Abstract

Reduced renal blood flow (RBF) is considered central to the pathogenesis of septic acute renal failure (ARF). However, no controlled experimental studies have continuously assessed RBF during the development of severe septic ARF. We conducted a sequential animal study in seven female Merino sheep. Flow probes were implanted around the pulmonary and left renal arteries. Two weeks later, systemic hemodynamics and RBF were monitored continuously during a 48-h control period and, after a week, during a 48-h period of hyperdynamic sepsis induced by continuous Escherichia coli infusion. Infusion of E. coli induced hyperdynamic sepsis with significantly increased cardiac output (3.8±0.4 vs 9.8±1.1 l/min; P24 h). Accordingly, we have developed a reproducible model of sustained, hyperdynamic sepsis. Using this model, we have investigated the changes in renal hemodynamics and function in this setting and now report our findings. All animals were in a healthy state before commencing the experiments (Table 1). In six of the seven sheep, the Escherichia coli infusion was continued for 48 h. One sheep died 12 h after the induction of sepsis. The sheep developed tachypnea, tachycardia, and a temperature of >41°C, and began to use the accessory muscles of respiration. The white blood cells decreased after 48 h of E. coli infusion to 1600±800/μl compared with 5400±2900/μl in the control period (P24 h) sepsis, the condition typically seen in critically ill humans. In the first, Cumming et al.36.Cumming A.D. Driedger A.A. McDonald J.W. et al.Vasoactive hormones in the renal response to systemic sepsis.Am J Kidney Dis. 1988; 11: 23-32Abstract Full Text PDF PubMed Scopus (69) Google Scholar induced sepsis in sheep by cecal ligation and perforation. These authors found a 70% decrease in glomerular filteration rate but had no control animals and did not measure RBF or renal vascular conductance directly. Instead of measuring RBF via flow probes, radioactive labelled para-aminohippurate was used to determine the renal plasma flow. Particularly in the septic setting, this technique might not be sufficiently accurate.37.Rector F. Goyal S. Rosenberg I.K. et al.Sepsis: a mechanism for vasodilatation in the kidney.Ann Surg. 1973; 178: 222-226Crossref PubMed Scopus (36) Google Scholar, 38.Lucas C.E. Rector F.E. Werner M. et al.Altered renal homeostasis with acute sepsis. Clinical significance.Arch Surg. 1973; 106: 444-449Crossref PubMed Scopus (56) Google Scholar Nonetheless, in this study, renal plasma flow did not decrease.36.Cumming A.D. Driedger A.A. McDonald J.W. et al.Vasoactive hormones in the renal response to systemic sepsis.Am J Kidney Dis. 1988; 11: 23-32Abstract Full Text PDF PubMed Scopus (69) Google Scholar The second and third studies, by Weber et al.,39.Weber A. Schwieger I.M. Poinsot O. et al.Sequential changes in renal oxygen consumption and sodium transport during hyperdynamic sepsis in sheep.Am J Physiol. 1992; 262: F965-F971PubMed Google Scholar, 40.Weber A. Schwieger I.M. Poinsot O. et al.Time course of systemic and renal plasma prostanoid concentrations and renal function in ovine hyperdynamic sepsis.Clin Sci (London). 1994; 86: 599-610Crossref PubMed Scopus (11) Google Scholar were also uncontrolled in design and sepsis was induced by continuous endotoxin infusion. In both studies, only two animals survived to be assessed at 48 h. In both studies, however, once a hyperdynamic state had developed after 24 h and in those animals that were still alive, RBF increased39.Weber A. Schwieger I.M. Poinsot O. et al.Sequential changes in renal oxygen consumption and sodium transport during hyperdynamic sepsis in sheep.Am J Physiol. 1992; 262: F965-F971PubMed Google Scholar and renal vascular resistance decreased.40.Weber A. Schwieger I.M. Poinsot O. et al.Time course of systemic and renal plasma prostanoid concentrations and renal function in ovine hyperdynamic sepsis.Clin Sci (London). 