THROMBOSIS RESEARCH 62; 5564,1991 0049-3848/91 $3.00 + .OO Printed in the USA. Copyright (c) 1991 Pergamon Press pk. All rights reserved.
IMPAIRED FIBRINOLYSIS AND POSTOPERATIVE THROMBOEMBOLISM IN ORTHOPEDIC PATIENTS. Bengt I. Et&son*,
Elsa Eriksson” and Bo l&berg”.
*Institutions of Orthopedics and ’Surgery, Gothenburg University, Gstra sjukhuset, S-416 85 Goteborg, Sweden.
in revised form 3.1.1991
by Editor U. Abildgaard)
ABSTRACT The incidence of deep vein thrombosis @VT) and pulmonary embolism was studied prospectively in patients undergoing elective total hip replacement. 96 patients were randomly allocated to receive either low molecular weight heparin (LMWH) or unfractionated heparin (UFH). All patients had bilateral phlebography and pulmonary perfusion/ventilation scintigraphy lo-12 days after surgery. The following fibrinolytic variables were analysed in plasma and related to thromboembolism: tissue plasminogen activator (t-PA) activity, t-PA antigen (t-PA Ag), plasminogen activator inhibitor (PAI- 1) activity and PAI- 1 antigen (PAI- 1 Ag). No significant difference was found, regarding the fibrinolytic response to surgery, between patients treated with LMWH and UFH. The level of PAI- 1 activity was significantly increased before operation in patients developing DVT as compared to non-DVT patients (p < 0.03). Immediately after surgery and in the morning the first postoperative day the levels of PAI- activity, PAI- Ag and t-PA Ag were positively correlated to thromboembolism. PAI- activity was the only preoperative fibrinolytic variable correlated to thromboembolism. ODUCTION The incidence of thromboembolism after major orthopedic surgery is high in spite of current prophylaxis. Preoperative identification of patients at risk of developing thromboembolism can provide one possibility to offer a more effective prophylactic regimen. Investigation of the fibrinolytic response to surgery can contribute to our understanding of thrombopathogenetic mechanisms. New techniques, using specific immunological and activity methods for quantitative measurement of tissue plasminogen activator (t-PA) and its inhibitor plasminogen activator inhibitor (PAI-l), have increased the possibilities for such studies. The balance between t-PA and its inhibitor PAI- determines the fibrinolytic activity in blood. The vascular endothelium produces both t-PA and PAI- 1. Defective fibrinolytic system is often due to an impaired extrinsic tibrinolytic pathway, i.e. reduced synthesis or release of t-PA from the vessel wall, or increased inhibition of t-PA . The inhibitor, PAI-1, seems to play a central role in the regulation of fibrinolysis after trauma (1,2). Defect fibrinolysis predisposes to spontaneous or recurrent (‘idiopathic’) thromboembolism and there are increasing evidence that the fibrinolytic system plays an -important role in the pathogenesis of postoperative DVT (3 - 9). Key Words:
Heparin, Low molecular weight heparin, Fibrinolysis, t-PA, PAI-1, Deep vein thrombosis, orthopedic Surgery, Total hip replacement 55
Vol. 62, Nos. l/2
The main purpose of this study was to test the hypothesis that a defective fibrinolytic system predisposes for DVT after orthopedic surgery.
METHODS Patients 96 consecutive patients, 68 f 8 years of age, admitted for elective total hip replacement were randomly allocated to receive subcutaneous injections of either low molecular weight heparin (LMWH) (Fragmin, Kabi, Stockholm, Sweden) 5,000 IU once daily (n = 49) or unfractionated sodium heparin (UFH) (Heparin, Kabi, Stockholm, Sweden) 5,000 IU three times daily (n = 47) for 10 days. All patients had bilateral phlebography and pulmonary perfusion/ ventilation scintigraphy lo-12 days after operation. Details of these techniques have been reported previously (10). A clinical follow up was performed 6 weeks after surgery. The protocol was approved by the Ethics Committe of the Gothenburg University and the Swedish National Board of Health and Welfare. The patients gave their informed consent before entering the study.
Biochemical analysis Cubital venous blood was sampled before operation, immediately after operation, the first postoperative day and one week after surgery. Venous blood was collected without application of tourniquet, between 8-10 a.m. after 15 min rest, except for the immediate postoperative sample. The blood was anticoagulated with one volume of 0.13 mol/l trisodiumcitrate for 9 volumes of blood. The samples were immediately transported on ice, centrifugated at 15,000 g and 4’C for 5 min, quick frozen in aliquots and stored at -70°C for batch assay. t-PA activity was determined with a spectrophotometric parabolic rate assay at two dilutions, l/2000 and l/4000 (11). The CV between days was 8.9% for 28.7 lU/ml and 19.2% for 0.5 IU/ml at 10 occasions. PAI- 1 activity was determined with a single point poly-D-lysine stimulator assay at two dilutions l/1000 and l/2000 respectively (12). The CV between days was 14.7% for 7.7 U/ml at 10 occasions. One unit of PAI- activity corresponded to the amount that inhibited 1 IU of single chain t-PA. Reference value for 40 normal individuals was 4.0 + 2.8 U/ml. t-PA Ag was measured by Imulyse t-PA Ag ELISA kit (Biopool AB, UmeIL, Sweden). The CV between days was 14.6% for 22.7 ng/ml at 10 occasions. Reference value for normal individuals was 3-10 n&nl. PAI- 1 Ag was measured by Tint Elize kit (Biopool, Ume& Sweden) whith a polyclonal antibody against PAI- that recognize the latent, free and complexed form to about the same extent. The CV between days was 14.7% for 11.0 rig/ml at lOoccasions. A pooled normal plasma had a value of 10.3 ng/ml.
