Continuous intravenous administration of heparin followed by oral administration of warfarin is the standard treatment for thromboembolism4,6,8-12. Anticoagulation therapy prevents the formation of additional thrombi while allowing the fibrinolytic system to dissolve thrombi that have already formed8. Barritt and Jordan2, in a prospective study of thirty-five patients who had a pulmonary embolism, reported five deaths in a group of nineteen patients who received no treatment compared with no deaths in a group of sixteen patients who received intravenous heparin therapy. Similarly, in a retrospective review of 103 patients, Kanis13 reported that seventeen (77 percent) of twenty-two patients who did not receive anticoagulation therapy for the treatment of a pulmonary embolism died compared with only seven (9 percent) of eighty-one patients who received intravenous administration of heparin, oral administration of warfarin, or a combination of the two. A number of other reports in the orthopaedic literature have supported the efficacy of intravenous heparin therapy in the prevention of mortality after thromboembolic events14,15,19.
Despite the known effectiveness of anticoagulation therapy in reducing morbidity and mortality following thromboembolism, there has been concern among orthopaedic surgeons that the use of intravenous heparin therapy in the immediate postoperative period is associated with an unacceptably high rate of complications7,14,15. Specifically, Patterson et al.15 reported a complication rate of 51 percent (eighteen of thirty-five) among patients who were managed with intravenous heparin therapy within five days after total joint replacement. Other than avoiding anticoagulation within one week after an operation15, we are not aware of any guidelines for avoiding bleeding complications if intravenous heparin therapy must be used. In addition, we are not aware of any study in which the postoperative courses of patients who received intravenous heparin therapy for the treatment of a thromboembolic event were directly compared with those of a control group of patients who did not receive therapeutic anticoagulation—with use of the same criteria for complications in both groups—in order to better define the true prevalence of complications.
The purposes of the present study were (1) to determine whether the rate of bleeding complications associated with the use of intravenous heparin therapy after total hip or knee arthroplasty is as high as has been reported, (2) to assess whether patients who receive intravenous heparin therapy have a higher rate of bleeding complications than those who do not receive therapeutic anticoagulation for the treatment of a thromboembolic event, and (3) to identify factors predictive of bleeding events so that such problems can be avoided.
*Although none of the authors has received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this article, benefits have been or will be received but are directed solely to a research fund, foundation, educational institution, or other nonprofit organization with which one or more of the authors is associated. No funds were received in support of this study.
†Musculoskeletal Research Center, Room 1500, Department of Orthopaedic Surgery (C. J. D. V., L. M. J., J. I., S. A. S., and P. E. D. C.), Epidemiology Unit (R. N. H.), and Department of Medicine (D. J. S.), New York University—Hospital for Joint Diseases, 301 East 17th Street, New York, N.Y. 10003. E-mail address for C. J. Della Valle: craigdv@yahoo.com. E-mail address for P. E. Di Cesare: PEDicesare@aol.com.
A retrospective analysis of forty-four consecutive patients who received intravenous heparin therapy for the treatment of a postoperative thromboembolic event following a primary unilateral total hip or knee arthroplasty between August 1994 and January 1998 was performed. Thirty-seven patients received anticoagulation therapy because of a pulmonary embolism; three, because of an isolated proximal deep venous thrombosis; and four, because of both of these complications. This group of patients was compared with a control group of 376 consecutive patients who had a primary unilateral total hip or knee arthroplasty between January and June 1996 and were managed with prophylactic anticoagulation therapy only (enoxaparin; thirty milligrams administered subcutaneously twelve hours postoperatively and every twelve hours thereafter until the time of discharge). The control group received enoxaparin for a mean of seven days (range, four to nineteen days) postoperatively.
