Principles of Ultrafiltration

Achieving fluid balance by removal is the most frequently requested application for dialytic intervention and is considered the simplest form of continuous therapy. Fluid is drawn from the blood space across a semipermeable membrane. The fluid removed, or ultrafiltrate, has the characteristics of plasma water. With knowledge of the sieving coefficients of a particular membrane for various solutes, the ultrafiltrate can be used to determine the composition of serum and can help avoid an excessive number of blood draws.

In prescribing UF, a specific volume of fluid loss should be determined, with the UF rate (Q_F) set to achieve that loss over a set time frame. It cannot be overemphasized that this form of therapy is by nature slow. The steady, constant loss of fluid at a rate that does not exceed the plasma-refilling rate gives this form of therapy its hemodynamic stability. If extremely rapid UF in a short time frame is the therapeutic intent, intermittent forms of pump-driven UF are more efficient and therefore the treatment of choice.

The artificial membranes usually employed have a high UF coefficient (K_Uf), allowing water to pass quite freely. Any pressure difference between the blood side and the ultrafiltrate side of the filter results in fluid passage. Higher pressures in the blood compartment of the filter result in net fluid flow from the blood to the ultrafiltrate compartment. This flow is enhanced by applying negative pressures to the ultrafiltrate compartment through gravity or by pumped mechanical suction. This pressure should be held constant and not be subject to rapid variations. If transmembrane pressures are too high, membrane rupture and blood loss may result.

Common UF rates range from 100 to 400 mL/hour. Larger amounts may be obtained if there is a need for rapid fluid removal. Automated continuous machines control UF through a volume-driven system, establishing a fixed loss of a determined amount of fluid from the system over a given time period (usually on an hourly basis).

The blood flow rate (Q_B) has a significant effect on any UF system. During UF, plasma water is removed from the blood as it moves through the dialysis filter, thereby increasing the viscosity of blood in the filter. If the viscosity increases excessively, the system will clot. Therefore, whenever UF is being prescribed, an appropriate Q_B must be prescribed in order to accommodate the prescribed UF.

Careful attention should be given to the amount of access recirculation. Rates greater than 15% are associated with a greater incidence of clotting. This tendency is more evident at higher UF rates (>300 mL/hour), at which the returning blood tends to have a higher hematocrit, creating a more viscous, afferent blood flow.

Principles of Hemofiltration

As more fluid removal is performed, additional fluid may be needed to replace that which is lost. This is termed hemofiltration or plasma water exchange. Fluid replacement can be delivered into the blood circuit either prefilter, before UF has occurred (predilutional hemofiltration), or after fluid has been removed by the filter (postdilutional hemofiltration).

In predilutional hemofiltration, ultrafiltrated fluid reflects the mixture of blood and replacement solution. To use ultrafiltrate as a surrogate for blood sampling (see previous discussion) would be problematic because correction has to be made for the degree of dilution and for the electrolytic composition of the replacement fluid. Because the oncotic pressure of blood is reduced within the filter, a greater rate of fluid removal is possible at the same transmembrane hydrostatic pressure. This increased rate of fluid removal is offset by dilution of plasma solute. In other words, the overall mass transfer of uremic toxins is reduced, and higher rates of fluid exchange are required to compensate. It is common to have exchange rates of 30 to 40 L/day. Predilutional hemofiltration may also be used for patients with high hematocrit levels in an effort to reduce clotting episodes.

Postdilutional fluid replacement has the advantage of being easier to perform, with lower rates of fluid exchange compared with the predilutional system. One problem with this form of replacement is the increased oncotic pressures at the venous end of the filter. With high rates of exchange or high degrees of access recirculation, blood viscosity may be increased to the extent that clotting may occur. Therefore, fluid exchange rates may be dictated by such factors as hematocrit, blood flow, and access recirculation. Table 18.3 compares predilutional and postdilutional CVVH.

The end point of hemofiltrative therapy is determined by the balance between solutes removed with the ultrafiltrate and those replaced with the substitution fluid. A blood pump is used to increase blood flow and allow higher filtration rates. Infusion pumps are used for the delivery of replacement solution.

The system should be kept below a filtration fraction of 15% in postdilution and 30% in predilution hemofiltration for efficient operation and a lower risk of clotting.

Principles of Hemodialysis

In hemodialysis (HD), the flow of dialysate is countercurrent to that of blood to maximize transmembrane concentration differences across all blood concentrations and at all levels of the filter. Blood flow (100 to 300 mL per minute) is maintained well above the usual dialysate flow rates (15 to 30 mL per minute). By contrast, in IHD, blood, rather than dialysate flow, is the limiting factor in diffusive clearance. Clearance of low-molecular-weight substances (e.g., urea, creatinine) is “flow-dependent” because there is little resistance to transmembrane movement posed by the porous membrane. Substances of larger molecular weight (e.g., β₂-microglobulin, vitamin B₁₂) are relatively slow in crossing the dialyzer membrane and are “membrane-dependent” molecules. Using the high-flux membrane characteristics generally employed with continuous therapies, substances with molecular weights (masses) of 20,000 to 30,000 D are transferred at rates that have an inverse relationship to their molecular weights.

Electrolytes, urea, and creatinine easily cross membranes at a rate that is directly proportional to membrane surface area, temperature, and concentration difference, and inversely proportional to viscosity, distance from the membrane, and molecular size. Changing the concentration of various elements in the dialysate alters solute balance. Balance is achieved, however, only between transferable particles. Protein-bound solutes are not subject to the concentration gradients that drive the molecular transport across the membrane. This concept is the basis for altered drug kinetics when patients are subjected to continuous supportive therapy.

