Summary

The burden of thromboembolism (TE) in severe sepsis is largely unknown. We assessed the prevalence of venous and arterial TE in patients with severe sepsis over a four-week period. We performed a retrospective analysis of a pooled database of three randomized, placebo-controlled trials of two novel pharmacological agents for the treatment of severe sepsis, drotrecogin alfa (activated) (DrotAA)and secretory phospholipase A2 inhibitor (sPLA2I). The study was conducted at intensive care units of the participating institutions. A total of 2,649 patients with known or suspected infection and sepsis-associated acute organ dysfunction were enrolled in the three trials and were assigned to treatment groups (DrotAA=850; sPLA2I =578; placebo=1221). The database was queried for venous and arterial TE, using investigator reports of serious adverse events. Eighty-four of 2,649 patients (3.2%; 95% confidence interval, 2.5% to 3.9%) developed at least one thromboembolic event over 28 days. Nearly threequarters of episodes were atheroembolic (n=62); 25% involved the deep venous system (n=25). Ischemic stroke (n=30) and venous thromboembolism (n=25) each occurred in about 1% of patients. Ischemic stroke and acute coronary syndrome had a higher peak incidence during the first five days compared to venous TE onset, which was more constant over the 28-day period. Subgroup analysis by pooled treatment groups yielded TE rates of 2.0% (DrotAA), 3.5% (placebo), and 4.0% (sPLA2I), respectively. Clinically manifest TE occurred in about 3% of severe sepsis patients treated in the intensive care unit over a 28-day period. Arterial TE may be more common than previously recognized. More accurate estimates of TE prevalence and relationship to sepsis await future studies.