Tone Enden, Department of Hematology, Oslo University Hospital, Ullevål, 0407 Oslo, Norway. Tel.: +47 917 16 584; fax: +47 230 16 211. E-mail: email@example.com
Summary. Background: Approximately one in four patients with acute proximal deep vein thrombosis (DVT) given anticoagulation and compression therapy develop post-thrombotic syndrome (PTS). Accelerated removal of thrombus by thrombolytic agents may increase patency and prevent PTS. Objectives: To assess short-term efficacy of additional catheter-directed thrombolysis (CDT) compared with standard treatment alone. Patients and methods: Open, multicenter, randomized, controlled trial. Patients (18–75 years) with iliofemoral DVT and symptoms < 21 days were randomized to receive additional CDT or standard treatment alone. After 6 months, iliofemoral patency was investigated using duplex ultrasound and air-plethysmography assessed by an investigator blinded to previous treatment. Results: One hundred and three patients (64 men, mean age 52 years) were allocated additional CDT (n = 50) or standard treatment alone (n = 53). After CDT, grade III (complete) lysis was achieved in 24 and grade II (50%–90%) lysis in 20 patients. One patient suffered major bleeding and two had clinically relevant bleeding related to the CDT procedure. After 6 months, iliofemoral patency was found in 32 (64.0%) in the CDT group vs. 19 (35.8%) controls, corresponding to an absolute risk reduction (RR) of 28.2% (95% CI: 9.7%–46.7%; P = 0.004). Venous obstruction was found in 10 (20.0%) in the CDT group vs. 26 (49.1%) controls; absolute RR 29.1% (95% CI: 20.0%–38.0%; P = 0.004). Femoral venous insufficiency did not differ between the two groups. Conclusions: After 6 months, additional CDT increased iliofemoral patency from 36% to 64%. The ongoing long-term follow-up of this study will document whether patency is related to improved functional outcome.
Current guidelines for treatment of deep vein thrombosis (DVT) of the lower limb recommend anticoagulant therapy for the prevention of thrombus extension and recurrence, pulmonary embolism and death (grade IA recommendation) . Daily wear of class II (30 mmHg) elastic compression stockings (ECS) decreases the risk of a chronically reduced functional outcome, that is, post-thrombotic syndrome (PTS). Still one in four will develop PTS within 2 years after the thrombotic event [2,3]. There is therefore a great need for improved therapy that can improve long-term functional outcome. A recent systematic review including 12 studies found that additional thrombolytic therapy for rapid dissolution of thrombotic material in acute DVT may offer advantages in terms of reducing PTS and maintaining venous patency . However, systemic thrombolytic therapy is associated with an unacceptably high risk of bleeding and is not recommended for the treatment of acute DVT . Catheter-directed thrombolysis (CDT) with local delivery of the thrombolytic agent significantly reduces the total dose required for dissolution of a venous thrombus. A case series on CDT revealed a high rate of technically successful thrombolysis at the expense of only a small additional increase in bleeding complications [5–9]. Just one very small randomized controlled trial with short-term follow-up has been performed . Based on this evidence the recent 2008 guidelines of the American College of Chest Physicians were modified to recommend CDT in selected patients (grade IIB recommendation) . However, properly designed multi-center randomized controlled trials adhering to recent reporting standards are required to enable an evidence-based practice for venous thrombolytic therapy [11,12]. Accordingly, the ongoing Catheter-directed Venous Thrombolysis in acute iliofemoral vein thrombosis trial (the CaVenT Study) aims to evaluate safety and efficacy of site-directed thrombolysis. The present study reports on short-term results with emphasis on venous patency after 6 months.
Patients and methods
Patients were recruited from 19 hospitals within the South-Eastern Norway Regional Health Authority, which serves a population of 2.6 million people. Patients aged 18–75 years with a first-time objectively verified iliofemoral DVT (± popliteal and calf vein thrombosis) were eligible for inclusion if symptoms had lasted < 21 days. Complete inclusion and exclusion criteria are presented in Table 1. Prior to treatment allocation, written informed consent was obtained. The study protocol was approved by the Regional Committee for Medical and Health Research Ethics (East) and the Norwegian Medicine Agency.
