Donald M. Arnold, MD; Susan R. Kahn, MD, MSc; and Ian Shrier, MD, PhD

Transcription

1 Missed Opportunities for Prevention of Venous Thromboembolism* An Evaluation of the Use of Thromboprophylaxis Guidelines Donald M. Arnold, MD; Susan R. Kahn, MD, MSc; and Ian Shrier, MD, PhD Objectives: To identify and characterize cases of potentially preventable venous thromboembolism (VTE): cases for which thromboprophylaxis was indicated, according to the American College of Chest Physicians (ACCP) consensus guidelines for VTE prevention, yet was administered inadequately. Design: A historical cohort study to examine all cases of deep vein thrombosis and pulmonary embolism from 1996 to 1997 at a large teaching hospital. Of these, we determined the proportion that was potentially preventable. We examined the reasons for inadequacy of prophylaxis and the setting in which preventable VTE occurred. Results: Of 253 objectively diagnosed cases of VTE in 245 patients, 44 cases (17.4%) were considered potentially preventable. This represented two thirds of all VTE cases for which thromboprophylaxis had been indicated (n 65). Of preventable cases, the most frequent reason for inadequacy of prophylaxis was omission of prophylaxis (47.7%), followed by inadequate duration of prophylaxis (22.7%), and by incorrect type of prophylaxis (20.5%). Surgical and medical indications for thromboprophylaxis that were common among preventable cases included nonorthopedic surgery, admission to hospital for pneumonia, and stroke with lower limb paralysis. Underlying risk factors for VTE that were common among preventable cases included recent immobility, active cancer, and obesity. Conclusions: One of six cases of all VTE and two of three cases of VTE for which thromboprophylaxis had been indicated could potentially have been prevented had physicians followed the recommended ACCP guidelines. Inadequacy of prophylaxis was most often caused by omission of prophylaxis. Missed opportunities for prevention occurred most commonly in the settings of nonorthopedic surgery, pneumonia, and stroke. (CHEST 2001; 120: ) Key words: guideline implementation; prevention; thromboprophylaxis; venous thromboembolism Abbreviations: ACCP American College of Chest Physicians; DVT deep vein thrombosis; PE pulmonary embolism; VTE venous thromboembolism Venous thromboembolism (VTE) is a major cause of morbidity and mortality in hospitalized patients. It is estimated that 250,000 patients are hospitalized annually in the United States with VTE 1, and the overall crude mortality rate from pulmonary embolism (PE) at 3 months has been *From the Department of Medicine (Dr. Arnold) and the Center for Clinical Epidemiology and Community Studies (Drs. Kahn and Shrier), Sir Mortimer B. Davis Jewish General Hospital, Montreal, Canada. Drs. Kahn and Shrier are Clinical Research Scholars supported by the Fonds de la Recherche en Santé du Québec, Canada. Manuscript received July 11, 2000; revision accepted June 13, Correspondence to: Susan R. Kahn, MD, MSc, Center for Clinical Epidemiology and Community Studies, Sir Mortimer B. Davis Jewish General Hospital, 3755 Cote Ste, Catherine Rd, Room A-118.1, Montreal, Quebec, Canada, H3T 1E2; susank@epid.jgh.mcgill.ca reported to be as high as 17.4%. 2 Given that deep vein thrombosis (DVT) is often clinically silent and PE may be rapidly fatal, prevention is the most effective strategy to reduce the burden of VTE and has been clearly shown to be cost-effective through a reduction in both fatal complications and treatment requirements. 3 Every 3 years since 1986, the American College of Chest Physicians (ACCP) has published comprehensive guidelines for the prevention of VTE, most recently published in January The guidelines identify risk groups of patients who should receive thromboprophylaxis, and they recommend the type of prophylaxis that is most appropriate for each risk group. The recommendations are formulated by experts in the field after critical review of the published literature and are categorized on the basis 1964 Clinical Investigations

