• 
  

  Indications for the initiation of renal replacement therapy (RRT) remain reactive, often waiting until potentially life-threatening complications/thresholds have been met.

  
  
• 
  

  The goal of renal replacement therapy should be to provide “renal support” to facilitate the other aspects of care of the critically ill patient (fluid balance, nutritional support, etc).

  
  
• 
  

  Retrospective and observational studies suggest that the early initiation of RRT may improve patient outcomes; however, definitive randomized, controlled trials have yet to be performed.

  
  
• 
  

  In the setting of acute kidney injury (AKI), no specific RRT modality (intermittent, continuous, or peritoneal) provides a mortality benefit over another. However, certain clinical scenarios (eg, hepatic failure, increased intracranial pressure) may mandate a specific modality.

  
  
• 
  

  In the setting AKI, randomized controlled trials have demonstrated that a minimum dose of 25 mL/kg/h of continuous renal replacement therapy (CRRT) be delivered in order to improve patient survival. Data on dosing of intermittent dialysis suggest prescription of a minimum of three treatments per week.

  
  
• 
  

  No singular method of systemic or regional anticoagulation, in the setting of AKI requiring renal replacement therapy, has demonstrated superiority. Several options including heparin, citrate, and no anticoagulation remain extremely common and each has their own risks and benefits.

  
  
• 
  

  In the setting of AKI requiring RRT, nutritional support consistent with the current ESPEN guidelines and monitoring of parameters of nutritional status in critically ill patients are appropriate.

  
  
• 
  

  Depending on the modality of RRT (intermittent, continuous, or peritoneal), dosing strategies for medications (including antimicrobials) differ significantly.

  
  
• 
  

  Adherence to dosing guidelines is critical to ensure that the targeted therapeutic dose is delivered in the setting of AKI and RRT, as inappropriate dosing has a significant impact on patient outcomes and increases the risk of mortality.

Despite advances in medicine and critical care, the nephrology community has yet to develop a consistent, proven intervention to predictably prevent or hasten the recovery of all forms of acute kidney injury (AKI), including its most severe form, acute tubular necrosis (ATN). Thus, care for the patient with AKI is focused on supportive measures including treatment of the underlying disease state and, when needed, renal replacement therapy (RRT). While advances in nephrology have not identified a consistent therapy for the prevention or improved recovery for AKI, there have been considerable advances in the field of RRT.

RRT, in this setting, refers to the use of extracorporeal support to remove solutes and water. The current available modalities of RRT are intermittent hemodialysis (IHD), peritoneal dialysis (PD), and the various blood-based modalities of continuous renal replacement therapy (CRRT). CRRT modalities include continuous venovenous hemodialysis (CVVHD), continuous venovenous hemofiltration (CVVH), and combination therapies, that is continuous hemodiafiltration (CVVHDF). The advances in technology with readily available large bore temporary and tunneled venous catheters and blood pumps have made the use of arteriovenous circuits, in the form of continuous arteriovenous hemofiltration/hemodialysis (CAVH/CAVHD) essentially obsolete.

The general principles underlying these various modalities remain the same: Solutes and water move across a semipermeable membrane and are ultimately removed from the body. The process by which solute and water transfer occur differs based on the modality of RRT. Dialysis operates on the principle of diffusion, that is solutes move across a semipermeable membrane down their concentration gradient (moving from higher concentration to lower concentration). This is utilized in modern hemodialysis techniques with blood flowing adjacent to a dialysate solution separated by a biocompatible filtering membrane. To maximize the concentration gradient between the blood and dialysate space, the dialysate flow is countercurrent to the flow of blood. Diffusion-based clearance of solutes remains limited by the principles governing all diffusion; the movement of solutes will depend not only concentration gradient, but also on the size of the solute. Smaller solutes are more diffusible, because they randomly move more in solution than large solutes.

