Claudine Jurkovitz, MD, MPH

Hospital‐acquired venous thrombus embolism (VTE) is a pressing patient health and safety issue and has been identified as a causal factor in preventable deaths in the hospital setting.[1, 2] More than 540,000 hospitalizations with VTE occur each year among adults in the United States.[3] The number of adults with VTE is anticipated to increase from 0.95 million in 2006 to 1.82 million in 2050.[4] The Institute of Medicine has defined failure to provide adequate thromboprophylaxis to hospitalized, at‐risk patients as a medical error.[2, 5] The American College of Chest Physicians guidelines state that thromboprophylaxis is highly effective at preventing deep vein thrombosis (DVT) and proximal DVT, highly effective at preventing symptomatic VTE and fatal pulmonary emboli (PE), and that the prevention of DVT also prevents PE.[6] Where anticoagulation is contraindicated, mechanical methods of thromboprophylaxis are recommended as preferable to no thromboprophylaxis, with careful attention directed toward ensuring the proper use of, and optimal adherence with, mechanical prophylaxis.[7, 8] In our institution, concerns about the existence of asymptomatic clots being propagated into PEs by the placement of pneumatic compression boots (PCBs), led to routine performance of duplex Doppler ultrasound with compression (DUSC) before applying PCBs to those patients who were admitted and who were deemed to have a contraindication to anticoagulation prophylaxis. The recently released (April 2012) American College of Radiology Choosing Wisely list of practices specifically recommends forgoing imaging for DVT and PE in the absence of risk factors.[9] The recommendations do not specifically address screening for DVT prior to the initiation of prophylaxis. The goal of this prospective observational study, conducted prior to the Choosing Wisely campaign, was to verify our hypothesis that the prevalence of asymptomatic DVTs was very low, and provide our clinicians with evidence to allay concerns about placement of PCBs without imaging, allowing a practice pattern that would reduce costs without impacting patient safety.

METHODS

RESULTS

A total of 1136 consecutive records were examined; 4 records were excluded from the analysis because they had a diagnosed PE prior to US, and 35 records were excluded because the US was performed beyond 48 hours after admission. The final dataset included 1097 hospital admissions for 1071 patients. Of the 1097 admissions, 759 (69.2%) originated from the emergency department (ED). It is important to note that 70,161 hospital admissions occurred during the same time period, of which 36,363 (51.8%) were admissions that started in the ED. The proportion of patients requiring mechanical DVT prophylaxis is therefore very small (<5%), assuming that a large number of the patients with unplanned admissions would require DVT prophylaxis.

Of the 1071 patients in the final analytical dataset, 544 (50.8%) were male, the mean age was 65.5 years, the mean BMI was 28.7 (median, 27.0) (Table 1), and the majority of the patients were white. US was performed within 24 hours in 712 (66.5%) patients, and 665 (62.1%) had Medicare. An asymptomatic DVT was detected by DUSC in 19 patients (1.8%). None of the clinical and demographic characteristics were statistically different between those with DVT and without (Table 1).

Patients with DVT had at least 1 risk factor; 16 (84.2%) of them had 2 or more risk factors. In addition, the presence of 2 or more risk factors was much more frequent among those with DVT than among those without (84.2% [16/19] vs 58.4% [614/1052], P=0.03).

As shown in Table 2, a history of DVT or PE and ambulatory dysfunction are the only risk factors associated with DVT at admission. In addition, the prevalence of DVT increases as the number of risk factors increases (Table 3). The prevalence is much higher in those who had 4 or more risk factors than among those with fewer than 4 risk factors (12.2% [6/49] vs 1.3% [13/1022], P=0.0001).

Results of the logistic regression, similar to those of the nonadjusted analysis, showed that the only risk factors independently associated with the discovery of a DVT upon DUSC were the presence of ambulatory dysfunction (odds ratio [OR]: 2.99, 95% confidence interval [CI]: 1.13‐7.90) and a history of DVT or PE (OR: 10.51, 95% CI: 3.90‐28.31) (Table 4).

We estimated a savings for Medicare of approximately $266,000 to $280,000 ($261.07 1071 DUSC or $261.07 1022 [after excluding the patients with 4 or more risk factors]) over 13 months had the DUSC not being conducted.

