Ann Vasc Dis Vol.5, No.3; 2012; pp 328 333 2012 Annals of Vascular Diseases doi: 10.3400/avd.oa.12.00049 Original Article Deep Vein Thrombosis in Orthopedic Surgery of the Lower Extremities Masatoshi Motohashi, MD, PhD, 1 Akira Adachi, MD, 2 Ko Takigami, MD, PhD, 2 Keishu Yasuda, MD, PhD, 2 Masayuki Inoue, MD, 3 Shigeyuki Sasaki, MD, PhD, 4 and Yoshiro Matsui, MD, PhD 5 To prevent pulmonary embolism due to deep venous thrombosis (DVT), we have treated 611 patients undergoing orthopedic surgery of the lower extremities with our protocol including pre- and postoperative ultrasonic venous screening and anticoagulant therapy if necessary. A total of 118 patients (19.3%) developed DVT. Among demographic and clinical factors, the site of operation (knee joint surgery: odds ratio 5.17), age (>60: odds ratio 3.91), and operation time (>120 minutes: odds ratio 4.52) were identified as significant risk factors of development of DVT. One patient received an infusion of urokinase for DVT of femoral vein, but no patients developed serious postoperative bleeding or pulmonary thromboembolisms. (*English Translation of J Jpn Coll Angiol, 2010, 50: 95-100.) Keywords: deep venous thrombosis, pulmonary embolism, ultrasonic venous screening Introduction Deep venous thrombosis (DVT) occurs frequently as a postoperative complication, particularly after orthopedic surgery, and increasing number of patients were referred to the outpatient clinic of vascular surgery. DVT has been diagnosed by venography, but there have recently been a number of reports suggesting the efficacy of lower limb vein ultrasonography in diagnosis of DVT. 1 Department of Cardiovascular Surgery, Hakodate Central Hospital, Hakodate, Hokkaido, Japan 2 Department of Cardiovascular Surgery, NTT East Corporation Sapporo Hospital, Sapporo, Hokkaido, Japan 3 Department of Orthopaedic Surgery, NTT East Corporation Sapporo Hospital, Sapporo, Hokkaido, Japan 4 Division of Medical Sciences, Health Sciences University of Hokkaido, Ishikari-gun, Hokkaido, Japan 5 Department of Cardiovascular Surgery, Hokkaido University Hospital, Sapporo, Hokkaido, Japan Received: April 10, 2012; Accepted: June 22, 2012 Corresponding author: Masatoshi Motohashi, MD, PhD. Department of Cardiovascular Surgery, Hakodate Central Hospital, 33-2, Honcho, Hakodate, Hokkaido 040-0011, Japan Tel: +81-138-52-1231, Fax: +81-138-54-7520 E-mail: motohasi.masatosi@orange.plala.or.jp *This article is English Translation of J Jpn Coll Angiol 2010; 50: 95-100. Ultrasonic screening of lower limb veins is also performed before and after surgery at our department. In this study, we validated our protocol for the early detection of DVT and prevention of pulmonary embolism primarily by ultrasonography of lower limb veins for patients undergoing orthopedic surgery of the lower limbs. Subjects and Methods Subjects were 611 consecutive patients, consisting of 462 who underwent hip surgery (HS group) and 149 patients who underwent knee surgery (KS group) by a single surgeon at our hospital in 2005 2006. They were 124 males and 587 females, with age of 56.6 ± 0.6 [mean ± standard error of the mean (SEM)] years. Surgical procedures were total hip arthroplasty (THA) in 426 patients, repair of fracture in 21, THA + fracture repair in 12, and others in 3 in HS group; total knee arthroplasty (TKA) in 50, repair of fracture in 17, and arthroplasty in 82 in the KS group. Mean operation time in all patients was 99.9 ± 1.9 min (Table 1). For preoperative screening and postoperative evaluation for DVT, bilateral lower limb venous ultrasonography was carried out in all patients within 1 week before and after surgery by a technician authorized as a vascular 328 Annals of Vascular Diseases Vol.5, No.3 (2012)
