Abstract
Thromboprophylaxis remains often underused in hospitalized patients. In 2001, a cohort study done at our institution, a tertiary care center in Montreal, Canada, showed that 67.7% of VTE cases necessitating thromboprophylaxis were potentially preventable with adequate use of American College of Chest Physicians (ACCP) guidelines. Following implementation of an institution-wide policy in 2005, we assessed the changes in the rate of potentially preventable VTE. We conducted a retrospective cohort study including all hospitalized patients with objectively diagnosed VTE in 2010 at our institution. Each case was classified as preventable (thromboprophylaxis indicated but inadequately administered), non-preventable (thromboprophylaxis indicated and correctly administered), spontaneous (thromboprophylaxis not indicated), and ineligible (contraindication to thromboprophylaxis). The results were compared to those obtained in 2001. Of the 230 cases of VTE, 55 cases were classified as potentially preventable (23.9%), 85 were non-preventable (37.0%), 74 were spontaneous (32.2%) and 16 (7.0%) were ineligible. Of the 140 cases requiring thromboprophylaxis, 39.3% were potentially preventable. The potentially preventable cases were mostly due to omission of thromboprophylaxis (50.9%), occurred during general medical admissions (74.5%), and the most common VTE risk factor was cancer (47.2%). In conclusion, we demonstrate a lower frequency of potentially preventable cases in 2010 compared to 2001 (39.3% vs 67.7%, respectively), partially due to physician education and adoption of an institution-wide policy. However, patients with medical indications for thromboprophylaxis, particularly those with cancer, are more prone to having preventable VTE, indicating an area for potential improvement.
Keywords
Introduction
Venous thromboembolism (VTE) is a significant cause of mortality and morbidity in hospitalized patients and is costly to the health-care system. In the United States, the incidence of VTE is estimated at 100 persons per 100,000 each year, with death occurring in about 6% of patients with deep vein thrombosis (DVT) and 12% of patients with pulmonary embolism (PE) within the first month of diagnosis. 1 More recently, a Canadian population-based study reported the incidence rate of VTE, DVT and PE in the Canadian province of Quebec to be 1.22, 0.78, and 0.45, respectively, per 1000 person-years. 2 Treatment of VTE is also expensive: each DVT and PE costs an estimated $10,000 and $16,000, respectively. 3
Prophylactic antithrombotic therapy for the prevention of VTE has been shown to be the most effective method to reduce the health and economic burden of this disease. 4 The American College of Chest Physicians (ACCP) has been instrumental in developing specific guidelines for the use of prophylactic antithrombotic therapy. 5 First published in 1986 and revised every few years thereafter, these guidelines serve to risk-stratify patients with regards to the development of VTE, and to identify the most appropriate prevention strategy 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 of the strength of the supporting evidence. These evidence-based recommendations are generally accepted as the standard of care for VTE prevention.
However, several studies done following the publication of these guidelines have consistently shown underuse of thromboprophylaxis in hospitalized patients.6,7,8,9,10 In 2001, our institution carried out a study of physician practices following the publication of the 1995 ACCP guidelines and showed that 17.4% of all cases of VTE in hospitalized patients, representing two-thirds of cases for which thromboprophylaxis was indicated, were potentially preventable. 6 Furthermore, we identified certain patient groups for which the application of VTE prevention guidelines was particularly inadequate or inconsistent. These included patients admitted for non-orthopedic surgery, pneumonia and stroke.
In view of these disconcerting results, in 2005 we instated an institution-wide thromboprophylaxis policy that was approved by our Medical Executive Committee and distributed to all physicians. We also developed and distributed pocket cards for nurses, residents and staff physicians and organized an annual dedicated VTE Awareness Month with several grand rounds presentations and educational activities targeting all health-care professionals. Also, we assisted in developing pre-printed admission orders for surgical patients which included tick boxes for various thromboprophylaxis options. Five years following its implementation, we assessed the impact of the thromboprophylaxis policy in changing the proportion of VTE cases that were potentially preventable (i.e. cases of DVT or PE which would theoretically not have occurred had the ACCP guidelines been applied as recommended).
Materials and methods
Study design
We conducted a retrospective chart review study of all patients with objectively diagnosed DVT or PE who were admitted to the Jewish General Hospital, a 637-bed university-affiliated hospital located in Montreal, between January 2010 and December 2010. The cohort included patients who were admitted with a primary diagnosis of VTE, as well as patients who developed VTE during hospitalization. Each case of VTE was assessed for indications for thromboprophylaxis and, if present, the adequacy of the thromboprophylaxis regimen prescribed. All charts were reviewed by two independent reviewers who received training by the principal investigator on the use of the same standardized case-report form as that used in the 2001 study to obtain information on the method of VTE diagnosis, patient characteristics, risk factors for VTE, indications for thromboprophylaxis and the regimen received. The information gathered allowed us to classify the patients into different risk groups as defined by the 2008 ACCP guidelines. 5 Based on these data, each case of VTE was categorized as preventable, non-preventable, spontaneous, or ineligible for thromboprophylaxis, as detailed below. The study received expedited approval by the hospital’s ethics committee for an anonymized chart review.
