The diagnosis of deep venous thrombosis and pulmonary embolism in medical-surgical intensive care unit patients

Development and Validation of an ICU-Venous Thromboembolism Prediction Model Using Machine Learning Approaches: A Multicenter Study

Purpose

The purpose of this study was to establish and validate machine learning-based models for predicting the risk of venous thromboembolism (VTE) in intensive care unit (ICU) patients.

Patients and Methods

The clinical data of 1494 ICU patients who underwent Doppler ultrasonography or venography between December 2020 and March 2023 were extracted from three tertiary hospitals. The Boruta algorithm was used to screen the essential variables associated with VTE. Five machine learning algorithms were employed: Random Forest (RF), eXtreme Gradient Boosting (XGBoost), Support Vector Machine (SVM), Gradient Boosting Decision Tree (GBDT), and Logistic Regression (LR). Hyperparameter optimization was conducted on the predictive model of the training dataset. The performance in the validation dataset was measured using indicators, including the area under curve (AUC) of the receiver operating characteristic (ROC) curve, specificity, and F1 score. Finally, the optimal model was interpreted using the SHapley Additive exPlanation (SHAP) package.

Results

The incidence of VTE among the ICU patients in this study was 26.04%. We screened 19 crucial features for the risk prediction model development. Among the five models, the RF model performed best, with an AUC of 0.788 (95% CI: 0.738-0.838), an accuracy of 0.759 (95% CI: 0.709-0.809), a sensitivity of 0.633, and a Brier score of 0.166.

Conclusion

A machine learning-based model for prediction of VTE in ICU patients were successfully developed, which could assist clinical medical staff in identifying high-risk populations for VTE in the early stages so that prevention measures can be implemented to reduce the burden on the ICU patients.

Low incidence of deep vein thrombosis in critically ill medical patients in Thais: a prospective study

Background

Critically ill medical patients face a heightened risk of developing venous thromboembolism. In Thailand, routine thromboprophylaxis is not employed. The incidence of deep vein thrombosis (DVT) in the medical intensive care unit (ICU) has not been elucidated in the Thai population.

Objectives

The aims were to evaluate the incidence of DVT and identify associated risk factors in critically ill medical patients.

Methods

A single-center, prospective cohort study was conducted from 2019 to 2020. Consecutive patients underwent screening for proximal DVT by duplex ultrasound of both legs.

Results

A total of 200 patients were enrolled, with 115 being male (57%). The mean (SD) age was 66.5 (16.4) years. The mean (SD) Acute Physiology and Chronic Health Evaluation II score was 27 (8). The cumulative incidence of DVT over 5 days was 7% (95% CI, 3.4%-10.6%). No clinically or radiologically diagnosed pulmonary embolism occurred in patients with DVT. No independent risk factor associated with DVT was identified. Hospital mortality in those with and those without DVT was 42.9% and 32.3%, respectively. There was no significant difference in the length of ICU or hospital stay or inpatient mortality between those with and those without DVT.

Conclusion

Without thromboprophylaxis, the incidence of DVT in the Thai population remains low. A strategy of screening ultrasound 5 to 7 days after admission to the ICU may be a suitable alternative to anticoagulant prophylaxis in critically ill Thai patients without symptoms of venous thromboembolism.

Comparing the application of three thrombosis risk assessment models in patients with acute poisoning: A cross-sectional survey

Background

Patients with acute toxic hemoperfusion are prone to deep vein thrombosis. However, there is no risk assessment model for thrombosis in patients with acute toxic hemoperfusion. Therefore, we compared three commonly used risk assessment models for deep vein thrombosis to determine the model most suitable for assessment of deep vein thrombosis in patients with acute toxic hemoperfusion.

Methods

Caprini, Autar, and Padua thrombosis risk assessment models were used to assess the risk of deep vein thrombosis in patients with acute poisoning and hemoperfusion admitted to a grade A hospital in Shandong province from October 2017 to February 2019. The predictive values of the three models were compared using receiver operating characteristic (ROC) curve analysis.

