A Pilot Randomized Double-blind Comparison of a Low-Molecular-Weight Heparin, a Nonsteroidal Anti-inflammatory Agent, and Placebo in the Treatment of Superficial Vein Thrombosis FREE

SUPERFICIAL VEIN thrombosis of the lower limbs is a common disease, usually considered benign. However, studies^1- 13 suggest that the frequency of association between superficial and deep vein thromboses ranges from 3% to 40%. Furthermore, pulmonary embolism has been clinically suspected to occur in 2% to 13% of patients with superficial vein thrombosis^2- 3,8,14 and in 20% to 33% of patients when it has been checked by lung scanning.^4,15

Systematic duplex ultrasonography has been proposed in the initial management of superficial vein thrombosis, to detect the presence of any underlying deep vein thrombosis,^{6,8- 9,11,13,16- 18} enabling prompt anticoagulant therapy. When superficial vein thrombosis occurs alone, treatment is generally ambulatory; various therapies have been proposed, such as active ambulation, graduated compression stockings, nonsteroidal anti-inflammatory agents, antithrombotic therapy, and, in selected cases, surgical intervention.^{1- 3,13,17- 22} However, none of these treatments has been evaluated in adequately designed trials.

Because superficial vein thrombosis may extend into the deep venous system and potentially engender pulmonary embolism, antithrombotic therapy might be the best choice. Low-molecular-weight heparins are the reference treatment in the preventive²³ and curative²⁴ management of venous thromboembolism, even in patients at a high risk of recurrence,²⁵ taking a subcutaneous once- or twice-daily dosage regimen.^18,26 Owing to their anti-inflammatory and antiplatelet properties, oral nonsteroidal anti-inflammatory agents may also be useful in treating superficial vein thrombosis.^18,22

We, therefore, performed a double-blind randomized trial to determine the frequency of venous thromboembolic complications in patients with superficial vein thrombosis and to evaluate, compared with placebo, the efficacy and safety of low-molecular-weight heparin and of a nonsteroidal anti-inflammatory agent in preventing these complications.

Hospitalized or nonhospitalized patients older than 18 years, weighing 45 to 110 kg, with acute superficial vein thrombosis of the lower limbs at least 5 cm long on duplex ultrasonography were considered for inclusion in the study.

Patients were excluded if they presented 2 or more superficial vein thromboses or superficial vein thrombosis following sclerotherapy, deep vein thrombosis on initial duplex ultrasonography, or documented pulmonary embolism. Moreover, women of childbearing age were excluded if they were pregnant, breastfeeding, or not using contraception. Other reasons for exclusion were known thrombophilia, uncontrolled arterial hypertension (systolic blood pressure >180 mm Hg, diastolic blood pressure >110 mm Hg, or both), previous or active peptic ulcer, bacterial endocarditis, stroke within the previous 3 months, other conditions favoring hemorrhage, history of hypersensitivity to heparins, heparin-induced thrombocytopenia, hypersensitivity to paracetamol or nonsteroidal anti-inflammatory agents, serum creatinine concentration above 1.81 mg/dL (>160 µmol/L), platelet count below 100 × 10³/µL, prothrombin ratio below 60%, or contraindication to elastic bandages or support stockings. Finally, patients who required anticoagulant therapy, ligation of the saphenofemoral junction, or thrombectomy and those who received any type of anticoagulant therapy or nonsteroidal anti-inflammatory agents for more than 48 hours were excluded.

Patients were centrally randomly assigned to receive, for 8 to 12 days, one of the following treatments: subcutaneous injections of enoxaparin sodium, 40 mg; enoxaparin, 1.5 mg/kg body weight (Lovenox, Clexane, or Klexane; Laboratoire Aventis, Paris, France); oral tenoxicam; or placebo, once daily, beginning within 24 hours after randomization. Follow-up visits were scheduled at day 10 ± 2 (ie, between days 8 and 12) and at 3 months (ie, between days 75 and 97). During follow-up, patients were instructed to report any symptoms or signs of venous thromboembolism or any other clinical event occurring since treatment completion. Investigators were encouraged to discontinue any antithrombotic treatment during follow-up in all patients without a thromboembolic complication between days 1 and 12, whereas treatment was left to the investigator's discretion if such an event had occurred during the study.

The study was conducted according to the ethical principles stated in the Declaration of Helsinki and local regulations. The protocol was approved by an independent ethics committee, and written informed consent was obtained from all patients before randomization.

