Chapter 31: Renal Replacement Therapy

Almost 60 years ago, RRT was first used to treat acute kidney injury (AKI) during the Korean War.¹ Within 25 years, continuous therapies were first attempted in Germany.² RRT modalities have since expanded, becoming commonplace in most hospitals of the developed world. In AKI, when preventative and supportive management fails we turn to RRT.

In the past decade, admissions for AKI increased with a doubling in the incidence of severe AKI (Figure 31–1).³ In parallel, approximately 200,000 patients required RRT in 2012.³ The incidence of RRT-requiring AKI now surpasses that of end-stage renal disease (ESRD).⁴ What's more, the morbidity and mortality of AKI requiring RRT extends beyond hospitalization.^3,5

The necessity for RRT in the ICU arises in 1 of 3 clinical situations: AKI, critically ill ESRD patients,⁶ and drug or toxic overdoses. RRT for AKI is dramatic but less common than most think. The incidence of AKI in the ICU is relatively low between 6% and 19%, but can be higher depending on population studied and risk, injury, failure loss, and end-stage kidney disease (RIFLE) criteria.^7,8 RRT needs occur in up to 5% of AKI.^8,9 Unfortunately, mortality continues to remain high, between 24% and 50%.^3,8,10 AKI incrementally adds to short- and long-term mortality, and especially when RRT is required,^6,9 with up to 14% requiring chronic dialysis on discharge.^5,8,11

Prescribing RRT in the critically ill is complex and ideally should involve clear communication between nephrologist and intensivist.¹² The choice of modality, the goals, its timing, as well as patient-specific findings, are collectively crucial in the decision for RRT.¹² Understanding the shortcomings and function of the modalities must also be a part of the process. In centers where selection is limited to intermittent hemodialysis (IHD), adaptations are not only viable but presently adequate.¹³

Clearance of any substance depends on modality characteristics, (convection vs diffusion), size, time on RRT, flow rates (both blood and dialysate), and dialyzer membrane characteristics (Figure 31–2).¹⁴ This coupled with patient acuity, hemodynamics, nutritional status, volume status, and diagnosis determines eventual success.

RRT attempts to support the kidneys' responsibilities in maintaining homeostasis. To some extent it succeeds in volume control and to a lesser extent in solute and acid/base homeostasis. The various modalities, whether peritoneal or hemodialysis, utilize 2 transport mechanisms in providing renal replacement: diffusion and convection. These forces result in solute clearance and plasma water removal or UF. Other renal responsibilities such as calcium/phosphate balance and anemia control are not provided with RRT.

Clearance, removal of solute from plasma, is most efficiently accomplished by diffusion. Namely plasma is cleared of certain solutes across a semipermeable membrane (the dialyzer), as it travels down a concentration gradient. Each fluid medium (plasma and dialysate), flows in a direction counter current to the other, maintaining the concentration gradient between plasma and dialysate. Hence a plasma potassium of 7 mEq/L approaches that of the dialysate's 2 mEq/L over time. Dialysate can be adjusted to approximate the desired goal of net solute change. Clearance is determined by the solutes' serum concentration, molecular weight, and dialysis membrane characteristics (pore size and surface area). Diffusion becomes relatively inefficient as solute molecular weight increases, while with convection large-molecular-weight solute clearance is more effective.¹⁵ Diffusion is a key mechanism of clearance in RRT using dialysate such as IHD.

Convection, which is preferred in treating the critically ill, is capable of removing large quantities of “plasma” water or UF. Blood hydrostatic pressure in the dialyzer supports a transmembrane pressure, through which UF is generated. Accompanying UF is the clearance of solute via solvent drag, signifying that this fluid is not pure water but laden with solute. In fact, convection plays a significant role in providing clearance in continuous RRTs. UF rate is what determines clearance and is dependent on blood flow rate. Though inefficient in providing clearance, convention by adding the luxury of 24 hours surpasses what IHD can do in 4 (Figure 31–3).

An added benefit with convection is the enhanced clearance of larger molecules or middle molecules (500-5000 Da), felt to include uremic toxins. There are over 115 compounds identified as toxins that correlate more closely with uremic symptoms such as encephalopathy.¹⁶ Since middle molecule clearance depends on the duration of a plasma-dialysate interface, continuous RRTs are more effective. Though at present conjectural, middle molecule clearance may in time provide for an important element of treatment in AKI due to sepsis. For instance, evidence indicates that continuous modalities may enhance clearance of various mediators of inflammation (see later).

