Brian G. Rubin, MD, Jeffrey M. Reilly, MD, Gregorio A. Sicard, MD, and Mitchell D. Botney, MD, St. Louis, Mo.

Transcription

1 Care of patients with deep thrombosis in an academic Limitations and lessons venous medical center: Brian G. Rubin, MD, Jeffrey M. Reilly, MD, Gregorio A. Sicard, MD, and Mitchell D. Botney, MD, St. Louis, Mo. Purpose: The primary goal of our study was to review the quality of care in patients with deep vein thrombosis, with emphasis on identifying recurrent and remediable problems. Secondary goals were (1) to evaluate the use of the vascular laboratory and (2) to characterize our patient population with deep vein thrombosis and to identify a subset of patients with tmcomplicated deep vein thrombosis who might be candidates for outpatient therapy in the future. Methods: A retrospective review was performed for all patients with deep vein thrombosis diagnosed with duplex scanning who were treated as inpatients from January 1993 through March Results: Fifty-four (16%) of 306 duplex scans were positive; 50 patients were treated as inpatients. Forty percent of patients had uncomplicated deep vein thrombosis that was potentially treatable on an outpatient basis. Mean time to obtain a therapeutic partial thromboplastin time was 22 hours (range 4 to 54 hours). Ten (20%) patients had inferior vena cava filters placed. The in-hospital mortality rate was 4%. Management problems occurred in 18 (36%) patients and included difficulty titrating anticoagulation, (10) physician failure to provide treatment after diagnosis, (five) and inappropriate use or complication of inferior vena cava filter placement (three). Conclusions: Venous duplex examination is liberally but appropriately used. The primary remediable problem resulting in suboptimal management is difficulty titrating anticoagulation; inappropriate placement of inferior vena cava filters and physician failure to provide treatment also occur. In the future a substantial number of patients may be suitable for outpatient therapy. (J VASC SURG 1994;20: ) Lower extremity venous duplex scanning is an appropriate initial test for patients with suspected venous thrombosis. Because of the accuracy of this examination, it is now the first and often the only examination performed to confirm or eliminate the diagnosis in patients with suspected deep vein thrombosis (DVT).~ An interdisciplinary group of vascular surgeons and pulmonary physicians reviewed the care of patients with venous thrombosis admitted to a university medical center hospital. This article reports From the Section of Vascular Surgery, Department of Surgery, and the Division of Respiratory and Critical Care, Department of Medicine, Washington University School of Medicine, St. Louis. Presented at the Sixth Annual Meeting of the American Venous Forum, Wailea, Maul, Hawaii, Feb , Reprint requests: Brian Rubin, MD, Department of Surgery, 216 S. KAngshighway, St. Louis, MO Copyright 1994 by The Society for Vascular Surgery and International Society for Cardiovascular Surgery, North American Chapter /94/$ /6/ the results of our review of hospitalized patients with DVT with the use of DVT identified with duplex scanning as the primary inclusion criterion. The primary goal of this study was to review the type and quality of care rendered after DVT diagnosis, with emphasis on identifying recurrent and remediable problems. Because of the importance of rapid attainment of therapeutic levels of anticoagulation, we focused attention on identifying problems in the delivery of care that reproducibly delayed efficient therapy. Our other goals were (1) to evaluate the role of the vascular laboratory with regards to both the appropriateness of test use and the clinical ramifications of test results, and (2) to characterize our patient population in an attempt to identify a subset of patients with uncomplicated DVT who might in the future be candidates for outpatient therapy. MATERIAL AND METHODS The noninvasive vascular laboratory of The Washington University Medical Center maintains a

