Secondary Logo

Journal Logo


Perioperative Acute Renal Failure

Sadovnikoff, Nicholas, MD

International Anesthesiology Clinics: January 2001 - Volume 39 - Issue 1 - p 95-109
  • Free


Although significant advances in the perioperative and intraoperative care of patients undergoing major vascular and cardiac surgery have been achieved over the last several decades, the occurrence of perioperative acute renal failure (PARF) continues to plague the postoperative course of these patients. The magnitude of this problem is substantial when it occurs because it is associated with marked increases in morbidity and mortality. Anesthesiologists often use intraoperative strategies with the intention of providing renal protection; however, no substantive scientific evidence has validated any such intervention. Consequently, the commonly used practices aimed at preserving the kidney are based on small studies, anecdotal information, or personal experience. In fact, the incidence of PARF in this population and its impact on outcome have changed little in the last 20 years. The purpose of this article is to summarize the statistics regarding the incidence and consequences of PARF, review what is known about the pathophysiology of PARF, evaluate the techniques currently used to provide renal protection, and scan the horizon for evidence of more effective therapies to prevent this complication in the future.

Incidence and Impact

Identifying the exact incidence of PARF is rendered difficult by lack of consensus of definition. In one systematic review of 28 studies examining perioperative risk factors for PARF, no two studies used the same criteria to define renal failure. 1 Whatever definition is used, however, it is evident that the problem is far from trivial. Several large studies examined the statistical dimensions of this persistent issue, obtaining similar results.

The best studied and most uniform population consists of patients undergoing cardiopulmonary bypass (CPB). Although estimates of PARF in this population vary from 1–15%, even 1% of the roughly 600,000 CPB cases done annually represents 6,000 new cases of renal failure, constituting a serious public health concern. Conlon and colleagues 2 studied 2,843 consecutive adults undergoing CPB, of whom 7.9% experienced renal dysfunction and 0.7% required dialysis. In Mangano and associates 3 prospective observational study of 2,222 adults undergoing CPB, 7.7% of patients experienced renal failure and 1.4% needed dialysis. Chertow and others, 4 using a data base of 43,643 patients who had undergone CPB, found a 1.1% incidence of PARF requiring dialysis. The effect of PARF on mortality in these studies is profound. In Conlon and colleagues's study, mortality in the absence of renal failure was 1% but rose to 14% with renal dysfunction and to 28% if dialysis had to be instituted. Similarly, Mangano and coworkers found a mortality of 0.9% without renal failure, 19% with renal failure, and 63% in patients requiring dialysis. Chertow and associates also reported 63.7% mortality when dialysis was needed, although the mortality without renal failure was higher than in the other studies at 4.3%.

Recent data regarding thoracic aortic surgery are less voluminous but no more reassuring. Godet and others 5 reported on 475 consecutive thoracoabdominal aneurysm (TAA) repairs over a 12-year period. Renal failure occurred in 25% and dialysis was needed in 8%, but the mortality of these groups was not reported. Kashyap and associates 6 studied 183 TAA repairs, and found an 11.5% incidence of renal failure and 2.7% rate of dialysis. The odds ratio for death in the first 30 days postoperatively was 9.2 for the 11.5% of patients in whom renal failure developed compared with those who did not. The data for abdominal aortic surgery are even scantier. Braams and associates 7 reported a 69% mortality in 42 patients who experienced renal failure, requiring dialysis after abdominal aortic aneurysm (AAA) repair. Johnston, 8 in a study of 666 unruptured AAA repairs, reported a 5.4% incidence of PARF, with 0.6% requiring dialysis.

The uniformly gloomy statistics once renal failure is established are not unique to the surgical population. Levy and others 9 published a case-control study in which 183 patients who experienced contrast media–associated renal failure were compared with 174 matched controls. The mortality in the renal failure group was 34% compared with 7% in the control group. The authors concluded that renal failure increases the risk of nonrenal complications that lead to death. These statistics beg for the development of strategies to protect the kidneys from perioperative injury; unfortunately, to date, no such strategy has been shown to be effective.

