Comparison of outcomes from sepsis between patients with and without pre-existing left ventricular dysfunction: a case-control analysis

The aim of this study was to determine if there are differences between patients with pre-existing left ventricular dysfunction and those with normal antecedent left ventricular function during a sepsis episode in terms of in-hospital mortality and mortality risk factors when treated in accordance with a sepsis treatment algorithm. We performed a retrospective case-control analysis of patients selected from a quality improvement database of 1,717 patients hospitalized with sepsis between 1 January 2005 and 30 June 2010. In this study, 197 patients with pre-existing left ventricular systolic dysfunction and sepsis were compared to 197 case-matched patients with normal prior cardiac function and sepsis. In-hospital mortality rates (P = 0.117) and intubation rates at 24 hours (P = 0.687) were not significantly different between cases and controls. There was no correlation between the amount of intravenous fluid administered over the first 24 hours and the PaO2/FiO2 ratio at 24 hours in either cases or controls (r2 = 0.019 and r2 = 0.001, respectively). Mortality risk factors for cases included intubation status (P = 0.016, OR = 0.356 for no intubation), compliance with a sepsis bundle (P = 0.008, OR = 3.516 for failed compliance), a source of infection other than the lung (P = 0.019, OR = 2.782), and the initial mixed venous oxygen saturation (P = 0.004, OR = 0.997). Risk factors for controls were the initial platelet count (P = 0.028, OR = 0.997) and the serum lactate level (P = 0.048, OR = 1.104). Patients with pre-existing left ventricular dysfunction who died had a lower initial mean mixed venous oxygen saturation than those who survived (61 ± 18% versus 70 ± 16%, P = 0.002). Clinical outcomes were not different between septic patients with pre-existing left ventricular dysfunction and those with no cardiac disease. There was no correlation between fluid administration and oxygenation at 24 hours in either cohort. The mortality risk factor profile of patients with pre-existing left ventricular dysfunction was different when compared with control patients, and may be related to oxygen delivery determinants.