Utility of lower extremity ultrasound prior to application of serial compression device in patients with COVID-19: “choosing wisely” initiative at a major referral center in the Middle East

Abstract

Objective

See More

The link between venous thromboembolism (VTE) and coronavirus disease 2019 (COVID-19) infection has been consolidated by many studies in the literature. The increased risk of VTE among COVID-19 patients, on the one hand, and the morbidity that can be associated with the ICU course, on the other hand, promote quality care among this patient population. “Choosing wisely” is a quality improvement initiative that emphasizes the importance of assessing the utility of diagnostic tests. Our study was a “Choosing Wisely” single-center initiative aimed at assessing the utility of LEUS among COVID-19 patients who are treated at AUBMC, a major referral center in the Middle East.

Methods

Data from hospitalized COVID-19 patients who underwent LEUS during the pandemic between 2019 and 2021 at our institution were retrospectively analyzed. LEUS was ordered to screen for preexisting DVT prior to the application of mechanical DVT prophylaxis via a serial compression device (SCD) or to rule out suspected DVT. Data on patients’ demographics, comorbidities, and mortality were also retrieved.

Results

A total of 179 patients were included in this study. The mean age of the patients was 66.09 ± 16.587 years, and 108 (60.3%) of our patients were men. Ninety-four (52.5%) patients underwent LEUS for asymptomatic DVT screening prior to SCD placement, and 84 (46.9%) patients underwent LEUS to rule out suspected DVT in the context of other causes, namely, prolonged hospital stay, immobilization, and other hypercoagulable risk factors. Among the 94 patients who underwent LEUS screening, 12 (12.76%) patients were found to have DVT, and SCD placement was consequently aborted. Half of these patients had an IVC filter placed afterward. A previous history of DVT or pulmonary embolism (PE) was strongly associated with DVT occurrence in ICU and non-ICU patients. The mortality rate was 88 (49.2%) among the studied population, which was the highest among the ICU patients (88 (69.8%) with p < 0.001).

Conclusion

Compared with the available literature, we report a greater incidence of asymptomatic DVT among COVID-19 patients, including those screened prior to SCD. We suggest that the clinical utility of LEUS for this patient population outweighs its cost and presumed low benefit. Prospective studies with larger sample sizes are needed to further assess the utility of LEUS and promote the “Choosing Wisely” initiative.

Introduction

Multiple studies have reported an increased incidence of venous thromboembolism (VTE) among this patient population following the coronavirus disease 2019 (COVID-19) pandemic [1, 2]. This was particularly relevant for patients with moderate to severe COVID-19 illness [3,4,5]. Studies suggest that a combination of endothelial dysfunction [5] and hypercoagulability could account for the prothrombotic state among infected patients. Importantly, the decreased mobility associated with admission to the intensive care unit (ICU), sedation, and/or invasive mechanical ventilation, along with any preexisting malignancy or hypercoagulable state, constitute important predisposing factors for VTE, thus further complicating the course of COVID-19 illness [4, 6, 7].

The increased risk of VTE among COVID-19 patients, on one hand, and the morbidity that can be associated with the ICU course, on the other hand, promote quality care among this patient population. A significant number of these patients require prophylactic deep venous thrombosis (DVT) protection with serial compression device (SCD) placement, anticoagulation, or both [2]. SCDs have been associated with the risk of thrombus dislodgment and VTE in cases of preexisting DVT [8, 9]. As a result, many institutional policies recommend routine screening for DVTs prior to applying SCDs. While screening for DVT with lower extremity ultrasound (LEUS) has been suggested and practiced in the ICU setting, there is no clear evidence of its cost-effectiveness and clinical utility [10, 11]. Our colleagues, Younes et al., reported the low utility of LEUS, with only 7% of ICU patients testing positive for DVT at the American University of Beirut Medical Center (AUBMC) in 2017 [12]. To date, no strong evidence supports or refutes this practice for COVID-19 patients in either the ICU or non-ICU setting. As a result, the utility of LEUS in this patient population and its cost-effectiveness remains a subject of debate.