Convective clearance, used in hemofiltration, is an alternative means of achieving solute clearance that generally provides better clearance of larger size solutes (see below, “Modality of RRT: Convective Versus Diffusive”). Ultrafiltration operates on the principle that water will move across a semipermeable membrane from a higher-pressure system to a lower-pressure system. Solutes dissolved in ultrafiltered water also move across the membrane via “solvent drag,” or convection. Importantly, solute clearance by ultrafiltration requires high volumes of water movement. Thus, low-volume hemofiltration, as utilized in slow continuous ultrafiltration (SCUF), is effective at removing water, with limited solute removal (or “clearance”). Conversely, high-volume hemofiltration, such as used in CVVH, is effective at removing solutes, but large amounts of water are removed and fluid must be returned to maintain blood volume in the form of a replacement solution. The replacement solution also contains supplemental electrolytes (eg, potassium, phosphorus, calcium) and a buffer (lactate or bicarbonate) to prevent iatrogenic depletion of these solutes, in addition to treating metabolic acidosis, and both diluting and removing circulating uremic solutes. The use of acute peritoneal dialysis in the setting of AKI has largely fallen out of favor in many countries, although still commonly used in critically ill children and for adult acute RRT in developing countries. In the large, multicenter international observational study of the epidemiology of AKI in the setting of critical illness conducted by the Beginning and Ending Supportive Therapy for the Kidney (BEST for the Kidney) investigators, only 40 of 1258 (3.2%) individuals requiring RRT underwent PD or SCUF.¹ Thus, the focus of this chapter will be on the use of blood-based extracorporeal therapies to achieve small solute clearance, including IHD and the forms of CRRT outlined above. The choice of modality (intermittent vs continuous therapy and diffusive vs convective therapy) remains controversial and the data supporting the different modalities will be outlined in more detail in the sections below. Nevertheless, the goal of all RRT therapy remains the same; ameliorate the severe metabolic and volume derangements that contribute to the poor prognosis of AKI in the setting of critical illness. Combinations of these complementary therapies are commonly used to support such patients at various stages of their acute illness and recovery.

The indications for renal replacement therapy vary between the clear and the obscure. Medical students and physicians-in-training are routinely instructed that there are some uncontroversial, standard “acute indications for hemodialysis” (see Table 98-1).

These established “indications,” however, are severely limited. Firstly, they are reactive in nature, as they aim to avert potentially life-threatening complications of renal dysfunction as they become clinically problematic. Secondly, while some of the indications are objective and readily apparent (ie, hyperkalemia with ECG changes or pulmonary edema requiring mechanical ventilatory support), others are potentially subjective and nonspecific (the clinical diagnosis of uremia).

In recent years, many clinicians have opted to initiate RRT earlier in the evolving course of AKI, attempting to be more proactive and reduce the burden of the complications before they become acutely life threatening. Similarly, investigators have studied various criteria for the initiation of RRT to identify the optimal time of initiation, mostly by conducting retrospective analyses of RRT datasets. Initial studies focused on the degree of azotemia at RRT initiation to compare “early” versus “late” initiation of RRT. Uncontrolled data on “prophylactic” RRT in the setting of posttraumatic renal failure (initiation of dialysis prior to blood urea nitrogen (BUN) reaching 200 mg/dL) suggested marked improvement in survival as well as neuromuscular, metabolic, and hematologic consequences of renal dysfunction.² The first controlled trial by Conger and colleagues confirmed these findings; 18 individuals with posttraumatic AKI were randomized to a more intensive hemodialysis therapy to maintain BUN <60 mg/dL and serum creatinine (SCr) <5 mg/dL versus holding the initiation of hemodialysis until BUN >150 mg/dL and SCr >10 mg/dL or other complications developed (hyperkalemia, volume overload, or uremic encephalopathy). Five of eight patients (64%) in the intensive dialysis arm survived, versus 2 of 10 patients (20%) in the conservative arm, a difference that did not reach statistical significance (p = 0.14).³ While the study was small, the results were consistent with the retrospective findings, and supported to the notion that “prophylactic” hemodialysis may be helpful to reduce the complications of kidney injury.