Table 1 Demographic and clinical data of 611 patients Age (years) 56.6 ± 0.6 (range: 14 92) Male / Female 124/487 Op. procedure Hip 462 Total hip arthroplasty (THA) 426 Fracture 21 Fracture & Replacement 12 others 3 Knee 149 Total knee arthroplasty (TKA) 50 Fracture 17 Fracture & Replacement 0 Plasty 82 Op. time (min) 99.9 ± 1.9 (range: 7 340) D-dimer 8.8 ± 0.2 (range: 0.1 47.2) DVT: deep venous thrombosis The averaged data are presented as the mean ± SEM (standard error of the mean). Deep Vein Thrombosis, Ultrasonic Screening examination specialist. Actually, ultrasonography was performed on the 2nd postoperative day in most patients despite the 1 week margin due to the setting of the day of surgery. Ultrasonography of the lower limb veins was performed using SONOS5500 PHILIPS 7.5 10 mhz. The examination sites were the bilateral soleal, fibular, gastrocnemius, popliteal, and femoral veins. Search for thrombi was made by checking the blood flow in the color Doppler mode. Body position during ultrasonography was supine, prone, or standing depending on the examination site unless it was impossible to adopt a posture due to fracture, etc. Thrombi were judged to be present when solid echoes were observed within veins or when venous lumen was not collapsed despite compression with echo probe. Organized thrombi were considered to be present when the brightness of the echo was higher than the surrounding area (Fig. 1). According to the Guidelines for the Prevention of Pulmonary Thromboembolism/Deep Vein Thrombosis, 1) the following treatments depending on the degree of DVT risk were conducted in both HS and KS groups. An elastic stocking was applied to the intact limb, and intermittent pneumatic compression was applied to the affected limb until the beginning of ambulation after 24 hours using calf pumps in HS group and foot pumps Fig. 1 Hyperechoic mass in vein. in KS group. Our protocol was not indicated for patients who had pre-existing DVT and those who had a history of DVT with floating thrombi detected by preoperative ultrasonography. If thrombi were detected by preoperative ultrasonography, or if there was a history of DVT (all histories including those under the care of other physicians and symptoms in the past), continuous infusion of heparin Annals of Vascular Diseases Vol.5, No.3 (2012) 329
Motohashi M, et al. Table 2 Comparisons of demographic and clinical data according to the presence or absence of DVT DVT ( ) (n = 493) DVT (+) (n = 118) p-value Age (years) 55.1 ± 0.7 63.2 ± 1.7 <0.0001 Sex Male 100 24 Female 393 94 NS Op. procedure Hip surgery 406 56 Knee surgery 87 62 <0.0001 Op. time (min) 93.7 ± 1.9 125.7 ± 5.4 <0.0001 D-dimer 7.9 ± 0.2 12.8 ± 0.9 <0.0001 DVT: deep venous thrombosis The averaged data are presented as the mean ± SEM (standard error of the mean). Table 3 Logistic regression of deep venous thrombosis (DVT) with several risk factors Risk Factor Coefficient SE Odds Ratio 95% Confidence Interval p-value Age (>60) 0.214 0.030 3.91 2.56 5.97 <0.0001 Sex 0.109 0.038 1.00 0.61 1.65 0.004 Op. procedure 0.242 0.048 5.17 3.36 7.94 <0.0001 Op. time (>120) 0.129 0.046 4.52 2.96 6.91 0.005 SE: standard error at 10000 U/day was initiated immediately after surgery. Warfarin was administered simultaneously but heparin was discontinued when the INR reached 2. Regardless of the preoperative condition, if thrombi developed after surgery, warfarin administration was also initiated with INR 2 as a target. In all patients administered warfarin, ultrasonography was performed after 3 months. Warfarin administration was discontinued if no thrombus was noted, warfarin administration was discontinued and an antiplatelet agent (aspirin) was administered for 3 months if thrombi were organized. Warfarin administration was continued for 3 more months if no change was observed in the thrombi. D-dimer (D-D) was measured in all patients within 1 week after surgery. Data were compared retrospectively between the patients with and without postoperative DVT formation. Statistical analysis was performed using SPSS ver.12.0 J at the 5% level of significance. Results (1) Comparison of clinical features between those who developed DVT and those who did not, and evaluation of factors of DVT DVT occurred in 118 (19.3%) of the 611 patients. It occurred in areas proximal to the femoral region in 4 (2 in HS group and 2 in KS group) and in the lower leg in 114. For comparisons, the subjects were divided into those who developed DVT [DVT(+) group, n = 118] and those who did not [DVT( ) group, n = 493] (Table 2). Age was significantly higher in the DVT(+) group [63.2 ± 1.7 vs. 55.1 ± 0.7 (years)], but gender difference was not significant. The prevalence of DVT was significantly higher in KS group. In the DVT(+) group, operation time was significantly longer, D-D level was significantly higher,. Table 3 summarizes the results of multiple regression analysis of causative factors of DVT. From the regression coefficient and odds ratio, the site of surgery (knee), old age (>60 years), and prolonged surgery (operation time >120 min) were particularly related to the high risk of DVT. 330 Annals of Vascular Diseases Vol.5, No.3 (2012)