This was a double-blind, double-dummy, controlled trial. Study medications were packaged in boxes of identical appearance, one for each patient. All boxes had visually identical contents: 12 prefilled single-dose syringes; 12 ampules, with syringes and needles; and 12 scored tablets (ie, 1 active treatment and matching placebos of the 2 others for patients assigned to receive 1 of the 3 active treatments and placebos only for patients assigned to receive placebo). Furthermore, each treatment box contained open-label paracetamol to be used in case of pain. The active treatments were 40 mg of enoxaparin in 0.4 mL of water for injectable preparations per syringe (equivalent to 4000 IU of anti–factor Xa activity), 1.5 mg/kg of enoxaparin (equivalent to 150 IU of anti–factor Xa activity per kilogram; using a concentration with 200 mg in 1.0 mL per ampule), or 20 mg of tenoxicam per tablet. In outpatients, injections were performed by nurses.

In all patients, treatment with elastic bandages or support stockings was initiated on day 1 of therapy and continued for at least 15 days. The patient was free to take paracetamol as needed. Throughout the treatment period, dextran; mannitol; thrombolytic treatment; any other anticoagulant, antiplatelet, or nonsteroidal anti-inflammatory agents; lithium carbonate and lithium gluconate therapy; and methotrexate were prohibited.

The primary efficacy outcome was deep venous thromboembolism (defined as deep vein thrombosis, pulmonary embolism, or both) between days 1 and 12. The secondary efficacy outcomes were deep venous thromboembolism between days 1 and 97 and the combined end points of deep venous thromboembolism and superficial vein thrombosis recurrence or extension toward the saphenofemoral junction between days 1 and 12 and between days 1 and 97. Patients were examined for deep and superficial vein thromboses by systematic duplex ultrasonography of the legs between days 8 and 12, or whenever thrombosis was clinically suspected. In case of doubt, the diagnosis of deep vein thrombosis was confirmed by phlebography. A suspected pulmonary embolism was confirmed by high-probability lung scanning, pulmonary angiography,²⁷ or helical computed tomography, or at autopsy. Recurrent superficial vein thrombosis was defined as a new episode in any other superficial venous location or at the same location as the initial episode, provided that the disappearance of the previous thrombus had been proved by duplex ultrasonography. Extension of the initial superficial vein thrombosis was diagnosed by duplex ultrasonography if there was a 2-cm extension or more of the initial thrombus toward the saphenofemoral junction within the superficial venous system. All duplex ultrasonographic examinations were performed according to a predefined standardized procedure, including visualization of the saphenous veins and 21 sites of the deep venous system.^16,28

Safety outcomes were death, major and minor hemorrhage, thrombocytopenia, and any other adverse event. Hemorrhage was classified as major when it was overt and necessitated transfusion of 2 U or more of packed red blood cells or whole blood, when the hemoglobin concentration decreased by 2 g/dL or more compared with baseline, or if bleeding was retroperitoneal, intracranial, or fatal or led to permanent study treatment discontinuation. Hemorrhage was defined as minor if it was overt but did not meet the criteria for major hemorrhage. Thrombocytopenia was defined as a decrease in platelet count of at least 50% from baseline or a platelet count below 100 × 10³/µL. Outcomes were reviewed blindly by an independent critical event committee.

To maintain an overall 2-sided significance level of .05 in the analysis of the primary outcome, the nominal significance level in 3 comparisons with placebo was adjusted according to the Bonferroni method (α = .015). Assuming a venous thromboembolism incidence by day 12 of 20% in the placebo group and 10% in 1 of the 3 other treatment groups and an overall power of 85%, 315 patients were needed in each group (total, 1260 patients). However, the study was interrupted when, after the enrollment of 427 patients over 3 years, the steering committee, unaware of the results, decided that the slow recruitment rate was not compatible with continuation of the study.

All analyses were performed on an intention-to-treat basis by comparing each of the active treatment groups with the placebo group. Patients with missing duplex ultrasonographic data between days 8 and 12 were not evaluable for analysis of the primary outcome.

The 98.5% confidence interval of the difference between each of the 3 active treatment groups and the placebo group was calculated, with normal approximation to a binomial distribution. A 2-tailed P value below .015 was considered statistically significant. We used a 2-sided χ² test or the Fisher exact test (where appropriate) for qualitative variables and the t test or comparison of medians for quantitative variables.