RRT indications include the critically ill ESRD patient, the critically ill AKI patient, and the toxic overloaded patient. Only the first 2 are actual renal replacement.

The Critically Ill Esrd Patient: Maintenance Dialysis

RRT for ESRD patients may be the most common reason (up to 40% of all patients receiving RRT) for its use in the ICU.^6,17 The etiology is most often related to ESRD comorbidities, primarily cardiovascular problems (31% of ICU ESRD cases) and most commonly arrhythmias and congestive heart failure.⁶ Admissions may result from missed dialysis and/or compliance issues resulting in volume or electrolyte imbalances. ICU monitoring provides ventilator support, or if hyperkalemic, telemetry for possible arrhythmias, or cardiogenic shock. After prompt RRT the patient is stabilized, often resulting in relatively short admissions.

ICU care may also be required for acute events unrelated to ESRD, such as sepsis (15% of ICU ESRD cases), trauma, or complications from surgery. They have twice the rate of sepsis and higher readmission rates than nonrenal failure patients.^6,17 Mortality is lower than AKI patients requiring RRT, but higher than the general population.

The Critically Ill Aki Patient: Bridge Rrt

In treating AKI, RRT is supportive, bridging the patient while treating the initiating event. About 7.5% of hospitalized patients with AKI require RRT.⁵ In the ICU, AKI is most often due to sepsis thought to be from ATN, though glomerular hemodynamics play a significant role.^8,18 The need for RRT is relatively uncommon in AKI when all RIFLE classes are included. It requires a process of involved decision making between nephrologist and intensivist: recognizing the patient's needs; weighing issues of hypotension and anticoagulation; selecting the proper RRT; and balancing them with realistic capabilities. This includes possibly avoiding RRT, waiting for renal recovery, or not even considering it due to futility.

Emergent RRT

Clear indications for immediate RRT include life-threatening hyperkalemia, severe metabolic acidosis, and volume overload; overt symptomatic manifestations of uremia such as pericarditis and encephalopathy; and intoxications (lithium, aspirin, ethylene glycol, methanol, metformin, amanita).

Adequate treatment of uremia through the clearance of uremic toxins is unclear since such toxins are not clinically measured.¹⁶ Since, for every small increment of creatinine there is a significant rise in mortality,¹⁹ uremic toxins may be clinically more important than realized. Unfortunately, we are limited to assuming uremia in the azotemic patient with a pericardial rub, or encephalopathy. Given the multiple reasons for mental status changes in the ICU, the encephalopathy is particularly difficult to discern and ideally RRT should be initiated well before uremic symptoms occur.

In uremic pericarditis, treatment is daily, continued RRT. When entertaining the diagnosis, it is important to evaluate and rule out, cardiac tamponade. Aggressive RRT without initial decompression may decrease preload and result in shock.

Early or urgent RRT

RRT is initiated early in patients whose renal function is not expected to quickly improve due to severity of illness: multiorgan failure, AKI unresponsive to resuscitation, high FENa, rising azotemia (without plateau of creatinine or urea levels), and oliguria all suggestive of ATN. In such “extreme” conditions the concept of early RRT to replace renal function and/or support multiorgan dysfunction may make most sense and benefit.²⁰ RRT maintains volume despite hemodynamic, nutritional, antibiotic, and other infusion needs, while providing balance of electrolytes and pH.

The decision for early RRT remains a difficult one especially since the benefits to mortality or otherwise is not clear.²¹ A meta-analysis of 30 RCTs and 8 prospective cohort studies was inconclusive,⁹ while a recent meta-analysis of 23 studies (1960-2006), suggested some benefit.²² Importantly, these trials differed considerably in clinical settings and definitions of early RRT. Many excluded patients that recovered or died without starting RRT, rendering a recommendation for early RRT difficult.

The question becomes more complicated prompting clinical discussion in less dramatic presentations, such as nonoliguric AKI,²³ and/or slow rising azotemia (risk or injury on the RIFLE scale). Since there is no key single marker to use as a starting point for RRT, the decision is based on the multiple facets of the presentation.