DISCUSSION

This study shows that discovering an asymptomatic DVT is relatively rare (<2%) in patients arriving at the hospital for all causes of admission, even taking into account multiple risk factors that increase the risk for DVTs. The study strongly supports the practice of placing compression devices as soon as possible for those patients who have a contraindication to anticoagulant prophylaxis. Along with reducing the delay to placement while awaiting the test, there is significant cost reduction to the healthcare system by not doing DUSC. There appears to be no need for diagnostic studies prior to the placement of these devices unless the patient has more than 3 risk factors or there is a history of previous DVT or ambulatory dysfunction. This study strongly supports the premise that patients are not arriving with DVTs, but are developing them in the hospital.[1, 2, 10] The 1.8% prevalence of asymptomatic DVT in this study is somewhat lower than that found in other studies. The Prophylaxis for Thromboembolism in Critical Care Trial (PROTECT) tested dalteparin vs unfractionated heparin on 3764 patients in the intensive care unit. Initial screening done to rule out DVT found that 3.5% of patients receiving dalteparin and 3.4% receiving unfractionated heparin had proximal DVTs.[8] Other Investigators used venous compression ultrasound examinations of the lower limbs to determine that 5.5% of patients hospitalized in a medical unit have an asymptomatic DVT of the lower limbs on admission.[5] A limitation of that study is the inclusion of all thrombo emboli, specifically those found in the calf (19 out of 21, or 90%). However, if one eliminates the calf venous thrombi, not considered risk factors for PE, the prevalence of DVT (0.85%) is about half that of our observed 1.8%.

In common with previous studies, a history of previous thromboembolic disease was clearly the most significant of many evaluated risk factors for DVT.[5, 6, 10] Ambulatory dysfunction was also a statistically significant risk factor that was likely under‐reported here because of the inexact documentation in many of the medical records. Interestingly, a history of active malignancy did not prove to be a significant risk factor, contrary to other study reports.[5, 6, 10]

The frequency of asymptomatic DVT appears to increase with the accumulation of risk factors. An asymptomatic DVT existed in 1.3% of the patients with 3 or fewer risk factors, compared with 12.2% of those with 4 or 5 risk factors. It is possible that a higher number of risk factors for DVT would be an indication for obtaining a DUSC prior to the placement of PCBs, although the small number of patients with more than 3 risk factors in our study population may limit the strength of this observation.

CONCLUSION

Our data strongly suggest, in alignment with recent recommendations, that there is no need to perform screening DUSC prior to the placement of prophylactic compression devices among hospital admissions who have contraindications to anticoagulation. Rather, efforts should be focused on implementing systems to ensure rapid placement of these compression devices at the time of admission for those patients who cannot receive anticoagulation prophylaxis. Evaluation for DVT may be of value if there is a history of previous DVT or PE, ambulatory dysfunction, or more than 3 risk factors, as the information may change the therapeutic approach. Current guidelines recommend the measurement of D‐dimers as a screening tool for DVT.[11]

Hospital‐acquired venous thrombus embolism (VTE) is a pressing patient health and safety issue and has been identified as a causal factor in preventable deaths in the hospital setting.[1, 2] More than 540,000 hospitalizations with VTE occur each year among adults in the United States.[3] The number of adults with VTE is anticipated to increase from 0.95 million in 2006 to 1.82 million in 2050.[4] The Institute of Medicine has defined failure to provide adequate thromboprophylaxis to hospitalized, at‐risk patients as a medical error.[2, 5] The American College of Chest Physicians guidelines state that thromboprophylaxis is highly effective at preventing deep vein thrombosis (DVT) and proximal DVT, highly effective at preventing symptomatic VTE and fatal pulmonary emboli (PE), and that the prevention of DVT also prevents PE.[6] Where anticoagulation is contraindicated, mechanical methods of thromboprophylaxis are recommended as preferable to no thromboprophylaxis, with careful attention directed toward ensuring the proper use of, and optimal adherence with, mechanical prophylaxis.[7, 8] In our institution, concerns about the existence of asymptomatic clots being propagated into PEs by the placement of pneumatic compression boots (PCBs), led to routine performance of duplex Doppler ultrasound with compression (DUSC) before applying PCBs to those patients who were admitted and who were deemed to have a contraindication to anticoagulation prophylaxis. The recently released (April 2012) American College of Radiology Choosing Wisely list of practices specifically recommends forgoing imaging for DVT and PE in the absence of risk factors.[9] The recommendations do not specifically address screening for DVT prior to the initiation of prophylaxis. The goal of this prospective observational study, conducted prior to the Choosing Wisely campaign, was to verify our hypothesis that the prevalence of asymptomatic DVTs was very low, and provide our clinicians with evidence to allay concerns about placement of PCBs without imaging, allowing a practice pattern that would reduce costs without impacting patient safety.

