Prescription of continuous kidney replacement therapy in acute kidney injury in adults

INTRODUCTION — Kidney replacement therapy (KRT) is commonly required in patients with severe acute kidney injury (AKI). Acute KRTs include intermittent hemodialysis, peritoneal dialysis, continuous kidney replacement therapies (CKRTs), and hybrid therapies such as prolonged intermittent kidney replacement therapies (PIKRTs).

CKRT is widely used for acute KRT in the intensive care units (ICUs) in resource-abundant nations.

This topic provides recommendations for the prescription of CKRT. Anticoagulation and medication dosing are discussed elsewhere. (See "Anticoagulation for continuous kidney replacement therapy" and "Drug removal in continuous kidney replacement therapy".)

Acute intermittent hemodialysis, peritoneal dialysis, and PIKRT are discussed elsewhere. (See "Acute hemodialysis prescription" and "Use of peritoneal dialysis (PD) for the treatment of acute kidney injury (AKI) in adults" and "Kidney replacement therapy (dialysis) in acute kidney injury in adults: Indications, timing, and dialysis dose" and "Prolonged intermittent kidney replacement therapy".)

IMPLEMENTATION OF CKRT PROGRAM — CKRT is a complex intervention that is applied to complicated, critically ill patients [1,2]. The provision of CKRT should be coordinated among multiple disciplines including critical care medicine, nephrology, nursing, pharmacy, and nutrition support teams [1,2].

We believe that institutions should adhere to a written protocol developed by a panel of multidisciplinary experts within the institution. The protocol should address the key components of the CKRT prescription including vascular access, anticoagulation, CKRT mode, dose, and CKRT solutions. Making programmatic decisions regarding these components will foster consistency and improve quality.

While there is a nearly limitless combination of modes (convection, diffusion, both), machines, circuits, effluent flow rates, fluid removal strategies, and vascular access devices, having multiple individual approaches between CKRT prescribers leads to wide variability in practice and increases the possibility for errors in medication dosing, anticoagulation strategy, machine set-up, and nursing operations.

Furthermore, ideally, institutions are encouraged to develop and monitor CKRT quality indicators that track outcomes such as CKRT circuit survival duration, small-solute clearance, bleeding events, interruptions to treatment and downtime (ie, time during which CKRT is not being delivered to the patient), catheter dysfunction, fluid management, and patient mortality. While there are no data to date that such a tracking program improves patient outcomes, there are data that tracking programs improve CKRT quality [3,4].

VASCULAR ACCESS — A well-functioning vascular access is critical to maintaining CKRT circuit function since CKRT will not work adequately with a suboptimal access. (See 'CKRT blood flow rate' below.)

The preferred insertion site, proper catheter size, configuration, length, and depth and proper technique for placement are addressed in depth elsewhere. (See "Central venous catheters for acute and chronic hemodialysis access and their management".)

Deeper catheters that can be inserted into the larger central veins ending at the right atrium or caval-atrial junction (ie, internal jugular catheters) or the abdominal inferior vena cava (ie, femoral catheters) improve CKRT circuit performance and are preferred [5].

Some clinicians use specialized triple-lumen dialysis catheters for CKRT, although this is not our preferred option for most patients. While triple-lumen catheters may be effective in some patients, the third lumen decreases the internal diameter of the two dialysis lumens, which could limit blood flows.

If a triple-lumen catheter is used, the third lumen should not be used for lifesaving medications (such as antimicrobials) while CKRT is ongoing. The third lumen should only be used for medications without CKRT-related clearance (ie, intravenous crystalloid or colloid boluses or infusions, systemic heparin infusion, blood product administration, etc). It is theoretically possible that medications infused via the third lumen may be more rapidly cleared by dialysis than if the medication was administered via a different access. However, there are no published data on the magnitude of first-pass clearance resulting from recirculation between the medication infusion lumen and the dialysis outflow lumen.