2 Vohtme 20, Number 5 Rubin et al. 699 computer registry of all patients studied since Patients treated between January 1, 1993, and March 31, 1993, with DVT identified with duplex scanning were retrospectively identified. Patients who did not undergo venous duplex imaging are not included in this study. The duplex scanning technique used in our laboratory employs a 5 MHz probe and color duplex scanner (Acuson 128 XP-10, Acuson, Inc., Mountainview, Calif.). Both legs were examined from the common femoral vein to the ankle level in continuity in both longitudinal and transverse planes. Attempts were always made to visualize the posterior tibial and peroneal veins; an image of the anterior tibial venous segment was not routinely obtained unless the anterior compartment of the leg was symptomatic. Compression of the underlying venous segment was performed at continuous levels over the entire length of the leg. Incompressibility of the venous segment was used for the diagnosis of DVT. Doppler flow characteristics were recorded including whether flow was spontaneous, phasic, augmented by distal compression, or pulsatile. Valvular competence was evaluated by response to proximal compression. Patients who were not hospitalized because of or at the time of their DVT diagnosis were excluded from this study. Four patients were treated as outpatients and are thus excluded from our analysis. Because patients were identified solely on the basis of venous duplex examination results, patients with DVT who did not undergo noninvasive imaging were excluded from our analysis. Patient demographics and risk factors for development of venous thrombosis were collected. The location of venous thrombi was tabulated. The days from symptom onset to diagnosis were collected, as was the number of hours from the time of duplex scanning diagnosis until a therapeutic anticoagulant level was documented. For patients being treated with intravenous heparin therapy, a partial thromboplastin time (PTT) of 1.5 to 2.5 times the control value was deemed appropriate therapy. 2 Subsequent conversion to warfarin was recorded with an acceptable prothrombin time of 15.5 to 21.5 seconds (international normalized ratio 2.0 to 3.0). a Acceptable indications for placement of an inferior vena cava filter included (1) contraindication to initiation of anticoagulant therapy, (2) complications of anticoagulant therapy requiring its discontinuation, and (3) pulmonary embolism despite therapeutic levels of anticoadflation. Care that varied from standard management algorithms was noted, as were problems achieving adequate levels of anticoagulation expeditiously, problems titrating anticoagulation, and indications and complications of inferior vena caval filtration. In the future uncomplicated DVT may be treated at home or with an abbreviated hospital stay, most probably with low molecular weight heparin-based regimens. Because of our interest in initiating a trial of outpatient treatment of DVT, patients were evaluated retrospectively regarding their candidacy for such therapy. Patients were considered candidates for outpatient therapy if their sole criterion for hospital admission was for treatment of an uncomplicated deep vein thrombosis. Patients with pulmonary symptoms and patients who were hospitalized for other reasons and subsequently experienced a DVT were not considered potential candidates for outpatient therapy. RESULTS Patient inclusion, exclusion, and mortality. From January 1993 through March 1993, 306 lower extremity venous duplex examinations were performed. Two hundred seventy-one (89%) studies were performed for signs or symptoms consistent with lower extremity venous thrombosis or acute pulmonary embolism, whereas 35 (11%) studies were screening evaluations. Fifty-four (16%) of the 306 studies demonstrated venous thrombus located in at least one venous segment between the level of the inguinal ligament and the malleoli. Four patients were not hospitalized and are excluded from this review. The medical records of all 50 inpatients were available for chart review and form the basis of this report. Forty-eight (96%) of 50 patients survived until hospital discharge for a hospital mortality rate of 4%. Demographics, risk factors, outpatient candidacy. Mean patient age was 68 years (range 25 to 93 years). Recognized risk factors included malignancy, recent major abdominal, pelvic, or orthopaedic surgery, history of deep venous thrombosis, immobilization, dehydration, inflammatory bowel disease, 4 and pregnancy. These risk factors were present in 46 (92%) of the 50 patients. Some patients had more than one risk factor. Table I tabulates the more common predisposing factors in our patient population. Twenty (40%) patients fulfilled the criteria to be considered as potential candidates for DVT treatment as outpatients. Distribution of venous thrombi. Venous thrombi were noted to occur with equivalent frequency in the right leg, left leg, or both legs (Table II). Clot was usually present in more than one venous segment with the superficial femoral and popliteal