Deep Vein Thrombosis, Ultrasonic Screening Table 4 Comparisons of demographic and clinical data according to the site of operation HS Group (n = 462) KS Group (n = 149) p-value Age (years) 59.3 ± 0.5 48.4 ± 1.8 <0.0001 Sex Male [DVT (+)] 59 (2) 65 (22) Female [DVT (+)] 403 (54) 84 (40) <0.0001 Difference of DVT development between male & female Significant difference of male < female (p <0.05) Tendency of male <female (p <0.1) Op. time (min) 83.5 ± 1.4 150.5 ± 4.2 <0.0001 DVT (+) 95.1 ± 5.6 153.4 ± 7.3 DVT ( ) 81.9 ± 1.4 148.5 ± 5.0 Difference of operation time between DVT (+) and DVT p <0.01 NS ( ) in each group DVT (+) 56 (12.1%) 62 (41.6%) <0.0001 Site of DVT Same side as operation 25 (44.6%) 57 (91.9%) Contralateral or both lower extremities 31 (55.4%) 5 (8.1%) <0.001 D-dimer 9.1 ± 0.2 7.9 ± 0.7 <0.05 DVT: deep venous thrombosis The averaged data are presented as the mean±sem (standard error of the mean). (2) Comparison of clinical features based on the site of surgery Paying attention to the site of surgery (hip or knee) as a factor particularly affecting the occurrence of DVT, we classified the patients into those who underwent hip surgery (HS group, n = 462) and those who underwent knee surgery (KS group, n = 149), and more closely evaluated the occurrence of DVT. Table 4 summarizes the results. HS group consisted of 59 males and 403 females with an age of 59.3 ± 0.5 years, and KS group consisted of 65 males and 84 females with an age of 48.4 ± 1.8 years. Age was significantly higher in HS group (p <0.001). Unlike the results obtained in (1) showing that the risk of DVT was higher in KS group and in older patients, the mean age was lower in KS group at a higher risk. Regarding HS group, DVT occurred in 56 (12.1%) of the 462 patients. Gender-wise, it occurred in 2 (3.4%) of the 59 males and 54 (13.4%) of the 403 females, being more frequent in females. In KS group, DVT occurred in 62 (41.6%) of the 149 patients, an incidence significantly higher than in HS group. Gender-wise, it occurred in 22 (33.8%) of the 65 males and 40 (47.6%) of the 84 females, the difference was not significant (p = 0.097) (Table 4). As for the relationship between the side of surgery and side of the occurrence of DVT, DVT occurred on the same side of surgery in 82 (69.5%) of 118 patients, and on the other side or bilaterally in 36 (30.5%). The side of DVT occurrence differed markedly between HS and KS groups. While DVT occurred on the same side of surgery in 57 (91.9%) of the 62 patients in KS group, it occurred on the other side or bilaterally in 31 (55.4%) of the 56 patients in HS group (p <0.001). D-D level was higher in HS group with a lower incidence of DVT, also contradicting the results suggested in (1) showing that D-D level increases with the incidence of DVT. The relationship between the operation time and incidence of DVT showed no marked difference according to the site of surgery. While the mean operation time was 95.1 ± 5.6 min in those who developed DVT but 81.9 ± 1.4 min in those who did not develop DVT with a significant difference (p <0.05) in HS group, it showed no significant difference between those who developed DVT and those who did not in KS group with a higher incidence of DVT. (3) Courses of treatment and outcomes in patients found to have thrombi by lower limb venous ultrasonography Of the 118 patients who developed DVT, thrombi were detected by preoperative ultrasonography or there was a history of DVT in 23, with none showing a floating thrombus. Heparin and warfarin were used postoperatively according to the protocol. None of those administered heparin showed postoperative hemorrhage that required Annals of Vascular Diseases Vol.5, No.3 (2012) 331
