Safely Ruling Out Deep Venous Thrombosis in Primary Care

Context Primary care physicians need office-based methods to rule out suspected deep venous thrombosis (DVT). Contribution The authors conducted a management trial of a prediction rule that uses clinical findings and a point-of-care d-dimer test to identify patients at very low risk for suspected DVT. They managed 1028 patients from approximately 300 primary care practices according to the rule, which identified nearly half (49%) to be at low enough risk to withhold imaging tests and anticoagulation treatment. In 3 months, 1.4% (95% CI, 0.6% to 2.9%) of low-risk patients had venous thromboembolism. Caution There was no control group. Authors relied on symptoms to detect subsequent venous thromboembolism. Implication Office-based methods can safely rule out DVT. The Editors Each year, more than 140000 persons in the United Kingdom present to their primary care physician with signs and symptoms suggestive of deep venous thrombosis (DVT) of the leg (1, 2). Because DVT is a potentially life-threatening disorder, current practice is to refer all patients for diagnostic testing services. These services are readily available, use noninvasive tests (such as, ultrasonography and d-dimer testing), and provide the referring physician with the assurance that DVT is not missed (3, 4). However, many studies have revealed that 80% to 90% of these referred patients do not have DVT (46). Therefore, it would be ideal to safely exclude DVT at initial presentation in a large proportion of these patients and thereby avoid referral. The recent introduction of rapid point-of-care d-dimer assays that can be included in a specific clinical decision rule makes it possible to do a diagnostic work-up in a primary care setting (79). We recently found that the use of a decision ruleinitially developed and validated in secondary carewas not accurate enough for primary care patients suspected of having DVT because the prevalence of thrombosis was still 2.9% among patients with a low probability (based on the Wells score and a normal quantitative d-dimer) compared with 0.9% reported in the original publication by Wells and colleagues (10, 11). Therefore, we developed and validated such a decision rule specifically for the primary care setting (10, 12) that included clinical items and the d-dimer assay result. A major difference between the rules, taking into account the additional use of d-dimer for low-probability patients in the Wells rule, is the replacement of the subjective phrase alternative diagnosis more likely with the more objective phrase absence of leg trauma. In primary care, the category of low probability based on the new rule had a 0.7% prevalence for thrombosis (13). However, as Reilly and Evans (14) recently outlined, development and validation studies should be followed by a prospective impact or management study demonstrating that the rule could be used by physicians to direct care before the rule is implemented in daily practice. Therefore, we conducted this study in a large series of consecutive patients in a primary care setting to evaluate the safety and efficiency of excluding DVT by using a clinical decision rule and a point-of-care d-dimer assay. In addition, we measured the yield of ultrasonography in the referred patients. Methods Study Overview In this prospective study in primary care, we managed patients suspected of having DVT by using a clinical decision rule that included a point-of-care d-dimer test. We did not refer patients with a low probability of DVT for further testing or administer treatment; we followed them for 3 months to record the incidence of venous thromboembolism (10, 12). Setting and Patients We invited the affiliated general practices of the 3 academic centers (who organized the study) to participate. Approximately 300 general practitioners agreed to participate. From March 2005 to January 2007, consecutive patients who presented with clinically suspected DVT were eligible for the study on the basis of the presence of at least 1 of 3 lower extremity symptoms: swelling, redness, or pain. We excluded patients if they were younger than 18 years, received anticoagulant treatment (that is, vitamin K antagonists or low-molecular-weight heparin) at presentation, or declined to participate. We obtained written informed consent, and the local institutional review boards approved the study. Diagnostic Strategy General practitioners applied a clinical decision rule, provided on a worksheet, to all study patients. This clinical decision rule was developed to safely exclude clinically suspected DVT in primary care patients. It included clinical items and a d-dimer assay result (Table 1) (10, 12). Table 1. Clinical Decision Rule Because we aimed to improve the management of patients suspected of having DVT in a primary care setting, we explicitly selected a rapid point-of-care d-dimer assay (Clearview Simplify d-dimer assay, Inverness Medical, Bedford, United Kingdom) (8, 