We also do not use the third lumen to infuse calcium in patients receiving citrate anticoagulation. In our experience, this has been associated with recirculation, inaccurate calcium values, and less effective regional citrate anticoagulation. (See "Anticoagulation for continuous kidney replacement therapy", section on 'Preferential use of regional citrate anticoagulation'.)

For patients with end-stage kidney disease (ESKD), we do not use the arteriovenous fistula (AVF) or arteriovenous graft (AVG) unless no other access is possible, although there have been published data in such patients [6]. There is a potential risk of vascular damage to the access with rigid needles, and needle dislodgement could cause blood loss, especially with required patient repositioning in the intensive care unit (ICU). However, AVF and AVGs can be safely used if performing prolonged intermittent kidney replacement therapy (PIKRT) with either a CKRT machine or traditional dialysis platform. (See "Continuous kidney replacement therapy in acute kidney injury", section on 'Vascular access'.)

HEMOFILTER — The hemofilter (also referred to as hemodialyzer) is the same device for all CKRT modalities. High-permeability, high-flux biocompatible membranes are used for all modalities of CKRT. The typical membrane materials used are polyacrylonitrile (AN69), polyarylethersulfone (PAES), and polyethersulfone (PES). There are no data suggesting that one type of membrane is better.

Theoretically, because of their negative charge, polyacrylonitrile membranes may allow more adsorption and removal of middle-molecular-weight solutes, such as cytokines. However, no difference in outcomes has been demonstrated.

The polyacrylonitrile membranes can cause bradykinin release. An untreated AN69 membrane should not be used in patients with recent or ongoing angiotensin-converting enzyme (ACE) inhibitor use, as this has been reported to cause anaphylaxis [7] (see "Biochemical mechanisms involved in blood-hemodialysis membrane interactions", section on 'Blood-membrane interactions'). However, AN69 surface-treated (ST) membranes can be used safely with these medications. The AN69 ST membrane is coated with a polycationic saline solution (PEI), which reduces the surface electronegativity and avoids bradykinin generation [8]. The AN69 ST membrane is not available in the US [8]. (See "Reactions to the hemodialysis membrane".)

CKRT PRESCRIPTION — The prescription includes the choice of CKRT modality, method of anticoagulation (if this is used), filtration fraction, blood flow rate, dose, CKRT replacement or dialysis solution, and the fluid removal rate.

CKRT modality — CKRT modalities include continuous venovenous hemofiltration (CVVH), continuous venovenous hemodialysis (CVVHD), and continuous venovenous hemodiafiltration (CVVHDF).

The modalities are distinguished by their underlying mechanism of solute removal. CVVH utilizes convection, whereas CVVHD utilizes diffusion. CVVHDF uses a combination of both convection and diffusion. (See "Continuous kidney replacement therapy in acute kidney injury", section on 'Definition of CKRT modality'.)

We use CVVHD or CVVHDF rather than CVVH because diffusive therapies are able to deliver a higher KRT dose without driving up the filtration fraction to unacceptably high values. When blood flow, hematocrit, and total effluent flow rates are held constant, purely convective modes of therapy (such as CVVH) always have a higher filtration fraction compared with diffusive therapies.

The filtration fraction is the fraction of plasma water that is removed from blood during ultrafiltration. High filtration fractions are associated with increased circuit clotting [9]. (See 'Filtration fraction' below.)

However, many centers use CVVH, and the modality is often selected based on the CKRT device availability at a given institution. There are no studies that suggest better clinical outcomes with a specific modality. In a meta-analysis of 19 randomized, controlled trials that compared hemofiltration with hemodialysis for patients with acute kidney injury (including 16 studies of continuous therapies), there was no difference in survival, dialysis dependence, organ dysfunction, or vasopressor use among survivors [10]. (See "Continuous kidney replacement therapy in acute kidney injury", section on 'Definition of CKRT modality'.)