“Choosing Wisely” is a national initiative led by the American Board of Internal Medicine Foundation and supported by specialty societies such as the American Society of Hematology (ASH) [13]. It aims to reduce the overuse of diagnostic tests and treatments that offer limited benefit, especially when weighed against their cost, potential complications, and overall utility [14]. In the context of VTE prophylaxis, the initiative encourages clinicians to critically assess the necessity of tests like LEUS by identifying patient populations who are most likely to benefit. This aligns with the goal of tailoring care based on risk stratification, clinical judgment, and available evidence.

Our study was a single-center “Choosing Wisely” initiative aimed at evaluating the screening utility of LEUS in hospitalized COVID-19 patients prior to SCD application. By identifying asymptomatic DVTs in patients who otherwise would have received mechanical prophylaxis (SCD), we sought to assess whether this screening approach can contribute to safer, evidence-driven care for COVID-19 patients. A secondary aim was to assess the incidence and risk of developing DVT in COVID-19 patients admitted to the ICU and non-ICU wards while focusing on the prevalence of DVT detected by LEUS in patients planned for SCD placement.

Methodology and research design

Study design

During the study period (March 2020–December 2021), 2578 patients tested positive for COVID-19 at our institution. Of those, 387 patients were hospitalized (ICU or non-ICU wards). Among hospitalized patients, 179 patients underwent LEUS and were included in the analysis. Confirmed COVID-19 patients are patients who tested positive for COVID-19 through reverse transcriptase‒polymerase chain reaction (RT‒PCR) in a tertiary care center. Clinical characteristics and patient outcomes were evaluated. The inclusion criteria were as follows: (1) more than 18 years of age, (2) positive RT‒PCR result for COVID-19, and (3) underwent lower extremity ultrasound. The exclusion criteria were as follows: (1) negative RT‒PCR results for COVID-19; (2) patients with missing data and incomplete medical records; and (3) patients with symptoms of DVT (calf swelling or collateral superficial veins presenting localized tenderness along the deep venous system). The study was conducted according to the principles outlined in the Declaration of Helsinki and was approved by our Institutional Review Board, BIO-2022-0037.

Data collection and variables

Clinical characteristics, demographics, laboratory variables, and patient outcomes were collected manually from patients’ electronic medical records. Primary laboratory variables collected included D-dimer, international normalized ratio (INR), prothrombin time (PT), activated partial thromboplastin time (aPTT), white blood cell count (WBC), platelet count, and C-reactive protein (CRP). The Wells score was calculated retrospectively based on chart review and clinical documentation at the time of LEUS. The score includes clinical criteria such as active cancer, recent immobilization, localized tenderness, and previous DVT, among others. Patients were categorized as low, moderate, or high probability for DVT. Acute respiratory distress syndrome (ARDS) was defined based on the Berlin criteria: acute onset, bilateral pulmonary infiltrates not fully explained by cardiac failure or fluid overload, and hypoxemia with PaO2/FiO2 ≤ 300 mmHg. ARDS outcome was recorded during the hospital course. The grouping principle was based on hospital location at the time of ultrasound examination, patients were categorized into either the ICU or the non-ICU COVID-19 floors. Hospital mortality was defined as death following ICU admission in ICU patients, and as death during hospitalization in non-ICU patients.