Between March 6, 1996, and December 15, 1998, 436 patients were enrolled in 49 centers. The logs, available from 35 centers (in which 325 patients were enrolled), recorded 867 patients meeting the inclusion criteria. Of these patients, 513 (59.2%) were excluded because of associated deep vein thrombosis or pulmonary embolism (n = 120), an indication for venous surgery (n = 116), prohibited treatment before inclusion (n = 110), impracticable follow-up (n = 62), contraindications to study treatments (n = 41), 2 or more superficial vein thromboses or superficial vein thrombosis following sclerotherapy (n = 31), or other reasons (n = 33). Of the 354 eligible patients, 29 declined to participate.

Of the 436 patients enrolled, 7 withdrew their consent, 1 in the tenoxicam group and 2 in each of the 3 other groups. Moreover, data from one center, which enrolled 2 patients (1 each in the placebo and tenoxicam groups), were excluded from analysis by the steering committee because access to source documents was not possible. Thus, 427 patients were included in the intention-to-treat population. Baseline characteristics did not differ significantly between the placebo group and any active treatment group (Table 1). Overall, 97.7% of the patients had at least 1 risk factor for venous thromboembolism, 73.8% had at least 2 risk factors, and 71.7% had moderate or severe chronic venous insufficiency, according to the classification of Porter et al.²⁹

All patients received the study treatment outside the hospital. Systematic duplex ultrasonography was not feasible between days 8 and 12 in 11 patients. All but 1 patient (in the placebo group) underwent follow-up at 3 months, 412 in person and 14 by telephone.

Results on venous thromboembolism are detailed in Table 2, Table 3, and Table 4. In the placebo group, the incidence of deep venous thromboembolism by day 12 was 3.6% (4 of 111 patients). The incidence was lower in the other 3 groups: 40-mg enoxaparin, 0.9% (1 of 109 patients); 1.5-mg/kg enoxaparin, 1.0% (1 of 102 patients); and tenoxicam, 2.1% (2 of 94 patients). However, the differences between each of the active treatment groups and the placebo group were not statistically significant (P = .37, .37, and .39, respectively [Table 4]). The incidence of deep and superficial venous thromboembolism by day 12 was significantly reduced in all active treatment groups, from 30.6% (34 of 111 patients) in the placebo group to 8.3% (9 of 109 patients) in the 40-mg enoxaparin group (P<.001), 6.9% (7 of 102 patients) in the 1.5-mg/kg enoxaparin group (P<.001), and 14.9% (14 of 94 patients) in the tenoxicam group (P<.01). At the 3-month follow-up, the trend in favor of the active treatment groups with respect to deep venous thromboembolism disappeared, whereas the incidence of deep and superficial venous thromboembolism was still significantly reduced (40-mg enoxaparin, P = .005; 1.5-mg/kg enoxaparin, P = .003; and tenoxicam, P = .005 [Table 4]). Figure 1 shows that most venous thromboembolic events occurred between days 1 and 12 for those taking placebo and tenoxicam, but mainly between days 12 and 97 for those taking enoxaparin, after its discontinuation.

The mean ± SD duration of study treatment was 9.6 ± 1.9 days, and did not differ between the groups; 404 (94.6%) of the patients received treatment for the scheduled duration of 8 to 12 days. All patients with deep venous thromboembolism by day 12 subsequently received antithrombotic therapy. Among the 56 patients with isolated superficial vein thrombosis recurrence or extension toward the saphenofemoral junction by day 12, 35 (62.5%) subsequently received such treatment. None of these 35 patients presented any symptomatic deep venous thromboembolism during follow-up, whereas 5 (23.8%) of the remaining 21 untreated patients experienced such an event.

The percentage of patients taking paracetamol was 48.6% (54 of 111 patients) in the placebo group, 50.5% (55 of 109 patients) in the 40-mg enoxaparin group, 48.6% (51 of 105 patients) in the 1.5-mg/kg enoxaparin group, and 42.3% (41 of 97 patients) in the tenoxicam group.

No death, major hemorrhage, or heparin-induced thrombocytopenia occurred during the study. Two patients had a minor hemorrhage several weeks after discontinuation of study treatment, 1 each in the placebo and the 1.5-mg/kg enoxaparin groups. Only 1 patient, in the 1.5-mg/kg enoxaparin group, presented with asymptomatic thrombocytopenia on day 5, which resolved spontaneously while study treatment was maintained, which was not considered a heparin-induced thrombocytopenia.

There were no significant differences between any active treatment group and the placebo group in the incidence of any other adverse event during treatment or follow-up.

Our study suggests that superficial vein thrombosis may not be as benign as initially thought. Indeed, the 4.5% incidence of deep venous thromboembolism in the placebo group at 3 months falls within the range reported in previous smaller open studies.^1- 15 This 4.5% incidence is probably underestimated because patients with associated deep vein thrombosis or pulmonary embolism at enrollment were excluded, a situation found in 120 (13.8%) of the 867 patients screened, as shown by the analysis of the study log. Moreover, most patients with venous thromboembolism by day 12, whether symptomatic or not, received curative doses of anticoagulant therapy, which could have prevented the occurrence of symptomatic deep venous thromboembolism during follow-up. The use of a placebo group was considered ethically justifiable because the risk of venous thromboembolism in such patients was not precisely known and there was no established method of treatment for superficial vein thrombosis. Two previous studies,^30- 31 one published after the start of our study, reported the beneficial effect of an antithrombotic treatment in patients with superficial vein thrombosis, but both have various methodological limitations and their results cannot be considered definitive.

Overall, the benefit of active antithrombotic treatments over placebo observed in our study is in agreement with previous data.³¹ There was a clinically meaningful reduction of the incidence of deep venous thromboembolism in the 3 active treatment groups compared with the placebo group. However, these reductions were not statistically significant, possibly due to a lack of power. The statistically significant reduction of the incidence of deep and superficial events by day 12 achieved with 40-mg and 1.5-mg/kg enoxaparin and tenoxicam, compared with placebo, was mainly driven by their effect on the occurrence of superficial vein events (ie, the recurrence of superficial vein thrombosis and its extension toward the saphenofemoral junction). The clinical relevance of using ultrasonographically detected superficial vein thrombosis recurrence or extension toward the saphenofemoral junction as an end point may be questioned. However, the 27% (30 of 112 patients [Table 2]) incidence of these events found in the placebo group of our study at day 12 is of the same order of magnitude as the 40% incidence reported at 3 months in a previous study,³¹ suggesting that detection by ultrasonography is reproducible. Moreover, in our study, the outcome of patients with such events differed according to whether they were treated for this event. Indeed, 23.8% of the patients left untreated after superficial vein thrombosis recurrence or extension toward the saphenofemoral junction by day 12 had symptomatic deep venous thromboembolism during follow-up, compared with none when antithrombotic treatment was instituted. Finally, superficial vein thrombosis is painful, and we believe that reducing recurrence or extension is a valid therapeutic goal. There was no difference with respect to day 12 outcomes between the active treatment groups. However, there was a trend in favor of the low-molecular-weight heparin relative to the nonsteroidal anti-inflammatory agent, as reported in a previous study,³⁰ and the curative dosage regimen of enoxaparin did not seem to bring additional benefit over its preventive dosage regimen. Conversely, there was an expected trend in favor of tenoxicam regarding the number of patients taking paracetamol.

Because most patients with symptomatic or asymptomatic deep and superficial venous thromboembolism by day 12 received antithrombotic treatment, results of efficacy at 3 months are difficult to interpret. Possible modification of the natural history of the disease may at least partly account for the disappearance of the favorable trends observed in the active treatment groups regarding the incidence of deep venous thromboembolism at day 97. Interestingly, most symptomatic thromboembolic events occurred after the end of study treatment in the enoxaparin groups, and a rebound phenomenon cannot be ruled out. Another explanation may be that the duration of treatment was too short for some patients and that treatment was discontinued when they were still at risk for venous thromboembolism. A duration of 8 to 12 days was chosen to match the usual duration of treatment of superficial vein thrombosis in France.²¹ Future studies could examine the benefit of prolonging treatment for 1 month in these patients, because in our study most thromboembolic complications occurred within 30 days of the initial diagnosis.

Overall, these preliminary results indicate that patients with superficial vein thrombosis of the lower limbs are at significant risk for deep and superficial venous thromboembolic complications. In the absence of a statistically significant reduction in the incidence of deep venous thromboembolic complications with a low-molecular-weight heparin or with an oral nonsteroidal anti-inflammatory agent, simple surveillance may remain the treatment of choice for superficial vein thrombosis. However, treatment with a low-molecular-weight heparin or with an oral nonsteroidal anti-inflammatory agent tends to reduce the incidence of deep venous thromboembolic complications, and is effective and safe in reducing the incidence of recurrence or the extension of superficial venous thrombosis. A low-molecular-weight heparin seems more effective than nonsteroidal anti-inflammatory agents, and a curative dosage regimen of a low-molecular-weight heparin does not seem to bring additional benefit over a preventive dosage regimen. Further investigations in larger patient populations are warranted to confirm these results and to examine whether a longer duration of antithrombotic treatment would reduce the incidence of deep venous thromboembolism.

