Surveillance or no surveillance ultrasonography for deep vein thrombosis and outcomes of critically ill patients: a pre-planned sub-study of the PREVENT trial

Purpose We examined the association between surveillance for deep vein thrombosis (DVT) among medical-surgical critically ill patients by twice-weekly ultrasonography and 90-day all-cause mortality. Methods This was a pre-planned sub-study of the Pneumatic Compression for Preventing Venous Thromboembolism (PREVENT) trial (Clinicaltrials.gov: NCT02040103) that compared addition of intermittent pneumatic compression (IPC) to pharmacologic prophylaxis versus pharmacologic prophylaxis alone. The surveillance group included enrolled patients in the trial, while the non-surveillance group included eligible non-enrolled patients. Using logistic regression and Cox proportional hazards models, we examined the association of surveillance with the primary outcome of 90-day mortality. Secondary outcomes were DVT and pulmonary embolism (PE). Results The surveillance group consisted of 1682 patients and the non-surveillance group included 383 patients. Using Cox proportional hazards model with bootstrapping, surveillance was associated with a decrease in 90-day mortality (adjusted HR 0.75; 95% CI 0.57, 0.98). Surveillance was associated with earlier diagnosis of DVT [(median 4 days (IQR 2, 10) vs. 20 days (IQR 16, 22)] and PE [median 4 days (IQR 2.5, 5) vs. 7.5 days (IQR 6.1, 28.9)]. There was an increase in diagnosis of DVT (adjusted HR 5.49; 95% CI 2.92, 13.02) with no change in frequency in diagnosis of PE (adjusted HR 0.56; 95% CI 0.19, 1.91). Conclusions Twice-weekly surveillance ultrasonography was associated with an increase in DVT detection, reduction in diagnostic testing for non-lower limb DVT and PE, earlier diagnosis of DVT and PE, and lower 90-day mortality.