The total study group of 420 patients included 290 women and 130 men. The mean age was sixty-five years (range, twenty-three to ninety-two years), and the mean weight was eighty-one kilograms (range, forty-four to 146 kilograms). Two hundred and eleven patients had a primary total hip arthroplasty (147 patients had the acetabular component inserted without cement and the femoral component inserted with cement, fifty-two had both components inserted without cement, and twelve had both components inserted with cement), and 209 patients had a primary total knee arthroplasty (all components were inserted with cement). Total hip arthroplasty was performed because of osteoarthritis in 161 patients, osteonecrosis in twenty, rheumatoid arthritis in thirteen, posttraumatic arthritis in five, developmental dysplasia of the hip in three, acute fracture of the femoral neck in two, bone tumor in two, nonunion of the femoral neck in two, Paget disease in one, psoriatic arthritis in one, and ankylosing spondylitis in one. Total knee arthroplasty was performed because of osteoarthritis in 189 patients, rheumatoid arthritis in sixteen, osteonecrosis in two, and posttraumatic arthritis in two. Two hundred and fifty procedures were performed with spinal anesthesia and 170, with general anesthesia. The mean duration of the operation was 152 minutes (range, sixty to 240 minutes).
The forty-four patients who received intravenous heparin therapy were identified by means of a review of the medical records of 322 consecutive patients who had ventilation-perfusion scanning at our institution between August 1994 and January 1998. Two hundred and seventy-eight patients, including 238 who had a negative result on scanning, twenty-four who had had a procedure other than primary total hip or knee arthroplasty, fourteen who had received a filter in the inferior vena cava without adjunctive intravenous heparin therapy, and two who had received anticoagulation therapy more than two weeks postoperatively, were excluded from the present study. The forty-four patients who were included in the study group were evaluated by one of us (D. J. S.), a pulmonary-trained medical intensivist who also supervised the anticoagulation therapy. The evaluation included duplex ultrasonography for assessment of the deep venous system of the lower extremities as well as ventilation-perfusion scanning. Ten patients also were evaluated with pulmonary angiography.
Intravenous heparin therapy was begun at a mean of five days (range, one to thirteen days) postoperatively, with thirty patients (68 percent) receiving the initial dose of heparin on or before the fourth postoperative day. Thirty-six patients (82 percent) received a bolus of 5000 units of heparin intravenously at the initiation of anticoagulation therapy. The mean initial dose of heparin was 966 units per hour (range, 600 to 1200 units per hour), and the mean therapeutic dose was 1206 units per hour (range, 600 to 3100 units per hour). A therapeutic level of anticoagulation (an activated partial thromboplastin time of more than 1.5 times the control value) was achieved within twenty-four hours after the initiation of therapy in 64 percent (twenty-eight) of the forty-four patients. Oral administration of warfarin was initiated at a mean of three days (range, zero to thirty days) after the initiation of intravenous administration of heparin. Intravenous heparin therapy was discontinued when the international normalized ratio was greater than 2.0; the mean duration of intravenous heparin therapy was seven days (range, two to thirty-one days). Only one patient received intravenous heparin therapy for fewer than five days; this patient received such therapy for only two days secondary to a bleeding complication that necessitated the insertion of a filter in the inferior vena cava. Overall, seven (16 percent) of the forty-four patients had insertion of a filter in the inferior vena cava as part of the treatment of thromboembolism. Two patients received a filter because of recurrent thromboembolic events despite intravenous heparin therapy; two, because of bleeding complications secondary to intravenous heparin therapy; two, because they were thought to have poor cardiopulmonary reserve with severe hypoxemia; and one, because a pulmonary embolism was diagnosed on the first postoperative day and the attending surgeon believed that anticoagulation therapy was contraindicated at that time (intravenous heparin therapy subsequently was started on the fourth postoperative day).
Data Analysis
The forty-four patients who required intravenous heparin therapy for the treatment of a thromboembolic complication were compared with a control group of 376 patients who did not have a thromboembolic complication and who received enoxaparin as prophylactic anticoagulation only. The two groups were compared with regard to demographic parameters, the prevalence of complications, total transfusion requirements, and the duration of hospitalization. Complications were defined as wound-related problems (hematoma, infection, or dehiscence), nonoperative hemorrhage (gastrointestinal, intracranial, intraocular, or retroperitoneal bleeding), a total transfusion requirement of more than two units of packed red blood cells, thrombocytopenia (a platelet count of less than 100,000 per deciliter or a decrease of more than 200,000 per deciliter), excessive wound drainage (lasting for more than eight days or necessitating discontinuation of the anticoagulant medication), or other minor bleeding episodes. In the group of patients who received anticoagulation therapy because of a thromboembolic event, complications that occurred after the initiation of heparin therapy were recorded.