2 of the strength of the supporting evidence. These evidence-based recommendations are generally considered to be the standard of care for DVT and for PE prevention. However, studies 2,5,6 suggest that, in practice, implementation of the guidelines may be inconsistent and inadequate. In this study, we attempted to identify the obstacles that hinder the success of VTE prevention at our institution by examining the rate of preventable VTE during a 1-year time period. Preventable VTE was defined as objectively diagnosed DVT or PE that occurred in a setting in which thromboprophylaxis was indicated but was either administered inadequately or not administered at all. The indications for thromboprophylaxis that were used in this study were outlined in the 1995 ACCP guidelines on VTE prevention, 7 according to the most recent version of the guidelines available at the time of this study. Among preventable cases of VTE, we examined the reasons for inadequacy of prophylaxis by comparing the prophylaxis regimens used with the prophylaxis regimens recommended by the guidelines. In addition, we compared preventable VTE with nonpreventable VTE, ie, VTE that occurred despite proper adherence to the guidelines, with respect to patient characteristics and the settings in which VTE occurred. Materials and Methods We conducted a historical cohort study in which we performed a chart review on all patients with objectively diagnosed DVT or PE, who were admitted to the Sir Mortimer B. Davis Jewish General Hospital in Montreal, a 637-bed McGill University teaching hospital, between October 1996 and October This 1-year period was chosen because (1) it allowed sufficient time for dissemination and implementation of the 1995 ACCP guidelines on VTE prevention, 7 and (2) it preceded the introduction of an emergency department-based outpatient DVT treatment program, for which recorded clinical information might have been less complete than for VTE patients admitted to hospital. Prior to its initiation, the Research Ethics Committee of the hospital approved the study protocol. Each episode of VTE was reviewed to determine whether it occurred in a setting in which thromboprophylaxis had been indicated and, if so, whether thromboprophylaxis was adequately administered. To assess indication for and adequacy of thromboprophylaxis, we collected data that allowed us to categorize patients into different risk categories for VTE as specified in the 1995 ACCP guidelines. 7 Using a standardized data collection form, we abstracted chart data on method of VTE diagnosis, patient characteristics, risk factors for VTE, indications for prophylaxis, and the thromboprophylaxis received. Based on these data, each case of VTE was categorized as preventable, nonpreventable, spontaneous, or ineligible for thromboprophylaxis as detailed below. VTE Diagnosis Objective diagnosis of DVT was established by Doppler ultrasonography, compression ultrasonography, venography, or autopsy. A thrombosis was considered proximal if it involved the popliteal vein or any segment proximal to it. The diagnosis of PE was established by ventilation/perfusion scanning read as high probability for PE with a moderate or high clinical suspicion, pulmonary angiography, infused CT, echocardiography, or autopsy. Patient Characteristics and Risk Factors for VTE Patient age and gender were recorded. Data on VTE risk factors, as outlined by the ACCP guidelines, were extracted. These included prior DVT or PE; history of cancer and active treatments thereof; use of oral contraceptive pill or hormone replacement therapy; known hypercoagulable state; obesity; varicose veins; fracture of the pelvis, hip, or lower extremity; immobility for at least 3 days before VTE diagnosis; and the presence of an indwelling central venous catheter. Indications for Thromboprophylaxis The ACCP guidelines identify clinical events, both surgical and medical, for which prophylactic antithrombotic therapy is indicated. Surgical indications for thromboprophylaxis include elective total hip or total knee arthroplasty, surgical hip fracture repair, general surgery (which comprises vascular, urologic, and cardiac surgical procedures and any orthopedic procedures other than total hip arthroplasty, total knee arthroplasty, and surgery for repair of hip fracture), intracranial neurosurgery, and multiple trauma. General surgical procedures were graded as low risk, moderate risk, high risk, or very high risk based on the patient s age, the number of risk factors for VTE, and the duration of the surgery, as outlined in the ACCP guidelines. 