Acute Kidney Injury

Published on 26/03/2015 by admin

Last modified 22/04/2025

Print this page

This article have been viewed 3440 times

114 Acute Kidney Injury

Acute kidney injury (AKI) is characterized by an abrupt decrease in the glomerular filtration rate (GFR) that results in accumulation of nitrogenous waste products and an inability to maintain fluid and electrolyte homeostasis.¹ AKI can result from decreased renal perfusion not severe enough to cause cellular injury; an ischemic, toxic, or obstructive injury of the renal tubule; a tubulointerstitial process with inflammation and edema; or a primary reduction in the filtering capacity of the glomerulus. If renal tubular and glomerular function is intact, but solute clearance is limited by factors compromising renal perfusion, the injury is termed prerenal azotemia. If renal dysfunction is related to obstruction of the urinary outflow tract, it is termed postrenal azotemia. AKI due to a primary intrarenal cause is called intrinsic renal injury or renal azotemia. Prerenal azotemic and intrinsic renal injury due to ischemia and nephrotoxins are responsible for most episodes of AKI.^2,³

Renal blood flow is approximately 1200 mL/min and constitutes 20% of cardiac output. Given this apparently generous perfusion, it may seem surprising that the kidneys are so susceptible to hemodynamic insults. The majority of this perfusion (80%-90%), however, is to the renal cortex, where glomerular filtration occurs. The medulla is designed to concentrate and dilute urine. During urine concentration, the high osmotic gradient required for reabsorption of water is associated with a low rate of blood flow. In fact, oxygen tension in the outer medulla in the region of the metabolically active thick ascending limb of Henle is only around 10 mm Hg.⁴ This combination of low blood flow and oxygen tension in a metabolically active environment makes the kidneys very susceptible to ischemic injury.

Prerenal Causes

Prerenal azotemia is a consequence of reduction in renal perfusion without cellular injury. As such, this is a reversible process if the underlying cause is corrected. It may be secondary to decreased blood volume, as occurs with vomiting, dehydration, and hemorrhage, or it may be due to a reduction in the effective arterial blood volume, as in congestive heart failure and cirrhosis. Further, the administration of medications that interfere with the normal autoregulatory ability of the kidney can contribute to prerenal azotemia. In settings of diminished renal perfusion, administration of nonsteroidal antiinflammatory drugs (NSAIDs) or angiotensin-converting enzyme (ACE) inhibitors can precipitate overt prerenal azotemia.³

During prerenal azotemia, the renin-angiotensin-aldosterone system becomes activated secondary to a decrease in renal blood flow accompanied by increased activity of the adrenergic nervous system. Increased levels of angiotensin II and adrenergic activation serve to increase the proximal reabsorption of sodium, whereas aldosterone increases sodium reabsorption in the distal tubule. Together these actions decrease urine sodium concentration to less than 20 mmol/L and fractional excretion of sodium (FE_Na) to less than 1%.⁵

Prerenal azotemia accounts for approximately 70% of community-acquired cases of AKI ⁶ and 40% of hospital-acquired cases.⁷ Therefore, prerenal causes should be excluded in all cases of AKI. Therapy of prerenal AKI involves reversing the underlying cause, such as volume replacement or discontinuation of offending agents.

Postrenal Causes

Postrenal AKI occurs when there is bilateral (or unilateral in the case of a single kidney) obstruction of urine flow. Intratubular pressure increases and in turn decreases net glomerular filtration pressure. Obstruction of urine flow is a relatively uncommon cause of AKI and is more common in the community than in the intensive care unit (ICU). Several series have placed the incidence of postrenal AKI at 3% to 25% of all cases of AKI.^8,^9,¹⁰ Postrenal AKI can be divided into renal and extrarenal causes. Extrarenal causes include prostatic disease, pelvic malignancy, and retroperitoneal disorders. Intrarenal causes include crystal deposition, as occurs in ethylene glycol ingestion, or uric acid nephropathy in tumor lysis syndrome. Cast formation and tubular obstruction also occur in light-chain diseases such as multiple myeloma.

Postrenal causes of AKI should be evaluated with renal ultrasonography and measurement of postvoid residual urine in the bladder (>50 mL is abnormal). It is important to rule out these causes rapidly, because the potential for renal recovery is inversely related to the duration of obstruction.¹¹

Intrarenal Causes

Intrarenal causes of AKI can be classified according to the anatomic location of the injury: glomerulus, tubule, vasculature, or interstitium. Suspicion of glomerulonephritis or vasculitis should be raised in a patient with renal failure who has an active urine sediment with red cells and red cell casts. In contrast, acute interstitial nephritis classically presents with pyuria and white cell casts in the urine; on occasion, hematuria is also present. Most cases of AKI from interstitial nephritis are drug related, commonly due to antibiotics or NSAIDs. Recovery usually occurs with removal of the offending agent and may be hastened by a short course of steroids, such as 60 to 80 mg of prednisone for 10 days. Tubular injury is most often either ischemic or toxic in nature and presents as acute tubular necrosis (ATN). This is the most common form of AKI encountered in the hospital and ICU ^10,^12,¹³ and is the focus of this chapter.

In ischemic AKI, there is both tubular and vascular injury. In the tubules, an increase in intracellular calcium after ischemic injury activates the cysteine proteases calpain and caspase. This leads to necrosis and apoptosis as well as relocation of Na⁺/K⁺-ATPase from the basolateral membrane to the cytosol. This relocation interferes with normal vectorial transport of sodium and increases distal delivery of sodium chloride (NaCl). An increase in delivery of NaCl to the macula densa in the distal tubule activates tubuloglomerular feedback and further decreases GFR. Further, ischemia increases production of nitric oxide, which also causes cellular damage and detachment of epithelial cells from the basement membrane. Much of the deleterious action of nitric oxide is mediated through the generation of peroxynitrite from the combination of reactive oxygen species and nitric oxide. Cellular detachment is responsible for cast formation and tubular obstruction. These mechanisms all independently contribute to the decrease in renal function seen in ATN.¹⁴

In ischemic injury, the vascular endothelium is damaged and displays an exaggerated response to vasoconstrictor stimuli such as angiotensin II and endothelin-1 and a decreased response to vasodilators such as acetylcholine and bradykinin. In addition, there is a loss of autoregulatory capability. This loss of autoregulation in the setting of otherwise minor hemodynamic changes is likely responsible for the fresh ischemic lesions often seen on biopsy when recovery from AKI is delayed.¹⁵

The kidney’s susceptibility to toxic injury can be attributed to its functional properties. The kidneys receive 20% to 25% of the cardiac output, and there is extensive reabsorptive capacity as well as concentrating ability. All these factors contribute to the delivery of large amounts of toxin to the tubular epithelial cells. In addition, there is extensive biotransformation, generating toxic metabolites, and the high energy consumption with marginal oxygen delivery renders the tubules susceptible to toxic injury.¹⁶

An increasingly common form of AKI in the hospital is secondary to the use of contrast media. Nash and colleagues found contrast nephropathy to be the third most common form of AKI in the hospital.⁷ The pathogenesis involves both hemodynamic and toxic effects. Contrast media cause renal vasoconstriction and medullary ischemia as well as direct tubular toxicity.¹⁷ Patients with preexisting renal disease and diabetes are at high risk, as are patients who are volume depleted.

Differentiation of ATN from prerenal azotemia can be aided by evaluating urinary indices (Table 114-1).¹⁸ In established ATN, tubular function is impaired, and tubular sodium reabsorption is hindered. This results in a urine sodium value greater than 40 mmol/L and an FE_Na greater than 2%. Urine concentrating ability is also abnormal, resulting in isosthenuria with urine osmolality less than 350 mOsm/kg H₂O.¹⁹ However, a low FE_Na may be seen in entities causing ATN, such as rhabdomyolysis and myoglobinuria,²⁰ as well as in contrast-mediated AKI²¹ and sepsis.²² In patients with prerenal azotemia who are treated with diuretics that may obscure the FE_Na, fractional excretion of urea (FE_Urea) or urine-to-plasma ratio of creatinine may be more discriminatory. An FE_Urea less than 35% or a urine-to-plasma ratio of creatinine higher than 15 is indicative of prerenal azotemia.²³ However, a subsequent study indicates that in patients with AKI administered diuretics, the distinction between transient and persistent AKI cannot be made accurately by means of FE_Urea because it lacks specificity.²⁴

TABLE 114-1 Laboratory and Microscopic Findings in Prerenal Azotemia and Acute Tubular Necrosis

Laboratory Test	Prerenal Azotemia	Acute Tubular Necrosis
Urine osmolality (mOsm/kg H₂O)	>500	<400
Urine sodium (mEq/L)	<20	>40
Urine plasma/creatinine ratio	>40	<20
Fractional excretion of sodium (%)	<1	>2
Urinary sediment	Normal, occasional hyaline cast	Renal tubular epithelial cells, granular and muddy brown casts

Data from Esson ML, Schrier RW. Diagnosis and treatment of acute tubular necrosis. Ann Intern Med 2002;137:744-52.

Epidemiology

When the RIFLE criteria (risk, injury, failure, loss, end-stage renal failure) are employed, AKI is a common complication occurring in up to a third of ICU patients and is usually a manifestation of multiorgan failure syndrome.²⁵^–²⁷

The most common cause of intrinsic renal failure is ATN.³ Specific causes of ATN can be classified as hemodynamically mediated AKI, such as in prolonged prerenal azotemia, hypotension, and sepsis; toxic AKI, secondary to antibiotics, chemotherapeutic agents, and contrast media; or postsurgical AKI. In a large prospective analysis by Liano and coworkers, sepsis was the most common cause (35%); postsurgical (25%) and toxic (31%) causes were also common.¹⁰ Many, if not most, patients have a multifactorial cause of AKI (Figure 114-1). Despite ever-improving supportive interventions in the ICU, the mortality rate for AKI has not changed in the last 3 decades, remaining at 40% to 80% depending on the study.²⁸ It has been hypothesized that this continued poor prognosis is due to the changing patient population cared for in the ICU. Today, patients are older with greater comorbidities, and their renal disease most often develops in the setting of multiorgan failure.^10,²⁹ This high incidence of multiorgan failure has made it difficult to discern whether AKI itself causes increased mortality or whether it is a marker of severely ill patients. Several recent studies have found that AKI does in fact contribute to excess mortality in the setting of contrast nephropathy and cardiac surgery.^30,³¹ In those patients who do survive, there is significant morbidity, with about 33% requiring long-term renal replacement therapy (RRT) and 28% requiring long-term institutionalization.³² As explained later, increasing RIFLE severity grades correspond with increasing mortality in patients. Hoste et al. reported that patients with a maximum score of RISK had a mortality rate of 8.8%, compared to 11.4% for INJURY and 26.3% for FAILURE. On the other hand, patients who had no evidence of AKI had a mortality rate of 5.5%.³³

Figure 114-1 Pathophysiology of ischemic acute kidney injury (AKI). Interacting microvascular and tubular events contribute to the physiology of ischemic acute renal failure. PGE2, prostaglandin E2.

(Adapted from Bonventre JV, Weinberg JM. Recent advances in the pathophysiology of ischemic acute renal failure. J Am Soc Nephrol 2003;14:2199-210.)

The risk of developing AKI in the ICU was evaluated by de Mendonca and associates, who found that seven characteristics, if present on admission, were associated with a high risk of developing AKI (Table 114-2).²⁹ Several other studies addressed risk factors for mortality in the setting of AKI.^7,^10,^12,³⁴ As indicated in Table 114-3, the risk of death in those with AKI is increased by the presence of nonrenal organ failure; more severe renal dysfunction, as indicated by oliguria; sepsis; advanced age; and male gender. Liano and colleagues found that as the number of organ failures increased, mortality increased.¹⁰ With two organ failures, mortality was 53%; this increased to 80% with three organ failures and 100% with five organ failures.

TABLE 114-2 Risk Factors for Developing Acute Kidney Injury

Age >65 years

Infection on admission

Cardiovascular failure

Cirrhosis

Respiratory failure

Chronic heart failure

Lymphoma or leukemia

Adapted from de Mendonca A et al. Acute renal failure in the ICU: risk factors and outcome evaluated by the SOFA score. Intensive Care Med 2000;26:915-21.

TABLE 114-3 Risk Factors for Mortality in Acute Kidney Injury

Higher severity index score

Age > 65 years

Male gender

Oliguric acute renal failure

Sepsis

Nonrenal organ failure: (cardiovascular, hepatic or respiratory failure)

Thrombocytopenia

Mechanical ventilation

Prior compromised health status

To further stratify the probability of death in critically ill patients, several severity-of-illness scoring systems have been developed. These indices help compare patients enrolled in clinical trials and better utilize finite resources to help those patients with the best chance of recovery. In large populations, these scoring indices have been successful in predicting outcome ³⁵; however, they do not discriminate well in patients with AKI.³⁶ The renal parameters used in these scores consist of blood urea nitrogen (BUN), serum creatinine, and total urine output per day. With the latest version of the Acute Physiology and Chronic Health Evaluation (APACHE III), oliguric AKI constitutes just 12.7% of the maximal score, thereby underestimating the effect of AKI on mortality.³⁷ Further, there is no correction for patients with AKI and a low serum creatinine, who also have a poor outcome, probably reflective of poor nutritional status.³⁷ An attempt has therefore been made to develop more disease-specific indices, such as Liano and colleagues’ individual severity index, the Cleveland Clinic Foundation severity score, and the Project to Improve Care in Acute Renal Disease index. The majority of these indices were developed at single centers, and few have been validated outside the original institution. Also, the patient populations to which the indices were applied have differed, such as using all AKI patients or only dialyzed patients. Thus, there is no completely generalizable, validated bedside predictor for mortality in AKI patients.

Definition

Acute renal failure (ARF) has traditionally been defined as an abrupt decrease in GFR with resultant retention of urea and other nitrogenous waste products along with dysregulation of body fluids and electrolytes. However, this is only a qualitative definition and not very helpful clinically, where a quantitative definition is required. Until recently, no agreement existed about how to best define, characterize, and study acute renal failure. This lack of a standard definition has been a major hindrance to the progress of clinical and basic research in this field. The term acute kidney injury was proposed by the Acute Kidney Injury Network (AKIN) as an alternative to ARF in order to encompass the entire range of failure based on recent data showing that a small change in serum creatinine influences outcome. The Acute Dialysis Quality Initiative (ADQI) was created to develop consensus and evidence-based guidelines for treatment and prevention of acute renal failure, with the goal of comparing studies and advancing research.³⁸ The ADQI group proposed a consensus categorized definition—the RIFLE criteria³⁹—which were validated and shown to correlate with hospital mortality and patient outcomes in several populations in large international databases. Subsequently, AKIN proposed a revision of the RIFLE criteria^40,^41,⁴² to better account for small changes in serum creatinine not captured by RIFLE. The following modifications were made (Table 114-4):

1 The RIFLE severity grading system was reconfigured, with the R category becoming stage 1, I becoming stage 2, and F becoming stage 3, which also includes anyone who receives acute RRT, irrespective of their preceding serum creatinine increase or urine output.

2 The diagnosis of AKI could be made with a period of oliguria of at least 6 hours or a serum creatinine increase of ≥0.3 mg/dL from baseline. AKIN criteria caution that adequate volume resuscitation should be ascertained and urinary tract obstruction ruled out prior to using urine output to identify AKI. However, it should be noted that the change in urine flow is less helpful as a diagnostic criterion because of the high incidence of nonoliguric AKI.⁴³ Inclusion of the aforementioned absolute serum creatinine increase in the AKI definition was based on the repeated finding from several large studies that such serum creatinine increments are associated with increased mortality.^44,⁴⁵

3 Diagnostic increments of serum creatinine should occur during a period of no more than 48 hours, compared to 7 days for the RIFLE criteria.

