Abstract
BACKGROUND: Studies have shown routine ultrasound surveillance (RUSS) will facilitate deep vein thrombosis (DVT) detection in patients with trauma and reduce the subsequent incidence of pulmonary embolism (PE); however, the findings were inconsistent. In adults with trauma at a high risk of venous thromboembolism, this systematic review and meta-analysis compared RUSS outcomes with those of "no RUSS." METHODS: Three databases were screened for relevant articles from inception to October 18, 2021. Randomized controlled trials (RCTs) and observational studies comparing RUSS with no RUSS were included. We used relative risks (RRs), odds ratios (ORs), and mean differences to pool effect estimates for dichotomous and continuous outcomes. The cochrane risk of bias or the risk of bias in non-randomized studies of interventions were used to assess bias risk. The grading of recommendations, assessment, development, and evaluation framework assessed the certainty of the evidence. FINDINGS: Out of 1685 articles, 5 met the inclusion criteria (RCT: 1; observational studies: 4). Observational studies suggested RUSS is associated with higher odds of DVT detection (OR, 4.87; 95% confidence interval [CI], 3.13-7.57; very low certainty). Whereas higher risks of DVT were associated with RUSS in the RCT (distal DVT: RR, 15.48; 95% CI, 7.62-31.48; low certainty, and proximal DVT: RR, 2.37; 95% CI, 1.04-5.39; very low certainty). Reduced odds of PE risk were observed with the RUSS (OR, 0.47; 95% CI, 0.24-0.91; very low certainty). Observational studies indicated that RUSS had an uncertain effect on mortality (OR, 0.46; 95% CI, 0.06-3.49). In the RCT, times to proximal and distal DVT diagnoses were shorter with RUSS (proximal DVT, mean difference 2.25 days shorter [95% CI, 5.74-1.24]; distal DVT, mean differences 1.56 days shorter [95% CI, 4.22-1.12]; very low certainty for both). Increasing bleeding risk was not linked to the RUSS group (RR, 1.24; 95% CI, 0.31-4.92). INTERPRETATION: The RUSS efficacy in adults with trauma at high risk for venous thromboembolism showed that it increases DVT detection, decreases PE incidence, and shortens the time to DVT diagnosis, with an uncertain impact on mortality. The evidence is low or very low in certainty because of bias, inconsistency, imprecision, and indirectness.