1994; 86: 599-610Crossref PubMed Scopus (11) Google Scholar Our findings are important because there is a widely proposed paradigm that, in sepsis, renal ischemia occurs owing to renal vasoconstriction and a consequent decrease in RBF.12.Badr K.F. Sepsis-associated renal vasoconstriction: potential targets for future therapy.Am J Kidney Dis. 1992; 20: 207-213Abstract Full Text PDF PubMed Scopus (61) Google Scholar, 13.De Vriese A.S. Bourgeois M. Pharmacologic treatment of acute renal failure in sepsis.Curr Opin Crit Care. 2003; 9: 474-480Crossref PubMed Scopus (18) Google Scholar, 14.Schrier R.W. Wang W. Acute renal failure and sepsis.N Engl J Med. 2004; 351: 159-169Crossref PubMed Scopus (874) Google Scholar Contrary to this paradigm, we found that septic ARF in our sheep occurred in the setting of marked renal vascular vasodilatation and a marked increase in RBF. Given that 90% of RBF is delivered to the glomeruli, in order for such renal vascular vasodilatation to occur, both afferent and efferent arteriolar vasodilatation should logically occur. We found that in this setting CC, a surrogate of glomerular filtration rate, decreased. Glomerular filteration rate is mostly dependent upon intraglomerular pressure. The relationship between afferent and efferent arteriolar tone controls such pressure. Thus, it is physiologically logical to expect glomerular filteration rate to decrease if the efferent arteriole dilates more than the afferent arteriole. We speculate that this might have happened in our animals. This theory is supported by the finding that the filtration fraction in the sepsis group decreased significantly despite an increase in renal plasma flow. Nonetheless, it is not possible to know whether the loss of CC seen in our animals is indeed secondary to such hemodynamic changes alone, whether other factors played a role or even whether such hemodynamic changes represent an epiphenomenon. However, a decrease in the fractional excretion of sodium, a marker for relatively intact tubular function, was seen over time compared to the control period. This significant decrease in FeNa has also been described in septic humans.41.Vaz A.J. Low fractional excretion of urine sodium in acute renal failure due to sepsis.Arch Intern Med. 1983; 143: 738-739Crossref PubMed Scopus (51) Google Scholar Similarly, decreased fractional excretion of urea nitrogen, another marker of relatively preserved tubular function,42.Carvounis C.P. Nisar S. Guro-Razuman S. Significance of the fractional excretion of urea in the differential diagnosis of acute renal failure.Kidney Int. 2002; 62: 2223-2229Abstract Full Text Full Text PDF PubMed Scopus (228) Google Scholar was also observed during sepsis in our animals. The dramatic increase in RBF in sepsis might have been associated with intrarenal shunting and a change in intrarenal blood flow distribution.43.Brezis M. Rosen S. Hypoxia of the renal medulla – its implications for disease.N Engl J Med. 1995; 332: 647-655Crossref PubMed Scopus (895) Google Scholar Thus, ischemia to the medulla might have occurred despite increased global RBF. However, in a recent study, using laser Doppler flow probes, no significant change in the intrarenal distribution of blood flow could be demonstrated in sepsis.22.Di Giantomasso D. Morimatsu H. May C.N. et al.Intrarenal blood flow distribution in hyperdynamic septic shock: effect of norepinephrine.Crit Care Med. 2003; 31: 2509-2513Crossref PubMed Scopus (96) Google Scholar Similarly, studies using microspheres could not detect any significant change in intrarenal blood flow distribution in sepsis.44.Bone H.G. Schenarts P.J. Fischer S.R. et al.Pyridoxalated hemoglobin polyoxyethylene conjugate reverses hyperdynamic circulation in septic sheep.J Appl Physiol. 1998; 84: 1991-1999PubMed Google Scholar, 45.Booke M. Armstrong C. Hinder F. et al.The effects of propofol on hemodynamics and renal blood flow in healthy and in septic sheep, and combined with fentanyl in septic sheep.Anesth Analg. 