METHODS

RESULTS

A total of 1136 consecutive records were examined; 4 records were excluded from the analysis because they had a diagnosed PE prior to US, and 35 records were excluded because the US was performed beyond 48 hours after admission. The final dataset included 1097 hospital admissions for 1071 patients. Of the 1097 admissions, 759 (69.2%) originated from the emergency department (ED). It is important to note that 70,161 hospital admissions occurred during the same time period, of which 36,363 (51.8%) were admissions that started in the ED. The proportion of patients requiring mechanical DVT prophylaxis is therefore very small (<5%), assuming that a large number of the patients with unplanned admissions would require DVT prophylaxis.

Of the 1071 patients in the final analytical dataset, 544 (50.8%) were male, the mean age was 65.5 years, the mean BMI was 28.7 (median, 27.0) (Table 1), and the majority of the patients were white. US was performed within 24 hours in 712 (66.5%) patients, and 665 (62.1%) had Medicare. An asymptomatic DVT was detected by DUSC in 19 patients (1.8%). None of the clinical and demographic characteristics were statistically different between those with DVT and without (Table 1).

Patients with DVT had at least 1 risk factor; 16 (84.2%) of them had 2 or more risk factors. In addition, the presence of 2 or more risk factors was much more frequent among those with DVT than among those without (84.2% [16/19] vs 58.4% [614/1052], P=0.03).

As shown in Table 2, a history of DVT or PE and ambulatory dysfunction are the only risk factors associated with DVT at admission. In addition, the prevalence of DVT increases as the number of risk factors increases (Table 3). The prevalence is much higher in those who had 4 or more risk factors than among those with fewer than 4 risk factors (12.2% [6/49] vs 1.3% [13/1022], P=0.0001).

Results of the logistic regression, similar to those of the nonadjusted analysis, showed that the only risk factors independently associated with the discovery of a DVT upon DUSC were the presence of ambulatory dysfunction (odds ratio [OR]: 2.99, 95% confidence interval [CI]: 1.13‐7.90) and a history of DVT or PE (OR: 10.51, 95% CI: 3.90‐28.31) (Table 4).

We estimated a savings for Medicare of approximately $266,000 to $280,000 ($261.07 1071 DUSC or $261.07 1022 [after excluding the patients with 4 or more risk factors]) over 13 months had the DUSC not being conducted.

DISCUSSION

This study shows that discovering an asymptomatic DVT is relatively rare (<2%) in patients arriving at the hospital for all causes of admission, even taking into account multiple risk factors that increase the risk for DVTs. The study strongly supports the practice of placing compression devices as soon as possible for those patients who have a contraindication to anticoagulant prophylaxis. Along with reducing the delay to placement while awaiting the test, there is significant cost reduction to the healthcare system by not doing DUSC. There appears to be no need for diagnostic studies prior to the placement of these devices unless the patient has more than 3 risk factors or there is a history of previous DVT or ambulatory dysfunction. This study strongly supports the premise that patients are not arriving with DVTs, but are developing them in the hospital.[1, 2, 10] The 1.8% prevalence of asymptomatic DVT in this study is somewhat lower than that found in other studies. The Prophylaxis for Thromboembolism in Critical Care Trial (PROTECT) tested dalteparin vs unfractionated heparin on 3764 patients in the intensive care unit. Initial screening done to rule out DVT found that 3.5% of patients receiving dalteparin and 3.4% receiving unfractionated heparin had proximal DVTs.[8] Other Investigators used venous compression ultrasound examinations of the lower limbs to determine that 5.5% of patients hospitalized in a medical unit have an asymptomatic DVT of the lower limbs on admission.[5] A limitation of that study is the inclusion of all thrombo emboli, specifically those found in the calf (19 out of 21, or 90%). However, if one eliminates the calf venous thrombi, not considered risk factors for PE, the prevalence of DVT (0.85%) is about half that of our observed 1.8%.

