Chapter 41: Acute Renal Failure

Although acute renal failure (ARF) is conceptually simple to characterize as “an abrupt decrease in renal function,” a commonly accepted definition of ARF emerged only very recently.

An international group of experts convened in 2002 to develop a consensus definition of ARF in order to advance research in the field. They devised a staging classification for ARF—using changes in serum creatinine concentration and hourly urine output—and gave it the acronym RIFLE (Risk, Injury, Failure, Loss, and End-Stage) (Table 41-1).¹ Many investigators have since validated that classification scheme in various clinical settings, and a recent meta-analysis showed a graded impact on mortality by RIFLE stage.²

The RIFLE criteria represented real progress in the field of ARF research. Nonetheless, shortcomings of RIFLE were soon recognized: there was frequent discordance between the stage assignment by the serum creatinine and the urine output criteria.³ In addition, smaller changes in serum creatinine (≥ 0.3 mg/dL)⁴ over a short time (48 hours)⁵ were shown to be associated with important clinical outcomes.

With the goal of refining the ARF definition, a second international group met in 2004. Their consensus produced the Acute Kidney Injury Network (AKIN) criteria.⁶ They recommended that the term acute kidney injury (AKI) replace ARF, to reflect the fact that failure is at the extreme end of a broad spectrum of injury to the kidney.⁶ They defined stages of AKI based on smaller changes in serum creatinine concentration (0.3 mg/dL) over shorter time period (48 hours). Table 41-1 compares the RIFLE and AKIN criteria.

The AKIN and RIFLE criteria agreed substantially in the classification of a large number of critically ill patients, and similarly correlated with mortality.⁷ (Whether mortality is the proper outcome by which to validate a definition of AKI is open to question.⁸)

It is important to recognize that these definitions of ARF or AKI were devised to facilitate research. From a practical standpoint, clinicians will diagnose AKI by a rise in the serum creatinine concentration and BUN over hours to days. Using this practical criterion, or either of the consensus definitions, our patient has AKI.

AKI is common in critically ill patients, and its frequency increases with severity of illness.⁷ Overall, its occurrence varies from 5% in all hospitalized patients,⁹ to 10% to 20% in patients after cardiac surgery,¹⁰ to 20% in patients with sepsis, and up to 50% in patients with positive blood cultures.¹¹ AKI was diagnosed in 23% of patients with acute subarachnoid hemorrhage¹² and in 14% to 27% of patients after acute stroke.^{13, 14}

Risk factors for AKI in patients with strokes include age, underlying CKD, heart disease, stroke type (hemorrhagic more than ischemic), and severity.^{13, 14} In patients with subarachnoid hemorrhage, poor Hunt and Hess score on admission and higher admission serum creatinine concentration were associated with AKI (Table 41-2).¹²

AKI is associated with worse outcomes in critically ill patients as it aggravates morbidity and both short- and long-term mortality. In addition to the direct effects of kidney dysfunction (eg, volume overload, electrolyte and acid-base derangements, uremia), it also affects other organs at a distance both in space and time.

Animal data show that AKI induces or worsens acute lung injury by increasing pulmonary vascular permeability and neutrophil infiltration¹⁵ via cytokine production.^{16, 17, 18} Moreover, it leads to cardiac dysfunction by stimulating inflammation,¹⁹ fibrosis,²⁰ and apoptosis.¹⁵ The neurologic system may be affected by AKI, with reported increases in cerebral vascular permeability, inflammation, and functional impairment.²⁰ In addition, uremic toxins also have deleterious effects on the bone marrow and on liver function.²¹ Thus, it is not surprising that AKI is reported to have an independent association with mortality.^{11, 22, 23, 24}

Until recently, it was generally thought that if patients survived their episode of AKI, their renal function was likely to recover to its baseline. It is now increasingly apparent that AKI predisposes to the development of CKD^{22, 25} and end-stage (dialysis-dependent) renal disease (ESRD),^{25, 26, 27} the latter being more common in patients with underlying CKD and in older patients.

In a prospective study of patients admitted with acute subarachnoid hemorrhage, AKI was significantly associated with increased mortality rate.¹² In patients with acute stroke, AKI worsened both short-term¹⁴ and long-term mortality.¹³

The etiology of AKI can be divided into prerenal, postrenal, and renal causes (Table 41-3).

Prerenal failure is AKI due to altered renal hemodynamics, such that if normal renal perfusion were restored, the AKI would resolve immediately. Prerenal states often are evident by history (eg, history of negative fluid balance or hemorrhage) or physical examination. It is important to recognize, however, that physical assessment of volume status may be quite unreliable,²⁸ and perhaps all the more so in critically ill patients, where the usual indices (vital signs, edema) may be misleading. In oliguric patients (urine output < 0.5 mL/kg per hour), the fractional excretion of sodium (FE_Na) may improve diagnostic accuracy: an FE_Na less than 1% is evidence of avid sodium reabsorption by functional renal tubules and is consistent with a prerenal state.^{29, 30} FE_Na, expressed as a percentage, is calculated from a “spot” urine and simultaneous plasma sample as follows:

where U_Na is the urine sodium concentration, P_Na is the plasma sodium concentration (mEq/L), U_Cr is the urine creatinine concentration, and P_Cr is the plasma or serum creatinine concentration (mg/dL). Note that the FE_Na may be high in a volume-depleted patient who is undergoing diuresis²⁹ and may be low with some causes of intrinsic renal failure, most notably radiocontrast³¹ and pigmenturia.³²

Prerenal failure due to “renal autoregulatory failure” is a common phenomenon. The kidney depends on a variety of humoral mediators of vascular tone—among which are components of the renin-angiotensin system and vasodilatory prostaglandins—to autoregulate its blood flow over a wide range of mean arterial pressure and to autoregulate its glomerular filtration rate (GFR) over a wide range of renal blood flow. This complex autoregulation serves to maintain GFR in states of hemodynamic compromise.³³ Medications that interrupt renal autoregulation (eg, angiotensin-converting enzyme [ACE] inhibitors, angiotensin receptor blockers, and prostaglandin synthesis inhibitors) lead to prerenal AKI in the setting of modest volume depletion or hypotension that would otherwise be well tolerated.³⁴ Our patient was taking both an ACE inhibitor and a cyclooxygenase (COX) inhibitor and so is at risk for prerenal failure.

Postrenal failure is due to obstruction of the urinary tract distal to the renal collecting ducts. Common causes are listed in Table 41-3. In elderly men, the most common cause is bladder outlet obstruction from prostate enlargement. Stroke may amplify the tendency for bladder outlet obstruction. After stroke, bladder dysfunction is common, and bladder-emptying disorders may be more common after ischemic than hemorrhagic stroke.³⁵ Other factors that may predispose to bladder outlet obstruction include underlying autonomic neuropathy (eg, from diabetes mellitus) and medications with anticholinergic activity. Renal ultrasound has over 95% sensitivity for detecting urinary tract obstruction, manifested by dilation of the urinary collecting system (hydronephrosis with or without hydroureter).³⁶

The classification of intrinsic or renal parenchymal causes of AKI follows a histologic nosology (see Table 41-3). Diagnosis is guided by urinalysis with microscopic examination of the sediment (Table 41-4).³⁷

Among the types of intrinsic renal failure, acute tubular necrosis (ATN) is by far the most common in the intensive care unit (ICU).^{38, 39} ATN may be caused by a nephrotoxic or ischemic insult, or a combination of the two.^{39, 40} Common nephrotoxins include radiocontrast dye, aminoglycoside antibiotics, amphotericin, and endogenous pigments such as myoglobin and free hemoglobin. Ischemic insults include shock, sepsis, and major vascular surgery that may compromise renal perfusion. The likelihood of having ATN is proportional to the number of granular casts on urine microscopy; the positive predictive value of finding any casts approaches 100%.³⁷

Of particular relevance to the neurology ICU (but not a factor in the current case) is the role of mannitol in the development of AKI. Mannitol infusion for treatment of increased intracranial pressure (ICP) is recognized to be associated with AKI,^{41, 42} especially with the use of very high doses.^{42, 43} High-dose mannitol seems to have a vasoconstrictive effect and also induces “osmotic nephrosis” in renal proximal tubule cells.⁴⁴ AKI also has been reported to result from hypertonic saline infusion for treatment of increased ICP.⁴⁵

Acute glomerulonephritis (AGN) and acute renal vasculitis are rare causes of AKI in critically ill patients.^{11, 46} AGN or renal vasculitis in a patient with a critical brain disorder should trigger a search for a systemic disease that might explain both phenomena, such as systemic lupus erythematosus and cryoglobulinemia. The suggested serologic evaluation of a patient suspected of having AGN or vasculitis is presented in Table 41-5.

Acute interstitial nephritis (AIN) is most commonly seen as an allergic reaction to a medication. The list of medications associated with AIN is long⁴⁷; frequent culprits include β-lactam and sulfonamide antibiotics, loop-acting and thiazide diuretics, allopurinol, and nonsteroidal anti-inflammatory drugs.^{47, 48} Other causes of AIN include infections of various types and systemic conditions, such as lupus erythematosus, Sjögren syndrome, and sarcoidosis.⁴⁸ Classic associations with AIN include fever, generalized rash, and peripheral eosinophilia. Each of these is present in fewer than 50% of patients with drug-induced AIN, however, and the triad occurs in fewer than 5% of cases.⁴⁷ Thus, AIN should be suspected in any patient with otherwise unexplained AKI and an active urine sediment.

Aminoglycosides such as gentamicin can induce ATN in up to 20% of cases. Among the frequently used agents, amikacin may be less nephrotoxic than gentamicin or tobramycin.

The drugs accumulate in the renal proximal tubule cells and cause cellular necrosis.⁴⁹ Risk factors for nephrotoxicity are advanced age, liver failure, renal failure, peak concentration, and volume depletion.^{49, 50}

With aminoglycoside nephrotoxicity, the serum creatinine concentration starts to rise a week to 10 days after beginning treatment. The AKI often is accompanied by hypokalemia, hypomagnesemia, and a non-anion-gap metabolic acidosis.

Prevention of aminoglycoside-induced AKI is best achieved by minimizing drug exposure, single daily dosing (in patients with normal baseline renal function), frequent pharmacokinetic monitoring, and daily serum creatinine concentration.⁴⁹

The time course of AKI in this case is incompatible with aminoglycoside nephrotoxicity.

Radiocontrast-induced nephropathy (RCIN) is the third most common cause of AKI in hospitalized patients.⁵¹ Its exact incidence is reported to be quite variable, and it depends on the stringency of the definition of AKI and the number and severity of risk factors prevalent in the population under consideration.⁵² Risk factors for RCIN are shown in Table 41-6.^{52, 53, 54, 55, 56, 57} The incidence of RCIN in patients with mild CKD is less than 5%,^{58, 59} and as high as 50% in patients with advanced CKD with diabetes mellitus.⁶⁰ All other factors being equal, it appears to be less common with intravenous than intra-arterial radiocontrast administration.^{61, 62} One retrospective study, using a liberal definition of AKI, found a 3% incidence of radiocontrast-induced injury in unselected patients with acute stroke undergoing CT angiography; none of the patients in that study required renal replacement therapy (RRT).⁶³

After injection of iodinated contrast, the glomerular filtration rate falls immediately, but the consequent accumulation of creatinine in the blood may be slow enough that a change in serum creatinine concentration is not detectable for 24 or 48 hours. The serum creatinine concentration usually reaches a plateau within a few days to a week, followed by resolution in 2 to 4 weeks. About 10% of patients require RRT to manage the consequences of AKI⁶⁴; however, permanent dialysis dependence solely as a consequence of RCIN is rare.

Because RCIN results from a discrete, premeditated event, it has been considered the paradigm for prevention (Table 41-7). Notwithstanding decades of research into its prevention, however, few interventions have shown convincing benefit.^{55, 65} The only surefire way to prevent RCIN is to avoid contrast administration. Short of that, only two interventions have proven unequivocally effective: intravenous fluids and manipulation of the radiocontrast agents.

Intravenous administration of sodium chloride reduces the risk of RCIN. Normal (0.9%) saline seems to be more effective than half-normal (0.45%) saline, when administered at 1 mL/kg/h starting before the contrast injection and continuing for 24 hours.⁶⁶ There has been debate as to whether sodium bicarbonate is more effective than normal saline for the prevention of RCIN. Studies to date have yielded conflicting results,^{67, 68} so there is no basis for recommending sodium bicarbonate. Decompensated congestive heart failure probably is just as great a risk for RCIN as volume depletion. Thus, intravenous saline must be used cautiously in patients with impaired heart or renal function, and should be avoided in patients who are already fluid overloaded. The use of intravenous furosemide before contrast administration has been shown to be deleterious,⁶⁹ and this has led to the recommendation that even oral diuretics be held before contrast procedures.

The risk of RCIN appears to be related to the chemistry of the radiocontrast agent, its route of administration, and its dose. Radiocontrast agents vary according to their osmolality. Conventional high-osmolality agents have an osmolality of about 1200 mOsm/kg, and so-called low-osmolality agents about 600 mOsm/kg. The newest, iso-osmolar agents have an osmolality of about 300 mOsm/kg. There is little doubt that high-osmolality agents are more nephrotoxic than low-osmolality agents.^{70, 71} The additional benefits of iso-osmolar agents in this regard are unclear.^{72, 73} Recent studies have shown that intra-arterial rather than intravenous administration of contrast is associated with a greater risk of ARF.^{61, 62} Finally, the risk of RCIN in high-risk patients seems to be proportional to the volume of contrast administered.⁵³

Other interventions intended to prevent RCIN have failed to show a consistent benefit. N-acetylcysteine (NAC, Mucomyst) showed great promise early on,⁷⁴ but subsequent studies have been equivocal.^{75, 76, 77, 78} Because oral NAC appears to have little if any toxicity, its use in this setting cannot be strongly discouraged. Other interventions shown to be of no benefit for prophylaxis of RCIN include natriuretic peptides,⁷⁹ dopaminergic agents,⁵⁵ and cessation of ACE inhibitors or angiotensin receptor blockers.⁸⁰ A single-center study on the use of prophylactic hemofiltration showed a reduction in 1-year mortality,⁸¹ an outcome of dubious relationship to the intervention, and one that does not justify routine use of this invasive modality. Because COX inhibitors predispose animals to RCIN,⁸² it is prudent to withhold them in patients about to receive a radiocontrast infusion.

In the present case, radiocontrast administration is very likely to have contributed to the development of nephrotoxic ATN. The additional probable cause is sepsis leading to ischemic ATN. This sort of multifactorial AKI is quite typical of critically ill patients.

Table 41-8 shows the general approach to medical management of the patient with AKI.

Hemodynamic optimization will, by definition, reverse prerenal AKI⁸³ and may reduce the likelihood of progression from prerenal failure to ATN.⁸⁴ Regardless of the impact on renal function, however, critically ill patients in shock need adequate intravascular volume resuscitation in order to restore organ perfusion and cardiac output. The amounts and duration of fluid therapy are still controversial: fluid administration is clearly beneficial early in the course of the disease,⁸⁵ but overaggressive volume resuscitation has been associated with worse outcomes.⁸⁶ Furthermore, only half of the patients with circulatory failure respond to fluid,⁸⁷ and the determination of fluid responsiveness can be challenging.⁸⁸

Hence, the need for fluid administration has to be carefully evaluated in order to not place patients at risk of pulmonary edema or abdominal compartment syndrome.⁸⁹ There does not seem to be any superiority of colloid over crystalloid fluid administration⁹⁰; in fact, use of the colloid plasma expander hydroxyethyl starch has been associated with increased risk of AKI.⁹¹

Loop diuretics have been used to augment urine output and potentially treat AKI. Although they can facilitate fluid management, they do not shorten the course of AKI, reduce the need for dialysis, or improve hospital mortality rate.⁹²

Low-dose dopamine has achieved mythic status for its purported effect to selectively improve renal perfusion, and thus to cure renal failure. So sturdy is this myth that low-dose dopamine has even been given the eponym “renal dose dopamine.” A body of literature has accumulated over many years, however, showing conclusively that dopamine is of no benefit whatsoever in AKI, and may be associated with gut necrosis and digital gangrene.^{93, 94} Indeed, low-dose dopamine seems to worsen renal hemodynamics in patients with ARF.⁹⁵ We strongly discourage the use of low-dose dopamine in AKI.

In cases of probable or possible AIN, the culprit medication should be stopped immediately. Glucocorticoid therapy appears to be beneficial in about half of patients if initiated within the first 2 weeks of diagnosis,⁹⁶ and is indicated in cases where the diagnosis is highly likely or confirmed by renal biopsy.

Dosage of all medications should be reviewed for the impact of the decreased renal function on their elimination. In that regard, it is important to recognize that formulas for estimating glomerular filtration rate (eg, the MDRD formula⁹⁷) and creatinine clearance (eg, the Cockroft and Gault formula⁹⁸) cannot be used with a serum creatinine concentration that is rising or falling on a daily basis. These formulas apply only in steady-state conditions; they will overestimate renal function with a rising serum creatinine concentration, resulting in overdosage.

Because of the central role of the kidney in maintaining homeostasis of “the internal environment,” disorders of sodium, potassium, calcium, phosphorus, and acid-base balance are common in patients with AKI. Serum chemistries should be monitored at least daily; more frequent monitoring would depend on the severity of the renal dysfunction and comorbidities.

Adequate nutrition should be ensured, with a daily caloric intake of 25 to 35 kcal/kg, with 0.8 g/kg of protein for patients not yet on RRT.⁹⁹