Results

The risk assessment model scores of Caprini, Autar, and Padua were 7.55 ± 1.76, 8.63 ± 2.36, and 3.92 ± 0.55, respectively. The Caprini risk assessment model was significantly different (p < 0.05) in high-risk patients in the thrombus and non-thrombotic groups; the difference between the other two models was not significant (p > 0.05). The areas under the ROC curve of the Caprini, Autar, and Padua risk assessment models were 0.673, 0.585, and 0.535, respectively. The difference in areas under the ROC curve between the Caprini risk assessment model and the Autar risk assessment model as well as the Padua risk assessment model was significant (p < 0.05), but the areas under the ROC curve of the Autar risk assessment model and the Padua risk assessment model were not statistically significant (p > 0.05). The Caprini risk assessment model had a sensitivity of 91.9%, specificity of 33.1%, and a Youden index of 0.249. The sensitivity and specificity of Autar’s risk assessment model were 37.0 and 77.2%, respectively, and the Youden index was 0.141. The Padua risk assessment model had a sensitivity of 91.3%, specificity of 15.0%, and a Youden index of 0.063.

Conclusion

The three thrombosis risk assessment models were not suitable for patients with acute poisoning and hemoperfusion.

Stroke in SARS-CoV-2 Infection: A Pictorial Overview of the Pathoetiology

Since the early days of the pandemic, there have been several reports of cerebrovascular complications during the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. Numerous studies proposed a role for SARS-CoV-2 in igniting stroke. In this review, we focused on the pathoetiology of stroke among the infected patients. We pictured the results of the SARS-CoV-2 invasion to the central nervous system (CNS) via neuronal and hematogenous routes, in addition to viral infection in peripheral tissues with extensive crosstalk with the CNS. SARS-CoV-2 infection results in pro-inflammatory cytokine and chemokine release and activation of the immune system, COVID-19-associated coagulopathy, endotheliitis and vasculitis, hypoxia, imbalance in the renin-angiotensin system, and cardiovascular complications that all may lead to the incidence of stroke. Critically ill patients, those with pre-existing comorbidities and patients taking certain medications, such as drugs with elevated risk for arrhythmia or thrombophilia, are more susceptible to a stroke after SARS-CoV-2 infection. By providing a pictorial narrative review, we illustrated these associations in detail to broaden the scope of our understanding of stroke in SARS-CoV-2-infected patients. We also discussed the role of antiplatelets and anticoagulants for stroke prevention and the need for a personalized approach among patients with SARS-CoV-2 infection.

Clinical characteristics of acute lower extremity deep venous thrombosis diagnosed by duplex in patients hospitalized for coronavirus disease 2019

OBJECTIVE

Little is known about coronavirus disease 2019 (COVID-19)-associated hypercoagulability. We sought to characterize patients with deep venous thrombosis (DVT) identified after admission for COVID-19.

METHODS

All adult patients admitted to Montefiore Medical Center from March 1, 2020, to April 10, 2020, and undergoing lower extremity venous duplex for DVT evaluation were included. Patients admitted with suspicion of COVID-19 were divided into severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) positive and SARS-CoV-2 negative groups based on in-hospital test results. Patients without clinical suspicion for COVID-19 were not tested. A retrospective case-control study design was used to identify potential risk factors for DVT in patients with COVID-19. Demographic, radiographic, and laboratory values were abstracted and analyzed.

RESULTS

During the study period, 3404 patients with confirmed COVID-19 were admitted to the hospital. Of the 135 SARS-CoV-2 patients who underwent duplex scanning, there were 18 (13.3%) noted to have DVT compared with 72 of the 711 patients (10.1%) who were either SARS-CoV-2 negative or untested. The odds ratio for DVT in COVID-19 was 1.35 (95% confidence interval, 0.78-2.34; P = .289). Baseline characteristics for COVID-19 patients with and without DVT were overall similar. COVID-19 patients with DVT had an elevated median first d-dimer (18.88 μg/mL [interquartile range (IQR), 7.79-20.00] vs 2.55 μg/mL [IQR, 1.45-6.28]; P = .002; reference value, <0.5 μg/mL), average in-hospital d-dimer (median, 11.93 μg/mL [IQR, 8.25-16.97] vs 3.54 μg/mL [IQR, 2.05-8.53]; P < .001) and median fibrinogen level (501.0 [IQR, 440.0-629.0] vs 654.5 [IQR, 535.8-780.0]; P = .002; reference range, 187-502 mg/dL). There was a trend to significance for COVID-19 patients with DVT compared with without DVT in median d-dimer levels at the time of the duplex (13.61 μg/mL [IQR, 4.04-19.97] vs 3.58 μg/mL [IQR, 2.51-9.62]; P = .055) and median ferritin levels (1679.0 ng/mL [IQR, 1168.0-2577.0] vs 1103.0 ng/mL [IQR, 703.5-2076.5]; P = .055; reference range, 25-270 ng/mL). Twelve of the 18 patients with COVID who developed DVT did so despite chemical thromboprophylaxis, and 2 developed DVT despite therapeutic anticoagulation CONCLUSIONS: We found only a modestly increased risk of DVT in patients with COVID-19, likely underestimated owing to limitations in duplex testing early in the epidemic. Elevated d-dimer and a less elevated fibrinogen are associated with DVT in patients with COVID-19 who seem to form thrombus despite conventional chemical thromboprophylaxis. Additionally, an increasing d-dimer over time may be a reflection of the development of DVT in patients with COVID-19.