7 Medical indications for thromboprophylaxis included admission to the hospital for myocardial infarction, congestive heart failure or chest infection, stroke with lower extremity paralysis, and acute spinal cord injury with paralysis. Data on these clinical events were extracted. If any events occurred during the 6 weeks before VTE diagnosis, we considered them to be possible precipitants of VTE, and prophylaxis was indicated; otherwise, prophylaxis was not indicated. Adequacy of Thromboprophylaxis The ACCP guidelines recommend thromboprophylaxis according to risk category. Several acceptable types of prophylaxis are suggested, and the recommended dose, frequency, and duration are specified in the guidelines. To determine the adequacy of thromboprophylaxis, for each case of VTE we compared the actual prophylaxis used with the ACCP guideline recommendations. A priori, thromboprophylaxis was considered adequate if all of the following conditions were met: (1) the prophylaxis used was the same type and at least the same dose and frequency of dosing as that recommended in the guidelines; (2) prophylaxis was administered regularly for at least 7 days, until hospital discharge, or until the patient was ambulatory; and (3) prophylaxis was initiated within 24 h of the precipitating event (eg, surgery, myocardial infarction). Otherwise, thromboprophylaxis was considered inadequate. Classification of VTE Based on the above-mentioned data, each case of VTE was categorized as preventable, nonpreventable, spontaneous, or ineligible for our analysis. Preventable cases of VTE were defined as VTE that occurred in a setting in which prophylactic antithrombotic therapy had been indicated but was administered inadequately or omitted. Nonpreventable cases of VTE were CHEST / 120 / 6/ DECEMBER,

3 those that occurred despite adequate thromboprophylaxis. Spontaneous cases were those cases of VTE that occurred in the absence of a documented indication for thromboprophylaxis. Ineligible cases consisted of cases of VTE that occurred in patients who were ineligible for thromboprophylaxis either because of a contraindication or because the VTE occurred in another institution and the details of the event and of the use of prophylaxis were unavailable. Analysis The rate of preventable VTE and the specific reasons for inadequacy of thromboprophylaxis were examined. The clinical characteristics of preventable vs nonpreventable cases were compared using a comparison of simple proportions. Results During the 1-year study period, 253 cases of acute VTE were objectively diagnosed in 245 patients (Table 1). The mean ( SD) age was years, and 57.6% of case subjects were women. There were 48 cases of PE and 230 cases of DVT, most of which were proximal (62.6%) and unilateral (78.7%). Twenty-five patients had both DVT and PE concomitantly. Of the 253 cases of VTE, 44 cases (17.4%) were classified as preventable, 21 cases (8.3%) were nonpreventable, 179 cases (70.8%) were spontaneous, and 9 cases (3.6%) were ineligible for thromboprophylaxis. Therefore, of the 65 cases of VTE for which thromboprophylaxis had been indicated, inadequate prophylaxis was administered in 44 cases (67.7%). The reasons for inadequacy of thromboprophylaxis in the 44 preventable cases of VTE were examined (Fig 1). In 21 cases (47.7%), prophylaxis was omitted entirely. In 10 cases (22.7%), prophylaxis was administered for an inadequate duration of time. In nine cases (20.5%), the type of prophylaxis administered was other than that stipulated in the ACCP guidelines. Table 2 outlines the various types of prophylaxis used that were considered inadequate and compares them with the recommended type of prophylaxis according to each indication. In six cases (13.6%), the frequency of daily dosing of the anticoagulant was inadequate. In three cases (6.8%), prophylaxis was initiated after a delay of 24 h from the time of the precipitating clinical event. In an additional three cases (6.8%), the dosing of anticoagulant was less than that recommended in the guidelines; this included two patients who received warfarin appropriately (one after total hip replacement surgery and one after surgical repair of a hip fracture), but the measured international normalized ratio was never therapeutic until the seventh postoperative day. In six cases, there were two reasons, and in one case there were three reasons for prophylaxis being classified as inadequate. Table 1 Characteristics of Patients and VTE Cases and Category Assignment of VTE Cases (n 253)* Characteristics Data Study cohort Patients 245 Cases of VTE 253 Mean age, yr Female patients, % 57.6 Objectively confirmed VTE (n 253) DVT 230 (91.0) PE 48 (19.0) DVT and PE 25 (9.9) DVT characteristics (n 230) Proximal DVT 144 (62.6) Distal DVT 86 (37.4) Unilateral DVT 181 (78.7) Bilateral DVT 49 (21.3) Case category (n 253) Preventable 44 (17.4) Nonpreventable 21 (8.3) Spontaneous 179 (70.8) Ineligible for thromboprophylaxis 9 (3.6) *Data are presented as mean SD or No. (%) unless otherwise indicated. Figure 1. Reasons for inadequacy of prophylaxis in preventable VTE (n 44) Clinical Investigations

