Answers are grounded in this paper and cite their sources.
You are the chief medical officer of a freestanding rehabilitation hospital. As a part of this hospital, there is a large inpatient traumatic brain injury unit. This unit is staffed by 2 physicians, and patients are admitted to their respective services based on census. Therefore, it appears that the 2 physicians essentially have the same types of patients. However, you note that there is a large discrepancy between the 2 physicians regarding the number of lower extremity venous Doppler ultrasonography tests that are ordered. Further investigation reveals that one of the physicians screens every patient admitted for a possible lower extremity deep vein thrombosis, whereas the other physician does not but relies on a combination of history, examination findings, comorbidities, and functional status to determine relative risk. You are not sure which approach is more appropriate but believe that there should be some standardization of clinical practice regarding this aspect of patient care. Therefore, you ask the 2 physicians to present information that supports their respective practices regarding the screening for deep vein thrombosis to help with the determination of how to proceed in the future. Their responses are now presented. Venous thromboembolism (VTE), including deep vein thrombosis (DVT) and pulmonary embolus (PE), is a life-threatening complication for patients who experienced traumatic brain injury (TBI) [1-3]. Due to concerns over hemorrhagic complications, no consensus exists regarding VTE anticoagulant prophylaxis and treatment in the presence of intracranial hemorrhage among patients with TBI [4, 5]. Physicians lack evidence-based guidance to prevent and treat DVT and its sequelae in survivors of TBI. Several studies have looked at the incidence of DVT in patients with TBI on admission to rehabilitation. Estimates of the prevalence on admission range from 5.5%-20% and as high as 54% in the acute care setting [1-9]. In 1986, participants at a National Institutes of Health consensus conference regarding VTE concluded that preventive measures need to be tailored according to the patient's disease and degree of risk [10]. More recently, the American College of Chest Physicians recommended that the choice of diagnostic test should be based on pretest probability rather than performing the same diagnostic test for all patients [11]. Three categories of tests are typically used to determine the probability of DVT: (1) clinical probability assessment based on patient history and clinical findings, (2) D-dimer assays, and (3) imaging studies, most commonly venous duplex ultrasonography (VDU) and less frequently venography, computed tomography, or magnetic resonance imaging. Objective testing for DVT is crucial because clinical assessment alone is unreliable and the consequences of misdiagnosis are serious [12]. Contrast venography is the benchmark for the diagnosis of DVT. Due to the risks associated with venography, it is unsuitable for routine use in patients who present with suspected DVT. Clinical prediction or pretest probability estimates (structured and based on specific criteria or unstructured and empirical) are able to stratify patients into groups according to their probability of DVT. Several structured scoring systems have been developed; the most well studied system is the Wells score [1]. The Wells score incorporates signs, symptoms, and risk factors for DVT to categorize patients as having a low, moderate, or high probability of DVT. The Wells score has limitations. Interobserver reliability has not been widely evaluated. Greenfield et al [13] developed a risk assessment profile for trauma patients. Gearhart et al [14] showed that patients with a risk assessment profile score of 5 or more were 3 times more likely to experience the development of VTE than patients with a risk assessment profile score of less than 5. D-dimer, a degradation product of cross-linked fibrin, is typically elevated in patients with acute DVT. However, because D-dimer levels may also be increased in a variety of nonthrombotic disorders (eg, malignancy, disseminated intravascular coagulation, increasing age, infection, pregnancy, after surgery or trauma, inflammatory conditions, atrial fibrillation, and stroke), D-dimer is a sensitive but nonspecific marker for VTE. Consequently, although a positive result is not useful in confirming the diagnosis of DVT, a negative result can aid in the exclusion of this diagnosis. In hospitalized and other acutely ill patients commonly affected by the conditions listed above, D-dimer testing has less usefulness because of the high frequency of false-positive results. A wide variety of D-dimer assays are available. In patients with clinically suspected DVT, D-dimer has a 91% sensitivity and a 55% specificity for DVT, although performance varies substantially among assays and populations. D-dimer specificity is dependent on pretest clinical probability, being higher in patients with a low clinical probability of DVT [11]. VDU is the most widely used imaging study for the diagnosis of DVT [11]. Proximal VDU assesses compressibility of the femoral and popliteal veins but does not look at the distal vasculature. Whether whole-leg or only proximal VDU is done is controversial. Although distal DVT seldom leads to PE, DVTs may propagate proximally. Additional considerations, such as pretest probability assessment (discussed above), D-dimer testing, or a second proximal VDU performed 5-7 days later, can be considered. Whole-leg VDU as a stand-alone test may result in treatment of distal DVT that will not extend and may carry the risk of overtreatment. VDU has 94% sensitivity for proximal DVT, 64% sensitivity for distal