Multiple prospective, randomized studies have compared outcomes of AKI, supported using either IHD or CKRT [15-19]. As examples:

●In a multicenter study, 166 patients with AKI were randomly assigned to IHD or CKRT [16]. CKRT was associated with significantly higher all-cause mortality at 28 days (59.5 versus 41.5 percent) and in-hospital mortality (65.5 versus 47.6 percent). However, despite randomization, patients randomly assigned to CKRT were significantly more likely to have higher Acute Physiologic and Chronic Health Evaluation (APACHE) III scores and liver failure. After adjustment for these characteristics, there was no increased risk of death with CKRT (adjusted odds of death 1.58, CI 0.7-3.3).

●In the HemoDiafe Study (a prospective, multicenter, French study), 360 patients with AKI and multi-organ dysfunction syndrome were randomly assigned to IHD or continuous venovenous hemodiafiltration (CVVHDF) [15]. The primary endpoint was survival at 60 days. Severity of illness was similar in both randomized groups, protocol adherence was good, both groups used the same dialysis membranes, and there was a low rate of crossover from intermittent to continuous therapies (3.3 percent). At 60 days, survival was the same in both groups (32 and 33 percent in the intermittent and continuous groups, respectively). In addition, both therapies were associated with similar rates of hypotension, including the group of hemodynamically unstable patients.

While the HemoDiafe Study was the largest and most rigorously conducted randomized, controlled trial comparing modality of KRT in AKI, definitive conclusions must be tempered by specific aspects of the trial design and execution. Most importantly, the mean duration of the hemodialysis treatments were longer (5.5 hours) than is typically employed in clinical practice in the United States. Applicability of these results, therefore, need to tempered based on local procedures for the use of IHD in hemodynamically unstable patients. Importantly, however, this study demonstrates that is possible to successfully perform IHD in practically all patients with AKI given the findings of similar hypotensive rates with stable and unstable patients and the low crossover rate from intermittent to continuous therapy.

●The Clinical Trial Comparing Continuous versus Intermittent Hemodialysis in Intensive Care Unit Patients (CONVINT) trial was a single-center, randomized, controlled trial of 252 critically ill patients in Germany who were randomly assigned to continuous venovenous hemofiltration (CVVH) at a prescribed dose of 35 mL/kg/hour or daily hemodialysis [19]. Nineteen and one-half percent of the 128 patients randomized to IHD were switched to CVVH after a mean of 4.4±12 days for severe hypotension or volume management, while 45.9 percent of patients randomized to CVVH were switched to IHD after a mean of 6.2±5.6 days because of repeated filter clotting, bleeding necessitating discontinuation of anticoagulation, metabolic control, thrombocytopenia, or improvement in overall status and need for mobilization.

Survival 14 days after discontinuation of KRT was 39.5 percent in the IHD group as compared with 43.9 percent in the CVVH group, with no difference in in-hospital mortality or mortality at 30 days of follow-up. Similarly, there was no difference in recovery of kidney function.

Meta-analyses that compared outcomes with CKRT and IHD have also been performed [20-26]. Overall, no survival benefit can be attributed to either modality.

●Recovery of kidney function – Recovery of kidney function appears to be the same with CKRT and IHD. Although some observational studies report better recovery of kidney function with CKRT [27-30], these reports only evaluated kidney function recovery in patients who survived, thereby failing to account for mortality differences between groups. When the analysis combined mortality and nonrecovery of kidney function, both groups showed similar recovery of function [31,32]. Randomized trials have also found no such benefit with CKRT [15,17,18]. These observations were confirmed in a meta-analysis that included 3971 survivors of KRT-requiring AKI. A pooled analysis of 16 observational studies (n = 3499) showed a higher rate of dialysis dependence associated with IHD, but analysis of seven randomized trials (n = 472) showed no difference in recovery of kidney function between groups [25]. Similar findings were reported in a second meta-analysis [26].

