Danaparoid has been shown in animal studies to be more effective than standard heparin or two different low-molecular-weight heparin preparations in preventing the extension of experimentally induced venous thrombi [4]. It has been both safe and effective in the prophylaxis of deep venous thrombosis in patients having cancer surgery [5], hip-fracture surgery [6], or hip-replacement surgery [7] and in patients with nonhemorrhagic stroke [8]. It has been used as an anticoagulant during hemodialysis [9, 10] and in patients with heparin-induced thrombocytopenia [11, 12] or disseminated intravascular coagulation [13]. Data from studies of the treatment of deep venous thrombosis in patients with hemorrhagic stroke indicate that treatment with danaparoid can prevent the extension of venous thromboembolism without aggravating cerebral bleeding [14]. No study has yet assessed the efficacy and safety of danaparoid in the treatment of patients presenting with acute deep venous thrombosis or pulmonary embolism. Danaparoid has a bioavailability of 100% after subcutaneous administration; the bioavailability of unfractionated heparin after subcutaneous injection is only 20% to 30%. Therefore, danaparoid is particularly suitable for subcutaneous administration, much like the low-molecular-weight heparin preparations [15-19], which have a bioavailablity of approximately 90%. Our study was designed to assess the efficacy and safety of two doses of subcutaneously administered danaparoid and of continuous intravenous administration of unfractionated heparin as initial treatment in patients presenting with acute proximal deep venous thrombosis of the leg, pulmonary embolism, or both.

Methods

Top Methods Results Discussion Author & Article Info References

Study Design

Our study was a randomized, open, parallel-group clinical trial done in one university hospital and two university-affiliated hospitals in the Netherlands. The study protocol and forms giving informed consent were approved by the institutional review board at each hospital.

Patients

All patients gave witnessed informed consent before being entered into the study. Men and women 18 years of age or older who presented with clinical symptoms of acute proximal deep venous thrombosis of the leg or pulmonary embolism of no more than 7 days duration were eligible.

Patients were excluded if they had had intracranial bleeding within 2 months or resuscitation by external chest compression within 48 hours; if they were allergic to heparin; if they were pregnant; if they were receiving treatment with coumarin derivatives; if they had been treated with thrombolytic drugs within 7 days; or if they were receiving ongoing treatment with aspirin, nonsteroidal anti-inflammatory drugs, dextran, or fibrinolytic drugs. The provisional diagnosis of venous thrombosis or pulmonary embolism had to be confirmed within 48 hours after the start of study treatment by compression ultrasonography or contrast venography (whichever could be done soonest) or by ventilation-perfusion lung scan. Treatment was discontinued and the patient was excluded from the study if the clinical diagnosis was not confirmed. Enrollment began in March 1991 and continued through August 1992.

Dosing Schedule

The efficacy and safety of two dosing schedules of danaparoid were compared with the efficacy and safety of continuous intravenous unfractionated heparin. The schedule for low-dose danaparoid was 1250 anti-factor Xa units given as an intravenous bolus, followed by subcutaneous doses of 1250 anti-factor Xa units every 12 hours. The schedule for high-dose danaparoid was 2000 anti-factor Xa units given as an intravenous bolus, followed by subcutaneous doses of 2000 anti-factor Xa units every 12 hours. The first subcutaneous injection was simultaneous with the intravenous bolus injection. The second subcutaneous injection was given at the time of the first of the routine twice-daily injections, unless this was within 6 hours of the first injection. Unfractionated heparin was given intravenously as a loading dose of 2500 U and was followed by an initial maintenance dose of 30 000 U every 24 hours. This maintenance dose was adjusted to reach activated partial thromboplastin x 2.5 to 3.5 times the control values; these times were equivalent to a heparin level of 0.25 to 0.40 U/mL. This was measured daily and 4 hours after any dose adjustment. Study treatment was given for at least 5 days and was continued until an international normalized ratio of at least 3.0 was achieved with oral anticoagulation therapy, which was started 48 hours after the initiation of study treatment. The oral anticoagulant dose was calculated daily using the Thrombotest (Nyegaard and Co., Oslo, Norway; international sensitivity index, 0.94); this was done each morning using plasma samples taken before the morning dose had been given. If the international normalized ratio was below the target level after 8 days of study treatment, the attending physician decided whether to continue heparin therapy, continue danaparoid therapy, or switch patients being treated with danaparoid to intravenous heparin.

