Introduction: The frequency, predisposition and significance of venous thrombi outside the legs are unclear. Our objective was to describe the incidence, risk factors, and consequences of non-leg venous thrombi in the ICU. Hypothesis: We hypothesized that non-leg-thrombi would occur in 2% of patients, would have no specific risk factors, and would not increase mortality. Methods: We conducted nested observational study in a 3746 patient thromboprophylaxis trial (NCT00182143). Agreement between 2 thrombus adjudicators on was calculated using chance-corrected agreement. We conducted a Cox proportional hazard multivariable analysis introducing possible baseline (APACHE II score, body mass index, malignancy) and time-dependent (vasopressors. statins) risk factors. We describe consequences of these thrombi in terms of pulmonary embolism (PE) risk and duration of ICU and hospital stay. To examine whether a proximal non-leg-deep vein thrombosis (DVT) influenced ICU and hospital mortality risks, we conducted a multivariable analysis adjusted for age, APACHE II score, mechanical ventilation, vasopressors, dialysis, and platelet count <50,000x109/L. Results: 84 (2.2%) patients were diagnosed with a non-leg thrombus (deep or superficial). Most were incident; of these, 59 (1.6%) were deep and 75 (2.0%) were both deep and superficial. Adjudicator agreement on the classification was excellent (kappa 0.91). All patients with a non-leg DVT had a central venous catheter (CVC) in ICU, compared to only 86.2% of those without a non-leg DVT. The only independent risk factor for non-leg DVT was cancer (malignancy in the last 5 years; hazard ratio 2.22 (1.06-4.65); p=0.034). Compared to patients without a non-leg DVT, those who had one were more likely to develop a PE (14.9% vs 1.9%, p<0.001) and had a longer stay in ICU (19 vs 9 days, p <0.001) and hospital (39.5 vs 21 days, p<0.001). Adjusting for confounders, a non-leg DVT did not increase ICU mortality (odds ratio 1.09, 0.62-1.92; p=0.757). Conclusions: Non-leg venous thrombi occur in about 2% of ICU patients, often with CVCs or cancer. Although associated with PE and longer lengths of stay, non-leg venous thrombi do not independently increase risk of death.