●Other differences – CKRT may be associated with the following advantages compared with IHD:

•Enhanced hemodynamic stability, which may be particularly beneficial in hemodynamically unstable patients [17]. Hemodynamic stability is thought to be related to slower solute and volume removal and the effects of modest hypothermia often associated with CKRT.

•More consistent net salt and water removal, particularly in hemodynamically unstable patients, thereby permitting more gradual correction of volume overload and nutritional requirements [17,33].

•Enhanced clearance of inflammatory mediators, which may provide benefit in septic patients, particularly using convective modes of continuous therapy [34-36]. However, a meta-analysis of convection versus diffusion demonstrated no benefit to convection [37]. Open-label, randomized, controlled trials have also shown no benefit of high-volume hemofiltration in sepsis [38] or cardiogenic shock following cardiac surgery [9]. Several factors may contribute to this lack of benefit. Although convective therapy may provide enhanced clearance of proinflammatory mediators, it may also result in removal of beneficial antiinflammatory mediators. In addition, the maximal achieved extracorporeal clearance of these mediators is low relative to the rates of generation and endogenous clearance.

•Among patients with cerebral edema, acute brain injury, or fulminant hepatic failure, continuous therapy may be associated with better preservation of cerebral perfusion [39].

Prolonged intermittent kidney replacement therapy — Although prolonged intermittent kidney replacement therapy (PIKRT) has been shown to have similar hemodynamic effects and provides similar metabolic control as CKRT, there are few data comparing outcomes with either IHD or CKRT [40]. In a single-center study of 60 patients treated with either PIKRT or CKRT, PIKRT was associated with comparable or better clinical outcomes [41]. Similar results were observed in a second single-center study comparing PIKRT to CVVH in 232 critically ill patients [42]. There was no difference in 90-day all-cause mortality between the PIKRT group and the CVVH group. Patients treated with PIKRT required fewer days of mechanical ventilation, required fewer days in the intensive care unit (ICU), and received fewer blood transfusions, resulting in an overall lower cost of therapy. A subsequent meta-analysis of 7 randomized, controlled trials and 10 observational studies comparing PIKRT with CKRT found no difference in mortality or recovery of kidney function associated with modality of therapy [43]. These results were confirmed in a second meta-analysis restricted to randomized trials that found no difference in either mortality or recovery of kidney function [26]. (See "Prolonged intermittent kidney replacement therapy", section on 'Definition and indications for PIKRT'.)

Peritoneal dialysis — Peritoneal dialysis (PD) has a long history of use in the treatment of AKI, but is now rarely used in the United States and many other countries. A comparison of outcomes with PD relative to other modalities is presented separately. (See "Use of peritoneal dialysis (PD) for the treatment of acute kidney injury (AKI) in adults".)

Acute PD, where available, is a reasonable alternative to modalities of extracorporeal kidney support (ie, IHD, PIKRT, and CKRT). While PD may be used in the management of AKI in most settings, it may be of particular benefit in the following settings [44]:

●Patients who develop AKI in areas where equipment to perform hemodialysis, PIKRT, or CKRT are unavailable (eg, resource-limited setting) or in times of scarcity of supplies and/or surge in demand for kidney support (eg, natural disasters, pandemics)

●Patients who are at a high risk for bleeding with anticoagulation

●Patients for whom hemodialysis access could not be secured

●Patients with heart failure refractory to medical management and who require dialysis for AKI

●Patients with cirrhosis and ascites

However, acute PD should generally be avoided among the following patient groups [44]:

●Patients who have undergone recent abdominal surgery – Acute PD may be a suboptimal choice of modality among patients who have had abdominal surgery within a few weeks prior to development of dialysis-requiring AKI. This is because abdominal surgery leads to violation of the peritoneal cavity, potentially interfering with peritoneal membrane function. In addition, patients may have abdominal drains placed, which increase the incidence of infection of the PD catheter and confound fluid accounting with PD. Among such patients, we prefer waiting to initiate PD for at least 30 days after the surgery or until abdominal drains are out, whichever occurs first [45-47]. Even among patients with a history of remote abdominal surgery, the presence of an abdominal hernia or intra-abdominal adhesions may affect efficacy of acute PD.