Traditional markers of AKI, serum creatinine and blood urea nitrogen (BUN) are limited in this regard, though, a lower mortality risk has been reported when starting RRT at lower BUN levels.^26,27 Urea generation (BUN levels) varies depending upon on protein intake and catabolism²⁶. Its highly variable volume of distribution (Vd) is exemplified by the BUN rebound seen in sepsis after RRT. What's more, dialysis efficacy using urea kinetics has not been assessed in either AKI or in the critically ill.¹⁴

Pursuing this question, the BEST study²⁸ separated azotemic patients by BUN, creatinine, and time in ICU. The answers were mixed. In the BUN-based analysis, there was no outcome difference in timing of RRT.²⁸ “Late” initiation based on creatinine had lower mortality while those initiated early based on ICU duration prior to RRT had lower mortality. These observations were confounded due to some having chronic kidney disease (CKD) and variability in patient presentations. However, the study did suggest that early RRT in patients with preexisting CKD may impart a higher survival advantage.²⁸

Another reported reason in delaying RRT is the perception of impending renal recovery.²⁵ RRT is invasive and can result in complications from access placement, hypotension, electrolyte abnormalities, and arrhythmias.^13,29 Hemodynamic instability is common in the ICU and the intensivist needs to balance RRT safety with its need. Hypotension during RRT, along with issues of dialyzer bioincompatibility may delay renal recovery.³⁰ Dialyzer membranes induce monocyte-derived proinflammatory cytokines (interleukin [IL] 1, 6, 8, tumor necrosis factor alpha [TNF-α]), which can increase renal toxicity. So-called biocompatible membranes may lessen these responses. Therefore when taken into account a prudent decision may be to wait and monitor volume, potassium, protein intake, and supplement bicarbonate as needed, selecting extended or continuous RRTs if condition worsens. Recently published AKIKI trial with early vs late initiations of RRT in critically ill patients suggest delaying initiation based on clinical indications may avoid the need for RRT. ELAIN trial suggests reduced 90 days mortality with early RRT initiation.^25,26

Finally, delays can also arise from issues of logistic support involving nursing and staff availability as well as their timeliness.²⁵

Fluid overload

In severe sepsis and AKI, positive fluid balance is independently associated with increased mortality.^31,32 Recent studies suggest fluid overload to be more critical indication for initiating RRT earlier. In both the pediatric and adult literature, fluid overloaded patients had better outcomes with early RRT even after severity of illness adjustments.^31,33 Furthermore, survivors had lower fluid accumulation at RRT initiation compared to nonsurvivors even with RRT modality and severity adjustments. Always influential, volume status may now provide a greater urgency for early RRT.

Fluid overload is also associated with more days on dialysis, and greater ICU and hospital length of stay.^28,34 Though adjusted for severity, it is not clear the difference in volumes suggest hemodynamically dissimilar patients or that volume should be applied more conservatively and/or ultrafiltrated more aggressively.

Cessation

Similar to commencement, no prespecified single finding or estimated glomerular filtration rate (eGFR) exists to trigger cessation. Actually, attempting to assess GFR is inaccurate when renal function is not in steady state. As patient acuity improves, the first decision is to transfer to IHD. In deciding, it is important to consider if IHD can support present and daily volume load. Moreover, drug dosing must be adjusted to IHD or level of renal function.

In general, RRT is terminated when urine output increases and creatinine is stable or decreases. The return of renal function is often preceded by an improvement in the patient's overall condition. Continuing may not only be unnecessary but any RRT, and possibly more so IHD, may delay renal recovery from hypotension and/or continued inflammation. Conversely, continued RRT may be of benefit for maximizing uremic toxin clearance.³⁵ The answer is not known.

The coupling of adequate solute and UF control with satisfactory patient outcomes are the goals of RRT. Practical issues that include nursing scheduling and cost are coupled to the patient's presentation when selecting the RRT modality (Table 31–1).

Peritoneal dialysis

In North America, peritoneal dialysis (PD) was extensively used into the 1990s to treat AKI in the ICU.³⁶ PD requires placement of an intra-abdominal catheter, with continuous exchanges of high dextrose dialysate into the abdominal cavity. PD, therefore, is a continuous RRT. The high osmolarity dextrose generates UF, and hemodynamically it tends to fare better than IHD. Catheters were temporarily placed at the bedside by nephrologists, or Tenckhoff catheters were placed by surgeons. In the following decades, acute PD was largely abandoned due to inefficient solute clearance, infection concerns, and its contraindication in patients with abdominal problems or recent surgery. PD may also be difficult in ventilated patients or in severe obstructive lung disease.