Prevention of thrombotic risk in hospitalized patients with COVID-19 and hemostasis monitoring

COVID-19 is an infection induced by the SARS-CoV-2 coronavirus, and severe forms can lead to acute respiratory distress syndrome (ARDS) requiring intensive care unit (ICU) management. Severe forms are associated with coagulation changes, mainly characterized by an increase in D-dimer and fibrinogen levels, with a higher risk of thrombosis, particularly pulmonary embolism. The impact of obesity in severe COVID-19 has also been highlighted.In this context, standard doses of low molecular weight heparin (LMWH) may be inadequate in ICU patients, with obesity, major inflammation, and hypercoagulability. We therefore urgently developed proposals on the prevention of thromboembolism and monitoring of hemostasis in hospitalized patients with COVID-19.Four levels of thromboembolic risk were defined according to the severity of COVID-19 reflected by oxygen requirement and treatment, the body mass index, and other risk factors. Monitoring of hemostasis (including fibrinogen and D-dimer levels) every 48 h is proposed. Standard doses of LMWH (e.g., enoxaparin 4000 IU/24 h SC) are proposed in case of intermediate thrombotic risk (BMI < 30 kg/m2, no other risk factors and no ARDS). In all obese patients (high thrombotic risk), adjusted prophylaxis with intermediate doses of LMWH (e.g., enoxaparin 4000 IU/12 h SC or 6000 IU/12 h SC if weight > 120 kg), or unfractionated heparin (UFH) if renal insufficiency (200 IU/kg/24 h, IV), is proposed. The thrombotic risk was defined as very high in obese patients with ARDS and added risk factors for thromboembolism, and also in case of extracorporeal membrane oxygenation (ECMO), unexplained catheter thrombosis, dialysis filter thrombosis, or marked inflammatory syndrome and/or hypercoagulability (e.g., fibrinogen > 8 g/l and/or D-dimers > 3 μg/ml). In ICU patients, it is sometimes difficult to confirm a diagnosis of thrombosis, and curative anticoagulant treatment may also be discussed on a probabilistic basis. In all these situations, therapeutic doses of LMWH, or UFH in case of renal insufficiency with monitoring of anti-Xa activity, are proposed.In conclusion, intensification of heparin treatment should be considered in the context of COVID-19 on the basis of clinical and biological criteria of severity, especially in severely ill ventilated patients, for whom the diagnosis of pulmonary embolism cannot be easily confirmed.

Prevalence and Risk Factors for Thrombotic Complications Following Venovenous Extracorporeal Membrane Oxygenation: A CT Scan Study

OBJECTIVES

The aims of this study were to: 1) analyze the cannula-associated deep vein thrombosis frequency after venovenous extracorporeal membrane oxygenation using a CT scan and 2) identify the associated risk factors for cannula-associated deep vein thrombosis.

DESIGN

Retrospective observational analysis at a single center.

SETTING

Tertiary referral university teaching hospital.

PATIENTS

Patients under venovenous extracorporeal membrane oxygenation with a femorofemoral or femorojugular cannulation admitted for acute respiratory distress syndrome or primary graft dysfunction after pulmonary transplantation. CT scan was performed within 4 days after decannulation.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

We included 105 of 228 patients screened. Bacterial pneumonia was the main indication of venovenous extracorporeal membrane oxygenation (46.7%). CT scans were performed at a median of 2 days (1-3 d) after decannulation. Cannula-associated deep vein thrombosis was found in 75 patients (71.4%) despite it having a mean activated partial thromboplastin time ratio of 1.60 ± 0.31. Femorofemoral cannulation induced femoral cannula-associated deep vein thrombosis more frequently than femorojugular cannulation (69.2% vs 63.1%, respectively; p = 0.04). Seventeen of the 105 patients (16.2%) had a pulmonary embolism. Multivariate logistic regression analysis showed that higher the percentage of thrombocytopenia less than 100 G/L during extracorporeal membrane oxygenation period, lower the risk for developing cannula-associated deep vein thrombosis (hazard ratio, 0.98; 95% CI, 0.98-1.00; p = 0.02).