DVT, and 94% specificity [11]. Identification of risk factors for development of DVT in trauma patients has received considerable attention. Knudson et al [15] identified 6 risk factors that were associated independently with the risk of VTE in trauma patients, including age older than 40 years, lower-extremity fracture, major head injury, more than 3 ventilator days, venous injury, and major surgical procedure. Geerts et al [2], in their prospective study, found that major head injury was not associated with DVT. They also did not find a significant association between DVT and gender, injury severity score, pelvic fracture, and amount of blood transfused; they reported that the independent predictors of DVT were age, blood transfusion, surgery, fracture of the femur or tibia, and spinal cord injury [2]. Kudsk et al [16] also did not find an increase of DVT incidence with injury severity score, but they found that it increased with increasing age. Different mechanisms have been proposed that lead to the increased risk of DVT formation among patients with TBI, including immobilization, higher levels of circulating tissue factor, activation of the extrinsic pathway, and elevated plasma levels of von Willebrand factor [17]. Ekeh et al [7] studied patients with moderate-to-severe TBI while they were in acute care. Isolated head injuries have a lower incidence. Older age, male gender, higher injury severity score, and the presence of a lower extremity injury were strong predictors for developing DVT [7]. Meythaler et al [6] found that concurrent lower extremity or pelvic fractures, gender, and age were not associated with an increase in DVT development in the TBI population. Yablon et al [1] confirmed that observations of lower extremity long-bone fractures, age, and Glasgow Coma Scale score (GCS) are not associated with greater risk among patients with TBI admitted to rehabilitation. Among patients with TBI, only D-dimer level and longer postinjury duration were associated with DVT [1]. Yamaki et al [18] used the Wells pretest probability score combined with specific D-dimer levels to decrease use of VDU. The combination of a specific D-dimer level with the clinical probability score was most effective in patients with low-risk pretest probability for excluding DVT. Yablon et al [1] combined VDU and the D-dimer assay as an admission screening protocol. In this study, the D-dimer assay demonstrated high sensitivity to DVT and correlated strongly with VDU evidence of DVT. At the highest titers, D-dimer positive predictive values approached the sensitivity of VDU to detect occult DVT [1]. In recognizing the significant financial constraints of our current health care system, significant effort needs to go into analyzing the utility of medical tests. Several studies have examined the cost-effectiveness of DVT screening in various high-risk populations. Schwarcz et al [19] evaluated the routine postoperative VDU after total joint arthroplasty. VDU rarely found DVT after total joint arthroplasty when effective prophylaxis was used. Furthermore, surveillance scanning did not enable a reliable prediction of PE. They concluded that postoperative inpatient surveillance duplex scans for DVT provide very minimal benefit and that a routine screening program is not clinically useful for patients managed with effective DVT prophylaxis [19]. The American College of Chest Physicians Evidence-Based Clinical Practice Guidelines in 2012 recommended against screening VDU in asymptomatic patients after major orthopedic surgery because it did not result in fewer symptomatic postdischarge VTE. However, screening for asymptomatic DVT appeared to cause harm by leading to unnecessary anticoagulation for several months, which resulted in a higher risk of major bleeding in this population [20]. Perone et al [21] evaluated the cost-effectiveness of 4 strategies for the diagnosis of DVT. The strategies used various combinations of assessment of a patient's clinical probability of having DVT, serial lower limb VDU, and measurement of plasma D-dimer levels. The least-expensive and most cost-effective option was to perform D-dimer as the initial test, followed by a single ultrasound if the D-dimer level was abnormal without any substantial increase in mortality. Serial ultrasonography was the least cost effective [21]. Meythaler et al [6] studied patients with acquired brain injury who were admitted to rehabilitation and who had not been treated with prophylactic anticoagulation or intermittent anticoagulation since their TBI. Screening VDU on admission to rehabilitation revealed a proximal DVT prevalence of 7.8%. The cost of screening in this population was estimated to be significantly lower than the cost of saving someone's life [6]. Goodacre et al [22] analyzed the diagnostic accuracy of noninvasive tests for DVT in patients with clinically suspected DVT or patients who are asymptomatic and at high risk. They also evaluated practical diagnostic algorithms for DVT and estimated the diagnostic accuracy, clinical effectiveness, and cost-effectiveness of each. Two algorithms were identified that offered high net benefit and would be feasible in most hospitals. Both involved using a combination of the Wells score, D-dimer, and VDU. Further diagnostic testing for patients with a low Wells score and negative D-dimer was unlikely to represent a cost-effective use of resources [22]. Lennox et al [23] studied consecutive patients who were referred to the vascular laboratory with clinically suspected DVT. They investigated the use of D-dimer and a clinical risk assessment score to determine whether a practical and cost-effective clinical model could be developed for the diagnosis and exclusion of DVT in symptomatic patients (Table 1). A combination