Prevention of acute renal failure in the

intensive care unit

Kevin W Finkel, Patrick T Murray, and Andrew Shaw

INTRODUCTION initiating potentially beneficial treatment. For

vasoactive agents such as dopamine and atrial
Acute renal failure (ARF) as a result of acute tubu- natriuretic peptide (ANP) to be effective it is possi-
lar necrosis (ATN) may develop in 10% of all ble they need to be given early in ARF, when renal
patients admitted to the intensive care unit (ICU), ischemia is prominent. These agents may provide
and is associated with high morbidity and mortal- no benefit when given at a later stage of ATN. In
ity rates.'-Wespite increased understanding of the the majority of studies on ARF in ICU patients,
pathophysiologic mechanisms operative in ATN, treatment is given late in its c ~ u r s e . ' ~The
, ' ~lack of
and the success of various treatments in reversing efficacy of the various interventions could be the
or ameliorating ATN in animal models, there are result of this delay, either because the mechanism
currently no $oven therapies in human ARF, in of decreased GFR is no longer responsive to ther-
which mortality rates remain in excess of 5 0 Y ~ . ~ apy or simply because too much functional renal
Arguments offered to explain the discrepancy mass has already been lost. Therefore, prevention
between laboratorv and clinical studies include the of ATN remains a cornerstone in the treatment of
lack of an appropriate model of human disease in the critically ill patient.
the ICU, and the untimeliness of intervention in Prevention of ATN in patients in the ICU
clinical trials.5hThe mechanisms of ARF in the ICU requires an understanding of certain key issues:
are complex, involving issues of renal hypoperfu-
1. the pathophysiologic mechanism of a particu-
sion, nephrotoxic insults, and accompanying acute
lar insult
and chronic illness, often in the face of the systemic
2. risk factors for the development of ARF in the
inflammatory response syndrome (SIRS) or sepsis.
ICU
Additionally, the use of positive pressure ventila-
3. the contribution of acute and chronic organ
tion and administration of vasopressor medica-
dysfunction to renal failure
tions can have adverse effects on renal
4. the effect of therapies such as vasopressors
The various animal models used to study ATN
and positive pressure ventilation on renal
lack these complexities. Furthermore, ATN is an
function
evolving process. Early on, renal ischemia and
5. the appropriate hemodynamic profile for a
redistribution of renal blood flow play a prominent
particular patient that optimizes renal blood
role.'"ater, apoptosis and inflammation are key
flow (RBF).
players." Therefore, there is probably a limited
time window for different therapeutic agents to Unfortunately, many of these factors are poorly
provide benefit" (Fig. 5.1). In animal studies, the understood or unknown.
timing of a particular insult is known and therapy
is applied without reliance on first detecting a fall
in glomerular filtration rate (GFR). On the other GENERAL PRINCIPLES
hand, in clinical trials, because of the dependence
011 using changes in serum creatinine levels to In the ICU, patients are exposed to multiple poten-
cletect a fall in GFR, there is an inherent delay in tial renal insults. Frank ischemia or relative renal
72 R E N A L FAILURE IN THE I C U

;1 -
,.zc
--
BBM loss
- exfoliation
/ tubular obstruct~on
cell Injury

--s I ;-.r2v,
,~~~;,~~,-~: .,iJ.,
cc .iI .* :
.2 / .>
. ++**. __--
--A --' -4

5 CMJ Hypoxi= i .>, --,*--

Microvascular injury
obstruction P
.*,, ..- ,....
--. -,'.3d %e9a
- inflammation Maintenance redifferentiation
coagulopathy dedifferentiation
migration
- repolarization
proliferation
0
0 1 2 3 4 5 6 7 8 9 1 0
Days

Figure 5.1 Relationship between the clinical phases and the cellular phases of ischem~cacute renal failure (ARF), and the tern
poral impact on organ function as represented by the glomerular filtration rate IGFR). Prerenal azotemia exists when a reduc
tion in renal blood flow causes a reduction in GFR. A variety of cellular and vascular adaptations maintain renal epithelial cell
Integrity during this phase. The initiation phase occurs when a further reduction in renal blood flow results in cellular injury,
particularly the renal tubular epithelial cells, and a continued decline in GFR. Vascular and inflammatory processes that
contribute to further cell injury and a further decline in GFR usher in the proposed extens~onphase. During the malntenancc
phase, GFR reaches a stable nadir as cellular repair processes are initiated in order to maintain and re-establish organ integrity.
The recovery phase is marked by a return of normal cell and organ function that results in an improvement in GFR. (CMJ
corticomedullary junction, BBM = brush border membrane.) (Reproduced with permission from Sutton et al.12)

hypoperfusion from capillary leak syndrome, aminoglycoside antibiotics is associated with

hypoalbuminemia, and diuretics are frequently lower risk of nephrotoxicity and equivalent
present in ICU patients. Potential exogenous antimicrobial efficacy compared to multiplr,
insults to the kidneys, including myoglobin, radio- dosing strategies.I5 Drug modifications such as
contrast media, aminoglycoside antibiotics, and nonionic radiocontrast and lipid-emulsified
amphotericin B preparatibns, often produce both amphotericin B may also reduce the incidence oi
ischemic and nephrotoxic damage. Distant organ ATN.Ih,l7 Aggressive diuresis must be avoided
dysfunction can contribute to the complexity by when possible, particularly in conjunction with the
adversely affecting renal function, as occurs in use of angiotensin-converting enzyme (ACE)
hepatorenal syndrome and veno-occlusive disease inhibitors or angiotensin receptor blockers, and
of the liver (VOD). Therefore, it is difficult to must be accompanied by careful monitoring of
devise a simple prevention strategy for all patients fluid balance and renal function. Monitoring
in the ICU (Fig. 5.2; Box 5.1). the serum levels of potentially nephrotoxic drugs
Avoidance bf potential nephrotoxins such as such as aminoglycosides, cyclosporin A (CSA),
intravenous radiocontrast, aminoglycoside antibi- tacrolimus, and vancomycin is recommended,
otics, and antifungal agents is prudent, when pos- although studies proving that therapeutic drug
sible. Although nonsteroidal anti-inflammatory monitoring decreases the incidence of ATN are
drugs (NSAIDs), including cyclooxygenase-2 lacking. General measures that may decrease the
(COX-2) inhibitors, generally have a low nephro- incidence of ARF are listed in Box 5.1. The role of
toxic risk, the potential renal vasoconstrictive hemodynamic monitoring and support with fluids
effect of these agents may be significant in selected and vasoactive drugs in the prevention of ARF in
patients, such as those with sepsis, heart failure, the ICU is clearly important. The relationship
cirrhosis, nephrotic syndrome, volume depletion, between renal perfusion and function is discussed
and hypoalbuminemia. Single daily dosing of in depth in Chapter 4, but will be discussed in
 PREl:Er\;TION O F PICUTE REN.4L FAILURE 73

lntravascular volume depletion and hypotension

Gastrointestinal, renal, and dermal losses;
hemorrhage; shock

Large-vessel renal vascular disease

Decreased effective intravascular volume Renal artery thrombosis or embolism,
Congestive heart failure, operative arterial cross-clamping,
cirrhosis, nephrosis, peritonitis
/ Generalized \
or localized reduction in
renal blood flow
Medications
Cyclosporine,tacrolimus, Small-vessel renal vascular disease
angiotensin-converting enzyme inhibitors, Vasculitis, atherombolism, hemolytic-uremic
nonsteroidal anti-inflammatory drugs, syndrome, malignant hypertension, scleroderma,
radiocontrast agents, amphotericin 8 pre-eclampsia, sickle cell anemia, hypercalcemia,
transplant rejection

Hepatorenal syndrome 1{I Sepsis

Ischemic
acute renal failure

Figure 5.2 Condit~onsthat lead to ischemic acute renal failure. A wide spectrum of clinical conditions can result In a gener-
alized or localized reduction in renal blood flow, thus increasing the likelhood of ischemic acute renal failure. The most com-
mon condition leading to ischemic acute renal fa~lureis severe and sustained prerenal azotemia. Kidney ischem~aand acute
renal failure are often the result of a combination of factors. (Reproduced with permission from Thadhani R, Pascual M,
Bonventre JV. Acute renal fa~lure.N Engl J Med 1996;334:1448-60.)

terms of ARF prophylaxis in this chapter, along meas~res.'~:" An analogous situation is found in
with other approaches to avoid or ameliorate ARF attempting to predict mortality rates in ICU
in critically ill patients. patients with ARF. Although scoring methods such
as APACHE I1 and APACHE I11 are reasonably
good at predicting overall mortality in ICU
PREDICTION OF ACUTE RENAL FAILURE patients, they routinely underestimate the rate in
the presence of ARF."," Predictive scores designed
For prevention of ARF to be an achievable goal, it specifically for ICU patients with renal failure,
is imperative that there is a means of accurately such as the Cleveland Clinic Score and the Liano
predicting the development of ARF. Although Score, have better accuracy, but may not be
much research continues to center on predicting applicable to other medical
outcome in patients who experience ARF in the
ICU, surprisingly little work has been done in the
area of predicting who will develop ARF. To date, RENAL BLOOD FLOW
a reliable prediction model is not available. Studies
using multiple linear or logistic regression analysis Providing adequate renal perfusion in the face of
suggest that a combination of factors such as age, critical illness appears to be an appropriate goal.
hypotension, hypoxia, the presence of two of four Often, patients in the ICU already have some com-
markers of SIRS, use of vasopressors, positive ponent of prerenal failure that can progress to
pressure ventilation, chronic renal failure, and sep- frank ischemic ATN if not corrected. However,
sis have some prediction value, but none have the what amount of perfusion is needed, where it
~'rt~cisionto allow application of true preventive should be distributed, and how to measure
74 R E N A L FAILURE I N THE ICU

sodium retention in congestive heart failure (CHI:!

Box 5.1 Strategies to decrease ARF in the ICU In CHF, although retention of sodium to impro\ t ,
cardiac preIoad is an adaptive response to prr
Proven ceived poor perfusion, it is detrimental to t h l
Avoidance of nephrotoxins patient and must be treated with diuretics. 111
S~ngledaily dosing of aminoglycosides contrast, increasing RBF with vasodilators in tin,
Drug modifications:
presence of ATN may reverse a protective hemo
dynamic response of injured kidneys; this may b t ~
Liposomal amphotericin B
an inappropriate therapy for what has bee11
Nonionic radiocontrast
termed 'acute renal success'.'"
Hydration: There are no reliable indicators of adequate RBI:,
Radiocontrast except (presumably) normal renal function. Urinc-
C~splatln output correlates best with degree of renal injur\'
Rhabdomyolysis (i.e. oliguric ATN has a worse prognosis, thouglit
Tumor lysis syndrome to reflect more severe renal injury). Except in thts
Tight glycemic control in the critically ill case of prerenal failure, where a volume infusion
reverses the low flow state, urine output is a n
Possibly effective unreliable marker of renal perfusion. Most often,
Mon~toringdrug levels
predetermined mean arterial pressure (MAP) read-
ing is considered proof of adequate renal perfu-
Adequate renal perfusion
sion. Vasopressor algorithms used in the ICL
Low-dose fenoldopam
usually specify an arbitrary MAP goal for titration
Nacetylcysteine (radiocontrast) of the medications. However, the appropriate
Alkalinizatlon of the urine: MAP for a critically ill patient with ARF is not
Rhabdomyolysis defined. The three major determinants of renal
Methotrexate blood flow are cardiac output, intravascular
Tumor lysis syndrome volume, and renal perfusion pressure.

Ineffective
Low-dose dopamine Cardiac output
Mannitol
Increasing cardiac output by administration ot
Diuretics
intravenous fluid or cardiac inotropes should
Aminophylline
increase renal blood flow in low flow states such as
Atrial natriuretic peptide cardiogenic shock and prerenal azotemia. In
patients with ARF, decreased renal perfusion is
the result of an imbalance between local renal
vasoconstrictor and vasodilator influences. This
adequacy are all unknown factors. A fall in renal imbalance causes both global renal hypoperfusion
perfusion is both a result and a response to injury. and shunting of blood flow away from the renal
A decrease or a redistribution of blood flow away medulla. There is no evidence that increasing car-
from a damaged tubule to prevent solute loss is an diac output to 'supranormal' levels counteracts
adaptive response to guard against hypovolemic this imbalance. Most studies in critically ill
shock.'"hen enough nephron loss occurs, patients comparing normal to enhanced cardiac
nitrogenous waste accumulates and ARF ensues. output measured by oxygen delivery or cardiac
As discussed in Chapter 4, indiscriminate index have failed to show benefit on survival
increases in blood flow could have deleterious rates." One recent study of aggressive hemo-
effects on the course of ATN by increasing solute dynamic management of 263 septic patients on
loss, maintaining toxin exposure to injured arrival to the emergency center did demonstrate a
tubules, or promoting reperfusion oxidant injury. survival benefit, but effects on renal function were
In fact, although renal vasoconstriction in ATN not described, and its relevance to preventing ARF
could be viewed as a maladaptive response per- in established ICU patients is ~ n k n o w n . ' ~
petuating ARF, this phenomenon should not be Currently, there are no randomized controlled
considered a purely inappropriate response like trials comparing different cardiac outputs and the
 PREVENTION O F ACUTE RENAL FAILURE 75

subsequent development of ARF in high-risk mum fluid resuscitation regimen remains a dis-
individuals. puted topic. For the time being it appears that
treatment of the underlying diagnosis, usually
sepsis, and general supportive efforts are the
lntravascular fluid expansion mainstays of therapy.

Both animal and human studies suggest that

intravascular fluid expansion decreases the risk of Renal perfusion pressure
ARF from radiocontrast agents and various
nephrotoxic i n ~ u l t s . ' ~ ~ ~ ~ o w ewhat v e r , role it In the critically ill patient attention is usually
plays in preventing ATN in patients in the ICU focused on improving cardiac output and intra-
with multiple risk factors is purely speculative. vascular volume to maintain adequate perfusion to
Because ICU patients usually have capillary leak the heart and brain, the 'vital organs.' What is
syndrome and impaired pulmonary function, often unappreciated is that in the normal kidney
indiscriminate fluid administration can have dele- loss of autoregulation of renal blood flow occurs
terious effects. Therefore, measuring volume status at a MAP of 75-80 mmHg, and that the loss of
in ICU patients often requires invasive monitoring autoregulation of GFR occurs at approximately
of right atrial pressure or pulmonary artery occlu- 80-85 mmHg.'7," In the face of long-standing
sion pressure. At this time, no data are available to hypertension present in many ICU patients,
indicate that a certain degree of intravascular fill- autoregulation fails at even higher MAP levels.
ing is more protective of renal function than Therefore, 1CU protocols that titrate vasopressors
another. to an MAP of 65 or 70 mmHg can result in per-
The putative superiority of colloid over crystal- sistent renal ischemia (Fig. 5.3). There are no ran-
loid fluids in the resuscitation of the critically ill domized controlled trials comparing MAP values
patient has been a source of considerable contro- and the development of ARF in ICU patients.
versy. Aggressive hydration with crystalloid solu- However, the current treatment strategies may be
tions such as 0.9% sodium chloride can worsen inappropriate in some individuals.
interstitial edema and pulmonary function. Col-
loidal solutions such as various starches and
human albumin might appear to be attractive PHARMACOLOGIC AGENTS
alternatives, but there is little solid evidence of
their superiority in clinical Systematic N-acetylcysteine (NAC, Mucomyst)
reviews of randomized controlled trials comparing
crystalloids with colloids have yielded conflicting NAC is an antioxidant and causes renal vasodila-
results. Some trials have found an increased mor- tion by generating increased levels of nitric oxide
tality rate associated with the administration of (NO). Based on these effects, NAC was used in
human albumin and hydroxyethylstarch, whereas several small human trials for the prevention of
others have n ~ t . ~ Most' , ~ ' recently, a large random- ARF from radiocontrast agents in high-risk indi-
ized, controlled prospective trial of albumin vs viduals. The trials have shown that its administra-
saline in almost 7000 critically ill patients found no tion results in a significantly smaller change from
benefit of one over the other." Specifically, there baseline of serum creatinine values compared to
was no demonstrable effect on mortality, renal changes in the placebo 4 @ Arecent meta-
function, or the frequency of renal replacement analysis of 8 randomized controlled trials sup-
therapy. Of note, patients with cirrhosis were ported the use of NAC to prevent radiocontrast
excluded from this trial, and limited data suggest nephropathy (RCN).41However, whether or not it
that albumin is useful to prevent ARF in cirrhotic prevents severe ARF and the need for dialysis has
patients with spontaneous bacterial peritonitis (see not been determined. Furthermore, emerging liter-
b e l o u ~ ) , or
? ~ undergoing large volume paracente- ature suggests that NAC causes a GFR-independent
sis. As discussed below, for prevention of radio- decrease in serum creatinine,?' perhaps by inhibit-
contrast nephropathy normal saline is superior to ing creatine phosphokinase-mediated generation
half-normal saline," and an equimolar solution of of ~ r e a t i n i n e This
. ~ ~ may explain the puzzling but
sodium bicarbonate is superior to normal saline.36 consistent decrease in serum creatinine below
Otherwise, despite the importance of fluid therapy baseline levels observed in NAC-treated patients
in the prevention of ARF, the nature of the opti- wit11 chronic kidney disease undergoing RCh:
76 RENAL FAILURE IN THE ICU
i
: ;
1.5 -
Autoregulatory range Figure 5.3 Autoregulation of renal blooi!
flow (RBF). Autoregulation maintains RBF I
and glomerular filtration rate (GFR) over <I t
wide range of mean arterial pressure (MAP:
levels by modulation of afferent arteriolal
resistance. RBF increases by only 10% w h e ~ '
MAP increases by 50% from 1 0 0 to 150
mmHg, but decreases precip~touslywhev
MAP falls below 70-80 mmHg. (Reproducetl
with permission from Vander AJ: Renal
physiology, 5th edn. McGraw-Hill, New York.
1995. In: Chapter 2: Renal blood flow ant1
glomerular filtration (Figure 2-4, page 34).)

0 40 80 120 160 200 240

BP (mmHg)

prophylaxis, which was previously assumed to be associated with radiocontrast administration,

an effect of volume expansion. It is thus conceiv- repair of aortic aneurysms, orthotopic liver trans-
able that the putative renoprotective effect of NAC plantation, unilateral nephrectomy, renal trans-
in contrast nephropathy is an artifact. NAC has plantation, and chemotherapy with i n t e r f e r ~ n . ~ ~ , ~ "
also been administered intravenously to 100 criti- Yet despite more than 20 years of clinical experi-
cally ill patients in a randomized placebo- ence, prevention trials with low-dose dopamint,
controlled trial to prevent progression of multiple have all been small, inadequately randomized, ot
organ dysfunction syndrome (MODS).@There was limited statistical power, and with endpoints ol
no significant difference between the groups in questionable clinical significance. Furthermore,
mortality rate, days of inotropic support, mechani- there is concern for the potential harmful effects ot
cal ventilation, ICU length of stay, or development dopamine, even at low doses. It can trigger tachy-
of ARE arrhythmias and myocardial ischemia, decreasc
At this point, it is probably harmless to provide intestinal blood flow, cause hypothyroidism, and
NAC prior to radiocontrast administration in suppress T-cell function.4547 It has also been
patients at risk for the development of ARF pend- shown to increase the risk of RCN when given
ing further research, provided that it is combined prophylactically to patients with diabetic
with appropriate volume expansion, and that the nephropathy."
effect of NAC on serum creatinine is remembered Numerous trials using low-dose dopamine to
while assessing post-dye renal function. However, treat established ATN have also been reported in
there are no convincing data to support the routine the last several years and suggest its use is benefi-
administration of NAC in RCN or other clinical ~ial.~j,.'~,."
However, most studies were either
circumstances to prevent the development of ATN, uncontrolled case series or small randomized trials
and it may soon be proven ineffective for such with Iimited statistical power. Kellum and Decker
purposes. found no benefit of dopamine for prevention or
therapy of ARF in an adequately powered meta-
ana1ysis.j" More recently, a large randomized
placebo-controlled trial in 328 critically ill patients,
with early ARF sufficiently powered to detect o
Low-dose dopamine administration (1-3 pg/kg/ small benefit, reported no effect of low-dose
min) to normal individuals causes renal vasodila- dopamine on renal function, need for dialysis, ICU
tion and increased GFR, and acts as a proximal or hospital length of stay, or m ~ r t a l i t y . ~These
'
tubular diuretic. Due to these effects, numerous findings combined with the aforementioned
studies have used low-dose dopamine to either potential deleterious effects of low-dose dopaminr
prevent or treat ATN in a variety of clinical set- are strong arguments for abandoning its use
tings. It has been given as prophylaxis for ARF entirely for the prevention and therapy of ARF.
 P R E V E N T I O N O F A C U T E R E N A L FAILUlZE 77

Low-dose fenoldopam mesylate Diuretics

Fenoldopam mesylate is a pure dopamine type-1 Furosemide is a loop diuretic and vasodilator that
receptor agonist that has similar hernodynamic may decrease oxygen consumption in the loop of
effects to dopamine in the kidney without a- and P- Henle by inhibiting sodium transport, thus poten-
adrenergic stimulation. Limited trials suggested tially lessening ischemic injury. By increasing
that administration of fenoldopam mesylate urinary flow, it may also reduce intratubular
reduced the occurrence of ARF from radiocontrast obstruction and backleak of filtrate. Based on these
agents and following aortic aneurysm repair.",53 properties, furosemide might be expected to pre-
However, a recently reported large randomized vent ARF (Fig. 5.4). However, there are little data to
controlled trial of fenoldopam mesylate to prevent support its use. Furosemide was found to be
contrast nephropathy in 315 patients demonstrated ineffective or harmful when used to prevent
that its administration had no beneficial effect on ARF after cardiac surgery, and to increase the
urine output, change in serum creatinine levels, risk of ARF when given to prevent contrast
incidence of ARF, or need for dialysis.*' Promising nephr~pathy.'~,'~
data from a recent randomized, placebo-controlled Similarly, there is little evidence of benefit from
pilot trial of low-dose fenoldopam mesylate in 155 diuretic therapy in established ARF. A single recent
ICU patients with early ATN showed that study in 100 patients with oliguric ARF after car-
fenoldopam patients tended to have lower 21-day diac surgery suggested that a cocktail infusion of
mortality rates and decreased need for dialysis, but furosemide, mannitol, and dopamine improved
the study was ~ n d e r p o w e r e d A
. ~ larger
~ study is renal function postcardiac surgery compared to
required to determine if fenoldopam ameliorates intermittent loop diuretics alone.57In patients with
the course of ATN. established ATN, several studies have found no

Tubular damage
(proximal tubules and
ascending thick limb) ,

renin-angiotensin
iq by casts backleak

I endothelin
t wll,

+
4 lntratubular
fluid flow

? Direct glomerular
I
Oliguria
effect

Figure 5.4 Pathophysiologic mechanisms of acute renal failure (ARF). Tubular damage by ischemia, nephrotoxins, or both,
leads to decreased glomerular filtration rate (GFR) by a combination of mechanisms. (1)Renal vasoconstriction via activation
of tubuloglomerular feedback, and decreased vasodilator substances (PGI,, prostacyclin; NO, nitric oxide), is a prominent func-
tional mechanism of decreased GFR in acute tubular necrosis (ATN). (2) Backpressure from tubular obstruction by casts directly
decreases GFR. (3) Backleak of glomerular filtrate into peritubular capillaries decreases the efficiency of glomerular filtration,
effectively decreasing GFR. (4) There is increasing evidence for a role of interstitial inflammation in the extension phase of ATN.
(51 Direct glomerular effects (mesangial contraction, decreased filtration surface area) may also play a role in decreasing GFR
in the presence of ATN.
78 R E K A L F.4ILLRE IN T H E ICU

