Objective: To determine the validity of severity of illness scoring systems for patients with thrombotic thrombocytopaenic purpura (TTP) treated in an intensive care unit (ICU) and to determine ICU outcomes in this population. Design: Retrospective, cohort analysis. Setting: Tertiary care, university affiliated medical center. Subjects: Nine consecutive patients with TTP and a control cohort of six subjects who underwent autologous haematopoietic stem cell transplantation (AHSCT) and were admitted to the ICU during the same 3 year time period. Interventions: Patient surveillance and data collection. Endpoints: Survival to hospital discharge and need for mechanical ventilation (MV), and mortality. Measurements and results: The study cohort included nine TTP patients (mean age 45.3 years, 88.9% female); 3 (33.3%) survived. No subject requiring mechanical ventilation (n = 6) survived (p = 0.012). Only the need for MV separated TTP survivors from non-survivors. At both admission to the ICU and on ICU day three, no index of disease activity (haemoglobin, platelet count, lactate dehydrogenase, or creatinine) differed between survivors and non-survivors. The APACHE II score was significantly lower in TTP subjects (13.1 ± 1.8 vs. 29.3 ± 6.1, p < 0.001) as was the Multiple Organ Dysfunction Score (5.8 ± 1.0 vs. 9.8 ± 1.3, p<0.001). However, overall mortality and the need for MV were similar between the TTP and AHSCT groups. Conclusions: Patients with TTP requiring ICU care face significant mortality. Outcomes for patients with TTP who require MV are dismal. Severity of illness scoring systems may underestimate mortality for patients with TTP.