Patients with severe multiple injuries often develop sepsis. The renal response to sepsis is, for the most part, a renal adaptation to an altered systemic circulatory status. Those factors that have created the most controversy regarding this response, therefore, are an extension of the controversy regarding the cardiovascular response to sepsis.37 Historically, sepsis has been characterized as a hypodynamic state with increased TPR and decreased CO; the kidneys, traditionally, shared in this response with increased RVR and decreased RBF, especially outer cortical (CI) flow.38 These traditional concepts, however, reflect endotoxin studies in animals and a poor understanding of fluid shifts in septic patients.
The empirical recognition that septic patients respond better to a fluid challenge questioned these traditional views.39 Clinical studies in severely septic patients showed an expanded PV, increased CO, decreased TPR, increased IFS volume, and increased urine volume.37,39 This hemodynamic response has been called the “hyperdynamic state of sepsis”. The kidney often shares in this hyperdynamic state.39
Polyuria exceeding 2 L per day is seen in many severely septic patients if resuscitation is initiated early enough to prevent renal ischemia and shutdown.39,40 Furthermore, polyuria may be inappropriate and persist despite fluid restriction that causes hypovolemia and hypotension.40 Since a vital renal function is regulation of PV by both diuresis and water conservation, this “inappropriate polyuria” soon after a septic insult is an unusual paradox.40 The physiology and danger of this phenomenon are illustrated in the following example (Fig. 59-5).
This 27-year-old woman presented with a 3-day history of lower abdominal pain, distention, fever, tenderness, leukocytosis, and a lower quadrant mass. Shortly thereafter, she became hypotensive and agitated; fluids containing vasopressors were administered. Vital signs responded to large-volume intravenous replacement. Laparotomy revealed massive peritoneal spillage from a ruptured viscus. Postoperatively, she exhibited marked fluid sequestration, weight gain, increased CVP (22–25 cm H2O), and polyuria (>300 mL/h); she received large volumes of intravenous fluid to prevent hypotension and was given inotropes. Her urine sodium concentration was 47 mEq/L. Despite objections from the surgical team, a fluid restriction regimen was instituted to “protect” the lungs (Fig. 59-5). Within 4 hours, the blood pressure fell to 90/70 mm Hg and the pulse rose to 125/min as the urine volume decreased from 300 to 50 mL/h. The pooled 4-hour urine specimen, which averaged 140 mL/h, contained 12 mEq sodium. By hour 28, hypotension (80/60 mm Hg) and tachycardia (155/min) worsened and she became anuric; the CVP fell to 21 cm H2O. The urine sodium just prior to ARF was 5 mEq/L. She remained anuric and died of respiratory failure 2 days later (Fig. 59-5).
This brief summary shows how the polyuric state of sepsis persists after the initiation of fluid restriction and contributes to hypotension; thus, it is inappropriate polyuria. Studies on patients with inappropriate polyuria of sepsis show an elevated CO, normal GFR, and normal TRPF associated with normal RBF distribution to the outer cortex, inner medulla–outer cortex, and inner medulla.40 Past experiences indicate that fluid restriction leading to ARF in comparable septic patients almost always is fatal despite the availability of HD. Since pulmonary function generally shows little improvement with fluid restriction during the initial 24 hours of a massive septic insult, the pulmonary insufficiency appears to result from the sepsis, per se, rather than a PV overload. Patients with this inappropriate polyuric syndrome must be monitored closely. When the urine sodium level falls below 10 mEq/L or the FENA is less than 1, expansion of PV to enhance renal perfusion is indicated.39,40
Mechanism of Inappropriate Polyuria of Sepsis
The polyuria of sepsis was initially observed by Ladd during the Korean conflict; he postulated osmotic diuresis as the etiology.12 Hyperosmolemia (between 290 and 330 mOs/L; normal is 270–278 mOs/L) has been confirmed in septic patients; furthermore, this hyperosmolemia cannot be explained by the combined osmolar effects of serum sodium, blood urea nitrogen (BUN), and serum glucose concentrations that are used to measure calculated serum osmolality (Calc Osm) by the following formula: Calc Osm = serum sodium × 1.86 + BUN (mg/dL)/2.5 + serum glucose (mg/dL)/18. The Calc Osm in septic patients, typically, is normal (250–265 mOs/L).39,40 The osmolar difference (OsmDiff) equals the actual serum osmolarity minus the Calc Osm and is usually increased in severely septic patients, averaging 47 mOs/L compared with a normal of 15 mOs/L.39 This increased OsmDiff reflects the severity of sepsis and correlates closely with mortality rates and the degree of polyuria. Contrary to Ladd’s theory that the polyuria is due to osmotic diuresis, however, COsm in septic patients averages about 3 mL/min, which is the upper limit of normal.39 The positive correlations between the OsmDiff and polyuria are probably spurious in that both factors are indices of the severity of sepsis. Hyperosmolemia, per se, however, may contribute to the polyuria by a mechanism not related to diuresis. It is a potent vasodilator in many vascular beds, including the kidney, and may facilitate diuresis by renal vasodilatation and its described effects.39,41
The polyuria of sepsis might be due, partially, to a juxtamedullary “washout” secondary to selective decrease of RBF to the outer cortex (CI) with maintained flow to the juxtamedullary nephrons (CII) prior to volume replacement. This reduces the inner medullary interstitial (CIII) osmotic pressure, thereby temporarily “paralyzing” the countercurrent mechanism. This explanation would explain polyuria that occurs during the early postresuscitation interval when the inner medullary osmolar gradient becomes reestablished. Two factors have negated this postulate. First, the period of increased RBF and decreased RVR in septic patients is transient, usually lasting 48–96 hours; in contrast, the polyuria may last beyond that time in patients subjected to a second septic insult that is not associated with hypotension. Second, RBF distribution measurements by Xe disappearance during the period of inappropriate polyuria, but 72 hours after the initial insult when the TPR was normal, did not support this conclusion.40 Each patient had normal intrarenal distribution to all three renal components (Fig. 59-6).40
This patient with extensive peritonitis had inappropriate polyuria (6 mL/min; 360 mL/h) despite marginal blood pressure and normal RPF (1,200 mL/min), GFR (95 mL/min), and RBF distribution to CI, CII, and CIII. This inappropriate polyuria subsided after 36 hours and he recovered.