Theoretically, pure convective therapies (CVVH) and, to some degree, combined convective and diffusive therapies (CVVHDF) may remove larger-sized molecular-weight solutes slightly better than dialytic or diffusive therapies (CVVHD), although the clinical importance of this is uncertain [1,11-13].

Anticoagulation — Common options for anticoagulation include regional citrate anticoagulation (RCA) and systemic unfractionated heparin [14]. Many clinicians use no anticoagulation, at least initially. This issue is discussed elsewhere. (See "Anticoagulation for continuous kidney replacement therapy".)

CKRT dose — The dose is defined by the effluent flow rate. The effluent is the waste fluid that comes out of the outflow port of the dialysate/ultrafiltrate compartment of the hemofilter [15]. For patients on CVVH, the effluent consists of ultrafiltration volume (ie, plasma water moved by convection across the hemofilter membrane). For patients on CVVHD, the effluent consists mostly of spent dialysate and, to a smaller degree, ultrafiltration volume generated by convection.

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. Exceptions are patients with severe metabolic derangements (such as hyperkalemia or acidemia) that require more urgent correction over 24 to 36 hours. For such patients, we start with a higher CKRT dose [2]. For example, among some patients with severe metabolic acidemia (pH <7.1), we start with an effluent flow rate of 35 to 60 mL/kg/hour until acidosis is partially corrected.

Once the severe metabolic derangements are improved, we decrease the prescribed CKRT dose to approximately 25 mL/kg/hour. This is consistent with the 2012 Kidney Disease: Improving Global Outcomes (KDIGO) guidelines [14].

Large clinical trials have not shown a benefit of higher CKRT dose (>35 to 40 mL/kg/hour) compared with standard CKRT dose (20 to 25 mL/kg/hour) when applied over the entire duration of a patient's CKRT course (ie, days to weeks) [16,17]. In addition, higher doses of CKRT, particularly if prolonged, may lead to protein malnutrition, severe deficiency of many vitamins and micronutrients, and inadequate antimicrobial drug levels [16,17].

Among patients who start with a higher CKRT dose, it is difficult to define an exact threshold of laboratory abnormalities at which the CKRT dose may be decreased. Usually, the decision to reduce dose depends on stability or improvement of laboratory values (ie, the potassium remains stable and within normal parameters; pH and bicarbonate are normal or near normal and blood urea nitrogen is steadily declining).

Filtration fraction — The filtration fraction is the proportion of plasma water entering the dialyzer that is moved by ultrafiltration (convection) across the dialysis membrane. Stated differently, as defined above, it is the fraction of water that is removed from blood. We maintain a filtration fraction <20 percent. Higher fractions are associated with increased circuit clotting, presumably from hemoconcentration and blood protein-membrane interactions within the hemofilter [9].

Filtration fraction is arithmetically defined as follows:

The ultrafiltration flow rate is the rate at which plasma water is transferred across the membrane driven by a pressure gradient between blood and dialysate/ultrafiltrate compartments.

The total ultrafiltration flow rate equals the sum of the replacement fluid flow rate(s) and the patient fluid removal rate set on the machine. Notably, this definition of ultrafiltration rate is different from that used in intermittent hemodialysis, where it refers only to the rate/volume at which fluid is removed from the patient.

The plasma water flow rate is the rate at which plasma water is delivered to the dialyzer or hemofilter. It is equal to:

The blood flow rate  x  (1 - Hematocrit) + the prefilter replacement fluid flow rate + any other prepump infusion rate (such as citrate)

A relatively low filtration fraction can be maintained by:

●Keeping the ultrafiltration flow (convection) rate low

●Increasing blood flow rate (which determines plasma water flow rate), providing catheter function can support higher flows

●Using prefilter replacement fluid in CVVH or CVVHDF

Keeping the ultrafiltration rate low may require the addition of a diffusive component of clearance, especially when a relatively high effluent flow rate (ie, >2 L/hour) is required based on patients' weights or clinical needs (see 'CKRT modality' above). In pure convective therapies, such as CVVH, solute movement across the membrane is dependent upon the same force that drives ultrafiltration (ie, transmembrane pressure). In order to maintain adequate solute movement, transmembrane pressure and, thus, ultrafiltration must be relatively high. In diffusive therapies, solute movement is largely independent of the ultrafiltration rate because solute movement is driven by passive diffusion down concentration gradients and not by the transmembrane pressure. The ultrafiltration rate may be kept relatively low and still maintain solute movement.