Other data that were recorded for the group that received heparin included the time-course of anticoagulation; the number of doses of heparin used; and serial measurements of the level of hemoglobin, the platelet count, and coagulation parameters. Any recurrence of thromboembolic disease despite treatment was also noted. In addition, an analysis was performed to compare the patients who did and did not have a bleeding or wound-related complication while receiving intravenous heparin therapy in order to determine if there were any factors that were predictive of these events.
Statistical analysis was performed with use of the Student t test, the Fisher exact test, or the chi-square test where appropriate; the level of significance was set at p < 0.05. Power analysis revealed that, with the sample size available (with use of a beta of 0.80 and an alpha of 0.05), a 15 percent difference in the rate of bleeding complications between the two groups of patients would be detected as significant.
Analysis of Patients Who Received Intravenous Heparin Therapy for the Treatment of Thromboembolism and Those Who Received Prophylactic Anticoagulation Only
A comparison of the forty-four patients who required intravenous heparin therapy for the treatment of a thromboembolic event following primary total hip or knee arthroplasty and the 376 patients who received only prophylactic anticoagulation with enoxaparin revealed that the two groups had similar demographic characteristics. The mean age was sixty-six years in the heparin group compared with sixty-five years in the control group (p = 0.36). Women accounted for 73 percent (thirty-two) of the forty-four patients in the heparin group compared with 69 percent (260) of the 376 patients in the control group (p = 0.57). The mean weight was eighty-two kilograms in the heparin group compared with eighty-one kilograms in the control group (p = 0.74), and the mean percentage of the ideal body weight was 143 percent in the heparin group compared with 140 percent in the control group (p = 0.60). The mean duration of the operation was 177 minutes in the heparin group compared with 149 minutes in the control group (p = 0.002), and the mean intraoperative blood loss was 551 milliliters in the heparin group compared with 447 milliliters in the control group (p = 0.09). Spinal anesthesia was used for 45 percent (twenty) of the forty-four patients in the heparin group compared with 61 percent (230) of the 376 patients in the control group (p = 0.04).
Four (9 percent) of the forty-four patients in the heparin group had a total of five complications. Three of these patients had received the initial dose of heparin on the second postoperative day, and the fourth had received the initial dose on the twelfth postoperative day. The complications included a requirement for more than two units of packed red blood cells (two patients), gastrointestinal bleeding (one patient), epistaxis (one patient), and excessive wound drainage (one patient). Two patients (5 percent) had recurrent nonfatal pulmonary emboli, which were diagnosed on the basis of new perfusion defects seen on repeat ventilation-perfusion scanning; both patients were managed with the insertion of a filter in the inferior vena cava. As therapeutic anticoagulation had been accomplished for both of these patients (the partial thromboplastin time was more than 1.5 times the control value), the recurrent thromboembolic events could not be attributed to inadequate anticoagulation. Heparin therapy was discontinued in two patients who had a bleeding complication; both of these patients were managed with the insertion of a filter in the inferior vena cava. There were no instances of wound hematoma or of intraocular, retroperitoneal, or intracranial bleeding. No patient died, and no patient required operative intervention for the treatment of a bleeding or wound-related complication.
Twenty-three (6 percent) of the 376 patients in the control group had a total of twenty-four complications. With the numbers available, this rate of complications could not be shown to be significantly different (p = 0.44) from the 9 percent rate in the heparin group. The complications included excessive wound drainage (ten patients), a requirement for more than two units of packed red blood cells (eight patients), thrombocytopenia (four patients), and gastrointestinal bleeding (two patients). There were no instances of wound hematoma or of intraocular, retroperitoneal, or intracranial bleeding. No patient died, and no patient required operative intervention for the treatment of a bleeding or wound-related complication.
The mean transfusion requirement in the heparin group (1.8 units of packed red blood cells) was significantly greater than that in the control group (0.8 unit) (p < 0.0001). The mean duration of hospitalization in the heparin group (fifteen days; range, seven to forty-one days) was significantly longer than that in the control group (seven days; range, four to nineteen days) (p < 0.0001).