In CVVHD, dialysate flow rates remain the most influential factor in determining urea clearance.¹⁷ Dialysate usually is delivered via pumps at rates of 15 to 40 mL per minute. Given adequate blood flow through the circuit, one can see why the limiting factor for flow-dependent transfer is the relatively low dialysate flow rate. Blood flow and filter membrane have limited effects on the diffusion of molecules compared with the potential of dialysate flow changes.

In addition to the diaysate inflow, the dialysate outflow also must be controlled. By setting the outflow rate higher than dialysate inflow rates, one can create a negative transmembrane pressure promoting UF across the dialysis membrane. This difference in flow is used to establish the rate of UF. An increase or decrease in this flow difference would increase or decrease the rate of fluid loss. Dialysate flow is external to and independent of blood flow, and therefore, one may see continued dialysate flow in a system with virtually no blood flow. Although decreases in hemofiltration flow rates may indicate system clotting, dialysate flow rate changes have no predictive value for clotting and may continue despite blood-side occlusion.

Continuous Renal Replacement Therapy Versus Intermittent Therapy

The clinical presentation and circumstances may favor either intermittent or continuous therapies (Table 18.4). There are many theoretical benefits that may favor the use of cRRT over intermittent forms of therapy, such as improved hemodynamic stability, faster resolution of fluid overload, and increased dialysis dose delivery (Table 18.5). Box 18.1 lists some nonrenal indications for using cRRT. However, clinical trials demonstrating an evidence-based assessment of these potential advantages are still lacking.

Hemodynamic stability is one of the most important advantages for the use of cRRT over intermittent modalities. Slower removal of solute and fluid from the intravascular space by continuous techniques should allow adequate time for refilling from the interstitium and intracellular space, theoretically minimizing therapy-induced hypotension. There are longer term implications for renal recovery, with IHD-related hemodynamic instability potentially predisposing to recurrent renal injury. The data from rigorous, comparative studies seem to lead to varying conclusions, however.

Continuous therapies are able to deliver a higher dose of clearance compared to intermittent therapies. The concept of dose will be covered later in this chapter. Intermittent therapy may have to be provided at a high frequency to produce equivalent levels of solute removal.

Definitive data to support many of the suspected advantages of cRRT are still lacking. In fact, Box 18.2 lists some of the disadvantages of cRRT. Interpretation of much of the published data has been hampered by retrospective analysis, the use of historical control groups, incomplete randomization, incomplete descriptions of patient populations and dialysis dose delivery, and study group–control group heterogeneity.

In the absence of a solid evidence base, how should one decide between prescribing continuous or intermittent therapy? The basic indications for delivering renal support remain unchanged (Box 18.3) and range from the most frequent request for fluid balance to more esoteric requirements such as toxin removal. Common considerations in choosing to apply intermittent or continuous support are listed in Table 18.3, being mindful of numerous relative advantages and disadvantages of each (see Table 18.5).

Hybrid therapies that combine cRRT and IHD techniques are described in the literature as extended daily dialysis, SLED, or prolonged daily intermittent RRT. These therapies use standard IHD equipment to apply lower solute clearances and UF rates for prolonged periods and aim to combine the desirable features of each modality—reduced rate of UF for improved hemodynamic stability, low efficiency solute removal to minimize solute disequilibrium, longer treatment duration to achieve prescribed dialysis dose, and intermittency for the convenience of diagnostic and therapeutic procedures during scheduled downtime.

Sustained low-efficiency dialysis is typically performed with low blood flows of about 200 mL per minute and dialysate flows of 100 to 300 mL per minute. Clearance is predominantly diffusive; however, available systems may also combine diffusive and convective clearance via HDF. Overall, hybrid therapy provides a high dose of dialysis with minimal urea disequilibrium and good control of electrolytes, with survival similar to that predicted by a variety of illness severity scores.

Continuous Hemodialysis Versus Continuous Hemofiltration

Deciding between the forms of dialytic therapy is dependent on a number of factors: rate of catabolism, mean arterial pressure with resultant blood flow rates, UF rates required or desired, and access choice. The patient’s primary diagnosis also influences which modality is chosen, as some data suggest an improved outcome with the use of hemofiltration in some patients with multiple organ failure. Therapy effect on such vastly differing substances as urea and creatinine, molecules of “middle” molecular weight (β₂-microglobulin, vitamin B₁₂), cytokines (tumor necrosis factor-α, interleukins), and hormones (endothelin, angiotensin) is the subject of intense research on potential differences in therapy choice and its influence on patient outcome.

Assuming a flawless period of therapy delivered by all modalities, urea clearance would have a great influence on therapy choice. Circuit variations, anticoagulation, “downtime” of the differing techniques, clotting frequency, and blood and dialysate flow rates all should be considered in the choice. Interruptions to therapy decrease overall treatment effectiveness.

Blood urea nitrogen and creatinine are frequently used as indicators of the need for renal support. Urea or creatinine appearance rates can be used as a gauge of the quantity of therapy required. Patients with high urea generation rates should receive HD. The only influence on urea removal in hemofiltration is fluid exchange rates; the increased need for fluid removal carries with it a need for greater differences in fluid exchanges, limiting the replacement volume. Given a maximal UF rate dictated by the system’s blood flow rate (maximal filtration fraction) and by the patient’s hematocrit level, solute clearances are restricted and may be inadequate for that particular patient’s needs. Selection of HD allows for the UF rates needed without compromising solute clearance because dialysate flow rates are generally not as limited.

The influence of the dialysate rate on dialyzer clearance is depicted in Figure 18.7. The higher the rate, the more effective the urea clearance, until one reaches a rate approximating the effective plasma flow of the system. The stability of continuous therapies lies in their low rates of exchange over a longer period. The increase in clearance rates must be balanced against the desire to provide stable therapy. If one desires a high rate of removal in a short time, then an intermittent therapy should be employed.

A drawback to diffusive methods of delivery is the relatively low removal rates of the higher-molecular-weight substances. As noted earlier, convection provides greater clearance for substances 5000 to 20,000 D. HD may have limitations when the goal of therapy is targeted toward these larger molecules. This may be the basis for the early reports of improved outcome with patients who were subjected to HDF techniques. The suggested advantage comes from the removal of cytokines and the influence on endotoxin adsorption.

Figure 18.7 depicts an approach to determining the modality of cRRT based on the patient presentation.

Dose

Dose in RRT refers to how much of a measure of the quantity of a representative marker solute is removed from a patient. The concept of dose is used to gauge the adequacy of a given treatment. Urea clearance has been the standard molecular marker for IHD. Urea is chosen as an easily measured surrogate for low-molecular-weight products of metabolism. Measuring total urea in the effluent fluid and continuous plasma urea concentration could allow calculation of clearance. However, this is cumbersome and approximations of dose are instead estimated from flow rates of dialysate.

The current recommendation for a minimum dose of RRT supports the delivery of at least 20 mL/kg per hour of CVVH, CVVHD, or CVVHDF. Usually this will require a prescribed dose of 25 to 30 mL/kg per hour (if expecting treatment downtime or other interruptions in treatment, see “Prescribed Versus Delivered Dose”). Although higher doses of dialysis might be beneficial in selected patients, there is no evidence at this time of any benefit for a higher dose.

Other approaches for determining the dose of dialysis include using the clearance of molecules other than urea. Biochemical parameters such as the correction of electrolyte disturbances, the clearance of larger middle-weight molecules (such as β₂-microglobulin), normalized protein catabolism ratio (nPCR), the anion gap, or the strong ion gap have been suggested as a “marker.” Clinical parameters of measuring dialysis include fluid balance, improvement and respiratory function, and nutritional markers. However, all of these markers remain investigational.

Prescribed Versus Delivered Dose

It is recommended that the prescription of dialysis should exceed that which is calculated to be adequate because there is a difference between what is prescribed and what is actually delivered. In a number of cases the delivered dose of dialysis is much smaller than what is prescribed.18,19 This can be due to interruptions in treatment due to hemodynamic instability, patient testing, circuit clotting, access recirculation, or other reasons why a patient may need to be disconnected from a machine.

Circuits and Material

Fluids

Solutions of fluid take the form of either dialysate or replacement fluid. When hemofiltration is performed, replacement fluid composition dictates the resultant concentration of electrolytes. Similarly, knowing the composition of dialysate is important to determine to what equilibrium the electrolytes will settle. The sieving coefficients for relevant electrolytes and blood components are listed in Figure 18.6. Elements with a negative charge that are small enough to cross the membrane do so at greater than unity. This apparent active transport is actually accomplished through a dynamic process similar to the Gibbs-Donnan effect seen in stagnant fluid balance. Because negatively charged proteins are unable to cross the membrane, chloride and bicarbonate move against a concentration gradient to maintain electrical neutrality. The exaggerated loss of these elements must be reflected in the replacement solution used.

Calcium and magnesium levels should be monitored closely, and replacement should be initiated at an early stage. Because therapy is effective in removing phosphate and generally fluids do not contain phosphate, deficiencies develop, requiring supplementation. In general, these electrolytes are monitored every 12 hours.

As a general rule, to avoid undue hemoconcentration, the filtration rate for postdilutional hemofiltration should be no more than 15% of plasma flow through the filter. A 30% filtration fraction may be allowed in predilutional techniques.

Most fluids either contain lactate or bicarbonate at the base buffer. Other substances have also been considered, for example, the use of acetate or citrate. The use of citrate has an added advantage of being an anticoagulant. Acetate, although previously used extensively in long-term HD, may give rise to vasodilation, myocardial depression, and increased oxygen consumption.²⁴ As a result acetate should be avoided for cRRT.

Lactate undergoes hepatic conversion to bicarbonate on an equimolar basis. Although more stable than bicarbonate, there are numerous theoretical disadvantages to its use in cRRT. Although bicarbonate-based and lactate-based solutions can correct acidosis, bicarbonate is preferred because lactate-buffered fluids may require intravenous bicarbonate supplementation to achieve the same bicarbonate level.²⁵ The metabolism of exogenous lactate may be impaired in critical illness, with accumulation giving rise to a paradoxic metabolic acidosis.²⁶ Hyperlactatemia carries with it potential negative hemodynamic effects and metabolic complications, including increased protein catabolism and reduced adenosine triphosphate regeneration.²⁷ Complications of lactate overload are more likely to develop in patients with liver impairment and poor peripheral perfusion, particularly with high-volume treatment. Lactate intolerance is defined arbitrarily as a greater than 5 mmol/L increase in lactate levels during therapy.

Bicarbonate, although the more physiologic anion, also has specific drawbacks. It exists in solution with other ions in a state of equilibrium under specific physical conditions of temperature and pressure:

When CO₂ outgassing from the solution occurs (e.g., delivering bicarbonate using an open-top container), overall bicarbonate concentration may be reduced. Additionally, calcium and magnesium can precipitate out as insoluble carbonate compounds when sterilized with the buffer. Many bicarbonate-based solutions are produced with lower concentrations of both cations to help ameliorate this problem, with final mixing of the electrolyte and bicarbonate solutions just before use. A novel adaptation has electrolyte and bicarbonate solutions housed and sterilized in separate chambers of the same bag. A connecting valve is broken just before use to mix the two fluids. A final caveat to the use of bicarbonate is its apparent predilection to bacterial growth—at least in liquid bicarbonate concentrates used in long-term dialysis.

From the data available, although it seems that lactate-based substitution or dialysate fluids may be used safely in many patients, they should be avoided in patients with lactic acidosis or hyperlactatemia and in patients with hepatic failure. Bicarbonate-buffered solutions should be used in these cases.