Study treatment was randomized as follows. Consecutively numbered, identical boxes were kept in each hospital pharmacy; each box contained one of the three treatments, randomized per hospital. After giving informed consent, a patient was treated with medication from the next consecutive box. The investigators were blinded to the randomization schedule. Only when the study medication for one individual patient was delivered did the treatment become known. After being assigned to treatment, patients were excluded from the efficacy analysis only if diagnosis of deep venous thrombosis or pulmonary embolism could not be confirmed within 48 hours of admission.

Evaluations and Scheduling

The primary method of assessment for recurrence or extension was repeated ultrasonography of the leg, contrast venography, ventilation-perfusion scanning, or both contrast venography and ventilation-perfusion scanning. Assessment was done after at least 5 days and at most 8 days of study treatment, within 24 hours after stopping treatment, or if clinically indicated. Institutional physicians, who were blinded to treatment assignments, interpreted venograms, ultrasonograms, and lung scans. Clinical evidence of recurrence or extension was defined as documented clinical circumstances suggestive of venous thromboembolic disease leading to the discontinuation of study treatment. A daily physical examination (including tests for hemoglobin level, platelet count, and leukocyte count) and a urinalysis to test for erythrocyte count were done. Liver function tests were done and creatinine levels were measured before and at the end of study treatment. Plasma used to measure amidolytic anti-factor Xa activity was collected at the time of screening and at treatment days 2 and 4 (before and 2.5 hours after the morning injection on both days). This plasma was frozen at –20°C and stored until assay. Plasma samples were collected at the time of screening for determination of activated partial thromboplastin times before study treatment.

Follow-up assessment was done 2 months after the initiation of study treatment to gather information on state of health, recurrence or extension of venous thromboembolism, and bleeding complications.

Compression Ultrasonography

To establish the extent of thrombosis using ultrasonography [20], the deep venous system was divided into six segments: lower popliteal, upper popliteal, inferior femoral, mid-femoral, upper femoral, and common femoral veins. Each patient was first examined in the supine position so that the superficial femoral, common femoral, and iliac vein segments could be assessed. Patients were then examined in the prone position so that the proximal and distal popliteal vein segments could be assessed. The calf veins were not routinely examined. Compression and decompression maneuvers were done to test for compressibility (patency) of the vein, and duplex scanning was used as an additional tool for detecting and locating blood flow through the veins [21]. The results (presence or absence of occlusion per segment) were recorded in a standardized manner. The examiners were blinded to treatment assignment.

Contrast Venography

Venography was done according to a modification of the method described by Rabinov and Paulin [22]. A dorsal vein was cannulated with a 21-gauge or smaller needle, and contrast medium was infused while the patient was on a tilt table. Tourniquets were used as needed. In each case, the repeated venogram was evaluated (without knowledge of the treatment received) for the size of the clot. The amount of residual clot was then compared with the amount of original thrombus in each of the venous segments. The venograms were reevaluated by two radiologists who were blinded in order to maximize the accuracy of interpretation [23].

Ventilation-Perfusion Scanning

Pulmonary embolism was diagnosed using ventilation-perfusion lung scanning with Krypton-81m- and Technetium-99m-loaded macro-aggregates [24]. The high-probability scan accurately predicts the presence of pulmonary embolism in approximately 90% of patients with this condition, and its results may be accepted as the definitive diagnosis if alternative diagnoses are unlikely [25]. For our purposes, high-probability scans were considered diagnostic for pulmonary embolism. The scans were reviewed by two experienced specialists in nuclear medicine; they used standard diagnostic charts to optimize interpretation [26] and were blinded to the treatment given.