Lower extremity duplex ultrasound assessment

LEUS examinations were conducted using high-frequency linear transducers and followed a standardized institutional protocol under the supervision of board-certified radiologists. The color Doppler flow imaging technique assessed compressibility of the common femoral, superficial femoral, popliteal, and posterior tibial veins, and evaluated spontaneous flow, phasicity, and augmentation. A non-compressible vein with absence of flow or incomplete color filling was considered positive for DVT. All studies were performed at bedside to minimize patient transport and infection risk. Indeed, LEUS was done in the ICU for ICU patients and on the COVID-19 floors for non-ICU patients. Venous duplexes were obtained in patients with suspected DVT on the basis of a clinical assessment or prior to SCD application. As per our institutional DVT prophylaxis protocol, patients who were not candidates for anticoagulation prophylaxis were offered SCDs as mechanical prophylaxis. Typically, such patients receive a duplex scan of the lower extremity venous duplex prior to the application of SCD prophylaxis. Additionally, SCDs have also been used as supplementary interventions for medical anticoagulation in severely ill patients.

Statistical analysis

Statistical analysis was performed using SPSS software, version 28. Categorical variables are summarized as frequencies and percentages [N (%)], while continuous variables are reported as mean ± standard deviation if normally distributed or median with interquartile range (IQR) if not. Normality was assessed using the Shapiro–Wilk test.

Group comparisons between ICU and non-ICU patients were conducted using Chi-square tests for categorical variables and independent samples t-tests or Mann–Whitney U tests for continuous variables, as appropriate. To identify independent predictors of DVT, we first performed univariate logistic regression for all clinically relevant variables. Variables with a p-value < 0.1 or those with established clinical significance were considered for multivariable modeling. To minimize overfitting, we applied the rule of 10 events per variable, limiting the model to three predictors based on our 30 outcome events. The final multivariable model included age, D-dimer level, and history of DVT or PE, selected for their strong association with thromboembolic risk and clinical relevance. Statistical significance was set at p < 0.05.

Results

Patient characteristics

Over the period from March 1st, 2020, to December 31st, 2021, 179 adult patients had COVID-19 infection and underwent LEUS, thus meeting our study inclusion criteria. Among those, 126 were ICU patients, and 53 were admitted to regular COVID-19 floors. One hundred and eight patients in the total population were males (60.3%), constituting 61.1% of the ICU patients and 31 (58.5%) of the non-ICU patients (p value = 0.744). The most prevalent comorbidities were hypertension (108 patients; 60.3%), diabetes (76 patients; 42.5%), and malignancies (58 patients; 32.4%), among others. There were 86 smokers (48%) in the total population, but there was no statistically significant difference in their distribution among ICU patients (56 smokers (44.4%) versus 67 nonsmokers (53.2%) and non-ICU patients (30 smokers (56.6%) versus 23 nonsmokers (43.3%) (P value 0.178)). Similarly, the difference in the Wells score between the ICU and non-ICU patients was not statistically significant, with a reported p value of 0.480. Further details are provided in Table 1.

Table 1 Baseline characteristics and clinical variables

During hospitalization, 81 patients (45.3%) had SCD placement either as an adjunct to anticoagulation in severely ill and immobilized patients or as an alternative to anticoagulation when contraindicated. Among those patients, 71 (56.3%) were ICU patients, whereas 10 (18.9%) were non-ICU patients. Routine blood chemistry data at admission and coagulative studies were similar on average between ICU and non-ICU patients, with differences not being statistically significant, as shown in Table 1.

Utility of LEUS

According to our institutional guidelines, all patients planned for SCD placement should undergo LEUS to rule out an underlying preexisting DVT. Our primary goal was to assess the utility of LUES in detecting asymptomatic DVT. Among the 179 patients with COVID-19 who underwent LEUS, 94 (52.5%) had LEUS performed for asymptomatic DVT screening prior to SCD placement. The remaining patients were distributed as follows: 84 (46.9%) patients underwent LEUS to rule out suspected DVT in the context of other causes, namely, prolonged hospital stay, immobilization, other hypercoagulable risk factors, and 1 (0.6%) patient had LEUS for elevated D-dimer. Of those 94 patients who underwent LEUS for screening for DVT prior to SCD application, 12 (12.76%) patients were found to have DVT, and SCD placement was consequently aborted. Half of these patients had an IVC filter placed afterward.