of the risk assessment score and D-dimer testing provided an effective means of excluding proximal DVT in symptomatic patients. They concluded that application of a clinical diagnostic model had the potential for saving a large proportion of unnecessary duplex scans with the associated time and costs [23]. Given the potentially fatal consequences of undetected DVT after failed prophylaxis, some clinicians routinely screen patients with TBI who are asymptomatic with VDU. However, VDU is costly and may have limited sensitivity for asymptomatic DVT. There is no consensus regarding appropriate screening, prophylaxis, or treatment during acute rehabilitation. Carlile et al [4] surveyed the medical directors of the 16 Traumatic Brain Injury Model System to determine national patterns of screening, prophylaxis, and treatment of DVT after TBI. This survey revealed significant disparity in practices of screening, prophylaxis, and treatment of VTE in patients with TBI among these centers. In the acute trauma setting, 33% of the reporting Traumatic Brain Injury Model System centers used VDU for routine screening for DVT. Of the acute rehabilitation hospitals, 56% routinely screen for DVT by using VDU within 72 hours of admission. One center reported screening patients with VDU before they arrive on the rehabilitation unit. Thirteen percent of the surveyed rehabilitation hospitals use D-dimer for its negative predictive value in addition to VDU for routine screening for DVT, whereas 31% routinely use leg circumference measurements for DVT surveillance to augment their VDU and clinical examinations to identify DVT [4]. Carlile et al [5] conducted a prospective observational study and evaluated prophylactic anticoagulation during rehabilitation in patients with TBI. There was a pre-rehabilitation incidence of 9.5%, and VDU screening on rehabilitation admission revealed a lower incidence of previously unsuspected DVTs of 5.5%. These researchers concluded that, given the number of DVTs present before rehabilitation, screening and prophylaxis during acute care may be more important [5]. Ideally, patients with TBI should have screening in the acute care setting to avoid using rehabilitation resources and missing therapy. Screening of all patients with TBI on admission to rehabilitation by using VDU uses significant valuable resources unnecessarily. The vast majority of these patients will screen negative on VDU. A stratified screening program based on an individual patient's risk has merit. The literature supports using pretest probability estimates and the negative predictive value of D-dimer testing as screening tools. VDU then can be applied to the high-risk strata of patients with TBI, which will save resources without increasing the patient's risk. The risk of deep vein thrombosis (DVT) in patients after traumatic brain injury (TBI) is not trivial. One prospective study of DVT after TBI found an incidence of 54% whereas found that DVT of patients with moderate-to-severe TBI [2]. There are no widely screening to DVT in patients as they are admitted to an acute inpatient rehabilitation unit. Therefore, there is a wide discrepancy of practices among which range from screening all patients for DVT their admission to the unit to screening only at a high risk of DVT The associated with DVT formation is potentially their in of patients with an DVT will a pulmonary embolus and to of will be fatal The treatment of DVTs in the TBI population is because it the use of which may be or in the of given at a prophylactic are not without risk. One study showed that, in the of patients with TBI computed of the head within one of admission to an acute care setting showed of their all of had received DVT prophylaxis [4]. only is the of to screen to the of the individual but the of screening also is not [5]. A venous duplex ultrasonography (VDU) is the most commonly used screening because it is low and low but venography or other diagnostic are used as There also are among as to whether the veins of the lower are studied or only a The physician in this screens patient as or the of the patient's specific The second physician only screens patients who are high risk or who or of a DVT of high-risk patients may with a history of a to DVT prophylaxis, a history of being in a or a history of DVT in the will to the as an and to the as a Both have although believe the approach is that of the is a practice among to DVT prophylaxis of a hemorrhage is a of time for development of a Although various for may be used during this they are only if they are for hours with a of during any is to that these are with during this of when tests and other may frequently of the patient to various of the hospital, the of the In this is the are on rehabilitation because patients are to be among for a of 3 hours in has been identified as a to of [6]. The this is that a can during the and then be of the a patient the rehabilitation and on functional for is the time since injury, this an for can be that a VDU be before any to prevent of a potential DVT in this There also is the in the literature that TBI is associated with development of DVT independent of the use of DVT prophylaxis, which the of screening on admission including who have been on prophylactic of In it is not for a patient with TBI to arrive on the acute rehabilitation unit with an in the a patient with a high-risk DVT for anticoagulation is and the of high-risk major trauma in which the was are not without the potential for serious in their an may be that patients in these have been are at an high risk of the sequelae of DVT. This population of patients should be as well to the rehabilitation unit. only against the of they not the DVTs are not of potential addition to risks of DVT venous and