Experimental studies on the effects of sepsis on renal hemodynamics show a decrease in RVR and an increase in RBF.41,42 Hermreck et al. demonstrated that dogs made septic by the introduction of enteric flora into the muscles of the hind limb developed increased RBF as measured by an electromagnetic flow meter.41 Ravikant and Lucas monitored renal circulation with this hind limb sepsis model, using radioactive microspheres; they showed a hyperdynamic state with increased CO and RBF despite a lower MAP and CVP when compared with control animals.42 Ravikant and Lucas also showed a decrease in EPAH indicating that some blood entering the kidney was shunted into the renal vein without actively engaging in renal metabolism.42 The reduced EPAH in septic animals led to clinical studies in which renal vein sampling was used to measure EPAH in septic patients.40 These measurements showed a significant reduction in EPAH, calculated RBF using measured EPAH, and documented decreased RVR and increased RBF early in the course of septic patients who were fully resuscitated and had normal vital signs.40 Decreased EPAH in septic animals and patients parallels similar findings in human volunteers receiving pyrogens intravenously.43 Both septic patients and volunteers receiving pyrogens have a comparable increase in RBF, urine output, and CNa; the same is seen in animals and humans after vasodilation with phenoxybenzamine administration.43
Hermreck et al. implicated a diabetes insipidus–like syndrome.41 They blocked the polyuria in dogs with infected hind limbs by the systemic administration of ADH. The ADH dosage, however, was greater than that which produces a specific renal effect independent of systemic vasoconstriction. Using renal physiologic dosages of ADH (1–2 μg/kg/h) injected into the renal arteries of septic patients with inappropriate polyuria, no effect was seen on the polyuria.40 This syndrome has also been attributed to a distal tubular blockade of aldosterone receptor sites; calculated aldosterone-like effect on the renal tubules (CH2O/(CH2O + CNa) in septic patients with positive free water clearance, however, shows no such impairment. Indeed, most septic patients can reabsorb sodium with great efficiency right up to the development of AORF, and few have a positive CH2O (C. E. Lucas and I. K. Rosenberg, Surgical renal study unit, unpublished data).39
The therapeutic implications of this syndrome are apparent. Clinicians must recognize that “a good urine output” of 35–50 mL/h in the early septic period may not be adequate and may reflect PV deficiency. Likewise, a large urine output exceeding 200 mL/h need not reflect overload.39 Polyuria, together with large insensible losses, IFS expansion, and increased gastric losses over a period of days, may lead to an incipient hypovolemia that goes unrecognized prior to a “sudden” vascular collapse. Careful monitoring of intake and output provides clues of impending collapse; blood pressure, pulse, pulse pressure, and urine sodium concentration must be monitored closely.39
Renal Response to Hypodynamic Sepsis
Not all septic patients develop the hyperdynamic state.37,39 Indeed, up to 35% of patients with sepsis develop a low CO and a high TPR. The renal response to “hypodynamic” sepsis consists of increased RVR, decreased RBF, and decreased GFR reflecting, primarily, a decrease in outer cortical (CI) flow. This leads to a low CNa, COsm, and urine output. Therapy for the oliguria must be directed toward supporting the myocardium and maintaining PV. Persistent oliguria in fully resuscitated patients requires induced diuresis with mannitol or a loop diuretic.35 When oliguric ARF is developing, a loop diuretic, such as furosemide, beginning at 40 mg and doubling this dose every 30–60 minutes to a total of 3,200 mg, may prevent renal shutdown. If a renal response does occur with this regimen, it usually occurs by the time the furosemide dosage has reached 160 mg.35 When there is no response, the end result is oliguric ARF.
Significance of Free Water Clearance
Free water clearance (CH2O) represents the difference between urine output and osmolar clearance (COsm). The addition of CH2O measurements provides more insight into renal concentrating capacity.44 Concentrating impairment, as reflected by the excretion of a dilute urine, is due to an intrinsic renal dysfunction. This defect may proceed azotemia and oliguria by 24–72 hours.45 Excess CH2O (CH2O > −0.25 mL/min) in critically ill patients is associated with a reduction in PV, GFR, and ERPF.45 This subgroup of patients has the highest incidence of subsequent renal failure (21%). When the CH2O approaches unity (±0.25 mL/min), careful monitoring will help determine whether this is a normal response to increased PV or to an intrinsic tubular defect predicting subsequent ARF.44 Patients with a rising CH2O, without evidence of PV expansion, need to be watched carefully for toxic renal insult, particularly from associated medicines such as renal toxic antimicrobials (C. E. Lucas and I. K. Rosenberg, Surgical renal study unit, unpublished data). The effect of the random addition of albumin (HSA) to a resuscitation regimen also affects CH2O.45 The significant decrease in CNa associated with HSA-supplemented resuscitation causes a rise in CH2O. This is due to both increased peritubular oncotic pressure and decreased filtered sodium load. These changes are mediated through distal nephron sodium–potassium exchange.45