3 700 Rubin et al November 1994 Table I. Risk factors in 50 patients with deep vein thrombosis identified with duplex scanning Previous DVT/PE 18 Recent surgical precedure 18 Malignancy 17 Prolonged immobilization 8 Inflammatory bowel disease 1 Pregnancy I DVT, Deep vein thrombosis; PE, pulmonary embolism. Table II. Distribution of venous thrombi by limb involved (n -- 50) Limb(s) involved No. of limbs (%) Right leg 17 (34) Left leg 17 (34) Both legs 16 (32) Table III. Distribution of venous thrombi by venous segment (n = 50) Right Left Common femoral vein Superficial femoral vein Popliteal vein Tibial/peroneal veins 11 9 Deep femoral vein 1 0 Table IV. Distribution of isolated venous thrombi (n = 50) Right Left Common femoral vein 0 1 Superficial femoral vein 5 3 Popliteal vein 4 3 Tibial/peroneal veins 1 1 Deep femoral vein 0 0 veins most frequently involved. DVT distribution by venous segment is presented in Table III. Venous thrombi involving only a single venous segment in a leg were deemed "isolated venous thrombi" and were also noted to occur most commonly in the superficial femoral and popliteal venous segments (Table IV). In two patients DVT was limited to infrapopliteal venous segments only. Time to diagnosis. Patients often noted symptoms for several days to weeks before seeking evaluation by their physician or before the physician obtained a diagnostic examination. Six of the 50 patients are excluded from this portion of the analysis either because their DVT was old (four) as documented by previous venous duplex scans in our laboratory with identical findings or because their symptoms were too vague or chronic to identify time of onset (two). In 21 patients a duplex scan was obtained on the same day as the onset of their signs or symptoms. Twenty-three patients had some delay between the onset of symptoms and the time of diagnostic evaluation, ranging from 1 to 30 days (mean, 7 days, median, 3 days). Eight (16%) patients had an interval of more than 7 days from symptom onset to DVT diagnosis. Therapy. An outline of the treatment plans is presented in Fig. Nine (18%) of the 50 patients did not receive any therapy. In four patients the DVT diagnosed with duplex was a chronic rather than an acute process. Two other patients had terminal illnesses at the time of diagnosis; one patient died, and the second patient was discharged to hospice. One patient had a previously placed inferior vena cava filter and a contraindication to anticoagulation, and no new therapy was initiated despite the development of a new venous thrombus. Most worrisome, two patients had thrombus limited to the superficial femoral vein. In both instances the physicians thought that the superficial femoral vein was part of the superficial rather than deep venous system, and thus these two patients were not treated with anticoagulation. Forty-one patients had some treatment initiated after the diagnosis of deep vein thrombosis. Anticoagulant therapy was selected in 38 patients, and primary IVC filtration was chosen for three patients. Anticoagulant therapy was initiated in 38 (76%) patients. Thirty-five (70%) patients were initially treated with heparin. Thirty-two of these 35 patients received an initial intravenous heparin bolus usually consisting of 5000 U but ranging from 3000 to 10,000 U. Thirty-four of the 35 patients were placed on a heparin drip, whereas one patient was treated with intermittent subcutaneous heparin injections. The initial heparin doses were based on physician estimates of patient requirements rather than with a nomogram. Heparin dosing was subsequently based on measurement of PTT performed several times daily. The mean time to obtain a therapeutic PTT was 22 hours (range 4 to 54 hours after initiating heparin therapy). Sixty-nine percent of patients received anticoagulant therapy within 24 hours of initiation of therapy. After certain orthopedic surgical procedures, patients underwent screening venous duplex examination as part of a standard protocol. In six patients unsuspected deep vein thromboses were