Recently, the modality has been reassessed in the ICU setting, where the use of high-volume PD was examined.³⁷ One study which excluded severely hypercatabolic patients, had few complications (7.5%) and 12% had peritonitis. In certain critically ill patients, PD may be a viable alternative,³⁷ where its characteristic continuous slow clearance may better suit patients with elevated ICP.³⁸

Intermittent hemodialysis

This is essentially conventional outpatient dialysis averaging 4 hours, 3 times per week, with blood flow rates (BFR) up to 400 cc/min. Due to the higher BFR and dialysate flow rates (DFRs) it is more likely to induce hypotension than other modalities. Therefore, the traditional arguments against using IHD in the ICU have been relatively less hemodynamic stability and clearance compared to continuous RRT. Furthermore, IHD is also a concern in situations where rapid changes in solute (sodium or potassium) or osmoles (elevated ICP) can be detrimental (see below). Financially, the use of roller blood pumps and rapid clearance requires skilled dialysis nursing, increasing costs.

Earlier studies demonstrated IHD a hemodynamic risk compared to continuous RRT (CRRT). Since then, advancements such as in-line bicarbonate baths and biocompatible membranes have improved IHD stability.^30,39 Presently, IHD is the most commonly selected RRT in American ICUs, utilized in 57% of cases.⁴⁰

Furthermore, in the severely ill, nephrologists can improvise IHD by slowing BFR, increasing hours and/or increasing its use to 6 d/wk. This increases net clearance while maintaining hemodynamic stability, essentially becoming sustained low-efficiency daily dialysis (SLEDD) (see later). What's more, adding a separate UF procedure (blood through the dialyzer without counter current dialysate, excising diffusion), facilitates volume removal with less hypotension. UF is better tolerated when solute and osmole clearance is not concomitant.

In the critically ill requiring the use of vasopressor agents, IHD may still be attempted by increasing the dose of the vasopressor as needed. The suggested guidelines in Table 31–2 can largely be met by most institutions.

Continuous Renal Replacement Therapy

Continuous modalities (see Table 31–1) are considered in hemodynamically unstable patients less likely to tolerate abrupt fluid shifts associated with IHD. CRRT is distinguished by lower BFR, allowing for greater stability in treating fluid overload. Indeed, fluid accumulation is more likely seen with IHD than CRRT.³³ Also, acidosis and volume control are more consistent with CRRT. Both convective and diffusion modes of solute clearance can be utilized depending on CRRT type. Surprisingly, outcome benefits of CRRT over IHD have not been demonstrated.

Patients for which CRRT is preferred include those with hypotension as in severe sepsis, cirrhosis, liver transplant, and congestive heart failure (CHF). It also benefits patients with markedly elevated BUN and/or ICP due to gradual solute removal avoiding disequilibrium.

As implied, CRRT is administered continuously for as long as required. In actuality, due to interruptions for diagnostic and/or therapeutic interventions treatment time is closer to 18 hours. The long duration on the CRRT circuit cools the blood, resulting in reflexive vasoconstriction supporting hemodynamics.⁴¹ CRRT can be used in operating rooms during prolonged procedures such as liver transplantation, and requires frequent hemodynamic, laboratory and electrolyte assessments.

Continuous Arteriovenous Hemofiltration

Arteriovenous modalities (continuous arteriovenous hemofiltration [CAVH] and continuous arteriovenous hemodialysis [CAVHD]) required placement of catheters in an artery and central vein. This system utilized the patient's arterial pressure rather than a rolling pump for blood flow. CAVH is rarely used since the advent of continuous venovenous hemofiltration (CVVH) due to more invasive arterial access, variable BFR, and increased filter clotting issues.

CVVH

Essentially through convection, clearance is dependent on large UF rates. To maintain hemodynamic and electrolyte stability large volumes (between 1 and 3 L/h) of electrolyte solutions are necessary. Frequent close attention to volume and electrolyte status is essential. Effluent flow is reported as mL/kg/h. The difference between infused volume and UF generated is the net fluid balance. Recent studies have not demonstrated superiority of a particular flow over another.²⁹

REPLACEMENT FLUIDS:—Replacement fluids (RFs) are provided as prefilled sterilized bags by the equipment manufacturer. They have varying concentrations of sodium, magnesium, calcium, and potassium, and allow for modification. Either bicarbonate or lactate is the alkali source. Ideally, lactate-based RF should be avoided in liver failure and/or lactic acidosis, and periodic lactate measurements are needed with bicarbonate replacement when higher than 5 mmol/L.⁴² Bicarbonate has become the buffer of choice despite issues with storage and preparation.