Motohashi M, et al. additional treatment including transfusion. Thrombi were also detected by postoperative ultrasonography in 12 (52.2%) of these patients. 23 patients consisted of 5 males and 18 females, of whom 1 and 11 (61.1%), postoperative developed DVT respectively. Of these patients, one in KS group who showed floating thrombi in femoral vein was administered urokinase. Eventually, of the 118 patients who developed DVT, medication was discontinued following ultrasonography 3 months after surgery in 64 (54.2%), and an antiplatelet agent was administered in 46 (39.0%). Warfarin administration for 3 months or longer was necessary in 16 (13.6%), but all thrombi were observed in distal regions such as the soleus, fibular, and gastrocnemius veins. Sustained administration of warfarin were performed in 1 patient with atrial fibrillation and one patient with mechanical valve. None developed pulmonary embolism. Discussion DVT is known to occur frequently after surgery, particularly orthopedic and gynecological surgery, and to cause pulmonary embolism, which often leads to a serious outcome. 2) As shown in the Guidelines for the Prevention of Pulmonary Thromboembolism/Deep Vein Thrombosis, 1) the prevention of DVT and its early diagnosis before the occurrence of pulmonary thromboembolism are important. Venography is effective for its diagnosis, but it is not appropriate as a screening method, because it has problems such as the induction of allergy and thrombus formation by the contrast medium, 3) and because it is difficult to perform frequently in a short period. Recently, ultrasonography has been reported to be useful for the diagnosis of DVT. 4,5) Yamada et al. 6) reported that the sensitivity and specificity of ultrasonographic assessment were 91 and 95%, respectively, and were comparable to those of venography. Ro et al. 7) also reported that thrombi in the soleal veins enlarged and caused pulmonary embolism in patients showing no swelling of the lower limb, and suggested the importance of the early detection of DVT. Ultrasonography, which is noninvasive and can be performed repeatedly even at the bedside, is considered to be suitable for screening. We have used it at our facility since 2004. The incidence of DVT has been reported to be 23.5% in overall, and to be 23% 33% in patients after hip replacement and 44% 58% in those after knee replacement. In our series, the incidence of DVT was 19.3% in all, 12.1% in hip surgery, and 41.6% in knee surgery. At our facility, the contents of surgery somewhat differed compared with the literature, e.g., knee surgery was TKA in many patients. However, the outcomes were comparable to previous reports, so that screening by ultrasonography is not considered to be inferior in sensitivity. According to the results of this study, the site of surgery (knee surgery, odds ratio 5.17), age (>60 years, odds ratio 3.91), and operation time (>120 min, odds ratio 4.52) were detected as factors that markedly affected the occurrence of DVT in all patients. Since D-D appears when thrombi are lysed in the thrombolytic process, the result that it was higher in those with DVT is reasonable. However, when the results are viewed according to the site of surgery, they were more or less contrary to our expectations. The incidence of DVT in all patients was 19.3% in both males and females, but it was higher in females both in those after hip surgery (p <0.05) and those after knee surgery (p <0.1). Moreover, DVT occurred in 61.1% of the females compared with 20% of the males with positive ultrasonographic findings of thrombi or a history of DVT. Therefore, attention to DVT is necessary in females, particularly those who were found to have thrombi on preoperative ultrasonography or had a history of DVT. Also, the risk of DVT increased with age in all patients, but the incidence of DVT was higher in younger patients after knee surgery, indicating that the site of surgery (knee surgery) is a more important factor than age. This is explained more clearly when the laterality of DVT is evaluated. In knee surgery, a tourniquet was applied to the limb that was operated on in all patients to reduce hemorrhage, and the duration of its application was nearly equal to the operation time. In most patients (91.9%) who underwent knee surgery, DVT occurred on the same side as surgery, so DVT is considered to have been caused by direct compression of the veins with the tourniquet. Since there is a report that the volume of hemorrhage could be reduced, and the visibility of surgical field could be secured, by hypotensive anesthesia without a tourniquet, 8) it is considered desirable to avoid the use of a tourniquet as much as possible in patients who tolerate hypotensive anesthesia. On the other hand, in those who underwent hip surgery, DVT often occurred on the other side or bilaterally, probably because of the rest and changes in the coagulation system due to surgery. D-D level was higher in HS group with a lower incidence of DVT, but mechanisms such as thrombus formation and thrombolysis unrelated to DVT may be present due to the high invasiveness of surgery. The evaluation of the usefulness of the protocol that 332 Annals of Vascular Diseases Vol.5, No.3 (2012)
