Sepsis

Sepsis is the most common cause of new-onset acute renal failure in the postoperative period (see Chapter 74 ). Renal function may deteriorate progressively without defined episodes of hypotension. In addition, sepsis predisposes the kidney to further ischemic and nephrotoxic insults, such as concomitant aminoglycoside use.^[147] Acute renal failure itself and hemodialysis perpetuate sepsis by activating leukocytes. Sepsis may induce renal damage by hypotension and vasomotor nephropathy and through the direct and indirect effects of endotoxin.

A reasonable amount of evidence suggests that renal autoregulation may be impaired in sepsis and that RBF and GFR decline pari passu with systemic vascular resistance and mean arterial pressure. Hypotension in turn sets off a cascade of neurohormonal responses (sympathoadrenal activity; renin-angiotensin, vasopressin, and thromboxane activation) that result in decreased RBF, GFR, sodium excretion, and urine flow. The severity of renal dysfunction appears to be directly related to the severity of sepsis and the degree of plasma renin activation.^[148]

Renal dysfunction in sepsis is characterized as a vasomotor nephropathy, which implies renal vasoconstriction in the face of an increased total cardiac index. Renal vasoconstriction, mesangial cell contraction, a decreased ultrafiltration coefficient, and decreased GFR are induced by endotoxin and by compounds activated in sepsis.^[149] These compounds include endothelin and eicosanoids such as thromboxane, PGF₂ , and the leukotrienes C₄ and D₄ . PGF₂ , which mimics the action of thromboxane, is formed during leukostasis when arachidonic acid is oxidized by free oxygen radicals.

Endotoxin causes leukocyte arachidonic acid to undergo lipoxygenation to form leukotrienes; it also impairs their biliary elimination. In addition, experimental infusion of lipopolysaccharide (endotoxin) directly decreases RBF, GFR, and tubular concentrating ability and increases urinary loss of tubular enzymes. It causes sequestration of leukocytes in peritubular capillaries, induces endothelial lesions by releasing neutrophil-derived elastase (enhanced by reperfusion injury), and potentiates renal ischemia such that brief hemodynamic instability causes rapid loss of renal function. It also potentiates nephrotoxicity. The changes caused by endotoxin are mimicked by the cytokine tumor necrosis factor-α.

It has been estimated that nephrotoxic renal insufficiency will develop in about 10% to 26% of septic patients receiving aminoglycoside antibiotics. ^[130] Aminoglycoside nephrotoxicity is enhanced by the interaction of fever, renal vasoconstriction, hypovolemia, and endotoxin. In the presence of these and other risk factors (see earlier), the use of alternative non-nephrotoxic antibiotics should be considered to cover gram-negative infections, including penicillins (ticarcillin), cephalosporins (ceftazidime), carbapenems (imipenem), or monobactams (aztreonam).

Renal Protection in Septic Shock (see Chapter 74 )

Anti-inflammatory Drugs

Cumming and colleagues^[150] demonstrated a marked renal protective effect of a selective thromboxane synthetase inhibitor (U63557A) on laparotomy in a volume-loaded sheep peritonitis model. Administration of the selective inhibitor either before or 30 minutes after surgery prevented deterioration in creatinine clearance, urinary sodium excretion, and the urine flow rate. Beneficial effects have also been observed with aprotinin, perhaps through its anti-inflammatory action.^[44] In contrast, nonselective COX inhibition by NSAIDs worsens renal function in sepsis by decreasing synthesis of the renal vasodilator prostacyclin.^[151]

806

The use of pharmacologic doses of methylprednisolone in septic shock was discredited by two large multicenter studies showing no beneficial effect on outcome.^{[152] [153]} Moreover, patients who received steroids had significant increases in BUN but not in serum creatinine, thus suggesting a prerenal state induced by increased protein catabolism.^[154] Other potentially adverse effects of high-dose steroids include impaired mitochondrial function, impaired leukocyte function, and inhibition of phospholipase A₂ resulting in decreased synthesis of intrarenal vasodilator prostaglandins.