Clinical Analysis

Outcome assessment was done after at least 5 and no more than 8 study treatment days. Follow-up data were collected 2 months after the initiation of study treatment.

The primary determinant of efficacy was recurrence, which was defined as one or more of the following: an increase in size of a thrombus, the new development of deep venous thrombosis detected by repeated sonography or venography, or one or more new high-probability defects on repeated perfusion scintigraphy of the lung (with a normal ventilation scan and without infiltrative abnormalities on the chest radiograph). The examination method initially used to establish the diagnosis (ultrasonography or phlebography of the leg and ventilation-perfusion lung scan) was repeated within 24 hours after stopping the study treatment at least 5 and no more than 8 days after treatment had been initiated. Patients withdrawn from the study by their attending physicians because of clinical evidence of new or extending thromboembolism were considered to have had recurrence; confirmation by objective tests was sought whenever possible.

The primary determinant of safety was the incidence of major bleeding, which was defined as fatal bleeding, intracranial bleeding, or bleeding with or without defined focus causing a decrease in hemoglobin level of more than 1.5 mmol/L (2.42 g/dL) within 72 hours. We also assessed the incidence of minor bleeding, which was defined as any overt bleeding (including hematomas larger than 1 cm² at puncture sites) that did not meet the criteria for major bleeding.

Statistical Analysis

Statistical testing of the null hypothesis (there is no difference in events between patients receiving danaparoid and those receiving heparin) compared with the alternative hypothesis (there is a difference in events) was done two-sided at a significance level of 5%. Differences between the groups were tested for statistical significance using analysis of variance (ANOVA), the chi-square test, and the Fisher exact test, as appropriate. Whenever the variables were not normally distributed, the Kruskal-Wallis test and the Wilcoxon ranked-sum test were used to analyze the data. No corrections were made for multiple tests, so P values close to 0.05 for comparisons of secondary variables may represent spurious findings. Our analysis primarily involved patients receiving full treatment because of the dose-finding character of the study. Statistical analysis was done using STATA (Computing Resource Center, Santa Monica, California); 95% CIs for relative risk were calculated using CIA (British Medical Journal, London, United Kingdom).

All consecutive eligible outpatients and patients admitted to the medical departments of the three hospitals were invited to take part in the study. Although this was an open study, the efficacy and safety variables were statistically analyzed because the assessors of efficacy determinants were blinded and because assessments for the safety determinants were considered to be insensitive to subjective bias.

Results

Top Methods Results Discussion Author & Article Info References

Patient Characteristics

Two hundred and nine of 213 consecutive patients (98%) consented to participate, were randomly assigned to treatment, and received initial study medication. In 15 of the 209 patients, objective tests excluded venous thromboembolism and study treatment was discontinued. None of these patients had venous thromboembolism or bleeding complications during therapy or at follow-up. Six of the remaining 194 patients were excluded because of major protocol violations during the study: the concomitant use of subcutaneous heparin (1 patient), accidental overdose of heparin because of a labeling error (1 patient), combination of two treatments (1 patient), a treatment period too short in duration (1 patient), and a comorbid condition that did not allow adequate conduct of the trial (2 patients). Thus, 188 of 209 patients (90%) received treatment according to the study protocol and were included in the evaluable-patient efficacy analysis; their characteristics are shown in Table 1. Of these patients, 158 had clinical and objective evidence of deep venous thrombosis; 30 had evidence of pulmonary embolism; and 15 had evidence of both thrombosis and embolism. Abnormal results on a ventilation-perfusion scan were found in 66 of 158 patients (42%) with proven deep venous thrombosis who did not have signs or symptoms of pulmonary embolism. Deep venous thrombosis was detected in 7 of 30 patients (23%) who presented clinically with pulmonary embolism only. Sixty-five patients received low-dose danaparoid, 63 received high-dose danaparoid, and 60 received unfractionated heparin. The treatment groups were well balanced for baseline characteristics, except that the group receiving low-dose danaparoid had a preponderance of limb paralysis and immobilization, and the group receiving heparin had a low incidence of risk factors for bleeding. The clinical and biochemical characteristics of the three treatment groups did not differ at study entry (Table 1).