VTE diagnosis
DVT was diagnosed using duplex ultrasonography, venography or autopsy. DVT was classified as distal if it only involved veins distal to the popliteal vein. PE was diagnosed using contrast-enhanced computed tomography (CT), a ventilation/perfusion scan, pulmonary angiography or autopsy.
Patient characteristics and risk factors for VTE and bleeding
Patient age, sex, and weight were recorded. Risk factors as identified in the ACCP guidelines and occurring within 3 months of hospitalization were extracted, including prior VTE, age >40 years, history of cancer or active treatment thereof, obesity, hip, pelvis or leg fracture, pregnancy, and history of immobility or paralysis for at least 3 days prior to VTE diagnosis.
Contraindications to prophylactic antithrombotic therapy and potential risk factors for bleeding that occurred within 3 months of hospitalization were also recorded. These included bleeding disorders, thrombocytopenia, hemorrhagic stroke, hepatic failure, bacterial endocarditis, and active gastrointestinal bleed.
Indications for thromboprophylaxis
According to the ACCP guidelines, there are clinical situations in surgical and medical patients for which thromboprophylaxis is indicated. 5 Surgical indications for thromboprophylaxis include elective hip and knee repair, hip fracture surgery, spine surgery, general surgery, peripheral vascular surgery, neurosurgery, and trauma. Medical indications for thromboprophylaxis include admission for congestive heart failure, respiratory distress, any active malignancy, burns and admission to a critical care unit. Patients were stratified according to ACCP guidelines as having a low, moderate, or high risk for VTE depending on the reason for admission, the surgical procedure, and the number of risk factors for VTE. Although not included in the 2008 Chest guidelines, we considered rivaroxaban an adequate thromboprophylactic agent after orthopedic surgery in light of new supporting evidence that emerged between 2008 and the time of this study.
Adequacy of thromboprophylaxis
The ACCP guidelines recommend stratifying patients according to their VTE risk and for each risk group propose thromboprophylaxis regimens, including drug, dose, frequency and duration. Hence, for a given patient, thromboprophylaxis was considered to be adequate if all the following criteria were met: (1) the thromboprophylactic agent used was the same type, and at least the same dose and frequency as that proposed by the ACCP guidelines; (2) it was given continuously for at least 7 days, until hospital discharge or the patient was ambulatory; and (3) it was initiated within 24 hours of a clinical event associated with risk of thrombosis (e.g. surgery, myocardial infarction) or at hospital presentation.
Classification of VTE
We used the same definitions to classify VTE as those used in our 2001 publication. 6 Based on the above data, each case of VTE was categorized as potentially preventable, non-preventable, spontaneous, or ineligible for our analysis:
Statistical analysis
We describe the proportions of cases of VTE that were classified as preventable, non-preventable, spontaneous and ineligible based on ACCP guidelines. The rate of preventable VTE and the specific reasons for inadequacy of thromboprophylaxis were examined, and compared to the results obtained in our 2001 publication. 6
Results
During the 1-year study period, 230 cases of VTE were objectively diagnosed in 219 patients (Table 1). The mean (±SD) age was 70.2 ± 15.2 years, and 55.7% of patients were female. There were 51 cases of PE, 49 cases of concomitant PE and DVT, and 130 cases of isolated DVT, most of which were unilateral (74.9%). There was a near equal distribution of proximal (53.1%) versus distal (44.7%) DVTs. The high prevalence of distal DVT suggests a lower threshold for ordering calf vein imaging (i.e. even in patients with minimal symptoms). However, it should be noted that at our institution, all Doppler studies undertaken to rule out DVT image both the proximal and calf veins. Of the 230 cases of VTE, 193 (83.9%) cases were admitted from the emergency department, 13 (5.6%) were transferred from other institutions, and 24 (10.4%) were elective admissions to a surgical ward.
Characteristics of patients, venous thromboembolism (VTE), and category of VTE.
Data shown are n (%) unless otherwise indicated.
VTE, venous thromboembolism; SD, standard deviation; DVT, deep vein thrombosis; PE, pulmonary embolism.