Among the 94 patients who underwent screening via LEUS, 81 patients were in the ICU, and 13 were on the COVID-19 regular floor. Nine (11.1%) of the 81 patients in the ICU reported having DVT, whereas 3 (23.1%) of the 13 in the non-ICU setting reported having DVT. The difference in DVT incidence between these two subgroups was not statistically significant (p value: 0.230).

Patient outcomes

During their hospital stay, 12 (6.7%) of 179 patients developed a pulmonary embolism (PE) event between the ICU (9 patients, i.e., 7.1% of the ICU patients) and the non-ICU setting (3 patients, 5.7%) (P value = 0.717). With respect to DVT, 27 (15.1%) out of 179 patients developed DVT. When categorized into ICU and non-ICU settings, 20 (15.9%) of the ICU patients had DVT, whereas 7 (13.2%) had non-ICU DVT (p value = 0.649). Hospital mortality was 0% in the non-ICU setting and 69.8% in the ICU setting, with an overall hospital death of 49.2% in the total population (p < 0.001). Finally, the incidence of ARDS was greater in ICU patients (52 (41.3%)) than in non-ICU patients (1.9%) (p < 0.001) (Table 2).

Table 2 Patient outcomes among patients in the ICU versus non-ICU units

When patients were categorized into those with or without malignancy (Table 3), both had similar rates of DVT (15.5% versus 14.9%, p = 1) and PE occurrence (5.2% versus 7.4%, p = 0.75). Additionally, there was no statistically significant difference in hospital the incidence of ARDS (34.5% versus 27.3%, p = 0.38) or hospital mortality (51.7% versus 50.4%, p = 0.87).

Table 3 Hospital outcomes in patients with or without previous malignancy

Baseline differences among patients with or without DVT occurrence

As shown in Table 4, patients who developed DVT during admission were more likely to have a history of DVT or PE prior to admission (both p < 0.001). There was statistically significant difference in other baseline characteristics among patients with or without DVT occurrence.

Table 4 Differences in baseline characteristics among patients with or without occurrence of deep vein thrombosis

Multivariate logistic regression analysis

On univariate analysis, history of DVT or PE was significantly associated with increased risk of DVT occurrence during hospital admission (Odds ratio [OR] 6.6, p < 0.001). Higher D-dimer levels were also associated with increased risk of DVT but did not reach statistical significance (p = 0.09). On multivariable analysis, history of DVT/PE remained significantly associated with higher risk of DVT occurrence (OR 8.1, p = 0.001). Age and D-dimer levels were not significantly associated with increased DVT risk. Additional data are provided in Table 5.

Table 5 Multivariate logistic regression assessing the association between DVT occurrence and multiple risk factors

Discussion

To our knowledge, this is the first study to evaluate the utility of screening LEUS prior to SCD placement in hospitalized COVID-19 patients. We observed a DVT incidence of 15.1% in the overall cohort and 12.76% among patients screened specifically before SCD application. Importantly, SCD placement was aborted in all patients with a positive DVT result, and half required IVC filter placement. These findings underscore the clinical relevance of LEUS in identifying asymptomatic thrombi and guiding management in this high-risk population. This targeted approach is well-aligned with the “Choosing Wisely” initiative, which advocates for evidence-based, high-impact diagnostic strategies and discourages indiscriminate use of low-yield testing.

Owing to its strikingly expansive course, the COVID-19 pandemic has posed a heavy burden on the medical sector worldwide [15]. Consequently, many therapeutic strategies have been implemented on the basis of anecdotal clinical impressions in the face of a restrictive limitation in time, resources, and evidence. LEUS screening is one such example. The predilection of COVID-19 patients with VTE has been extensively reported in the literature [4, 7, 16,17,18,19]. Several studies have focused on identifying the factors that may increase the risk of venous thromboembolism in COVID-19 patients to help identify a patient population that might benefit from screening via duplex ultrasonography [20,21,22]. This pushed many medical centers to use LEUS as a screening tool in patients admitted with COVID-19 pneumonia given the high prevalence of thrombotic events in this critical population [5, 23].