TABLE 114-4 Contrast Between Acute Kidney Injury Staging By Rifle and Akin Systems*

Existing evidence supports the validity of both RIFLE and AKIN criteria to identify groups of hospitalized patients with increased risk of death and/or need for RRT.^39,^40,⁴⁶ Staging of AKI is relevant because with increased stage of AKI, the risk of death increases. Moreover, there is now mounting evidence of long-term risk of subsequent development of cardiovascular disease or chronic kidney disease and mortality even after resolution of AKI.⁴⁷ Lo et al. recently studied the long-term sequelae of AKI in a retrospective analysis of the large Kaiser Permanente database using the years 1996-2003.⁴⁸ This paper explored AKI and its correlation with long-term kidney disease and mortality in comparison with enrollees of the same healthcare organization who did not develop AKI and served as controls. Compared with controls, patients who suffered dialysis-dependent AKI during their hospitalization had a 28-fold increased risk of developing stage 4 or 5 CKD. There was also a more than twofold long-term risk of death in this group.

Given the difficulties of measuring function as an index of injury, there has been a search for identifying kidney injury markers of critically ill patients. This approach would be optimal because it could identify patients early in the course of AKI who would benefit from intervention. Several biomarkers have been proposed and are currently being investigated.⁴⁹^–⁵¹ These biomarkers include:

1 Serum cystatin C ^52,⁵³

2 Urinary interleukin (IL)-18 ⁵⁴ and tubular enzymes such as the intestinal form of alkaline phosphatase, N-acetyl-α-glucosaminidase, and alanine aminopeptidase.⁵⁵

3 Neutrophil gelatinase–associated lipocalin (NGAL)⁵⁶

4 Kidney injury molecule 1 (KIM-1)⁵⁷

Further studies are required to establish any of these or other potential biomarkers as practical diagnostic tools in the early clinical diagnosis of AKI. Such tools should facilitate timely and aggressive therapeutic interventions. Currently, urine sediment (i.e., epithelial cells) and urine indices (i.e., FE_Na) are early and sensitive harbingers of ATN.

Treatment

In light of its dismal outcome, it is imperative that therapies to prevent or ameliorate AKI be developed. To that end, several trials for both prevention and treatment of AKI have been conducted with multiple agents.

With the increasing use of contrast agents in diagnostic and therapeutic procedures, prevention of contrast-mediated nephropathy has been studied extensively. Intravenous fluids have long been used to prevent contrast nephropathy, but in patients with chronic renal insufficiency, the incidence is still high. Therefore, multiple other agents have been studied. Solomon and coworkers found that both furosemide and mannitol when given with saline produced a worse outcome than saline alone in patients with chronic renal insufficiency.⁵⁸ Dopamine⁵⁹ and atrial natriuretic peptide⁶⁰ have also failed to reduce contrast nephropathy. Two agents, acetylcysteine⁶¹ and fenoldopam,⁶² were found to decrease the incidence of contrast nephropathy in high-risk patients, but these findings were not verified in a study by Allaqaband and associates.⁶³ In that trial, acetylcysteine and fenoldopam offered no additional benefit in patients with chronic renal insufficiency undergoing cardiovascular procedures. Landoni et al. recently performed a meta-analysis of 16 randomized trials of fenoldopam versus placebo or dopamine in 1290 patients who were in a variety of ICU or perioperative settings. They found that fenoldopam reduced the need for renal replacement and mortality in patients with AKI. However, because of the small size and heterogeneity of the studies included, a large multicenter appropriately powered trial will be needed to better define the role of fenoldopam in AKI.⁶⁴ Currently, our recommendation for preventing contrast nephropathy in high-risk patients (Table 114-5) is adequate hydration, preferably with isotonic sodium bicarbonate,^65,^66,⁶⁷ administration of 1200 mg acetylcysteine orally twice daily the day before and day of the procedure (given its tolerability and relative low cost), and the use of low-osmolar⁶⁸ or iso-osmolar⁶⁹ contrast media.

TABLE 114-5 Risk Factors for Contrast Nephropathy

Preexisting renal impairment

Diabetes mellitus

Decrease in effective arterial volume (congestive heart failure, volume depletion, cirrhosis)

High dose of contrast media

Concurrent use of nephrotoxic agents (nonsteroidal antiinflammatory drugs, angiotensin-converting enzyme inhibitors)

Dopamine has long been used to treat AKI. The renal effects of dopamine include an increase in GFR and an increase in sodium and water excretion. Clinically, the first response is an increase in diuresis.⁷⁰ These responses occur in patients with normal renal function, but it is unknown whether they are also seen in those with AKI. In patients with early renal dysfunction (serum creatinine > 1.8 mg/dL or urine output < 0.5 mL/kg/h), dopamine did not alter peak serum creatinine or the need for RRT.⁷¹ This was confirmed in a meta-analysis to determine whether progression of AKI, need for RRT, or mortality were affected by dopamine.⁷²

Aside from its lack of efficacy in AKI, dopamine has deleterious side effects. It hastened the onset of gut ischemia in an experimental model,⁷³ and clinically it worsened contrast nephropathy.⁷⁴ In cardiac surgery patients, dopamine was independently associated with an increased risk of postoperative atrial fibrillation.⁷⁵ Higher doses may increase mortality,⁷⁶ perhaps by worsening myocardial ischemia.⁷⁷ Therefore, low-dose dopamine currently has no role in the treatment or prevention of AKI.

Diuretics are also frequently used in patients with AKI, especially in an attempt to convert oliguric into nonoliguric AKI, given the improved prognosis of the latter.⁷⁸^–⁸⁰ Loop diuretics, most commonly furosemide, inhibit Na⁺/K⁺-ATPase in the thick ascending loop of Henle and therefore decrease the active reabsorption of sodium. Theoretically, this has some potential benefits, such as decreasing energy expenditure and increasing flow rate to flush out tubular casts. In the experimental setting, loop diuretics can be protective if administered before the insult. However, even when patients are successfully converted to nonoliguria, there is no reduction in the need for RRT or mortality.^81,⁸² Cantarovich and colleagues studied the role of high-dose loop diuretics in a placebo-controlled clinical trial of 388 dialysis-requiring AKI patients. Despite the increase in urine output, there were no differences between the two groups in terms of patient survival, renal recovery rates, number of dialysis sessions required, or time on dialysis. In addition, cardiac surgery patients and patients with contrast nephropathy who were treated with furosemide had a worse outcome.^83,⁸⁴ A study by Mehta and coworkers found an increased mortality in AKI patients treated with diuretics.⁸⁵ It is unclear why this occurred, but the authors speculated about a possible nephrotoxic effect of diuretics or a delay in the initiation of RRT because of increased urine output. However, the increased mortality occurred in patients who were not diuretic responsive, likely because of more severe AKI. These patients already had a worse prognosis, and whether diuretics may have worsened the outcome is unknown. Ho et al. recently conducted a comprehensive systematic review of the use of furosemide in AKI.⁸⁶ They have shown that furosemide is not associated with any significant clinical benefits in the prevention or treatment of ARF in adults. High doses may be associated with an increased risk of ototoxicity. Although the use of loop diuretics in early or established AKI facilitates management of fluid balance, hyperkalemia, and hypercalcemia, and is indicated for these clinical purposes, any putative role in prevention or amelioration of AKI course is unproven. Therefore, if diuretics are temporarily employed for such indications, care must be taken to avoid delaying initiation of dialysis if clinically necessary.