Table 1. Inclusion and exclusion criteria
Age 18–75 years
Onset of symptoms < 21 days
Objectively verified deep vein thrombosis localized in the upper half of the thigh, the common iliac vein or the combined iliofemoral segment
Anticoagulant therapy prior to trial entry for > 7 days
Contraindications to thrombolytic therapy, including bleeding diathesis
Indications for thrombolytic therapy, for example, phlegmacia coerolia dolens or isolated vena cava thrombosis
Severe anemia (hemoglobin < 8 g dL−1)
Thrombocytopenia (platelets < 80 × 109 L−1)
Severe renal failure – estimated creatinine clearance < 30 mL min−1
Severe hypertension, that is, persistant systolic blood pressure > 160 mmHg or diastolic blood pressure > 100 mmHg
Pregnancy and thrombosis ≤7 days postpartum
Less than 14 days postsurgery or post-trauma
History of subarachnoidal or intracerebral bleeding
Disease with life expectancy < 24 months
Drug abuse or mental disease that may interfere with treatment and follow-up
Former ipsilateral proximal deep vein thrombosis
Malignant disease requiring chemotherapy
Any thrombolytic therapy within 7 days prior to trial inclusion
The study design has previously been published . Patients were randomized to conventional treatment alone or CDT in addition to conventional treatment. A random block allocation sequence for each hospital with stratification for involvement of the pelvic veins and block size of six was generated using the Web site Randomization.com (http://www.randomization.com). Enrolment and treatment assignment were performed by the local trial investigator by picking the lowest number of sealed, opaque-numbered envelopes.
The CaVenT Study was designed with two primary end points: long-term functional efficacy end point was frequency of PTS after 24 months, and short-term descriptive efficacy end point was patency after 6 months . For sample size calculation, we assumed that expected prevalence of PTS after 2 years was 25% in those allocated conventional therapy compared with 10% in those given additional CDT. With a significance level of 5% and a statistical power of 80%, 200 patients (100 in each group) must be included in the study. For short-term efficacy and with the same power, we calculated that at least 38 patients were required in each treatment arm to detect an increase in venous patency from 50% in patients allocated conventional treatment to 80% in those given additional CDT. Using this background, we decided a priori to analyze results on short-term patency on a subpopulation of the first 100 patients with 6 months data.
Antithrombotic and thrombolytic therapy
Anticoagulant therapy was given in accordance with local routines based on international guidelines using low-molecular-weight heparin (LMWH) followed by oral warfarin for at least 6 months with a target international normalized ratio (INR) of 2.0–3.0 . Patients allocated CDT were transferred to the nearest of four interventional centers, and CDT was started on the next workday. Meanwhile the patients received subcutaneous LMWH.
Before the CDT procedure, LMWH was discontinued for at least 8 h, and oral anticoagulants were discontinued to obtain an INR < 1.5. At the start of CDT, an intravenous bolus dose of unfractionated heparin (UFH), 5000 U, was given followed by a continuous intravenous UFH infusion at 15 U kg−1 h−1. The UFH dose was adjusted to keep activated partial thromboplastin time (aPTT) (Cephotest®; Axis-Shield, Oslo, Norway) at 1.2–1.7 times prolongation, that is, at 40–60 s, during CDT.
After applying a local anesthetic, an introducer was inserted into an appropriate vein, preferentially the popliteal vein, guided by ultrasound. At the discretion of the operator, the calf or inguinal veins were other options for venous access. Venography was performed to determine the extent of the thrombus. The wire and catheter were advanced above the proximal part of the thrombus and fitting-sized perfusion catheters, for example, 10, 20, 30, or 50 cm, were positioned. CDT was discontinued if introduction of the catheter through the occluded segment was not successful. Next, 20 mg of alteplase (Actilyse®; Boehringer-Ingelheim, Ingelheim am Rhein, Germany) diluted in 500 mL 0.9% NaCl was infused at 0.01 mg kg−1h−1 with a maximal dose of 20 mg per 24 h and maximal duration of 96 h. Treatment continued in a medical ward. Blood pressure and pulse were recorded and puncture site was inspected four times daily. Hemostasis was also monitored by daily analysis of hemoglobin, fibrinogen and platelet counts, and APTT was monitored twice daily for adjustment of heparin dose.
Thrombolysis was assessed daily by venography using a contrast injection through the introducer and/or perfusion catheter. Each vein segment, that is, IVC, the common iliac vein, the external iliac vein, the common femoral vein, the proximal and distal superficial femoral veins, and the popliteal vein, were given a score, where 0 = open vein, 1 = partly occluded vein, and 2 = completely occluded vein. Total thrombus score before and after lysis was calculated by adding the segmental scores. The difference between the pre- and post-lysis thrombus scores divided by the pre-lysis score gave the grade of thrombolysis; grade I = < 50%; grade II = 50%–90%, and grade III = complete thrombolysis . Duration of CDT and use of adjunctive angioplasty and stents to establish flow and obtain < 50% residual stenosis were at the discretion of the operator.