limb may be that for the of the in the of a patient with a DVT. for is an important in risk and of the Although there is the potential of the of a duplex to an this low and not will in all of the as There may be some to the if there is of a negative VDU on the before then an can be to on an screen in of an of therapy. The in that the of is a which a for DVT formation and which screening on of rehabilitation all the more The of the is not without cause for of and to be in with the an can be to unnecessary testing as to not with time in or initial The cost of unnecessary tests is However, as physicians to the health and of our patients, not to and our medical A pulmonary embolus would be more to a patient an than missing of would have to identify patient and risk factors that the risk of DVT formation after TBI. One found that head injuries have a lower incidence of DVT whereas older age, male gender, a higher score on the injury severity and the presence of a lower extremity injury are strong predictors of developing DVT [2]. older high-risk groups as with the surgical trauma, malignancy, fractures, trauma of the lower and The of the has some the number of tests performed on the unit may save health care although an would need to be done that at the costs associated with the treatment of DVTs not on admission but rather they have or their potential for and one study did not find that the diagnosis of DVT increased rehabilitation of or functional after TBI [11]. Therefore, it is possible that the cost of treatment of an DVT may not be significant alone to this a physician to practice as a or be of the for with to the or the of without are among the most of an acute DVT. is very to in a patient with TBI and varies on the patient's level of and as well as and A patient who is to in or examination or who reliability of or no responses is not to provide an of The that with the approach is the lack of a of routinely used on rehabilitation to stratify a patient's risk for current DVT. This that the relies on or of evidence or other on which to a would be to a of criteria that the range of typically on a it is unlikely that such a it would the the in a in screen the can to this In all of the above, it is to that the screening approach is the and most to that all DVTs are as as possible in this high-risk population. with on the risk of DVT for patients admitted for rehabilitation and the lack of for DVT risk prophylaxis and with use of a is but will not all rehabilitation studies estimated that 5.5%-20% of patients with TBI have a DVT [1]. every patient with TBI on admission to rehabilitation is the of having a lower extremity of the lower as recommended by then the vast majority of tests will be These patients will have valuable rehabilitation time and be to unnecessary and duplex screening for DVT is not the in other with high risk of DVT. This patients with lower extremity orthopedic and hospitalized patients. American College of Chest Physicians that the choice of diagnostic test should be based on pretest probability rather than performing the same diagnostic test in all patients [2]. The of our TBI population a tailored to sure that patients with TBI are the most Screening all patients with of the lower on admission to rehabilitation has not been to lower risk of during rehabilitation. The risk of hemorrhage in an asymptomatic DVT in this population during rehabilitation also needs approach is to tests without the patient as an not that the current health care such any Given the resources for inpatient rehabilitation, less and less screening be considered. The lack of consensus on screening patients in the TBI Model System the need for study of this This lack of of screening would the value and risks of duplex screening in a Given the high of DVT on admission to rehabilitation, a of patients being before acute care may be more appropriate [1]. A prospective study of the setting care and value of various means of screening combinations of screening will be to this present a structured literature that the of the diagnostic testing and the proposed to risk of DVT that screening is but on how to go the and their on screening in the of their They have referred to as the according to has with this is it relies on and use of an to pretest probability of having a DVT, the Wells score [1]. This has not been in the TBI population and may not have the of this admission to the rehabilitation unit is in to the there are a of and that are performed by and the on with and using as an and the one more of to this time a in which may be low, which to a in diagnosis of DVT. you look at the Wells score patients will be to be at low risk risk or high risk In the in this it is unlikely that a patient is admitted to an acute rehabilitation unit that in TBI without having been 3 days or had major surgery within 4 which is an in the a score value of would the patient at a VDU be would save valuable time to the screening and for a VDU to be because the would one be done also that a result of a D-dimer assay be into the determination of whether or not to perform VDU. Given the low specificity of this marker in trauma the number of with the high number of on the Wells this approach will in most to the to VDU In these initial screening tests may save the time and cost associated with VDU, in the most likely is that a patient will now have had a D-dimer a Wells score and followed by a VDU. these costs is the that VDU by to one to these other screens the risks for and other sequelae of DVT that have Although and well the by and has not believe that all patients should be given VDU to the brain injury unit.
No takes yet. Share an insight, caveat, or question.
Greenwald et al. (2013) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: