A comprehensive assessment of multi-system responses to a renal inoculation of uropathogenic E. coli in swine

BACKGROUND: The systemic responses to infection and its progression to sepsis remains poorly understood. Progress in the field has been stifled by the shortcomings of experimental models which include poor replication of the human condition. To address these challenges, we developed and piloted a novel large animal model of severe infection that is capable of generating multi-system clinically relevant data. METHODS: Male swine (n = 5) were anesthetized, mechanically ventilated, and surgically instrumented for continuous hemodynamic monitoring and serial blood sampling. Animals were inoculated with uropathogenic E. coli by direct injection into the renal parenchyma and were maintained until a priori endpoints were met. The natural history of the infection was studied. Animals were not resuscitated. Multi-system data were collected hourly to 6 hours; all animals were euthanized at predetermined physiologic endpoints. RESULTS: Core body temperature progressively increased from mean (SD) 37.9(0.8)°C at baseline to 43.0(1.2)°C at experiment termination (p = 0.006). Mean arterial pressure did not begin to decline until 6h post inoculation, dropping from 86(9) mmHg at baseline to 28(5) mmHg (p = 0.005) at termination. Blood glucose progressively declined but lactate levels did not elevate until the last hours of the experiment. There were also temporal changes in whole blood concentrations of a number of metabolites including increases in the catecholamine precursors, tyrosine (p = 0.005) and phenylalanine (p = 0.005). Lung, liver, and kidney function parameters worsened as infection progressed and at study termination there was histopathological evidence of injury in these end-organs. CONCLUSION: We demonstrate a versatile, multi-system, longitudinal, swine model of infection that could be used to further our understanding of the mechanisms that underlie infection-induced multi-organ dysfunction and failure, optimize resuscitation protocols and test therapeutic interventions. Such a model could improve translation of findings from the bench to the bedside, circumventing a significant obstacle in sepsis research.