●Patients who have diaphragmatic pleuroperitoneal connections – Performing PD among patients with a diaphragmatic pleuroperitoneal connection may result in a large pleural effusion and respiratory compromise. It should be noted that such connections may result from cardiac or other thoracic surgeries. (See "Evaluation and management of pleural effusions following cardiac surgery".)

●Severe or impending respiratory failure – Among patients with severe or impending respiratory failure, instillation of the peritoneal dialysate can increase intra-abdominal pressure and limit diaphragmatic excursion, leading to respiratory compromise. However, using a cycler with small fill volumes (eg, 1 liter) and frequent exchanges (every 90 minutes) may be feasible in some patients. In addition, among patients with pulmonary edema, acute PD may not be sufficient to relieve hypervolemia, further placing such patients at risk for respiratory failure.

●Extremely high or rapidly rising potassium – Potassium clearance with PD is slower than with hemodialysis or hemofiltration [48]. Thus, acute PD is not an ideal modality among patients who have conditions causing life-threatening hyperkalemia or rapidly rising potassium levels (eg, rhabdomyolysis, tumor lysis syndrome, or hypercatabolic states).

●Intra-abdominal sepsis or abdominal wall cellulitis – Patients with peritonitis may have difficult and unpredictable clearance of solutes due to increased membrane permeability, often resulting in diminished ultrafiltration. In addition, the PD catheter can become infected from placement in an infected peritoneal cavity and remain a source for ongoing infection, ultimately hindering cure.

Patients with other conditions such as severe gastroesophageal reflux disease, pregnancy, or uncontrolled diabetes should be evaluated on a case-by-case basis for suitability of acute PD [49]. These are not absolute contraindications to acute PD but may require prescription modifications or other adjustments to ensure safe delivery of PD.

OPTIMAL DOSING

Intermittent hemodialysis — Dosing in intermittent hemodialysis (IHD) is based upon the dose delivered per session, as well as the frequency of sessions. Thus, outcomes may vary based upon differences in dose per session, as applied to a fixed treatment schedule, or with differences in treatment schedule, as applied to a fixed dose per session. In addition, alterations in the dose per session as well as in the dialysis schedule can also be evaluated.

There have been no studies that have prospectively evaluated the impact of differences in dose per session in patients undergoing IHD on a fixed schedule, such as three times per week. Some data suggest that dosing may have an impact on patients with intermediate levels of disease severity. For example, in a retrospective study of 844 patients in intensive care units (ICUs) with AKI that used Kt/V as a measure of the delivered dose of acute IHD, improved survival was observed with a higher Kt/V (>1) among patients with intermediate levels of illness severity [50]. By comparison, among those either extremely ill or not very ill, the intensity of the dialysis treatments was less influential on outcome.

In one study that evaluated the impact of frequency of IHD on outcomes, 160 patients with AKI were assigned in an alternating fashion to either daily or every-other-day hemodialysis [27]. Enrolled patients were likely to have intermediate levels of illness severity, which was supported by both Acute Physiologic and Chronic Health Evaluation (APACHE) scores and the concurrent offering of continuous kidney replacement therapy (CKRT) at the study center. Compared with every-other-day dialysis, daily therapy was associated with a significant reduction in mortality (28 versus 46 percent), fewer hypotensive episodes during hemodialysis, and more rapid resolution of AKI (mean 9 versus 16 days). The delivered dialysis dose in the every other day group, as assessed by single-pool Kt/V, was low (0.94 per dialysis). By comparison, the daily dialysis group received the same per treatment delivered dose, but it was delivered twice as frequently. Thus, this study may have best concluded that inadequate therapy is associated with increased mortality.

