INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Several studies have shown that low molecular weight heparins (LMWH) are at least as effective as unfractionated heparin in the prophylaxis of deep-vein thrombosis (DVT) in major surgery patients (1). However, few studies have been conducted on the prevention of venous thromboembolism in nonsurgical patients (10).

The incidence of DVT in medical disorders varies considerably. Although it is generally estimated to be about 25% without prevention (20), it can range from 8 to 25% for patients with mild disease who are incompletely immobilized (19, 21), to 40% for patients with hemiplegia after a nonhemorrhagic, ischemic, cerebrovascular event (22). These differences are a reflection of the populations at study and the different assessment methods used.

The few studies that have compared LMWH with unfractionated heparin in medical patients with various diseases (10, 13, 14, 25) have not demonstrated any significant difference in efficacy, but only a trend towards improved safety with LMWH. No study has yet to provide a rationale for the routine use of preventive heparin therapy in such patients. In a recent review, four large studies were analyzed (26); 15,795 patients hospitalized in general medicine departments and at risk for DVT were treated with prophylactic doses of heparin. The figuring of odds ratio for the effect of low-dose heparin on mortality is 0.91 (95 CI: 0.80 to 1.04). Thus, these recent studies do not exclude the possibility that prophylactic treatment of general medical patients may reduce the mortality by a noteworthy 20% rate.

Although there are no precise data on the prevalence of thromboembolic disease in patients hospitalized for pulmonary disease, it has been estimated at 8 to 25% (19, 27, 28). There is no consensus regarding the efficacy of thromboembolism prophylaxis in critically ill patients. Further, no clear guidelines as to drug dose and regimen can yet be derived from clinical studies. As a consequence, it was thought to be ethical to include a placebo group in this study, which was designed to evaluate the therapeutic benefits of the LMW heparin nadroparin calcium (Fraxiparine) in patients with acute, decompensated chronic obstructive pulmonary disease (COPD) who required mechanical ventilation, using clinical examination and Doppler ultrasonography to detect thromboembolic events.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Design

This prospective, randomized, double-blind, comparative trial involved 34 medical intensive care units in France. From March 1992 through April 1995, consecutive patients admitted with acute, respiratory-decompensated COPD, who required mechanical ventilation, were enrolled if they were 40 to 80 yr of age and weighed 45 to 110 kg. Patients were included if they had: a history of a confirmed DVT within the previous 6 mo or presence of signs of a DVT on the Doppler ultrasonography at inclusion, an organic lesion that could bleed, i.e., an active gastroduodenal ulcer or a recent hemorrhagic cerebrovascular accident; severe liver failure leading to a decrease of the prothrombin time (PT) to less than 50% (normal values between 70 and 100%); severe renal impairment (serum creatinine > 300 µmol/L); confirmed or uncontrolled hypertension (diastolic blood pressure > 120 mm Hg); a congenital or acquired coagulation disorder; a history of hypersensitivity or thrombocytopenia to heparins of any type; were contraindicated to anticoagulant therapy, venography, or angiography; or were receiving any form of acetylsalicylic acid, ticlopidine, or oral anticoagulants. Written, informed consent was obtained, and the Ethics Committee of Lyons, France approved the study protocol.

Patients were randomized to receive either nadroparin (Fraxiparine; Sanofi-Winthrop, Gentilly, France) or matching placebo by subcutaneous injection once daily. Taking previous clinical experience with nadroparin in high-risk surgical patients (8) into consideration, dosage was based on patients' body weight (3,800 AXa IU, i.e., 0.4 ml for 45 to 70 kg; 5,700 AXa, i.e., 0.6 ml for 71 to 110 kg). Nadroparin was supplied as a concentrated solution of 9,500 AXa IU/ml in disposable, prefilled syringes. Placebo 0.9% physiological saline was supplied in an identical manner.

Study treatment started immediately after enrollment and continued until the patient could be weaned from mechanical ventilation. By convention in the protocol, the duration of study treatment could not exceed 21 ± 1 d. The time between the start of mechanical ventilation, enrollment, and first injection of LMWH was as short as possible, and, in any case, was supposed not to exceed 24 h. Patients with less than 48 h of ventilator were excluded from the study. If the mechanical ventilation had to be continued more than 21 ± 1 d after inclusion, the studied treatment was stopped at 21 ± 1 d after inclusion. The study observation period was then considered to be ended and the patients were evaluated for primary end point (bilateral venography) and included in the intention-to-treat analysis.

Patients were examined daily for DVT on usual signs or symptoms (pain, heat, redness, edema), hemorrhage, and other adverse events. Doppler ultrasonography was performed before inclusion and weekly during the study, and in all cases of clinically suspected DVT. Examinations, performed by a single operator in each center using Duplex compression ultrasonography, were bilateral, compared both legs, and included systematic study of each venous segment. Thrombosis was suspected by the presence of echogenic endoluminal material and/or the absence of total vein compressibility.