Deep Vein Thrombosis, Ultrasonic Screening we prepared for this study is limited, because only comparison with the existing literature is possible. However, as mentioned above, the sensitivity and specificity of the detection of thrombi were comparable to those by venography. Warfarin and heparin, which were employed in this protocol, have no direct thrombolytic activity but prevent new thrombogenesis and are considered to be effective for the prevention of the release of floating thrombi from primary thrombi. Although evaluation is limited, because we did not perform pulmonary blood flow scintigraphy, etc., none of the 611 patients in this study developed pulmonary embolism that required treatment. Since floating thrombi in femoral vein derived from primary thrombi in the lower limb (which were detected in 1 patient and treated with urokinase in this study) have been suggested to be an important cause of pulmonary embolism, our protocol of examining the patients more frequently by venous ultrasonography, which can be performed more readily than venography, is considered to be more practical. Conclusion Screening for DVT was performed by lower limb venous ultrasonography before and after surgery in 611 patients undergoing orthopedic surgery of the lower limbs. Patients with high risk of DVT, and/ or development of DVT were treated using a protocol including anticoagulant therapy. The current results suggested that the site of surgery, age, and a long operation time were risk factors of the occurrence of DVT. Patients who underwent knee surgery under at a higher risk of DVT were younger, and gender appeared to be a more important factor (the incidence of DVT was higher in females). It is considered important to avoid the use of a tourniquet as much as possible in patients undergoing knee surgery. As a result of management using our protocol, no major hemorrhage or pulmonary embolism was observed postoperatively, but treatment with urokinase was needed in 1 patient in whom thrombi were detected in femoral vein. Since floating thrombi have been reported to be an important cause of pulmonary embolism, Screening is considered to be more simple and practical. Disclosure Statement The authors have no conflict of interest. References 1) Committee for the Preparation of the Guidelines for the Prevention of Pulmonary Thromboembolism/Deep Vein Thrombosis (Deep Vein Embolism) Guidelines for the Prevention of Pulmonary Thromboembolism/ Deep Vein Thrombosis. Tokyo. Medical Front International Limited 2004; 1: 96. 2) Monreal M, Ruíz J, Olazabal A, et al. Deep venous thrombosis and the risk of pulmonary embolism. A systematic study. Chest 1992; 102: 677-81. 3) Bettmann MA, Robbins A, Braun SD, et al. Contrast venography of the leg: diagnostic efficacy, tolerance, and complication rates with ionic and nonionic contrast media. Radiology 1987; 165: 113-6. 4) Raghavendra BN, Horii SC, Hilton S, et al. Deep venous thrombosis: detection by probe compression of veins. J Ultrasound Med 1986; 5: 89-95. 5) Hamper UM, DeJong MR, Scoutt LM. Ultrasound evaluation of the lower extremity veins. Radiol Clin North Am 2007; 45: 525-47 ix. 6) Yamada N, Fujioka H, Yazu T, et al. Diagnosis of Deep vein thrombosis as source of pulmonary thromboembolism. Comparison between Real-Time B-Mode ultrasonography and contrast venography. Jpn J Phlebol T(1) 1996; 7: 23-7. 7) Ro A, Kageyama N, Tanifuji T, et al. Histopathologital features about deep vein thrombosis resulting in fetal pulmonary thromboembolism from forensic autopsy. Jpn J Phlebol 2004; 15: 365-9. 8) Maruhasi Y, Okamoto S, Toribatake Y. Venous thromboembolism in TKA with anticoagulant therapy and without tourniquet. Jpn J Orthop Assoc 2009: s256. Annals of Vascular Diseases Vol.5, No.3 (2012) 333