Preoperative Initiation of Deep-Vein Thrombosis Prophylaxis with a Low-Molecular-Weight Heparin (Nadroparin) in Elective Hip Replacement

Postoperative

Versus

Gualtiero Palareti, M.D., *Battista Borghi, M.D., Nicoletta Leali, M.D., Rita Golfieri, M.D., *M.

Sergio Coccheri, M.D., Montebugnoli, M.D.,

†Giovanni Inghilleri,

Enrico Maria Pogliani, ∥

M.D., ‡Alessandro Ronzio, M.D., §Tiziano Barbui, M.D., M.D., Gianfranco Di Nino, and ¶Lorenzo Spotorno, M.D., for

the CITO

Study Group

S. Orsola, University of Bologna, and *Istituti Ortopedici Rizzoli, Bologna; †Istituto Ortopedico Gaetano Pini and ‡Istituto Ortopedico Galeazzi, Milan; §Ospedali Riuniti, Bergamo; ∥ Ospedale S. Gerardo, University of Milan, Monza; and ¶ Ospedale S. Corona, Pietra Ligure, Italy

Ospedale

Summary: The aim of this multicenter, randomized, double-blind study performed in patients undergoing elective hip surgery was to compare the efficacy and safety of prophylaxis with low-molecular-weight heparin (LMWH) (Nadroparin, 7,500 anti-Xa IC units for the first 3 days and 10,000 from the fourth day on, s.c. o.i.d.) begun in one group shortly after surgery and in the other 12 h before operation, as is usually recommended. Preoperative administration (drug or placebo) was the only difference between the two groups. Deep vein thrombosis (DVT) was detected by bilateral venography 10-15 days after surgery. The study investigated 179 patients (55 men), 40-80 years old, in seven Italian orthopedic centers. In 131 patients efficacy analysis was possible because of adequate bilateral venography. All 179 patients were evaluated for bleeding complications. The prevalence of thrombotic complications was similar in the two groups. Proximal DVT was found in 8.4% of patients (10.8% and 6.1% in the preop and postop groups, respec-

Venous thromboembolism (VTE) is

tively ; difference

not statistically significant). Distal DVT recorded in 30.5% of patients (30.8% and 30.3% in the pre- and postop groups, respectively). DVTs were more common in patients ≥65 years old (54.2% versus 28.4%, p < 0.05); no significant differences were detected in terms of other characteristics. No significant differences were recorded in the number or type of bleeding was

complications: major (nonfatal) bleeding episodes were reported in five patients (2.8%, two and three in the preand postop groups); minor bleeding was noted in 25 (13.9%, 14 and 11in the pre- and postop groups). In conclusion, the present study suggests that a LMWH regimen started postoperatively is no less effective in preventing DVT in elective hip replacement than the classical regimen started preoperatively. Surprisingly, postoperative commencement offered no significant advantage in terms of bleeding complications. Key Words: Low-molecular-weight heparin—Deep vein thrombosis— Prophylaxis—Hip replacement—Bleeding complications.

a common

occurrence

complication in patients undergoing elective total hip replacement. The incidence of phlebographically proven deep-vein thrombosis (DVT) in the absence of drug prophylaxis is as high as 46-55% in the different studies (1); fatal pulmonary embolism (PE)

occurs

in 1-3% of

cases

of VTE (1); however, it has not been

possible to completely eliminate the induced risk of potential bleeding. Many orthopedic surgeons are, in fact, reluctant to make extensive use of effective antithrombotic drugs, mainly for fear of perioperative bleeding (3). In patients who have undergone joint replacement, bleeding at the site of operation with subsequent wound hematoma can be a significant problem, since it may result in failure of the procedure. The incidence of bleeding complications can be lessened either by using safer antithrombotic

(2). Several forms of

prophylaxis have been evaluated. Many of them have proved highly effective in lowering the rate of

agents [such Manuscript

received November 8, 1994; accepted

as

low-molecular-weight heparins

unfractionated heparin or oral anticoagulants] (1) or by delaying the start of prophylaxis until early after surgery. It is generally agreed that prophylaxis should be started before surgery, to blunt intraoperative and

(LMWHs)

January 6,

1995.

Address correspondence and reprint requests to Dr. G. Palareti at Department of Angiology and Blood Coagulation, University Hospital S. Orsola, Via Massarenti 9, 40138 Bologna,

Italy.