4 ]'OUKNAL OF VASCULAR SURGERY Volume 20, Number 5 Rubin et al EXCLUDED I I 50 INCLUDED I 38 ANTICOAGULATED I Fig. 1. Initial and subsequent therapies of patients DVT identified with venous duplex scanning are depicted. Fifty-four patients were initially identified, four of whom were not hospitalized and were therefore excluded from this report. identified; however, three patients were already receiving anticoagulant therapy with their warfarin anticoagulation. The three patients not yet adequately receiving anticoagulant therapy with warfarin had heparin added temporarily. These six patients were maintained with warfarin therapy for 3 months after the operation for therapy of their DVT. Three (8%) patients had bleeding while receiving anticoagulant therapy that was substantial enough to mandate cessation of their anticoagulant therapy. No serious complications occurred as a consequence of bleeding other than termination of anticoagulation. Two (5%) other patients had clinical findings consistent with pulmonary embolism despite therapeutic levels of anticoagulation. Heparin-associated thrombocytopenia requiring discontinuation of heparin therapy did not occur in any patient. Similarly, symptomatic thrombus propagation or development of new venous thrombosis was not noted in any patient. These event rates reflect clinically detected in-hospital occurrences only. In three patients a clear contraindication to initiation of anticoagulant therapy was present at the time of DVT diagnosis, and all three patients underwent placement of inferior vena cava filtration devices. In addition to the three IVC filters placed as primary therapy for DVT, seven additional patients from the group initially treated with anticoagulant therapy subsequently underwent placement of IVC Table V. Indication for IVC filter placement (n = 10) Contraindication to anticoagulation 3 PE despite therapeutic anticoagulation 2 Anticoagulation stopped because of bleeding 2 Not indicated 2 Unable to administer anticoagulation I PE, Pulmonary embolism. filtration devices. Indications for subsequent placement of an IVC filter included pulmonary emboli despite therapeutic anticoagulation (two), bleeding requiring discontinuation of anticoagulation (two), no acceptable indication (two), and unsuccessful anticoagulation despite massive doses of warfarin (one). The indications for placement of inferior vena cava interruption devices for all patients in this series are presented in Table V. Management errors. Management errors occurred in 18 (36%) patients. Six patients had PTT values (on heparin) greater than 2.5 times control or INR values (on warfarin) greater than 3.0 for more than 72 consecutive hours, with three of these patients requiring treatment with fresh-frozen plasma or vitamin K to correct the coagulopathy. Four patients received subtherapeutic anticoagulant therapy throughout their hospital stay. Five patients were not treated after diagnosis of an acute DVT.

5 702 Rubin et al. November 1994 Reasons for this included erroneous belief that the superficial femoral vein was not part of the deep venous system (two patients), diagnosis made but test results overlooked by care providers (one patient), therapy not begun for 4 days after diagnosis (one patient), and anticoagulant therapy stopped after 24 hours because of bleeding, without IVC filter placement. Errors also occurred relative to IVC filtration in three patients; two patients had IVC filters placed without appropriate indications, and one patient experienced a pnettmothorax after filter placement via an internal jugular vein approach. DISCUSSION The yield of positive duplex scans was surprisingly low at 16% for the entire group, and this percentage remains essentially unchanged when fractionated into examinations on patients with symptoms (44 of 271 examinations positive, 16%) or screening exams after orthopaedic surgical procedure (six of 35 examinations positive, 17%). The relatively low percent of positive studies suggests that our clinicians have a low threshold for ordering the examination, although an alternative explanation would bc that they are poor at identifying patients at risk for development of DVT. Two recent reports from other large university hospitals have shown that 25% s and 26% 6 of studies were positive in apparently comparable patient populations. Given our clinicians' high index of suspicion leading to a low threshold for obtaining venous duplex scans, the prolonged time from symptom onset to diagnosis presumably reflects patient delays in presentation, although physician failure to refer for definitive diagnostic evaluation expeditiously cannot be entirely excluded. Our patients with positive duplex scans usually had identifiable DVT risk factors, suggesting a potential role for selection criteria in an effort to obtain a higher yield of positive venous duplex scans. To date no validated set of clinical criteria are available that will generate a higher yield of positive venous duplex examinations without compromising overall diagnostic sensitivity. Though the question of which patients warrant scanning remains unresolved, we believe the limits of the extent of the duplex examination that are required can be answered. The superficial and popliteal segments are most frequently