4 Table 2 Cases of VTE in Which the Type of Prophylaxis Was Considered Inadequate, and the Prophylaxis Recommended by the ACCP Guidelines According to Indication* Cases Indication for Prophylaxis Prophylaxis Administered Recommended Prophylaxis 1 TKA LDUH, 3,500 U sc tid; warfarin, 7.5 mg/d po LMWH sc fixed dose bid, or IPC 2 TKA LDUH, 3,500 U sc tid for 1 d, then 5,000 U LMWH sc fixed dose bid, or IPC sc tid for 4 d; warfarin, 2.5 mg/d po 3 Hip fracture surgery LDUH, 3,500 U sc warfarin for 3 days, then 2500 U sc tid for 4 d, then 5000 U sc bid LMWH sc fixed dose bid, or oral anticoagulation (INR ); adjuvant IPC may provide additional benefit 4 Very-high-risk general surgery laparotomy LDUH, 5,000 U sc bid LDUH and IPC, or LMWH and IPC, or dextran and IPC 5 Very-high-risk general surgery CABG LDUH, 5,000 U sc bid for 10 d, then 5,000 Usctidfor5d LDUH and IPC, or LMWH and IPC, or dextran and IPC 6 Moderate-risk general surgery vascular surgery ASA, 325 mg/d po for 2 d ES or IPC or LDUH, 5,000 U sc bid 7 Neurosurgery craniotomy ES IPC with or without ES or LDUH; IPC and LDUH may be more effective in combination 8 CVA and paralysis ECASA, 325 mg/d po LDUH or LMWH; IPC and ES are also probably effective 9 Pneumonia Warfarin for 4 d, INR never therapeutic LDUH or LMWH *TKA total knee arthroplasty; CABG coronary artery bypass grafting; CVA cerebral vascular accident; LDUH low-dose unfractionated heparin; sc subcutaneous; ASA acetylsalicylic acid; ECASA enteric-coated acetylsalicylic acid; INR international normalized ratio; IPC intermittent pneumatic compression; LMWH low-molecular-weight heparin; ES elastic stocking. Clinical characteristics of cases of preventable vs nonpreventable VTE were compared (Table 3). Both groups were similar with respect to mean age and sex distribution. Bilateral DVT occurred more frequently in the nonpreventable group compared with the preventable group (55.0% vs 22.7%, respectively), as did distal DVT (70.0% vs 56.8%, respectively), whereas PE occurred with similar frequencies in the nonpreventable and preventable groups. Risk factors for VTE that were more frequent among patients in the preventable group compared with the nonpreventable group included immobility for the 3 days before VTE diagnosis (29.5% vs 9.5%, respectively), concomitant diagnosis of active cancer (22.7% vs 14.3%, respectively), and obesity (18.2% vs 9.5%, respectively). A number of surgical indications for thromboprophylaxis were more prevalent in the preventable group compared with the nonpreventable group: prior nonorthopedic general surgery (52.3% vs 19%, respectively), which included moderate-risk, high-risk, and very-high-risk general surgery; and prior neurosurgery. Conversely, prior orthopedic surgery was more prevalent in the nonpreventable group: total hip arthroplasty (33.3% vs 2.3%, for the nonpreventable and preventable groups respectively), surgery for hip fracture repair (19% vs 9.1%, respectively), and total knee arthroplasty (14.3% vs 4.5%, respectively). With regard to medical indications for thromboprophylaxis, admission to the hospital for pneumonia or for stroke with lower limb paralysis preceded preventable VTE exclusively (six cases and three cases, respectively); although there were only a few cases of VTE occurring in the context of congestive heart failure or myocardial infarction, most were nonpreventable. The clinical context and the prophylaxis regimen used for each case of nonpreventable VTE are shown in Table 4. Fourteen of the 21 cases of nonpreventable VTE occurred in the setting of orthopedic surgery (66.7%). Three cases occurred after moderate-risk surgery (all of which were coronary bypass graft surgery), three cases occurred in the context of an admission to a hospital for myocardial infarction or congestive heart failure, and one case occurred after very-high-risk general surgery. The majority of VTE in this group was lone distal DVT (61.9%). All patients in the nonpreventable group who had undergone orthopedic surgery and the one patient who had undergone very-high-risk general surgery received appropriate thromboprophylaxis with lowmolecular-weight heparin, enoxaparin (Lovenox; Aventis Pharma; Montréal, Québec) at a dose of 30 mg subcutaneously bid for 7 days or until the diagnosis of VTE was established. The patients who had undergone coronary artery bypass graft surgery were receiving appropriate prophylaxis with lowdose unfractionated heparin at a dose of 5,000 U subcutaneously bid daily at the time of VTE diagnosis. All three medical patients with nonpreventable VTE had been receiving full anticoagulation before CHEST / 120 / 6/ DECEMBER,