In common with previous studies, a history of previous thromboembolic disease was clearly the most significant of many evaluated risk factors for DVT.[5, 6, 10] Ambulatory dysfunction was also a statistically significant risk factor that was likely under‐reported here because of the inexact documentation in many of the medical records. Interestingly, a history of active malignancy did not prove to be a significant risk factor, contrary to other study reports.[5, 6, 10]

The frequency of asymptomatic DVT appears to increase with the accumulation of risk factors. An asymptomatic DVT existed in 1.3% of the patients with 3 or fewer risk factors, compared with 12.2% of those with 4 or 5 risk factors. It is possible that a higher number of risk factors for DVT would be an indication for obtaining a DUSC prior to the placement of PCBs, although the small number of patients with more than 3 risk factors in our study population may limit the strength of this observation.

CONCLUSION

Our data strongly suggest, in alignment with recent recommendations, that there is no need to perform screening DUSC prior to the placement of prophylactic compression devices among hospital admissions who have contraindications to anticoagulation. Rather, efforts should be focused on implementing systems to ensure rapid placement of these compression devices at the time of admission for those patients who cannot receive anticoagulation prophylaxis. Evaluation for DVT may be of value if there is a history of previous DVT or PE, ambulatory dysfunction, or more than 3 risk factors, as the information may change the therapeutic approach. Current guidelines recommend the measurement of D‐dimers as a screening tool for DVT.[11]

Hospital‐acquired venous thrombus embolism (VTE) is a pressing patient health and safety issue and has been identified as a causal factor in preventable deaths in the hospital setting.[1, 2] More than 540,000 hospitalizations with VTE occur each year among adults in the United States.[3] The number of adults with VTE is anticipated to increase from 0.95 million in 2006 to 1.82 million in 2050.[4] The Institute of Medicine has defined failure to provide adequate thromboprophylaxis to hospitalized, at‐risk patients as a medical error.[2, 5] The American College of Chest Physicians guidelines state that thromboprophylaxis is highly effective at preventing deep vein thrombosis (DVT) and proximal DVT, highly effective at preventing symptomatic VTE and fatal pulmonary emboli (PE), and that the prevention of DVT also prevents PE.[6] Where anticoagulation is contraindicated, mechanical methods of thromboprophylaxis are recommended as preferable to no thromboprophylaxis, with careful attention directed toward ensuring the proper use of, and optimal adherence with, mechanical prophylaxis.[7, 8] In our institution, concerns about the existence of asymptomatic clots being propagated into PEs by the placement of pneumatic compression boots (PCBs), led to routine performance of duplex Doppler ultrasound with compression (DUSC) before applying PCBs to those patients who were admitted and who were deemed to have a contraindication to anticoagulation prophylaxis. The recently released (April 2012) American College of Radiology Choosing Wisely list of practices specifically recommends forgoing imaging for DVT and PE in the absence of risk factors.[9] The recommendations do not specifically address screening for DVT prior to the initiation of prophylaxis. The goal of this prospective observational study, conducted prior to the Choosing Wisely campaign, was to verify our hypothesis that the prevalence of asymptomatic DVTs was very low, and provide our clinicians with evidence to allay concerns about placement of PCBs without imaging, allowing a practice pattern that would reduce costs without impacting patient safety.

METHODS

RESULTS

A total of 1136 consecutive records were examined; 4 records were excluded from the analysis because they had a diagnosed PE prior to US, and 35 records were excluded because the US was performed beyond 48 hours after admission. The final dataset included 1097 hospital admissions for 1071 patients. Of the 1097 admissions, 759 (69.2%) originated from the emergency department (ED). It is important to note that 70,161 hospital admissions occurred during the same time period, of which 36,363 (51.8%) were admissions that started in the ED. The proportion of patients requiring mechanical DVT prophylaxis is therefore very small (<5%), assuming that a large number of the patients with unplanned admissions would require DVT prophylaxis.

Of the 1071 patients in the final analytical dataset, 544 (50.8%) were male, the mean age was 65.5 years, the mean BMI was 28.7 (median, 27.0) (Table 1), and the majority of the patients were white. US was performed within 24 hours in 712 (66.5%) patients, and 665 (62.1%) had Medicare. An asymptomatic DVT was detected by DUSC in 19 patients (1.8%). None of the clinical and demographic characteristics were statistically different between those with DVT and without (Table 1).

