Introduction: The objective of the study was to describe, in medical-surgical patients in the intensive care unit (ICU), the prevalence (on ICU admission), incidence (in the ICU), risk factors and consequences of TCP. Hypothesis: TCP is frequent, multifactorial and associated with an increased risk of bleeding, transfusions and mortality. Methods: 3746 patients were enrolled in 67 centers to a thromboprophylaxis trial comparing heparin vs low molecular weight heparin (LMWH) (PROTECT, clinicaltrials.gov NCT00182143). Patients with platelet counts <75x109/L or severe coagulopathy were excluded. We conducted Cox regression analyses to a) identify the independent baseline and time-dependent predictors of 3 incident TCP thresholds; and b) examine the effect of 3 TCP thresholds on clinical outcomes. Results: For TCP thresholds of <50x109/L (severe), <100x109/L (moderate) and <150x109/L (mild), prevalence was: 0.2%, 6.8%, and 26.2% and incidence was: 4.9%, 11.5% and 22.1%. Consistent TCP predictors across platelet thresholds were: APACHE II score (severe HR 1.60 95%CI 1.29-1.97; moderate HR 1.52 95%CI 1.34-1.73; mild HR 1.25 95%CI 1.12-1.40), vasopressors (severe HR 3.32 95%CI 2.24-4.91; moderate HR 2.86 95%CI 2.28-3.59; mild HR 1.81 95%CI 1.53-2.15) and dialysis (severe HR 3.62 95%CI 2.49-5.27; moderate TCP 2.39 95%CI 1.84-3.10; mild TCP 2.79 95%CI 2.20-3.53). The risk of moderate TCP was lower in patients prophylaxed with LWMH (HR 0.72 95%CI 0.56-0.92) and medical admissions (HR 0.68; 95%CI 0.53-0.87), but higher in patients with liver disease (HR 3.03; 95%CI 2.00-4.61). Heparin-induced TCP was associated with an increased risk of mild (HR 7.31; 95%CI 2.68-19.92) and moderate TCP (HR 9.18; 95%CI 3.72-22.62). For each platelet threshold, TCP was associated with significantly more bleeding (any or major), transfusions (red cells, platelets or plasma), and mortality (ICU or hospital). Conclusions: Most consistent TCP risk factors are non-modifiable, including illness severity, advanced life support and organ dysfunction. LMWH lowers the risk of moderate TCP and HIT increases the risk of mild or moderate TCP. Patients with mild, moderate or severe TCP are more likely to bleed, receive transfusions, and die than other patients.