During CDT, concomitant use of antithrombotic agents other than UFH was discontinued. Overt bleeding and symptoms suspect of bleeding or pulmonary embolism were dealt with according to local routines. Any complication of clinical relevance was reported to the Safety and Monitoring Committee. There were no established stopping rules for the study. Bleeding was classified as major if it was overt with a decrease of ≥ 2 g dL−1 hemoglobin, led to transfusion of ≥ 2 units packed red blood cells, was retroperitoneal, intracranial, or in a critical organ, or if it contributed to death. Clinically relevant non-major bleeding included, for example, intervention for epistaxis, a visible large hematoma, or spontaneous macroscopic hematuria . All other hemorrhages were categorized as trivial.
A weight-adjusted full therapeutic dose of subcutaneous LMWH given twice daily was initiated 1 h after removal of catheters. Oral anticoagulation was then established as described earlier and according to guidelines.
Immediately after CDT or otherwise as soon as possible after entering the study, all patients received knee-high ECS class II and were advised daily wear for 24 months.
Assessment at 6 months follow-up
Short-term end-point evaluation by non-invasive assessment of the veins of all patients in both treatment groups was performed after 6 months ± 2 weeks by one angiologist (C.E.S.), who was unaware of the patient’s treatment allocation and medical history. To further secure an unbiased evaluation of outcomes, the patients were explicitly informed not to reveal which treatment they had received . There was no adjudication committee. Uncertainties were agreed by consensus among three of the investigators (N.E.K., P.M.S. and T.E.).
The venous system was evaluated using ultrasound and air-plethysmography. Gray-scale ultrasound was used for detection of post-thrombotic iliofemoral wall-thickening and intraluminal hyperechoic structures. Compression ultrasound assessed compressibility of the femoral vein. Doppler ultrasound was used to evaluate iliofemoral venous flow and femoral venous insufficiency. Insufficiency was evaluated with the patient in the standing position, and reflux was defined as reversal of the velocity curve lasting longer than 0.5 s after standardized distal pneumatic decompression . Functional obstruction of the veins was assessed using air plethysmography .
Venous patency was defined as a composite outcome measure where patients having any of the following were classified as not having regained iliofemoral venous patency: partial or complete incompressibility of the femoral vein, no flow in pelvic or femoral vein and/or functional venous obstruction. Patients with duplicate femoral veins with normal compressibility and flow in at least one course and without functional obstruction were considered successfully recanalized.
When comparing dichotomous variables in the two treatment groups, a two-sided χ2-test was used. When comparing continuous variables, a two-sided Mann–Whitney test was used. Correlations were calculated using Pearson’s product–moment coefficient. Findings with P-values < 0.05 were considered statistically significant. The statistical analysis was performed using the statistical package spss, version 15 (SPSS Inc., Chicago, IL, USA).
Role of the funding source
Study sponsors had no involvement in study design, collection, analyses, and interpretation of data, or writing and submission for publication.
Figure 1 shows the trial profile of the 118 patients recruited and randomized between January 2006 and January 2008. Patients’ baseline characteristics of each treatment group are summarized in Table 2. Based on enrolment rate complete recruitment of 200 patients in the CaVenT Study is expected late 2009.
Table 2. Characteristics of treatment groups
Catheter-directed thrombolysis group (n = 50)
Standard treatment group (n = 53)
Data are mean (SD) or n (%). *Based on routine diagnostic imaging before recruitment. †Ipsilateral proximal deep vein thrombosis exclusion criterion. ‡One homozygous factor V Leiden. §Protein S combined with homozygous prothrombin G20210A and prothrombin G20210A combined with lupus antiocoagulant (both in CDT group), protein C combined with factor V Leiden and protein S combined with prothrombin G20210A (both in standard treatment group).
