Relevance of immobility as a risk factor for symptomatic proximal and isolated distal deep vein thrombosis in acutely ill medical inpatients

Machine learning used to determine features of importance linked to overnight stay after patellar tendon repair

Purpose

The purpose of this study is to determine if machine learning is an effective method to identify features of patients who may need a longer postoperative stay following a patellar tendon repair.

Methods

The American College of Surgeons National Quality Improvement Program (ACS-NSQIP) was used to collect 1173 patients who underwent patellar tendon repair. Machine learning (ML) was then applied to determine features of importance in this patient population. Several algorithms were used: Random Forest, Artificial Neural Network, Gradient Boosting, and Support Vector Machine. These were then compared to the American Society of Anesthesiologists (ASA) classification system based logistic regression as a control.

Results

Random Forest (RF) was determined to be the best performing algorithm, with an AUC of 0.72, accuracy of 77.66 %, and precision of 0.79, and recall of 0.96. All other algorithms performed similarly to the control. RF gave the highest permutation feature importance to age (PFI 0.25), BMI (PFI 0.19), ASA classification (PFI 0.14), hematocrit (PFI 0.12), and height (PFI 0.11).

Conclusions

This study shows that machine learning can be used as a tool to identify features of importance for length of postoperative stay in patients undergoing patellar tendon repair. RF was found to be a better performing model than logistic regression at determining patients predisposed to longer length of stay as determined by AUC. This supported the study's hypothesis that ML can provide an effective method for identifying features of importance in patients requiring a longer postoperative stay after patellar tendon repair.

Prevalence and Risk Factors of Deep Venous Thrombosis in Intensive Inpatient Neurorehabilitation Unit

Venous thromboembolism (VTE) (deep vein thrombosis and its complication, pulmonary embolism) is a major cause of morbidity and mortality in hospitalized patients and about 7% of these cases are due to immobility secondary to a neurological impairment. Acquired brain injury (ABI) has also been recognized as one of the main risk factors for VTE. Numerous epidemiological studies have been conducted to assess the risk factors for VTE in institutionalized polytrauma patients, although there is a lack of information about neurorehabilitation wards. Since VTE is often undiagnosed, this prospective study aimed to determine the prevalence and clinical characteristics of lower-limb deep venous thrombosis (DVT) in ABI patients at neurorehabilitation admission.

METHODS

ABI patients were screened for DVT on admission to the intensive rehabilitation unit (IRU) with compression ultrasonography and basal D-dimer assay and were daily clinically monitored until discharge. A total of 127 consecutive ABI patients (mean age: 60.1 ± 17.6 years; 63% male; time from event: 30.9 ± 22.1 days; rehabilitation time in IRU: 84.6 ± 58.4 days) were enrolled.

RESULTS

On admission to the IRU, the DVT prevalence was about 8.6%. The mean D-dimer level in patients with DVT was significantly higher than in patients without DVT (6 ± 0.9 vs. 1.97 ± 1.61, p-value = 0.0001). ABI patients with DVT did not show any significant clinical characteristics with respect to ABI without DVT, although a prevalence of hemorrhagic strokes and patients originating from the Intensive Care Unit and Neurosurgery ward was revealed. During the rehabilitation period, patients with DVT showed a significant difference in pharmacological DVT prophylaxis (high prevalence of nadroparin with 27.3% vs. 1.7%, p-value = 0.04) and a prevalence of transfers in critical awards (36% versus 9.5% of patients without DVT, p-value = 0.05). The mortality rate was similar in the two groups.

CONCLUSIONS

Our research offers a more comprehensive view of the clinical development of DVT patients and confirms the prevalence rate of DVT in ABI patients as determined upon IRU admission. According to our findings, screening these individuals regularly at the time of rehabilitation admission may help identify asymptomatic DVT quickly and initiate the proper treatment to avoid potentially fatal consequences. However, to avoid time-consuming general ultrasonography observation, a more precise selection of patients entering the rehabilitation ward is required.

Perioperative Cerebrospinal Fluid Drain Placement Does Not Increase Venous Thromboembolism Risk After Thoracic and Fenestrated Endovascular Aortic Repair

BACKGROUND

Venous thromboembolism (VTE) incidence after thoracic and fenestrated endovascular aortic repair (TEVAR/FEVAR) is high (up to 6-7%) relative to other vascular procedures; however, the etiology for this discrepancy remains unknown. Notably, patients undergoing TEVAR/FEVAR commonly receive cerebrospinal fluid drains (CSFDs) for neuroprotection, requiring interruption of perioperative anticoagulation and prolonged immobility. We hypothesized that CSFDs are a risk factor for VTE after TEVAR/FEVAR.

METHODS

Consecutive TEVAR/FEVAR patients at a single center were reviewed (2011-2020). Cerebrospinal fluid drains (CSFDs) were placed based on surgeon preference preoperatively or for spinal cord ischemia (SCI) rescue therapy postoperatively. The primary end-point was VTE occurrence, defined as any new deep venous thrombosis (DVT) or pulmonary embolism (PE) confirmed on imaging within 30 days postoperatively. Routine postoperative VTE screening was not performed. Patients with and without VTE, and subjects with and without CSFDs were compared. Logistic regression was used to explore associations between VTE incidence and CSFD exposure.

RESULTS

Eight hundred ninety-seven patients underwent TEVAR/FEVAR and 43% (n = 387) received a CSFD at some point during their care (preoperative: 94% [n = 365/387]; postoperative SCI rescue therapy: 6% [n = 22/387]). CSFD patients were more likely to have previous aortic surgery (44% vs. 37%; P = 0.028) and received more postoperative blood products (780 vs. 405 mL; P = 0.005). The overall VTE incidence was 2.2% (n = 20). 70% (14) patients with VTE had DVT, 50% (10) had PE, and 20% (4) had DVT and PE. Among TEVAR/FEVAR patients with VTE, 65% (n = 13) were symptomatic. Most VTEs (90%, n = 18) were identified inhospital and the median time to diagnosis was 12.5 (interquartile range 7.5-18) days postoperatively. Patients with VTE were more likely to have nonelective surgery (95% vs. 41%; P < 0.001), had higher American Society of Anesthesiologists classification (4.1 vs. 3.7; P < 0.001), required longer intensive care unit admission (24 vs. 12 days; P < 0.001), and received more blood products (1,386 vs. 559 mL; P < 0.001). Venous thromboembolism (VTE) incidence was 1.8% in CSFD patients compared to 3.5% in non-CSFD patients (odds ratio 0.70 [95% confidence interval 0.28-1.78, P = 0.300). However, patients receiving CSFDs postoperatively for SCI rescue therapy had significantly greater VTE incidence (9.1% vs. 1.1%; P = 0.044).

CONCLUSIONS

CSFD placement was not associated with an increased risk of VTE in patients undergoing TEVAR/FEVAR. Venous thromboembolism (VTE) risk was greater in patients undergoing nonelective surgery and those with complicated perioperative courses. Venous thromboembolism (VTE) risk was greater in patients receiving therapeutic CSFDs compared to prophylactic CSFDs, highlighting the importance of careful patient selection for prophylactic CSFD placement.

Cannula associated deep vein thromboses in COVID-19 patients supported with VV ECMO

BACKGROUND

VV ECMO is increasingly used as a rescue strategy for hypercarbic and hypoxic respiratory failure refractory to conventional management, and more than 14,000 patients with COVID-19 related respiratory failure have been supported with VV ECMO to date. One of the known complications of VV ECMO support is the development of cannula-associated deep vein thromboses (CaDVT). The purpose of this study was to identify the incidence of CaDVT in COVID-19 patients supported with VV ECMO as compared to non-COVID-19 patients. We hypothesized that due to the hypercoagulable state and longer duration of VV ECMO support required for patients with COVID-19, a higher incidence of CaDVT would be observed in these patients.

METHODS

This is a single center, retrospective observational study. About 291 non-trauma adult patients who were cannulated for VV ECMO and managed at our institution from January 1, 2014 to January 10, 2022 were included. The primary outcome was the presence of CaDVT 24 h after decannulation in COVID-19 versus non-COVID-19 patients. Our secondary outcome was continued presence of DVT on follow up imaging. CaDVT were defined as venous thrombi detected at prior cannulation sites.

RESULTS

Both groups had a high incidence of CaDVT. There was no significant difference in the incidence of CaDVT in COVID-19 patients compared to non-COVID-19 patients (95% vs 88%, p = 0.13). Patients with COVID-19 had an increased incidence of persistent CaDVT on repeat imaging (78% vs 56%, p = 0.03).

CONCLUSION

Given the high number of post-decannulation CaDVT in both groups, routine screening should be a part of post ECMO care in both populations. Repeat venous duplex ultrasound should be performed to assess for the need for ongoing treatment given the high incidence of CaDVT that persisted on repeat duplex scans.

Age-adjusted D-dimer, clinical pre-test probability-adjusted D-dimer, and whole leg ultrasound in ruling out suspected proximal and calf deep venous thrombosis

D-dimer (DD) and ultrasonography (US) are part of the diagnostic workup for lower-extremity deep vein thrombosis (DVT). Recent studies have shown that adjusting DD level cut-offs by age or clinical pre-test probability (PTP) decreases the use of US. We compared diagnostic accuracy of PTP-adjusted DD and age-adjusted DD in 3883 patients (F: 61.1%; age: 65.3 ± 16.8 y) referred to our unit for clinically suspected DVT. All patients underwent clinical evaluation, DD, and US. