Fifty dysfunctional central venous catheters proven radiographically to be occluded by thrombus were blindly randomized to be injected with either 2 mg recombinant tissue plasminogen activator (t-PA) or 10,000 units of urokinase (UK) and allowed to incubate for 2 h. A second dose was allowed if catheter function was not restored with the first injection. Repeat radiograph contrast injection was done when catheter function was restored or after 2 doses of study drug were administered, whichever occurred first. Thirteen of 22 catheters randomized to UK had full function restored compared to 25 of 28 randomized to t-PA (p = 0.013). Radiographic contrast injection showed 7 catheters randomized to UK had complete resolution of the thrombus compared to 17 randomized to t-PA (p = 0.042). Four catheters randomized to UK had complete resolution of the thrombus after a single dose compared to 13 randomized to t-PA (p = 0.036). A novel dose of 2 mg of t-PA restored catheter function more reliably and dissolved thrombi faster than twice the standard, FDA-approved dose of UK.