The RF with the added clearance of solute has effects on blood composition and requires periodic monitoring of blood chemistries. It is added either before the dialyzer (predilution/inflow) or after blood has passed through the dialyzer (postdilution/outflow).

Predilution RF lowers (dilutes) solute concentration and theoretically decreases CVVH clearance efficiency. The UF is not 100% saturated with urea having been diluted by RF prior to passage through the dialyzer/filter. Advantages lie in requirements of low BFR and less clotting of circuit also due to the dilutional effect of the RF.

In postdilution outflow mode, UF rate should not be more than 20% of BFR to avoid excess hemoconcentration and clotting of the circuit. This requires higher BFR of 150 to 200 mL/min to achieve adequate fluid removal in excess of 25 L/d. Predilution inflow overcomes this problem at the expense of clearance efficiency but overall better treatment due to less clotting and more fluid processed for UF generation.⁴³ As an example, 35 L of daily replacement fluid addition in predilution/inflow mode dilutes the blood by 15% at a BFR of 140 mL/min. (35 L/d = 24 mL/min; 24/140 + 24 = 15%).

Continuous Venovenous Hemodiafiltration

Diffusion is now introduced enhancing clearance, by running dialysate countercurrent to blood flow in the extracorporeal circuit. As a combined modality, it requires large volumes of RF in maintaining hemodynamic and electrolyte stability. Again, close attention to the volume and electrolyte status is essential. Comparing CVVH to continuous venovenous hemodiafiltration (CVVHDF) meta-analysis has failed to show any benefit in mortality, renal recovery, vasopressor use, or organ failure.⁴⁴

Anticoagulation

All RRT requires anticoagulation to avoid circuit clotting that can result in blood loss, fluid administration to flush the circuit, and minimize blood loss. Treatment disruption to replace the circuit, associated costs for dialyzer and line changes, and resource utilization of personnel, since many institutions mandate dialysis staff to replace circuits. Due to lower BFRs, CRRT have higher clotting risks than IHD.

As in IHD, heparin is the preferred anticoagulant. It is intravenous (IV) bolus administered in the venous line, 2000 to 5000 units, allowing a few minutes to mix. Maintenance is 500 to1000 units/h infused by roller pump into the arterial line. Partial thromboplastin time (PTT) is measured via arterial and venous lines every 6 hours to maintain PTT 40 to 45 or greater than 65, respectively.

In patients with heparin-induced thrombocytopenia direct thrombin inhibitors are used.⁴⁵ Argatroban is preferred in renal patients at 0.5 to 1 μg/kg/min, due to its liver metabolism.⁴⁶ Lepirudin is renal eliminated and may be administered as bolus or infusion (0.005-0.025 mg/kg/h). A target activated partial thromboplastin time (aPTT) greater than 1.5 to 2 times normal avoids excessive bleeding and ensures anticoagulation. Fresh frozen plasma is employed for reversal of bleeding attributed to these direct thrombin inhibitors.

Anticoagulation should be reviewed daily between intensivist and nephrologist, and if contraindicated, heparin-free treatments are possible. In SLEDD periodic, every 15 to 30 minutes, saline flushes are instituted. In CVVH, RF is infused prefilter to avoid hemoconcentration.⁴⁷ Heparin-free extracorporeal circuits on average clot over 8 hours. Decrease in dialysate/serum BUN levels to less than 0.6 may indicate imminent clotting.