Controlled trials regarding timing of initiation of RRT since the initial study by Conger have been limited and the results have failed to provide definitive direction on optimal timing. Rather than using markers of azotemia as the strict criteria for randomization, Bouman and others conducted a more recent randomized controlled trial of initiation of RRT using a more comprehensive strategy. One hundred six adult subjects were randomized to one of three strategies: (a) early, high-volume hemofiltration, (b) early, low-volume hemofiltration, or (c) late, low-volume hemofiltration. Patients were critically ill and eligible for randomization if they met the following criteria: urine output <30 mL/h for more than 6 hours despite adequate circulatory support (central venous pressure [CVP] or pulmonary artery occlusion pressure [PAOP] >12 mm Hg), addition of any dose of norepinephrine or phosphodiesterase inhibitors or >5µg/kg/min of dobutamine or dopamine and challenge with high-dose diuretics (>500 mg of furosemide in <6 hours), creatinine clearance (CrCl) of <20 mL/min in a 3-hour urine collection and receiving mechanical ventilation. Individuals with preexisting chronic kidney disease (Cockroft-Gault estimated creatinine clearance <30 mL/min), AKI secondary to glomerulonephritis, tubulointerstitial nephritis, post-renal obstruction, surgical renal artery occlusion, or preexisting advanced liver disease or AIDS were excluded. Early initiation was defined as initiation of CRRT within 12 hours of meeting inclusion criteria. Late initiation was defined as implementation once conventional criteria for RRT were met (BUN >112 mg/dL, potassium >6.5 mmol/L or severe cardiogenic pulmonary edema requiring high-level ventilatory support). The study also compared dose of therapy—high volume defined as blood flow rate of 200 mL/min and hemofiltration rate of >3 L/h and low volume defined as blood flow rate of 150 mL/min and hemofiltration rate of 1 to 1.5 L/h. The mean time from meeting inclusion criteria to initiation of CRRT was 7 hours in the early group and 42 hours in the late group. There was no baseline difference in severity of illness scores (at ICU admission or study inclusion), vasoactive support or creatinine clearance (at study inclusion) between the early and late groups. The investigators found no difference in survival (ICU, hospital, or 28-day) or duration of renal failure, mechanical ventilation, or hospitalization between the early and late groups. Although this was a very small and underpowered trial, and essentially a pilot study, it remains the only prospective, randomized, controlled trial of RRT initiation timing in the modern era.⁴

Despite the absence of convincing data supporting or refuting early initiation of RRT, further prospective trials have been lacking. Additional data supporting the early initiation of RRT come from retrospective, observational studies examining the use of RRT in AKI occurring in a variety of clinical settings, including sepsis, post-cardiac surgery and combined liver and renal failure. Overall, the data appear to support earlier RRT initiation.^5-8 However, given the retrospective nature of these studies, the variability in criteria used to define “early” versus “late” timing, and the diversity of patient populations, it is impossible to use this literature to provide a strong, evidence-based recommendation for early initiation of RRT. The literature remains conflicting and the primary limitations in study design have been the continued reliance on markers of clearance to identify individuals with kidney injury and the common practice of waiting for significant complications to develop prior to initiating RRT. Levels of certain serum chemistries (eg, potassium, phosphorus, bicarbonate) are affected by issues not directly related to the severity of AKI. They are also influenced by dietary intake, choice of fluid administration, and medication use; thus, their utility as thresholds for initiation of RRT is questionable (Table 98-2).

The goal of initiation of RRT should move beyond the notion of a simple “replacement therapy”, used reactively to remove the waste products and excess fluid that accumulate in AKI. Rather, as Mehta writes, the goal of RRT should be to provide “renal support” and facilitate the other aspects of care of the critically ill patient including early nutritional support, restoration or preservation of euvolemia, maintenance of acid-base balance, maintenance of respiratory gas exchange, and prevention of the accumulation of endogenous and exogenous (ie, medications/metabolites and poisons) toxins.⁹

Unfortunately, the available literature does not provide specific, objective guidelines for how to integrate these additional clinical factors (volume excess, nutritional support, etc) into the decision-making process of initiating RRT. Nevertheless, the available literature does emphasize the potential deleterious effects of the complications of AKI and the potential benefits of full supportive measures. Specifically, data from the PICARD study, a multicenter, prospective observational study of patients with AKI in the setting of critical illness, observed that individuals with AKI and fluid overload (defined as >10% increase in fluid as compared to admission weight) had greater mortality: in-hospital—48% versus 35%, p = 0.01, 30-day—37% versus 25%, p = 0.02, and 60-day—46% versus 32%, p = 0.006.¹⁰ An observational study of more than 17,000 individuals with AKI and concomitant critical illness, using multivariate stepwise logistic regression the use of enteral nutritional support, compared to all other nutritional support options, was associated with improved survival (OR 0.86; p < 0.001).¹¹

Guidelines regarding initiation of RRT in the setting of AKI and critical illness should, therefore, take into account the complete aspects of treating these complex patients as well as objective definitions of AKI. It is likely that until the development and acceptance of standardized and reproducible criteria to initiate RRT, the standard approach to timing of RRT initiation will continue to be individualized without use of standardized criteria, and accordingly there will be a high degree of practice variability.