In contrast, the Veterans Affairs (VA)/National Institutes of Health (NIH) Acute Renal Failure Trial Network (ATN) study, which allowed patients to switch between intermittent and either CKRT or prolonged intermittent kidney replacement therapy (PIKRT) as hemodynamic status changed over time, did not find a difference in mortality associated with a more intensive dosing strategy for KRT [51]. This outcome was independent of the percentage of time treated using IHD, with no difference in mortality based upon intensity of therapy among the 247 patients who only received IHD [52]. This study is discussed in detail in the next section. (See 'Continuous kidney replacement therapy' below.)

The 2012 Kidney Disease: Improving Global Outcomes (KDIGO) guidelines for AKI recommend delivering a Kt/V of 3.9 per week for patients undergoing intermittent therapy [14]. While these results are loosely based on the results of the ATN study, several caveats need to be considered. In the ATN study, the targeted dose of IHD in both treatment arms was a Kt/V of 1.2 to 1.4 per treatment, with a median delivered Kt/V of 1.3 per treatment [51]. The weekly dose of dialysis recommended in the KDIGO guidelines is the arithmetic sum of the median dose in the less intensive arm, summed over the course of a week. Since a median delivered Kt/V of 1.3 implies that half of treatments had a delivered Kt/V of less than this value, it is not clear that a per-treatment dose of 1.3 represents the appropriate target. In addition, the approach of taking the arithmetic sum of the individual treatment dose to calculate a weekly dose is not consistent with urea kinetic modeling; the weekly dose provided by six treatments with a Kt/V of 0.65 is not equivalent to three treatments with a Kt/V of 1.3 [53]. We therefore recommend that, if IHD is provided three times per week, the targeted dose of therapy should be a Kt/V of ≥1.2 per treatment, with monitoring of the delivered dose of therapy. If this minimum dose is achieved, there is no evidence that more frequent hemodialysis is associated with improved outcomes, unless necessitated for specific acute indications (eg, hyperkalemia). Conversely, if a Kt/V of ≥1.2 per treatment cannot be achieved, treatment frequency should be increased. Given issues related to the assessment of volume of distribution of urea in acutely ill patients, measurement of urea reduction ratio (URR) may be a reasonable alternative to assessment of Kt/V. In a post-hoc analysis of data from the ATN study, a URR >0.67 corresponded to a Kt/V >1.2 with a sensitivity of 0.99 and a specificity of 0.77 [54].

If more frequent IHD is provided, the targeted dose per treatment may be lower; however, the optimal dose is not defined. Surveys of practice in the United States suggest that monitoring of the delivered dose of hemodialysis is only infrequently assessed [55].

More frequent hemodialysis may be required for volume management in patients with one or more of the following: severe volume overload; a hypercatabolic state in which thrice-weekly hemodialysis is inadequate to provide control of azotemia, acidosis, or hyperkalemia; and when the targeted delivered dose of therapy cannot be achieved.

The Hanover Dialysis Outcome (HAND-OUT) study compared extended-duration dialysis (EDD), provided for approximately eight hours per day, with a more intensive regimen where additional eight-hour treatment sessions were provided to maintain the blood urea nitrogen (BUN) <42 mg/dL [56]. No difference in survival or recovery of kidney function was observed with more intensive treatment.

Continuous kidney replacement therapy — For patients on CKRT, we prescribe an effluent flow rate of approximately 25 mL/kg/hour in order to achieve (despite interruptions and CKRT downtime, which are inevitable) a minimum effluent rate of 20 mL/kg/hour over a 24-hour period. This is consistent with the 2012 KDIGO guidelines [14]. (See "Prescription of continuous kidney replacement therapy in acute kidney injury in adults", section on 'CKRT dose'.)

Outcomes of an increased dose of CKRT have been assessed in several randomized, controlled trials and three meta-analyses [8,57-62].