Regional citrate avoids systemic anticoagulation, and is superior in circuit survival and bleeding complications.⁹ Citrate can be administered either before or after blood has been exposed to the filter. Commercially available solutions include ACD-A which has 3% trisodium citrate, citric acid, and dextrose. Using calcium-free RF decreases the amount of citrate required for anticoagulation. Periodic assessment of bicarbonate levels and ionized Calcium (iCA), both from the extracorporeal circuit and patient, is mandatory to avoid serious hypocalcemia. The citrate administration or calcium supplementation is titrated to keep the postfilter iCA between 1.21 and 1.45 mmol/L. Infusion of calcium into circuit avoids delivery of hypocalcemic blood to patient. Citrate toxicity can be a concern in patients with liver disease or lactic acidosis.⁴⁸

Staffing CRRT requires intense attention to patient care, necessitating one-to-one nurse-to-patient ratios. Elevating the head of the bed or turning the patient can compromise flow and risk clotting, making positioning an issue. In the case of femoral access, better flow is obtained in the supine position.⁴⁷ Also, some institutions require hemodialysis staff to assist at CRRT commencement, interruption, or termination.

Equipment CVVH equipment is smaller in size since water purification filters in RRT requiring dialysate (IHD, CVVHDF, or SLEDD) are not necessary. The basic operating principle is similar to IHD as a blood pump is utilized to circulate blood through the dialyzer/filter to generate UF (see Figure 31–3). CVVH machines are now equipped with temperature regulation mechanisms resulting in added hemodynamic effects of vasoconstriction at lower temperatures. The febrile response to infection can therefore be missed. Despite impressive advancements in practicality and compactness of machines, it is still necessary to interrupt CRRT for transportation.

Access Any preexisting ESRD access can be utilized for CRRT. In AKI requires placement of a double-lumen catheter in either the femoral or internal jugular (IJ) vein regardless of RRT. These catheters have staggered openings allowing blood flow out from the patient via the proximal (insertion site) opening and return via the distal opening. This arrangement diminishes the degree of mixing or recirculation, and hence, inefficiency.

In general, site selection for catheter placement follows similar risks such as bleeding and infection. Subclavian and left IJ veins are avoided when possible due to the angulation necessary for proper placement and flow. The stiffer catheter risks vessel injury, compromising blood flow and clotting. Furthermore, the subclavian catheter can cause stenosis resulting in high venous pressures, making the creation of a permanent AV access difficult. Up to 14% of AKI patients may need permanent chronic dialysis.^5,8

Sustained Low-Efficiency Daily Dialysis

Also called extended dialysis (ED), it is a hybrid of CRRT and IHD. It is increasingly utilized throughout the world,⁴⁹ as well as in the United States, where about 25% of ICU providers reported using SLEDD in 2007.⁴⁰

SLEDD generally runs daily for 8 to 10 hours with low DFR (200-400 cc/min) and BFR (150-200 cc/min). It can utilize the IHD machinery with added software to support a lower BFR allowing for a more gradual solute and/or fluid removal. Time on SLEDD can be extended up to 24 hours, to provide for greater clearance according to perceived needs by providers.⁵⁰ The filters have smaller surface areas (1 m² compared with 1.5 m² for IHD), lower UF coefficients (10 mL/h/mm of Hg compared with 45), and lower K0As (maximal theoretical clearance of dialyzer of 600 mL/min compared with 1000 mL/min), but are similarly composed of polysulfone.

SLEDD's hybrid properties support hemodynamics, enhance clearance, and increase convenience while lowering costs. This, as well as its versatility, accounts for a rising popularity.⁴⁹ For instance, it often is utilized overnight (nocturnal dialysis) when the probability for interruption from procedures or radiologic examinations is less likely. Such maneuvers better guarantee the prescribed dialysis without interruption. Still another important benefit is the decreased need for heparin, not being continuous.

Most importantly, these benefits come without compromising outcomes in AKI as compared to CRRT.⁴⁹ In a recent study,⁵⁰ the net 24-hour UF and fluid balances were similar between CVVH and SLEDD, as were hypotensive episodes. Furthermore, SLEDD was associated with a fewer days on mechanical ventilation and in the ICU.