5 Table 3 Comparison of Preventable and Nonpreventable Cases* Variables Preventable (n 44) Nonpreventable (n 21) Patient characteristics Mean age, yr Female patients 23 (52.3) 13 (61.9) VTE characteristics DVT Proximal 19 (43.2) 6 (30.0) Distal 25 (56.8) 14 (70.0) Unilateral 34 (77.3) 9 (45.0) Bilateral 10 (22.7) 11 (55.0) PE 4 (9.1) 2 (9.5) VTE risk factors Immobility 13 (29.5) 2 (9.5) Cancer 10 (22.7) 3 (14.3) Obesity 8 (18.2) 2 (9.5) Leg fracture 7 (15.9) 4 (19.0) Hormone replacement therapy use 3 (6.8) 1 (4.8) Varicosities 3 (6.8) 1 (4.8) Prior DVT 1 (2.3) 1 (4.8) Surgical indications for VTE prophylaxis Nonorthopedic general surgery 23 (52.3) 4 (19.0) Moderate risk 8 (18.2) 3 (14.3) High risk 11 (25.0) 0 Very high risk 4 (9.1) 1 (4.8) Neurosurgery 4 (9.1) 0 Total hip replacement 1 (2.3) 7 (33.3) Hip fracture 4 (9.1) 4 (19.0) Total knee arthroplasty 2 (4.5) 3 (14.3) Medical indications for VTE prophylaxis Pneumonia 6 (13.6) 0 Stroke with paralysis 3 (6.8) 0 Heart failure 1 (2.3) 2 (9.5) Myocardial infarction 0 1 (4.8) *Data are presented as No. (%) unless otherwise indicated. their VTE diagnosis either with warfarin or unfractionated heparin to maintain an international normalized ratio or partial thromboplastin time, respectively, in the therapeutic range. VTE was suspected in this group of patients after the onset of typical symptoms, and the diagnosis was confirmed by objective testing. Doppler ultrasonography was used to confirm the diagnosis in all but one case of nonpreventable DVT, and compression ultrasonography was used for the remaining case. The one case of PE in this group was confirmed by high-probability ventilation/perfusion lung scan. Discussion VTE occurs despite adequate thromboprophylaxis at a rate that varies according to the indication for therapy. 7 Some authors 2,8 have suggested that the incidence of these breakthrough cases may represent a limitation in efficacy of the recommended guidelines. In our series, if VTE prevention guidelines had been implemented as recommended, then all VTE for which thromboprophylaxis had been indicated should have fallen into the nonpreventable group. We found that only one third of cases (21 of 65 cases) of VTE with a clear indication for prophylaxis were nonpreventable, whereas two thirds of such VTE cases (44 of 65 cases) occurred in the context of inadequate prophylaxis (ie, were potentially preventable). Of all cases of VTE in our series, one sixth of cases (44 of 253 cases) were potentially preventable. These results indicate that there is a need for improvement in thromboprophylaxis implementation. Our study has attempted to uncover some of the obstacles that may be hindering effective thromboprophylaxis usage. We found that inadequacy of prophylaxis was most often caused by omission of prophylaxis. The reasons for this may include physicians being unaware of the ACCP guideline recommendations, or, in certain clinical situations, physicians forgetting to consider thromboprophylaxis or being deterred by the perceived risk of bleeding Clinical Investigations

6 Table 4 Characteristics of Nonpreventable VTE* Cases Context of VTE VTE Risk Factors Prophylaxis Type VTE Type 1 Total hip replacement None Enoxaparin, 30 mg sc bid for 4 d (until DVT diagnosed) Proximal 2 Total hip replacement None Enoxaparin, 30 mg sc bid for 7 d Distal 3 Total hip replacement None Enoxaparin, 30 mg sc bid for 7 d Distal 4 Total hip replacement None Enoxaparin, 30 mg sc bid for 7 d PE, distal 5 Total hip replacement None Enoxaparin, 30 mg sc bid for 4 d (until DVT diagnosed) Distal 6 Total hip replacement Known malignancy Enoxaparin, 30 mg sc bid for 6 d (until DVT diagnosed) Distal 7 Total hip replacement Immobility 3 d Enoxaparin, 30 mg sc bid for 7 d Proximal 8 Total knee None Enoxaparin, 30 mg sc bid for 7 d Proximal arthroplasty 9 Total knee Hormone replacement therapy Enoxaparin, 30 mg sc bid for 5 d (until DVT diagnosed) Distal arthroplasty 10 Total knee Obesity Enoxaparin, 30 mg sc bid for 6 d (until DVT diagnosed) Distal arthroplasty 11 Hip fracture surgery None Enoxaparin, 30 