Most studies found that, compared with standard-intensity dialysis, higher-intensity dialysis did not result in improved survival or clinical benefits:

●In the United States VA/NIH ATN study, all 1124 patients were treated with IHD, CKRT, or PIKRT based upon hemodynamic status [51]. Patients were randomly assigned to one of two dosing arms:

•Intensive therapy – Hemodialysis and PIKRT were given six times per week with a target Kt/V of 1.2 to 1.4 per treatment (median delivered dose of 1.3 per treatment), while CKRT was provided with an effluent flow rate of 35 mL/kg per hour.

•Less intensive therapy – Hemodialysis and PIKRT were given three times per week with a target Kt/V of 1.2 to 1.4 per treatment (median delivered dose of 1.3 per treatment), while CKRT was provided with a flow rate of 20 mL/kg per hour.

The death rate at day 60 was the same for both groups (53.6 percent with intensive therapy and 51.5 percent with less intensive therapy). In addition, the duration of KRT and the rates of recovery of kidney function or other organ failure were similar for both treatment arms. The group that received intensive therapy had an increased number of hypotensive episodes. Thus, more intensive therapy beyond that obtained with a standard thrice-weekly regimen (with a target Kt/V of 1.2 to 1.4 per treatment) or standard CKRT (with an effluent flow rate of 20 mL/kg per hour) does not improve clinical outcomes.

●In the Randomized Evaluation of Normal versus Augmented Level of Replacement Therapy (RENAL) study (a trial in Australia and New Zealand), 1508 patients with AKI were randomly assigned to continuous venovenous hemodiafiltration (CVVHDF) at an effluent flow of either 25 or 40 mL/kg/hour [63]. At 90 days, mortality was the same in each group (44.7 percent, odds ratio [OR] 1.00, 95% CI 0.31-1.23). In addition, the incidence of patients who continued to receive KRT at 90 days was similar with both dialysis doses (6.8 and 4.4 percent of higher- and lower-intensity groups [OR 1.59, 95% CI 0.86-2.92]).

●Three meta-analyses (3841 patients and 8 trials, 3999 patients and 12 trials, and 2402 patients and 5 trials) found that more intensive therapy did not improve survival compared with less intensive regimens [60,61]. There was significant trial heterogeneity in all meta-analyses.

●Two additional studies evaluating even higher doses of continuous venovenous hemofiltration (CVVH; 70 to 80 mL/kg/hour) in patients with severe sepsis [38] or shock after cardiac surgery [9] found no benefit associated with further augmentation of KRT dose.

Observational studies have suggested that the actual delivered effluent volume during CKRT is substantially less than the prescribed dose. In the Dose Response Multicentre International Collaborative Initiative (DO-RE-MI) study of 338 patients treated with CKRT, for example, the median delivered dose of CKRT was 27 mL/kg/hour despite a median prescribed dose of 34.3 mL/kg/hour [64]. In addition, the actual time on therapy each day in both the ATN and RENAL studies probably exceeded the time on therapy achieved in clinical practice due to enhanced attention to minimizing interruptions in therapy. We therefore suggest that the prescribed dose exceed the desired delivered dose by a factor of approximately 20 to 25 percent to adjust for interruptions in study therapy.

In a patient-level meta-analysis of seven randomized, controlled trials comparing more intensive with less intensive KRT in AKI, more intensive KRT was associated with delayed recovery of kidney function, particularly in patients in whom the initial modality of KRT was CKRT, but was not associated with mortality [65]. More intensive therapy is also associated with a higher rate of electrolyte disturbances, such as hypophosphatemia, and has been associated with prolongation of mechanical ventilator support [66].

DISCONTINUATION OF THERAPY — KRT is usually continued until the patient manifests evidence of recovery of kidney function. Most often, recovery is assessed based on empiric data. In oliguric patients, the primary manifestation of recovery of kidney function is an increase in urine output; however, this finding may not be apparent in patients who are nonoliguric. Recovery of kidney function may also be manifest by a progressive decline in serum creatinine concentration after initial attainment of stable values (assessed daily during continuous kidney replacement therapy [CKRT] or predialysis in patients managed with intermittent hemodialysis [IHD]) despite a constant dose of KRT. More objective assessment of recovery of kidney function can be obtained by measurement of creatinine clearance. As an example, in the Acute Renal Failure Trial Network (ATN) study, creatinine clearance was assessed on six-hour timed urine collections obtained when the urine output exceeded 30 mL/hour [51]. Since the serum creatinine concentration may not be constant during the collection, the average concentration can be estimated by measuring serum creatinine at the beginning and end of the timed collection or based on the midpoint serum creatinine concentration. A precise level of kidney function needed to allow discontinuation of KRT has not been established; however, a creatinine clearance <12 mL/min is probably inadequate to allow discontinuation of therapy. In the ATN study, KRT was discontinued when the measured creatinine clearance exceeded 20 mL/min and was left to the discretion of providers when in the range of 12 to 20 mL/min [51].

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Acute kidney injury in adults".)

SUMMARY AND RECOMMENDATIONS

●Urgent indications – Urgent indications for kidney replacement therapy (KRT) in patients with acute kidney injury (AKI) include volume overload refractory to diuretics, severe hyperkalemia (serum potassium >6.5 mEq/L), severe metabolic acidosis (pH <7.1) despite medical management, uremia, and toxic overdose of a dialyzable drug. (See 'Urgent indications' above.)

●Timing of elective initiation – Even if one of the urgent indications is not satisfied, we electively initiate KRT in patients with AKI that is unlikely to resolve quickly and who, in addition, have one or more of the following (see 'Timing of elective initiation' above):

•Serum potassium >6.0 mEq/L despite medical management, or >5.5 mEq/L if there is ongoing tissue breakdown (eg, rhabdomyolysis, crush injury, tumor lysis syndrome), or ongoing potassium absorption (eg, due to severe gastrointestinal bleeding).

•Severe metabolic acidosis (pH <7.15) without reversible cause (eg, ketoacidosis) and despite optimal medical management (eg, intravenous sodium bicarbonate therapy as volume status permits). The benefit of KRT in patients with severe metabolic acidosis due to lactic acidosis is controversial.

•Hypervolemic patients who remain in persistent positive fluid balance despite aggressive attempts at diuresis, particularly if oxygen requirements are increasing.

We recommend against early initiation of KRT (ie, initiation before the patient develops an indication) (Grade 1B), since early initiation results in increased health care utilization but does not improve, and may worsen, outcomes in patients with severe AKI. On the other hand, excessive delay in initiation of KRT does not provide greater benefits and may be associated with an increased risk of adverse outcomes.

●Optimal modality – Data do not support the superiority of either continuous kidney replacement therapy (CKRT) or intermittent hemodialysis (IHD). Thus, the selection of modality of KRT should be based upon local expertise and experience in combination with the needs of the individual patient. If CKRT is administered, we recommend the use of venovenous circuits rather than arteriovenous circuits (Grade 1B). (See 'Optimal modality' above.)

●Optimal dosing – We suggest the following strategies for dosing of KRT:

•We recommend that IHD be provided on a three times per week schedule (alternate days), with monitoring of the delivered dose of dialysis to ensure delivery of a Kt/V of ≥1.2 per treatment, corresponding to a urea reduction ratio ≥0.67 (Grade 1B). More frequent treatments may be necessary for volume management or if the target delivered dose of therapy cannot be achieved. (See 'Intermittent hemodialysis' above.)

•We recommend that CKRT be provided with a delivered effluent flow rate (sum of hemofiltration rate and dialysate flow rate) of ≥20 mL/kg/hour (Grade 1B). In order to ensure delivery of this flow rate, we prescribe an effluent flow rate of approximately 25 mL/kg/hour. (See 'Continuous kidney replacement therapy' above.)