15, 16). This allowed the general practitioner to use the decision rule outside of office hours and during house calls. We drew a capillary blood sample by using the finger-prick method (15). The test result was considered abnormal if, next to the control band, a second band appeared within 10 minutes (15). Participating physicians and their assistants received a single, brief instruction on how to use the d-dimer assay and the clinical rule. Physicians calculated the score for each patient by using the clinical decision rule (Table 1) (10, 12) and managed patients accordingly. Those with a score of 3 or less did not receive anticoagulant treatment or a referral for ultrasonography, but they were instructed to contact their general practitioner if symptoms became worse. Patients with a score of 4 or greater received a referral for ultrasonography. Deep venous thrombosis was considered present when 1 of the proximal veins of the lower extremities was noncompressible on ultrasonography (4). All patients visited their general practitioner between days 5 and 9 for reevaluation. Three months after entering the study, all patients received a questionnaire addressing signs and symptoms of (recurrent) venous thromboembolism. We contacted patients who did not respond (30%) through their general practitioners. If we had any suspicion of a (recurrent) venous thromboembolic event during the 3-month follow-up, based on the information presented in the questionnaire, we retrieved additional medical information of patients from their general practitioners, including letters from hospital specialists. Outcome Measure We defined the primary outcome as the incidence of symptomatic venous thromboembolism during 3-month follow-up. This included fatal pulmonary embolism, nonfatal pulmonary embolism, and DVT. An independent adjudication committee, unaware of the patient's result of the clinical decision rule, evaluated all suspected venous thromboembolic events and deaths. A diagnosis of pulmonary embolism or DVT was based on a prioridefined and generally accepted criteria (Appendix Table) (17). Deaths were classified as caused by pulmonary embolism when autopsy was done if an objective test result was positive for pulmonary embolism before death or if pulmonary embolism could not be confidently excluded as the cause of death (17). Appendix Table. Adjudication Criteria Used for the AMUSE Study Statistical Analysis On the basis of an expected incidence of venous thromboembolism in 1% of patients (those with a score 3) during 3-month follow-up and the exclusion of a predetermined incidence of 4% or more, we calculated that 488 patients needed to be included in this low-risk group (type I error, 0.05; type II error, 0.2). The primary analysis was about the incidence (with exact 95% CI) of symptomatic venous thromboembolism during 3-month follow-up in the group of patients with a score of 3 or less who were not referred for further testing or treatment. In addition, we calculated the percentage of patients with a score of 3 or less. Furthermore, we calculated the probability of venous thromboembolism on leg ultrasonography at baseline or during follow-up, according to the results of the clinical decision rule without the d-dimer assay result, as well as the d-dimer assay result alone. For this purpose, a cutoff of 3 or less was also used because on the basis of this cutoff, patients with a negative d-dimer result should still receive a referral for ultrasonography. Role of the Funding Source The study was funded by the Netherlands Organization for Scientific Research. The funding source had no role in the design, conduct, or reporting of the study or in the decision to submit the manuscript for publication. Results Patients We assessed 1086 consecutive patients with clinically suspected DVT. We excluded 58 patients (5.3%) because of predefined exclusion criteria (Figure). Table 2 shows characteristics of the 1028 study patients, including the items of the clinical decision rule. The mean age was 58 years, and 37% were men. Suspicion of DVT that led to study inclusion was based most commonly on leg pain (87%) and leg swelling (78%). Figure. Study flow diagram. CDR = clinical decision rule; DVT = deep venous thrombosis; DVT+ = deep venous thrombosis confirmed by ultrasonography; DVT = deep venous thrombosis excluded by ultrasonography; LMWH = low-molecular-weight heparin; PE = pulmonary embolism; VTE = venous thromboembolism. * All patients were referred for ultrasonography on the day of presentation. No clinical events occurred in this group. Clearview Simplify d-dimer assay, Inverness Medical, Bedford, United Kingdom. Incidence of VTE, 1.4% (95% CI, 0.6% to 2.9%). If the 1 patient lost to follow-up had developed VTE, the percentage missed by the procedure would have been 1.6% instead of 1.4%. Table 2. Demographic and Clinical Characteristics of the Study Sample (n= 1028) Results of the Clinical Decision Rule In 500 of 1028 patients (49%), the score was 3 or les