Traditionally, adult sepsis has been considered a contraindication to extracorporeal life support (ECLS). The objective of this study was to review the authors' institutional experience with a subgroup of adult patients requiring ECLS for severe respiratory failure and sepsis. Hospital records from 100 consecutive adult patients with respiratory failure placed on ECLS between 1990 and 1996 were retrospectively reviewed. Patients with sepsis as a primary indication were identified, and blood culture data reviewed. Data were analyzed with t tests and chi-square and are presented as mean +/- standard deviation. Multiple logistic regression determined the impact of sepsis and positive blood cultures (PBCs) on survival. Fourteen patients required ECLS for sepsis; 36 had PBCs during hospitalization (15 before or during ECLS). Septic patients had lower pre-ECLS PaO2/FIO2 ratios (septic: 53 +/- 14 mmHg, nonseptic: 70 +/- 68 mmHg, p = 0.04). Patients with PBCs before or during ECLS were younger (PBC: 29 +/- 6 years, no PBC: 35 +/- 13 years, p = 0.003), remained on ECLS longer (PBC: 485 +/- 336 hours, no PBC: 232 +/- 212 hours, p = 0.01), and were more frequently cannulated within 12 hours (PBC: 15/15, no PBC 60/85 p = 0.02). Neither group differed in organ dysfunction (incidence or type), frequency of respiratory recovery, or survival. Neither sepsis nor positive blood cultures were independently predictive of mortality. Sepsis and positive blood cultures do not adversely affect outcome in adult patients with respiratory failure requiring ECLS.