mg sc bid for 7 d Proximal 12 Hip fracture surgery None Enoxaparin, 30 mg sc bid for 7 d Distal 13 Hip fracture surgery Varicose veins Enoxaparin, 30 mg sc bid for 2 d (until DVT diagnosed) Distal 14 Hip fracture surgery Immobility 3 d Enoxaparin, 30 mg sc bid for 6 d (until DVT diagnosed) Distal 15 Very-high-risk general Known malignancy Enoxaparin, 30 mg sc bid for 6 d (until DVT diagnosed) Distal surgery 16 Moderate-risk general None Unfractionated heparin, 5,000 U sc bid for 21 d Distal surgery 17 Moderate-risk general None Unfractionated heparin, 5,000 U sc bid for 1 d Distal surgery 18 Moderate-risk general None Unfractionated heparin, 5,000 U sc bid for 18 d Distal surgery 19 Hospital admission for myocardial infarction None Long-term warfarin therapy, INR 2.6 Proximal 20 Hospital admission for congestive heart failure 21 Hospital admission for congestive heart failure None Prior DVT, known malignancy, obesity *PTT partial thromboplastin time; see Table 2 for expansion of other abbreviations. Full-dose unfractionated heparin for 2 d (80 U/kg bolus then 18 U/kg/h infusion), PTT s Long-term warfarin therapy, INR 3.6 Proximal PE Missed opportunities for the prevention of VTE occurred most commonly in the settings of general surgery, neurosurgery, pneumonia, and stroke. We hypothesize that the perceived risk of VTE is low for patients undergoing nonorthopedic surgery, despite evidence from pooled data showing that the overall incidence of VTE among general surgical patients is approximately 20%. 4 Similarly, a low perceived risk of VTE may have contributed to missed opportunities for thromboprophylaxis of patients with pneumonia or stroke in our series, despite current evidence that indicates a 15% incidence of VTE among medical patients. 9 For patients undergoing neurosurgical procedures and patients with fresh strokes, opportunities for VTE prevention may have been missed because of a perception that the bleeding risk attributable to these patients precluded the use of thromboprophylaxis. There are sufficient data, however, to refute these concerns There was no obvious gap in VTE prophylaxis for cases occurring in the context of orthopedic surgery, perhaps because the early experience with VTE prophylaxis emerged from this clinical setting and because orthopedic health professionals more routinely implement thromboprophylaxis. Our results show that risk factors more common among patients with preventable VTE included immobility, cancer, and obesity. Physicians may not be taking these clinical characteristics into account when attempting to stratify patients into risk groups for VTE, especially before surgical procedures or during hospital admissions. As a result, adequate thromboprophylaxis may be overlooked. The cases of DVT or PE in the nonpreventable group occurred despite adequate prophylaxis. This group represents thromboprophylaxis failures and deserves special attention. Nonpreventable VTE occurred most commonly in the setting of orthopedic surgery, despite 7 days of proper prophylaxis with low-molecular-weight heparin, administered in rec- CHEST / 120 / 6/ DECEMBER,

7 ommended doses. The majority of nonpreventable thromboses were lone distal DVT. The effectiveness of thromboprophylaxis is known not to be perfect. Goldhaber et al 13 recently studied a cohort of predominantly medical patients and reported that 52% of patients developed VTE despite antecedent prophylaxis (although the adequacy of the prophylaxis used with respect to the indication is not discussed). The association that we found between high-risk surgical procedures and the occurrence of nonpreventable VTE suggests that the current guidelines for DVT/PE prophylaxis may be inadequate, particularly for those situations that may confer a heightened thrombogenic state. However, the fact that most DVTs that occurred in the nonpreventable group were distal thromboses may imply that the current guidelines for thromboprophylaxis, although not perfectly effective, may be sufficient to prevent the more dangerous proximal DVT and PE. One nonpreventable case of VTE occurred after only 1 day of thromboprophylaxis. It is possible that VTE was present before the initiation of prophylaxis in that case and would therefore not represent a treatment failure. Nonetheless, the charted information suggested that this was a de novo nonpreventable VTE. Three cases of nonpreventable VTE occurred despite full therapeutic range anticoagulation with either warfarin or full-dose unfractionated