In a recent meta-analysis by Moores et al., the overall estimated pooled incidence of VTE was close to 17% among COVID-19 patients admitted to ICUs and non-ICUs [24]. This is in accordance with our retrospective study where the incidence of VTE was estimated at 21.8%, with 15.1% being DVTs and the remaining 6.7% being PEs. When categorized into ICU and non-ICU patients, the incidences of asymptomatic DVTs in our study were 15.9% and 13.2%, respectively, without statistically significant differences, highlighting the relatively high incidence of DVT among admitted COVID-19 patients in general. Similarly, there was no difference in the incidence of PE or DVT among patients with or without malignancy [25, 26]. A study by Pieralli et al. reported that the incidence of asymptomatic DVTs in non-ICU settings is 13.7%, which is very close to that reported in our study (13.2%) [27]. For patients admitted to ICU settings, an observational study by Voicu et al. reported a DVT incidence of 35%, which was higher than that reported in our study (15.9%). These findings, in addition to our results, suggest that the incidence of asymptomatic DVT is high not only in critically ill patients with COVID-19 but also in non-ICU and non-cancer patients.

Our data demonstrated that having a previous history of DVT or PE is strongly associated with DVT occurrence in both ICU and non-ICU settings. In addition, ARDS was also found to be significantly associated with DVT occurrence in the ICU setting only. Hence, given the high incidence of asymptomatic DVTs among ICU and non-ICU patients, LEUS appears to be crucial for the early detection of DVTs in patients admitted for COVID-19 pneumonia, particularly those with a high propensity for developing lower extremity DVTs (with a previous history of DVT or PE).

Although the Wells score was retrospectively calculated for all patients in our cohort, it was not the primary determinant for ordering LEUS during the study period. Among those screened, 0.6% had a Wells score of 0, 59.2% had a score of 1, 35.2% had a score of 2, and the remainder had a score of 3, indicating that the majority fell into the intermediate-risk category. In practice, the decision to perform LEUS, particularly prior to SCD placement, was influenced by a combination of factors including contraindications to anticoagulation, prolonged immobility, ICU admission, clinical judgment, and institutional protocols aimed at minimizing the risk of thrombus dislodgment. While LEUS screening was not applied universally to all admitted COVID-19 patients due to the absence of strong supporting guidelines at the time, clinicians at our institution maintained a low threshold for initiating LEUS in this high-risk population. Our findings support this cautious approach, given the substantial incidence of asymptomatic DVTs identified among screened patients, many of which led to meaningful changes in clinical management.

In our institution, SCDs are mostly used in patients with contraindications to anticoagulation. SCDs are also applied in severely ill and immobilized patients, providing an added benefit to anticoagulation medications. All patients in our institution who are planning for SCD application usually undergo screening LEUS to rule out possible asymptomatic DVTs. Many guidelines argue against this practice because of the absence of adequate evidence regarding the effect of LEUS screening on the prognosis and mortality of patients planned for SCD placement [10, 11]. Such guidelines cannot be reliably extended to the COVID-19 population given the prothrombotic state of affected individuals. To our knowledge, no previous studies have addressed this issue in the COVID-19 era. Hence, reporting the incidence of DVT in this specific population is highly important.

We report an incidence of asymptomatic DVT of 12 (12.76%) among all patients planned for SCD application during our study period. Younes et al. conducted a similar study at our institution in 2017, where they assessed the incidence of asymptomatic DVTs in patients prior to SCD placement in the ICU. At that time, they reported an incidence of 7%, which is lower than that obtained in our study [12]. This clearly highlights the rise in the incidence of DVT from the pre-COVID-19 era to the COVID-19 era within this specific population.