Left-sided deep vein thrombosis
Femoral deep vein thrombosis*
Iliofemoral deep vein thrombosis*
Duration of symptoms (days)
Duplicate femoral vein
Recurrent venous thrombosis
D-dimer at 6 months (mg L−1)
Daily wear of compression stockings class II
Comorbidity of the lower limb
Familial disposition of various veins
Transient risk factors for venous thrombosis
Surgery previous 3 months
Trauma previous 3 months
Infection in the previous 6 weeks
Pregnancy in the previous 3 months
Hormonal replacement therapy
Oral contraceptive pill
Permanent risk factors for venous thrombosis
Previous venous thrombosis†
Previously known malignancy at baseline
Inflammatory bowel disease
First degree relative with venous thrombosis
Factor V Leiden
Factor V Leiden and prothrombin G20210A
Protein C deficiency
Protein S deficiency
Any other combination of the above§
Four patients in the CDT arm withdrew from the study before CDT was initiated. After randomization, four patients in ther CDT group were found not to satisfy eligibility criteria; two patients had exclusion criterion, that is, one alcohol abuse with bleeding tendency and one previous traumatic cerebral bleeding; and two did not have any signs of acute thrombosis on venography during CDT procedure, consequently not satisfying inclusion criteria and representing false–positive findings on routine compression ultrasound. These eight patients were excluded from the analyses.
Missing outcome data on seven more participants were not included in the primary analysis as these data were assumed to be missing independent of treatment received; one patient in the CDT group was not reached, one control withdrew from the study, three controls died of cancer (of whom two were diagnosed during follow-up), and two controls had inconclusive assessments of the veins because of obesity. The statistical analyses were performed on 103 patients; 53 in the control group received standard treatment alone, whereas 50 received interventional therapy, including one procedure without establishment of thrombolytic infusion as a result of technical failure from agenesis of the vena cava inferior. One patient did not receive CDT at the operator’s discretion owing to DVT in the distal femoral vein only at the initiation of CDT procedure. These two patients resulted in ineffective (grade I) lysis.
In the CDT group, 50%–90% lysis (grade II lysis) was achieved in 20 patients, and complete lysis (grade III) in 24 patients. Lysis grade did not correlate with duration of symptoms (results not shown). Two patients received CDT for 6 days; both resulted in grade II lysis. Among the four CDT procedures resulting in grade I lysis, two were prematurely ended because of bleeding complications; one developed compartment syndrome of the calf, and one encountered clinically relevant bleeding from the inguinal puncture site. Mean duration of CDT was 2.3 (SD 1.2) days. Angioplasty of the iliofemoral vein segment with balloon only or additional stenting was performed in 13 and eight patients (of whom two were femoral stents), respectively.
Hemostasis monitoring by daily analysis of hemoglobin, fibrinogen, D-dimer, INR and platelet count did not reveal or indicate occult bleeding in any of the patients undergoing CDT, or lead to modification of the therapy. A total of 10 overt bleeding complications were reported in relation to the 49 CDT procedures. Major complications were reported in two patients; one required surgery because of compartment syndrome of the calf after local bleeding, and one patient contracted durable and partial impairment of sensibility of the foot immediately after the procedure. Two patients suffered clinically relevant non-major bleeding; one developed an inguinal hematoma of approximately 100 cm2 in relation to an inguinal puncture site, and one developed a > 100 cm2 subcutaneous hematoma of the abdomen as a result of erroneous administration of heparin (five times prescribed dosage). Minor complications without clinical relevance occurred in five patients with bleeding from the puncture site, one patient suffered epistaxis lasting 3 days, and one developed a small subcutaneous hematoma of the shoulder. There were no pulmonary embolizations or deaths related to CDT.
Two patients experienced bleeding complications related to antiocagulation during follow-up; one female reported an episode of gynecological bleeding and one male needed surgical evacuation of a hematoma in relation to a muscle rupture of the calf after 5 months on warfarin. None of these had received CDT. There were no bleeding complications during acute standard treatment.
Efficacy outcomes at 6 months
Patency of the iliofemoral vein segment was found in 32 (64.0%) patients in the CDT group and 19 (35.8%) in the control group, corresponding to an absolute risk reduction of 28.2% (95% CI: 9.7%–46.7%; P = 0.004). When the same analysis was performed on all 118 randomized patients assuming no patency in patients with missing data, the difference between the treatment groups remained statistically significant (P = 0.026).
Venous obstruction as assessed by air plethysmography was found in 10 (20.0%) patients in the CDT group and in 26 (49.1%) controls, corresponding to an absolute risk reduction of 29.1% (95% CI: 20.0%–38.0%; P = 0.004). The two other subcategories of venous patency, that is, compressibility and flow, and other post-thrombotic changes of the iliofemoral veins, that is, wall thickening and echoic content of vein lumen, did not differ significantly between the two treatment arms (Table 3). Femoral venous insufficiency was found in 35 (66.0%) and 30 (60.0%) limbs among controls and in the CDT group, respectively (P = 0.53). Duration of femoral reflux time did not differ significantly between the two groups (P = 0.94).