Analysis of Patients Who Did and Did Not Have a Bleeding Complication During Intravenous Heparin Therapy
The number of patients who had a bleeding complication while receiving anticoagulation therapy for the treatment of a thromboembolic event was too small to allow for a statistical comparison between those who did and those who did not have such a complication during therapeutic anticoagulation. However, an analysis of these cases revealed that three of the four patients who had a bleeding complication had highly abnormal coagulation parameters at the time that the complication occurred. Two of these patients had a highly elevated activated partial thromboplastin time (ninety-eight and 135 seconds), and the third had a prothrombin time of 29.7 seconds (international normalized ratio, 5.7); these values are substantially higher than the levels of anticoagulation recommended for the treatment of a thromboembolic event (an activated partial thromboplastin time of 1.5 to 2.5 times the control value, or approximately forty to seventy seconds, and an international normalized ratio of 2.0 to 3.0). Three of the four patients who had a bleeding complication had received a bolus of heparin intravenously at the initiation of anticoagulation therapy, compared with thirty-three (83 percent) of the forty patients who did not have a bleeding complication.
Previous reports7,14,15,19 have suggested that the use of intravenous heparin therapy in the early postoperative period following a major orthopaedic operation is associated with a high rate of complications. Patterson et al.15 reported a complication rate of approximately 30 percent in a retrospective review of ninety-five patients who received heparin and warfarin after total joint arthroplasty and noted that the complication rate was 51 percent in the group of thirty-five patients who received anticoagulation therapy within five days after the operation. Bicalho et al.3, in a similar retrospective study, reported a complication in only three (11 percent) of twenty-eight patients who received intravenous heparin therapy within two weeks after total joint arthroplasty; only two of these complications (one wound hematoma and one case of transient thrombocytopenia) were hematological in nature. While both of those studies provide valuable information, neither included an analysis of whether the prevalence of complications in the study group was significantly different from that in a control group of patients who did not receive intravenous anticoagulation therapy with heparin for the treatment of a thromboembolic event.
To our knowledge, the present study is the first to compare the rate of bleeding complications in a group of patients who received anticoagulation therapy for the treatment of a thromboembolic event with that in a control group, with the same criteria for bleeding complications being used for both groups. In the present study, the rate of bleeding complications in the group that required intravenous heparin therapy for the treatment of a thromboembolic event was not significantly different from that in the group that received only prophylactic anticoagulation with enoxaparin; our findings might have been different, however, if an alternative form of thromboprophylaxis that caused fewer bleeding episodes had been used for comparison. In addition, with the sample size used in this study, a 15 percent difference in the rate of bleeding complications may have been detected as significant; accordingly, a smaller difference may have been detected as significant if the sample size had been larger. These findings suggest that the rate of bleeding complications associated with the use of intravenous heparin therapy for the treatment of thromboembolism in the early postoperative period following primary total joint arthroplasty may be more moderate than is commonly perceived and may be closer to the rate of 7 percent (two of twenty-eight) reported by Bicalho et al.3.
The patients in the heparin group had significantly greater transfusion requirements than did the patients in the control group (mean, 1.8 units of packed red blood cells compared with 0.8 unit) (p < 0.0001). This difference may be partially explained by the trend toward increased intraoperative blood loss in the patients who required intravenous heparin therapy (551 milliliters compared with 447 milliliters). While this increase in transfusion requirements is both clinically important and statistically significant, it appears to be justified given the efficacy of intravenous heparin therapy in the prevention of mortality after thromboembolic events2,13.
Heparin acts by binding to antithrombin III and potentiating its inhibitory effect on thrombin and activated factor X8,12. A bolus of heparin is usually given at the initiation of anticoagulation therapy, followed by continuous intravenous infusion. The hourly dose is then titrated to prolong the activated partial thromboplastin time to 1.5 to 2.5 times the control value12. Failure to exceed 1.5 times the activated partial thromboplastin time within twenty-four hours has been associated with a significant (five to fifteenfold) increase in the risk of both early and late recurrent thromboembolism (p < 0.001)10. There is wide variability in the patient response to heparin, necessitating individual dose titration and frequent monitoring16. Heparin is continued for five to ten days, whereas warfarin therapy is initiated with a goal of achieving an international normalized ratio of 2.0 to 3.011.