Since CVVHDF uses a combination of diffusion and convection, the filtration fraction lies between CVVH and CVVHD, depending on the relative contribution of convection and diffusion to the total effluent flow rate. In general, for a fixed effluent rate, the filtration fraction with CVVHDF is less than that with CVVH since a portion of the dose is provided by diffusive therapy.

The use of prefilter replacement fluid in CVVH or CVVHDF will help to maintain a lower filtration fraction since it increases the plasma water flow rate, at least compared with postfilter replacement fluid. However, this is usually not sufficient to prevent hemoconcentration and clotting of the hemofilter. Furthermore, prefilter replacement dilutes the blood and will decrease small-solute clearance compared with postfilter replacement fluid.

CKRT blood flow rate — For patients who are on anticoagulation, we try to maintain a blood flow rate of 200 mL/min (although blood flow rates of <200 mL/min are used frequently with regional citrate anticoagulation). Among patients who are not on anticoagulation, a higher blood flow rate (250 to 300 mL/min) may be required to maintain catheter patency and CKRT circuit life. However, at least one randomized study has shown that there is no difference in circuit failure rates between a blood flow rate of 150 and 250 mL/min [18].

Filtration fraction is inversely proportional to the blood flow. Consequently, low blood flow rates (<100 to 150 mL/min) can increase hemofilter and circuit failures due to stasis of blood and an increase in filtration fraction.

Blood flow rates higher than 250 to 300 mL/min may decrease the circuit lifespan if the vascular access cannot support these higher blood flow rates over many hours or days, as is common in critically ill patients. Poor catheter performance results in increased access and return pressure alarms, temporary blood pump stoppage, blood stasis, and increased frequency of circuit clotting. Additionally, with some CKRT manufacturers, CKRT hemofilter life is limited both by time and by the volume of blood processed. The maximum volume of blood processed is attained sooner with higher blood flow rates.

A change in the blood flow rate between 100 and 300 mL/min usually does not affect solute clearance. Solute clearance can be limited by either the blood flow rate or the effluent flow rate. Since the blood flow rate is almost always much greater than effluent flow rate, solute clearance is usually effluent flow rate limited. The exception is when the effluent flow rate is greater or equal to blood flow rate.

In general, for CVVHD, the blood flow rate should be ≥2.5 times the dialysate flow rate. This allows for complete saturation of the dialysate and preserves the linear relationship between dialysate rate and small-solute clearance. When using CVVH with postfilter replacement fluid, the blood flow rate should be ≥5 times the replacement fluid rate to optimize the filtration fraction. When using CVVH with prefilter replacement fluid, the blood flow rate should be ≥6 times the replacement fluid rate to optimize the solute clearance.

Among patients who are anticoagulated with RCA, faster blood flows also increase the amount of required citrate. This increases cost (as one has to purchase more citrate) and the risk of complications, as more citrate will enter the patient's systemic circulation. (See "Anticoagulation for continuous kidney replacement therapy", section on 'Preferential use of regional citrate anticoagulation'.)

Finally, the blood flow rate does not affect hemodynamic stability, since the volume of blood in the circuit at any one time does not change as blood flow rate changes.

CKRT solutions — Solutions used for replacement fluid and dialysate can either be custom compounded on site or purchased from commercial vendors. We prefer not to customize the CKRT solutions, in order to reduce risks associated with compounding [19,20].