Supranormal Oxygen Delivery

Over the last decade there has been considerable controversy in the concept of supranormal oxygen delivery to tissues to overcome the defect that exists in oxygen utilization by septic tissues.^[155] This approach consists of inotropic support and blood transfusion to drive global oxygen delivery (DO₂ ) to one of three end points: to a DO₂ level consistently found in survivors (600 mL/min/m² ), when global oxygen consumption (VO₂ ) no longer increases with increasing DO₂ (consumption independence), or when blood lactate levels start to decline. The benefits of supranormal oxygen delivery on outcome have been disputed, and high-dose inotropic and vasopressor support may themselves have adverse consequences. Moreover, renal DO₂ and VO₂ differ markedly from systemic indices. Renal VO₂ is largely determined by tubular metabolic function, which is regulated by fluid and electrolyte changes. In a volume-loaded septic porcine model, inotropic support with dobutamine increased systemic DO₂ and VO₂ but did not increase renal DO₂ and VO₂ .^[25] Furthermore, decreased global renal DO₂ does not appear to cause tubular damage, possibly because tubular work and VO₂ are decreased when the GFR declines.^[156]

Dopaminergic Drugs

Low-dose dopamine (1 to 3 µg/kg/min) is frequently administered to septic patients in the belief that it confers renal protection through renal vasodilation or perhaps by inhibition of the Na⁺ -K⁺ -ATPase pump and decreases in renal tubular VO₂ .

It is also administered in combination with more potent pressors (dobutamine, epinephrine, and norepinephrine) in sepsis in the hope of enhancing hepatic, renal, and mesenteric perfusion. Some animal data support this practice. In a nonseptic dog study in which RBF was measured by thermodilution, the addition of low-dose dopamine to a norepinephrine infusion of 0.2 to 1.6 µg/kg/min increased RBF by 40% to 50%.^[24] However, a subsequent study demonstrated that although low-dose dopamine increased RBF when added to an epinephrine infusion in healthy sheep, this benefit could not be found in an intraperitoneal sepsis model.^[157] Low-dose dopamine can increase hepatic DO₂ , but possibly at the expense of splanchnic oxygenation.

In patients with sepsis syndrome (signs of sepsis without hypotension), low-dose dopamine infusion doubled the urine flow rate and increased creatinine clearance by 60% without any change in systemic hemodynamics.^[158] However, the renal response to dopamine decreased significantly after 48 hours of dopamine infusion, possibly as a result of downregulation of renal dopaminergic receptors or diuresis-induced contraction of intravascular volume. In patients with established septic shock who required catecholamines for blood pressure support, low-dose dopamine did not alter systemic hemodynamics or renal function.

The prophylactic administration of low-dose dopamine in sepsis appears to have been laid to rest by a large prospective controlled study conducted by the Australian and New Zealand Intensive Care Society. They randomized 328 patients with signs of systemic inflammatory response syndrome and early renal dysfunction (oliguria or increasing serum creatinine) to dopamine, 2 µg/kg/min, or placebo. They found no differences in serum creatinine, dialysis requirement, intensive care unit or hospital length of stay, or overall mortality.^[159]

The potential role of dopexamine in septic shock remains speculative. Most studies have examined its role in splanchnic and hepatic perfusion rather than in renal protection. In animal models of sepsis, dopexamine has improved splanchnic and hepatic DO₂ , but its β₂ -adrenergic activity in causing tachycardia and hypotension may limit its application in clinical sepsis. Smithies and associates^[160] administered dopexamine to patients with sepsis syndrome, acute respiratory failure, and at least one other organ system failure. The cardiac index increased and gastric intramucosal pH (an index of splanchnic perfusion) improved significantly.