VTE category – preventable: cases of VTE for which thromboprophylaxis was inadequately given; non-preventable: cases of VTE for which thromboprophylaxis was given adequately; spontaneous: cases of VTE for which there were no indications for thromboprophylaxis; ineligible: cases of VTE for which insufficient data was available or VTE was diagnosed at another institution.
Of the DVTs in the 2010 cohort, 136 (76.0%) were symptomatic and 43 (24.0%) were asymptomatic.
Of the 230 cases of VTE, 55 (23.9%) were classified as potentially preventable, 85 (37.0%) as non-preventable, 74 (32.2%) as spontaneous, and 16 (7.0%) as ineligible. Therefore, of the 140 (i.e. 55+85) cases for which thromboprophylaxis was indicated, 39.3% were potentially preventable. The reasons for the inadequacy of prophylaxis in potentially preventable VTE cases are shown in Figure 1. For 28 (50.9%) cases, thromboprophylaxis was omitted entirely. In 22 (40.0%) cases, there was an inappropriate delay in initiation of thromboprophylaxis, and all were admitted through the emergency department. In four (7.3%) cases, thromboprophylaxis was administered for an inadequate duration. In one (1.8%) case, the dose of anticoagulant was less than that recommended in the guidelines. This patient had received warfarin following a congestive heart failure exacerbation, but the international normalized ratio (INR) was subtherapeutic (1.3) at the time of VTE diagnosis. There were no instances in which the type of prophylaxis administered was other than that stipulated by the ACCP guidelines.
Reasons for inadequacy of prophylaxis in preventable VTE (n=55).
Clinical characteristics of patients with preventable and non-preventable VTE cases were compared with the goal of identifying specific patient groups at increased risk for inadequate thromboprophylaxis or VTE despite proper prophylaxis (Table 2). The groups were of similar age and sex distribution. DVT was more often the form of VTE diagnosed in potentially preventable than in non-preventable cases (76.3% vs 57.6%, respectively), and among those with DVT, proximal DVT occurred more often in the preventable group than in the non-preventable group (53.2% vs 43.3%, respectively). Several VTE risk factors were more prevalent in the preventable group compared with the non-preventable group. These included cancer (47.3% vs 36.5%, respectively) and immobility (33.3% vs 16.5%, respectively). Surgical indications for VTE prophylaxis were notably more common in the non-preventable compared with the preventable group, including non-orthopedic surgery (29.4% vs 9.1%, respectively) and total knee replacement (11.8% vs 1.8%, respectively). Conversely, medical indications for thromboprophylaxis were more prevalent in the preventable group than in the non-preventable group (74.5% vs 43.5%, respectively), which included 16 (29.1%) patients with cancer-related complications and 10 (18.2%) patients with non-respiratory infections.
Comparison of non-preventable and preventable cases, present study.
Data shown are n (%) unless otherwise indicated.
VTE, venous thromboembolism; SD, standard deviation; DVT, deep vein thrombosis; PE, pulmonary embolism.
Medical indications for VTE prophylaxis include primary reason for admission, and do not reflect secondary reasons for hospitalization. Surgical indications for VTE include surgery for tumor resection. These patients were not included in the category ‘medical indications for VTE’. The denominator of the DVT characteristics is the total number of DVT.
The type and dosing of thromboprophylaxis used in non-preventable cases of VTE were examined (Table 3). Patients who had undergone orthopedic surgery had either received enoxaparin 30 mg sc bid or rivaroxaban 10 mg po qd for 10–30 days or until the diagnosis of VTE. Most patients who had medical indications for thromboprophylaxis received enoxaparin 30 mg sc qd or 40 mg sc qd or unfractionated heparin 5000 units sc bid for the duration of their hospital admission or until VTE diagnosis. In two cases of non-preventable VTE, warfarin was used for thromboprophylaxis, and the INRs were 2.9 and 2.7 at the time of VTE diagnosis. Multiple agents at prophylactic doses were used in two cases. Therapeutic doses of enoxaparin (1.5 mg/kg sc qd) were given in 13 cases, which included patients admitted with myocardial infarction or previously diagnosed VTE who developed a new VTE despite therapeutic anticoagulation.
Thromboprophylaxis regimens used in cases of non-preventable venous thromboembolism.
Data shown are n (%).
INR, international normalized ratio.
Therapeutic low molecular weight heparin indicated for acute coronary syndrome, atrial fibrillation or previous VTE.
Therapeutic warfarin indicated for atrial fibrillation or chronic thromboembolic disease.