There are multiple commercially available CKRT solutions with variable concentrations of electrolytes and glucose (table 1):

●Sodium ─ The sodium concentration in commercially available solutions ranges from 130 to 140 mEq/L. For most patients, the sodium concentration in CKRT solutions should be physiologic (ie, 135 to 140 mEq/L).

A lower sodium (ie, 130 mEq/L) may be used for patients receiving RCA anticoagulation in order to prevent hypernatremia since the infused citrate solution may be hypertonic. (See "Anticoagulation for continuous kidney replacement therapy", section on 'Preferential use of regional citrate anticoagulation'.)

●Potassium ─ The potassium concentration in standard solutions ranges from 0 to 4 mEq/L. We use a 4 mEq/L potassium concentration for all patients except those with severe hyperkalemia (defined as >6 mEq/L in the absence of electrocardiogram [ECG] changes or elevated to any degree in the setting of ECG changes consistent with hyperkalemia).

Either a 0 or 2 mEq potassium solution may be used to treat severe hyperkalemia, depending on which solution is available (many institutions stock a 4 mEq/L potassium solution and either, but not both, a 0 or 2 mEq/L potassium solution). For patients on CVVHDF, a 2 mEq K solution may be generated on the CKRT device by using both a 4 K solution and 0 K solution delivered at the same rate. For example, if one uses a 4 K prefilter replacement solution at a rate of 1200 mL/hour and a 0 K solution as dialysate at a rate of 1200 mL/hour, then the effective K concentration of the CKRT circuit is 2 mEq/L.

However, we emphasize that intermittent hemodialysis rather than CKRT is indicated for the treatment of severe hyperkalemia (ie, with ECG changes such as worsening peaked T-waves or QRS prolongation refractory to calcium supplementation), even if the patient requires vasopressors. Even with the highest effluent rates possible with the CKRT, bulk potassium removal per minute is much more efficient with standard intermittent hemodialysis. (See "Continuous kidney replacement therapy in acute kidney injury", section on 'Indications'.)

●Bicarbonate ─ We, and most other centers, use bicarbonate- rather than lactate-based CKRT solutions. Serum lactate levels are often higher when lactate-based solutions are used, particularly among patients with liver failure, and may confuse the clinical interpretation of blood lactate levels. However lactate-based solutions have been used effectively among patients with a baseline lactate level <4 mmol/L.

Standard solutions have a bicarbonate concentration ranging from 22 to 35 mEq/L. We use a "high" bicarbonate solution (ie, 32 to 35 mEq/L) in all patients except those who are treated with RCA. Among patients treated with RCA, we use "normal" bicarbonate solution (ie, 22 to 25 mEq/L). Metabolic alkalosis is a common side effect of RCA since administered citrate is converted to bicarbonate. (See "Anticoagulation for continuous kidney replacement therapy", section on 'Preferential use of regional citrate anticoagulation'.)

●Phosphate ─ Standard solutions contain either no phosphorus or 1 mmol/L phosphorus. We use phosphorus-containing solution if the serum phosphate is <4.5 mg/dL and phosphorus-free solution in all other patients. Small studies confirm our clinical experience that using phosphorus-containing CKRT solutions minimizes the risk of developing severe hypophosphatemia [21,22].

●Glucose ─ Standard solutions are either glucose free or contain between 100 to 110 mg/dL glucose. We, and most clinicians, use a solution with 100 mg/dL of glucose. Some clinicians have suggested using glucose-free solutions in order to improve glucose control among hyperglycemic patients. Additionally, phosphorus-containing CKRT solutions are glucose/dextrose free. When using glucose-free solutions, there is a small risk of hypoglycemia and/or euglycemic diabetic ketoacidosis.

●Calcium ─ Standard solutions are calcium free or contain 2.5 to 3.5 mEq/L calcium. We use calcium-free solution if RCA is used. We use a maximum calcium concentration of 2.5 mEq/L if the solution contains phosphorus. (See "Anticoagulation for continuous kidney replacement therapy", section on 'Regional citrate anticoagulation'.)