benefit of loop diuretics:~">l~eir use did not accel- and may even cause acute renal failure ('osmotic
erate renal recovery, decrease the need for dialysis, nephrosis').""
or reduce mortalitv. It was shown that the mortal-
ity rate of oliguric patients who responded to
furosemide with a diuresis was lower than those Atrial natriuretic peptide
who did However, the clinical characteris-
tics, severity of renal failure, and mortality rates Atrial natriuretic peptide causes vasodilation of
were similar in patients with either spontaneous the afferent arteriole and constriction of the effer-
nonoliguric ARF or patients who became nonolig- ent arteriole, resulting in an increased GFR. It also
uric after furosemide. This implies that those inhibits renal tubular sodium reabsorption. Most
patients able to respond to furosemide have less studies with ANP involved the treatment of
severe renal damage than nonresponders, rather established ATN. However, in two studies that
than deriving any true therapeutic benefit from administered ANP in renal transplant recipients
furosemide administration. Although administra- to prevent primary renal dysfunction, no benefit
tion of furosemide might facilitate improved fluid was f o ~ n d . ~As " , ~with
~ mannitol and low-dose
management if it induces a diuresis, a retrospec- d ~ p a r n i n e , ' "one
~ ~ study suggested that ANP pro-
tive review of a recent trial in critically ill patients phylaxis might worsen renal function in diabetic
with ATN raised concerns of possible harm from patients receiving radiocontrast agents."
loop diuretics in ARF. The authors found that Based on the positive results of small clinical
diuretic use was associated with an increased risk studies using ANP to treat ATN, a randomized
of death and nonrecovery of renal f ~ n c t i o n . ~ W o s t placebo-controlled trial of 504 critically ill patients
of the increased risk, however, was seen in those with ARF was conducted.'"espite the large size
patients unresponsive to high doses of diuretics, of the trial, ANP administration had no effect on
implying they had more severe disease. Therefore, 21-day dialysis-free survival, mortality, or change
diuretics should be used with caution in critically in plasma creatinine concentration. Of note, the
ill patients, and iatrogenic hypovolemia and super- mean serum creatinine values at enrollment (about
imposed prerenal azotemia must be avoided. 4.4-5 mg/dl) in this study confirm that interven-
Diuretics should be withdrawn if there is no tion in this trial was extremely late in the course of
response, to avoid ototoxicity. In patients who ATN. Although a subgroup analysis of the study
experience an increase in urine output, hypo- suggested that ANP might be beneficial in those
tension must be avoided, since kidneys with ATN patients with oliguric renal failure, a subsequent
are susceptible to further damage from decreases trial in patients with oliguric renal failure failed to
in perfusion pressure. To maintain the diuresis, a demonstrate any benefit of ANP.7%ypotension
continuous infusion of drug is probably preferable was significantly more common in ANP-treated
to intermittent bolus a d r n i n i s t r a t i ~ n . ~ ~ l t h o u g hpatients in this study, and may have negated any
there are no large randomized controlled trials, the potential benefit of renal vasodilation in these
overall evidence suggests that continuous infusion patients. Hence, there is no convincing evidence to
of diuretics as opposed to bolus administration is support the use of ANP in the prevention or treat-
more effective and associated with less toxicity and ment of ARF. A new, promising, but underpowered
delayed development of diuretic resistance. (61 patients) positive study of ANP to treat ARF
Mannitol is an osmotic diuretic that can immediately following cardiac surgery showed a
decrease cell swelling, scavenge free radicals, and decreased rate of postoperative renal replacement
cause renal vasodilatation by inducing intrarenal therapy compared to placebo-treated patient^;^? a
prostaglandin production.6h-Itmay be beneficial larger prospective trial in this setting appears
when added to organ preservation solutions warranted.
during renal transplantation and may protect
against ARF caused by rhabdomyolysis if given
extremely early.66-68Otherwise, mannitol has not Insulin-like growth factor-1
been shown to be useful in the prevention of ARF.
In fact, mannitol may aggravate ARF from radio- Insulin-like growth factor-1 (IGF-I) increases renal
contrast agentszR Furthermore, mannitol may blood flow and induces cell proliferation and dif-
precipitate pulmonary edema if given to volume- ferentiation.Inaddition,it reversesapoptosis. In ani-
overloaded patients who remain oliguric, can mal models, it ameliorates renal injury associated
exacerbate the hyperosmolar state of azotemia, with ischemia and may prevent injury following
 P R E V E N T I O N O F A C U T E R E N A L FAILURE 79

renal t r a n s p l a n t a t i o ~ ~However,
. ~ ~ , ~ ~ a recent small glucose level on admission is an independent pre-
clinical trial found no benefit of 1GF-I therapy for dictor of prognosis after myocardial infarction or
delayed graft function in postcadaveric renal of the need for coronary artery bypass grafting8?
transplant in humans.77 Furthermore, the in-vitro responsiveness of leuko-
1GF-I has been given to a small group of patients cytes stimulated by inflammatory mediators is
in a single trial for prophylaxis of ARF following inversely correlated with glycemic control.8hBased
aortic aneurysm repair.7R1GF-I was started post- on these findings, a randomized controlled trial
operatively in a randomized placebo-controlled was conducted involving 1500 ICU patients who
fashion; it was well tolerated, and produced a received either intensive or conventional glycemic
modest increase in the creatinine clearance in the control." All patients were receiving mechanical
treated group compared to the placebo group, pos- ventilation, the majority postoperatively. The
sibly by vasodilation rather than a 'trophic' effect. study was terminated early because the mortality
However, no patients developed ARF that necessi- rate in the intensive treatment group was signifi-
tated dialysis. Hence, the role, if any, for IGF-1 in cantly lower than in the conventional treatment
the prevention of ARF remains unknown. arm. Moreover, the incidence of severe renal insuf-
Based on the positive effects of 1GF-I in animal ficiency (peak serum creatinine >2.5 mg/dl; 11.2%
models of ATN, a randomized placebo-controlled conventional, 7.7% intensive, p = 0.04) and need
trial was conducted in 72 critically ill patients with for renal replacement therapy (8.25%conventional,
established ARF.'" The results showed there was 4.8% intensive, p = 0.007) were significantly lower
no difference in the two groups in post-treatment in the intensive glycemic control group. Whether
GFR, need for dialysis, or mortality, although it these results can be readily extrapolated to patients
should be noted that GFR at randomization was in a nonsurgical ICU or to those with other types of
only 6.4-8.7 ml/min, and ATN was possibly too critical illness is currently unknown. Similar trials
established for a successful intervention in this in other groups of patients will be necessary to
study population. In anuric patients, IGF-I admin- confirm the observed benefits.
istration was associated with a slower rate of
improvement in urine output and GFR. So, despite
the ample evidence that IGF-I accelerates renal Anti-TNF-a therapy
recovery in animal models of ARF, there is no
support for its use in humans. Tumor necrosis factor-a (TNF-a) is an inflamma-
tory cytokine that plays a pivotal role in the host
response to infection. In addition to systemic
Thyroxine effects, TNF-a may have specific renal effects. In-
vivo TNF-a infusion in animals or perfusion of the
The administration of thyroid hormone following isolated rat kidney with TNF-a decreased GFR.88
the initiation of ATN in a variety of ischemic and TNF-a caused leukocyte and fibrin accumulation
nephrotoxic animal models was found to be effec- in glomerular capillary lumens and induced apop-
tive in promoting recovery of renal f u i ~ c t i o n . ~ ~ tosis
- ~ ~ in glomenilar endothelial cells.R5A large num-
Based on these results, thyroxine was administered ber of studies in diverse animal models have
to 59 patients with ARF in a randomized placebo- shown that anti-TNF antibodies confer protection
controlled trial.s3 Patients were well matched in against the morbidity and mortality from both
baseline characteristics. Administration of thyrox- Gram-positive and Gram-negative sepsis, includ-
ine had no effect on any renal parameter. However, ing the development of ARF.9b5'Emerging animal
the trial was terminated early because of a signifi- data suggest that anti-TNF therapies may have the
cantly higher mortality rate in the patients who potential to prevent septic ARF, despite the failure
received thyroxine. of such therapies to improve mortality in sepsis
syndrome. For example, Cunningham and col-
leagues found that renal susceptibility to renal
Intensive insulin therapy injury in endotoxemic mice with or without TNF
receptor knockouts was associated with TNF
Hyperglycemia associated with insulin resistance receptor expression in the kidneys rather than the
is common in critically ill patients, independent of hosts, as determined by performing renal trans-
a history of diabetes mellitus.8~tudiesinvolving plants between knockouts and wild-type mice."
nondiabetic patients have found that the plasma On the other hand, they recently found that
80 R E N A L FAILURE I N THE ICU

.s~/sft.micexpression of Toll-like receptor4 is also ity reduction in the recombinant TFPI-treatvci

r&uired for development of endotoiemic ARE9' group.""' However, a recently completed pivot.11
Limited clinical data similarly suggest that TNF phase I11 trial of TFPI in 1700 patients with sepsiz
receptor expression may play a rolein the suscep- or severe sepsis failed to show a survival benefit;
tibility to ARF in patients with sepsis."l ARF incidence in the groups was not reported.""
Several large trials with neutralizing monoclonal Antithrombin 111 blocks several proteast.5
anti-TNF-a antibodies or soluble TNF recentor involved in coagulation and plasma levels arcb
fusion proteins have failed to consistentlv show usually markedly reduced in patients with sepsis.
significant survival benefits or a reduced incidence In a double-blind, placebo-controlled multicenter
of ARF in patients with sepsis."-"8 This apparent trial of 2300 patients with severe sepsis and
lack of efficacy may be related to the heterogeneity septic shock, high-dose antithrombin 111 had no
of the patient population. Age, immune status, and effect on 28-day mortality and was associated
genetic predisposition may all alter the int-lam- with an increased risk of hemorrhage when
matory reaction in critical illness and lead to co-administered with heparin.Io'
different responses to anticytokine therapies. A complex interaction between the pro-
Therefore, despite the apparent success of anti- inflammatory, coagulation, and fibrinolytic net-
TNF therapies in animal models in decreasing both works plays a pivotal role in organ damage ill
mortality and renal failure, the beneficial effects of sepsis. Although several strategies to inhibit
these strategies in humans are marginal at best. coagulation have been evaluated in sepsis, only the
administration of activated protein C has proveci
successful, and its efficacy may depend on its
Inhibitors of coagulation combined effects on coagulation, fibrinolysis, and
inflammation. Activated protein C is currently
Disseminated intravascular coagulation is com- approved for the treatment of severe sepsis,
mon in ICU patients and is asiociated with an although current trials are assessing its effective-
adverse prognosis. It is characterized by a general- ness in less-ill patients. Specific renoprotective
ized activation of the coagulation cascade, result- effects have not been demonstrated in humans,
ing in the intravascular formation of fibrin clots but emerging data suggest that activated
and endothelial damage. Impaired tissue blood protein C may be broadly protective against
supply contributes to o&an dysfunction, including ischemia-reperfusion injury in a variety o!
ARE Several agents that block coagulation at dif- setting^.'^"
ferent levels have been evaluated as adjunctive
therapy in sepsis.
Protein C is activated bv the thrombin- Nitric oxide synthase inhibition
thrombomodulin complex on endothelial cells and
inhibits thrombin generation. Besides its effects on Nitric oxide (NO) is the metabolic product of L-
coagulation, activated protein C has direct anti- arginine and is produced by three major NO
inflammatory properties, including impairment of synthase (NOS) isoforms: endothelial NOS
leukocyte adhesion to the endothelium and inhibi- (eNOS), neuronal NOS (nNOS), and inducible
tion of the production of inflammatory cytokines. NOS (iNOS). Within the kidney, eNOS is constitu-
In a randomized, multicenter trial conducted in tively expressed in endothelial cells and produces
2690 patients with severe sepsis, recombinant vascular relaxation and inhibition of leukocyte
human activated protein C significantly reduced adhesion and platelet aggregation. Excessive gen-
mortality.0gIt was particularly effective in the most eration of NO by iNOS has been implicated as an
seriously ill patients, as assessed by the APACHE I1 important mediator of the hemodynamic alter-
score, the number of failing organs, and the pres- ations in sepsis, particularly the vascular hypo-
ence of shock; effects on the incidence or course of responsiveness and v a ~ o d i l a t i o n . ' ~ ~
ARF were not described in study publications. Animal studies using nonspecific inhibitors of
Tissue factor forms a complex with factor VIIa NOS have yielded conflicting results, although use
and initiates thrombin generation. Tissue factor is of specific iNOS inhibitors has been more encour-
inhibited by a natural anticoagulant, tissue factor aging. The ubiquitous nature and the pleiotropic
pathway inhibitor (TFPI). A phase I1 trial compar- effects of the NO system, as well as its complex
ing placebo and recombinant TFPI in 210 patients alterations in sepsis and ARF, probably explain
with severe sepsis showed a trend toward mortal- why NOS inhibition fails to show reproducible
 PREVENTION OF ACL TE R E h A L F.4ILLRE 81

beneficial effects. NO release by the endothelial constrictors. Endotoxin and various inflammatory
cells of the renal microcirculation is essential to cytokines stimulate the synthesis of thromboxane
counterbalance the vasoconstrictor influences and A, and leukotrienes in the kidney and in
maintain RBF, to inhibit infiltration of leukocytes, inflammatory cells.'lu
and to prevent thrombosis. Several animal studies In animal models of sepsis, cyclooxygenase inhi-
have sho\v17 that whereas nonselective NOS inhibi- bition with indomethacin, selective thromboxane
tion worsens septic ARF while raising blood pres- inhibition, and leukotriene antagonism all had
sure, iNOS-selective agents improve both systemic beneficial effects on renal function.""-1" However,
hemodynamics and renal function.lu5 in 455 patients with sepsis, cyclooxygenase inhibi-
The results of a phase I1 and a pivotal phase 111 tion with intravenous ibuprofen reduced the syn-
trial of the nonspecific NO inhibitor L-NMMA thesis of thromboxane and prostacyclin, but it had
were recently reported.loh,'"In the phase I1 trial of no effect on the development of shock or renal fail-
315 patients with severe sepsis, compared to ure and did not improve survival."? In the absence
placebo, NO inhibition led to increased reversibil- of clinical studies with selective thromboxane or
ity of shock.'" However, in the phase 111 trial leukotriene inhibitors, no meaningful conclusions
involving over 800 patients, the trial was stopped on their potential benefit can be drawn.
prematurely when interim analysis demonstrated
increased mortality in the L-NMMA-treated
group."" This discordant result may be due to dif- Inhibition of leukocyte adhesion
ferences in patient enrollment and hemodynamic
management protocols, and the nonspecific nature The recruitment of circulating leukocytes into a tis-
of NOS inhibition. Future strategies that inhibit sue is directed by specific adhesive interactions
iNOS but amplify eNOS may prove beneficial. between the leukocyte and the vascular endothe-
lium. Selectins mediate the initial contact between
the leukocyte and the endothelium, and adherence
Endothelin antagonism and migration are mediated by interactions
between integrins on the leukocyte and surface
Endothelin-1 (ET-I) is a peptide with potent vaso- receptors on the endothelium such as intercellular
constrictor effects on the renal microcirculation, adhesion molecule-1 (ICAM-I ). Studies suggest
thereby reducing RBF and GFR. Experimental that during sepsis and ischemia leukocytes
studies with ET-receptor antagonists in animal infiltrate the kidneys, resulting in renal dysfunc-
models of ARF demonstrated improved renal tion, and provide a rationale for the inhibition of
function.10sHowever, no studies with ET receptor Ieukocyte recruitment in these settings.'14,"'
antagonists have been performed in patients with Treatment of experimental animals with anti-
sepsis. A nonseIecti1.e ET antagonist increased the ICAM-1 antibodies or with antisense oligonu-
risk of contrast nephropathy in patients with cleotides for ICAM-I prevents ischemic ATN and
chronic renal failure undergoing coronary angiog- ameliorates the functional and histologic injury
raphy,'" perhaps because ETB receptors cause associated with experimental ischemic or septic
vasodiIation and ETAreceptors cause vasoconstric- ARF.''6,"7Several mechanisms may be operative in
tion. This paradoxical finding may make it difficult leukocyte-mediated renal injury. Leukocytes
to perform studies of ET antagonism in patients release reactive oxygen species and enzymes that
with ARF and sepsis, although results may be may directly injure cells. The production of
better with an E T , receptor-selective agent. cytokines attracts additional inflammatory cells
and up-regulates adhesion molecules, creating a
cycle of injury. Release of vasoconstrictor arachi-
Inhibitors of arachidonic acid metabolism donic acid metabolites, as well as physical conges-
tion of medullary capillaries, contributes to
Metabolism of arachidonic acid by cyclooxpgenase persistent hypoxia. However, no results from
results in the generation of prostaglandins and human trials with antibodies to leukocyte adhe-
thromboxanes, whereas lipoxygenase yields sion molecules are available. Inhibition of leuko-
leukotrienes. Both prostaglandin E, and prosta- cyte recruitment is a potential promising approach
cyclin cause renal vasodilatation and natriuresis, in the treatment of septic ARF, but data in humans
whereas thromboxane A,, leukotrienes, and pro- are required before relevant conclusions can be
staglandins F, and H, are potent renal vaso- drawn.
82 R E N A L FAILURE I N T H E ICU

SPECIFIC CLINICAL SCENARIOS

Box 5.2 Risk factors for the development of

Radiocontrast agents radiocontrast nephropathy

An acute rise in serum creatinine levels following Chronic renal insufficiency (particularly in diabetics)
administration of radiocontrast material is defined Volume depletion
as contrast nephropathy. Clinical trials assessing High contrast dose
the effectiveness of various prevention strategies Myeloma kidney
have utilized different absolute changes in creati- Nephrotic syndrome
nine to define ARF, making comparisons between
Older age
studies difficult and obscuring the actual benefit
Congestive heart failure
achieved. Preventing increases in serum creatinine
values as small as 25% have been used to define Nephrotic syndrome
successful intervention instead of relying on end-
points such as hospital length of stay, need for
dialysis, or mortality rates. However, a retrospec-
tive review of 16 000 patients who received intra- high-risk patients undergoing coronary arteriogra-
venous radiocontrast showed that patients with a phy, although it has never been studied in a
50% or more increase in serum creatinine had a 6- placebo-controlled trial. In a study of 78 patients
fold higher mortality rate compared to those who with underlying renal dysfunction (mean baseline
did not.'ln Hence, using modest changes in serum creatinine of 2.1 mg/dl), randomization to 0.45%
creatinine as a surrogate for more serious compli- saline was superior to 0.45% saline plus either
cations may be a reasonable approach. As shown mannitol or furosemide in preventing a 0.5 mg/dl
in Box 5.2, numerous risk factors are known for the increase in serum creatinine levels.28A variety of
development of RCN, including chronic kidney renal vasodilators (dopamine, fenoldopam, ANP)
disease (especially diabetic nephropathy), volume have proven ineffective in addition to saline. A
depletion, uncompensated CHF, and high contrast more recent trial of comparable patients demon-
volume. strated that 0.9% saline was more beneficial than
The pathogenesis of ARF from radiocontrast is 0.45% saline in preventing contrast n e ~ h r o p a t h y . ~ ~
complex and incompletely ~ n d e r s t o o d . ~After
~ ~ - ~ ~ 'Most recently, administration of an equimolar
intravenous injection, a brief period of renal sodium bicarbonate solution was superior to
vasodilation is followed by intensive vasoconstric- normal saline for RCN prophylaxis, infused as
tion, in part mediated by endothelin and adeno- 3 mEq/L/h for 1 hour precontrast, then 1 ml/kg/h
sine. Medullary blood flow is more profoundly for 6 hours.'"t is interesting to note that experi-
affected than is cortical blood flow. Hence, con- mental data from animals support the concept that
comitant administration of drugs that affect RBF urinary alkalinization ameliorates renal ischemia-
particularly NSAIDs, may act synergistically with reperfusion injury, but the mechanism is
radiocontrast to produce ARE This vasoactive unexplained.
mechanism probably explains the increased risk of
contrast nephropathy observed in patients with NAC. NAC is a free radical scavenger and, by
CHF, volume depletion, and nephrotic syndrome, generating nitric oxide, a renal vasodilator. By
as well as the therapeutic benefit of saline loading. comparing changes in mean creatinine values for
Direct tubular toxicity also contributes to ARF treatment groups or absolute changes in creatinine
from radiocontrast. Proximal tubular cells exposed levels, several studies have demonstrated the
to contrast material demonstrate altered cellular effectiveness of Mucomyst (NAC) in preventing
metabolism and intracellular enzyme release, contrast nephropathy."g4" All studies to date have
probably mediated by oxygen free radicals and relied on these changes as surrogates for more
reactive oxygen species. meaningful endpoints such as need for dialysis
and mortality rates. A recent meta-analysis of 8
Prevention strategies randomized controlled trials involving 855
patients reported that the use of Mucomyst
Hydration. Intravenous administration of 0.45% reduced the risk of radiocontrast by 59%." How-
saline has long been the standard of care therapy to ever, emerging data suggest that NAC causes a
reduce the incidence of contrast nephropathy in decrement in serum creatinine (but not cystatin C)
 P R E V E N T I O N O F ACUTE R E N A L FAILURE 83

by a GFR-independent me~hanism,~' perhaps by better middle molecule removal with heniofiltra-

inhibiting creatinine phosphokinase function.43 tion, there is no plausible explanation for its posi-
Nevertheless, given the available data, oral tive effects in removing a low-molecular-weight
administration of Mucomyst to patients at high substance such as radiocontrast. Many of the con-
risk of developing contrast nephropatliy appears trol patients subsequently required dialysis for
warranted. pulmonary edema, probably as a result of the
hydration regimen used in these cardiac patients.
Although s u c l ~patients represent a therapeutic
Fenoldopam mesylate. Limited trials suggested challenge, provocation of pulmonary edema with
that administration of fenoldopani mesylate hydration does not prove the superiority of
reduced the occurrence of ARF from radiocontrast hemofiltration. It is also difficult to interpret the
agents.",'3 However, a large randomized placebo- incidence of contrast nephropathy when one
controlled trial (the CONTRAST Trial) was pub- group has a fall in serum creatinine as a direct
lished that refutes the evidence accumulated in result of the intervention. It is interesting to specu-
the previous reports. In 315 patients at risk for late that bicarbonate administration in the hemofil-
contrast nephropathy undergoing coronary tration group may have been another benefit of
arteriography who were given a standardized this therapy in RCN prophylaxis, in view of the
hydration regimen, randomization to low-dose subsequent data showing a protective effect of
fenoldopam had no effect on urine output, sodium bi~arbonate.'~It is also challenging to
change in serum creatinine levels up to 96 hours, explain how limited hemofiltration would
or the need for dialysis.'"his study is strong evi- decrease the rate of acute myocardial infarction,
dence against the indiscriminate administration ischemic stroke, and multiple organ failure, the
of fenoldopam mesylate for prevention of most common causes of death in the hydration
contrast nephropathy. group. More to the point, treatment details of the
placebo group were inadequately described. From
the available data, it is unclear if the placebo group
Dialysis. Radiocontrast is a small molecule did worse because of the lack of hemofiltration or
easily removed from the circulation by dialysis: because ideal medical management was not
therefore, there remains an interest in performing provided. Given the cost and invasiveness of
dialysis in high-risk individuals who receive radio- hemofiltration, further data will be required before
contrast to prevent the development of ARE In two it can be recommended as a preventive measure in
studies of postcontrast hemodialysis, the subse- contrast nephropathy.
quent need for dialysis for ARF was unchanged or In summary, volume expansion with saline or
increased by prophylactic h e m ~ d i a l ~ s i s . ' ~ ~ ~ ' ~ ' perhaps sodium bicarbonate remains the only
A more recent randomized trial of hemofiltra- proven method for RCN prophylaxis. Avoidance
tion started in high-risk patients before adminis- of iodinated radiocontrast agents in high-risk
tration of radiocontrast demonstrated a benefit of patients is preferred where possible, perhaps by
therapy.12-lIn this study, 114 consecutive patients using alternate imaging methods such as magnetic
with serum creatinine levels > 2 mg/dl under- resonance angiography with gadolinium. When
going coronary interventions were randomly radiocontrast is required, the least nephrotoxic
assigned to either hemofiltration in an ICU or iso- agent available is preferred. Currently, the iso-
tonic saline hydration at a rate of 1 ml/kg/h in a osmolar, dimeric, nonionic contrast medium,
step-down unit. Hernofiltration or hydration was iodixanol is thought to be the least nephrotoxic
initiated 4-8 hours before coronary intervention agent available. In a study of 129 diabetics with
and continued for 18-24 hours after the procedure chronic renal insufficiency, iodixanol caused signif-
was completed. A 25% or more increase in baseline icantly less RCN than the low-osmolar, nonionic,
serum creatinine level, need for temporary dialy- monomeric contrast medium, iohexol." Although
sis, in-hospital mortality rate, and I-year mortality many positive studies suggest benefit, the value of
rate were all significantly lower in the hemofiltra- NAC for RCN prevention is now uncertain.
tion group compared to the hydration group. Finally, although a single positive study suggest-
However, numerous concerns surround this study. ing the use of prophylactic hemofiltration has
Although arguments continue about the superior- added to the growing list of potentially effective
ity of hemofiltration compared to hemodialysis in preventive strategies, we do not regard this as a
treating critically ill patients with ARF because of proven approach.
84 RENAL FAILURE IN THE ICE

Aminoglycosides this finding."' Animal studies also suggest that

administration of calcium channel blockers or
Aminoglycosides cause ATN in approximately aminophylline are protective, although no human
10% of patients who are treated with them for trials have been r e p ~ r t e d . " ~ , ' ~ ~
more than 2-3 days. The serum creatinine typically
rises 7-10 days after the drug is initiated. Amino-
glycosides concentrate in the proximal tubular Cisplatin
cells, causing cellular damage. Risk factors for
aminoglycoside-induced ARF are advanced age, Nephrotoxicity is the most common dose-limiting
chronic kidney disease, volume depletion, liver side effect of cisplatin administration. The primary
disease, and prolonged use of the drug. site for clearance of cisplatin is the kidney. Cis-
Aminoglycosides have concentration-dependent platin asserts its toxicity on the tubules, resulting
bactericidal activity as well as a 'post-antibiotic' in a tubular wasting syndrome that is often severe.
effect on bacterial g r 0 ~ t h . ITherefore,
~~ single- The proximal tubule is most often affected but the
daily dosing is as effective as multiple daily dosing distal nephron is also vulnerable. The direct tubu-
in treating infections by Gram-negative bacteria. lar toxicity associated with cisplatin is exacerbated
The rate of renal cortical uptake of aminoglyco- in a low-chloride environment. In the intracellular
sides is saturable, so accumulation of d n ~ gis less compartment, chloride molecules are replaced
when given in one large dose rather than in with water molecules in the cis position of cis-
divided doses.'" Therefore, single daily dosing platin, forming hydroxyl radicals that injure the
of aminoglycosides should lower the incidence of neutrophilic binding sites on The
nephrotoxicity. In a randomized controlled trial of decline in GFR associated with cisplatin toxicity
123 patients, single daily dosing of gentamicin usually occurs 7-14 days after the exposure. Doses
was as effective in treating infection as multiple of cisplatin >50 mg/m2 are sufficient to cause
dosing, and significantly decreased the incidence renal insufficiency. The renal injury is typically
of ARF. Two meta-analyses also support this reversible but repeated doses of cisplatin in excess
c~nclusion.'~~''" of 100 mg/m2 may cause irreversible renal
damage.
Hydration and avoidance of concomitant
Amphotericin B nephrotoxins is the most effective way to prevent
cisplatin-induced nephrotoxicity. Numerous trials
Amphotericin B and its liposomal derivatives are a proposing various hydration regimens have been
common cause of ATN in the ICU, particularly in reported, but it appears that 3 L of normal saline
patients who have undergone bone marrow trans- over 8-10 hours before and after cisplatin is suffi-
plant. Eighty percent of patients who receive cient to avoid most toxicity associated with con-
amphotericin B will develop some degree of renal ventional doses.138,'" When cisplatin is given in
impairment. The initial nephrotoxic injury from very high doses (>I00 mg/m2), administration in
amphotericin B results from renal vasoconstriction 3% saline is also p r o t e ~ t i v e .Amifostine
'~~ has also
of the preglomerular arterioles and predisposes been shown to reduce cisplatin nephrot~xicity.'~'
the patient to an ischemic ins~lt.'~'~'" Direct tubu-
lar toxicity follows. The newer liposomal forms of
amphotericin B lack the solubilizing agent deoxy- Calcineurin inhibitors
cholate that contributes to tubular toxicity.
Although such agents have been shown to cause The calcineurin-inhibitors CSA and tacrolimus are
less nephrotoxicity, renal failure still develops in a widely used as immunosuppressants in solid
significant proportion of patients.I3' organ and bone marrow transplantation. CSA and
Effective strategies to prevent the nephrotoxicity tacrolimus cause both ARF and cl~ronicrenal fail-
of amphotericin B are not clearly defined due to ure. The nephrotoxicity seen in the critically ill
the lack of rigorous clinical trials and the inherent patient is the result of direct afferent arteriolar
complexities of patients who require amphotericin vasoconstriction, leading to a decrease in the
B. Several retrospective or uncontrolled prospec- glomerular filtration pressure and GFR. The vascu-
tive trials suggest that salt-loading reduces the risk lar effect associated with CSA and tacrolimus is
of n e p h r o t o ~ i c i t y . ~ ~
A, ' ~small,
~ randomized reversible with discontinuation of the drug. A dose
placebo-controlled trial of 20 patients confirmed reduction is sometimes enough to reverse the
 P R E V E N T 1 0 1 O F A C U T E R E N A L FAILURE 85

prerenal effect. Calcium channel blockers (CCBs)

decrease ischemic damage and reverse CSA- Box 5.3 Causes of rhabdomyolysis
induced renal vasoconstriction in animal mod-
e l ~ . ' ~However,
: small clinical trials in humans D~rectinjury
have had conflicting results.143,'"It has been sug- C r u s h injurles
gested that the response to CCBs may be Compression
due to their ability to increase plasma levels of Seizures
CSA, thereby decreasing acute rejection, as well as High-voltage eiectr~calinjury
directly modifying T-cell f ~ n c t i o n . ' ~ "
Hemolytic uremic syndrome (HUS) is a rare
Vascular ~nsufficiency
complication of CSA and tacrolimus therapy. The
mechanism of CSA- or tacrolimus-induced HUS is Compartment syndrome
direct damage to the vascular endothelium in a Shock
dose-denendent fashion. With discontinuation of
the drug, patients may have partial recovery.'4h Metabolic disorders
The utility of therapeutic plasma exchange in the Hypophosphatemia
treatment of HUS induced by calcineurin Hypokalemia
inhibitors has not been well studied Dtabetic ketoacidosis
Hypothyroidism
Myoglobin Toxic injury
Alcohol
The principal causes of pigment nephropathy are
(1)rhabdomyolysis and (2) hemoglobinuria due to Amfetamines
hemolysis. The majority of rhabdomyolysis cases Cocaine
are subclinical, with mild elevations in the creatine Snake venom
kinase, lactic dehydrogenase, or aspartate amino-
transferase enzyme levels. In severe cases, ARF Infections
may ensue from myoglobinuria. Commonly Toxic shock syndrome
encountered causes of rhabdomyolysis are listed in Influenza
Box 5.3. Rhabdomyolysis has been thought to Malarla
cause ATN through three mechanisms: renal vaso- HiV infection
constriction, intratubular cast formation, and
heme-mediated proximal tubular injury. It is also
Myopathies
known that oxidant stress is increased with the
release of heme proteins. Free heme proteins are McArdle's disease
suspected of reducing the formation of nitric oxide Phosphofructok~nase(PFK) deficiency
and increasing endothelin levels, which results in Carnitine palmityl-transferase deficiency
vasoconstriction and the decline in GFR. Dermatomyosit~s
Intratubular obstruction occurs with the inter- Polymyositis
action of myoglobin and Tamm-Horsfall protein in
an aciduric environment.14'
To prevent and treat the ARF of rhabdornyolysis, effect stem from its serving as a nonreabsorbed
aggressive hydration is effective. Alkalinization of solute which promotes an osmotic diuresis.14' It is
the urine has also been advocated to increase the not known whether or not alkalinization of the
solubility of the heme proteins in the urine, with urine is beneficial once a brisk diuresis is
the goal of achieving a urine pH >6.5. Alkaliniza- established with saline or diuretics.
tion may also reduce the production of reactive Mannitol has also been advocated for the treat-
oxygen species, thus reducing the oxidant stress.14h ment of pigment nephropathy based on experi-
Caution must be exercised when administering mental models oi myohemoglobinuric ARF. The
bicarbonate, since the resulting alkalemia can protective effect has been attributed to its diuretic,
worsen the hypocalcemia often present in patients renal vasodilatory, and hydroxyl sca\,enging prop-
with severe rhabdomyolysis and precipitate gener- erties.''['.'" However, in a glycerol model of myo-
alized seizures. Some of bicarbonate's salutary globinuric ARF, mannitol's protective effect could
86 RENAL FAILURE I N THE ICU

be completely ascribed to a solute diuresis and characterized by the development of hyperphos-

increased heme e~cretion."~ Also, although manni- phatemia, hypocalcemia, hvperuricemia, and
to1 is a potent renal vasodilator, it may actually hyperkalemia. Tumor lysis syndrome can occur
worsen cellular energetics during the induction of spontaneously during the rapid growth phase of
ARE If administered immediately after renal malignancies such as bulky lymphoblastomas and
ischemia, renal cortical ATP levels may abruptly Burkitt's and non-Burkitt's lymphomas that have
decline by increasing the metabolic cost of sodium extremely rapid cell turnover rate~.'~%orecom-
reabsorption by the loop of Henle."i-"3 Therefore, monly, it is seen when cytotoxic chemotherapy
intravenous fluids or perhaps furosemide may be induces lysis of malignant cells in patients with
the preferable diuretic. If mannitol is used, it is large tumor burdens. Tumor lysis syndrome has
essential to monitor serum osmolality to avoid a developed in patients with non-Hodgkin's lym-
hyperosmolar state. phoma, acute lymphoblastic leukemia, chronic
Because of its size, myoglobin removal by peri- myelogenous leukemia in blast crises, small cell
toneal dialysis or hemodialysis is poor, although a lung cancer, and metastatic breast cancer.Ih0 In
study by Amyot and colleagues suggests its clear- most patients the ARF is reversible after aggressive
ance may be enhanced by continuous hemofiltra- supportive therapy including dialysis.
tion."-' In addition, myoglobin levels fall The pathophysiology of ARF associated with
exponentially with cessation of muscular release tumor lysis syndrome is related to two main fac-
due to hepatic and splenic uptake.155It is also tors: (I) pre-existing volume depletion prior to the
unclear whether or not myoglobin is a major con- onset of renal failure and ( 2 ) the precipitation of
tributor to the pathogenesis of rhabdomyolysis- uric acid and calcium phosphate complexes in the
induced ARE Currently, there is no compelling renal tubules and t i ~ s u e . ' ~Patients
' may be
evidence for the use of extracorporeal therapies in volume-depleted from anorexia or nausea and
the treatment of myoglobinuric ARF. vomiting associated with the malignancy, or from
increased insensible losses from fever or tachy-
pnea. Therefore, it is important to establish brisk
Methotrexate flow of hypotonic urine to prevent or ameliorate
ARF associated with tumor lysis syndrome.
Methotrexate (MTX)-induced ARF is caused by the Hyperuricemia is either present before treatment
precipitation of the drug and its metabolites in the with chemotherapy or develops after therapy
tubular lumen and is also a tubular cyt~toxin.';~ despite prophylaxis with a l l o p ~ r i n o l . ~Uric~ ' acid
High doses of MTX ( > I g/m2) increase the risk of is nearly completely ionized at physiologic pH, but
ARE Once ARF develops, the excretion of MTX is becomes progressively insoluble in the acidic
reduced and the systemic toxicity of MTX is environment of the renal tubules. Precipitation of
increased. Hydration and high urine output are uric acid causes intratubular obstruction, leading
essential to preventing MTX renal toxicity. Isotonic to increased renal vascular resistance and
saline infusion and furosemide may be necessary to decreased GFR.Ih3 Moreover, a granulomatous
keep the urine output >I00 ml/h. Alkalinization of reaction to intraluminal uric acid crystals and
the urine to a pH >6.5 is also recommended, to necrosis of tubular epithelium can be found on
decrease tubular precipitation and increase renal biopsy specimens.
MTX clearance. Once ARF has developed, it may be Hyperphosphatemia and hypocalcemia also
necessary to remove the drug with dialysis. occur in tumor lysis syndrome. In patients who do
Hemodialysis using high blood flow rates with a not develop hyperuricemia in tumor lysis syn-
high-flux dialyzer is an effective method of remov- drome, ARF has been attributed to metastafic
ing m e t h 0 t r e ~ a t e . High-dose
l~~ leucovorin therapy intrarenal calcification or acute nephrocal~inosis.~~
can reduce the systemic toxicity associated with Tumor lysis with release of inorganic phosphate
MTX and ARF, and is routinely used when plasma causes acute hypocalcemia and metastatic
levels are excessive 48 hours post-dose.I5' '-X calcification, resulting in ARF.
Therefore, ARF associated with tumor lysis
syndrome is the result of the combination of vol-
Tumor lysis syndrome ume depletion in the face of urinary precipitation
of uric acid in the renal tubules and parenchyma,
Tumor lysis syndrome is often a dramatic presen- and acute nephrocalcinosis from severe hyper-
tation of ARF in patients with malignancy. It is phosphatemia. Since patients at risk for tumor
 P R E V E N T I O N O F A C U T E R E N A L FAILURE 87

lysis often have intra-abdominal lymphoma, uri- Liver failure, bacterial peritonitis, and
nary tract obstruction can be a contributing factor hepatorenal syndrome
in the development of ARF. Given the aforemen-
tioned pathogenetic factors for ARF, patients who Spontaneous bacterial peritonitis (SBP) is a com-
are undergoing treatment with malignancies who mon and severe disorder in patients with liver cir-
are likely to experience rapid cell lysis should rhosis and ascites. In one-third of patients, renal
receive vigorous intravenous hydration to main- impairment develops despite treatment of the
tain good urinary flow and urinary dilution. In infection with non-nephrotoxic antibiotics. Devel-
addition, because uric acid is very soluble at opment of renal failure is the best predictor of
physiologic acid urine pH, sodium bicarbonate hospital mortality in these patients and is thought
should be added to the intravenous fluid to to result from a decrease in the effective arterial
achieve a urinary pH >6.5. Since metabolic alka- blood volume caused by sepsis. In a randomized
losis can aggravate hypocalcemia, caution should trial of 126 patients with SBP, administration of
be exercised when using alkali in patients with albumin plus antibiotics compared to antibiotics
low serum calcium levels. It is advisable to stop alone significantly decreased the incidence of renal
the infusion if the serum bicarbonate level is >30 impairment, in-hospital mortality, and 3-month
mEq/L. Furthermore, it is conceivable that the m~rtality.'~ This study was not blinded, potentially
use of urinary alkalinization to prevent uric acid introducing bias, and the quantity of albumin used
nephropathy might actually precipitate or worsen was substantial, leading to significant cost. Also,
calcium phosphate-induced ARF. Allopurinol is there was no 'hydration' arm in the antibiotic-only
administered to inhibit uric acid formation. group that would provide proof of the superiority
Through its metabolite oxypurinol, allopurinol of albumin over crystalloid solutions. However,
inhibits xanthine oxidase and thereby blocks the these encouraging results will probably lead to
conversion of hypoxanthine and xanthine to uric further clinical investigations.
acid. During massive tumor lysis, uric acid excre- Hepatorenal syndrome (HRS) is a unique cause
tion can still increase despite the administration of renal vasoconstriction, with a decline in GFR in
of allopurinol, so that intravenous hydration is the face of normal renal histology that occurs in the
still necessary to prevent ARF. Since allopurinol setting of liver failure. The clinical picture associ-
and its metabolites are excreted in the urine, the ated with HRS is that of prerenal azotemia with
dose should be reduced in the face of impaired oliguria and low FENa (urinary fractional excre-
renal function. Uricase has been recently tion of sodium). In true HRS without confounding
approved for use in the United States. It converts renal injuries, the renal failure will resolve with
uric acid to water-soluble allantoin, thereby liver transplantation. The pathogenesis of HRS is
decreasing serum uric acid levels and urinary uric incompletely understood. Systemic and splanchnic
acid e ~ c r e t i o n . ' ~
The
' use of uricase may obviate vascular resistance is decreased, leading to a
the need for urinary alkalinization, but good decrease in the effective arterial blood volume and
urine flow with hydration should be maintained renal hypoperfusion. The compensatory hemo-
given the probability of pre-existing volume dynamic response to systemic vasodilation
depletion. includes an increase in the mediators of renal
Dialysis for ARF associated with tumor lysis vasoconstriction, including increased renin-
syndrome may be required for the traditional angiotensin-aldosterone activity, antidiuretic hor-
indications of fluid overload, hyperkalemia, mone (ADH) levels, sympathetic tone, and
hyperphosphatemia, or hyperuricemia unrespon- endothelin levels. The renal response is an increase
sive to medical management. There is some inter- in salt and water avidity, leading to worsening
est in using dialysis in patients at high risk of ascites and edema.167,1hs Liver transplant is the
tumor lysis syndrome to prevent the develop- definitive therapy for HRS. However, patients who
ment of renal failure. In a small trial involving develop HRS prior to transplant have worse graft
five children, continuous hemofiltration was and patient survival.1h9
started prior to administration of chemotherapy Newer pharmacologic therapy with vasopressin
and appeared to prevent renal failure in 80% of analogs (e.g. ornipressin and terlipressin), which
the p a t i e n t ~ . ' ~ W o w e v e rgiven
, that continuous are splanchnic vasoconstrictors, has shown some
dialysis is complicated, expensive, and not with- benefit.""-"' Ornipressin and albumin was admin-
out risk, its routine use as prophylaxis cannot be istered to a total of 16 patients with HRS for either
recommended. 3 or 15 days (8 patients in each group). The 3-day
88 RENAL FAILLRE IN THE I C E

regimen was associated with a normalization of improved early graft f u n c t i o n . " ~ o m e studies
the overactivity of renin-angiotensin and sympa- have suggested that intraoperative administration
thetic nervous svstems, ANP levels, and only a of mannitol decreases the incidence of ATN.179
slight improvement in renal function. ~ o w e " e r , Decreasing warm and cold ischemia times should
treatment for 15 days resulted in improved serum also decrease the occurrence of post-transplant
creatinine levels, renal plasma flow, and glomeru- delayed graft function. The use of the University of
lar filtration rate. Similar results were seen in 9 Wisconsin preservation solution during cold
patients who received intravenous terlipressin. ischemia reduces the incidence of delayed graft
However, a major complication associated with f u n ~ t i o n . ' Renal
~ ' ~ vasodilators such as low-dose
these medications is mesenteric ischemia. Oral dopamine and fenoldopam, IGF-1, and ANP have
midodrine (a selective a,-adrenergic agonist) in not shown significant benefit in small clinical
combination with octreotide also showed benefit trial^.^^,^^ As previously discussed, although
in renal function in a small series of patients.17' CCBs reverse CSA vasoconstriction and prevent
Mucomyst given intravenously to 12 patients ischemic injury in animal models, the benefit in
increased RBF without changing the hemody- clinical trials has been inconsistent. Current clini-
namic derangements
" associated with HRS.~;-' cal trials aimed at lessening reperfusion injury by
Several small studies have shown that transjugular blocking adhesion molecule interactions are
intrahepatic portosystemic shunting (TIPS) has ~ngoing.'~~,'~~
prolonged survival and improved renal function in
patients with HRS.175,176 Firm conclusions about
efficacy of these therapies await the results of Postoperative states
randomized controlled trials.
Many critically ill patients who undergo surgery
will go on to develop ARF, and a large proportion
Renal transplantation of these will require renal replacement therapy.18'
In addition to the traditional risk factors for ARF
Acute renal failure remains a common complica- present in any critically ill patient, the surgical
tion of renal transplantation. It may occur at any patient may be exposed to the potentially harmful
time point in the life of the transplant, although it effects of cardiopulmonary bypass, hypothermic
usually develops in the immediate postoperative circulatory arrest, or aortic c r o s ~ c l a m p .The
' ~ most
period and is referred to as delayed graft function. common cause of ARF in the surgical setting is
The causes of delayed graft function are listed in
Box 5.4, and include obstruction, volume deple-
tion, and acute rejection, although the most com- Box 5.4 Causes of delayed graft function in renal
mon etiology remains ischemic ATN. Risk factors
transplant
for ATN are advanced donor age, intraoperative or
postoperative hypotension, prolonged warm or
cold ischemia times, and initial high CSA dosage. Prerenal
Reperfusion injury as a result of direct endothelial Hypovolemia
trauma, oxygen free-radical generation, and Renal artery thrombosis
neutrophil activation also contributes to the
development of ATN. Renal
Patients with delayed graft function have longer Ischemic ATN
hospitalization rates and more complications, Hyperacute rejection
including a lower 5-year graft survival rate.17'
Acute rejection
Therefore, prevention of ischemic ATN in the post-
Acute calcineurin nephrotoxicity
transplant setting may prolong renal survival. As
in all patients at risk for the development of
ischemic ATN, optimization of hemodynamic Postrenal
parameters in both recipient and donor is a key Urinary tract obstruction
element and may require monitoring of central Lymphocele
venous pressures; Carlier and colleagues showed Urinary leak
that higher wedge pressures at the time of renal Ureteral necrosis
allograft revascularization were associated with
 I'REVENTION O F A C U T E R E N A L FAILURE 89

iscl~emicATN that occurs in the face of critical ill- monary bypass is a state of hypotension that trig-
ness and MODS. Often, there has been a preceding gers release of several renal vasoconstrictor agents,
period of relative renal hypoperfusion from either resulting in renal hypoperfusion. Although overt
true or effective volume depletion. ARF may develop in less than 5 % of patients with
normal preoperative renal function, tubular enzy-
muria suggests that subclinical injury is much
Risk assessment more common."' Preoperative left ventricular dys-
function and time on CPB increase the risk of ARF.
In terms of risk assessment and characterization, The type of CPB (pulsatile or nonpulsatile) and
there are no specific scoring systems that can pre- bypass pressure do not appear to be significant
dict accurately who will develop postoperative factors in determining renal outcome.
organ dy~function.'~'The American Society of Vascular surgery, particularly when cross-
Anesthesiologists (ASA) scoring system is among clamping the aorta is required, is associated with a
the simplest and most reproducible general risk higher incidence of postoperative ARF.I9' The best
assessment tools.'RhIt stratifies patients into high-, predictors of ARF are pre-existing renal insuffi-
moderate-, or low-risk categories but does not ciency and hemodynamic instability during
specify the actual type of harm. Several studies surgery.
have demonstrated that the presence of pre-
existing renal disease, CHF, obstructive jaundice, Preoperative period
diabetes mellitus, peripheral vascular disease, In general terms, the predisposition to a prerenal
hypertension, and coronary artery disease are all state can often be avoided by an overnight intra-
risk factors for developing postoperative ARF.187 venous fluid infusion. Surgical patients are typi-
This observation suggests that patient-related fac- cally starved overnight, and administration of a
tors have a significant impact on the development bowel preparation routine can compound the vol-
of postoperative complications. In addition, over- ume depletion. Intravenous fluid administration is
all risk can be thought of as having both genetic a simple means of preventing the development of
and environmental components. The environmen- a prerenal state from hypovolemia.
tal risk consists of the nature of the planned opera- More specific therapies begun preoperatively to
tion, its urgency, and the surgical skill of the prevent postoperative ARF have not been proven
personnel involved. The environmental risk effective. In part, because the incidence of post-
further entails those unpredictable events such as operative ARF is low, large numbers of patients
catastrophic intraoperative hemorrhage and tech- would need to be randomized in order to either
nical errors. Genetic risk is subtle and has only demonstrate a positive effect, or to conclusively
recently become recognized as important. The field show such an effect does not exist. This problem
of the genetics of complex-trait diseases is evolv- has plagued ARF clinical research in general, and
ing, and the best tools for its analysis are still under the surgical setting has been no exception. As pre-
debate. Currently, how genetic variability affects viously detailed, various trials using low-dose
an individual's response to underlying illness, dopamine, the diuretics furosemide and mannitol,
injury, treatment, and drug therapy is poorly growth factors, and ANP have failed to decrease
defined. However, genetic variability probably has the incidence of ARF. Small trials suggest that pre-
some effect on both the risk of developing a post- operative administration of fenoldopam mesylate
operative complication and on the response of the and N-acetylcysteine may decrease the develop-
individual to its treatment. ment of postoperative ARF,L""lU%utthese require
Certain operations carry a higher risk for ARF larger confirmatory effectiveness trials. However,
because of the nature of the surgical procedure and to date, there are no large, randomized controlled
the underlying medical condition of patients clinical trials examining these agents in the
requiring the surgery Patients undergoing surgery operative setting.
for either traumatic or thermal injuries are at high Recently, studies have been published suggest-
risk for ARF.IBnThe ARF is usually multifactorial in ing that for those patients with chronic renal
nature, with ischemia, rhabdomyolysis, nephro- impairment a period of elective renal replacement
toxins, and sepsis all having a contributory role. therapy may subsequently improve both renal and
Cardiac surgery requiring cardiopulmonary ovei-all outcome. In a single study of 44 patients
bypass (CPB) may cause some degree of renal dys- with chronic renal failure (mean serum creatinine
function in up to 50% of patient^.'"^"^ Cardiopul- level of 3.3 mg/dl) undergoing coronary artery
90 R E N A L F A I L L R E I N T H E ICU

bypass grafting (CABG) and CPB, randomization major surgery are not known because of a lack of
to prophylactic hemodialysis appeared to decrease appropriate studies. No studies have proven that
mortality hospital and ICU length of stay, and the intensive hemodynamic monitoring improves out-
incidence of postoperative ARF compared to stan- come.lyy~'OOIn fact, some studies actually suggest
dard care.''? However, the study was small and that the use of pulmonary artery catheters is asso-
details of medical therapy in the nondialysis arm ciated with worse For now, inter-
(such as the use of ACE inhibitors, diuretics, and ventions that maximize the chance of good renal
fluid balance) were not described. Furthermore, it outcome include avoidance or treatment of hypox-
is difficult to explain the benefits based on any emia, hypercarbia, hypotension, hyperglycemia,
sound physiologic principle. If frank fluid over- and anemia.
load, which should ha1.e prompted therapeutic
dialysis, was not present in these patients, then it is Postoperative period
hard to explain how limited small solute clearance Immediately after surgery, the emphasis on pre-
would decrease mortality or the incidence of post- vention of ARF shifts away from monitoring the
operative ARF. In fact, there is good evidence that minute-to-minute changes- in hemodynamics to
hemodialysis can prolong or perpetuate ARF by observation for bleeding, organ hypoperfusion,
triggering ischemia or i n f l a m m a t i ~ n . ' ~ In
~ ~ an
'~' infection, and coagulation complications. Hypox-
uncontrolled cohort study of patients on chronic emia may arise from shunting through atelectatic
hemodialysis, intensive dialysis prior to cardiotho- areas of lung, or hypoventilation from narcotic
racic surgery in 13 consecutive patients resulted in analgesics
" o r residual neuromuscular blockade.
similar postoperative outcomes compared to Impaired oxygen delivery can exacerbate any
patients with normal renal function.lo8 Both of reduction in cardiac output, leading to decreased
these studies probably demonstrate that in renal renal perfusion. Large fluid shifts may accompany
failure patients, aggressive management of fluid major surgery and can manifest as severe anemia,
overload and electrolyte imbalances results in electrolyte disturbances, acid-base abnormalities,
better postoperative outcomes. However, proof or changes in cognition.
that dialysis is superior to medical management Later in the postoperative course, the greatest
will depend on the results of large, randomized risk is posed by the development of severe sepsis
controlled clinical trials. syndrome, which can cause postoperative ARF
from ischemic ATN. Semis- induces vascular
Intraoperative period endothelial changes, organ and tissue ischemia,
Most anesthetic agents cause dose-dependent cellular apoptosis and necrosis, as well as coagula-
venous and arterial vasodilatation, with an accom- tion abnormalities that exacerbate and accelerate
panying reduction in cardiac pre- and afterload. the pathologic p r o c e ~ s . ~To
" date, activated protein
Hypotension may be exacerbated by neural C has been the only agent administered to critically
impairment from spinal or epidural blockade. A ill surgical patients that has significantly improved
mild reduction in blood pressure leads to less survival rates in the ICU.""
bleeding and is usually well tolerated, since anes-
thetic agents generally also reduce oxygen
demand. Typically, crystalloid or colloid solutions Abdominal compartment syndrome
are administered to expand the intravascular space
in response to the mild drop in blood pressure that Abdominal compartment syndrome (ACS) was
almost universally accompanies induction of anes- first reported in 1876, in a paper describing thtl
thesia. A urine output of 1-2 ml/kg/h during sur- reduction in urine flow associated with elevated
gery is considered evidence of adequate organ intra-abdominal pressure (IAP)."'.' Acute increases
perfusion, although adequate urine output cer- in IAP are deleterious for both intra-abdominal
tainly does not preclude the presence of renal and distant organ function, including the kid-
hypoperfusion and ischemia. Hemodynamic mon- neys.'"' Acutely, the abdomen functions as a closed
itoring, including central venous pressure (CVP), space; thus, any increase in the volume of its con-
cardiac index, systemic vascular resistance, and tents leads to a rise in compartmental pressure.
pulmonary artery occlusion pressure, is often used Intra-abdominal hemodynamics are compromised
in critically ill patients. However, the ideal blood when the IAP approaches and then exceeds
pressure, cardiac output, vascular tone, and 10 cmH,O. ACS is present when the IAP reaches
intravascular volume for patients undergoing 20-25 c m ~ , ~ and , unless decompressed,
 PREVENTION O F ACUTE RENAL FAILURE 91

irreversible organ failure may r e s ~ l t . ' ~ACS

" is Multisystem trauma
characterized by an acute rise in IAP, coupled with
evidence of organ dysfunction, usually reduced Altliough the incidence of ARF follo~vingmajor
urine output. Tlie pathogenesis of this reduction in trauma is generally low, it is associated with a high
urine formation is complex and attributed to three mortality rate.20"2"" Opinion has varied as to
major factors: whether pre-existing medical conditions such as
diabetes and hypertension are more important
1. Compression of the great veins reduces risk factors than the timeliness and type of
venous return to the right lieart, which mani- resuscitation.:"'
fests as relative hypovolemia and can become Mechanical ventilation with high PEEP (positive
frank volume depletion if third spacing of end-expiratory pressure), hernoperitoneum, and
fluid is extensive. Increased renal venous back rhabdomyolysis have been identified as the three
pressure associated wit11 high CVP may cause conditions most strongly associated with ARF after
renal vein compression. major trauma.?" Rhabdomyolysis may be caused
2. Direct pressure on the renal cortex can shunt by extensive skeletal muscle damage and wors-
blood away from the vulnerable cortico- ened by subsequent infection, fever, drugs, and
medullary junction by altering renal vascular elcctrolyte abnormalities. The degree of elevation
resistance and induce an ischemic injury in of serum creatine kinase is usually proportional to
this area. the extent of the muscle injury. Acute renal failure
3. Direct pressure on the ureters can lead to develops in rhabdomyolysis because of tubular
obstructive nephropathy, particularly if there obstruction, oxidant injury, and renal vasoconstric-
is a predisposing condition such as extensive tion, and thus preventive measures have been
lymphadenopathy. largely aimed at promoting tubular tlow and
restoring renal p e r f u ~ i o n . ' The
~ ~ single most
Regardless of the underlying cause, a reduction in important feature of ART; prevention is volunic
urine output t azotemia in the presence of a expansion. Alkalinization of the urine is usually
measured 1AP over 15 cmHzO is certainly cause for recommended also, since this improves myoglobin
concern and should prompt intervention. washout, prevents lipid peroxidation, and miiii-
1AP can be measured in a number of ways, niizes renal vasoconstriction.'"; The use of manni-
although the transvesical method is the most corn- to1 is controversial. Although it has theoretic
monly In this technique, 50-100 nil of ster- benefits such as reduced compartment pressure
ile saline solution is instilled through the bladder and cellular swelling, and mild antioxidant prop-
erties, it has not been shown to produce results
catheter illto the bladder and the drainage port
superior to those of volunie expansion alone."?
then clamped distal to the sampling port. A con-
ventional pressure monitoring line is then attached
by hypodermic needle through the sampling port
and the pressure measured directly, with tlie sym-
physis pubis taken as tlie zero reference point. This
technique is straightforward, reliable, simple, and Box 5.5 Causes of Increased Intra-abdominal
reproducible. An estimate of IAP can also be pressure
obtained by elevating the bladder catheter above - - - - -

the bed, since the pressure of urine in the bladder Peritoneal tissue edema (trauma,peritonitis)
will then equilibrate with that in the tubing. Tliis
Fluid overload in shock
method represents a simple screening tool, and
Retroperitoneal hematoma
thus should be used in any patient who may be at
risk of ACS. Surgical trauma
Tlie causes of an increased 1AP are listed in Box Reperfusion injury after bowel ~schem~a
55. ACS should be managed with attention to Pancreatitis
preservation uf underlying organ function, and is lleus or obstruction
~ ~ s u a l ltreated
y with urgent surgical deconipres- Abdominal packing to control hemorrhage
51011.Clearly, this is a highly complex situation and Abdominal closure under tension
skilled surgical supervision is mandatory if the Severe ascites
~ ~ ~ i t i cis~ ultin~ately
iit to do well.
92 R E N A L F A I L U R E IN THE ICE

SUMMARY 14. Hirschberg R, Kopple J , Lipsett P, rt sl. Multlienter clinical trial of

recombinant human ~nsulin-likegrowth factor I in pltients with
acute renal t'lilure. K ~ d n e yInt 1999;55:2423-32
Acute renal failure as a result of ATN is a common
15. Hntala R, Dinh T, Cook DJ. Once-dallv an~inoglycosldedosing in
development in critically ill medical and surgical ~mmunocompetent ~ d u l t s :a metd-analysis. Ann Intern Med
patients that is associated with significant morbid- 1996;124:717-25.
ity and mortality. Despite improvements in the Ih. B ~ r r e t tBI, Cdrllsle EJ. Met,3,lnalysis of the rel.rt~venephrotoxlcity
understanding of the pathophysiologic mecha- of high- nnd low-osmolality lodinated contrast media. Rad~ology
1993.18R:l71-X.
nisms of ARF and the numerous agents effective in 17. Sorkine P, Nagnr H, Weinbroum A, et al. Administration ot
treating ARF in animal models, there are currently amphoter~cin B In lipid emulsion decreases nephrotoxicity:
no effective treatments for ATN in human subjects. results ot a prospective. randomized, c o ~ ~ t r o l l study
ed in critically
Therefore, prevention of ARF remains the best way 111patients. Crit Care Med 1~96;24:1311-15.
18. Shusterman N, Strom BL, hlurrav TG, et al. Risk factors and
to improve renal dysfunction-related outcome in
outcome of hospital-acquired ncute renal fnllure. Clinical
the ICU. To date, most specific pharmacologic epidemiologic study Am J Mrd 198;;83:65-71
agents used to prevent ARF have been proven 19. d e btendonca A, Vincent JL, Suter PM, et a!. Acute r m a l failure in
ineffective in clinical trials. Until further data are the ICE: risk factors and outcome evaluated by the SOFA score
available, it seems that simple measures such as Intensive Care Med 2000;26:915-21.
20. Erlksen 8 0 . Hoff KR, Solberg S. Prediction of acute rend1 failure
maintaining adequate renal perfusion, avoiding after cardiac surgery: retrospective cross-\'alidation of a clinic,ll
excessively aggressive diuresis and nephrotoxins, ~ l g o r i t h m Nephrol
. Dial Transplant 20(13;18:77-81
and monitoring levels of potentially nephrotoxic 21. Dounia CE, Redekop WK, van der Meulen JH, et a1 Predicting
drugs are important goals in the care of the mortality in intensive care patients wlth acute renal failure
treated with dialysis. J Am Soc Nephrol lQ97;S:lIl-17.
critically ill patient. 22. Halstenberg WK, Goormastic M, Paganini EP Validity of four
models for predicting outcome in critically ill acute renal failure
patients. Clin Nephrol 19%';47:814.
23. Liano F, Gallego A, Pascu,ll J . et a1. Prognosis of acute tubular
REFERENCES necrosis: an extended prospectively contrasted study. Nephron
1993;63.21-31.
1. Hou SH, Bushinsky DA, Wish JB, Cohen JJ. Harrington IT. 21. Thurau K, Boyldn JW. Acute renal success. The unexpected loglc
Hospital-acquired renal insutticiency: 1' prospective s t u d y A m J oi oliguria in acute renal failure. Am J .Med 1976;61.308-15.
Med 1983;74:243-8. 25. Gattinoni L, Brazzi L, Pelosi P, et al. .A trial 01 goal-orlented hemo-
2. Albright RC, Jr. Acute renal failure: a practical update. Mayo Clin dvnamlc therapy in cr~tically ill patients. SvO2 Collaborative
Proc 2001;76:67-74. Group. N Engl J Med 1995;333:102532.
3. N.+,h K, Hafeez A, Hou S. Hospital-acquired renal ~nsufficiency. 26. Rivers E, Nguyen B, Havstad S, et dl. Earlv goal-directed therapy
Am J Kidney Dis 2002;39:93M. In the treatment o i severe sepsis and septic shock. N Engl [ Med
4. Brivet FG, Kleinknecht DJ, Loirat P, Landais PI. Acute renal fail- 2001,345:136?-77.
ure In intensive care units--causes, outcome. and p r o p o s t i c iac- 27. Stein RS, Albridge I<, Lenox RK, Ray CV, Rexner JM. Neplirotoxi-
tors o i hospital mortality; a prospective, multicenter study. city in leukemic patients receiving empirical amphotericln B and
French Study Group o n Acute Renal Failure. Crit Care Med am~noglycosides.South Med J 1988;811095-9.
199b;24:192-8. 28 Solomon R, Werner C . Mann D, D'Elia J, Silva P Effects of saline,
5,Hevnian SN, Lieberthal W, Rogiers P, Bonventre JV. Animal mod- mannltol, and fi~rosemideto prevent acute decreases in renal
els of acute tubular necrosis. Curr Opin Crit Care 2002;?:52b-14. function induced bv radiocontrast ,?gents. N Engl J bled
6. Rosen 5, Hevman SN. Difficulties In understanding human '.icute 1994;331:1416-20.
tubular necrosis': limited data and flciwedanimal models. Kidney 29. Schierhout G, Roberts I. Fluid resuscitation with colloid or crys-
Int 2001;60:12204. tdlloid solutions in critically ill patients: a svstemc~ticrelriew of
7.Annat G, Viale JP. Bui Xuan B, et al. Eifect of PEEP ventilation o n randomised trials. BhlJ 1998;316:9614.
renal function, plasma renin, aldosterone. neurophvsins and uri- 30. Choi PT, YIP G, Quinonez LG, Cook DJ. Crystalloids vs, colloids
nary ADH, and prostaglandins. ,\nesthesiology 1983;58:13h+l 111 tluld resuscitation: a systematic review. Crit Care Med
8,Pannu N,Mehta RL. Mechanical ventil,jtion , ~ n drenal function. 1999;27:200-10.
.in area tor concern' Am J Kidney Dis ?002;39:61b-21. 31. Boldt J, Mullcr M, Mentges D, Papsdorf Ivl, Hempeimann G Vol-
9 i c h e t z .V. Vasopressors and the kidney Blood Purif ume therapy in the critically ill: is there .i difference? lntensive
?002;211:213-51. Care Med 1998;24:28-36.
10. Hevman SN. Fuchs S, Brezis M. The role oi medullary ischemia in 32. Schortgen F, Ldcherade JC. Bruneel F, et .ll. Effects of hydrox-
acute renal tailure. New Horiz 1995;3:597-607. vetlivlstarch and gelatin on renal iunction In severe sepsis: a
11. Wellv KJ, Molitons BA. Acute renal f,jilure in the new millenniuni: multicentre randomised study. Lancet 2001;357 911-16.
time to consider combination therapy Semin Nephroi 33. Finier 5, Bellomo R, Boyce N, et al. A comparison of albumin and
2000;20:+19. s.lline for iluid resuscitation in the Intensive care unit. N Engl J
12. Sutton TA, Fisher CJ, Molitoris BA. M~crovastularendothelial Med 2004;350:2247-5h
injurv 'lnd dysfunction during ischemic acute rencl1 fa~lure. 34. Sort P, Nav.isa M,Arroyo V, et ,jl. Effect of intravenous ,jlbumin
Kidney Int 2002;62: 153944. o n renal ~mpairmentand mortality in patients with cirrhosis and
13. Allgren RL, .Marbury TC, Rahman SN, et nl. .Anxitide In acute spontaneous bacterial peritonit~s.N Engl J Med 1Q99;?41:403-9.
tubular necrosis. Auriculin Anantide Acute Renal F,lilure Study 3. Mueller C, Buerkle G, Buettner HJ, et a 1 Prevention of contrast
Group. N Engl J Med 1997;336:828-31. medi'l-associated nrpliropathv: randomized comparison of 2
 PREVENTION O F ACUTE REKAL FAILURE 93

hvdratlon regimens in 1620 patients undergoing coronary 5 5 . Tumlin JA, Finkel, K, Murray P, Shaw A. Dopamine 1 receptor
angioplasty. Arch Intern Med ?0112:lh2:329-36. agonists In early acute tubular necrosis: A prospective random-
36. Merten GJ, Burgess M'r, Gr.?\. Li', et al. Prevention of contrast- ized double blind piaceho-controlled trial of tenoldopam
induced nephropathy with sodium bicarbonate: a randomized mesylate. J Am Soc Nephrtd (abstract) 2003;14.
controlled trial. JAMA 2004;291:2328-34. 56. Lassnigg A. Donner E, Grubhofer G. et al. Lack of renoprotective
17. Shiplev RE, Study RS Changes in renal blood t l o ~ .extract~on
. of eftects of dopamine and h r o s e m i d e during cardiac surgery. J .Am
inulin, gloerular tiltration rate, tissue pressure. a n d urlne flow Soc Nephrol 2000.11:97-104.
~ \ ~ i tacute
h alterations 111 renal artery pressures Am J rhysiol 57. Sirivella S, GielchinsLy 1, Parsonnet V. Mannitol, hirosemidt.. and
1951,167:676-$8. dopamine infusion in postoperative renal failure complicating
36. Bersten .4D, Holt AW. Vasoactive drugs and the importance of cardiac surgery. Ann Thorac Surg 2000;b9:501+.
renal perhision pressure. New Horiz 1995.3:650+1 56. Cantarovich F, Galli C. Benedetti L, et al. High dose trusemide in
34. Durham JD, Caputo C, Dokko 1, et al. A randomized controlled established acute renal failure. Br Med J 1973;4:449-50
trial 01 N-acetylcysteine to prevent contrast nephropathy in 59. Minuth AN, Terrell JB, Jr., S u l i WN. Acute renal failure: a study
cardiac angiography. Kidney Int 7002;62.2202-7. of the course and prognosis o i 104 p a t ~ e n t sand of the role of
40 Kay J, Chow WH, C l ~ a nTM, et al. Acetylcyste~netor prevention turosemlde. Am I Med Sci 1976:271:317-24.
of acute deterioration of renal hinction following elective coro- h0. Kleinknecht D. Ganeval D, Gonzalez-Duque LA, Fermanian J .
nary angiography and intenention: a randon~ized controlled Furosemide in acute oliguric renal failure. A controlled trial.
trial. JAMA 2003;289:553-8. Nephron 1976;17:51-8.
41. Alonso A, Lau J, Jaber BL, Weintraub A. Sarnah .MI. Prevention ot 61. Bra\>-n CB. Ogg CS, Cameron JS. High dose frusemide in acute
radiocontrast nephropathy \ v ~ t hk-act~tvlcysteinein patients with renal failure. a controlled trlal. Clin Nephrol 198l;lS:YO+.
chronic kidney disease: a meta-analysis of randomized. 62 Shilliday IR, Quinn KJ, Allison ME. Loop diuretics In the nlan-
controlled trials. Am J Kidney Dis 2004;43:1-9. agement of acute renal failure: a prospecti\?, double-blind.
42. Hoftmann ti. Fischereder M, Kruger B, Drohnik Lj!, Kramer BK. placebo-controlled, randomized study. Nephrol Dial Transplant
The value of X-acetylcysteine in the prelrention of radiocontrast 1997;12:2592+.
agent-induced nephropathy seems questionable. J Am Soc 63. Anderson RJ, Linas SL. Berns AS, et al. Nonoliguric acute renal
Nephrol 2004;15.407-10. failurt,. N Engl J Med 1977;296:113+8.
64. Mehta RL. Pascual MT, Soroko S, Chertow GM. Diuretics, mor-
43. Genet S, Kale RK, Baquer NZ. Ettects of free radicals o n cytosolic
tal~ty,and nonrecover). of renal function in acute renal failure.
creatine kinase activities and protection by antioxidant enzymes
JAMA 200?;288:2547-53
and sulfliydryl compounds. Mol Cell Biochem ?000;210:?3-8.
65. Martin SJ, Danziger LH. Continuous Infusion of loop diurtttlcs in
44. Molnar Z, Shearer E, Lowe D. N-Acetylcystrine treatment to pre-
the critically ill: a review ot the literature. Crit Care Mrd
vent the progression o f multisystem organ iailure: a prospective.
1994;22:132>9.
randomized. placebo-controlled study Crit Care Med 1999:
66. Schetz M. Should lve use diuretics in acute renal failure? Best
?7:1100-4.
Pract Re5 Clin Anaesthesiol 2004:18:73-89.
45. Burton CJ. Tomson CR. Can the use ot loiv-dose dopamine tor
67. Bonventre JV. Weinberg JM. Kidney preservation ex \ i v o for
treatment of acute renal failure he justified? Postgrad Med J
transplantation. Annu Rev Med 1992,43:323-53.
1999:75.269-74.
68. Better OS, Rubinstein 1, Winaver JM, Knochel JP. Mannitol
46. Denton MD. Chertow GM, Bradv HR 'Renal-dose' dopaniine
therapy revisited (1940-1 997). K ~ d n e yInt 1997;52886-94.
tor the treatment of acute renal tallure: scientific rationale,
69. Visweslvaran P. Massin EK, Duhose TD, J r Mannitol-induced
experimental studies aitd clinical trials Kidney Int 1996: acute renal failure. J Am Soc Nephrol 199?:8:1028-33.
50:4-14. 70. Sands JM, Neylan IF, Olson RA, et al. Atrial natr~ureticfactor does
47. Segal JM, Phang PT, Walley KR L o n ' 4 o s e dopamine hastens not improve theoutcome of cadaveric renal transplantation. J A m
onset of gut ischemia in a porcine model of hemorrhagic shock. J Soc Nephrol 1991;1:10814.
Appl Phvsiol 1992;73:1159-6. 71 Ratcliffe PJ. Richardson A], Kirhy JE, et al. Eiiect of intravenous
48. Weisberg LS. Kurnik PB, K u r n i l BR. Risk of radiocontrast infusion of atriopeptin 3 on immediate renal allograft f u n c t ~ o n .
nephropathy in patients with and without diabetes mell~tiis. Kidney Int 1491;39:1@-8.
Kidney Int 1994;45:25Y45. 72. Kurnik BR, Allgren RL, Genter FC, et al Prospective studv of
49. Graziani G, Casati S, Cantaluppi A. Dopamine-furosemide ther- atrial natriuretic peptide for the prevention of radio-
apy In acute renal failure. I'roc EDTA 1962;19:31Q-24. contrast-induced nephropathy. A m J Kidney DIS 1998;31:b74-80.
50 Kellum !A, M Decker J . Use of dopamine in acute renal tailure- a 73. Lewis J, Salem MM, Chertow GM, et al. Atrial natriuretic tactor in
meta-analysis Crit Care Med 21101;29:152&31 oliguric acute renal fallure. Anaritide Acute Renal Failure Study
51. Bellomo R, Chapman M, Finfer S, Hickllng K Myburgh J. Low- Group. An1 J Kidney Dis 2000;36:767-74.
dose dopamlne in patients with early renal dystilnction. a 74. Sward K. Valsson F, Odencrants P, Saniuelsson 0. Ricksten SE.
placebo-controlled randomised trial. Australian and New Recombinant human atrial natriuretic peptide in lschemic acute
Zealand Intensive Care Society (AKZICS) Cllnlcal Trial, Group renal lailure: a randomized placebo-controlled tnal. Crit Care
Lancet 2000;356:213943. Med 2004;32:1310-15.
52. Tumlin JA, Wang A. Murray PT, Mathur VS. Fenoldopam 75 Petrinec D, Reilly JM.Sicard GA, et al. Insulin-like growth tactor-
mesylate blocks reductions in renal plasma flow after radiocon- 1 attenuates delayed gratt tunction In a canine renal autotrans-
trast d y e infusion: a pilot trlal in the prevention of contrast plantation model Surgery 1996:120:221-5;discussion2 2 5 4 .
nephropathy. A m Heart J 2002;141:894-9113. 76. Miller SB. Martin DR, Kissane J . Hammerman MR. Insulin-likr
53. Sheinbaum R. Ignacio C, Safi HJ, Estrera A. Contemporary strate- growth factor I accelerates recovery from ischemic acute tubular
gies to preserve renal h n c t l o n during cardiac and vascular necrosis in the rat. Proc Natl Acad Sci ti S A 1992;8911876-80.
surgerv. Rev Cardiovasc Med 2003;4 Suppl 1:SZl-8.
--
I , . Hladunewich MA, Corngan G, Derhy GC, et al. A randomized,

51.Stone GM'. McCullougl~PA, Tumlul JA, et al. Fenoidopam mesy- placebo-controlled trial of IGF-l tor delayed gratt h n c t ~ o n :a
late tor the pre\rentioil ot contrast-induced nephropathy: a human model to study postischemic ARF. Kidney Int
randonuzed controlled trlal. IAMA 2003:290:?2R4-~1 ?00;:64:593-602.