Comparison of results of this study with our 2001 publication
We compared the results from the 2010 cohort to those obtained in 2001 5 (Table 1). Although there is an increase in the absolute number of potentially preventable cases from the 2001 to 2010 cohort (44 to 55, respectively), there was a marked improvement in the ratio of preventable to non-preventable cases of VTE. In our 2001 publication, 67.7% of cases (i.e. two in three cases) for which thromboprophylaxis was indicated were deemed potentially preventable, while in 2010, 39.3% of such cases (i.e. 1 in 2.5 cases) were considered potentially preventable. Omission of thromboprophylaxis was the primary reason for inadequacy of VTE prevention in both 2001 and 2010 (47.7% vs 50.9%, respectively). Also, in 2010, there were no potentially preventable cases in which patients received the wrong thromboprophylactic agent compared to the 2001 publication, in which nine such cases were reported. Unlike the 2001 publication, in which we reported only three (14.3%) cases of potentially preventable VTE that had medical indications for thromboprophylaxis, 41 (74.5%) such cases were noted in the present publication. The most common medical indication for thromboprophylaxis was cancer-related complications (16 of 55 cases of potentially preventable VTE; 29.1%) followed by non-respiratory infections (10 of 55 cases of potentially preventable VTE; 18.1%), while in 2001 we reported pneumonia (13.6%) and stroke (6.8%) as being the most common medical indication for thromboprophylaxis.
Discussion
The underuse of thromboprophylaxis in hospitalized patients is well documented.9,10 In 2001, we reported that preventable VTE represented two-thirds of cases for which thromboprophylaxis was indicated, and one in six cases of all VTE. 6 Partly in response to these findings, our institution developed and implemented a hospital-wide thromboprophylaxis policy in 2005. The policy provides guidance on VTE risk assessment, choice of anticoagulant and contraindications to pharmacologic thromboprophylaxis in various patient groups (e.g. medicine, orthopedic surgery, general surgery).
Five years following the implementation of this policy, we reassessed the preventability of VTE at our institution. We acknowledge that our ability to compare the 2001 and 2010 cohorts was constrained by the substantial changes in the ACCP recommendations from 1995 to 2008. Of note, the latter recommendations have expanded significantly to now include specific patient groups such as subspecialty surgeries (gynecology, urology, vascular, etc.) and cancer patients that were not included in the 1995 guidelines. Hence, the number of patients for whom prophylaxis is indicated has increased significantly, which might account for the increase in the absolute number of potentially preventable and non-preventable cases between the 2001 and 2010 studies.
Nevertheless, when looking specifically at high-risk patients, we noted substantial improvement, in that 39.3% of all cases for which thromboprophylaxis was indicated were deemed to be potentially preventable, compared with 67.7% in 2001. There are several possible reasons behind the improved adherence to the ACCP guidelines. First, the past decade has seen the emergence of several new anticoagulation agents, and the subject of VTE has moved to the forefront of medical discussion, including issuance of a US Surgeon General’s report on VTE. 11 The subsequent sensitization of physicians and allied health professionals to the issue of thromboprophylaxis may play a large part in better guideline adherence. Second, as stated above, the ACCP guidelines have evolved in the past decade, becoming more inclusive while providing more detailed recommendations for specific patient groups. Finally, implementation of an institution-wide policy allowed for a more protocolized approach to thromboprophylaxis. We further improved physician education at our institution by organizing several grand rounds presentations on VTE each year and holding an annual VTE Awareness Month, which includes educational workshops, public lectures and informational kiosks.
Nevertheless, despite the specific changes implemented in our institution, we found that 55 cases of VTE documented in 2010 were potentially preventable. We attempted to elucidate the reasons behind the persistent insufficient use of thromboprophylaxis. As in the 2001 publication, the most common reason for inadequate thromboprophylaxis was complete omission of a thromboprophylactic agent. The second most common reason for inadequate thromboprophylaxis was inappropriate delay in the administration of thromboprophylaxis. In an overburdened health-care system in which prolonged stays in the emergency room may commonly exceed 24 hours, our results indicate that there may be a place for initiation of thromboprophylaxis in the emergency department.