Patient fluid removal rate — We target an hourly fluid balance ranging from net even to net negative 150 to 200 mL/hour, thereby yielding a 24-hour fluid balance of even to negative 2 to 4 L. Occasionally, we suspend patient fluid removal completely and allow an hourly net positive fluid balance; this is especially common in situations where there are high ongoing sources of nonurine, non-CKRT fluid losses such as from surgical wounds/drains, severe burns, active hemorrhage, etc.

The net goal and the rate of fluid removal are determined by the clinical condition. Almost all patients requiring CKRT have some degree of volume overload [23,24], which contributes to mortality and morbidity [23-30]. However, hemodynamic status usually dictates the rate of fluid removal, and clinical discretion is critically important [31]. We believe it is most effective for the collaborative team (ie, nephrology team plus intensive care unit [ICU] team) to target a goal hourly fluid balance. The bedside ICU nurse then adjusts the CKRT fluid removal rate as needed to achieve the hourly fluid balance goals.

As fluid is removed and a net negative fluid balance is achieved, clinicians should monitor a patient's hemodynamic status and slow or suspend fluid removal if there are signs of intolerance (ie, decreasing cardiac output, a progressive and concerning increase in vasopressor needs, etc).

In some patients, fluid removal may need to be prioritized over mild increases in vasopressor needs or prolonging vasopressor exposure, such as those with severe acute respiratory distress syndrome (ARDS) complicated by fluid overload, as an example. (See "Acute respiratory distress syndrome: Fluid management, pharmacotherapy, and supportive care in adults", section on 'Conservative fluid management'.)

Bedside tools (such as pulse pressure or stroke volume variability, bedside echocardiography, and inferior vena cava volume) may be used to assist in volume assessment if questions develop regarding the pace of fluid removal. (See "Novel tools for hemodynamic monitoring in critically ill patients with shock".)

Observational data indicate that the relationship between net fluid removal rate and mortality has a U-shaped curve, with the lowest mortality associated with net ultrafiltration rates of 1.0 to 1.75 mL/kg/hour [32,33]. However, it should be noted that these studies only evaluate the set fluid removal rate on the CKRT machine rather than the net fluid balance and have numerous competing contributions to higher mortality rates. As an example, patients requiring higher daily intake (ie, continuous infusions, bolus medications) will require a higher set CKRT machine fluid removal rate to achieve a net even fluid balance, and, generally, those critically ill patients requiring more medications and higher infusion rates have higher acuity of illness (ie, Sequential Organ Failure Assessment scores) and higher mortality rates.

LABORATORY MONITORING — We monitor electrolytes and acid-base status every 6 to 12 hours initially. If the patient remains stable with minimal changes in electrolytes at 24 to 48 hours, measurements of electrolytes can be decreased to every 12 to 24 hours.

When using regional citrate anticoagulation (RCA), more frequent monitoring is required. (See "Anticoagulation for continuous kidney replacement therapy", section on 'Monitoring'.)

COMPLICATIONS — Complications of CKRT include electrolyte, mineral, and acid-base imbalances; hypotension; infection; bleeding; and hypothermia [34]. Subtherapeutic antibiotic concentrations are a frequent (and often unrecognized) complication of CKRT; thus, careful attention to antimicrobial dosing is required in those on CKRT. (See "Drug removal in continuous kidney replacement therapy".)

Electrolyte, mineral, and acid-base imbalances — Laboratory abnormalities include hypophosphatemia, hypokalemia, hypomagnesemia, and, less commonly, hypocalcemia.

Hypophosphatemia, hypokalemia, and hypomagnesemia are the most commonly cited laboratory complications [16,35]. In general, if the electrolyte or mineral concentration is not in physiologic range in the CKRT solution, loss of that electrolyte or mineral in the effluent will occur and will require repletion. The more physiologic the electrolyte concentrations are in the CKRT solutions, the less likely replacement will be needed.