heparin therapy. All three cases occurred in the context of a medical admission to hospital, and all three cases were either proximal DVT or PE. Medical patients may be at particularly high risk for VTE because thromboprophylaxis failure occurred despite full anticoagulation and the thromboses that occurred were more serious. Stratton et al 5 evaluated the extent of adoption of grade A guideline recommendations in high-risk surgical patients, and found that although 89.3% of patients received some type of thromboprophylaxis, only 63.6% received prophylaxis that conformed to grade A recommendations. Other studies 2,6,14 18 have examined physician patterns of usage of antithrombotic therapy in general, and most studies 2,6,14 16 concluded that thromboprophylaxis is underused. Three surveys aimed at UK orthopedic surgeons, published between 1989 and 1995, showed that 14%, 27.5%, and 25% of respondents, respectively, did not administer any form of thromboprophylaxis after total hip replacement. Thromboprophylaxis usage in the orthopedic setting has since increased. A 1994 survey of 5,000 US orthopedic surgeons concluded that 95% of respondents administered some type of thromboprophylactic therapy for patients undergoing total hip or total knee replacement surgery, 17 and a 1996 chart audit demonstrated that 93% of patients undergoing total hip replacement received some type of prophylaxis. 18 However, in other settings, thromboprophylaxis continues to be underused. A large review 6 in 1998 of US Medicare patients undergoing major abdominothoracic surgery found that VTE prophylaxis was implemented in only 38% of patients. Analysis of the International Cooperative Pulmonary Embolism Registry, 2 a registry of 2,454 patients with PE established to identify factors associated with death from PE, concluded that one half of the 708 patients who underwent surgery in the preceding 2 months did not receive perioperative thromboprophylaxis, although the types of surgeries were not specified. Our study has several limitations. First, because of the historical cohort design, we were not able to directly assess physician practice patterns in the usage of thromboprophylaxis because data on the total number of patients who received prophylaxis were unavailable. Therefore, we do not have the denominator required to determine how consistently clinicians implement the ACCP guidelines and how many VTEs were averted (and therefore excluded from our series) because of proper usage of the guidelines. However, we were able to assess the rate of preventable cases among patients with VTE, and to identify clinical characteristics that were most often associated with preventable VTE. Second, we may have imprecisely estimated the degree of risk of VTE in certain situations. In determining the level of risk (low to very high) for any patient who underwent a general surgical procedure, we adhered strictly to the definitions offered in the ACCP guidelines, which list clinical risk factors for VTE and attribute the same degree of importance to each one. In doing so, the overall risk to a particular patient may be overestimated or underestimated because certain risk factors contribute to a greater extent than others. 19 Third, in assessing whether prophylaxis was adequate or inadequate, we adhered strictly to the definitions outlined in the guidelines. As a result, inadequacy of prophylaxis may have been overestimated because significant clinical differences may not exist for certain aspects of adequate prophylaxis, for example, twice vs three times daily dosing of unfractionated heparin in certain situations. Fourth, clinically manifest VTE was used as our end point and, as such, the rate of VTE in our study population may have been underestimated because it is known that a proportion of DVT and PE are clinically silent. 7 However, because the ACCP guidelines were developed to reduce the incidence of clinically important thromboembolic disease, we believe that clinically manifest VTE was the appropriate endpoint to use. Finally, we may have underestimated the number of preventable VTE cases because none of the VTE cases that occurred in a setting in which 1970 Clinical Investigations