involved, similar to findings reported by Markel et al. 5 In our series, however, we were unable to confirm their findings of greater tendency towards left leg involvement. Indeed, our experience suggests equal propensity for right, left, or bilateral leg involvement. Moreover 18 (18%) isolated venous segments were identified in 100 lower extremities, including isolated tibial vein thrombi. Reports by both Markel et al.s and Messina et al. 6 have suggested the importance of scanning both the iliac and tibial venous segments in search of isolated thrombi, citing a significant yield ofthrombi that would otherwise have remained unidentified. However, when their series are combined, a total of only two isolated thrombi (0.5%) were seen in the IVC and iliac segments of 389 patients, with the isolated clot seen frequently in the tibial veins. Given the minimal likelihood of obtaining data that would have an impact on clinical decision making, we do not routinely obtain an image above the inguinal ligament. Both our report and the series cited previously justify our approach of obtaining an image from the inguinal ligament to the malleoli bilaterally, with little advantage to extending the examination proximally. The importance of rapid attainment of a therapeutic level of anticoagulation has been emphasized by Hull et al. 7,8 who noted a 15 times higher incidence of recurrent thromboembolism for patients with subtherapeutic heparin levels at 24 hours compared with that of patients with therapeutic levels. Wheeler et al.9 and Fennerty et al.10 documented subtherapeutic heparin levels at 24 hours in 60% to 78% of patients and suggested that lack of understanding of heparin kinetics and overcautious physician practices were responsible. Sixty-nine percent of patients treated with heparin in our series received anticoagulant therapy by 24 hours. Our experience parallels that ofraschke et al.n who found that 77% of patients received anticoagulant therapy by 24 hours with "standard care" regimens, compared with 97% of patients treated with weight-based nomograms and that the mean time to obtain adequate heparin levels in a standard care regimen was 22 hours. The substantial percent of patients in our series who keep receiving subtherapeutic anticoagulation while receiving heparin for 24 hours suggests that more aggressive heparin therapy early, institution ofa nomogram for heparin adjustment, or use of fixed-dose low molecular weight heparin preparations 12 will be necessary to avoid the risks of recurrent thromboembofism. Despite frequent subtherapeutic dosing our rates of recurrent thromboembolism are towards the lower end of the reported spectrum for patients receiving "standard" heparin regimens. 7,9,12 It must be noted that only clinically significant events occurring during hospitalization are tabulated, which probably results in substantial underestimation of the actual rate of occurrence during the entire period of anticoagulation.

6 Volume 20, Number 5 Rubin et al. 703 The question of whether a brief exposure to heparin in doses that result in a PTT time of longer than 2.5 times control substantially increases the risk of hemorrhage is open to speculation. 7 Problems with excessive anticoagulant therapy occurred in 10% of patients, although no clinical evidence of hemorrhage was noted in these patients. The combination of excessive or insufficient antithrombotic therapy, primarily resulting from problems with titration of anticoagulation during the first few days after DVT diagnosis, remains the single most common potentially remediable problem identified. The striking frequency of patients ultimately undergoing placement of IVC filters (20%) bears attention. Lack of understanding of appropriate indications and the ease of percutaneous insertion by nonsurgeons may be contributing factors to the increasing frequency of filter placement. Most (80%) of the inserted filters were appropriate, suggesting that contraindications to sustained anticoagulant therapy are more common than generally perceived. Despite the lack of mortality and modest morbidity (one pneumothorax) in our patients, ongoing monitoring of the appropriateness of IVC filters is indicated to prevent their overuse, t3 More disconcerting, however, were the reasons for some physicians' failure to treat patients appropriately after diagnosis. These lapses in care occurred despite an ongoing and active educational program for both attending and resident staff. Our findings have allowed us to focus subsequent educational programs, including refamiliarizing care providers with the nomenclature of the deep venous segments. Also, verbal and written confirmation between vascular laboratory personnel and medical care providers in all cases of acute DVT is now standard laboratory protocol. Despite a significant incidence of suboptimal management in this group of elderly patients with debilitation, hospital mortality rates remained surprisingly love at 4%. This appears quite acceptable when contrasted with the 12% in-hospital case fatalit T rate reported in the Worcester DVT study. 