With the numbers available, we were unable to perform a statistical comparison of the patients who did and did not have a bleeding complication during the course of treatment with heparin and warfarin. An analysis of these cases did reveal a trend in that three of the four patients who had a bleeding complication had coagulation parameters that were substantially higher than the values recommended for the treatment of thromboembolism. This finding suggests that supratherapeutic levels of anticoagulation may be responsible for some of the observed bleeding complications.
Patterson et al.15 found that patients who received anticoagulation therapy in the early postoperative period had a particularly high rate of bleeding complications. Although the small number of patients who had a bleeding complication in our study precluded a meaningful statistical analysis, a trend was noted in that three of the four patients who had a bleeding complication received the initial dose of heparin on the second postoperative day. These data suggest that the very early postoperative period is the riskiest time for the initiation of intravenous heparin therapy.
It has been suggested that a bolus of heparin should not be administered at the initiation of anticoagulation therapy in patients who have had an orthopaedic operation3,15. Bicalho et al.3 hypothesized that the lower rate of bleeding complications in their study (7 percent; two of twenty-eight) may have been partially related to the fact that only four of their patients received a bolus of heparin. The results of the present study suggest that the administration of a bolus of heparin (used for 82 percent of our patients) is compatible with a low rate of bleeding complications. As rapid anticoagulation (within twenty-four hours after the initiation of anticoagulation therapy) has been shown to reduce both the short and the long-term sequelae of thromboembolic events10, this strategy for rapidly achieving a therapeutic level of anticoagulation is justified if coagulation parameters are closely monitored. The relatively small sample size in the present study limited our ability to determine conclusively whether factors such as the administration of a bolus of heparin are related to bleeding complications in patients receiving anticoagulation therapy for the treatment of thromboembolic disease; this is a weakness of the present study. The retrospective nature of the present study also limited our ability to make causal inferences. Such issues would be better addressed in a large prospective, randomized study.
In the retrospective analysis by Bicalho et al.3, the rate of complications associated with the use of intravenous heparin therapy was compared with that associated with the use of a filter in the inferior vena cava. There were three complications (one wound hematoma, one transient thrombocytopenia, and one recurrent pulmonary embolism) in the group of twenty-eight patients who received intravenous heparin therapy, compared with one complication (premature opening of the filter in the jugular vein) in the group of twenty-four patients who received a filter in the inferior vena cava. The authors concluded that the insertion of a filter in the inferior vena cava is a viable alternative to heparin therapy in the early postoperative period. Although other authors have documented satisfactory early results in association with the use of these devices after orthopaedic operations1,3,5,20,21, a notable rate of both early and late complications has been reported1,3 and the long-term effects of these devices, particularly in younger patients, are unknown. At our institution, the use of these devices is reserved for patients who have had a complication related to anticoagulation therapy, those who have had recurrent thromboembolic events, and those who have a high risk of mortality in the event of additional pulmonary emboli.
The patients who required intravenous administration of heparin for the treatment of a thromboembolic event were found to have had a significantly longer operative time (p = 0.002) than those who did not have a postoperative thromboembolic event. This observation has been made by others, including Sharrock et al.18, and it supports the hypothesis that an extended operative time is associated with an increased rate of postoperative thromboembolic events. Patients who had a thromboembolic complication also had a significantly higher percentage of procedures performed under general as opposed to spinal anesthesia (p = 0.04) and had a trend toward greater intraoperative blood loss (p = 0.09). The beneficial effects of neuraxial anesthesia in preventing deep venous thrombosis (including reduced intraoperative blood loss, which also has been hypothesized to reduce the rate of postoperative thromboembolism), also has been previously noted17,18, and our data support these findings.
In the present study, the duration of hospitalization was greatly increased for patients who had a thromboembolic complication (fifteen compared with seven days; p < 0.0001). Similarly, Patterson et al.15 found that the duration of hospitalization averaged thirty-five days for patients who received intravenous heparin therapy for the treatment of a thromboembolic complication. Despite the increasing acceptance of the use of fractionated heparin as outpatient treatment for thromboembolic disease, there currently is no information to support the use of such treatment as a strategy to reduce the duration of hospitalization in the postoperative period following total hip or knee arthroplasty. These data underscore the need for effective thromboembolic prophylaxis in the postoperative period.
Note: The authors thank Adele Jazrawi, B.S., for assistance with data collection.