CONCLUSIONS

Cannula-associated deep vein thrombosis after venovenous extracorporeal membrane oxygenation is a frequent complication. This plead for a systematic vascular axis imaging after venovenous extracorporeal membrane oxygenation. Thrombocytopenia is associated with a reduction in the occurrence of thrombotic events.

Portable Ventilation/Perfusion Scanning is Useful for Evaluating Clinically Significant Pulmonary Embolism in the ICU Despite Abnormal Chest Radiography

OBJECTIVE

Computed tomography angiography is limited in the intensive care unit (ICU) due to renal insufficiency, hemodynamic instability, and difficulty transporting unstable patients. A portable ventilation/perfusion (V/Q) scan can be used. However, it is commonly believed that an abnormal chest radiograph can result in a nondiagnostic scan. In this retrospective study, we demonstrate that portable V/Q scans can be helpful in ruling in or out clinically significant pulmonary embolism (PE) despite an abnormal chest x-ray in the ICU.

DESIGN

Two physicians conducted chart reviews and original V/Q reports. A staff radiologist, with 40 years of experience, rated chest x-ray abnormalities using predetermined criteria.

SETTING

The study was conducted in the ICU.

PATIENTS

The first 100 consecutive patients with suspected PE who underwent a portable V/Q scan.

INTERVENTIONS

Those with a portable V/Q scan.

RESULTS

A normal baseline chest radiograph was found in only 6% of patients. Fifty-three percent had moderate, 24% had severe, and 10% had very-severe radiographic abnormalities. Despite the abnormal x-rays, 88% of the V/Q scans were low probability for a PE despite an average abnormal radiograph rating of moderate. A high-probability V/Q for PE was diagnosed in 3% of the population despite chest x-ray ratings of moderate to severe. Six patients had their empiric anticoagulation discontinued after obtaining the results of the V/Q scan, and no anticoagulation was started for PE after a low-probability V/Q scan.

CONCLUSION

Despite the large percentage of moderate-to-severe x-ray abnormalities, PE can still be diagnosed (high-probability scan) in the ICU with a portable V/Q scan. Although low-probability scans do not rule out acute PE, it appeared less likely that any patient with a low-probability V/Q scan had severe hypoxemia or hemodynamic instability due to a significant PE, which was useful to clinicians and allowed them to either stop or not start anticoagulation.

The impact of deep vein thrombosis in critically ill patients: a meta-analysis of major clinical outcomes

BACKGROUND

Critically ill patients appear to be at high risk of developing deep vein thrombosis (DVT) and pulmonary embolism during their stay in the intensive care unit (ICU). However, little is known about the clinical course of venous thromboembolism in the ICU setting. We therefore evaluated, through a systematic review of the literature, the available data on the impact of a diagnosis of DVT on hospital and ICU stay, duration of mechanical ventilation and mortality in critically ill patients. We also tried to determine whether currently adopted prophylactic measures need to be revised and improved in the ICU setting.

MATERIALS AND METHODS

MEDLINE and EMBASE databases were searched up to week 4 of June 2012. Two reviewers selected studies and extracted data. Pooled results are reported as relative risks and weighted mean differences and are presented with 95% confidence intervals (CI).

RESULTS

Seven studies for a total of 1,783 patients were included. A diagnosis of DVT was frequent in these patients with a mean rate of 12.7% (95% CI: 8.7-17.5%). DVT patients had longer ICU and hospital stays compared to those without DVT (7.28 days; 95% CI: 1.4-13.15; and 11.2 days; 95% CI: 3.82-18.63 days, respectively). The duration of mechanical ventilation was significantly increased in DVT patients (weighted mean difference: 4.85 days; 95% CI: 2.07-7.63). DVT patients had a marginally significant increase in the risk of hospital mortality (relative risk 1.31; 95% CI: 0.99-1.74; p=0.06), and a not statistically significant increase in the risk of ICU mortality (RR 1.64; 95% CI: 0.91-2.93; p=0.10).

CONCLUSIONS

A diagnosis of DVT upon ICU admission appears to affect clinically important outcomes including duration of ICU and hospital stay and hospital mortality. Larger, prospective studies are warranted.

Screening and prevention of venous thromboembolism in critically ill patients: a decision analysis and economic evaluation

RATIONALE

Venous thromboembolism is difficult to diagnose in critically ill patients and may increase morbidity and mortality.

OBJECTIVES

To evaluate the cost-effectiveness of strategies to reduce morbidity from venous thromboembolism in critically ill patients.