14 Though fixed doses of low molecular weight heparin preparations show equivalent efficacy with unfractionated intravenous heparin, 12 use of low molecular weight heparin preparations has not yet been adopted by our medical community. Because low molecular weight heparin regimens do not require titration of heparin dosing, this therapy could potentially be administered in an outpatient setting. Across the United States an unprecedented interest in limiting the cost of health care has arisen. Our awareness of this issue prompted us to consider initiation of a trial of limited stay or outpatient treatment for uncomplicated venous thromboses. The 40% of patients in our retrospective series who were candidates for such therapy suggests that substantial cost savings may be realized in the future without compromising patient care. Indeed, if problems associated with the initial titration of heparin were no longer a consideration, improved quality of care after DVT diagnosis may be the result. In conclusion, once duplex scanning confirms the diagnosis of acute DVT, suboptimal management occurs frequently, primarily resulting from difficulty in titrating anticoagulation. Other common problems noted were physician failure to provide appropriate therapy and poor understanding of the indications for IVC filter placement. Thus recurring patterns of suboptimal patient care have been identiffed and are being addressed within our institution by procedure monitoring, physician education, and efforts at improved communication at the time of diagnosis. Creation of a consultative service to advise on care of patients with venous thromboembolism is being pursued, along with institution-wide clinical trials of simplified but more effective anticoagulant regimens. REFERENCES 1. Aitken AGF, Godden DJ. Real-time ultrasound diagnosis of deep vein thrombosis: a comparison of venography. Clin Radiol 1987;38: Hyers TM, Hull RD, Weg JG. Antithrombofic therapy for venous thromboembolic disease. Chest 1992;102:408S- 425S. 3. Hirsh I, Dalen JE, Deykin D, Poller L. Oral anticoagulants: mechanism of action, clinical effectiveness, and optimal therapeutic range. Chest 1992;102:312S-326S. 4. Tsen TN, O'Brien JJ, Rubin BG, et al. Distinguishing characteristics of deep venous thrombosis (DVT) in patients with inflammatory bowel disease (IBD) compared to controls. Am J Gastroenterol 1993;104:A Markel A, Manzo RA, Bergelin RO, Stranduess DE. Pattern and distribution of thrombi in acute venous thrombosis. Arch Surg 1992;127: Messina LM, Sarpa SM, Smith MA, Greenfield LJ. Clinical significance of routine imaging of lilac and calf veins by color duplex scanning in patients suspected of having acute lower extremity deep venous thrombosis. Surgery 1993;114: Hull RD, Raskob GE, Rosenbloom D, et al. Optimal therapeutic level of heparin therapy in patients with venous thrombosis. Arch Intern Med 1992;152: Hull RD, Raskob GE, Hirsh J, et al. Continuous intravenous heparin compared with intermittent subcutaneous heparin in

7 704 Rubin et al. JOURNAL OF VASCULAR SURGERY November 1994 the initial treatment of proximal-vein thrombosis. N Engl J Med 1986;315: Wheeler AP, Jaquiss RDB, Newman JH. Physician practices in the treatment of pulmonary embolism and deep venous thrombosis. Arch Intern Med 1988;148: Fennerty AG, Thomas P, Backhouse G, et al. Audit of control ofheparin treatment. BMJ 1985;290: Raschke R_A, Reilly BM, Gnidry JR, et al. The weight-based heparin dosing nomogram compared with a "standard care" nomogram: a randomized controlled trial. Ann Intern Med 1993;119: Hull RD, Raskob GE, Pineo GF, et al. Subcutaneous low-molecular-weight heparin compared with continuous intravenous heparin in treatment of proximal-vein thrombosis. N Engl J Med 1992;326: Arnold TE, Karabinis VD, Mehta V, et al. Potential of overuse of the inferior vena cava filter. Surg Gynecol Obstet 1993; 177: Anderson FA, Wheeler HB, Goldberg RJ, et al. The population-based perspective of the hospital incidence and case-fatality rates of deep vein thrombosis and pulmonary embolism: the Worcester DVT study. Arch Intern Med 1991;151: Submitted March 2, 1994; accepted May 10, 1994.