In this context, “Choosing Wisely” supports using LEUS selectively in patients for whom the identification of asymptomatic DVTs could alter management decisions, such as withholding SCDs or initiating further prophylactic measures. Our findings, showing a 12.76% incidence of asymptomatic DVT in screened patients, many of whom subsequently had their management changed, align with this framework by highlighting a patient subgroup that clearly benefits from targeted screening. This reinforces the notion that LEUS screening in this context is not excessive but rather clinically actionable.

Practically, ordering a lower extremity venous duplex in a COVID-19 patient prior to SCD placement stresses the availability of resources, especially during a pandemic. Furthermore, such tests include exposure risk to health care professionals that would preferably be avoided if possible. As such, an elaborate assessment of the true incidence of DVT in COVID-19 patients in general and in those planning for SCD placement is urgently needed. On the basis of our findings, the need to screen LEUS in the COVID-19 era prior to SCD application can no longer be ignored given the indisputably increased incidence of asymptomatic DVTs in those patients, especially in patients with prior DVT or PE. Otherwise, such events would be missed, thus increasing the theoretical risk of pulmonary embolism with SCD application.

While our study did not include a formal economic analysis, it is important to consider the potential cost-saving implications of LEUS screening in hospitalized COVID-19 patients. Undiagnosed DVT can lead to serious complications such as PE, which are associated with higher morbidity, prolonged ICU stays, increased need for mechanical ventilation, and elevated healthcare costs [28]. By identifying asymptomatic DVT early, particularly before initiating mechanical prophylaxis with SCDs, LEUS may help prevent iatrogenic harm and reduce downstream resource utilization. Previous studies have shown that the treatment of VTE-related complications is significantly more costly than preventive strategies, especially in high-risk populations such as ICU patients with COVID-19 [28,29,30]. Therefore, although LEUS has an upfront cost, its ability to guide clinical decision-making and prevent adverse events could contribute to overall cost savings. Future prospective studies should incorporate cost-effectiveness models to more accurately quantify such impact. One strength of this study is that the reported incidence of DVT matches that reported in the literature. Additionally, our study is one of the very few that focuses on the utility of lower extremity ultrasound as a screening tool for deep vein thrombosis in general and in patients planning for SCD placement in particular. This question is worth considering the mixed results reported by other studies in the literature [5, 23].

Limitation

This was a retrospective, single-center study with a relatively small sample size, which may limit the generalizability of our findings. Notably, 46.3% (179 out of 387) of the total COVID-19 admissions at our institution during the study period underwent LEUS, introducing the possibility of selection bias. This is particularly relevant given that LEUS was not performed systematically, but rather based on clinical judgment and institutional protocols, especially for patients planned for SCD placement. Detailed data on anticoagulation use at the time of LEUS screening were not consistently available across the cohort, limiting our ability to assess its potential impact on thrombotic outcomes; as such, anticoagulation status was not included in the multivariable analysis and should be considered in future studies. Additionally, patients selected for mechanical prophylaxis rather than anticoagulation may represent a sicker cohort, further contributing to selection bias. As the study was conducted at a single tertiary care center in Lebanon, the findings may not be directly generalizable to other healthcare systems with different patient populations, resource constraints, or clinical workflows. Finally, the lack of a comparison group of COVID-19 patients who did not undergo LEUS limits our ability to assess the screening utility of LEUS across the broader inpatient population. While our findings support the targeted use of LEUS, a randomized controlled trial would be needed to validate its impact on outcomes, a study design that is unlikely to be feasible for ethical and logistical reasons in our setting.

Conclusion

Our study demonstrated a high prevalence of asymptomatic DVT in hospitalized COVID-19 patients, particularly among those screened prior to SCD placement, with a 12.76% detection rate leading to changes in clinical management. These findings support the selective use of lower extremity ultrasound as a targeted screening tool in high-risk patients, aligning with the principles of the “Choosing Wisely” initiative. Prospective, multicenter studies are warranted to further validate these results and evaluate the cost-effectiveness of this approach.