Anticoagulation

Circuit clotting is the most frequent cause of therapy interruption in cRRT. Various factors may account for the hypercoagulable state in a patient receiving cRRT.²⁸ Anticoagulation during cRRT is necessary to preserve the life of the extracorporeal circuit, to maximize the cRRT dose, and to minimize blood loss caused by clotting during cRRT. The ideal anticoagulant should have no effect on systemic hemostasis, no increase in hemorrhagic risk, only be limited to the extracorporeal circuit, optimize filter performance in circuit life, have a short half-life, be easily monitored, be easily reversible, and be inexpensive.²⁹ Options for anticoagulation of the circuit include using no anticoagulation, unfractionated heparin, low-molecular-weight heparin, citrate, and in rare circumstances thrombin antagonists or prostaglandins. Box 18.4 lists current strategies to prevent circuit clotting. According to a world survey, 44% of those surveyed preferred unfractionated heparin for anticoagulation.³⁰

The option of no anticoagulation is usually used in patients with intrinsic coagulopathies such as hepatic failure or low platelet count. Circuits are usually primed with saline or heparin. Intermittent normal saline flushes may be used as well. The rates of filter clotting using this method vary widely. The mean filter life lies between 16 and 70 hours if the patient is coagulopathic. In cases of severe coagulopathy, shorter filter life may be seen. The disadvantages of no anticoagulation include the need for increased UF, the risk of dialyzer fiber rupture, and the extra nursing workload.³¹ The hemoconcentration induced by high UF volumes in CVVH and CVVHDF promotes clotting but may be minimized by predilutional fluid replacement. This comes at the price, however, of the inefficiency of ultrafiltering a mixture of just-infused replacement fluid and plasma, the proportions of which are important considerations in the cRRT prescription.

Unfractionated heparin works by inactivating factors Xa and IIa. The molecular weights of unfractionated heparins range from 5 to 30 kDa. The half-life of unfractionated heparin is 90 minutes; however, this may be increased up to 3 hours in renal failure. The use of unfractionated heparin involves infusing continuous heparin at the arterial site of the circuit. Usually a bolus of 2000 to 5000 IU of heparin is used. Continuous infusion of heparin ranges between 5 and 20 units/kg per hour. In general, an activated partial thromboplastin time (aPTT) goal ranges between 34 and 45 seconds (1.5-2.0 times the normal). The reported circuit patency of this ranges between 20 and 40 hours. The advantages of using unfractionated heparin include the fact that it is effective, widely available, involves simple monitoring (i.e., aPTT), is easily reversed with protamine, is inexpensive, and has a short half-life. The disadvantages of the use of unfractionated heparin include systemic bleeding, unpredictable kinetics, aPTT not being a reliable predictor for bleeding, heparin resistance due to low antithrombin levels, and heparin-induced thrombocytopenia (HIT).

HIT may develop during heparin therapy.³² HIT begins with heparin exposure stimulating the formation of heparin-platelet factor 4 antibodies, usually 5 to 12 days after starting therapy. This triggers the release of procoagulant platelet particles. Both thrombosis and thrombocytopenia ensue and cause significant vascular complications. The prevalence of HIT varies among several subgroups, with greater incidence in surgical as compared with medical populations. HIT must be acknowledged for its intense predilection for thrombosis and suspected whenever thrombosis occurs after heparin exposure. In cRRT, this is manifested as recurrent filter clotting. The treatment of HIT mandates the cessation of all heparin exposure and the institution of an antithrombotic therapy, most commonly using a direct thrombin inhibitor. Current “diagnostic” tests, which primarily include functional and antigenic assays, have more of a confirmatory than diagnostic role in the management of HIT. Platelet aggregation studies are highly specific but lack sensitivity, so if they and heparin-induced platelet activation tests yield negative results, an enzyme-linked immunosorbent assay should be performed. Direct thrombin inhibitors are appropriate, evidence-based alternatives to heparin in patients with a history of HIT.

Direct thrombin inhibitors such as recombinant hirudin (lepirudin), danaparoid, and argatroban have been used in response to HIT.33,34 Bleeding resulting from overdosage of lepirudin or argatroban may be treated with the administration of fresh-frozen plasma. Hemofiltration with high-flux dialyzers also can reduce the plasma levels of hirudin.

Citrate regional anticoagulation works by chelating free calcium in the extracorporeal circuit and prevents the activation of calcium-dependent procoagulants. The anticoagulant effect of citrate is measured by ionized calcium levels. Anticoagulation is reversed by a calcium infusion. Normal blood levels of citrate are approximately 0.05 mmol/L. The bleeding time for a patient with citrate levels of 4 to 6 mmol/L can be infinite. Levels of 12 to 15 mmol/L are required for storing blood products for transfusion therapy. Citrate levels are usually performed in another facility or reference laboratory. Citrate has a plasma half-life of 5 minutes and is rapidly metabolized by the liver, kidney, and muscle cells. The extracorporeal clearance of citrate is the same as that of urea. The sieving coefficient ranges between 0.87 and 1.0. The clearance of citrate is the same for both CVVH and CVVHD.³⁵ The advantage of using citrate is that it is regional and avoids bleeding complications. It also doubles as a buffer. It is more effective than heparin in studies,^36,37 and it does not cause thrombocytopenia. Two early trials have shown that the use of citrate for regional anticoagulation yields no additional bleeding risk and leads to longer filter life.^36,37 The largest citrate randomized controlled trial used citrate for anticoagulation in a postdilutional CVVH modality with blood flows of 220 mL per minute and citrate concentration of 3 mmol/L.³⁸ The study showed a lower mortality rate in patients receiving citrate regional anticoagulation. It is hypothesized that citrate inhibits leukocytes esterase, which may yield immunologic benefits. The major disadvantage of citrate stems from its metabolic complications and the complex protocols that are involved in its administration. Metabolic consequences of citrate use include metabolic alkalosis from citrate overdose or toxicity, metabolic acidosis in a setting of severe liver disease or hypoperfusion, hypernatremia from hyperosmolar citrate solutions (4% sodium citrate), and hypocalcemia and hypercalcemia from inappropriate calcium supplementation. The risk factors for citrate toxicity include liver disease, nursing or pharmacy errors leading to overdose, shock liver, and severe hypoperfusion states. The detection of citrate toxicity then becomes extremely important. One should expect citrate toxicity when noticing a rising anion gap, worsening metabolic acidosis, a falling systemic ionized calcium, escalating calcium infusion requirements, or a total calcium/systemic ionized calcium ratio greater than 2.5 : 1.³⁹ Strategies to manage citrate toxicity include decreasing the citrate infusion rate, decreasing the blood pump speed, and increasing the dialysate flow rate.

The decision as to which citrate protocol to use depends on the available citrate solutions, the method of citrate delivery, and the cRRT circuit options. Available commercial citrate solutions include concentrations of sodium citrate between 1.32% and 4%. Regional citrate anticoagulation may be performed using 4% trisodium citrate or with ACD-A solution (anticoagulant citrate dextrose solution, form A) containing 3% combined trisodium citrate (2.2 g/100 mL), citric acid (0.73 g/100 mL), and dextrose (2.45 g/100 mL) (Baxter-Fenwal Healthcare Corp., Deerfield, IL). ACD-A solution is preferred over trisodium citrate for routine regional citrate anticoagulation because it is less hypertonic and commercially prepared, potentially reducing mixing errors and the complications associated with overinfusion. Regional citrate anticoagulation protocols differ in the type of citrate preparation used, the mode of dialysis, and the ability to customize dialysis solutions. There is a fixed relationship between the blood flow in citrate delivery. The titration of citrate delivery should be based on the ionized calcium level. The amount of citrate delivered to achieve a blood citrate concentration of 4 mmol/L depends on the blood flow.⁴⁰

In a typical circuit (Fig. 18.10) arterial blood leading from the patient is first infused with citrate, which chelates the free ionized calcium. This blood then enters the filter where a calcium-free dialysate is used. Postfilter ionized calcium is monitored and used to titrate the citrate rate to ensure anticoagulation. The goal is to keep the ionized calcium at less than 0.35 mmol/L. The returning blood combines with the venous blood in the body, which normalizes the ionized calcium and prevents systemic anticoagulation. Calcium is infused through a separate central line to replace the calcium lost in the ultrafiltrate. Citrate is metabolized primarily in the liver to bicarbonate and the bound calcium is released. For CVVH, citrate may or may not be a component of the replacement fluid.

A number of citrate protocols have been published. Protocols vary based on the number of fluid solutions utilized and commercial versus hospital specific fluid options. Published protocols include those from Massachusetts General Hospital,⁴¹ Gainesville,⁴² University of Alabama at Birmingham,⁴³ Sunnybrook,⁴⁴ and San Diego.⁴⁵

Low-molecular-weight heparins (such as enoxaparin, dalteparin, and nadroparin) have also been examined for their effect on circuit life.⁴⁶ Prostacyclin (prostaglandin I₂), a potent, short-acting, endogenous inhibitor of platelet aggregation, has also been studied.⁴⁷

Clinical experience with various agents and strategies influence choice. Table 18.6 reviews the advantages and disadvantages of each anticoagulation option.

Patient Selection

The question as to whether a patient needs renal replacement therapies is often difficult. There is no consensus as to the indication to start RRT, the criteria to start RRT, or the appropriate time for initiation of RRT. The conventional indications include volume overload, metabolic acidosis, hyperkalemia, uremia, azotemia without uremic manifestations, and drug overdose. The risks of dialysis include hypotension, complications of placing vascular access, and air embolism (extremely rare). Additional risks and complications are listed in Box 18.5 Patients in the ICU may have a number of conditions that impact the decision to initiate and prescribe RRT. In general, if patients have insufficient renal function, cRRT can be utilized to provide organ support much in the way that the ventilator is used to provide pulmonary support. Potential indications for renal support include nutrition, fluid removal in CHF, cancer chemotherapy, the treatment of respiratory acidosis in acute respiratory distress syndrome (ARDS), and fluid management in multiorgan failure.⁴⁸

Timing of Initiation and Discontinuation

The use of the traditional biomarkers such as creatinine may delay the initiation of RRT. As a result, criteria such as RIFLE (risk, injury, failure, loss of kidney function, endstage renal disease) have been used to grade the severity of one’s AKI. It is known that the net fluid balance affects the outcomes of critically ill patients; fluid overload itself may be a marker of more severe disease. The difficulty in answering the question as to when to initiate RRT is that we will never know whether those who received dialysis would have recovered renal function if they had not received the therapy. There is no prospective randomized controlled trial to answer this question.

Dialysis is discontinued when the signs of renal recovery are apparent. This often takes the form of increased urine output, the decrease in biomarkers, or improvement in the patient’s clinical status. cRRT may be switched to intermittent forms when the patient is hemodynamically stable and daily fluid balance would afford intermittent treatments (usually no more than 2 L positive per day). However, as with the initiation of dialysis there is no consensus as to when the therapy should be discontinued.

Renal Recovery

Recovery of renal function may manifest quite differently, depending on whether the patient was supported with intermittent or continuous techniques. Patients receiving cRRT may not show the typical signs of renal recovery because often their fluid balance is close to their premorbid weight. A decrease in the baseline serum creatinine without a change in the cRRT prescription may be the first sign of renal improvement. Urine creatinine and sodium may be more helpful in assessing patients receiving continuous rather than intermittent therapy. Recognizing these differences helps avoid extending therapy beyond what is needed and aids in hastening recovery when it has begun.

Pharmacokinetics During Continuous Renal Replacement Therapy

Altered drug kinetics is an important aspect of cRRT management, with some agents being cleared in significant quantities. Depending on membrane pore size, the passage of substances with molecular weights of 20,000 to 30,000 D is possible and may accommodate most drugs. As the clearance of drugs is limited to the free, non–protein-bound fraction, the degree of plasma protein binding dictates whether dialysis or filtration would result in significant removal. Pharmacokinetics also is highly dependent on the drug’s volume of distribution. Also, if the contribution of alternative elimination pathways to overall drug clearance is significant, then the clinical relevance of extracorporeal removal may be minimal.

A drug with low protein binding, a low volume of distribution, and low clearance by alternative pathways is one that would be cleared significantly by cRRT. Vancomycin and the aminoglycosides are good examples, and such agents require dosage adjustments if administered during continuous therapy.

Body clearance of drugs without significant tubular secretion or reabsorption is a linear function of creatinine clearance. Estimation of drug clearance by dialytic modalities is a more complex proposition and has to take account of dialysate and blood flow rates; the molecular weight of the drug; and membrane surface area, thickness, and composition. Nomograms used in the prediction of dialytic drug clearance may be used,⁴⁹ but these provide only a general guideline. The impact of molecular weight differs between dialytic techniques and may be lower in CVVHD, which usually uses high flux membranes at a low dialysate flow rate (Q_D). Under these circumstances, the Q_D may approximate Q_F in the previous equations.⁵⁰ Drug clearance during CVVHDF may be estimated by combining calculations of filtrative removal with calculations of dialytic removal, but the complex interplay between convection and diffusion is not fully appreciated by this approach.

A final confounder of pharmacokinetic predictability is the impact of membrane adsorption of the drug, which may be substantial with PAN/AN69 materials and may vary depending on the frequency of filter changes. Such filters that remain in situ for long enough can start to release their adsorbed drug back into the circulation. In adjusting dosing of a drug that is significantly cleared by cRRT, a choice must be made either to shorten the dosing interval (to maintain plasma levels) or to increase the dose (to optimize peak concentrations) depending on that agent’s mode of action. Formulas have been developed to estimate adjustments of dose and dosing interval.50,51

Published data do exist to help guide drug dosing during cRRT, but they come from a variety of sources and are often not standardized. The use of aids to estimate dose, such as those detailed previously, is required.

Complications of Therapy

Perhaps the most frequent problems encountered are electrolyte derangements (Table 18.7). With the use of hemofiltration, as we have seen earlier, the replacement solution determines the final electrolyte outcome. The replacement solution must reflect the sodium, chloride, and bicarbonate concentrations that one would like to achieve in patient serum and the relative loss via the hemofilter. The substance’s particular sieving coefficient can be used to determine therapeutic losses, following the formula:

where [ ]_s is the incoming blood concentration of the substance, S_coef is its sieving coefficient, and Q_F is the UF rate (see Fig. 18.6). Not only the actual replacement fluid, but all fluid must be taken into consideration. Frequently, drug vehicles with water or hypotonic hyperalimentation solutions are calculated in the fluid exchange, but not in the electrolyte balance. The resultant loss of sodium (in the hemofiltrate) and the replacement with hypotonic solution produce a true hyponatremia. The exaggerated loss of bicarbonate and chloride also produces variations if these balances are not considered in the total composition delivered to the patient.

Ionized calcium, magnesium, and phosphate also are lost during continuous hemofiltration and HD. This is a different situation from that of intermittent therapies, where these substances are usually retained and need to be subject to a limited intake. It is not unusual for patients to require the addition of magnesium, calcium, and phosphate to cover therapeutic losses and establish normal plasma levels. Serum levels should be checked frequently (up to every 12 hours), and replacement should be calculated as noted earlier.

Continuous HD procedures also require frequent electrolyte monitoring. Serum values generally reflect the dialysate concentration of that particular solute. Establishing a potassium floor of 4 mEq/L merely requires that the dialysate concentration of potassium also be set at 4 mEq/L. Common electrolyte problems seen in patients receiving continuous therapies are listed in Table 18.7. Box 18.5 lists other possible complications.

Ethical Considerations

Bioethical issues related to cRRT have been widely discussed. Many times the use of dialysis involves the decision of tapering aggressive interventions. Oftentimes, the kidney appears to be the last organ in the chain of multiorgan failure. From the family’s perspective the use of dialysis is usually not a predefined life-sustaining measure. The use of dialysis is often a source of conflict of attitudes between the treating teams of the patient. The decision to initiate dialysis should be a discussion between the treating team, the one prescribing cRRT, and the patient and caregivers. This can either be a long-term treatment for a short-term test case in the setting of longstanding chronic disease or set an advance illness. The discontinuation or withdrawal many times is the decision of the entire team.

Nonrenal Application of Continuous Therapies

The hemodynamic stability of continuous renal therapies has contributed to their use in situations in which fluid loss is desired but patient status has restricted more standard dialytic or other therapeutic interventions. The ability to manipulate a continuous extracorporeal blood circuit also has opened wide opportunities to assess a variety of different nonrenal applications (Fig. 18.11). The basic circuit has been incorporated into liver assist devices (currently undergoing clinical testing), and the ability to either warm or cool circulating blood has applications in clinical and experimental medicine. The adsorptive nature of different membrane types and structures also has been the focus for potential therapeutic interventions.

Perhaps the most frequently cited area of use has been isolated UF for the treatment of CHF. Several studies have described the utility of UF for the removal of extravascular lung water in models subjected to pulmonary damage or fluid overload.

The sieving and adsorptive qualities of the continuous extracorporeal circuit have found their way into the manipulation and eventual management of patients with sepsis or systemic inflammatory response syndrome. Numerous experimental studies have touted the ability of various membranes either to remove or to adsorb various cytokines. Tumor necrosis factor-α, interleukin 1, interleukin 6, and interleukin 8 have been the most studied, but endothelin, lipopolysaccharide fragments, and C3a and C5a also have been identified. High-volume zero-balance hemofiltration has been the focus of several animal experiments, in which control of systemic inflammatory response syndrome has been noted.

Clinical translation of this approach consistently has failed to confirm bench findings for many reasons. Extracorporeal removal of inflammatory mediators may be negligible in relation to endogenous turnover. Extracorporeal treatment was usually initiated within a short time frame after induction of experimental sepsis; this is almost impossible to achieve in the real world. Potentially beneficial substances, such as interleukin 10, water-soluble vitamins, and elements such as zinc or selenium, also may be cleared. As a result, clinical application of these techniques has been viewed with some caution.

The adsorptive qualities of specific membranes (AN-69, PAN) have been the most accepted mode of clinical attempts at cytokine control. Although early data seem to point to an impact in lowering blood cytokine levels, a cumbersome and costly exchange of circuit filters, with saturation seen at 2 to 4 hours, brings into question the practicality of this approach.

1. Nickolas, TL, Barasch, J, Devarajan, P. Biomarkers in acute and chronic kidney disease. Curr Opin Nephrol Hypertens. 2008; 17(2):127–132.

2. Hou, SH, Bushinsky, DA, Wish, JB, et al. Hospital-acquired renal insufficiency: A prospective study. Am J Med. 1983; 74(2):243–248.

3. Shusterman, N, Strom, BL, Murray, TG, et al. Risk factors and outcome of hospital-acquired acute renal failure. Clinical epidemiologic study. Am J Med. 1987; 83(1):65–71.

4. Nash, K, Hafeez, A, Hou, S. Hospital-acquired renal insufficiency. Am J Kidney Dis. 2002; 39(5):930–936.

5. Uchino, S, Kellum, JA, Bellomo, R, et al. Acute renal failure in critically ill patients: A multinational, multicenter study. JAMA. 2005; 294(7):813–818.

6. Chertow, GM, Soroko, SH, Paganini, EP, et al. Mortality after acute renal failure: Models for prognostic stratification and risk adjustment. Kidney Int. 2006; 70(6):1120–1126.

7. Bellomo, R. The epidemiology of acute renal failure: 1975 versus 2005. Curr Opin Crit Care. 2006; 12(6):557–560.

8. Ympa, YP, Sakr, Y, Reinhart, K, Vincent, JL. Has mortality from acute renal failure decreased? A systematic review of the literature. Am J Med. 2005; 118(8):827–832.

9. Chertow, GM, Burdick, E, Honour, M, et al. Acute kidney injury, mortality, length of stay, and costs in hospitalized patients. J Am Soc Nephrol. 2005; 16(11):3365–3370.

10. Scribner, BH, Buri, R, Caner, JE, et al. The treatment of chronic uremia by means of intermittent hemodialysis: A preliminary report. Trans Am Soc Artif Intern Organs. 1960; 6:114–122.

11. Hoste, EA, Schurgers, M. Epidemiology of acute kidney injury: How big is the problem? Crit Care Med. 2008; 36(4 Suppl):S146–S151.

12. Ishani, A, Xue, JL, Himmelfarb, J, et al. Acute kidney injury increases risk of ESRD among elderly. J Am Soc Nephrol. 2009; 20(1):223–228.

13. Hoyt, DB. cRRT in the area of cost containment: Is it justified? Am J Kidney Dis. 1997; 30(5 Suppl 4):S102–S104.

14. De Vriese, AS, Colardyn, FA, Philippe, JJ, et al. Cytokine removal during continuous hemofiltration in septic patients. J Am Soc Nephrol. 1999; 10(4):846–853.

15. Page, B, Vieillard-Baron, A, Chergui, K, et al. Early veno-venous haemodiafiltration for sepsis-related multiple organ failure. Crit Care. 2005; 9(6):R755–763.

16. Parikh, A, Shaw, A. The economics of renal failure and kidney disease in critically ill patients. Crit Care Clin. 2012; 28(1):99–111.

17. Bellomo, R. Do we know the optimal dose for renal replacement therapy in the intensive care unit? Kidney Int. 2006; 70(7):1202–1204.

18. Venkataraman, R, Kellum, JA, Palevsky, P. Dosing patterns for continuous renal replacement therapy at a large academic medical center in the United States. J Crit Care. 2002; 17(4):246–250.

19. Vesconi, S, Cruz, DN, Fumagalli, R, et al. Delivered dose of renal replacement therapy and mortality in critically ill patients with acute kidney injury. Crit Care. 2009; 13(2):R57.

20. Tapson, JS, Hoenich, NA, Wilkinson, R, Ward, MK. Dual lumen subclavian catheters for haemodialysis. Int J Artif Organs. 1985; 8(4):195–200.

21. Vijayan, A. Vascular access for continuous renal replacement therapy. Semin Dial. 2009; 22(2):133–136.

22. Kim, IB, Fealy, N, Baldwin, I, Bellomo, R. Insertion side, body position and circuit life during continuous renal replacement therapy with femoral vein access. Blood Purif. 2011; 31(1-3):42–46.

23. Abidi, SM, Khan, A, Fried, LF, et al. Factors influencing function of temporary dialysis catheters. Clin Nephrol. 2000; 53(3):199–205.

24. Maccariello, E, Rocha, E, Dalboni, MA, et al. Customized bicarbonate buffered dialysate and replacement solutions for continuous renal replacement therapies: Effect of crystallization on the measured levels of electrolytes and buffer. Artif Organs. 2001; 25(11):870–875.