●Hypophosphatemia ─ Hypophosphatemia is common and increases with the effluent flow rate [16,17]. In one randomized trial, hypophosphatemia occurred in >50 percent [16]. Hypophosphatemia on CKRT is associated with prolonged respiratory failure [36].

Hypophosphatemia can be avoided by using phosphate-containing dialysate and replacement fluid (see 'CKRT solutions' above). If hypophosphatemia develops despite the use of phosphate-containing solution, parenteral phosphorus supplementation may be required. Higher doses of parenteral phosphorus are usually required compared with patients who are not on CKRT since phosphorus is generally administered over four to six hours and ongoing CKRT removes a significant amount of the prescribed dose during that period. (See "Hypophosphatemia: Evaluation and treatment", section on 'Intravenous dosing'.)

●Hypokalemia ─ Hypokalemia is a common complication of CKRT. In one randomized trial, hypokalemia occurred in >23 percent of patients undergoing CKRT [16]. The risk increases with the effluent rate and with composition of the CKRT solution. In a randomized trial that compared intensive (ie, high effluent rate) to less intensive (lower effluent rate) KRT including continuous therapies, hypokalemia was present in 7.5 and 4.5 percent of intensive and less intensive KRT, respectively [17].

The risk of hypokalemia also depends on patient-specific factors such as nutrition (ie, composition of total parenteral nutrition and tube feeds, etc) and clinical status.

Hypokalemia can be minimized if CKRT replacement and/or dialysate solutions contain potassium 4 mEq/L. If hypokalemia develops despite the use of 4 mEq/L solution, we replete with intravenous potassium as for any other patient not on CKRT. (See "Clinical manifestations and treatment of hypokalemia in adults", section on 'Treatment'.)

Some clinicians may add potassium to the solutions. We prefer to use commercially available standardized solutions to avoid compounding errors, however. (See 'CKRT solutions' above.)

●Alkalosis ─ Citrate anticoagulation can cause either metabolic alkalosis or metabolic acidosis [37]. Metabolic alkalosis can occur in patients with adequate hepatic function and muscle perfusion, who are able to metabolize systemic citrate to bicarbonate [38]. Metabolic acidosis can occur in patients with acute hepatic failure or severe shock, who are not able to sufficiently metabolize accumulating systemic citrate. (See "Anticoagulation for continuous kidney replacement therapy", section on 'Regional citrate anticoagulation'.)

●Hypomagnesemia ─ Hypomagnesemia is commonly observed among patients on CKRT [34]. We treat hypomagnesemia with intravenous magnesium as in any other patient not on CKRT. Some clinicians may add magnesium to the CKRT solution.

●Hypernatremia ─ Hypernatremia may develop in patients on RCA if the CKRT solution contains standard amount of sodium (ie, 140 mEq/L). Among such patients, we use a solution with a lower sodium concentration of 130 mEq/L.

●Hypocalcemia ─ Hypocalcemia is uncommon unless citrate is being used for anticoagulation and dialysis or replacement fluids do not contain calcium. For patients specifically on citrate requiring a calcium infusion, calcium abnormalities are corrected by adjusting the calcium infusion. (See "Anticoagulation for continuous kidney replacement therapy", section on 'Regional citrate anticoagulation' and "Anticoagulation for continuous kidney replacement therapy", section on 'Other complications'.)

Hypotension — Generally, hypotension is less common with CKRT than with intermittent hemodialysis [34], although, in one randomized trial, hypotension occurred at similar rates in patients treated with CKRT versus hemodialysis (35 versus 39 percent, respectively) [39].

The net ultrafiltration rate determines the risk of hypotension. Hypotension occurs when the fluid removal rate exceeds the rate at which the intravascular space can be refilled. Patients with diabetic neuropathy, decreased ventricular ejection fraction, diastolic dysfunction, or sepsis are especially vulnerable to hypotension because refilling capacity (ie, the rate at which the intravascular space is refilled) is decreased. The patient's clinical condition and hemodynamic stability should be closely followed and the ultrafiltration rate adjusted throughout treatment in order to prevent and/or address hypotension. (See 'Patient fluid removal rate' above.)