Table 3. Non-invasive assessment of veins 6 months after iliofemoral deep vein thrombosis (n = 103)
Catheter-directed thrombolysis (n = 50) N (%)
Standard treatment (n = 53) N (%)
*Single or duplicate femoral vein.
Functional venous obstruction
Femoral venous insufficiency*
Other post-thrombotic changes
Among the 50 patients in the CDT arm, lysis grade or use of angioplasty/stent did not significantly correlate with 6-months patency. Five out of six patients with ineffective lysis (grade I) and approximately 60% of the patients with effective lysis (grade II and III) had regained iliofemoral patency. After CDT without adjunctive endovascular therapy, 17 out of 29 patients regained patency (Table 4). In patients who received adjunctive angioplasty or angioplasty with stent, patency was regained in nine out of 13, and six out of eight, respectively. Table 4 also shows lysis grade for CDT with and without angioplasty.
Table 4. Immediate lysis grade and 6-months patency after catheter-directed thrombolysis (CDT) (n = 50)
Immediate lysis grade* after CDT N
Patency after 6 months N (%)
*Grade I, <50% lysis; grade II, 50%–90% lysis; grade III, complete lysis. One patient with grade III lost to 6 months follow-up was not included. †Including one technical failure and one that did not receive thrombolytic infusion.
CDT only (n = 29)
CDT and balloon angioplasty (n = 13)
CDT, balloon angioplasty, and stent (n = 8)
The ongoing CaVenT Study is the first randomized controlled trial to evaluate relevant clinical effects of additional CDT in acute DVT of the lower limb, and additional CDT increased 6-months patency from 36% to 64% as compared with standard anticoagulant treatment and use of ECS.
It is believed that PTS evolves from continued venous hypertension as a result of persisting venous obstruction and/or venous insufficiency after inflammatory destruction of valves in response to acute DVT . From observational studies, it has been shown that repeated DVT of the limb is associated with PTS development [12,19]. Other factors, that is, characteristics and symptoms of index DVT, thrombophilia, overweight and gender, have been shown to be inconsistently associated with an increased risk of PTS [12,19]. Early recanalization is suggested to protect against valve destruction and development of PTS, but the relationship between early recanalization and long-term functional outcome of the limb has yet to be documented. We found that CDT reduced functional venous obstruction from 49% to 20%, whereas venous insufficiency was not significantly reduced and occurred in 60% after CDT and 66% controls. Different cut-off values have been used for ultrasonographic assessment of venous reflux, however, when increasing the cut-off to 1.0 s as suggested by Labropoulos and colleagues , still 52% and 51% in CDT and control group, respectively, had insufficiency (P = 0.92). This may indicate that if CDT reduces development of PTS this is because of a reduction in obstruction, but not insufficiency.
Among the 12 studies in the systematic review of Watson and colleagues  on thrombolysis for acute DVT, only six reported on patency after 6 months, including one study on local administration of a thrombolytic agent. The six studies combined indicated patency in 43% after thrombolytic therapy vs. 17% in the control group, and the trial of Elsharawy and colleagues  on 35 patients reported complete lysis in 72% after CDT vs. 12% in the control group. Other studies have found 6–12 months patency in 38%–50% after standard treatment . Three case series on CDT reported 6 months patency to be 83%, 85% and 100% [22–24]. These efficacy estimates are consistent with our findings.
The study of Elsharawy differed from the CaVenT study in some aspects; CDT included streptokinase and pulse-spray administration, reflux was assessed using plethysmography, and assessment of patency involved ultrasound and/or venography. Duration of anticoagulant therapy or use of ECS was not reported, and it is unclear how venous outflow obstruction related to patency. Our study allowed patients with up to 21-days history of DVT symptoms compared with < 10 days, however, mean duration of symptoms in patients receiving CDT was 6 days in both studies. The apparent smaller effect compared with Elsharawy’s study regarding the secondary efficacy estimates on venous obstruction and insufficiency, are likely to be explained by the relatively small sample sizes of the studies. In the CaVenT study, approximately 60% of the patients developed venous insufficiency after 6 months, and this is in line with previous reports [25–27].
The CDT procedure was successful in the great majority of patients; complete lysis was achieved in 48%, and 50%–90% lysis in 40% of the patients. This is consistent with previous reports [5–9]. One episode of major bleeding was seen related to the CDT procedure, indicating that safety was at least as good as previously reported. Bleeding complications are usually related to puncture of the vein at the start of CDT procedure, and are more likely to occur if the initial puncture did not secure venous access. This was the likely explanation of calf hematoma leading to the development of compartment syndrome. The mechanism explaining the partly, but durable sensibility impairment of the foot in one patient is unclear, but may relate to nerve damage during the procedure.
The CaVenT study is a properly designed, multicenter, randomized, controlled trial. However, the non-blinded design is unavoidably open to challenge for bias. Another limitation of our study is the lack of information on number and characteristics of patients screened for eligibility as this would give additional and valuable information on the generalizability of the results. Based on the inclusion and exclusion criteria of the CaVenT study, we believe that the recruited population is representative for patients potentially benefitting from CDT. The relatively slow recruitment rate may indicate selection bias, however, we have not been able to identify such, and no eligible patients received CDT during the inclusion period if not included in the CaVenT study.
Inherent in the CDT procedure is venography performed initially to map the extent of thrombi, and this allows for a more detailed and possibly more reliable visualization of the thrombi compared with routine imaging with ultrasound. Consequently, after randomization, two patients in the CDT arm were found not to fulfill the inclusion criteria of iliofemoral DVT. Two more patients were found to possess exclusion criteria, and this was probably because of additional clinical routines regarding contraindications to thrombolysis. This may introduce bias and a systematic difference between the treatment groups as similar patients may not have been found ineligible if allocated standard treatment. Another difference that may lead to bias between the two groups is the unfortunate fact that four patients decided to withdraw from the study after randomization to CDT, compared with only one in the control group. Possible explanations may be inadequate information at recruitment, reluctance to transfer to another hospital, or early improvement of symptoms with LMWH therapy while awaiting CDT start-up. We decided not to include these patients in the primary analysis as they were lost to follow-up before receiving treatment.
A matter of discussion is the handling of the missing data in the statistical analyses. The difference in patency between the treatment groups remained significant when we included all randomized patients by giving patients with missing data the outcome ‘no patency’ in the analyses. This finding indicates that our conclusion as regards the difference in patency cannot be explained by missing data.
We are also aware of the potential weakness of using envelope-based randomization as misconduct among investigators has been reported . To our knowledge there was no such misconduct in this study. The choice of randomization method was a matter of trial costs; computer-based randomization is hardly prone to misconduct, but also substantially more costly.
In conclusion, additional CDT increased patency 6 months after iliofemoral DVT, from 36% to 64%. Venous obstruction, but not venous insufficiency, was significantly reduced in the CDT group. The clinical relevance of these findings with regard to development of PTS is still unknown, but future long-term data from the CaVenT study will provide evidence on the functional outcome after CDT.
T.E.: study design, data collection, data analysis and writing of manuscript. N.-E.K.: original idea, study design, data analysis and manuscript review. L.S.: study design, statistical analysis and manuscript review. C.-E.S.: study design, data collection, and manuscript review. W.G.: study design, data collection and manuscript review. G.H.: study design, data collection and manuscript review. P.A.H.: study design, data collection and manuscript review. L.O.H.: study design, data collection and manuscript review. A.M.N.: study design, data collection and manuscript review. G.S.: study design, data collection and manuscript review. P.M.S.: original idea, study design, data analysis and manuscript review.
We would like to thank the patients who by agreeing to participate made the study possible. Thanks to the following physicians in South-Eastern Norway Regional Health Authority for invaluable contributions in recruitment of patients and collection of data: W. Aasebø, G. K. Bekken, Y. Benestad, Ø. Bukten, J. Dalgaard, O. Hagen, T. O. Isaksen, M. Kalbakken, E. Müller, M. Myrstad, E. Nyquist, H. Pihlstrøm, J. Rolke, V. Stenberg, A. Tveit, and P. O. Vandvik. Thanks for great help from the study nurses S. Foyn, K. Hulbekkmo og T. N. Moan. The original trial was registered at ClinicalTrials.gov (NCT 00251771). The study was supported by a grant from Eastern Norway Regional Health Authority Trust (fellowship to TE), the Research Council of Norway (grant no. 175465/V50), and the University of Oslo. Funders had no role in study design, data collection, analysis, interpretation, or writing and submission of article.
Disclosure of Conflict of Interests
The authors state that they have no conflict of interest.