18

versus

19

immediately postoperative hypercoagulability and prevent the growth of small fibrin deposits, which often arise perioperatively, especially in calf veins. However, commencing prophylaxis postoperatively may still be an effective and safe procedure for preventing the extension of fibrin deposits and favoring their spontaneous lysis, possibly with a lower risk of perioperative hemorrhage. Once it is proved clinically effective in reducing the rate of postoperative VTE, this &dquo;delayed&dquo; form of prophylaxis might be preferred in those types of surgery where hemorrhagic complications are of major concern (hip, neuro-, or eye surgery). It would especially help persuade surgeons to routinely employ VTE prophylaxis in high-risk patients. Some studies have already shown that postoperative LMWH prophylaxis is clinically effective in reducing VTE versus placebo or no prophylaxis in hip surgery (4-7). No study to date, however, has directly compared the effectiveness of postoperative versus preoperative prophylaxis. The aim of this multicenter, randomized, double-blind study of patients undergoing elective hip surgery was to compare the efficacy of prophylaxis with Nadroparin (Seleparina Italfarmaco, spA, Milan, Italy) begun shortly after surgery or, as usually recommended, 12 h before operation. MATERIAL AND METHODS Patients In seven Italian orthopedic centers, patients of both sexes, 40-80 years old and weighing 55-90 kg, who were undergoing elective hip replacement were included in the trial. Premenopausal women were excluded, as were patients with any of the following

clinical conditions: underlying bleeding disorder and/or altered routine blood clotting tests (prothrombin time or activated partial thromboplastin time > 1.3 ratio [patient/control time], platelet count < 100,000 mm); previous DVT of a leg and/or PE; postthrombotic syndrome or chronic venous insufficiency ; previous hip or knee surgery or traumatic fracture of a leg; contraindication to venography; renal insufficiency, with creatinine >3.5 mg/dl; severe high arterial blood pressure, with a diastolic level > 120 mm Hg; need of anticoagulant or antiplatelet treatment; or history of stroke or leg paralysis. Patients obliged to rest in bed for >2 days before surgery and requiring spinal anesthesia were also excluded. Concomitant therapy with drugs known to interfere with hemostasis, such as oral anticoagulants, aspirin, nonsteroidal antiinflammatory drugs, and dextran, was discontinued --10 days before surgery and during the study period. The trial protocol was approved by the institu-

tional review boards of the seven participating hospitals ; informed written consent was obtained from patients before admission to the study. Patients were operated on in the morning, under general anesthesia, with moderate intraoperative hypotension (reduction :::s;50% of the baseline values). Intra- and postoperative hemodilution as well as hemotransfusion or infusion regimens were left to each participating center. After surgery, all patients underwent early mobilization, elastic bandaging of the legs or graduated elastic compression stockings, and physical exercise. Table 1 reports the demographic and clinical characteristics of patients considered for efficacy analysis in the two treatment groups as well as their operative and transfusion regimen data. A 4- to 6-week postoperative follow-up of patients was undertaken at operating center or, if not possi-

ble, by telephone inquiry.

Prophylaxis regimens A single batch of Nadroparin (Seleparina; Italfarmaco S.p.A, Milan, Italy) was used in the study. Twelve hours before surgery the patients received, in double-blind fashion, either 0.3 ml of Nadroparin (7,500 anti-Xa IC units; preop group) or placebo (postop group) subcutaneously. Afterward, patients of both groups received 0.3 ml s.c. Nadroparin (7,500 anti-Xa IC units) in the evening of the day of operation (8-12 h after the end of surgery) and once a day (o.i.d.) in the morning of the first, second, and third postoperative days. From the fourth day on, for 14 days or until discharge, all patients received 0.4 ml s.c. o.i.d. Nadroparin (10,000 anti-Xa IC units). This increase in the dosage was adopted because it proved to be effective in a previous study of prophylaxis in hip surgery (8). Blind preoperative administration of drug or placebo was therefore the only difference in regimen between the two groups. In each participating center, patients were assigned to the two treatment groups by means of progressively numbered boxes; randomization was done in blocks of 10 boxes. Each box contained progressively numbered prefilled syringes sufficient for the first 10 days of treatment. Other syringes, prefilled with 0.4 ml Nadroparin, were available if prolongation of treatment was necessary. The first syringe of each box contained either the drug or placebo according to ran-

domization ; syringes were identical, and patients, nurses, and physicians were unaware of their content.