Efficacy

Assessment of efficacy was done in 188 patients (Table 2). Pairwise assessment of the leg veins was available in 173 patients (138 pairs of ultrasonograms and 35 pairs of venograms). In each of 179 patients, two lung scans were available for pairwise assessment. Both leg and lung assessments were done in 166 patients.

Extension of thrombus into at least one other vein segment was seen in 18 of 173 patients (10%). In patients with deep venous thrombosis, high-dose danaparoid reduced the frequency of recurrence or extension (3 of 58 patients [5%]) compared with heparin (6 of 54 patients [11%]; relative risk, 0.47 [CI, 0.12 to 1.77]). In patients receiving low-dose danaparoid, 9 of 61 patients had recurrence or extension (15%; relative risk, 1.33 [CI, 0.51 to 3.49] compared with heparin).

New high-probability defects were detected in 29 of 179 patients (16%). Patients receiving high-dose danaparoid had a reduced incidence of new defects (4 of 61 patients [7%]) compared with patients receiving heparin (14 of 58 [24%]; relative risk, 0.27 [CI, 0.09 to 0.78]), but patients receiving low-dose danaparoid (11 of 60 patients [18%]; relative risk, 0.76 [CI, 0.38 to 1.53]) did not.

Overall recurrence or extension of thromboembolism per patient was seen in 43 of 188 patients (23%). Incidence of overall recurrence was reduced in patients receiving high-dose danaparoid (8 of 63 patients [13%]) compared with patients receiving heparin (17 of 60 patients [28%]; relative risk, 0.45 [CI, 0.21 to 0.96]) but not in patients receiving low-dose danaparoid (18 of 65 patients [28%]; relative risk, 0.98 [CI, 0.56 to 1.72]) compared with those receiving heparin. Overall, extension or new disease was clinically silent: Clinical evidence of new pulmonary emboli was present in 3 patients (1 in each treatment group), clinical evidence of extension of leg venous thrombosis was present in 1 patient (receiving high-dose danaparoid), and clinical evidence of both thrombosis and embolism was present in 1 patient (receiving high-dose danaparoid). In these 5 patients, clinical progression was confirmed by objective testing. Death from pulmonary embolism was not seen.

Including the 6 patients with protocol violations in the analysis yielded similar results: The frequency of recurrence was 28% (19 of 67 patients) in patients receiving low-dose danaparoid, 13% (8 of 64 patients) in patients receiving high-dose danaparoid, and 29% (18 of 63 patients) in patients receiving heparin. Analysis of all 209 patients randomly assigned to treatment (including the 15 patients whose diagnosis of venous thromboembolism was rejected) did not alter the differences in outcome: Recurrence was 27% (19 of 70 patients) in patients receiving low-dose danaparoid, 12% (8 of 69 patients) in patients receiving high-dose danaparoid, and 26% (18 of 70 patients) in patients receiving heparin (relative risk for comparison between low-dose danaparoid and heparin, 1.06 [CI, 0.61 to 1.84]; relative risk for comparison between high-dose danaparoid and heparin, 0.51 [CI, 0.24 to 1.09]). Two months after the initiation of study treatment, the cumulative incidence of recurrence was 29% (19 of 65 patients) in patients receiving low-dose danaparoid, 14% (9 of 63 patients) in patients receiving high-dose danaparoid, and 28% (17 of 60 patients) in patients receiving intravenous heparin (relative risk for comparison of low-dose danaparoid and heparin, 1.06 [CI, 0.61 to 1.85]; relative risk for comparison of high-dose danaparoid and heparin, 0.50 (CI, 0.24 to 1.04).

Treatment with unfractionated heparin was adequate, as is evident from the activated partial thromboplastin time ratios. The overall mean ratio was 3.08, and no statistically significant difference was seen between the mean ratios for patients who did (3.02) and did not (3.08) have recurrence.

Safety

Analysis of safety was done in all 209 patients initially enrolled in the study. Major bleeding occurred in 4 patients (Table 3). One patient receiving heparin died of intracranial bleeding on treatment day 6. Autopsy showed a hemorrhage in a recent ischemic cerebrovascular accident and no evidence of recurrence of thromboembolism. At the moment of bleeding, the activated partial thromboplastin time ratio was 3.3 and the international normalized ratio was 2.9; both were within the therapeutic range. The hemoglobin levels of 3 patients decreased by more than 1.5 mmol/L without overt focus for bleeding: The hemoglobin level decreased from 7.9 mmol/L to 6.1 mmol/L in 1 patient receiving heparin, from 8.1 mmol/L to 6.5 mmol/L in 1 patient receiving low-dose danaparoid, and from 8.3 mmol/L to 6.7 mmol/L in 1 patient treated with high-dose danaparoid. Another patient received blood transfusions because of anemia caused by hemolysis. This patient had a history of severe autoimmune hemolytic anemia; he was considered to be nonevaluable for the end point anemia but was included in analysis for the other end points. Minor bleeding occurred in 38 of 71 patients (54%) receiving low-dose danaparoid, 35 of 68 patients (51%) receiving high-dose danaparoid, and 39 of 70 patients (56%) receiving heparin; episodes of minor bleeding were equally distributed over the groups. Adverse events caused by the administration of heparin or danaparoid were not seen. During trial treatment, 2 more patients treated with heparin died, 1 of prostate carcinoma and 1 of cardiac arrest. Two months after the initiation of study treatment, 1 patient who had received low-dose danaparoid had died of progressive cancer; 2 patients who had received high-dose danaparoid had died of progressive cancer and sepsis, respectively; and 1 patient who had received heparin had died of cardiopulmonary arrest.

No major bleeding complications were seen in the 15 patients whose diagnosis of venous thromboembolism was refuted by objective testing or in the 6 patients with major protocol violations. Two months after the initiation of study treatment, no further bleeding complications were seen.

Discussion

Top Methods Results Discussion Author & Article Info References

Danaparoid is a mixture of low-molecular-weight heparan, dermatan, and chondroitin sulfates, which are natural sulfated glycosaminoglycuronans derived from animal intestinal mucosa. In humans, this preparation has antithrombotic activity with apparently minimal bleeding side effects and has been shown to be effective in the prevention of venous thromboembolism in many different categories of patients [5-8]. The combined actions of heparan and dermatan sulfate suppress factor IX activation more effectively than heparin, and they delay the onset of factor X activation almost as effectively as heparin [27]. Most of the drug's anti-factor Xa activity resides in a fraction of heparan sulfate with high affinity. Dermatan sulfate provides relatively modest anti-factor IIa activity. Danaparoid inhibits exogenous thrombin by both antithrombin III and heparin cofactor II; unfractionated and low-molecular-weight heparins primarily inhibit exogenous thrombin by antithrombin III [28]. The overall antithrombotic activity appears to be enhanced by an unresolved endothelial cellular mechanism. These properties interact and provide the overall anticoagulant and antithrombotic effect of the drug, but they have different pharmacokinetic profiles and elimination half-life values. Optimal dosing regimens must be derived from clinical rather than pharmacokinetic studies because there are no physicochemical methods for the assay of separate components.

The hemorrhagic effects of danaparoid have been assessed in several experimental bleeding models [1] measuring blood loss after standardized trauma. The results of these models indicate that the benefit:risk (antithrombotic activity:bleeding) ratio of danaparoid is better than that of heparin [1]. The reduced hemorrhagic effect of danaparoid is evident as a lower initial bleeding-enhancing effect, and it may be caused by differences in effect on both coagulation and platelet function. Compared with heparin and low-molecular-weight heparins, danaparoid inhibits thrombin generation linearly over a much broader range of concentrations. Danaparoid has little influence on platelet function, whereas heparin impairs the formation of primary hemostatic plugs, retards platelet degranulation, and inhibits platelet adherence [1].

Danaparoid has minimal cross-reactivity (about 10%) in tests for heparin-associated antibody done in plasma from patients with heparin-associated thrombocytopenia (low-molecular-weight heparins have cross-reactivity of about 90%) and thus appears to be the appropriate drug for the treatment of these patients [11]. Its usefulness is evident from a large compilation of case reports [11], but no randomized trials have been done.

Despite the considerable experience that has been gained in thromboprophylaxis and in the treatment of patients with heparin-induced thrombocytopenia, no formal trials have addressed the efficacy and safety of danaparoid in the initial treatment of venous thromboembolism in a general setting of medical patients. Our study was done primarily to define the optimal dose of danaparoid for further clinical testing in the treatment of venous thromboembolism. After we excluded 15 patients whose diagnosis of thromboembolism was not confirmed, the treatment groups remained well balanced. The treatment regimens for danaparoid were based on previous pharmacokinetic [2, 29] and clinical studies [3, 30].

Outcome measurements for efficacy were clinical and substitute end points for extension or recurrence of venous thromboembolism, assessed using clinical evidence and by comparing venograms or ultrasonograms and lung scans done before and after study treatment. Only 5 patients had clinical recurrence. In our study, the main benefit of danaparoid was a significant reduction in the occurrence of new pulmonary emboli as evident from the repeated ventilation-perfusion lung scans. A large percentage of patients receiving heparin had such emboli (14 of 58 patients [24%]), confirming other observations [31]. To maintain the quality of the assessment of the ventilation-perfusion scans, central evaluation of the scans was done by a panel of specialists in nuclear medicine who were blinded to the treatment assignment and who used predetermined criteria [26]. The value and importance of decreasing the frequency of new defects on lung scans or imaging tests of the lower extremities, without evidence of an effect on clinical end points, are unclear. Therefore, the positive results seen with danaparoid in this study need to be confirmed in a study large enough to use clinical end points as measures of efficacy. Such a study should also include a follow-up of 6 months, given the unexpected finding of differences in mortality between patients treated with unfractionated heparin and those treated with low-molecular-weight heparin [19].

Early adequate anticoagulation is crucial in the prevention of new emboli in patients treated with intravenous heparin [32]. The recommended intravenous heparin loading dose is 5000 U, which should be followed by continuous intravenous heparin infusion to maintain an activated partial thromboplastin time 1.5 to 2.5 times the control values; this is equivalent to a heparin level of 0.2 to 0.4 U/mL [33]. We chose 2500 U as a loading dose, and elongation of activated partial thromboplastin time 2.5 to 3.5 times the control values, equivalent to heparin levels of 0.25 to 0.40 U/mL. The observed values make it unlikely that the benefit of treatment with the higher dose of danaparoid resulted from underdosage in the group treated with heparin. In accordance with common practice in the Netherlands, oral anticoagulation therapy was started 48 hours after the initiation of study treatment, and an international normalized ratio of at least 3.0 was pursued. This target is higher than that given in recent recommendations [34]. The early initiation and high target level of anticoagulation may explain the low incidence of clinical recurrence compared with findings in a recent meta-analysis that reported recurrence rates of 2.82% in patients treated with low-molecular-weight heparins and 4.63% in patients treated with unfractionated heparin [35].