Standard laboratory tests were performed at enrollment and the day after the final treatment (end of study), and in the event of early permanent discontinuation. Laboratory tests included: complete blood count with leukocyte differential, hemoglobin, hematocrit, activated partial thromboplastin time (APTT), PT, serum electrolytes, and creatinine. Platelets were counted twice per week.

Treatment was discontinued if a major critical event (lack of therapeutic efficacy or serious adverse event) or a major protocol violation occurred. In case of lack efficacy, or DVT, treatment at a therapeutic dose was usually initiated.

Assessment of Outcome

The primary efficacy criterion was incidence of DVT diagnosed by venography. Venography was performed at normal planned completion, in cases of early permanent discontinuation, or during the study, for positive, doubtful, or uninterpretable Doppler ultrasonography. Early permanent discontinuation of treatment was due to either a lack of efficacy, a serious adverse event (SAE), patient decision, an intercurrent event presenting a potential risk for the patient, or a major protocol violation (including noncompliance with therapy). Diagnosis was confirmed by two independent radiological experts' assessments of the venographies. Experts were not informed of the treatment allocated or the conclusions of the radiologist who performed venography.

Presence of pulmonary embolism was clinically assessed by daily physical examinations. This was confirmed by digital or conventional pulmonary angiography only in cases with clinical features suggestive of this diagnosis.

Safety criteria included the incidence of major or minor hemorrhage (29). Hemorrhage was considered major when it was overt and was associated with a decrease in hemoglobin concentration of 2 g/dl or more compared with the baseline value, when it necessitated a transfusion of two or more units of packed red cells, when it was retroperitoneal or intracranial, or when the investigator decided to end the treatment with heparin because of his judgment on the benefit/risk ratio. Minor hemorrhages were those not considered major. Safety criteria included severe thrombocytopenia (i.e., platelet count < 50,000 cells/mm³ with or without clinical signs, platelet count between 50,000 and 100,000 cells/mm³ with clinical signs, or a 50% decrease compared with the baseline reference count); and any other treatment-related adverse events. Adverse events were defined as serious if they caused death, were life threatening, or prolonged hospital stay. The Committee on Critical Events (members were independent of the study and unaware of the nature of the treatment administered) assessed whether serious adverse events were treatment-related. Laboratory safety was assessed at enrollment and at end-of-study assessments.

Statistical Analysis

Efficacy was analyzed using evaluable patients, i.e., those with a venography (see Figure 1), and safety was analyzed using an intention-to-treat (ITT) analysis in all patients who received at least one injection. All statistical tests (unpaired, two-sided t test for quantitative variables; chi-square test for discrete variables or Fisher's exact test when conditions for using a chi-square test were not met; and Mann-Whitney nonparametric U test for ordinal variables) were two-tailed, with an alpha error of 5%.

View larger version (24K): [in this window] [in a new window]   Figure 1.   Patients evaluable for the primary efficacy criterion. EPD = early permanent discontinuation.

Published data suggest that patients with acute, decompensated COPD who received no prophylaxis dose have a potential incidence of symptomatic DVT ranging between 8 and 25% (21, 27, 28). With a beta error of 10% and an alpha error of 5%, these estimates may be reduced to 3 and 8%, respectively. Thus, 200 patients were required for each group. Given the uncertainty of these premises, an interim analysis of data was performed to review whether the study should continue after the first 100 patients had been enrolled in each group. As a result, the study was stopped after 223 patients had been enrolled.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Description of Treated Population

All 223 enrolled patients were randomized, but two (one in each group) did not receive any treatment, giving 221 treated patients (Figure 1). In the nadroparin group, 56% received 0.4 ml and 44% received 0.6 ml, whereas 64 and 36% of those in the placebo groups received 0.4 and 0.6 ml, respectively.

Both groups were comparable for all demographic and clinical characteristics, except for age; patients receiving nadroparin were significantly (p = 0.02) older that those receiving placebo (Table 1). The primary cause of COPD was chronic bronchitis. Risk factors for venous thromboembolism were also similar in both groups, as were all laboratory parameters measured at inclusion.

Mean duration of therapy was 11.9 ± 6.0 (SD) and 11.4 ± 6.0 (SD) days in nadroparin and placebo groups, respectively. There was no significant difference in the numbers of patients with an early permanent discontinuation of therapy in the two groups (Table 2). However, patients receiving nadroparin were less likely to discontinue treatment because of DVT than were those receiving placebo (three versus 11; p = 0.06). The decrease in sample size over time was similar in both groups.