The widespread availability of sophisticated modern dialysis technologies including tunneled and temporary venous catheters; blood pumps that are able to maintain adequate blood flows to prevent thrombosis; standardized, portable dialysate and replacement solutions; and RRT equipment platforms that can be operated by well-trained nursing staff without a dedicated dialysis background has increased the popularity of continuous modalities of RRT (CRRT). In the study conducted by the BEST Kidney investigators of AKI in the ICU, 1006/1258 (80%) of individuals received CRRT.¹ Nevertheless, the optimal choice of modality remains controversial, and may be limited by resources available at a given institution.

The rationale for the use of continuous modalities is based primarily on the common presence of hemodynamic instability of critically ill patients with AKI, which is often exacerbated by IHD. Initially described in clinical use in 1977, continuous arteriovenous hemofiltration (CAVH) provided a means of fluid and solute removal in patients with systemic hypotension.¹² Since its initial inception, the use of CRRT (initially CAVH and subsequently using continuous hemodialysis—CAVHD) increased, and investigators worldwide reported its effectiveness at facilitating both solute clearance and volume removal in the setting of AKI, especially in the critically ill.^13-15 Despite the widespread use, reported success, and improved fluid balance/solute clearance, experimental studies directly comparing the use of continuous versus intermittent RRT have been limited, and have not shown clear superiority of CRRT over intermittent RRT in the majority of studies.

In 2001, the results of the first direct comparison trial of intermittent versus continuous RRT in the setting of AKI and critical illness was reported. This study of 166 critically ill adults with AKI was conducted at four US medical centers. AKI was defined as a BUN ≥40 mg/dL or SCr ≥2 mg/dL for those without baseline values and an increase of ≥1 mg/dL from a baseline for those with known prior serum creatinine values. Subjects were randomized to receive IHD or CRRT (CAVHDF for the initial 2 years, and CVVHDF for the subsequent years of the study). Importantly, individuals were required to have a mean arterial pressures (MAP) >70 mm Hg (with or without vasopressor support) to be eligible for randomization, so the population most likely to be selected for CRRT in clinical practice were excluded from this trial. The baseline characteristics of the subjects were generally similar, except individuals randomized to CRRT were more likely to have liver failure and had higher mean APACHE III scores. Unadjusted mortality was higher in the group randomized to CRRT: 59.5% versus 41.5%, p < 0.02 at 28-day and 65.5% versus 47.6, p < 0.02 in-hospital. However, after adjusting for the fact that individuals randomized to CRRT were more severely ill, the survival difference was eliminated. Specifically, when each tertile of APACHE III scores was examined and mortality was compared between IHD and CRRT, no difference was seen.¹⁶ The importance of this study is that it highlighted that while a randomized trial of modality of RRT could be executed, enrollment would remain limited by underlying illness. Approximately 21% of individuals who were initially screened and underwent RRT were not included in the study because they failed to meet the criteria for hemodynamic stability. Thus, while no benefit to CRRT was observed in the study, there were a significant number of individuals who could not safely receive IHD. No data were given on the outcome of these individuals and whether their ability to receive RRT ultimately influenced their outcomes is unknown.

The results of subsequent, randomized controlled trials have demonstrated similar outcomes. A single-center study conducted at the Cleveland Clinic randomized 80 critically ill adults with AKI requiring RRT to IHD or CVVHD. Importantly, the subjects were randomized according to severity of illness (high or low as determined by the Cleveland Clinic Foundation severity of illness score). Furthermore, while exclusion criteria were similar to other studies (eg, individuals previously receiving dialysis were excluded), no individuals were excluded for hemodynamic instability. Although the study failed to demonstrate a significant mortality benefit for either modality (67.5% of patients died in the CVVHD group vs 70% of the patients in the IHD group; p = NS), the study was inadequately powered for this end point. However, the two groups did differ significantly in their hemodynamic response to RRT, and in their achievement of fluid balance control. In the 72 hours after initiation of RRT, MAP fell in the individuals receiving IHD while there was no change in MAP in the individuals receiving CVVHD. MAP at day 3 of RRT was higher in the CVVHD group versus the IHD group (79.9 ± 9.3 mm Hg vs 74.2 ± 10 mm Hg) despite achieving a greater negative fluid balance over the first 3 days (median −4.005 L vs ±1.539 L; p < 0.001).¹⁷