The vast majority of missed opportunities for VTE prevention occurred in the setting of medical admission. The most striking difference from the 2001 publication is the preponderance of cancer-related complications. This may represent the increasing prevalence of cancer as a reason for admission to hospital. Also, the suboptimal application of ACCP guidelines in medical patients highlights the challenges of attaining adequate thromboprophylaxis in this heterogeneous population, in whom VTE risk stratification can be challenging. In a recent Cochrane review done by our research group, interventions such as ‘alerts’ and pre-printed orders were shown to improve adherence to thromboprophylaxis guidelines. 12 Also, the risk factors ‘acute medical illness’ and ‘immobilization’ are imprecisely defined in the current guidelines. Finally, four cases of potentially preventable cases of VTE occurred in a palliative care setting, which highlights a unique subgroup of patients in which thromboprophylaxis may need to be evaluated on a case-by-case basis despite an incidence of VTE as high as 5.5%, 13 significantly higher than previously reported rates of DVT (0.8%) and PE (0.4%) in hospitalized medical patients. 13
Cases of non-preventable VTE are those that occur despite adequate prophylaxis, and hence represent the failure of prophylaxis. These occurred most frequently in the setting of non-orthopedic surgeries, including surgery for malignancy resection, cancer-related medical complications, and elective total knee arthroplasties. These results identify areas in which there is still room to improve the effectiveness of thromboprophylactic agents and approaches. Of note, in the case of cancer surgical patients, none received low-dose unfractionated heparin (LDUH) in a TID dosing, which was suggested for particularly high-risk patients in the 2008 guidelines. 5 This could have contributed to some VTE events.
Overall, the single most frequent risk factor in preventable and non-preventable VTE was active malignancy (47.3% and 36.5% of cases, respectively). This highlights the high prevalence of cancer in hospitalized patients and the challenges in providing adequate thromboprophylaxis to this population. Physicians may continue to have a low perceived risk for VTE in surgical and medical patients with cancer despite evidence to the contrary. 14 Further, the high rate of active malignancy in non-preventable cases of VTE may point to suboptimal efficacy of current thromboprophylaxis regimens in this high-risk population.
The incidence of VTE in cancer patients is estimated to be as high as 1 in 200, although the incidence rate varies widely with type of tumor, extent of disease, and cancer histology. 15 Patients with an active malignancy undergoing a surgical procedure are twice as likely to develop VTE in the perioperative period as those without an active malignancy. 5 Importantly, cancer is also an independent risk factor for thromboprophylaxis failure. In the ARISTOS study, Agnelli et al. examined the incidence rate of VTE post-surgical intervention in cancer patients, and reported 37 cases out of 50 cases (74%) of confirmed VTE occurred in the setting of active thromboprophylaxis. 16 The adequacy of VTE thromboprophylaxis, however, was not specified, although the authors did mention that in-hospital thromboprophylaxis was given in 81.7% of patients, of which low molecular weight heparin prophylaxis was started before surgery in 71.1% of the cases. Similarly, in 2005, Kakkar et al. did a post hoc analysis of a randomized control study comparing the efficacy and safety of certoparin with unfractionated heparin as thromboprophylaxis in post-surgical patients. 17 They reported a high rate of fatal pulmonary embolism in cancer patients as compared to patients with no active malignancy (0.33% vs 0.09%, respectively; relative risk (RR), 3.7, 95% confidence interval (CI), 1.80, 7.77, p=0.0001). Our results seem to support the high rate of thromboprophylaxis failure in high-risk cancer patients.
There are several limitations to our study. First, due to the retrospective design of our study, we could not determine the total number of patients who received thromboprophylaxis during the time period studied. Hence, our study only gives a partial view of physicians’ adherence to thromboprophylaxis guidelines, as it focuses on instances in which inadequate thromboprophylaxis was delivered. However, we were able to highlight clinical settings and patient characteristics for which inadequate thromboprophylaxis was more frequent. Second, in stratifying patients as being at low or high risk of VTE, we gave equal weight to all risk factors defined in the ACCP guidelines. However, in the clinical setting, some risk factors may influence VTE risk more than others. Third, our endpoint was objectively diagnosed VTE, and hence did not include instances in which there was a high suspicion of VTE but no confirmatory test. This might have underestimated the true number of cases of VTE during the time period studied. Finally, due to its retrospective design, we were unable to ascertain if there were instances of delayed or missed doses among patients who received thromboprophylaxis.
In summary, our study provided a unique opportunity to examine real-world changes with regards to thromboprophylaxis delivery. Using institution-specific data, we documented improved rates of potentially preventable VTE from 2001 to 2010, which reflect the past decade’s significant changes in the approach to thromboprophylaxis and may also be ascribed to the implementation of our hospital-wide VTE prevention policy. However, thromboprophylaxis remains inadequate for non-orthopedic surgery and medical admissions, and patients with active malignancy are prone to thromboprophylaxis failure. This highlights areas in which thromboprophylaxis agents and approaches may be improved.
Footnotes
Declaration of conflicting interest
There are no conflicts of interest to declare for Dr Kim A Ma and Ms Eva Cohen. Dr Susan R Kahn has received honoraria from Sanofi and has attended advisory board meetings for Sanofi and Boehringer Ingelheim.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sector.