Purpose: To determine how the interval between ultrasonography (US) and a definitive diagnosis affects the accuracy of duplex US for the diagnosis of hepatic artery thrombosis after transplantation.
Materials and methods: Retrospective chart review was performed in 202 patients with liver transplants. Results of duplex US were compared with those of angiography, surgery, or autopsy.
Results: Sensitivity for US was 54% up to 1 month before and 82% on the day of the definitive diagnostic study. Specificity remained relatively constant over time (86%-87%). Accuracy reached 85%, and negative predictive value reached 93%. Qualitative changes on the sonogram enabled prediction of subsequent thrombosis. Rejection, recurrent hepatitis, aberrant anatomy, hypotension, arterial collaterals, and operator error caused false-positive or false-negative results.
Conclusion: Although duplex US remains a good screen for hepatic artery thrombosis, angiography is strongly recommended. Serial US studies are necessary for diagnosis. Sensitivity improves over time as the clinical picture clarifies. Qualitative changes on the sonogram (the syndrome of impending thrombosis) enable prediction of subsequent hepatic artery thrombosis.