The Hemodiafe study group, a multicenter study consortium including French medical centers, also conducted a randomized trial of intermittent versus continuous renal replacement therapy. One hundred eighty-four adult subjects were randomized to IHD and 175 individuals were randomized to CVVHDF. Importantly, the study excluded individuals with a SAPS II score less than 37, focusing on individuals with AKI and multiorgan system failure in the ICU; however, no comment was given with regard to exclusion of patients with hemodynamic instability. Similar to the results of previous studies, no difference in mortality was seen between the two groups—41.8%, 31.5%, and 27.2% in the intermittent group versus 38.9%, 32.6%, and 28.5% in the continuous group at 28, 60, and 90 days, respectively; p = NS for all comparisons. Further, there was no difference in frequency of hypotensive episodes (39% in IHD group vs 35% in the CVVHDF group; p = 0.47), despite similar mean net fluid removal on days of therapy (2213 mL in the IHD group vs 2107 mL in the CVVHDF group). The total net fluid balance, accounting for days not receiving therapy, was not reported.¹⁸

Finally, the Stuivenberg Hospital Acute Renal Failure (SHARF) project also conducted a randomized trial of intermittent versus continuous renal replacement therapy. Similar to the study conducted by the Cleveland Clinic, randomization was according to severity of illness as determined by the SHARF severity of illness score. The investigators randomized a total 316 adults, 144 to intermittent dialysis and 172 to CVVH. One hundred twenty-four of the eligible patients were excluded from randomization due to “medical reasons”—primarily coagulation or hemodynamic disturbance. The groups at baseline were similar and hospital mortality was similar in the two groups—58.1% in the CVVH group and 62.5% in the IHD group.¹⁹

Taken together, the collective results of the clinical trials conducted to date comparing intermittent versus continuous renal replacement therapy do not demonstrate a mortality benefit or a significant impact on recovery of renal function for either modality. Even when other effects of RRT are assessed, that is fluid balance and adequacy of clearance, the benefits of CRRT do not translate into mortality differences. The implications of the study results for clinicians are unclear. Should the results guide clinicians to only use IHD in that it allows more patient mobility and has lower cost? Have the conducted studies been adequately powered to demonstrate a mortality difference when the overall mortality in the studies is lower than observed mortality in nonexperimental trials? Further complicating our assessment of the findings of experimental trials, the results of a meta-analysis conducted in 2002 utilizing 13 studies and 1400 individuals including both observational and experimental designs concluded, after adjusting for severity of illness and quality of study, the relative risk of mortality was lower in patients receiving CRRT.²⁰

Rather than a nihilistic approach, an alternative method of interpreting the data is simply that we are studying the wrong question. Applying a “one-size-fits-all” approach to studying modality of RRT is flawed. In centers where both modalities—IHD and CRRT—are available, choice of therapy will be influenced by both patient and nonpatient factors that are not included in randomized controlled trials (ie, catheter function, safety of anticoagulation, patient mobility, nurse staffing, etc). Most importantly, the results of the available studies suggest that with use of either modality of RRT, practicing physicians are “doing no harm.” Individuals with severe hemodynamic instability or certain other special conditions outlined below may still benefit from use of CRRT over intermittent therapy. However, for the majority of individuals, even in the setting of multiorgan system failure, intermittent dialysis can adequately achieve solute clearance and control of volume balance. Rather than focusing on selecting one modality for all individuals, the goals of therapy—large amounts of volume removal, removal of ingested toxins, clearance of uremic solutes, etc, should be kept in mind and the therapy selected accordingly. The most important point is that in centers where both intermittent and continuous modalities are available in the ICU, they are used as complementary therapies. Apart from the special populations discussed below, such centers generally use IHD for hemodynamically stable ICU patients, or when rapid removal of potassium, toxins, or fluid is desired and the patient is sufficiently hemodynamically stable to tolerate aggressive dialytic therapy, reserving continuous modalities for periods of hemodynamic instability, particularly when associated with significant fluid overload. This personalized approach was successfully used in the ATN trial of RRT intensity (dose), which is discussed below.