Efficacy

Of the 221 treated patients, 52 (23.5%) were not assessed by venography (Figure 1), thus giving 169 patients (nadroparin, n = 84; placebo, n = 85) who were evaluable for the assessment of efficacy. The reason for the lack of assessment in 52 patients were similar in both groups and included: technical difficulty, contrast media hypersensitivity, septic shock, cardiogenic shock, congestive heart failure, arrhythmia, thrombocytopenia, bleeding, acute renal failure, patient refusal, and premature weaning of mechanical ventilation. No significant differences were observed between the two groups, as similar numbers in each group were not assessed by venography.

There was a significantly lower incidence of total DVT in patients receiving nadroparin than in those receiving placebo (15.5 versus 28.2%, respectively, p = 0.045) (Table 3). The patients receiving nadroparin and having a DVT numbered 13 out of 84; the patients receiving placebo and having a DVT numbered 24 out of 85. 

Proximal DVT (including extended DVT with both proximal and distal localizations) occurred in three patients in the nadroparin group versus seven patients in the placebo group. Because of the small size of the sample, no significant difference was reached on this criteria even when there was a trend in favor of nadroparin. This trend was reinforced by the fact that there was one (one of three) versus six (six of seven) patients with extended proximal DVT, respectively, in the nadroparin and the placebo groups.

Only one patient in the nadroparin group presented with clinical features of suspected PE during the trial. However, venography and pulmonary angiography were both normal. Another patient receiving nadroparin presented with acute respiratory insufficiency, followed by cardiovascular failure and died 48 h after the scheduled completion of the trial. Although PE was suspected in this patient it was not confirmed since neither pulmonary angiography nor autopsy were performed.

Safety

Clinical safety was analyzed on an ITT basis in the 221 treated patients (Table 4). One or more adverse events (regardless of severity or causal relationship to treatment) were experienced by 46.3% of patients receiving nadroparin compared with 39.8% of patients receiving placebo; the difference was not significant.

The most commonly reported adverse events in the nadroparin and placebo groups, respectively, were hemorrhage (25 versus 18, p = 0.18) and thrombocytopenia (10 versus 7, p = 0.39).

Early permanent discontinuation of treatment because of an adverse event occurred in 13 (12%) patients receiving nadroparin and in 10 (8.8%) of those receiving placebo; the difference was not significant.

Forty-nine patients experienced one or more serious adverse event. The most commonly reported event was cardiovascular (14 versus 9 patients in the nadroparin and placebo groups, respectively). This was mainly heart failure after septic shock.

Sixteen patients died during the study (eight in each group). The main cause of death was cardiovascular complications associated with infection.

No significant differences were observed in the hematologic or coagulation factors measured (Table 5).

Ten (9.3%) patients receiving nadroparin and seven (6.2%) receiving placebo had platelet counts < 100,000/mm³ or a 50% decrease compared with their baseline value; however, the difference was not significant. In patients receiving nadroparin, thrombocytopenia occurred concomitantly with septic shock (n = 5); an ischemic vascular event (n = 1); or with hematuria (n = 1). In the other three cases, treatment was continued and platelet count returned to normal in the following days. The Critical Events Committee considered three of these cases to be serious, and only one to be possibly related to nadroparin.

Thrombocytopenia in the seven patients receiving placebo was associated with septic shock (n = 3), DVT (n = 1), a bleeding event (n = 1), and fatal cardiogenic shock (n = 1). It was purely asymptomatic in one patient and did not worsen during the following days, despite continued treatment. The Critical Events Committee considered two of these cases to be serious, and only one to be possibly related to placebo.

Clinically significant abnormal hemoglobin levels (i.e., = 8 g/dl or a decrease of = 3 g/dl versus baseline) were found in 17 (15.7%) patients receiving nadroparin compared with 14 (12.4%) receiving placebo.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Although a relatively large number of studies have evaluated the efficacy and safety of LMWH in the prophylaxis of venous thromboembolism in surgery, there have been few in medical patients.

Estimates of the incidence of DVT in respiratory insufficiency range from 8.9 to 44% (21, 27, 28). However, the estimate of 44.7% was based on the detection of thrombosis by radiolabeled platelets at inclusion, a method that has good sensitivity but poor specificity DVT in acute, decompensated COPD is due to several concomitant factors: confinement to bed, sedation, right-sided heart failure, and venous stasis (25, 31, 32).

In a randomized, controlled trial, Cade (33) reported a significant decrease in the incidence of DVT in 119 intensive care patients with unfractionated heparin (5,000 IU twice a day) compared with placebo (13 versus 29%; p < 0.001).