Sustained continuous UF and fluid removal in severe CHF

Though technically not RRT, extracorporeal modalities for UF are considered in severe CHF with diuretic resistance. In such patients renal impairment is not the overwhelming reason to initiate sustained continuous UF (SCUF) but rather as an alternative to diuretics. To date various studies, the UNLOAD and others, have not demonstrated any renal or length-of-stay advantages.⁵¹ SCUF should probably be relegated to patients who failed medical therapy and are awaiting cardiac transplant (see Table 31–1).⁵²

Dosing

Unfortunately, significant questions involving RRT, namely the exact clearance provided, how much is necessary and when to provide it are poorly understood. Dosing should indicate a measured effectiveness of the ridding of waste products from a given volume of blood.¹⁴ Urea kinetic modeling, which is ESRD based, is questionable in assessing efficacy in the critically ill, given their increased catabolism and urea Vd.^14,26

Treatment dosing (volume of fluid processed) utilizes urea generation (g) over 24 hours (the difference in BUN levels 24 hours apart + loss in urine in nonoliguric AKI) as a marker, divided by the target BUN (g/L). For a 60-kg individual, water compartment is 33 L (0.55% of body weight). Total body urea = BUN level in g/L X's body water (add edema volume to this). The difference between 2 values 24 hours apart + urine loss of urea is total urea generation (g) in 24 hours. If urea production is 16 g/d and target BUN is 40 mg/dL (0.4 g/L), 16/0.4, 40 L of fluid needs to be processed to achieve the clearance and maintain BUN at 40 mg/dL. This however is not generally done when utilizing these modalities. In fact most providers are uncertain of prescribed dosing.⁵³

Past RRT trials have not used urea clearance as a study goal. Guidelines suggest a CRRT KT/V of 3.9/wk or an effluent volume of 20 to 25 mL/Kg/h.⁵⁴ To achieve this a higher prescription of 25 to 30 mL/kg/h may be necessary. What's more, CVVHDF studies suggest using doses at 35 mL/kg/h.

Antibiotic Management The most common cause of AKI in the ICU remains sepsis⁸ and despite advances, the mortality has not appreciably improved.^3,6,13,29 Though the reasons are likely multifactorial, recent attention on antibiotic (AB) dosing during RRT has identified an important shortcoming (Table 31–3).⁵⁵ High-flux membranes with larger surface areas and increased time on RRT (CRRT or SLEDD) remove ABs more efficiently compared to IHD. In these forms of RRT, eGFR provided may be up to 30 mL/min and as many as 25% of patients are not making pharmacokinetic targets.⁵⁵ Furthermore, difficulties in dosing are further compounded when considering modality variability and intermittent usage.⁵⁶ Given data supporting improved mortality with adequate AB dosing, this is a legitimate concern for septic patients on RRT.

Avoiding rapid changes in osmolarity and electrolyte levels

In patients with hemodynamic uncertainty it may be prudent to select a modality other than IHD. However, there are at least 3 scenarios where a gradual clearance rate may be wise even with hemodynamic stability: in the setting of acute coronary syndromes, elevated ICPs or hypo/hypernatremia. In such cases the issue is avoiding an abrupt solute or osmole change rather than hemodynamics.

Rapid decline of potassium levels in ESRD patients with acute myocardial infarction (MI) may increase the risk for arrhythmias.⁵⁷ Longer RRT time with lower BFR achieves a gradual safer decrease in potassium levels. Continuous RRT is ideal for such a presentation.

In patients with raised ICP or its potential, standard IHD should be avoided. Abrupt decreases in urea with resultant fluid shifts in the central nervous system (CNS) can result in worsening cerebral edema. Similarly, gradual decreases in urea concentration using continuous, sustained, or peritoneal dialysis are better tolerated.³⁸ These modalities also provide some insurance against hypotensive episodes, which by lowering cerebral perfusion pressure adds to the injury. Here too, if other modalities are not available, IHD slowed down to BFR of 150 to 200 cc/h as discussed earlier should be sufficient.

Finally, rapid normalization of sodium levels in patients with hyper- or hyponatremia can induce central pontine myelinolysis and/or edema. Likewise, dialysis disequilibrium syndrome is due to rapid reduction of high urea levels. Use of gradual RRT, close monitoring, and intravascular fluids can help guarantee safe normalization.

Endotoxin removal

There are multiple systems studying endotoxin removal in sepsis such as high-volume hemofiltration and hemoperfusion with polymyxin B embedded polystyrene absorbing surfaces. One issue of controversy is that anti-inflammatory cytokines and factors are removed as well.⁵⁸ Small studies have shown decreases in cytokines with inconsistent outcome effects.⁵⁹

Renal recovery

In general, AKI patients requiring RRT recover significant renal function, not requiring chronic dialysis. However, up to 14%, primarily those with baseline CKD, will continue RRT.^5,8 Recently, Wald et al⁶⁰ demonstrated that dialysis-requiring AKI was associated with an increased risk for later ESRD (adjusted heart rate [HR] 3.2; 95% confidence interval [CI] 2.7-3.9) compared with controls matched for age, gender, CKD status, need for mechanical ventilation, and a propensity score for dialysis-requiring AKI.