The rate of small-solute removal also affects in vivo fluid shifts independent of fluid removal. The rapid removal of small solutes may lead to a relative decrease in tonicity of the blood compared with the extravascular interstitial fluid, which leads to transient shifting of intravascular water to the extravascular compartments.

Hypothermia — Hypothermia may occur as a result of prolonged blood circulation in an extracorporeal circuit [40]. In a randomized trial, hypothermia occurred in 17 percent of patients undergoing CKRT compared with 5 percent of those treated with intermittent hemodialysis [34]. CKRT-induced hypothermia may mask the presence of fever. Blood warmers or external warming devices may be used to prevent excessive cooling [40].

Infection and bleeding — Infection and bleeding are well-recognized complications of KRT vascular access and are discussed elsewhere.

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

●General principles – The key components of the continuous kidney replacement therapy (CKRT) prescription include the modality, anticoagulation, dose, filtration fraction, blood flow rate, selection of CKRT solutions, and fluid removal rate. For many of these components, programmatic or institutional decisions will foster consistency and improve quality. (See 'Implementation of CKRT program' above.)

●Vascular access – A well-functioning vascular access is important to maintaining CKRT circuit function. We generally do not use triple-lumen dialysis catheters for CKRT, since the third lumen may decrease the internal diameter of the two dialysis lumens, limiting blood flows. If a triple-lumen catheter is used, the third lumen should not be used for lifesaving medications (such as antimicrobials) while CKRT is ongoing. Among patients with end-stage kidney disease (ESKD), we also do not use the arteriovenous (AV) access unless no other access is possible. (See 'Vascular access' above.)

●CKRT prescription

•Modality – CKRT modalities are distinguished by their underlying mechanism of solute removal (ie, convection versus diffusion). The specific CKRT modality should be selected based on the CKRT device availability and expertise at a given institution. We generally use diffusive therapies, particularly for larger patients who require a higher dose of CKRT. Diffusive therapies are more likely to provide an adequate dialysis dose (which is weight based) while maintaining the filtration fraction within an acceptable limit. (See 'CKRT modality' above.)

•Filtration fraction – Filtration fraction is defined as the ratio of ultrafiltration flow rate to plasma water flow rate. We maintain a filtration fraction <20 to 25 percent. Higher filtration fractions may increase the risk of circuit clotting, presumably from hemoconcentration and blood protein-membrane interactions. (See 'Filtration fraction' above.)

•Blood flow rate – For patients who are on anticoagulation, we try to maintain a blood flow rate of 200 mL/min. A higher blood flow rate (200 to 300 mL/min) is required if anticoagulation is not used in order to maintain catheter patency and CKRT circuit life. (See 'CKRT blood flow rate' above.)

•CKRT dose – The CKRT dose is expressed as effluent flow rate. We usually target an effluent flow rate of 25 to 30 mL/kg/hour in order to achieve a delivered flow rate of 20 to 25 mL/kg/hour. A benefit has not been shown with higher CKRT doses. An exception is among patients with severe electrolyte or acid-base disturbances who may require a higher dose, at least initially. (See 'CKRT dose' above.)

•CKRT solutions – Numerous commercially prepared CKRT solutions are available. We do not customize the CKRT solutions, due to the risk of error in compounding. The solution is selected based on laboratory values, which are checked frequently (table 1). (See 'CKRT solutions' above and 'Laboratory monitoring' above.)

•Patient fluid removal rate – We target an hourly fluid balance ranging from net even to net negative 150 to 200 mL/hour as the clinical condition dictates, yielding a 24-hour fluid balance of even to negative 2 to 4 L. The net fluid removal may need to be decreased if there are signs of intolerance (ie, decreasing cardiac output, concerning increase in vasopressor needs, etc). (See 'Patient fluid removal rate' above.)