Corresponding author and reprints: Hervé Decousus, MD, Clinical Pharmacological Unit, Hôpital Bellevue, Pavillon 5, 25 Boulevard Pasteur, 42055 Saint-Etienne CEDEX, France (e-mail: decousus@univ-st-etienne.fr).

Accepted for publication September 30, 2002.

This study was supported by a grant from Laboratoire Aventis.

Steering Committee

C. Boissier (study chair); J. M. Baud; S. Couzan; H. Decousus, MD; J. P. Favre; J. L. Guilmot; P. Ill; C. Janbon; G. Lance; A. Leizorovicz; J. M. Mollard; S. Perrot.

Writing Committee

H. Décousus, MD (chair); M.-F. Brégeault; J.-Y.Darmon; J.-M. Diamand; S. Laporte; A. Leizorovicz; P. Mismetti; S.Quenet.

Ultrasonography Quality Control Committee

A. Elias; J.-M. Mollard; S. Perrot.

Data Monitoring Committee

J. Ninet (chair); H. Bounameaux; I. Juhan-Vague.

Critical Event Committee

P. Carpentier (chair); P. Lacroix; F. Parent.

Coordinating and Statistical Analysis Center

Clinical Pharmacology Unit, Hôpital Bellevue, Saint-Etienne: H. Décousus, MD (director); F. Chambon, A.Devillard, P. Ducrozet, D. Leynadier, V. Souvignet (project management); S. Quenet, S. Laporte (statisticians); L. Arhès, A.Buchmüller, C. Dabrigeon, S. Lomazzi, C. Michalon, A. Picard (clinical research associates); C. Bernabé, E. Venet, C. Chauvet (secretaries).

Data Management Center

Clinical Pharmacology Unit, Hôpital Neuro-Cardiologique, Lyon: A. Leizorovicz (project director); S. Chabaud (statistician); F. Lorenzelli, N. Visèle (informatics); E. Gauthier (trial coordinator); A. Chérief (secretary).

Sponsor

Laboratoires Aventis, Paris: M.-F. Brégeault, S. Brugier, J.-Y. Darmon, P. Ill, N. Yvelin (project managment); S.Dorey, M. Saulière (clinical research associates); Tania Zamor (secretary).

Cosponsor

Association Française de Formation Continue en Angiologie, Paris.

Investigators

Grenoble: J.-M. Diamand; O. Pichot. Saint-Etienne: S. Couzan.Vertou: J.-F. Roux. Saint-Etienne: P. Patural. Echirolles: B. Roger. Limoges: P. Auboiroux. Malemort-en-Corrtèze: C.Fréchinos. Joué les Tours: A. Diard. Istres: A.-M. Saadoun. Le Mans: S. Eybalin; S. Leger-Markabi. Rézé: G. Fauvel; P. Gérard. Saint-Etienne: P. Conchonnet; L. Cretin; Y. Roussin. Arras: S. Aquilanti. Paris: J.-L. Gérard.Vénissieux: P. Paris; F. Vieville. La Ferté Bernard: J.-F. Londe. Chambéry: I. Tourvielle. Amboise: B. Blondeau. Toulouse: F. Carrie; M. Degeilh; M.-F. Fauroux; M. Giraud; A. Vinel. Lyon: F. Bezot; H. Goutal. Gap: J.-L. Bonneuil. Nantes: L. Said.Chalons en Champagne: M. Joly. Limoges: J.-L.Domenger. Clapier: D. Brisot. Valence: D. Jakob. Dieppe: H. Guenneguez; P. Ouvry. Versailles: J.-M.Baud. St Junien: P. Houles. Reims: R. Jacquet; D.Malgrange; A. Vinograd. Château d 'Olonne: D. Gayoux. Tarbes: C. Cledat-Wendel. Sèvres: H. Benoit. Saint-Etienne: P. Preynat. La Varenne St Hilaire: C. Garde. Strasbourg: P. Eberle; V. Knauer. Bagneux: J.-L. Sevestre. Foix: A. Viard. Le Havre: A. Demagny. Rouen: D. Dumesnil. Bernay: O. Aycard; N. Vechambre. Tours: O. Genre. Dijon: C. Perreaux. Saint-Etienne: A. Peyrard. Lyon: J. Berger. Pau: P. Foucault. Agen: M. Saucaze-Larame. Poitiers: B. Demiot. (Investigators are listed in order of the number of patients enrolled. All locations are in France.)