Belch and colleagues (12) conducted a study in 100 patients with pulmonary infection and/or heart failure who were confined to bed for at least 3 d. Routine screening for DVT was assessed with radiolabeled fibrinogen. The incidence of DVT was significantly greater in the untreated group than in the treated group (unfractionated heparin, 5,000 IU twice a day) (26 versus 4%, respectively; p < 0.01). But the recent review by Lederle (26) emphasizes the lack of positive controlled study in medical patients.

It was therefore decided to conduct a trial to evaluate the safety and efficacy of nadroparin in hospitalized patients with acute, decompensated COPD who required mechanical ventilation.

At the time of the study, there was no consensus on the use of heparin as prophylaxis for thromboembolic events in medical patients, thus justifying the inclusion of placebo. The only significant (p = 0.02) demographic difference between the two groups was age; patients receiving nadroparin were older than those receiving placebo (Table 1). However, there was no explanation for this other than a random event.

These relatively elderly patients, with a history of COPD for an average of 10 yr, presented with serious risk factors for DVT such as immobilization (100%), respiratory disease (100%), bronchial superinfection (74%), congestive heart failure (29%), age > 65 yr (50%), obesity (23%), venous insufficiency (13%), neoplastic disease (5%), and previous thromboembolic disorders (4%).

Exclusion criteria included recent suspected DVT and doubtful or uninterpretable Doppler ultrasonography at enrollment. Although the diagnosis of DVT with Doppler ultrasonography can be considered as good with 93% specificity and 95% sensitivity, as far as proximal DVTs are concerned, the difficulties encountered in performing the initial examination may have resulted in an underestimate of cases at enrollment (34). Only one patient was excluded for this reason, which implies a very low incidence of DVT in these potentially high-risk patients, in contrast to the much higher incidence observed during the study.

The incidence of total DVT in the nadroparin group was significantly lower than in the placebo group (15.5 versus 28.2%, p = 0.045), and the distribution of proximal and distal thrombi was not statistically different between groups. However, it should be noted that only one patient receiving nadroparin had an extended thrombus, i.e., proximal and distal, compared with six patients receiving placebo.

Detailed analysis of patients with uninterpretable venographies did not reveal any significant differences between the two groups. The reasons why venography was not performed were evenly distributed and therefore unlikely to have affected the results.

Although venography remains the reference method for the diagnosis of DVT in the lower limbs, the rate of agreement between observers using the Kappa test varies widely in literature (range, 51 to 95%) (35).

A comparison of the results from Doppler ultrasonography and venography at the end of the study revealed discrepancies, even after eliminating doubtful or uninterpretable results.

Nevertheless, our results appear to agree with those published in the literature (25, 28, 31), and prior consensus to determine examination methods, criteria for suspecting DVT (34), and using only one investigator per centre (35) tended to minimize an unavoidable center-related effect.

No proven pulmonary embolism was observed during the study; however, it was not systematically investigated by objective test. Clinical symptoms for pulmonary embolism are neither specific nor reliable. In addition, neither pulmonary angiography nor autopsy were routinely performed in patients who died. This probably accounts for the underestimate of risk: pulmonary embolism was thought to have contributed to only one patient's death, but both venography and pulmonary angiography were normal. In such patients, the starting point for pulmonary embolism can be a cardiac event (21, 30). Nonetheless, the high percentage of distal DVT could account for the fact that pulmonary embolism is rare (21); moreover, all patients with suspected DVT, no matter the way it was suspected, were immediately treated with curative doses of LMWH.

The incidence of reported adverse events was high in both groups, which can be expected since all patients had chronic lung disorders requiring hospitalization, and most were elderly. As a result, there were no significant differences between the two groups for total adverse events, serious adverse events, or those causing early permanent treatment discontinuation.

The most common adverse events were hemorrhage (often associated with septic shock) (n = 25) for patients receiving nadroparin, six serious; and n = 18 for those receiving placebo, three serious). Serious adverse events were considered (Committee on Critical Events) "possibly" or "likely" to be related to nadroparin in five cases of hemorrhage, one case of thrombocytopenia, and one case of "pneumonia"; and to placebo in two cases of bleeding, and one case of thrombocytopenia.

Mortality was high in both groups during the study. Most deaths were due to cardiovascular complications or nosocomial pneumopathies. Given the serious condition of these patients it is not surprising that neither treatment had any effect on mortality. In the frame of this clinical study, all patients had daily clinical checkups and weekly Doppler examinations, and any diagnosed DVT was immediately treated by curative dosage of heparin. This probably accounts for the very low rate of symptomatic pulmonary embolism observed. These results agree with those previously reported for elderly patients (21, 28, 30, 33).

In conclusion, medical patients at high risk of venous thromboembolism treated with nadroparin adjusted to body weight (mean duration of treatment, 11 d) gained definite therapeutic benefit; a 45% decrease in the incidence of DVT, which was not associated with a high incidence of serious bleeding or thrombocytopenia.