Statistics To test for differences between patients with development of thromboembolism and patients without (named: DVT and non-DVT patients) a non-parametric test, the two-tail Mann-Whitney U was used. The two-tail Wilcoxon signed rank test was used in the comparison of two groups in paired samples. Spear-man’s nonparametric test was used in the analysis of the correlation between t-PA activity and t-PA Ag, PAI- activity and PAI- Ag and the correlation between t-PA and PAI- 1. p < 0.05 was considered statistically significant.
Vol. 62, Nos. l/2
RESULTS Total hip replacement was performed in 96 patients, 56 women and 40 men, mean age 69 f 8.0 years. Cemented prosthesis was used in 79 cases and non-cemented in 17 cases. Epidural anesthesia was given to 89 patients and general anesthesia to 7 patients. The mean duration of operation was 123 f 24 min and the mean blood loss was 1572 f 684 ml. The study was part of a prospective, double-blind, randomized, controlled trial. All patients with correct assessment of thromboembolism and complete sets of blood samples were included. Thromboembolism was found in 36 out of 96 patients’(38%), 14 in the LMWI-I group and 22 in the UFH group. DVT was confined to the the calf veins in 15 patients. Popliteal DVT was found in 3 patients and femoral DVT in 17 patients. One patient was not exposed to phlebography, due to evident pulmonary embolism which was verified scintigraphically. Proximal thrombosis was significantly more frequent in the UFH group as compared to the LMWH group, 13 as compared to 4 (p < 0.05). Thirty-two patients had DVT in the operated extremity, six of these patients had bilateral thrombosis and in three cases the thrombi were confined to the non-operated limb. Pulmonary embolism was detected scintigraphically in 19 patients (20%), 7 in the LMWH group and 12 in the unfractionated heparin group. Two patients had symptoms of pulmonary embolism. No patient died during the follow-up period. The mean age of DVT and non-DVT patients was 70+7 and 68 f8 years respectively. There was no significant difference in incidence of DVT between patients treated with cemented as compared to non-cemented prosthesis. There was no correlation between duration of operation or blood loss and postoperative thromboembolism.
The level of PAI- activity was significantly increased before operation in patients developing DVT as compared to non-DVT patients (p c 0.03). Immediately postoperatively the level of PAIactivity was doubled in patients developing DVT but remained unchanged in non-DVT patients (p < 0.001). One day after surgery the level of PAI- activity was increased in both DVT and nonDVT patients but the difference between the two groups was still significant (p < 0.02). One week postoperatively there was no difference between the two groups.
Plasma levels of PAI- Activity (U/ml) in Patients undergoing Total Hip Replacement. No thromboembolism n = 60 Median Range Preoperatively 15 min. postop. 1 day postop. 7 days postop.
z.; 21:5 9.8
0 0 0 0
- 31.9 - 51.6 - 88.1 - 64.8
Thromboembolism n=36 MediaIl Range 0 9.3 * 0 15.2 *** 35.0 * 3.9 0 11.8
- 43.2 - 86.9 - 131.9 - 44.3
The statistical significance of differences between groups is indicated: *pcO.O5 and ***pcO.OOl. PAL1 activity c 0.05 (U/ml) = 0.
Vol. 62. Nos. It2
Preoperative PAL1 Ag was not significantly elevated in the DVT group compared to the non-DVT group. PAI- Ag increased in DVT patients immediately after operation and there was a significant difference between DVT and non-DVT patients (p < 0.01). The next day a significantly increased level was found in both categories of patients, but DVT patients had a significantly higher level of PAL1 Ag as compared to patients without DVT (p c 0.01). One week postoperatively there was no difference between the two groups. PAI- 1 Ag correlated well with PAI- activity both pre- and postoperatively (r = 0.7) (p c 0.0001).
Plasma levels of PAI- Antigen (&ml) in Patients undergoing Total Hip Replacement. No thromboembolism n = 60 Median Range Preoperatively 15 min. postop. 1 day postop. 7 days postop.
12.1 13.8 28.2 21.8
3.2 6.4 7.1 8.7
- 43.7 - 48.9 - 79.3 - 93.4
Thromboembolism n = 36 Median Range 16.7 23.7 ** 39.5 ** 25.6
4.6 4.9 10.4 9.8
- 49.9 - 97.4 - 126.0 - 61.8
The statistical significance of differences between groups is indicated: ** ~~0.01. t-PA activity did not show any difference between DVT and non-DVT patients. t-PA activity was significantly increased immediately after surgery, as compared to preoperative activity. This was followed by a significant decrease in the t-PA activity in the morning the first postoperative day. One week after surgery the t-PA activity did not differ from preoperative values.
Plasma levels of t-PA Activity (W/ml) in Patients undergoing Total Hip Replacement. No thromboembolism n = 60 Range Median Preoperatively 15 min. postop. 1 day postop. 7 days postop.
0.32 0.68 ** 0.05 aa 0.20
0 - 1.24 0 - 7.41 0 - 2.06 0 - 2.80
Thromboembolism n = 36 Median Range 0.23 0.47 O0 0.05 a4 0.19
0 - 1.27 0 - 6.67 0 - 0.79 0 - 1.67
The statistical significance of changes within groups, as compared to preoperative values, is indicated: ” ~~0.01. t-PA activity < 0.05 (IWml) = 0.
Vol. 62, Nos. 112
Preoperative values of t-PA Ag did not show any difference between DVT and non-DVT patients. After surgery t-PA Ag increased significantly in both DVT and non-DVT patients, but the DVT group had significantly higher median values at a11 postoperative recordings. A negative correlation was found between t-PA Ag and t-PA activity preoperatively (r = -0.4) (p