METHODS

A Markov decision analytic model to compare weekly compression ultrasound screening (screening) plus investigation for clinically suspected deep vein thrombosis (DVT) (case finding) versus case finding alone; and a hypothetical program to increase adherence to DVT prevention. Probabilities were derived from a systematic review of venous thromboembolism in medical-surgical intensive care unit patients. Costs (in 2010 $US) were obtained from hospitals in Canada, Australia, and the United States, and the medical literature. Analyses were conducted from a societal perspective over a lifetime horizon. Outcomes included costs, quality-adjusted life-years (QALY), and incremental cost-effectiveness ratios.

MEASUREMENTS AND MAIN RESULTS

In the base case, the rate of proximal DVT was 85 per 1,000 patients. Screening resulted in three fewer pulmonary emboli than case-finding alone but also two additional bleeding episodes, and cost $223,801 per QALY gained. In sensitivity analyses, screening cost less than $50,000 per QALY only if the probability of proximal DVT increased from a baseline of 8.5-16%. By comparison, increasing adherence to appropriate pharmacologic thromboprophylaxis by 10% resulted in 16 fewer DVTs, one fewer pulmonary emboli, and one additional heparin-induced thrombocytopenia and bleeding event, and cost $27,953 per QALY gained. Programs achieving increased adherence to best-practice venous thromboembolism prevention were cost-effective over a wide range of program costs and were robust in probabilistic sensitivity analyses.

CONCLUSIONS

Appropriate prophylaxis provides better value in terms of costs and health gains than routine screening for DVT. Resources should be targeted at optimizing thromboprophylaxis.

Pulmonary embolism in critically ill patients receiving antithrombotic prophylaxis: a clinical-pathologic study

PURPOSE

The true incidence of pulmonary embolism (PE) in critically ill adult patients receiving antithrombotic prophylaxis is unknown, as well as the impact on the outcome. The aim of this study was to assess the incidence of PE in a surgical and medical intensive care unit and to evaluate the presence of risk factors that could be helpful in identifying patients at higher risk of missed diagnosis.

MATERIALS AND METHODS

We retrospectively reviewed the autopsies and clinical data of all patients who died in our intensive care unit from 1996 to 2007. All patients received prophylaxis with subcutaneous low-molecular weight heparin.

RESULTS

Among the 600 autopsies, the clinical diagnosis of PE was confirmed in 13 patients (true positives) and not confirmed in 20 patients; in 73 patients, the PE was discovered only at the autopsy (false negatives [FNs]). The overall incidence of PE in our patients was 14.3%. Pulmonary embolism was considered the cause of death in 45% of FNs and 77% of true positives. Among all comorbidities, only a recent abdominal surgery and the presence of acute renal failure were associated with a higher risk of missed diagnosis. In the FN group, there was a significantly higher frequency of cases of septic shock.

CONCLUSIONS

Despite thromboprophylaxis, critically ill patients remain at risk for PE; and because of the difficulty in diagnosing it clinically, the death certificate diagnosis of PE underestimates the problem.

Computerised tomography for the detection of pulmonary emboli in intensive care patients--a retrospective cohort study

Pulmonary emboli are frequently considered as a cause for respiratory deterioration in intensive care unit (ICU) patients, however empirical observation suggests that computerised tomographic (CT) angiography is infrequently positive after the first 24 hours. This study aimed to determine the rate and risk factors for detection of pulmonary emboli by CT angiography in ICU patients. All patients undergoing CT angiography > 24 hours after ICU admission for respiratory deterioration from April 2000 until January 2004 were included. The positivity rate for pulmonary emboli was determined and risk factors analysed. Seven (6%) out of 113 CT angiograms were positive for pulmonary emboli. All were found in trauma patients. Comparing positive to negative scans, predefined risk factors including head injury (5/7 positive scans, 71% vs. 23/106 negative scans, 22%, P = 0.005), spine injury with neurological impairment (4/7, 57% vs. 9/106, 8%, P = 0.002) and lower limb injury (3/7, 43% vs. 12/106, 9%, P = 0.039) were significantly more frequent in patients with positive scans. Deep vein thrombosis prophylaxis was employed less frequently prior to a positive scan (in 3/7, 43% patients with positive scans vs. 91/106, 86% patients with negative scans P = 0.015). Only the predefined risk factors were independently associated with positive CT angiography on limited logistic regression (OR 24.7 per risk factor, 95% CI 2.38 to 255.1, P = 0.007). Pulmonary emboli were infrequently diagnosed using CT angiography in ICU patients admitted for more than 24 hours and found only in patients with recognised risk factors.