Lower Extremity Deep Venous Thrombosis: Vascular Laboratory Quality Assurance Without Correlation Between Ultrasound and Venography

Abstract

Venography is rarely available for comparison with ultrasonography (US) as a means for quality assurance (QA) in the detection of lower extremity venous thrombosis. New QA methods must be implemented. We compared results of multiple serial studies performed in the same extremity as a QA indicator. From a 3-year sample of close to 9,000 venous tests, we obtained a subset of 44 patients who had 331 tests in 71 lower extremities throughout the years. A positive or negative study preceded or followed by another positive or negative study was considered as a confirmed study. A negative or positive study not preceded or followed by a negative or positive study was considered as unconfirmed. Explanations were then sought to explain unconfirmed results. There were 169 (51%) and 124 (37%) confirmed positive and negative studies, respectively, and 13 (4%) and 25 (8%) unconfirmed positive and negative studies, respectively. Of the 13 unconfirmed positive tests, 2 were preceded by negative tests, 3 were preceded and followed by negative tests, and 8 were followed by negative tests. Of these 13 tests, 4 documented extensive venous thrombosis. Of the 25 unconfirmed negative tests, 11 followed treatment for venous thrombosis, 6 had recurrent thrombosis with intermittent lysis, and 8 were followed by positive tests. Considering the low probability of extensive thrombosis being a false-positive test, positive predictive value was 95% (173/182). Excluding 11 negative tests following treatment for venous thrombosis, negative predictive value was 90% (124/138) and accuracy was 93% (297/320). US versus US and literature US versus venography comparisons of these statistics were similar.