25. Kierdorf, HP, Leue, C, Arns, S. Lactate- or bicarbonate-buffered solutions in continuous extracorporeal renal replacement therapies. Kidney Int Suppl. 1999; 72:S32–S36.

26. Tan, HK, Uchino, S, Bellomo, R. The acid-base effects of continuous hemofiltration with lactate or bicarbonate buffered replacement fluids. Int J Artif Organs. 2003; 26(6):477–483.

27. Cole, L, Bellomo, R, Baldwin, I, et al. The impact of lactate-buffered high-volume hemofiltration on acid-base balance. Intensive Care Med. 2003; 29(7):1113–1120.

28. Bouman, CS, de Pont, AC, Meijers, JC, et al. The effects of continuous venovenous hemofiltration on coagulation activation. Crit Care. 2006; 10(5):R150.

29. Oudemans-van Straaten, HM, Fiaccadori, E, Baldwin, I. Anticoagulation for renal replacement therapy: Different methods to improve safety. Contrib Nephrol. 2010; 165:251–262.

30. Uchino, S, Bellomo, R, Morimatsu, H, et al. Continuous renal replacement therapy: A worldwide practice survey. The beginning and ending supportive therapy for the kidney (B. E. S. T. kidney) investigators. Intensive Care Med. 2007; 33(9):1563–1570.

31. Ramesh Prasad, GV, Palevsky, PM, Burr, R, et al. Factors affecting system clotting in continuous renal replacement therapy: Results of a randomized, controlled trial. Clin Nephrol. 2000; 53(1):55–60.

32. Hong, MS, Amanullah, AM. Heparin-induced thrombocytopenia: A practical review. Rev Cardiovasc Med. 2010; 11(1):13–25.

33. Gajra, A, Vajpayee, N, Smith, A, et al. Lepirudin for anticoagulation in patients with heparin-induced thrombocytopenia treated with continuous renal replacement therapy. Am J Hematol. 2007; 82(5):391–393.

34. Link, A, Girndt, M, Selejan, S, et al. Argatroban for anticoagulation in continuous renal replacement therapy. Crit Care Med. 2009; 37(1):105–110.

35. Chadha, V, Garg, U, Warady, BA, Alon, US. Citrate clearance in children receiving continuous venovenous renal replacement therapy. Pediatr Nephrol. 2002; 17(10):819–824.

36. Kutsogiannis, DJ, Gibney, RT, Stollery, D, Gao, J. Regional citrate versus systemic heparin anticoagulation for continuous renal replacement in critically ill patients. Kidney Int. 2005; 67(6):2361–2367.

37. Monchi, M, Berghmans, D, Ledoux, D, et al. Citrate vs. heparin for anticoagulation in continuous venovenous hemofiltration: A prospective randomized study. Intensive Care Med. 2004; 30(2):260–265.

38. Oudemans-van Straaten, HM, Bosman, RJ, Koopmans, M, et al. Citrate anticoagulation for continuous venovenous hemofiltration. Crit Care Med. 2009; 37(2):545–552.

39. Meier-Kriesche, HU, Gitomer, J, Finkel, K, DuBose, T. Increased total to ionized calcium ratio during continuous venovenous hemodialysis with regional citrate anticoagulation. Crit Care Med. 2001; 29(4):748–752.

40. Flanigan, MJ, Pillsbury, L, Sadewasser, G, Lim, VS. Regional hemodialysis anticoagulation: Hypertonic tri-sodium citrate or anticoagulant citrate dextrose-A. Am J Kidney Dis. 1996; 27(4):519–524.

41. Palsson, R, Niles, JL. Regional citrate anticoagulation in continuous venovenous hemofiltration in critically ill patients with a high risk of bleeding. Kidney Int. 1999; 55(5):1991–1997.

42. Munjal, S, Ejaz, AA. Regional citrate anticoagulation in continuous venovenous haemofiltration using commercial preparations. Nephrology (Carlton). 2006; 11(5):405–409.

43. Tolwani, AJ, Prendergast, MB, Speer, RR, et al. A practical citrate anticoagulation continuous venovenous hemodiafiltration protocol for metabolic control and high solute clearance. Clin J Am Soc Nephrol. 2006; 1(1):79–87.

44. Tobe, SW, Aujla, P, Walele, AA, et al. A novel regional citrate anticoagulation protocol for cRRT using only commercially available solutions. J Crit Care. 2003; 18(2):121–129.

45. Mehta, RL, McDonald, BR, Aguilar, MM, Ward, DM. Regional citrate anticoagulation for continuous arteriovenous hemodialysis in critically ill patients. Kidney Int. 1990; 38(5):976–981.

46. Tsang, DJ, Tuckfield, A, Macisaac, CM. Audit of safety and quality of the use of enoxaparin for anticoagulation in continuous renal replacement therapy. Crit Care Resusc. 2011; 13(1):24–27.

47. Davies, H, Leslie, G. Anticoagulation in cRRT: Agents and strategies in Australian ICUs. Aust Crit Care. 2007; 20(1):15–26.

48. Mehta, RL. Indications for dialysis in the ICU: Renal replacement vs. renal support. Blood Purif. 2001; 19(2):227–232.

49. Schetz, M. Drug dosing in continuous renal replacement therapy: General rules. Curr Opin Crit Care. 2007; 13(6):645–651.

50. Golper, TA. Update on drug sieving coefficients and dosing adjustments during continuous renal replacement therapies. Contrib Nephrol. 2001; 132:349–353.

51. Bohler, J, Donauer, J, Keller, F. Pharmacokinetic principles during continuous renal replacement therapy: Drugs and dosage. Kidney Int Suppl. 1999; 72:S24–S28.