Thrombogenesis in total hip replacement (THR) begins during surgery on the femur. This study assesses the effect of two doses of unfractionated intravenous heparin administered before femoral preparation during THR on circulating markers of thrombosis.
Seventy-five patients undergoing hybrid primary THR were randomly assigned to receive blinded intravenous injection of either saline or 10 or 20 U/kg of unfractionated heparin after insertion of the acetabular component. Central venous blood samples were assayed for prothrombin F1+2 (F1+2), thrombin-antithrombin complexes (TAT), fibrinopeptide A (FPA), and D-dimer.
No changes in the markers of thrombosis were noted after insertion of the acetabular component. During surgery on the femur, significant increases in all markers were noted in the saline group (P < 0.0001). Heparin did not affect D-dimer or TAT. Twenty units per kilogram of heparin significantly reduced the increase of F1+2 after relocation of the hip joint (P < 0.001). Administration of both 10 and 20 U/kg significantly reduced the increase in FPA during implantation of the femoral component (P < 0.0001). A fourfold increase in FPA was noted in 6 of 25 patients receiving 10 U/kg of heparin but in none receiving 20 U/kg (P = 0.03). Intraoperative heparin did not affect intra- or postoperative blood loss, postoperative hematocrit, or surgeon's subjective assessments of bleeding. No bleeding complications were noted.
This study demonstrates that 20 U/kg of heparin administered before surgery on the femur suppresses fibrin formation during primary THR. This finding provides the pathophysiologic basis for the clinical use of intraoperative heparin during THR.
PULMONARY embolism (PE) remains the most common cause of death after total hip replacement (THR). Postoperative anticoagulants decrease the risk of thromboembolism, but they are associated with an increased risk of wound hematoma or other bleeding complications. [2,3]Furthermore, low molecular weight heparin is associated with epidural hematoma, especially in orthopedic patients receiving epidural anesthesia. Pneumatic compression boots or foot pumps can be effective, but compliance can be poor. [7,8]A reliable means of decreasing the risk of thromboembolism and PE after THR, which also does not increase the risk of bleeding, would be beneficial.
During THR, there is minimal activation of thrombogenesis on the acetabulum, but an immediate activation of thrombosis occurs during surgery on the femur. [9,10]The increase in thrombogenesis is more pronounced using cemented rather than cementless femoral components and was related to time that the leg was held in flexion, adduction, and internal rotation, a position known to occlude the femoral vein. We subsequently demonstrated that an intravenous injection of 1,000 U heparin significantly suppressed the increase in one of the markers of thrombosis-fibrinopeptide A (FPA). FPA levels were not suppressed in 3 of the 12 patients who received 1,000 U heparin, suggesting that a higher dose may be needed in some patients.
In this study, a single intravenous bolus of heparin was administered on a weight-adjusted basis using two dosage regimens (10 and 20 U/kg) to determine which provides the most effective suppression of fibrin and, therefore, thrombus formation.
Materials and Methods
After Institutional Review Board approval, 75 patients undergoing primary hybrid THR signed informed consent forms to participate in the study protocol. Patients were randomly allocated into three groups to receive:(1) normal saline or (2) heparin in a dose of either 10 U/kg or (3) 20 U/kg, intravenously. The study medication had been prepared in the hospital pharmacy, the ampules identified by case number only, and the code was broken on completion of the study, facilitating a double-blinded study protocol.
All patients received hypotensive epidural anesthesia (HEA) as previously described. Patients were monitored with electrocardiography and pulse oximetry. Arterial pressure and central venous pressures (CVP) were monitored via radial artery and internal jugular catheters, respectively. CVP was maintained at 2-5 mmHg and mean arterial pressure (MAP) at 45-55 mmHg throughout surgery. Epidural anesthesia was performed at the L1-L2 interspace using 20-25 ml of 0.75% bupivacaine. Sedation was maintained with midazolam and fentanyl, as required to maintain patients comfortable but arousable. All patients breathed nasal oxygen, 3 1/min, spontaneously throughout surgery.
All surgical procedures were performed in the lateral decubitus position via the posterior approach. After femoral osteotomy, the acetabulum was reamed, and a cementless cup was inserted. Immediately thereafter, saline or heparin was administered. The leg was then flexed and internally rotated, the canal reamed, a trial prosthesis inserted and removed, and a cement restrictor placed. The canal was cleaned using pulsatile lavage and dried. Methylmethacrylate was injected into the canal in a retrograde fashion via a syringe, and then the femoral component was inserted. After hardening of the cement, the hip was relocated, and the wound was closed.