We assessed the safety of treatments using predetermined criteria for major and minor bleeding. This assessment could not be blinded because of the nature of the different therapeutic regimens. Major bleeding episodes were defined by objective parameters: death attributable to bleeding or a decrease in hemoglobin level of more than 1.5 mmol/L in a circumscribed time period. Few such episodes were seen: One fatal intracranial hemorrhage occurred during heparin therapy, and three patients (one in each treatment group) had decreases in hemoglobin level.

Risk factors for bleeding, as reported in the literature [36], were identified at baseline. Both of the groups receiving danaparoid had a preponderance of risk factors for bleeding Table 1 (32 of 65 patients [49%] receiving low-dose danaparoid and 31 of 63 patients [49%] receiving high-dose danaparoid) compared with the group receiving heparin (16 of 60 patients [27%]; odds ratio for comparison between low-dose danaparoid and heparin, 2.67 (CI, 1.26 to 5.56) and odds ratio for comparison between high-dose danaparoid and heparin, 2.66 [CI, 1.25 to 5.67]). This imbalance was not associated with more bleeding in patients treated with danaparoid.

The high incidence of minor bleeding, equal in all three study arms, may have been the result of our stringent criteria for minor bleeding and our careful daily examination of all patients.

The potential safety of treatment with danaparoid has been ascribed to the drug's high anti-factor Xa:anti-factor IIa ratio of 28:1, which is approximately 10 times higher than that of the low-molecular-weight heparin preparations. However, the anticoagulant activity of danaparoid is the result of the combined and interacting effects of the different components, which makes direct comparison with low-molecular-weight heparins difficult. The extent to which anti-factor Xa:anti-factor IIa ratios assayed in vitro determine anticoagulant properties relevant in vivo cannot currently be established.

Our study shows that subcutaneous administration of the low-molecular-weight heparinoid danaparoid, 2000 U every 12 hours, to patients with deep venous thrombosis or pulmonary embolism resulted in a lower incidence of recurrent thromboembolism. The limitations of our study were its size, which necessitated the use of substitution end points; the relatively short follow-up period of 2 months; and the patient population studied (medical patients, who have fewer potential bleeding complications than surgical patients). Despite these limitations and shortcomings, all trends (similar frequency of bleeding complications despite the presence of more risk factors for bleeding in the groups receiving danaparoid; reduction in recurrence of thrombosis and embolism; and reduction in recurrence of combined thromboembolism) suggest that high-dose danaparoid is more potent than standard intravenous unfractionated heparin.

Further clinical evaluation should consist of a randomized, double-blinded trial comparing danaparoid with unfractionated or low-molecular-weight heparin. This trial should be large enough to enable the use of clinical recurrences as efficacy end points, should include patients at high risk for bleeding (surgical and neurosurgical patients), and should have a 6-month follow-up.

A recent meta-analysis of 16 randomized controlled clinical trials involving 2045 patients, which compared the efficacy and safety of low-molecular-weight heparins and unfractionated heparin in the initial treatment of deep venous thrombosis, showed low-molecular-weight heparins to significantly reduce the incidence of thrombus extension (common odds ratio, 0.51 [CI, 0.32 to 0.83]) [35]. Nonsignificant trends in favor of low-molecular-weight heparins were seen for recurrence of thromboembolic events (common odds ratio, 0.66 [CI, 0.41 to 1.07]), major bleedings (common odds ratio, 0.65, [CI, 0.36 to 1.16]), and total mortality (common odds ratio, 0.72 [CI, 0.46 to 1.4]). These results indicate but do not prove that low-molecular-weight heparins have a higher benefit:risk ratio than unfractionated heparin in the treatment of venous thrombosis. Low-molecular-weight preparations may be preferred by clinicians because they are easy to administer and do not require laboratory-guided dose adjustments, but their clinical superiority remains to be confirmed in larger-scale clinical trials with adequate power. Similarly, the potential advantages of danaparoid, the intrinsic properties of which should theoretically produce a greater benefit:risk ratio than either unfractionated heparin or low-molecular-weight heparins, should be tested in adequate clinical trials.