Outcome measures DVT was assessed by bilateral ascending venography of the legs, which was performed between day 10 and day 15 postsurgery or earlier if there

20 TABLE 1. Clinical characteristics

a

symptoms

or

detectable in the venograms in any case. When only monolateral venograms were available, patients were considered for efficacy analysis only if DVT was detected in the examined leg. DVT was diagnosed when venography showed a constant intraluminal filling defect. Thromboses were rated as &dquo;proximal&dquo; when thrombi were located in the popliteal and/or femoral-iliac veins and as &dquo;distal&dquo; when thrombi were present only in the deep calf veins. Isolated thrombosis in the superficial communicating or muscular veins of the calf was not considered as a study end point since these veins do not belong to the deep venous system (10). was

two

study

groups

Lorenz formula.

signs of VTE. Guidelines for performing venography according to Rabinov and Paulin (9) were given to all radiologists; the method was, in fact, adopted in most centers. For technical reasons, however, in a few hospitals venography was performed with different techniques. All venograms were centrally evaluated by two independent radiologists unaware of the regimen assignment. If inadequately performed or not evaluable, venograms were discarded, and corresponding patients were excluded from efficacy analysis, unless DVT were

of the

Bleeding complications were classified as (a) major hemorrhages if they were intracranial, ocular (with reduction of visus), articular, retroperitoneal, and/ or associated with reduction of hemoglobin --2 g/dl need to transfuse 5=2 U of blood

or (b) minor were if hemorrhages they clinically evident but without the mentioned characteristics.

or a

Statistical analysis The incidences of DVT and bleeding episodes in the two treatment groups were compared using the test or Fisher’s exact test (significance level p 0.05, two-sided comparison) and calculating two-sided 95% confidence limits. The statistical analysis was performed using the software package SOLO (Version 4.0; BMDP, Los Angeles,

chi-square =

CA, U.S.A.). RESULTS of patients A total of 180 patients entered the

Population

study in the participating orthopedic departments, each of which enrolled from 11 to 65 patients. As reported in Table 2, one patient was excluded from the study seven

21 TABLE 2.

Study population (n

=

(four in each group), all with proximal and/or distal DVT at venography. Most DVTs were distal (40 patients, 78.4% of total DVTs); their incidence was almost identical in the two groups (Table 3). Proximal DVTs were detected in 11 of 131 patients (8.4% of the entire study group). They were noted in seven and four patients in the preop and postop groups, respectively; the

180)

difference, however,

a

In this case prophylaxis treatment was erroneously shifted from Nadroparin to unfractionated heparin the day after operation. b One patient had DVT detected on monolateral venography.

after randomization because of protocol violation (erroneous shift from Nadroparin to unfractionated heparin). Venography could not be performed in 14 patients (five and eight in preop and postop groups, respectively) for the following reasons: technical causes (presence of a plaster, n = 3), impossibility of performing a valid foot vein puncture (n 5), refusal of patient (n = 4), poor general conditions and/or bleeding complications (n 1), or early dis1). In nine patients (four and five in charge (n preop and postop groups, respectively), venography could be performed only in one leg; these patients were excluded from efficacy evaluation, except for one in whom DVT could be diagnosed anyway. Finally, in 26 patients (16 and 10 in preop and =

=

=

postop groups, respectively) phlebograms

were

considered by the two central radiologists inadequate to obtain a complete evaluation of the full deep-vein system; these patients were also excluded from final evaluation. Thus, 131 patients (65 and 66 in preop and postop groups, respectively) were eventually included in the efficacy analysis. As shown in Table 1, the two treatment groups were well matched in terms of all relevant characteristics. Prevalence and distribution of thrombosis As shown in Table 3, DVT was recorded in 51 of 131 patients at the time of efficacy analysis (38.9%). The total prevalence of DVT was very similar in the two groups (27 of 65 patients, or 41.5% in the preop and 24 of 66, or 36.4%, in the postop group). The 95% confidence limit of the difference between prevalences was - 0.11-0.22. In three patients, all in the postop group, pulmonary embolism was clinically suspected but was excluded on the basis of a normal perfusional scintigraphy; clinical symptoms remitted on subsequent days. Clinical signs of venous

thrombosis

were

detected in eight

patients

was

not

statistically signifi-

cant, and the 95% confidence interval of the difference between incidences was - 0.05-0.14. Table 4 shows the localization of DVT, which was very similar in the two groups. Most thrombotic episodes occurred in the deep calf vein system of the operated leg. With regard to the incidence of DVT in subgroups of patients (Table 5), thrombotic events were seen more frequently in patients aged ?65 years than in

patients (54.2% versus 28.4%, p < 0.05). significant differences could be detected in terms of other characteristics of the patients, such as sex, weight, presence of varicose veins, use of cement, operating time, use of intraoperative normovolemic hemodilution, preoperative autologous blood deposit, intraoperative blood salvage and erythrocyte reinfusion, and postoperative autoloyounger

No

gous blood transfusion.

Bleeding complications No significant differences could be detected in the number or type of bleeding complications found perioperatively in the two prophylaxis groups (Table 6). Five patients (2.8%, two and three in the preop and postop groups, respectively) experienced major (nonfatal) bleeding events in the early postoperative days, as assessed by a fall in hemoglobin level of >2 g/dl compared with the end-operation level. Bleeding was always local, presenting as increased blood loss from drains and/or the wound TABLE 3. Incidence

of deep-vein thrombosis efficacy analysis

at

CL, confidence limit; ns, not significant. Thrombosis of the popliteal vein and/or above. b Thrombosis of the deep calf veins (isolated muscular veins excluded). °

22 TABLE 4. Localization

a

Contralateral

leg proximal

+

both legs distal.

region; autologous blood was administered, and patients recovered without surgical revision. Minor bleeding events (increased drain volumes in three and wound hematoma in the others) were reported in 25 patients (13.9%, 14 and 11 in the preop and postop groups, respectively). Postoperative blood losses from drains were similar in the two groups, as were transfusion requirements. In 10 cases

cases

of deep-vein thrombosis

homologous blood was given postoperatively.

TABLE 5. Incidence of deep-vein thrombosis in the whole study group according to various possible risk factors

DISCUSSION surgery-a high-risk procedure for thromboembolic complications-prophylaxis with LMWH has proved to be more effective than low-dose heparin (LDH) ( 1,11,12), although it is no safer in terms of bleeding (11). Many orthopedic surgeons still do not routinely employ antithrombotic prophylaxis with effective drugs because of the fear of perioperative bleeding complications. In fact, they often assume that hemorrhage must be iatrogenic when prophylaxis with antithrombotic drugs is administered, especially if started preoperatively. It is likely that an effective prophylaxis with the first dose administered postoperatively would be better accepted by many surgeons (3). Since the early 1970s, when many studies established the efficacy of low-dose subcutaneous heparin in preventing postoperative VTE, preoperative administration of a first prophylactic dose has been recommended (13) and remains the usual regimen for either LDH or LMWH prophylaxis (14,15). Furthermore, a recent study by Planes et al. (6) comparing spinal and general anesthesia with different timing of prophylaxis regimens seemed to confirm that it is advantageous to start prophylaxis before rather than after surgery. Some studies, however, In

orthopedic

venous

Bleeding complications and blood loss in the prophylaxis groups at intention-to-treat analysis

TABLE 6. two

&dquo; DVT/total patients with deep vein thrombosis/total number of patients in the subgroup; n 131. RBC, red blood cell; Ht, hematocrit; ns, not significant. =

=

a

Increased blood loss from drains and/or wound hematoma. Increased drain volumes in three cases and wound hematoma in the others. b

23

have

already proved the efficacy of LMWH prophylaxis (versus placebo or LDH) commenced after orthopedic surgery (4-7), and the results of a recent survey showed that many orthopedic surgeons usually start prophylaxis during (26%) or even after (16%) operation (3). Despite these results, no study to date has directly addressed the problem of prophylaxis timing; to our knowledge, the present trial (multicenter, randomized, and double-blind) is the first to directly compare the efficacy and safety of prophylaxis begun before and after operation. Our study showed that there was no difference in either efficacy or safety between the two prophylaxis regimens. In fact, in the present study, the total DVT incidence was 38.9%, with no differences between the two study groups. This figure is much higher than that recorded by Leyvraz et al. (8), who used the same LMWH/dosage schedule adopted by us but found a DVT incidence as low as 12.6%. Our data, however, are in agreement with those obtained by the German Hip Arthroplasty Trial Group (GHAT) (16); they found a 33.1% DVT incidence using the same LMWH, though higher doses were administered during the first three treatment days (10,000 anti-Xa IC units). No confirmed cases of PE were recorded during the present study. With regard to proximal DVT, which is assumed to carry the highest risk of major pulmonary embolism (17), the total incidence recorded in our study was 8.4% (with no significant differences between the two subgroups). This figure is slightly lower than that recorded in the GHAT study (10.3%), though it is higher than the extremely low one obtained by Leyvraz et al. (2.8% (8) and consistent with those (ranging from 5.4% to 8.2%) recorded in other elective hip surgery prophylaxis studies, where different LMWH dosage schedules were

adopted (4,5,18-20). The relatively high incidence of distal DVT recorded in our study (~30% in both subgroups) raises certain questions. First, a number of venographies were rated as inadequate (unless found to be positive by another method) and the patients subsequently excluded from efficacy analysis. All these venograms had previously been rated normal by local radiologists. The failure to obtain readable venographies in a certain number of patients did not seem to bias the comparison between the two study groups, since the study was double-blind and the rate of inadequacy was no different. This circumstance might, however, have caused a reduction of the number of normal venographies included in the efficacy analysis. Second, there is the question of perioperative

transfusion regimens, often neglected in other prophylaxis studies. Most of the patients admitted to our trial (90%) underwent preoperative blood collection to lessen homologous blood use in case transfusions were required. In many patients intraoperative acute normovolemic hemodilution was performed, half the patients (88 of 180) having