Five blood samples were taken at specific times during surgery: Sample 1-baseline, before epidural injection; Sample 2-immediately after insertion of acetabulum; Sample 3-five minutes after insertion of femoral prosthesis; Sample 4-two minutes after hip relocation; Sample 5-thirty minutes after surgery. Blood samples were drawn from the central venous cannula in the following manner: 4 ml of blood was withdrawn via a stopcock as a discard, and 8 ml of blood was withdrawn into a 10-ml syringe. Samples for prothrombin F1+2 (F1+2), thrombin-antithrombin complexes (TAT), and D-dimer were placed into citrated tube (0.13 M sodium citrate, 9:1 volume to volume), and samples for fibrinopeptide A (FPA) quantification were collected in tubes containing an anticoagulant solution, provided with the Asserachrom (Diagnostica Stago, Franconville, France) fibrinopeptide A kit. The blood was then centrifuged at 2000g for 15 min at 4 [degree sign]C. Plasma samples were immediately stored at -70 [degree sign]C until ready for assay. Plasma samples for FPA were treated with bentonite twice before performing the assay. F1+2 was measured by enzyme-linked immunosorbent assay using the Enzygnost F1+2 kit (Behring Diagnostics Inc., Somerville, NJ). FPA, TAT, and D-dimer were measured by enzyme-linked immunosorbent assay using the respective Asserachrom kits (Diagnostica Stago). Patients did not receive blood during the observation period for markers of thrombosis.
To determine whether the doses of heparin had any effect on bleeding, the operating surgeons were asked to rank subjectively the degree of bleeding: during incision, with exposure of the hip joint, and while reaming the acetabulum and femur. “Bleeding” was ranked subjectively as less than expected, normal, or excessive. The volume of intravenous fluid infused, the intraoperative blood loss (by weighing sponges), and postoperative hemovac drainage were measured. Pre- and postoperative hematocrits were measured, and the number of units of autologous or allogenic blood infused was recorded. All wounds were inspected daily for evidence of wound hematoma.
The alpha is set to 0.05 for each variable: F1+2, TAT, FPA, and D-dimer. No correction has been applied for these four variables, which are probably not independent of one another. A repeated measures analysis of variance (ANOVA) over time with heparin dose as an independent factor was done for each variable. A Bonferroni correction was applied for the within factor of time. Contrasts were used to compare saline to the 10-U/kg dose and the 10-U/kg to the 20-U/kg dose overall, and simple effects models were used at each time point. To render the data compatible with the underlying hypothesis of a normal distribution, the FPA and D-dimer data were transformed using the natural logarithm of the observed value, and for TAT, the square root of the observed value was used. Differences in blood loss were measured by 2 x 2 Mann-Whitney test. Transfusion requirements were analyzed using Fisher exact test, and the ranking of bleeding was analyzed with the Kruskal-Wallis test.
All patients completed the study protocol. There were no significant differences in patient demographic variables, volumes of intravenous fluid administered, or the duration of surgery between groups (Table 1).
Indexes of bleeding were similar between groups. Pre- and postoperative values of hematocrit, intraoperative blood loss, and postoperative drainage were similar (Table 2). There was no difference in transfusion requirements between groups. There were no wound hematomas, and no patients returned to the operating room to explore the joint for hematoma or infection, nor were there any significant bleeding complications from other sites.
The surgeon's assessment of bleeding demonstrated no differences between groups (Table 3).
Blood levels of the four markers-F1+2, TAT, FPA, and D-dimer-are shown in Figure 1, Figure 2, Figure 3and Figure 4, respectively.
Prothrombin F1+2 (Figure 1) levels did not increase after insertion of the cup in any of the three groups. Thereafter, values increased in all groups (P < 0.0001). F1+2 was significantly lower in the 20-U/kg group compared with the saline group at each point thereafter (P = 0.03, 0.001, and 0.04 at samples 3, 4, and 5, respectively). There were no differences between the 10- and 20-U/kg groups.
There was no significant increase in TAT (Figure 2) after insertion of the acetabulum in any of the three groups. Thereafter, values increased in all groups (P < 0.0001). There were no significant differences between groups at any sample point.
There was no significant increase in FPA (Figure 3) after insertion of the cup in any of the three groups. With insertion of the femoral component, FPA increased sevenfold from 10.8 to 73 ng/ml (samples 2-4) in patients receiving saline; FPA increased twofold from 9.6 to 22.2 ng/ml (samples 2-4) in patients receiving 10 U/kg heparin but did not increase in patients receiving 20 U/kg heparin (samples 2-4). FPA levels at samples 3 and 4 were significantly lower in both groups receiving 10 and 20 U/kg heparin compared with the saline group (P < 0.0001). FPA at the end of surgery was similar in all groups.
Examination of individual responses demonstrated that FPA levels increased to a variable extent during insertion of the femoral component in patients receiving heparin (Figure 5). A fourfold increase in FPA was noted during insertion of the femoral component in 6 of 25 patients receiving 10 U/kg heparin but in none of those receiving 20 U/kg (P = 0.03).
D-dimer (Figure 4) did not increase after insertion of the acetabulum in any of the three groups. With insertion of the femoral component, D-dimer progressively increased (P < 0.0001). There were no significant differences between groups at any of the observation points.
In this study, heparin was administered as a single intravenous dose after implantation of the cup. Heparin was to provide anticoagulation during surgery on the femur because during this phase of the operation thrombogenesis may occur. There was no activation of thrombin formation during surgery on the acetabulum. This protocol has two potential advantages:(1) there is a time delay between inserting the epidural catheter and administering heparin, and (2) the initial phase of the operation can be performed without anticoagulation. In previous studies, we used heparin either throughout surgery or at two points in time. 
These data demonstrate statistically significant suppression of fibrin formation (as demonstrated by FPA levels) during insertion of the cemented femoral component with doses of 10 and 20 U/kg heparin used as a single intravenous bolus after implantation of the cup. Twenty units per kilogram suppressed FPA increases in all 25 patients, whereas of those given 10 U/kg, 6 of 25 patients exhibited a fourfold increase in FPA, suggesting incomplete thrombin inhibition in one fourth of the patients.
The clinically effective dose of heparin has been studied in in vitro models and in clinical practice. [13,14]Minimal increases in activated partial thromboplastin time (aPTT) are usually sufficient to prevent venous thrombosis. In a previous study, we demonstrated that 1,000 and 1,500 U heparin increased the activated clotting time by about 30% or 50%, respectively. In the present study, we demonstrated that 10 U/kg heparin suppressed increases in FPA, but not as reliably as 20 U/kg. FPA is a more precise index of heparin action than aPTT because FPA measures the degree of inhibition of fibrin formation directly. The aPTT was not measured in this study.
Unfractionated heparin is known to act at multiple sites to prevent coagulation. There was some suppression of F1+2 and TAT levels, suggesting that the heparin may have suppressed factor Xa activity to some extent. This could be caused by either a direct inhibitory action on factor Xa or possibly by suppressing factor VII activation through heparin's effect on tissue factor pathway inhibitor. [16,17]The predominant mechanism of action of unfractionated heparin in this study, however, appears to be by functioning as a thrombin inhibitor, as FPA formation was significantly suppressed, especially with the 20-U/kg dose. By contrast, low molecular weight heparin is believed to act primarily as an Xa inhibitor. 
In these low doses, heparin has a short biological effect. We previously measured indexes of heparin action and noted resolution of the heparin effect in 20-30 min. Thus, if the surgery on the femur is completed within 30 min, effective prophylaxis can be achieved. To minimize the risk of developing venous thrombosis, the duration of potential femoral venous occlusion should be kept to a minimum. [9,19]In vitro studies have demonstrated that 10 min of complete venous obstruction with a thrombogenic stimulus is sufficient to initiate thrombosis. [20,21]In a previous study, we noted that the increase in F1 + 2 was related to the duration of femoral venous occlusion. Thus, we believe that expeditious surgery facilities the efficacy of heparin.
There was no evidence that the doses of heparin used increased intra- or postoperative bleeding. Intraoperative bleeding was assessed subjectively by asking the operating surgeon to rank the degree of bleeding during four separate phases of surgery and by measuring the blood loss intraoperatively. Postoperative bleeding was measured by wound drainage. None of these variables were altered by intraoperative heparin. Finally, the hematocrit values postoperatively were similar between groups.
Perhaps the main virtue of selective intraoperative heparin is that precise anticoagulation is used only when the patient is initially at risk of forming deep vein thrombosis. Current strategies tend to ignore the intraoperative genesis of deep vein thrombosis and focus on prevention of clot propagation after surgery. [2,3,10]If the intraoperative genesis of thrombin is not prevented, stronger anticoagulants may be required later, which may result in a significant incidence of postoperative bleeding complications. [22,23]
This study demonstrates that 20 U/kg of heparin administered 1 or 2 min before surgery on the femur suppresses fibrin formation during primary THR. This finding provides the pathophysiologic basis of the clinical use of intraoperative heparin during THR. A prospective series of patients in which 15-20 U/kg of unfractionated heparin is administered after insertion of the acetabulum is currently being performed. Clinical PE and evidence of clinical bleeding are being measured to further define the role of this approach to thromboprophylaxis with THR.
The authors thank Dr. Margaret Peterson for her assistance with the statistical analysis of the data.