Atrial natriuretic peptide is a hormone secreted by the cardiac atria that increases GFR and glomerular filtration pressure by dilating the afferent arteriole and constricting the efferent arteriole.⁸⁷ It also decreases tubular reabsorption of sodium and chloride,⁸⁸ redistributes medullary blood flow,⁸⁹ disrupts tubuloglomerular feedback,⁹⁰ and reverses endothelin-induced vasoconstriction.⁹¹ Mentzer and colleagues studied the perioperative effects of nesiritide (BNP type) in 303 patients with left ventricular dysfunction who were undergoing coronary artery bypass graft.⁹² They demonstrated short-term benefits of nesiritide on perioperative renal function as assessed by an attenuated increase in levels of serum creatinine, a reduction in calculated GFR loss, and a greater urine output 24 hours after surgery. This trial and other reports that have studied administration of natriuretic peptides during cardiac surgery were recently reviewed by Murray, who emphasized that in addition to such surrogate renal endpoints, future studies must demonstrate beneficial effects on overall survival and/or dialysis-free survival.⁹³ Pending further studies, atrial natriuretic peptide cannot be recommended for prevention or therapy of ATN.

Hemodynamic Management

Intravascular volume is critical in maintaining hemodynamic stability, tissue oxygenation, and organ function.⁹⁴ In critically ill patients, it is increasingly being recognized that accurate assessment of volume status and appropriate use of fluid replacement may lead to better outcomes. In a study by Rivers and associates, it was shown that early goal-directed therapy (EGDT) based on optimizing the mixed control venous oxygen saturation in the first 6 hours resulted in decreased mortality in septic patients.⁹⁵ Subsequent studies have replicated those results,^96,⁹⁷ and one of them showed a significantly improved prevention of AKI in patients randomized to EGDT compared to the standard care group.⁹⁸ However, supranormal levels of cardiac index or mixed venous oxygen saturation did not decrease mortality.⁹⁹ In addition, studies have shown increased mortality in patients with positive fluid balance and acute respiratory distress syndrome (ARDS).¹⁰⁰^–¹⁰²

We have coined the term pseudo- or pre-ARDS to focus on a common and clinically important situation in ICUs. Just as prolonged prerenal azotemia may eventually lead to ischemic ATN, prolonged pseudo- or pre-ARDS may lead to ARDS in association with evidence of pulmonary capillary damage and stiff lungs, as diagnosed clinically by a decrease in pulmonary compliance. Thus, pseudo- or pre-ARDS describes a clinical syndrome of noncardiogenic pulmonary edema in the absence of evidence of decreased pulmonary compliance. Although many clinicians group these clinical entities together as ARDS—independent of pulmonary compliance—we believe that from a pathophysiologic, prognostic, and therapeutic viewpoint, these clinical entities may be substantially different.

Both pseudo-ARDS and ARDS are frequently associated with sepsis. Sepsis is a vasodilated state in which systemic vascular resistance decreases and cardiac output increases. Studies in renal experimental animals have shown that vasodilatation with an arterial vasodilator such as minoxidil is associated with an increased albumin distribution space and a failure of interstitial hydrostatic pressure to rise during saline administration.¹⁰³ These changes in interstitial Starling forces favor an increase in interstitial fluid volume during saline infusion. We frequently consult on ventilated ICU patients with AKI who have a 20-L positive fluid balance that has not been recognized in a quantitative sense because the pulmonary capillary wedge pressures are not considered elevated (<18 mm Hg). Excess saline fluid has been administered to resuscitate these vasodilated septic patients, leading to pulmonary edema, hypoxia, and ventilatory support. In the early stages, the majority of these patients do not have decreased pulmonary compliance (i.e., stiff lungs). However, these septic ICU patients with renal failure on prolonged respiratory support ultimately have a mortality as high as 80%. Patient mortality has been reported to begin increasing after 48 hours on a respirator. The potential barotrauma, oxygen toxicity, and pulmonary infections that may occur with prolonged ventilatory support frequently lead to stiff lungs and what virtually all authorities would term bona fide ARDS.

We believe that not distinguishing clinically between pseudo-ARDS and ARDS may be detrimental to ICU patients. Marked improvement in the pulmonary edema of pseudo-ARDS by diuresis or ultrafiltration may allow much earlier extubation and removal of ventilatory support before the development of pulmonary capillary damage and stiff lungs (i.e., ARDS). With ARDS and prolonged ventilatory support, a very high mortality occurs, particularly in the presence of renal failure and thus multiorgan failure. Recently, Bouchard and colleagues have reported results of a prospective multicenter observational study of 618 patients that aimed to determine whether fluid overload (>10% increase in body weight) in critically ill patients with AKI is associated with increased mortality. After adjustment for severity of illness, the study has shown that fluid overload was independently associated with mortality in those AKI patients who did and did not receive dialysis therapy.¹⁰⁴ A randomized study by the ARDS clinical trials network demonstrated that pulmonary function in critically ill patients was worse in those treated with a liberal fluid management strategy (to achieve a mean central venous pressure [CVP] of ∼12 mm Hg) than in those who were treated with a conservative strategy (to achieve a mean CVP ∼8 mm Hg).¹⁰⁵ Moreover, fewer patients in the conservative strategy group required dialysis than in the liberal strategy group. Several pediatric studies comprising more than 400 children have demonstrated an association between worsening fluid overload (higher than 10% to 20%) and mortality.¹⁰⁶^–¹⁰⁸ Thus, there are reasons to believe that fluid overload is not just a marker but rather a pathologic factor in the high mortality of critically ill patients with AKI. Prospective randomized clinical trials will be needed to confirm this possibility. Until such studies are available, however, we recommend the avoidance of fluid overload in patients with AKI on the basis of knowledge of body weight changes and cumulative fluid balance for these patients.¹⁰⁹

To aid in appropriate hemodynamic support, invasive monitoring has been used to guide therapy. Techniques such as the pulmonary artery catheter rely on measurement of filling pressures (e.g., CVP, pulmonary artery occlusion pressure) to estimate preload responsiveness. In critically ill patients, the relation between filling pressures and ventricular end-diastolic volume (preload) is often obscured by changes in ventricular compliance or changes in the pericardium or thorax.¹¹⁰ In addition, the pulmonary artery catheter has been linked to a worse outcome in patients.^111,¹¹² A positive response to fluid challenge can be predicted in mechanically ventilated patients by analyzing respiratory variations in pulse pressure. It has been shown that a change in pulse pressure greater than 15% during a single breath is more accurate in predicting an increase in cardiac output in response to volume loading than either right atrial pressure or wedge pressure.^113,¹¹⁴

Fluid management in critical illness is aimed at improving organ perfusion. However, in inflammatory states such as sepsis, there may be major fluid shifts resulting in tissue edema despite intravascular depletion. Aside from the inflammatory cascade, vasodilatation itself can result in an increase in interstitial fluid volume, likely secondary to albumin escape from the vasculature.¹¹⁵ There are currently no clinical methods to detect the presence of capillary leak, apart from fluid administration having no effect on intravascular volume.¹¹⁰ Therefore, if only transient improvements in hemodynamics occur with fluid administration, or if there is a continuing need for fluid, it is likely the patient will best be served by a change to vasopressor agents.

When volume replacement is indicated, there is controversy over the optimal type of fluid. Crystalloids are the most common form of volume replacement, but their effect on plasma volume is limited. Each liter of fluid administered increases plasma volume 200 mL, but the intravascular half-life is only 20 to 30 minutes.⁹⁴

Colloidal substances such as albumin, dextran, and hydroxyethyl starches, because they are macromolecules, are better retained within the intravascular space and have a greater effect on plasma volume. Albumin has been used for decades, but it is expensive and may cause an increase in mortality, according to the Cochrane Injuries Group.¹¹⁶ Nevertheless, a randomized controlled trial was conducted to compare human albumin with crystalloid in ICU patients (Saline versus Albumin Fluid Evaluation [SAFE] study). It indicated that albumin is safe, albeit no more effective than saline for fluid resuscitation. SAFE demonstrated no difference in renal outcomes, at least based on duration of RRT.¹¹⁷ Dextran cannot be recommended for plasma volume expansion because of serious side effects such as coagulation abnormalities¹¹⁸ and AKI.¹¹⁹

Hydroxyethyl starches (HESs) are polymers of amylopectin that vary in molecular weight and number of substitutions of hydroxyethyl groups. As molecular weight and number of substitutions increase, side effects also increase. HES 200/0.5 is a compound with a middle molecular weight and low substitution number. It has been studied in a number of situations such as perioperative volume replacement, cardiac surgery, trauma, and sepsis.¹²⁰^–¹²² A recent trial compared a “modern” HES preparation with a low-molecular-weight and low-molar substitution and a human albumin solution, given in cardiac surgery patients with preoperative compromised kidney function, showed that this type of HES solution had no negative influence on kidney integrity.¹²³ In another study (Efficacy of Volume Substitution and Insulin Therapy in Severe Sepsis [VISEP]), severely septic patients were randomly assigned to receive either 10% pentastarch, a low-molecular-weight hydroxyethyl starch (HES 200/0.5), or modified Ringer’s lactate for fluid resuscitation. HES appeared to be harmful, leading to higher rate and longer duration of AKI, and its toxicity increased with accumulating doses.¹²⁴ Aside from coagulation disorders, all hyperoncotic colloids may induce a pathologic entity known as osmotic nephrosis with potential impairment of kidney function.¹²⁵ A systematic review of randomized controlled trials (RCTs) on the use of HES for fluid management in patients with sepsis (totaling 1062 patients) showed an almost twofold increased risk of AKI with HES compared with crystalloids.¹²⁶ Lastly, a recent comprehensive Cochrane review concluded there is no evidence from RCTs that resuscitation with colloids instead of crystalloids reduces the risk of death in patients with trauma, burns, or following surgery.¹²⁷ There is even evidence that colloids may be associated with a higher incidence of AKI. Given the relative efficacy and safety of crystalloids, it is prudent to utilize them in fluid resuscitation and limit colloid use to the framework of clinical trials.

In sepsis and septic shock, there is hypotension despite normal or increased cardiac output.¹²⁸ The hypotension in sepsis is often unresponsive to fluid and requires administration of vasopressor agents. Because these agents cause vasoconstriction, there has been concern about their use in AKI. Norepinephrine causes a reduction in renal blood flow in healthy animals and humans.¹²⁹ The ultimate effect of norepinephrine on renal blood flow, however, depends on the resulting increase in blood pressure and vascular resistance. Norepinephrine increases blood pressure via an α₁-mediated increase in systemic vascular resistance and a β₁-mediated increase in cardiac output. The increase in resistance can potentially decrease cardiac output by increasing afterload. In the kidney, the effect on renal vascular resistance depends on the increase in systemic pressure, with a decreased renal sympathetic tone causing vasodilatation as well as an autoregulatory vasoconstriction secondary to increased perfusion pressure and α₁-mediated renal vasoconstriction.¹³⁰ In a nonrandomized study, it was demonstrated that norepinephrine increased arterial blood pressure, urine output, and GFR.¹³¹ A large randomized trial comparing dopamine to norepinephrine as initial vasopressor in patients with septic shock showed no significant differences between groups with regard to renal function or mortality, though norepinephrine was associated with less tachycardia in the first hours and was superior with regard to survival in cardiogenic shock patients (De Backer et al., in press).

Vasopressin is a hormone secreted by the posterior pituitary; it increases systemic vascular resistance by activating V_1a receptors on vascular smooth muscle. During septic shock, there is a biphasic response, with early high levels of endogenous vasopressin followed by a decrease.¹³² The renal effects of vasopressin are complex and involve an interplay between V₁ and V₂ receptors that regulates the antidiuretic function of vasopressin.¹³² Vasopressin is gaining attractiveness in the treatment of norepinephrine refractory shock patients.¹³³ It increases blood pressure and enhances diuresis in hypotensive oliguric patients but has not yet been proven to enhance survival nor been shown to prevent or ameliorate AKI in the critically ill.¹³⁴

It has been proposed that tight glycemic control can reduce the incidence and severity of AKI in critical patients. Recently Schetz et al. combined the renal endpoints of patients in a secondary analysis of two large randomized clinical trials.¹³⁵ They demonstrated that tight glycemic control significantly reduced the incidence of severe AKI from 7.6% to 4.5%. The need for RRT was not decreased in the overall population, but it was significantly lower in surgical ICU patients than in medical ICU patients. However, further studies have highlighted significant concerns regarding the effectiveness and safety of using intensive insulin therapy with tight glycemic control to prevent or ameliorate morbidity and mortality of AKI and other forms of organ injury. The international Normoglycemia in Intensive Care Evaluation and Survival Using Glucose Algorithm Regulation (NICE-SUGAR) study was recently published.¹³⁶ This large trial enrolled over 6000 patients and set out to definitively determine the risk/benefit of tight glycemic control in critically ill patients. It showed that in contrast to conventional insulin therapy, intensive glucose control increased mortality among these patients. A blood glucose target of ≤180 mg/dL resulted in lower mortality than a target of 81 to 108 mg/dL. It may therefore be prudent to use conventional insulin therapy in ICU patients at risk of AKI to target plasma glucose of less than 150 mg/dL, using a protocol to avoid hypoglycemia.

Nutritional Support

Nutritional support in patients with AKI does not differ significantly from that of critically ill patients in general. The goals of nutritional support are preservation of lean body mass, stimulation of immune competence, repair, and wound healing. AKI affects water, electrolyte, and acid-base balance, but it also induces a change in protein, carbohydrate, and lipid metabolism.¹³⁷ In patients with uncomplicated renal failure, oxygen consumption is approximately that of normal subjects. In the presence of sepsis or multiorgan failure, however, oxygen consumption is increased 20% to 30%.¹³⁸ Therefore, energy expenditure is determined more by the underlying disease. Energy substrate should not exceed this requirement, and it is better to err on the side of slight underfeeding than overfeeding. Patients with AKI should be supplemented with 20 to 30 kcal/kg body weight per day. Even in hypermetabolic states such as sepsis, energy expenditure is rarely greater than 130% of calculated basic energy expenditure. In a randomized trial in AKI patients, comparing 30 and 40 kcal/kg/d energy provision, the higher energy prescription did not induce a more positive nitrogen balance but was associated with a higher incidence of hyperglycemia and hypertriglyceridemia and more positive fluid balance.¹³⁹ Therefore, supplementation should not exceed 30 kcal/kg body weight per day.

The hallmark of metabolic alterations in AKI is activation of protein catabolism and release of amino acids from skeletal muscle. This process is responsible for the negative nitrogen balance encountered in critically ill patients. An underlying mechanism of protein catabolism is insulin resistance, which may be associated with increased mortality in AKI patients.¹⁴⁰ Plasma insulin levels are elevated, but maximal insulin-stimulated glucose uptake is decreased by 50%. This insulin resistance leads to stimulated hepatic gluconeogenesis fueled by protein catabolism.¹⁴¹ The elevated level of gluconeogenesis coupled with insulin resistance also frequently leads to hyperglycemia. Other factors such as inflammatory cytokines (namely, tumor necrosis factor) and catecholamines are also involved in the hypercatabolism.¹⁴² To combat malnutrition in this setting, it is often necessary to use nutritional supplementation in the form of enteral or parenteral feeding.

Enteral nutrition has become the standard form of nutritional support in critically ill patients. Enteral feeding helps maintain gastrointestinal function, including acting as a barrier to microorganisms. Two clinical studies have suggested that enteral feeding is associated with improved outcome and survival in ICU patients.^143,¹⁴⁴ A meta-analysis by Heyland and colleagues reviewed 26 randomized trials comparing total parenteral nutrition with standard care and found no survival benefit and possible harm in medical ICU patients fed parenterally.¹⁴⁵ Therefore, enteral support is recommended in critically ill patients with or without AKI.

Traditionally, nutrition has been delivered in the form of 50% to 80% carbohydrates. Recently, this has been the subject of study. In addition to providing calories, lipids also provide essential fatty acids. Essential fatty acids such as omega-3 polyunsaturated fatty acids and amino acids such as arginine have been found to stimulate the immune system. A prospective randomized trial of “immune-enhancing” enteral nutrition found that in patients who received adequate nutrition, those who received the immune-enhancing diet had a decrease in mortality and hospital stay.¹⁴⁶ This study did not address AKI patients, but it is likely they would also benefit.

In the past, protein restriction was employed in AKI patients to control uremia, but this is likely to be detrimental to the patient and results in a profoundly negative nitrogen balance.¹⁴⁷ With the advent of continuous modalities of RRT, it is possible to adequately supplement protein and control uremia. Therefore, some authors recommend aggressive protein replacement at 2.5 g/kg/d, as opposed to the standard 1 to 1.5 g/kg/d.¹⁴⁷ However, no compelling data are currently available concerning the efficacy and safety of such high protein intakes. Also, it is important to realize that hypercatabolism cannot be simply overcome by increasing protein or amino acid intake. We suggest administering 0.8 to 1.2 g/kg/d of protein in patients with AKI without the need for dialysis, and 1 to 1.5 g/kg/d in patients with AKI on RRT.

Indications for Nephrology Consultation

Currently there are wide variations in the timing of nephrology consultation in patients with AKI. Some physicians prefer to consult at the first rise in serum creatinine, whereas others wait until RRT is needed. In a study evaluating the effect of nephrology consultation on patient outcome, Mehta and associates found that a delay in nephrology consultation (>48 hours after ICU admission with AKI) led to higher mortality.¹⁴⁸ In this study, patients with delayed consultation had a lower serum creatinine concentration and higher urine output but more organ failure and higher total body water. In the multivariate analysis, delayed consultation was no longer significant, but the trend was there. Why would early consultation affect mortality? It could result from delayed recognition of renal failure. Higher total body water likely leads to tissue edema and organ dysfunction (i.e., pulmonary edema), so in ICU patients, early recognition of AKI and its appropriate management may lead to better outcomes.

Renal Replacement Therapy

Indications

As mentioned previously, up to a third of patients in the ICU develop AKI. Of those, 30% to 70% require RRT.^25,^26,²⁷ Many practitioners delay initiating RRT as long as possible because of concerns that dialysis may delay the recovery of renal function.^149,¹⁵⁰ The optimal timing of initiation of dialysis is not defined. There is little disagreement in commencing dialysis in the presence of life-threatening conditions such as diuretic-resistant volume overload, severe hyperkalemia, acidosis, azotemia, or overt symptoms and signs of uremia such as encephalopathy and pericarditis. Medical treatment approaches for hyperkalemia accomplish intracellular shifts. When intermittent hemodialysis is used to correct hyperkalemia after such measures have been utilized, dialytic potassium removal will be reduced, and greater levels of post-dialysis potassium can occur.¹⁵¹ Metabolic acidosis is common in severe AKI but can be corrected with bicarbonate and should rarely require urgent dialysis if not accompanied by volume overload or uremia.¹⁵² Some poisons, drug overdose, and toxic compounds can contribute to acid-base disturbances and AKI. In such cases, dialysis can be supportive and facilitate removal of these substances and their metabolites. In acute salicylate poisoning, RRT is indicated when serum concentration is above 100 mg/dL and the patient exhibits altered mental status, pulmonary or cerebral edema, renal impairment, fluid overload that prevents administration of sodium bicarbonate, or clinical deterioration despite aggressive and appropriate supportive care.¹⁵³ Ethylene glycol and methanol poisoning are important causes of anion-gap metabolic acidosis. Dialysis treatment has been shown to reduce development of subsequent AKI and organ dysfunction.¹⁵⁴ Metformin-associated lactic acidosis may be an indication for dialysis, especially in critical patients who are more prone to death. According to a recent study, these are patients who show a low pH (<6.9) and high serum lactate and metformin concentrations.¹⁵⁵

The level of azotemia at which RRT should begin is unknown. Several early retrospective studies that used blood urea or BUN suggested that early initiation of RRT resulted in survival improvements.^156,¹⁵⁷ More recent studies have continued to focus on BUN as the marker for starting dialysis. Single-center observational studies that were restricted to AKI after trauma¹⁵⁸ and coronary artery bypass surgery^159,¹⁶⁰ suggested a benefit to dialysis initiation at lower BUN concentrations. A prospective multicenter observational study analyzed dialysis initiation, as inferred by BUN concentration, in 243 geographically and ethnically diverse patients.¹⁶¹ Survival rates were slightly lower for patients who started dialysis at higher BUN concentrations, despite a lower burden of organ system failure. In a prospective multicenter observational trial study conducted at 54 ICUs in 23 countries, timing of RRT was stratified into “early” or “late” by median urea at the time RRT started and also categorized temporally from ICU admission into early (<2 days), delayed (2-5 days), or late (>5 days).¹⁶² Timing by serum urea showed no significant difference in mortality. However, when timing was analyzed in relationship to ICU admission, late RRT (this may also be late AKI) was associated with greater crude mortality and covariate-adjusted mortality. Overall, late RRT was associated with a longer duration of RRT and stay in hospital and greater dialysis dependence.

Serum concentrations of BUN and creatinine are recognized to be inherently subject to a multitude of factors other than kidney function, such as catabolic rate, volume status, age, race, and muscle mass. It would therefore be prudent not to base dialysis initiation decision on a single BUN and creatinine threshold, but rather on the broader clinical context and trends of laboratory tests. Finally, it is important to consider the volume status when deciding the time for initiating RRT, because volume overload, as previously elaborated, emerged as an important factor associated with mortality in AKI. Table 114-6 depicts accepted indications for initiating RRT in the ICU.

TABLE 114-6 Potential Indications for Renal Replacement Therapy in the ICU

* Intermittent hemodialysis removes K⁺ more efficiently than continuous modalities.

Adapted from Bellomo R, Ronco C. Continuous haemofiltration in the intensive care unit. Crit Care 2000;4:339-45.

Adequate Dosing

In chronic hemodialysis patients, adequacy of dialysis is primarily determined by the level of small-solute (urea) clearance. This is determined by the Kt/V formula, where K is the dialysis membrane clearance of urea, t is the time on dialysis, and V is the volume of distribution of urea, which is equal to total body water. In chronic hemodialysis, a Kt/V of 1.2 per session is considered adequate.¹⁶³ As can be seen from the formula, to increase urea clearance, one can increase the time on dialysis or increase the dialyzer clearance. Dialyzer clearance depends on blood flow and dialysate flow rates, as well as the inherent properties of the membrane.

In the United States, intermittent hemodialysis and continuous RRT are the most commonly used modalities of RRT, with sustained low-efficiency dialysis and other “hybrid” treatments used in fewer than 10% of patients. Intermittent hemodialysis is most commonly provided on a thrice-weekly or every-other-day schedule.¹⁶⁴ Concerning intermittent modalities, there is no standard Kt/V for adequate dialysis in AKI currently, but it has been suggested that a higher target Kt/V confers better patient outcomes. Schiffl and colleagues studied 160 patients with AKI who were divided into two groups: one received daily hemodialysis, and the other alternate-day hemodialysis. It was found that daily hemodialysis resulted in less hypotension, sepsis, gastrointestinal bleeding, and respiratory failure, as well as a significant decrease in mortality.¹⁶⁵ This study has been criticized because the Kt/V delivered to the alternate-day group was only 0.94, which is significantly less than the prescribed dose of 1.2. Therefore, the results could be explained by the fact that the alternate-day group received inadequate dialysis. In contrast, the VA/NIH Acute Renal Failure Trial Network Study (ATN study) did not find a benefit for a more intensive dosing strategy for RRT.¹⁶⁶ This study compared intermittent hemodialysis (hemodynamically stable patients) or sustained low-efficiency dialysis (hemodynamically unstable patients) performed 3 (less intensive) versus 6 (more intensive) times a week in 563 critically ill patients with AKI and failure of at least one nonrenal organ or sepsis. The prescribed Kt/V per session was 1.2 to 1.4, and the actual delivered mean dose was 1.3 in the less intensive arm. The 60-day mortality rate and percentage of patients recovering renal function were similar in both groups. The Hannover Dialysis Outcomes study was a prospective randomized parallel group study that used intensified extended dialysis (dosed to maintain plasma urea levels <90 mg/dL) versus standard dialysis (dosed to maintain plasma urea levels between 120 and 150 mg/dL) on 14- and 28-day mortality and renal function.¹⁶⁷ Mortality and frequency of renal function recovery were similar between the two groups. Based on these two well designed and performed clinical trials, it appears that increasing urea target clearances does not improve mortality or rates of renal recovery. Therefore, at least the smaller dose used in these trials should be pursued, with monitoring of the delivered dose of therapy to ascertain a minimum delivered Kt/V of 1.2 per treatment, or maintenance of plasma urea around 110 mg/dL when using extended or intermittent RRT in AKI patients. The significant difference between prescribed and delivered dialysis dose was studied by Evanson and coworkers, who found that the prescribed dose was a Kt/V of 1.25, whereas the dose delivered was only 1.04.¹⁶⁸ These authors found that the most significant factor predicting actual delivered dose was the patient’s predialysis weight. It follows that a higher weight in critically ill patients represents higher total body water and therefore a larger volume of distribution of urea. This would be expected to decrease the Kt/V if it were not accounted for in the prescription of dialysis.

CRRT has been advocated in patients with AKI because of its ability to more efficiently remove solute ¹⁶⁹ and provide hemodynamic stability,¹⁷⁰ but the optimal dosing in CRRT is not known. This form of RRT depends on convection, not diffusion, for solute clearance, which means that there is no dialysate involved, and solute is removed with water during ultrafiltration. Ronco and colleagues randomized 425 patients with AKI to increasing doses of continuous venovenous hemofiltration (CVVH).¹⁷¹ These investigators used three increasing doses of ultrafiltration—20, 35, and 45 mL/kg per hour—and found that mortality was 41%, 57%, and 58%, respectively. Survival was significantly lower in the 20 mL/kg group than in the other two groups. If the patient was septic, using the highest dose was beneficial.^170,¹⁷¹ A more recent study by Bouman et al. in severely ill, ventilated patients with high severity scores was unable to detect a difference in mortality between high ultrafiltration volume (48 mL/kg/h) and low ultrafiltration volume (20 mL/kg/h).¹⁷² In the mentioned VA/ATN study, an approximate number of 201 patients received predilutional continuous venovenous hemodiafiltration (CVVHDF) in the less intensive arm (mean delivered effluent of 22 mL/kg/min) and 179 in the intensive therapy group (mean delivered effluent of 36 mL/kg/min).¹⁶⁶ A higher dose of CRRT did not influence either mortality or renal recovery. The recently reported Randomized Evaluation of Normal versus Augmented Level of RRT (RENAL) study was conducted in 35 centers in Australia and New Zealand.¹⁷³ It compared the effects of postdilutional CVVHDF doses of 25 and 40 mL/kg/h in 24-day and 90-day mortality rates of 1508 critically ill AKI patients. Treatment with higher-intensity regimen did not reduce mortality at 90 days. In conclusion, the results of these two recent well-designed and executed large clinical trials (ATN and RENAL) did not show any benefit of higher CRRTs doses for critical AKI patients beyond a threshold dose necessary to optimize clinical outcome. Therefore, when using CRRT for treating such patients, a minimum dose to be targeted may be the minimal efficient one proved in those trials: 20 to 25 mL/kg/h. However, it is important to pay careful attention to ensure that interruptions of treatment in the ICU are minimized. Beyond small-solute clearance, other aspects of dialysis adequacy such as volume management should be well attended to.

Modality

When RRT is indicated in the ICU for severe AKI, physicians have to choose between intermittent techniques such as traditional intermittent hemodialysis (IHD; used in end-stage kidney disease), slow low-efficiency dialysis (SLED), or continuous therapies such as CVVH and peritoneal dialysis (PD) (Table 114-7). SLED is performed by utilizing dialysis machines to deliver a slow dialysate flow for periods ranging from 8 to 12 hours per day. Advantages with this technique include high hemodynamic tolerance, excellent solute-removal capability, and capacity to be instituted using regular hemodialysis machines without acquiring new equipment. Availability and expertise with the technique, as well as the hemodynamic status of the patient, are typical determining factors for modality choice for AKI.

TABLE 114-7 Practical Comparison of Acute Renal Replacement Therapy Modalities

RRT is required in severe AKI to remove uremic toxins and maintain fluid, electrolyte, and acid-base balance. CRRT and IHD are effective therapies that may be utilized and exchanged according to the hemodynamic status or coagulation problems of the patient. The effect of these modalities on patient outcomes has been evaluated. Lins et al. performed a multicenter randomized controlled trial to study the effect of intermittent versus continuous dialysis modalities for the treatment of 316 AKI patients who were admitted to the ICU.¹⁷⁴ They demonstrated that ICU stay, hospitalization, mortality, and renal recovery rates were not different between the groups. Moreover, two recent systematic reviews that collectively analyzed 45 studies found that outcomes were similar in critically ill AKI patients (stratified according to severity of illness) with CRRT and IHD for hemodynamically stable patients for relative risk of death, ICU mortality, in-hospital mortality, length of hospitalization, and requirement for chronic dialysis or renal recovery in survivors.^175,¹⁷⁶

Control of both uremia and volume is the major goal of RRT in AKI. A few studies have suggested that CRRT has advantages over intermittent therapies, including hemodynamic stability, improved survival, greater likelihood of renal recovery,^175,^177,¹⁷⁸ and better fluid balance.¹⁷⁹ IHD is complicated by hypotension in 20% to 30% of patients¹⁸⁰; in hemodynamically unstable patients, this can significantly limit therapy and delay recovery of renal function. Therefore, some clinicians favor initiating CRRT for hemodynamically unstable patients with AKI, but this has not been supported by a prospective randomized trial¹⁸¹ or the earlier-mentioned systematic reviews.^175,¹⁷⁶ Moreover, Bagshaw et al. recently performed a systematic review and meta-analysis of nine randomized trials and concluded that it is impossible to make definitive recommendations about the initial RRT modality because of numerous issues related to study design, conduct, and quality of these trials.¹⁸² The main disadvantage of CRRT is the need for prolonged anticoagulation. SLED is gaining popularity as an intermittent modality in ICUs because of the aforementioned multiple advantages. Two RCTs that compared SLED with CVVH or CVVHD^183,¹⁸⁴ found similar outcomes with regard to hemodynamic stability and uremic clearance; furthermore, a decreased anticoagulation requirement was reported for SLED.¹⁸⁴ Based on such evidence, all these modalities should be viewed as complementary. CRRT or SLED may be utilized for severe AKI with hemodynamic instability and transitioned to IHD once stability is attained. Peritoneal dialysis is an alternate modality for AKI where vascular access may be difficult, in conditions where anticoagulation may be problematic, in under-resourced regions, or following large disasters with mass casualties.¹⁸⁵ A prospective randomized study of daily IHD versus PD in 120 AKI patients showed no difference in survival or recovery of renal function.¹⁸⁶ These results contrast to a previous study that showed decreased survival associated with PD in comparison to CVVH¹⁸⁷ and suggest that PD remains an acceptable option to CRRT when dosed appropriately.

In certain situations, CRRT is preferable to IHD, including in patients with or at risk for increased intracranial pressure. Studies have shown that CRRT prevents the increase in intracranial pressure associated with intermittent RRT.^188,¹⁸⁹ The use of CRRT in patients with severe sepsis or septic shock has also received much attention. Sepsis is associated with hemodynamic instability, making CRRT an attractive option. It has been shown that CRRT has beneficial effects on hemodynamics in animal models of sepsis.¹⁹⁰ This is thought to be secondary to the removal of inflammatory cytokines by both convective and adsorptive measures. Hemofiltration membranes allow the ultrafiltration of mid-molecular-weight molecules such as cytokines. Further, the continuous blood/membrane contact allows the membrane to adsorb more mediators. There is some evidence that hemofiltration may provide some benefit in those with sepsis and AKI.^191,¹⁹² Larger and adequately powered studies are needed to better define the role of this modality in AKI and sepsis.

Despite potential hemodynamic advantage over IHD, CRRT has some disadvantages as well. With CRRT, there is generally a need for continuous anticoagulation to prevent clotting of the filter. Although this is usually done with low-dose heparin, there is the risk of bleeding or heparin-induced thrombocytopenia. When a patient is a bleeding risk, a trial of no anticoagulation can be carried out, or regional anticoagulation with citrate is used in some centers. CRRT also requires more nursing support and is considerably more expensive than IHD.¹⁹³

Dialysis Membrane

Early dialysis membranes were made of cellulose or its derivatives, and it has been shown that the hydroxyl radicals on the cellulose membranes were able to activate the complement system.¹⁹⁴ These older membranes were thus not biocompatible. Newer synthetic polymers are less able to activate the complement cascade and also have the ability to bind activated complement, thereby decreasing systemic effects.¹⁹⁵ Because of this decrease in immune activation, these membranes are considered biocompatible. In CRRT, there is continuous contact between the blood and membrane, making this interaction quite important. Despite mechanistic advantages, a recent meta-analysis found no advantage to biocompatible versus bio-incompatible membranes in terms of adult patient mortality and kidney function recovery rates.¹⁹⁶

Membranes may also play a role in blood purification beyond that of solute clearance. There is interest in the ability of membranes to adsorb and bind cytokines from the blood. This is particularly attractive in sepsis, when there is dysregulation of the immune system with both pro- and antiinflammatory effects. A recent prospective RCT evaluated the use of polymyxin B hemoperfusion in patients with abdominal sepsis to reduce circulating endotoxin levels and improve clinical outcomes. Patients were randomized to conventional therapy or conventional therapy plus two sessions of polymyxin B hemoperfusion. The study showed that polymyxin B hemoperfusion added to conventional therapy significantly improved hemodynamics and organ dysfunction and reduced 28-day mortality.¹⁹⁷ Abundant clinical studies show an association between inflammation and mortality in patients with AKI. Some investigators have suggested a potential role for using CRRT^198,¹⁹⁹ or IHD²⁰⁰ to remove cytokines and/or diminish the inflammatory response in septic patients with AKI. Such an approach is still experimental, and inadvertent removal of other potentially desirable middle molecules may be undertaken.

Buffer

In determining the adequacy of dialysis, factors other than solute clearance must be considered. One goal of RRT is to maintain normal acid-base balance in patients with AKI to prevent the complications of acidemia with regard to cardiovascular performance, hepatic metabolism, and hormonal response. To maintain normal pH, the dialysate must contain a buffer. Traditionally, the buffer choice was between bicarbonate and lactate, which metabolizes in the liver to bicarbonate on an equimolar basis under physiologic conditions. However, in critically ill patients with organ dysfunction and disordered tissue perfusion, it is possible that not all the anion will be converted to bicarbonate, resulting in increased serum lactate levels. Moreover, the increased lactate, without its redox partner pyruvate, can result in increased protein catabolism, myocardial depression,²⁰¹ and worsening acidosis in patients with preexisting lactic acidosis.^202,²⁰³

Bicarbonate-based solutions are currently the buffer of choice and are available in separated solutions that are mixed just before use. In a study by Barenbrock and colleagues, bicarbonate-based versus lactate-based fluid replacement was studied in patients with AKI treated with CVVH.²⁰⁴ They found that serum lactate concentration was significantly higher and bicarbonate lower in patients treated with lactate-based solution. In addition, they showed an increase in cardiovascular events and hypotension in patients treated with lactate solution.

Medication Dosing

During AKI, drugs normally eliminated by the kidney exhibit a markedly decreased clearance. The physiochemical characteristics of drugs affect their removal by dialysis and hemofiltration. The amounts of drug removed during these procedures can be sufficient to require supplemental dosing. For patients on hemodialysis, a supplemental dose of drug is most commonly given at the completion of the dialysis session.²⁰⁵ Drug clearance with CVVH is through convective transport, and it approximates the unbound drug concentration in plasma multiplied by the ultrafiltration rate.²⁰⁶ Drugs with molecular weights of less than 500 D are readily removed by either conventional hemodialysis or CVVH, but those with higher weights of 1000 to 5000 D are eliminated more efficiently by CVVH because of the use of high-flux membranes that allow the passage of larger molecules.

The volume of distribution greatly impacts the clearance of a drug, in that those with large distributions are likely to be more bound in the tissues. Therefore, only a small amount has access to the vasculature at any time. For these drugs, clearance with CVVH is greater than with intermittent therapies because of the continuous nature of the clearance.²⁰⁷ The extent of protein binding of a drug is important because the protein-drug complex is generally greater than 50,000 D. At this size, neither intermittent nor continuous therapies will efficiently remove the drug. However, the extent of protein binding is dependent on pH, uremia, concentration of free fatty acids, heparin therapy, and relative concentrations of drug and protein.²⁰⁸ In critically ill patients, serum albumin is often decreased, thereby making more drug available for clearance during RRT. Because of the potential toxicities, as well as the need to maintain therapeutic levels of multiple medications, it is important to consider and adjust medication dose during AKI and its therapy with RRT. Dosages of medications must be adjusted for the type of RRT, as well as for the specific characteristics of the drug.

Conclusion

Despite extensive clinical experience and improvements in supportive care, the mortality rate of critically ill patients with AKI has not changed over the last 3 decades. However, new information is emerging about the diagnosis and treatment of AKI. Table 114-8 summarizes current recommendations for the care of patients with AKI.¹⁸ These recommendations are based on evidence from clinical trials as well as clinical judgment.

TABLE 114-8 Recommendations for Evaluation and Treatment of Acute Kidney Injury

Evaluate patient for AKI when serum creatinine increases by >0.5 mg/dL.

Exclude prerenal causes (volume depletion, CHF, cirrhosis, NSAIDs, ACE inhibitors).

Exclude postrenal causes with renal ultrasonography and postvoid residual.

Review urine sediment (muddy brown casts, ATN; RBC casts, glomerulonephritis or vasculitis; pyuria, acute interstitial nephritis; bland sediment, prerenal or postrenal azotemia).

Evaluate urine electrolytes in absence of diuretics.

After exclusion of pre- and postrenal azotemia and confirmation of ATN by urine sediment and electrolytes, notify a nephrologist when serum creatinine >2 mg/dL.

Note the projected need for dialysis: oliguric ATN (urine volume <400 mL/24 h), 60%-70% of patients; nonoliguric ATN (urine volume >400 mL/24 h), 30%-40% of patients.

Avoid excessive fluid resuscitation leading to pseudo-ARDS, ventilator support, and multiorgan complications.

Avoid hypotension (generally there is no need to treat hypertension aggressively in the absence of hypertensive crisis).

Maintain fluid balance and treat hyperkalemia; do not use “renal-dose” dopamine.

Review active medications for necessary dose adjustments.

When indicated, use enteral rather than parenteral alimentation.

Discuss timing for initiation and mode of renal replacement with nephrologist (intermittent versus continuous hemodialysis and use of biocompatible membrane).

ACE, angiotensin-converting enzyme; ARDS, acute respiratory distress syndrome; AKI, acute renal failure; ATN, acute tubular necrosis; CHF, congestive heart failure; NSAIDs, nonsteroidal antiinflammatory drugs; RBC, red blood cell.

From Esson ML, Schrier RW. Diagnosis and treatment of acute tubular necrosis. Ann Intern Med 2002;137:744-52.