As to the benefits of one RRT modality versus another in supporting renal recovery, CRRT may be preferred. The suggestion is that better hemodynamic control, solute clearance, and maintenance of acid-base, may be more natural with CRRT. Though not seen in all studies, a recent review of 16 observational and 7 randomized controlled trials (RCTs) suggests CRRT benefits renal recovery (relative risk IHD vs CRRT of 1.73, P = 0.02).⁶¹ Significance was not seen when limiting the study to RCTs, and does not distinguish SLEDD from IHD. If this leads to less ESRD, then the use of CRRT may benefit quality of life as well as long-term costs.⁶²

Cost

RRT is one of the most expensive interventions used in an already cost burdensome ICU setting.⁶² Cost is primarily personnel, whose need is largely related to the selected modality's degree of complexity. Furthermore, RRT patients inevitably have longer lengths of stay.⁵ CRRT modalities being the most complicated and requiring the most supplies (replacement fluid, tubing, etc) are the most expensive, by as much as thousands of dollars per week.⁹ Of the RRTs, a number of studies have indicated SLEDD to be the least expensive, largely due to decreased nursing needs.⁵⁰

Mortality and outcomes

ICU patients with AKI requiring RRT have a higher mortality than other ICU patients including those with ESRD or AKI without need for RRT.^6,8 Despite over 35 years of experience and improvements,^2,30,39 there still remains an unfortunate mortality of between 25% and 50%.^3,6,10,13,29

The long-standing controversy is whether CRRT is superior to IHD in regards to mortality. At present the predominant evidence, randomized trials, and meta-analyses, do not support a mortality benefit for CRRT.⁹ Still the recent KDIGO AKI guidelines prefer CRRT (level 2 recommendation), over IHD, for hemodynamically unstable patients.⁵⁴ What's more, there does not seem to be a mortality advantage with larger delivered clearance doses.^9,13,29 Palevsky reviewed 2 cohorts of AKI patients receiving RRT, one attaining a higher intensity and therefore higher RRT dose compared to a second lower dose; no mortality advantage was demonstrated.¹³ Similarly, higher CVVH doses have not demonstrated any mortality benefits (Table 31–4).²⁹

Shortcomings: What Explains These Shortcomings is Likely Multifactorial

First the modalities themselves: Closer monitoring is vital. In the Palevsky study, which demonstrated a lack of benefit with intense RRT, it is notable that there was a greater need for vasopressor support in this group despite the use of continuous RRT.¹³ Moreover, there was significantly more hypokalemia and hypophosphatemia. The latter also seen in the higher UF group of the RENAAL study.²⁹ In the critically ill, electrolyte abnormalities and/or their rapidity⁵⁷ may result in morbidity and even mortality (Table 31–5).

Secondly, the process itself: the water in the dialysate, catheter and tubing plastic, and dialyzers are all capable of initiating inflammatory responses. Modifying dialysate water to ultrapure status, for instance can decrease inflammation and morbidity.⁶³ Finally as discussed earlier, better attention to antibiotic dosing is needed.

Thirdly, it may be that even with all these RRTs, present clearances are simply inadequate in the severely critically ill patient. Recent data demonstrate eGFR rising up by 40%, and over 130 cc/min in some critically ill patients.⁶⁴ Critically ill patients may need elevated clearances to help them survive their catabolic state, where the uremic toxin burden may be high.¹⁶ RRT modalities at best approach 30 cc/min or stage 4 CKD. Data that more intense RRT does not improve outcomes further support this.¹³ Even with higher-intensity RRT, our attempts may not even begin to provide what is needed.

Reviewing the shortcomings of present RRT systems prompts possible future research endeavors to optimize present systems. These include pharmacokinetic studies, improved biocompatible materials, and possibly water purification. Improving monitoring accuracy is also needed to minimize electrolyte and hemodynamic deficits.

Care in the design of proper future studies⁶⁵ is also paramount to understand outcomes. Biomarkers of AKI such as cystatin need to be studied to not only assess earlier response but if RRT is necessary. Recognition of the important consequences of AKI appears to be at hand and has prompted research in these areas.