1996; 82: 738-743PubMed Google Scholar, 46.Cronenwett J.L. Lindenauer S.M. Hemodynamic effects of cecal ligation sepsis in dogs.J Surg Res. 1982; 33: 324-331Abstract Full Text PDF PubMed Scopus (10) Google Scholar It is possible that medullary ischemia might yet occur in the setting of hyperemia because oxygen consumption increases more than oxygen delivery. This would require a greater than threefold increase in renal oxygen consumption. When renal oxygen consumption has been measured during experimental sepsis, it has been found to be only mildly increased,47.Bellomo R. Kellum J.A. Pinsky M.R. Transvisceral lactate fluxes during early endotoxemia.Chest. 1996; 110: 198-204Crossref PubMed Scopus (74) Google Scholar, 48.van Lambalgen A.A. Runge H.C. van den Bos G.C. et al.Regional lactate production in early canine endotoxin shock.Am J Physiol. 1988; 254: E45-E51PubMed Google Scholar whereas oxygen extraction remained unchanged.39.Weber A. Schwieger I.M. Poinsot O. et al.Sequential changes in renal oxygen consumption and sodium transport during hyperdynamic sepsis in sheep.Am J Physiol. 1992; 262: F965-F971PubMed Google Scholar Furthermore, a recent study of renal ATP in sheep with bacteremic septic shock found no decrease in high-energy phosphate compounds during sepsis.49.May C. Wan L. Williams J. et al.A technique for the measurement of renal ATP in a large animal model of septic shock.Int J Artif Organs. 2005; 28: 16-21PubMed Google Scholar Our study has several limitations. First, it is neither randomized nor double-blinded in design. However, it is controlled and the changes are so dramatic that it is inconceivable that they would represent an alpha error. Furthermore, the physiological changes are objective and not subject to bias. Second, we did not measure renal oxygen consumption, which would have been important in the interpretation of the changes in organ blood flow. However, measurement of regional oxygen consumption requires an acute preparation, because we have found that it is difficult to maintain the patency of chronically implanted venous cannulae and our goal was to study awake animals to minimize confounding variables. Third, our model does not completely reproduce severe human sepsis. For example, severe human sepsis is associated with an attributable mortality approaching 30%. However, whereas only one of the experimental animals died (15% mortality), two more were extremely ill at the time of euthanasia and would have likely died in the next 12 h. Fourth, no antibiotics were given to more closely simulate the human situation. Furthermore, only a few conditions in humans (e.g. endocarditis) are associated with almost constant bacteremia. In most other septic states, bacteremia is episodic. All of these differences between our model and human sepsis must be taken into account in the interpretation of our findings. To our knowledge, however, this is the first study to describe a model of ARF induced by hyperdynamic sepsis that mimics so many aspects of severe Gram-negative sepsis in humans: tachycardia, tachypnea, fever, leukopenia, hypotension, oliguria, high CO, and peripheral vasodilatation all sustained for 48 h and with significant mortality. Finally, we did not study renal histopathology. However, this was not the focus of the present investigation, in which we aimed to assess the effect of sustained sepsis on RBF, vascular conductance, and renal function. In conclusion, we have studied the effects of sustained Gram-negative bacteremia and sepsis on RBF, renal vascular conductance, and renal function. In addition to generalized peripheral vasodilatation with increased CO and decreased MAP, we found that such sepsis induced renal vasodilatation accompanied by a striking increase in RBF. Despite this marked increase in RBF, however, CC decreased significantly and serum creatinine increased fourfold. Our findings provide proof of concept that ARF can occur in mammals in the setting of hyperemia and highlight the need for renewed interest in investigations directed at measuring blood flow and renal vascular conductance in critically ill septic patients who develop ARF.