The Population at Risk

Risk factors for postoperative PARF have been intensively studied. The aforementioned systematic review of 28 studies found that preoperative renal dysfunction was the single consistent predictor of postoperative renal failure. The nature of the surgical procedure also significantly affects the incidence of PARF; the highest incidence occurs after CPB and vascular surgeries involving the aorta. The unifying theme in pathophysiology of PARF is renal ischemia, which is responsible for the bulk of PARF cases. PARF is commonly the consequence of more than one insult, and prerenal azotemia, which reflects renal hypoperfusion, is the most common predisposing factor. Hence, patients already suffering from a chronic ischemic state such as renal artery stenosis, volume depletion, or diabetes or a recent acute ischemic event such as hemorrhage or exposure to radiocontrast agents are more likely to suffer ARF when subsequently exposed to an intraoperative ischemic insult. Unfortunately, the patient population undergoing cardiac or aortic surgery also is characterized by high rates of diabetes, atherosclerotic disease, and recent exposure to radiocontrast dye as well as having a high incidence of preexisting renal insufficiency. Many other conditions may predispose the kidneys to ischemic injury including sepsis, cirrhosis, jaundice, hepatorenal syndrome, congestive heart failure, hemorrhage, shock, malignant hypertension, preeclampsia and toxemia, sickle cell anemia, collagen vascular diseases, and multiple myeloma. Many drugs also may potentiate the risk of ischemic renal injury through alterations in intrarenal hemodynamics, including angiotensin-converting enzyme inhibitors (ACEIs), nonsteroidal antiinflammatory drugs (NSAIDs), cyclosporine, tacrolimus, and amphotericin B. 10

Pathophysiology of Perioperative Acute Renal Failure

To understand best the pathophysiology of PARF, it is worthwhile to review some aspects of renal physiology. The kidneys normally receive 25% of cardiac output, by far the organ with the most luxuriant blood flow per tissue mass. Most of this blood flow is delivered to the cortex to maximize glomerular filtration. Flow to the renal medulla is substantially less generous, a condition necessary to keep from “washing out” the medullary concentration gradient. 11 Yet the most energy-requiring segment of the nephron is the medullary thick ascending loop of Henle (mTAL), which is located in the outer medulla. The tubular cells of this segment are very rich in mitochondria, because they are responsible for the bulk of active chloride ion reabsorption. This combination of low flow and high metabolic demand in this segment renders the outer medulla chronically on the verge of hypoxia. In an elegant series of experiments, Brezis and colleagues 12,13 used sensitive Clark-type oxygen (O2) microelectrodes to measure cortical and medullary oxygen tensions in rat kidneys. Laser Doppler probes were placed in the cortex and medulla as well to measure regional blood flow. The authors demonstrated baseline partial pressure of O2 (PO2) levels of 10–20 mm Hg in the medulla and 40–50 mm Hg in the cortex. They then measured the effect of hypotension on tissue oxygenation by exposing the rats to controlled hemorrhage, aortic ligation, or sodium nitroprusside. They found that, whereas cortical PO2 fell, medullary PO2 levels paradoxically rose. They then demonstrated that this was not due to redistribution of blood flow to favor the medulla, because the laser Doppler probes showed no increase in medullary flow. Further, the phenomenon occurred even when all major vasodilators were blocked, strongly suggesting that the tissue PO2 rose not because of enhanced O2 delivery but because of decreased medullary O2 utilization. The authors hypothesized that medullary O2 utilization decreased as a result of a decrease in glomerular filtration rate (GFR) leading to a decrease in distal tubular reabsorptive work. They additionally suggested that intrarenal mediators might directly inhibit transport in the mTAL in response to decreased renal perfusion. Supporting these data was another series of experiments where the tissue PO2 response to administration of diuretics was measured. When furosemide, which is known to block the active chloride reabsorption in the mTAL, was given, there was no effect on cortical PO2, but medullary PO2 rose much as with hypotension. Similar experiments with the carbonic anhydrase inhibitor acetazolamide revealed an increase in cortical PO2 but no effect on medullary PO2, consistent with this drug's proximal tubular site of action. The broad conclusion drawn from these experiments was that medullary blood flow correlated poorly with tissue oxygenation and that medullary PO2 was largely a function of local O2 demand. It follows, then, that preventing injury to the vulnerable renal medulla should involve not only preservation of oxygen delivery but attenuation of medullary oxygen demand. This is in contrast to the general thrust of traditional renal protective strategies, which emphasized maximizing renal blood flow and maintaining urine output. These experiments demonstrated that medullary PO2 could be improved in the setting of both low urine output (hypotension) and high urine output (furosemide); oliguria might, up to a point, reflect intact endogenous renal protective mechanisms rather than stand as evidence of dysfunction or injury. The common practice of using a desired minimum urine output as an end point for renal protection may thus be misguided.

The balance of oxygen supply and demand in the kidney is complicated by the fact that increasing renal blood flow also increases the volume of glomerular filtrate; this, in turn, may increase the oxygen demand of the kidney as more solute is delivered to the tubules for reabsorption, an energy-requiring process. Consequently, alterations in intrarenal regional blood flow distribution that promote cortical perfusion may significantly impair medullary blood supply. That is, if preglomerular capillary vasomotor tone is relaxed and postglomerular vasomotor tone is increased, the negative effects on the medulla will be twofold. Medullary blood flow will decrease as the result of postglomerular vasoconstriction, but medullary oxygen demand will increase because of the enhanced delivery of filtrate to the tubules. Thus, disturbances in intrarenal regional blood flow, even in the presence of relatively normal total renal blood flow, may lead to inadequate oxygen delivery in the medulla.

Normal intrarenal vascular tone is characterized by a balance of the effects of vasodilatory substances, including nitric oxide (NO), urodilatin, kinins, and prostaglandin (PG) E2, and vasoconstrictors, notably endothelin, catecholamines, and angiotensin II. Adenosine, released in ischemic tissues as adenosine triphosphate (ATP) is used and broken down, has a vasoconstrictor effect in the renal cortex and a vasodilating effect in the medulla. Its net effect in the setting of decreased renal perfusion is thus to attenuate medullary hypoxia both by decreasing glomerular filtration and subsequent reabsorption and by enhancing medullary blood flow (Fig. 1). In the setting of ischemia, decreases in NO production occur, which may have a proinflammatory effect 14 and result in vasoconstriction. 15 The renal injury in myoglobinuria seen with rhabdomyolysis may be mediated, at least in part, by the NO scavenging effect of free myoglobin with resultant vasoconstriction. Normal vascular balance may be disrupted by exogenous substances such as NSAIDs, which decrease PGE2 production with resultant loss of its medullary vasodilating effect. 16 Radiocontrast nephropathy has been shown to be an ischemic renal injury, caused by a combination of intense medullary vasoconstriction and increased delivery of solute to the tubules as a result of osmotic diuresis. 17 Experiments in animals support a role for disturbed regulation of multiple vasoactive substances, including prostaglandins, angiotensin II, NO, endothelin, and adenosine, but to date human data are lacking to define a causative mechanism. 17

Figure 1.
Figure 1.:
Schematic representation of the effects of adenosine on renal circulation. Note that the effect on A1 receptors is more prominent than on A2 receptors. O2 = oxygen; A1 = adenosine-1 receptor; A2 = adenosine-2 receptor; RBF = renal blood flow; GFR = glomerular filtration rate; Na+ = sodium ion; VO2 = oxygen consumption.

Ischemic renal injury affects the tubules at the level of the outer medulla, which include primarily the mTAL and the S3 portion of the proximal convoluted tubule. 11 Tubular cell death is now understood to be characterized both by necrosis and by apoptosis, or programmed cell death. 18 Necrosis results from profound cellular ATP depletion and is characterized by a predictable sequence of events. These begin with loss of cell polarity and of the epithelial brush border, followed by the loss of the integrity of tight junctions and the appearance of integrins such as intercellular adhesion molecule-1 (ICAM-1) on the cell surface. 19 These interact with leukocyte adhesion molecules to mediate an inflammatory response with release of cytotoxic mediators. 20 Cells then slough into the tubular lumen, further impairing already compromised filtrate flow and leading ultimately to complete tubular obstruction. This, in turn, causes an increase in tubular pressure proximal to the obstruction, with leakage of filtrate into the interstitium and further proximal tubular injury resulting. 10 Tubular cell apoptosis is also triggered by ischemia through as yet uncharacterized mechanisms, also resulting in cell loss, but there is no inflammatory component, and the resultant apoptotic bodies are phagocytosed by macrophages or surviving epithelial cells. 18 Apoptosis also is observed in the recovery phase during epithelial proliferation, where it likely serves a role in restoring a normal tubular structure. Clinically, the initial observation is a loss of urinary concentrating ability as the medullary gradient dissipates, followed by a decline in urine output as tubules become obstructed and denuded.

Current Practices for Renal Protection

Current strategies to provide renal protection are largely based on tradition, anecdotal information, extrapolation from animal models or other clinical scenarios, and animal data. No large randomized trials have assessed their effectiveness at preventing renal injury in the surgical patient. The most commonly used techniques are intravenous hydration, mannitol, “renal dose” dopamine, and loop diuretics, often in combination. The rationale and evidence for each of these are considered individually.

Intravenous Hydration

The most extensive literature in renal protection comes from the study of radiocontrast nephropathy. Although not exactly the same as a surgical insult, radiocontrast dye induces severe changes in intrarenal hemodynamics, which can lead to an ischemic injury comparable to that induced by CPB or aortic surgery. One of the only conclusions of these studies is that precontrast hydration is effective at lowering the rate of renal injury. The influential study of Solomon and colleagues 21 compared hydration alone with hydration plus mannitol and hydration plus furosemide in high-risk patients (creatinine > 1.6 or calculated creatinine clearance <60 mL/min) undergoing coronary angiography. Renal dysfunction occurred in 11% of the hydration-only group, 28% of the hydration plus mannitol group, and 40% of the hydration plus furosemide group. The hydration strategy consisted of 0.45% saline administered at 1 mL/kg/hr for 12 hours before the procedure. The study was conducted on a particularly small sample size (78 patients) and failed to report whether patients were taking ACEIs or NSAIDs, rendering the conclusions statistically suspect. Further, in the era of same-day surgery, 12 hours of preoperative intravenous hydration is not a logistically practical intervention and is unlikely to be used for elective surgeries. Nevertheless, volume depletion is a well-defined risk factor for PARF, and it is intuitively desirable to ensure adequate hydration before exposing the patient to the potential ischemic insult. The use of invasive monitoring to achieve this goal is widely practiced, but there is no consensus as to the superiority of pulmonary artery (PA) catheterization over central venous monitoring. Two small uncontrolled studies in the 1980s of patients undergoing abdominal aneurysm repair concluded that, when fluid management was guided by a PA catheter, renal impairment was prevented. 22,23 Subsequent studies in selected low-risk populations, however, failed to support this finding. 24,25 A study of patients undergoing abdominal aneurysm repair in whom PA catheters were placed concluded that data obtained from the catheter were infrequently used to guide management. 26 In 1997, the Pulmonary Artery Catheter Consensus Conference published its evidence-based recommendations regarding the use of PA catheterization perioperatively. 27 Given the paucity of data supporting routine PA catheter use in patients undergoing infrarenal aortic surgery, the authors concluded that patients with good left ventricular function undergoing infrarenal aortic surgery could be safely managed with central venous pressure monitoring only. On the basis of level IV (nonrandomized studies, historical controls, expert opinion) and level V (case series, uncontrolled studies, expert opinion) evidence, for patients with a left ventricular ejection fraction less than 50%, with “significant” coronary artery disease, or undergoing suprarenal cross-clamping, pulmonary catheterization was recommended. The same panel reviewed the evidence regarding PA catheter use in patients undergoing coronary artery bypass procedures and peripheral vascular surgery. In both of these populations, evidence was considered insufficient to support recommending routine use of pulmonary artery catheters.


Long a contentious subject, the role of “renal dose” dopamine for renal protection is becoming progressively questionable. The absence of convincing data showing a beneficial effect on renal outcome perioperatively has in the past not stopped its enthusiastic use, because urine output often is noted to increase with its administration. As already alluded to, however, urine production may not correlate with medullary oxygenation. Dopamine, by stimulation of DA1 receptors, causes an increase in renal cortical blood flow, leading to increased glomerular filtration, solute excretion, and urine output. 28 DA2 receptor activation decreases intrarenal norepinephrine excretion, further enhancing vasodilation and glomerular flow. These effects likely increase renal medullary oxygen requirements. On the other hand, blockade of DA1 and DA2 receptors does not affect the diuretic and natriuretic effect of dopamine 29; this implies a separate effect of dopamine, postulated to be an inhibition of the sodium-potassium ATPase in the proximal tubule and mTAL. 30 Such an inhibitory effect would decrease tubular energy requirements and medullary oxygen requirements. Thus, the net effect of dopamine on renal medullary oxygenation is unclear. Dopamine may also have undesirable cardiac effects such as tachycardia and dysrhythmias, and it may inhibit minute ventilation and adversely affect oxygen saturation. Perdue and colleagues, 31 in their extensive review of the literature regarding the use of dopamine in the perioperative period, concluded that the studies to date fail to show a benefit in preventing renal failure. In their 2000 report, Lassnigg and coworkers 32 described 126 adults with normal preoperative renal function who were treated during and after cardiac surgery with infusions of dopamine, furosemide, or placebo. They found no renal protective benefit from dopamine and a deleterious effect with furosemide. The study is hampered by small sample size and measures only serum and urine creatinine to assess for renal injury. Furthermore, the fairly long administration time (48 hours) of the study drugs extends well past the acute ischemic insult, deviating from the conventional use of these drugs in this setting. Tang and colleagues 33 randomized 40 cardiac surgical patients to receive 48-hour dopamine or placebo infusions, starting with induction of anesthesia. Using retinol-binding protein as a measure of tubular injury, they found marked elevations in the group receiving dopamine infusion compared with the placebo group. They concluded that subclinical renal tubular injury was exacerbated by dopamine infusion. Perhaps the most compelling evidence discrediting the concept of renal dose dopamine was the MacGregor and associates' 34 finding that identical weight-based dopamine infusions in nine healthy volunteers resulted in 10-to 75-fold variability in plasma dopamine levels. The notion that by dosing according to weight one can selectively stimulate dopaminergic receptors is essentially rendered untenable by these findings, which do not even account for the likelihood of interpatient variability in receptor populations as well.


Mannitol has traditionally been used in patients considered at high risk for ARF. Mannitol is believed to provide benefit by increasing tubular flow by augmenting filtrate volume through its osmotic diuretic effect. Additionally, once present in the filtrate, it diminishes tubular endothelial cell swelling by dehydration, increasing tubular diameter and decreasing resistance to flow. The enhanced urine flow is thought to help prevent tubular obstruction and the resultant further renal injury from occurring. The identification of mannitol as a hydroxyl radical “scavenger” makes it potentially attractive in the setting of ischemia-reperfusion injury. Initial animal studies formed the basis for the use of mannitol as a protective strategy, 35 and studies demonstrating protection of renal transplants support a protective role. 36,37 A small randomized, prospective study comparing mannitol with saline in patients undergoing infrarenal aortic aneurysm repair found significantly less glomerular and tubular injury, as measured by urinary albumin and N-acetylglucosaminidase levels respectively, in the mannitol group. 38 Although no large prospective, controlled studies in aortic or cardiac surgical patients have convincingly demonstrated these benefits, it is commonly administered in a single dose (0.5–1.0 g/kg) intraoperatively before the application of the cross-clamp in aortic surgery patients or included in the CPB prime in cardiac surgery. Appropriate attention is paid to monitoring of volume status and to maintenance of electrolyte balance, particularly potassium.


The rationale for use of furosemide as a renal protectant is clear from the prior discussion regarding the amelioration medullary hypoxia by inhibiting energy-dependent reabsorptive work in the mTAL. Although the experimental results in animal models are encouraging, they have yet to be translated into compelling evidence in humans. As mentioned, one problem may lie in how furosemide is used: if prolonged infusions are given that extend well past the surgical period, protective effects may be masked by the cumulative consequences of prolonged infusion. Furosemide has vasodilatory properties and causes cortical vasodilation. This may result in medullary hypoperfusion, which, over a prolonged period of time, may be more deleterious than protective. It seems prudent to use furosemide in a single large-bolus dose shortly before the anticipated ischemic stress, because the duration of action should be sufficient to carry through the CPB or cross-clamp time. Randomized studies evaluating furosemide as a sole renal protectant in surgical patients are lacking, but a number of other studies evaluated various pharmacological interventions in an attempt to prevent radiocontrast nephropathy. This includes the aforementioned study of Solomon and colleagues, in which hydration alone was found to be superior to hydration plus dopamine and hydration plus furosemide. Stevens and associates 39 studied 98 high-risk patients (creatinine >1.8) undergoing coronary angiography. Of these, 43 received a cocktail of crystalloid, furosemide bolus, dopamine infusion, and, if the pulmonary artery occlusion pressure was <20 on right heart catheterization, mannitol. The remaining 55 patients received only crystalloid, starting with the arrival in the angiography suite. There was no difference between the two groups in the mean rise in creatinine. Two control patients and five experimental group patients required dialysis. In a similar study, Weisberg and others 40 randomized 50 patients with chronic renal insufficiency, of which 24 were diabetic, into four groups before angiography; patients received saline placebo, dopamine, mannitol, or atrial natriuretic peptide. Measuring serum creatinine, they found that renal protection was afforded in the nondiabetic patients but not in the diabetic population, in which the agents appeared to worsen nephropathy. The results must be interpreted with caution because the sample size was exceedingly small.

Experimental Agents for Renal Protection

Other substances not routinely used but available have been tested as renal protectants in radiocontrast nephropathy, such as atrial natriuretic peptide and PGE1, known renal vasodilators. Atrial natriuretic peptide and the related endogenous compound urodilatin have renal vasodilating and natriuretic properties and could theoretically provide renal protective effects. An initial flurry of encouraging reports in patients after cardiac surgery created some enthusiasm for this agent. However, Kurnik and coworkers 41 found no difference comparing atrial natriuretic peptide and placebo in a randomized study of 247 patients with chronic renal insufficiency undergoing radiocontrast studies. No studies have been done in patients undergoing high-risk surgeries. Further large investigations are warranted before adoption of this agent as a perioperative strategy.

PGE1, an endogenous renal vasodilator, also is an attractive candidate for maintaining intrarenal hemodynamics in the face of an ischemic insult. Koch and others 42 compared PGE1 in three different doses with placebo in 130 patients with chronic renal insufficiency undergoing radiocontrast studies. Whereas serum creatinine rose more in the placebo group, measured creatinine clearances did not differ. Again, more compelling data are required to consider the routine use of this agent.

Another agent considered promising as a renal protectant is dopexamine, a dopaminergic (DA1 and DA2) agonist with β2-adrenergic activity, that has generated interest based on the hypothesis that it could improve renal perfusion without risk of α-adrenergic stimulation. As discussed, the simple augmenting of renal perfusion may conceivably have the effect of increasing medullary oxygen demand by increasing active chloride and sodium reabsorption in the mTAL. Unlike dopamine, dopexamine appears not to inhibit sodium reabsorption 43; it may, therefore, have the net effect of increasing medullary oxygen demand. Although one initial small study in aortic surgery patients provided encouraging results, 44 a more recent study demonstrated no effect on renal vascular resistance. 45 Although dopexamine is a potent antihypertensive agent and may cause hypotension, it lacks the myocardial side effects of dopamine and remains a potentially attractive agent. Larger well-designed trials are needed to assess its usefulness as a renal protectant.

Fenoldopam mesylate is the only commercially available selective DA1 receptor agonist, having been approved for the treatment of hypertension. Its complete lack of activity at DA2, α-adrenergic, or β-adrenergic receptors makes it attractive as an alternative to dopamine. It has been shown to reduce blood pressure in a dose-dependent manner while preserving renal perfusion and GFR. 46 Its inhibitory effects on sodium reabsorption in normotensive individuals are transient and modest 46 but may thus at least temporarily attenuate medullary oxygen demand while enhancing supply. It may have a more pronounced natriuretic effect in hypertensives. 47 One study of its use in a canine aortic cross-clamp model has generated some potentially interesting preliminary data, but large prospective trials evaluating fenoldopam as a renal protectant in humans have yet to be published.

Preliminary work is ongoing to evaluate insulin-like growth factor-1 (IGF-1) as a renal protectant. This endogenous substance has been demonstrated to improve recovery from renal failure in animal studies, but currently no definitive evidence supports its use on a prophylactic basis. Acidic fibroblast growth factor (FGF-1), which has been shown to attenuate myocardial ischemia-reperfusion injury, has also been shown to attenuate renal failure after experimental ischemia-reperfusion injury of the kidney in rats. 48 In this model, an NO synthase inhibitor blocked the protective effect of FGF-1, suggesting that the protective effect of FGF-1 may be mediated by the antiinflammatory and vasodilating effects of NO. 48

As noted, renal ischemia may be caused or aggravated by a vasoconstrictive effect on the renal microvasculature, with endothelin being a potent intrarenal vasoconstrictor. A study using a canine model of aortic surgery found a substantial renoprotective effect in dogs pretreated with an endothelin antagonist. 49 If such an agent can be found to be minimally toxic in humans, it would offer an attractive approach; however, the problem of systemic hypotension might prove limiting.

Frontiers in Renal Protection

Although some of the foregoing strategies may eventually be shown to have some efficacy in renal protection, it is my opinion that a recent direction of investigation will eventually bear fruit. That direction involves the related areas of ischemic preconditioning and adenosine receptors.

Perhaps the most exciting recent discovery, and one that might readily be brought into human experiments, is the observation that renal ischemic preconditioning has been shown by several investigators to protect the kidney from ischemic injury in rats. 50,51 In these experiments, short cycles of renal ischemia and reperfusion, administered before a longer ischemic period, provided substantial protection from the renal injury seen in animals given no preconditioning. Yet more intriguing is a report demonstrating that the effect of renal preconditioning can be mimicked by intravenous infusion of adenosine. In a well-conceived series of experiments, Lee and Emala 52 exposed rats to either ischemic preconditioning or adenosine infusion before a standardized ischemia-reperfusion period. Using specific agonists and antagonists for the A1, A2, and A3 adenosine receptors, they were able to demonstrate that the renal protective effect is mediated by the A1 adenosine receptor. Interestingly, the administration of a specific A1 receptor antagonist did not completely block the renal protection induced by ischemic preconditioning. They also found that, although A2 receptors were not involved in renal protection, activation of A3 receptors potentiated renal injury and blockade of those receptors had renal protective effects.

The beneficial effect of adenosine may be modulated through activation of potassium channels, as has been shown in myocytes; alternatively or additionally, it may be an effect of preglomerular vasoconstriction with resultant decrease in glomerular filtration, solute delivery to the tubules, sodium reabsorption, and medullary oxygen consumption.

These findings are of profound interest for several reasons. First, adenosine is an endogenous substance with which there is already extensive clinical experience. Second, it is a currently available agent that has passed all regulatory hurdles. Finally, it is an agent with an exceptionally short half-life whose use in hemodynamically unstable patients could be readily titrated. This makes it an extremely attractive candidate for use in clinical trials. The development of a safe A3 receptor blocking agent would also be a potentially valuable step in improving our renal protection armamentarium. This research invites the performance of comparable human studies of both preconditioning maneuvers and adenosine infusions in anticipation of a period of renal ischemia.


ARF after cardiac and aortic surgery continues to constitute a vexing problem. Once established, the adverse implications for morbidity and mortality are profound. The challenge for surgeons and anesthesiologists thus lies in developing techniques to avoid this complication. Unfortunately, little progress in its prevention has been made over the last three decades, and current renal protection strategies are unsatisfactory. Discoveries of the beneficial effects of ischemic preconditioning and the role of adenosine receptors provide therapeutic promise, however, and clinical studies are strongly indicated to assess the applicability of these findings to at-risk patient populations.


1. Novis BK, Roizen MF, Aronson S, et al. Association of preoperative risk factors with postoperative acute renal failure. Anesth Analg 1994; 78:143–149
2. Conlon PJ, Stafford-Smith M, White WD, et al. Acute renal failure following cardiac surgery. Nephrol Dial Transplant 1999; 14:1158–1162
3. Mangano CM, Diamonstone LS, Ramsay JG, et al. Renal dysfunction after myocardial revascularization: risk factors, adverse outcomes and hospital utilization. Ann Intern Med 1998; 128:194–203
4. Chertow GM, Lazarus JM, Christiansen CL, et al. Preoperative renal risk stratification. Circulation 1997; 95:878–884
5. Godet G, Fleron MH, Vicaut E, et al. Risk factors for acute postoperative renal failure in thoracic or thoracoabdominal aortic surgery: a prospective study. Anesth Analg 1997; 85:1227–1232
6. Kashyap VS, Cambria RP, Davison JK, et al. Renal failure after thoracoabdominal aortic surgery. J Vasc Surg 1997; 26:949–955
7. Braams R, Vossen V, Lisman BA, et al. Outcome in patients requiring renal replacement therapy after surgery for ruptured and non-ruptured aneurysm of the abdominal aorta. Eur J Endovasc Surg 1999; 18:323–327
8. Johnston KW. Multicenter prospective study of nonruptures abdominal aortic aneurysm: part II. Variables predicting morbidity and mortality. J Vasc Surg 1989; 9:437–447
9. Levy EM, Viscoli CM, Horwitz RI. The effect of acute renal failure on mortality: a cohort analysis. JAMA 1996; 275:1489–1494
10. Thadvani R, Pascual M, Bonventre JV. Acute renal failure. N Engl J Med 1996; 334:1449–1460
11. Brezis M, Rosen S. Hypoxia of the renal medulla: its implications for disease. N Engl J Med 1995; 332:647–655
12. Brezis M, Agmon Y, Epstein FH. Determinants of intrarenal oxygenation: 1. Effects of diuretics. Am J Physiol 1994; 267:F1059–F1062
13. Brezis M, Heyman SN, Epstein FH. Determinants of intrarenal oxygenation: 2. Hemodynamic effects. Am J Physiol 1994; 267:F1063–F1068
14. Lauriat S, Linas SL. The role of neutrophils in acute renal failure. Semin Nephrol 1998; 18:498–504
15. Yu L. Role of nitric oxide in acute renal failure. Ren Fail 1997; 19:213–216
16. Conger J. Hemodynamic factors in acute renal failure. Adv Ren Replace Ther 1997; 4(Suppl 1):25–37
17. Solomon R. Radiocontrast-induced nephropathy. Semin Nephrol 1998; 18:551–557
18. Lieberthal W, Koh JS, Levine JS. Necrosis and apoptosis in acute renal failure. Semin Nephrol 1998; 18:505–518
19. Rabb H, Postler G. Leukocyte adhesion molecules in ischaemic renal injury: kidney specific paradigms? Clin Exp Pharmacol Physiol 1998; 25:286–291
20. Molitoris BA, Marrs J. The role of cell adhesion molecules in ischemic acute renal failure. Am J Med 1999; 106:583–592
21. Solomon R, Werner C, Mann D, et al. Effects of saline, mannitol, and furosemide to prevent acute decreases in renal function induced by radiocontrast agents. N Engl J Med 1994; 331:1416–1420
22. Bush Jr, HL Huse JB, Johnson WC, et al. Prevention of renal insufficiency after abdominal aortic aneurysm resection by optimal volume loading. Arch Surg 1981; 116:1517–1524
23. Hesdorffer CS, Milne JF, Myers AM, et al. The value of Swan-Ganz catheterization and volume loading in preventing renal failure in patients undergoing abdominal aneurysmectomy. Clin Nephrol 1987; 28:272–276
24. Isaacson IJ, Lowdon JD, Berry AJ, et al. The value of pulmonary artery and central venous monitoring in patients undergoing abdominal aortic reconstructive surgery: a comparative study of two selected, randomized groups. J Vasc Surg 1990; 12:754–760
25. Joyce WP, Provan JL, Ameli FM, et al. The role of central hemodynamic monitoring in abdominal aortic surgery: a prospective randomized study. Eur J Vasc Surg 1990; 4:633–636
26. Mostafa G, Kumar M, Schlotthauer J, et al. The utility of hemodynamic measurements acquired by pulmonary artery catheterization. Am J Surg 1998; 175:293–296
27. Leibowitz AB, Belin Y. Pulmonary artery catheters and outcome in the perioperative period. New Horiz 1997; 5:214–221
28. Schwartz LB, Bissell MG, Murphy M, et al. Renal effects of dopamine in vascular surgical patients. J Vasc Surg 1988; 8:367–374
29. Luippold G, Osswald H, Muhlbauer B. Renal effects of exogenous dopamine: modulation by renal nerves and dopamine receptor antagonists. Naunyn Schmiedebergs Arch Pharmacol 1998; 358:445–451
30. Olsen NV. Effects of dopamine on renal haemodynamics, tubular function and sodium excretion in normal humans. Dan Med Bull 1998; 45:282–297
31. Perdue PW, Balser JR, Lipsett PA, et al. “Renal dose” dopamine in surgical patients: dogma or science? Ann Surg 1998; 227:470–473
32. Lassnigg A, Donner E, Grubhofer G, et al. Lack of renoprotective effects of dopamine and furosemide during cardiac surgery. J Am Soc Nephrol 2000; 11:97–104
33. Tang AT, El-Gamel A, Keevil B, et al. The effect of “renal-dose” dopamine on renal tubular function following cardiac surgery: assessed by measuring retinal binding protein (RBP). Eur J Cardiothorac Surg 1999; 15:717–721
34. MacGregor DA, Smith TE, Prielipp RC, et al. Pharmacokinetics of dopamine in healthy male subjects. Anesthesiology 2000; 92:338–346
35. Hanley MJ, Davidson K. Prior mannitol and furosemide infusion in a model of ischemic acute renal failure. Am J Physiol 1981; 241:F556–F564
36. Tiggler RG, Berden JH, Hoitsma AJ, et al. Prevention of acute tubular necrosis in cadaveric kidney transplantation by the combined use of mannitol and moderate hydration. Ann Surg 1985; 201:246–251
37. Weimar W, Geerlings W, Bijnen AB, et al. A controlled study on the effect of mannitol on immediate renal function after cadaver donor kidney transplantation. Transplantation 1983; 35:99–101
38. Nicholson ML, Baker DM, Hopkinson BR, et al. Randomized controlled trial of the effect of mannitol on renal reperfusion injury during aortic aneurysm surgery. Br J Surg 1996; 83:1230–1233
39. Stevens MA, McCullough PA, Tobin KJ, et al. A prospective randomized trial of prevention measures in patients at high risk for contrast nephropathy. J Am Coll Cardiol 1999; 33:403–411
40. Weisberg LS, Kurnik PB, Kurnik BR. Risk of radiocontrast nephropathy in patients with and without diabetes mellitus. Kidney Int 1994; 45:259–265
41. Kurnik BR, Allgren Rl, Genter FC, et al. Prospective study of atrial natriuretic peptide for the prevention of radiocontrast-induced nephropathy. Am J Kidney Dis 1998; 31:674–680
42. Koch J, Plum J, Grabensee B, et al. Prostaglandin E1: a new agent for the prevention of renal dysfunction in high risk patients caused by radiocontrast media? Nephrol Dial Transplant 2000; 15:43–49
43. Olsen NV, Lund J, Jensen PF, et al. Dopamine, dobutamine, and dopexamine: a comparison of renal effects in unanesthetized human volunteers. Anesthesiology 1993; 79:685–694
44. Welch M, Newstead CG, Smith JV, et al. Evaluation of dopexamine hydrochloride as a renoprotective agent during aortic surgery. Ann Vasc Surg 1995; 9:488–492
45. Sherry E, Tooley MA, Bolsin SN, et al. Effect of dopexamine hydrochloride on renal vascular resistance index and haemodynamic responses following coronary artery bypass graft surgery. Eur J Anaesthesiol 1997; 14:184–189
46. Mathur VS, Swan SK, Lambrecht LJ, et al. The effects of fenoldopam, a selective dopamine agonist on systemic and renal hemodynamics in normotensive subjects. Crit Care Med 1999; 27:1832–1837
47. O'Connell DP, Ragsdale NV, Boyd DG, et al. Differential human renal tubular responses to dopamine type 1 receptor stimulation are determined by blood pressure status. Hypertension 1997; 29:115–122
48. Cuevas P, Martinez-Coso V, Fu X, et al. Fibroblast growth factor protects the kidney against ischemia-reperfusion injury. Eur J Med Res 1999; 4:403–410
49. Krause SM, Walsh TF, Greenlee WJ, et al. Renal protection by a dual ETA/ETB endothelin antagonist L-754,142, after aortic cross-clamping in the dog. J Am Soc Nephrol 1997; 8:1061–1071
50. Toosy N, McMorris EL, Grace PA, et al. Ischaemic preconditioning protects the rat kidney from reperfusion injury. BJU Int 1999; 84:489–494
51. Cochrane J, Williams BT, Banerjee A, et al. Ischemic preconditioning attenuates functional, metabolic, and morphologic injury from ischemic renal failure in the rat. Ren Fail 1999; 21:135–145
52. Lee HT, Emala CW. Protective effects of renal ischemic preconditioning and adenosine pretreatment: role of A1 and A3 receptors. Am J Physiol Renal Physiol 2000; 278:F380–387
Copyright © 2001 Wolters Kluwer Health, Inc. All rights reserved.