Between December 2002 and February 2008, the consecutive patients who had a first episode of proximal DVT and who underwent CDT were included for this prospective study. The diagnosis of DVT was confirmed by the characteristic clinical manifestations and the Doppler ultrasound and/or venographic findings. The patients who were eligible for enrollment had documented symptoms of lower extremity swelling and/or pain, a duration of symptoms beyond 14 days and they were positive for D-dimer. The exclusion criteria were isolated infrapopliteal DVT, recurrent ipsilateral DVT, pre-existing leg ulcers or signs of chronic venous insufficiency, a short life expectancy and contraindications to the use of contrast media, anticoagulation or thrombolytic agents. The contraindications to thrombolytic agents included active internal bleeding, recent cerebrovascular accident, recent serious trauma, recent major surgery (less than 2 weeks), allergy to thrombolytic agents, recent gastrointestinal bleeding, uncontrolled hypertension, pregnancy, bacterial endocarditis and coagulopathy. The study protocol was approved by the local ethics committee (China).

With the patient in the prone position, venous access was obtained through the ipsilateral popliteal vein under ultrasound guidance. A 5-Fr vascular sheath (Terumo, Japan) was placed and direct venography was performed to estimate the severity and extent of the DVT under fluoroscopy. A 5-Fr multiple-side-hole lysis catheter (AngioDynamics, Queensbury, NY) was then gently inserted with the tip embedded in the distal extent of the thrombus, and the catheter and sheath were fixed and joined with a microinjection pump. Urokinase (Tianjin Biochem Pharma Co., LTD, China) was continuously pumped in at a low dose of 50,000-100,000 IU/8 hours (it was constituted in 50 ml of 0.9% saline) through the catheter with the same dose of urokinase being simultaneously pumped through the sheath. Surveillance venography was performed every 2-3 days to access the degree of thrombolysis and to adjust the catheter position to facilitate effective thrombolysis. The fibrinogen level was monitored at varying intervals, and if the fibrinogen level dropped below 100 mg/dL, then further urokinase infusion was postponed to allow the fibrinogen to recover, and then the urokinase infusion was restarted at 50% of the original dose. After thrombolysis, further adjunctive procedures were done if there was an underlying iliac vein stenosis of greater than 50%. The adjunctive procedures consisted of PTA and stent placement. Before catheterization, we recommended prophylactic placement of a retrievable inferior vena cava (IVC) filter under local anesthesia through the right jugular vein or the contralateral femoral vein, and especially when patients had symptomatic pulmonary embolism, or the thrombus was free-floating or it had propagated into the IVC. If there were large amounts of thrombus in the iliac vein before PTA, then placing an IVC filter was recommended (6). In general, the IVC filter was removed within two weeks after the placement.

In the hospital, low molecular weight heparin (LMWH; Fraxiparine, Glaxo Smith Kline, Suzhou, China) was administered subcutaneously at the rate of 0.4-0.6 ml for 12 hours initially before CDT. Warfarin sodium (QILU Pharma Co., Ji'nan, China) was routinely started prior to hospital discharge and this was continued for at least six months, and the dose was adjusted to maintain an international normalized ratio (INR) of 2.0 to 3.0. The LMWH was discontinued when the INR reached the target value for two consecutive days and the patients had received it for at least five days. After hospital discharge, all the patients were recommended to wear gradual compression stockings (below-knee, 30-40 mmHg, Sigvaris Inc., Peachtree City, GA) for more than one year.

During the treatment, we recorded the affected limbs' circumference and the peri-procedure complications. We measured the thigh circumference at 15 cm above the knee joint and that of the calf at 10 cm below the tibial tuberosity. Complications were classified by the outcome with using the SIR (Society of Interventional Radiology) classification system for complications (7, 8).

The venograms were retrospectively reviewed by a consultant physician who was 'blinded' to the study's purpose. Venous recanalization was determined by comparing the thrombus scores (9) of the pre- and post-treatment venography, and this was categorized as follows: "complete" recanalization for 95-100% restoration of patency, "partial" for 50-95% and "minimal" for less than 50% due to residual stenosis or organized thrombus (10, 11). The recanalization was calculated after completion of treatment, which included the CDT and any additional adjunctive procedures.

Technical success was defined as restoration of patency and flow with less than 50% residual stenosis, and clinical success was defined as significant resolution of lower extremity pain and swelling. Follow-up visits were scheduled at three and six months, and every six months thereafter. At each visit, a patient underwent a clinical evaluation according to a modified Villalta scale (12), and a Doppler ultrasound assessment of the affected venous patency and reflux by a designated experienced vascular ultrasound physician. The presence of five leg symptoms (pain, cramps, heaviness, pruritus and paresthesia) and six objective signs (pretibial edema, skin induration, hyperpigmentation, new venous ectasia, redness and pain during calf compression) was scored. Each of the signs and symptoms were rated as 0 (absent), 1 (mild), 2 (moderate) or 3 (severe). Clinical evaluation outcomes were classified as follows: a total score greater than 14 points or a venous ulcer was defined as severe post-thrombotic syndrome (PTS); 5-14 points was given for mild PTS and less than five points was given for no PTS. Valvular reflux is defined as greater than 0.5 seconds for the valve closure time after distal compression and release using an ultrasonic probe in a non-weight-bearing limb with the patient in the standing position (13). Primary patency was defined as confirmed patency and < 50% restenosis as documented by Doppler ultrasound (14).

Of the 142 eligible patients with an episode of non-acute proximal DVT of the lower extremity, 32 were excluded because of previous ipsilateral DVT (n = 8), pre-existing leg ulcers or venous insufficiency (n = 13), a poor life expectancy (n = 5) or the inability to attend follow-up visits (n = 6). The remaining 110 patients were enrolled this study. The baseline characteristics are reported in Table 1. Written informed consent was obtained from each enrolled patient or the family after the purpose and risk of the treatment were fully explained.

As determined at venography, iliofemoral DVT was present in 94 (84%) of 112 limbs and among these four (4%) limbs had thrombus propagating into the IVC, and femoropopliteal DVT occurred in 18 (16%) limbs. Eighty-two patients received retrieval IVC filter placement (32 OptEase Filters, Cordis, Miami, FL; 50 Aegis Filters, Seercare, Shenzhen, China). After lysis, 63 filters were retrieved (no evidence of clot was observed within the filters) and the other 19 were implanted permanently (filling defects were observed within six filters and the other 13 were abandoned and not retrieving due to economic considerations). Urokinase infusion was postponed and the dosage was adjusted in seven patients due to the fibrinogen progressively dropping below 100 mg/dL. The mean duration of CDT was 196 hours (range: 96-272 hours) and the average total dose of urokinase was 3.5 million IU (range: 1.8-5.7 million IU).

Percutaneous angioplasty with stent placement (Wallstent, Boston Scientific Co., Natick, MA) was necessary to treat residual stenoses and/or short occlusions in the iliac segments that were resistant to CDT in 25 (22%) limbs, and another five received iliac vein PTA alone for economic reasons. After completion of treatment, complete recanalization was achieved in 31 (28%) infusions and partial recanalization was achieved in 59 (53%), while minimal recanalization was achieved in 22 (19%). Thus, 90 (81%) treated limbs achieved significant recanalization (≥ 50% recanalization). In addition, venography indicated well-developed venous collaterals around the partially recanalized veins. Figures 1, 2, 3, 4 illustrate the representative cases.

At discharge from hospital, considerable improvement of the swelling and pain of the affected limb was achieved in 96% (108 of 112) of the limbs. The thigh circumference of the affected limb decreased by 4.0 ± 1.6 cm, while the calf circumference decreased by 3.6 ± 1.8 cm (Table 2); the rate of detumescence was up to 75%. The mean hospital stay was 11.3 days (range: 7-15 days). The mean hospital expense was $9,000 (range: $5,300-11,000).

Overall, there was an immediate technical success rate of 81% (90 of 112) and a clinical success rate of 96% for the treated limbs. The treatment outcomes are listed in Table 3.

During the hospital stay, 14 (13%) patients experienced minor bleeding, of which bleeding at the puncture site occurred in 10 (9%) patients, and four (4%) patients had mucosal bleeding (2 cases of hematuria, 1 vaginal bleeding and 1 nose bleeding). Forty-one (37%) patients complained of slight pain in the affected knees (intramuscular hemorrhage or hemarthrosis was excluded by ultrasound and X-ray examination). None suffered from intracranial hemorrhage, symptomatic pulmonary embolism or other procedure-related complications.

The average duration of follow-up was 15.2 months (range: 3-24 months). During the period, the vein was patent in 74 (67%) limbs, including 23 with stent placement. The overall stent patency was 92% (23 of 25); 37 (32%) limbs showed progression of the stenosis with luminal narrowing more than 50%, including three limbs with rethrombosis, four with PTA alone and two with stent placement, while asymptomatic iliac vein occlusion occurred in one limb with PTA alone. Twenty five (22%) limbs developed mild PTS that mainly presented as heaviness and edema of the affected limbs after activities, and mild pruritus or hyperpigmentation was also present in five limbs, but none of these five limbs had severe PTS. The median total PTS score was three (range: 0-9) and 52 limbs (47%) had a score of 0. Valvular reflux occurred in 24 (21%) limbs (Table 4).