While the above literature suggests that both intermittent and continuous therapies can be used in critically ill patients with AKI, a few clinical scenarios deserve special attention with regard to modality of therapy and may represent situations where one modality is superior.

ACUTE LIVER FAILURE

Acute liver failure (or fulminant hepatic failure) is characterized by laboratory abnormalities suggestive of hepatocyte injury, impairment of liver function (manifested by increased prothrombin time [PT]/international normalized ratio [INR] and bilirubin), and encephalopathy. Individuals with fulminant hepatic failure are at risk for increased intracranial pressure and cerebral herniation. Clearance of solutes and water via intermittent RRT in the setting of AKI and fulminant hepatic failure may have adverse effects on intracranial pressure, because rapid extracorporeal clearance of uremic solutes causes acute plasma hypoosmolality, shifting water into the brain. In an observational study of nine patients with fulminant hepatic failure, increased intracranial pressure (ICP) and AKI treated with RRT, investigators compared the effect of IHD versus CRRT on ICP and cerebral perfusion. The mean ICP increased from 9 ± 1.4 mm Hg to 13 ± 1.8 mm Hg (p < 0.05) in the first hour of an intermittent hemofiltration treatment versus no change in ICP (19 ± 4.8 to 18 ± 4 mm Hg) in the first hour of treatment with CAVH. Further, MAP also significantly declined in the first hour of intermittent hemofiltration treatment (93 ± 2 to 82 ± 2.1 mm Hg), whereas there was no change in MAP in individuals receiving CAVH. Overall, the cerebral perfusion pressure (MAP – ICP) declined approximately 27% in the individuals receiving intermittent hemofiltration versus no change in the group receiving CAVH. The study was not designed to demonstrate a mortality difference between the groups and no significant difference was observed.²¹ The small study population, single center, and relative severity of disease (individuals already had evidence of increased ICP) limit the generalizability of the study findings. Additionally given the limited data describing the effects of CRRT on ICP in the setting of hepatic failure, studies such as this one which utilized the outdated modality of CAVH technology still guide therapy. Nevertheless, cerebral perfusion pressure is a critical parameter in the setting of fulminant hepatic failure and interventions that risk cerebral perfusion pressure should be avoided if possible. The hemodynamic benefits of CRRT in the setting of AKI and liver failure were further documented in a small, randomized trial conducted at the same center as the study above. Thirty-two patients with fulminant hepatic failure, intracranial pressure monitoring, and oliguric AKI were randomized to intermittent hemofiltration or CRRT. Ultimately, 12 patients were randomized to intermittent therapy and 20 patients were randomized to CRRT (8 received CAVH, 12 received CAVHD). Hemodynamic parameters including right atrial pressure, systemic vascular resistance (SVR), cardiac index (CI), and tissue oxygen delivery (DO₂) were assessed along with ICP. During the first hour of intermittent hemofiltration, CI fell 15 ± 2% versus no change in the CRRT arm (3 ± 3%). CI did return back to the index value during the course of the intermittent treatment. MAP also fell during the intermittent treatment, 82 mm Hg to 74 mm Hg, p < 0.05, versus no change in the CRRT arm, 74 mm Hg to 74 mm Hg. Correspondingly, oxygen delivery fell in the intermittent versus continuous therapy arm. After the initiation of therapy, intracranial pressure fell in the CRRT group prior to returning back to the baseline value. Conversely, intracranial pressure increased in the intermittent therapy group and remained increased throughout the duration of the treatment. No difference in hospital outcome or mortality was reported.²² The results of the, albeit limited, literature on RRT in the setting of acute liver failure suggest that intermittent therapies increase ICP and, thus, increase the risk for cerebral herniation. Further, similar to other states of circulatory dysfunction, there is greater hemodynamic stability with CRRT versus intermittent therapy. Individuals with both AKI and fulminant hepatic failure may represent a specialized group of patients where, in attempts to “do no harm,” CRRT is the preferred modality of treatment.

ACUTE NEUROLOGIC INJURY

The data demonstrating the superior cerebral perfusion of CRRT over IHD in patients with fulminant hepatic failure may also apply to other forms of acute neurologic injury where cerebral perfusion pressure is already compromised such as intracranial hemorrhage, postneurosurgical, etc. While explicit data comparing cerebral perfusion in patients receiving CRRT or IHD do not exist, the same principles outlined above apply. IHD does induce increased in brain water, thereby increasing ICP.²³ In clinical scenarios where acute neurologic injury has occurred and lowering of CPP may exacerbate the injury, CRRT may be preferable as a modality. If CRRT is not available, IHD with reduced strategies to minimize solute shift—reduced blood flow rates, reduced dialysate flow rates and dialysate sodium modeling (keeping dialysate sodium >145 mmol/L)—should be employed.

ACID-BASE DISTURBANCE

RRT is an effective means of treating acid-base disturbances. Large amounts of a buffered solution (bicarbonate or lactate based) can be administered with concomitant removal of excess sodium, water as well as the removal of organic acids that accumulate in renal dysfunction. The critically ill represent a special population with regard to acid-base disturbance. Often, they have mixed acid-based disorders (concomitant metabolic and respiratory disturbances) and the metabolic acidosis results not only from the retention of the organic acids that accumulate in renal dysfunction, but also the generation of organic acids from anaerobic metabolism (ie, lactate, pyruvate, etc) and the other effects of systemic inflammation.²⁴ Given the increased complexity and, often, more severe form of acid-base disturbance in the setting of AKI and critical illness, the question of modality of RRT arises.

Unfortunately, little data exist to address this question. The previous studies outlined did not look at acid-base status as an entry criterion or as an outcome. The results of a retrospective, observational study suggest a potential benefit of CRRT for the control of acid-base disturbances. Forty-seven individuals with AKI in ICU treated with IHD were compared with 49 patients with AKI in the ICU treated with CVVHDF with regard to the effect of RRT on electrolyte and acid-base status. The two groups were comparable with regard to baseline arterial bicarbonate concentration and proportion of individuals with decreased arterial bicarbonate concentration. Patients treated with CVVHDF had a significant increase in the arterial bicarbonate concentration in the first 48 hours, whereas the individuals treated with IHD did not. Further, individuals treated with CVVHDF were more likely to maintain normal arterial bicarbonate concentration throughout the observation period than their counterparts receiving IHD (71.5% vs 59.2%; p = 0.007). The study did not examine the clinical significance of these differences.²⁵ Given that the patients were well matched and the patients treated with IHD were supported with the only modality available at the time (rather than individuals selected as “less sick”), the study findings support the superiority of CRRT in ameliorating acid-base disturbance. However, whether the improved laboratory parameters translate to better patient outcomes cannot be determined from this study.

To better understand the effect of RRT in treating acid-base disturbances, investigators have specifically examined the effect of RRT on lactate clearance. Two hundred patients with AKI and lactic acidosis treated with CVVH were examined retrospectively. While lactic acidosis resolved in 45% of individuals while they were receiving CRRT, resolution of lactic acidosis had no effect on survival. Further, the authors attributed most of the improvement in the lactic acidosis to spontaneous improvement in metabolism and improved endogenous lactate clearance rather than attributable to CVVH therapy.²⁶ A prospective study designed to examine lactate clearance by CRRT confirmed this conclusion. Ten individuals with AKI and stable plasma lactate levels treated with CVVH were infused with sodium lactate. Serial blood and effluent lactate concentrations were measured to determine total lactate clearance and the proportion of lactate clearance attributable to the CRRT. Total plasma lactate clearance had a mean of 1379 mL/min with a mean clearance of 24.2 mL/min attributable to extracorporeal therapy. Fractional lactate clearance achieved by CRRT ranged between 0.5% and 3.2%. While lactate clearance achieved by CRRT was equivalent to urea clearance, the extracorporeal circuit played a very small role in total lactate clearance.²⁷ The available data on RRT in treating acid-based disorders provide limited guidance regarding the superiority of one modality. The available data suggest that endogenous pathways of clearing intermediate metabolites (eg, lactate) are far superior to extracorporeal removal even with continuous therapies. Thus, while there is often much enthusiasm regarding the use of CRRT to correct acid-based disorders in the critically ill with AKI, available data do not suggest it significantly affects patient outcome.

HYPERPHOSPHATEMIA