Patients with DVT had at least 1 risk factor; 16 (84.2%) of them had 2 or more risk factors. In addition, the presence of 2 or more risk factors was much more frequent among those with DVT than among those without (84.2% [16/19] vs 58.4% [614/1052], P=0.03).

As shown in Table 2, a history of DVT or PE and ambulatory dysfunction are the only risk factors associated with DVT at admission. In addition, the prevalence of DVT increases as the number of risk factors increases (Table 3). The prevalence is much higher in those who had 4 or more risk factors than among those with fewer than 4 risk factors (12.2% [6/49] vs 1.3% [13/1022], P=0.0001).

Results of the logistic regression, similar to those of the nonadjusted analysis, showed that the only risk factors independently associated with the discovery of a DVT upon DUSC were the presence of ambulatory dysfunction (odds ratio [OR]: 2.99, 95% confidence interval [CI]: 1.13‐7.90) and a history of DVT or PE (OR: 10.51, 95% CI: 3.90‐28.31) (Table 4).

We estimated a savings for Medicare of approximately $266,000 to $280,000 ($261.07 1071 DUSC or $261.07 1022 [after excluding the patients with 4 or more risk factors]) over 13 months had the DUSC not being conducted.

DISCUSSION

This study shows that discovering an asymptomatic DVT is relatively rare (<2%) in patients arriving at the hospital for all causes of admission, even taking into account multiple risk factors that increase the risk for DVTs. The study strongly supports the practice of placing compression devices as soon as possible for those patients who have a contraindication to anticoagulant prophylaxis. Along with reducing the delay to placement while awaiting the test, there is significant cost reduction to the healthcare system by not doing DUSC. There appears to be no need for diagnostic studies prior to the placement of these devices unless the patient has more than 3 risk factors or there is a history of previous DVT or ambulatory dysfunction. This study strongly supports the premise that patients are not arriving with DVTs, but are developing them in the hospital.[1, 2, 10] The 1.8% prevalence of asymptomatic DVT in this study is somewhat lower than that found in other studies. The Prophylaxis for Thromboembolism in Critical Care Trial (PROTECT) tested dalteparin vs unfractionated heparin on 3764 patients in the intensive care unit. Initial screening done to rule out DVT found that 3.5% of patients receiving dalteparin and 3.4% receiving unfractionated heparin had proximal DVTs.[8] Other Investigators used venous compression ultrasound examinations of the lower limbs to determine that 5.5% of patients hospitalized in a medical unit have an asymptomatic DVT of the lower limbs on admission.[5] A limitation of that study is the inclusion of all thrombo emboli, specifically those found in the calf (19 out of 21, or 90%). However, if one eliminates the calf venous thrombi, not considered risk factors for PE, the prevalence of DVT (0.85%) is about half that of our observed 1.8%.

In common with previous studies, a history of previous thromboembolic disease was clearly the most significant of many evaluated risk factors for DVT.[5, 6, 10] Ambulatory dysfunction was also a statistically significant risk factor that was likely under‐reported here because of the inexact documentation in many of the medical records. Interestingly, a history of active malignancy did not prove to be a significant risk factor, contrary to other study reports.[5, 6, 10]

The frequency of asymptomatic DVT appears to increase with the accumulation of risk factors. An asymptomatic DVT existed in 1.3% of the patients with 3 or fewer risk factors, compared with 12.2% of those with 4 or 5 risk factors. It is possible that a higher number of risk factors for DVT would be an indication for obtaining a DUSC prior to the placement of PCBs, although the small number of patients with more than 3 risk factors in our study population may limit the strength of this observation.

CONCLUSION

Our data strongly suggest, in alignment with recent recommendations, that there is no need to perform screening DUSC prior to the placement of prophylactic compression devices among hospital admissions who have contraindications to anticoagulation. Rather, efforts should be focused on implementing systems to ensure rapid placement of these compression devices at the time of admission for those patients who cannot receive anticoagulation prophylaxis. Evaluation for DVT may be of value if there is a history of previous DVT or PE, ambulatory dysfunction, or more than 3 risk factors, as the information may change the therapeutic approach. Current guidelines recommend the measurement of D‐dimers as a screening tool for DVT.[11]