Norepinephrine

In patients with septic shock, profound hypotension, and oliguria, vasopressor therapy with norepinephrine may actually improve renal function by enhancing renal perfusion pressure. Desjars and colleagues^[7] evaluated a group of septic patients who remained oliguric despite volume resuscitation and the use of dopamine in doses of up 15 µg/kg/min. The addition of norepinephrine and reduction of dopamine to a low-dose level resulted in an improvement in mean arterial pressure from 50 to 70 mm Hg, a tripling of urine flow, and a doubling of creatinine clearance. Norepinephrine increased systemic vascular resistance with little change in the cardiac index or DO₂ . Subsequent studies have confirmed that the use of norepinephrine to keep mean arterial pressure greater than 60 mm Hg results in improved cardiac function (increase in stroke volume and decrease in heart rate) and GFR without deleterious effects on the cardiac index, oxygen extraction, or VO₂ . ^[161]

Large doses of norepinephrine may be required to achieve these goals because in septic shock, the peripheral vasculature is notoriously refractory to norepinephrine-induced vasoconstriction as a result of massive inducible nitric oxide release as well as vasopressin deficiency (see the next section). Nonetheless, these findings strongly support the concept that renal autoregulation is impaired in sepsis and that maintenance of adequate renal perfusion pressure is an important component of renal protection.

Arginine Vasopressin

Patients in vasodilatory shock have inappropriately low plasma levels of AVP and marked vascular sensitivity

807

Landry and coworkers^[164] observed unusual sensitivity to the vasoconstrictor effects of infused AVP in patients with septic shock and profound hypotension despite catecholamine infusion. Infusion of AVP at doses of 2.4 U/hr, less than one tenth of those used in the treatment of bleeding esophageal varices, resulted in a dramatic increase in systolic blood pressure from 92 ± 4 to 146 ± 4 mm Hg (mean ± SEM, P < .001), and catecholamine infusions were able to be discontinued. In an associated report,^[165] the authors observed that urine flow increased concomitantly in three of five patients, from an average of 30 to 110 mL/hr.

Plasma AVP levels were remarkably low (3.1 ± 1.0 pg/mL) and significantly lower than in a cohort of patients in cardiogenic shock who were also receiving catecholamines (22.7 ± 2.2 pg/mL, P < .001).^[164] It has been postulated that this "AVP deficiency" may be the result of excessive baroreceptor-mediated AVP release from sustained hypotension. This theory is strongly supported by a canine study demonstrating almost complete depletion of radiolabeled AVP in the posterior pituitary after 1 hour of hemorrhagic shock ( Fig. 20-21 ).^[162]

A second mechanism for sensitivity to AVP in septic patients is explained by its effect on the potassium-ATP (K_ATP ) channel.^[162] Intracellular acidosis, lactic acid accumulation, and ATP depletion close K_ATP channels in the sarcolemma of vascular smooth muscle. This closing of channels traps potassium outside the cell and hyperpolarizes the membrane, which in turn closes the calcium channels essential for norepinephrine-induced vasoconstriction. AVP binds at the K_ATP channel and opens it, thereby reversing membrane hyperpolarization and restoring sensitivity to norepinephrine.

In summary, the beneficial effect of AVP on renal function in sepsis may in part be due to enhanced renal

Figure 20-21 Baroreflex-induced arginine vasopressin (AVP) depletion in dogs. A, Section of the posterior pituitary gland (neurohypophysis) from a normal dog. The macrovesicles are stained with antivasopressin serum, a reaction indicative of replete stores of vasopressin. B, Section of the neurohypophysis from a dog after severe hemorrhagic hypotension (mean arterial pressure, <40 mm Hg) for 1 hour. The minimal staining with antivasopressin serum is indicative of profound baroreflex-induced depletion of AVP stores. (Reprinted, by permission, from Landry DW, Oliver JA: The pathogenesis of vasodilatory shock. N Engl J Med 345:588–595, 2001.)