8 thromboprophylaxis was contraindicated were considered preventable. All contraindications to anticoagulants in our series were attributable to an increased bleeding risk; however, clinical situations in which the bleeding risk becomes prohibitive and thromboprophylaxis is absolutely contraindicated have yet to be clearly defined. As such, omitting thromboprophylaxis in certain situations may not have been justifiable given that the benefit of VTE prevention often outweighs the risk of bleeding By examining preventable cases of VTE, our study identifies important opportunities for thromboprophylaxis that were missed, including the perioperative period for nonorthopedic surgery and neurosurgery, and the clinical setting of pneumonia, and stroke. Our study also represents a useful mechanism for improving clinical practice with institution-specific data, a method that has been shown to be an effective means of implementing change. 20,21 Further research should be directed toward determining factors contributing to inadequate thromboprophylaxis and toward improving the process of guideline implementation by practicing physicians, so that appropriate patients can benefit more widely from the available evidence on thromboprophylaxis efficacy. Finally, although not the primary aim of our study, we have shown, as have others, that even in patients receiving recommended thromboprophylaxis regimens, breakthrough VTE does occur. This highlights the fact that there is still room to improve existing thromboprophylaxis regimens. References 1 Goldhaber SZ. Pulmonary embolism. N Engl J Med 1998; 339: Goldhaber SZ, Visani L, De Rosa M. Acute pulmonary embolism: clinical outcomes in the International Cooperative Pulmonary Embolism Registry (ICOPER). Lancet 1999; 353: Oster G, Truden RL, Colditz GA. Cost-effective analysis of prophylaxis against deep-vein thrombosis in major orthopedic surgery. JAMA 1987; 257: Geerts WH, Heit JA, Clagett PG, et al. Prevention of venous thromboembolism. Chest 2001; 119:132S 175S 5 Stratton MA, Anderson FA, Bussey HI. Prevention of venous thromboembolism: adherence to the 1995 American College of Chest Physician consensus guidelines for surgical patients. Arch Intern Med 2000; 160: Bratzler DW, Raskob GE, Murray CK, et al. Underuse of venous thromboembolism prophylaxis for general surgery patients. Arch Intern Med 1998; 158: Clagett PG, Anderson FA Jr, Heit J, et al. Prevention of venous thromboembolism. Chest 1995; 106:313S 334S 8 Hirsch DR, Ingenito EP, Goldhaber SZ. Prevalence of deep vein thrombosis among patients in medical intensive care. JAMA 1995; 274: Samama MM, Cohen AT, Darmon J-Y, et al. A comparison of enoxaparin with placebo for the prevention of venous thromboembolism in acutely ill medical patients. N Engl J Med 1999; 341: Wen DY, Hall WA. Complications of subcutaneous low-dose heparin therapy in neurosurgical patients. Surg Neurol 1998; 50: Macdonald RL, Amidei C, Lin G, et al. Safety of perioperative subcutaneous heparin for prophylaxis of venous thromboembolism in patients undergoing craniotomy. Neurosurgery 1999; 45: McCarthy ST, Turner J. Low-dose subcutaneous heparin in the prevention of deep-vein thrombosis and pulmonary emboli following acute stroke. Age Ageing 1986; 15: Goldhaber SZ, Dunn K, MacDougall RC. New onset of venous thromboembolism among hospitalized patients at Brigham and Women s Hospital is caused more often by prophylaxis failure than by withholding treatment. Chest 2000; 118: Unwin AJ, Jones JR, Harries WJ. Current UK opinion on thromboprophylaxis in orthopaedic surgery: its use in routine total hip and knee arthroplasty. Ann R Coll Surg Engl 1995; 77: Brenkel IJ, Clancy MJ. Total hip replacement and antithrombotic prophylaxis. Br J Hosp Med 1989; 42: Owen TD, Coorsh J. The use of thromboprophylaxis in total hip replacement: are the attitudes of orthopaedic surgeons changing? J R Soc Med 1992; 85: Paiement GD. Practice patterns in prophylaxis of deep vein thrombosis among US orthopedic surgeons. Orthopedics 1994; 17:11S 13S 18 Anderson FA, Audet A-M. Physician practices in the prevention of deep vein thrombosis: the MassPRO DVT Study. Orthopedics 1996; 19: Heit JA, Silverstein MD, Mohr DN, et al. Risk factors for deep vein thrombosis and pulmonary embolism. Arch Intern Med 2000; 160: Anderson FA Jr, Wheeler HB, Goldberg RJ, et al. Changing clinical practice: prospective study of the impact of continuing medical education and quality assurance programs on the use of prophylaxis for venous thromboembolism. Arch Intern Med 1994; 154: Davis DA, Taylor-Vaisey A. Translating guidelines into practice: a systematic review of theoretic concepts, practical experience and research evidence in the adoption of clinical practice guidelines. Can Med Assoc J 1997; 157: CHEST / 120 / 6/ DECEMBER,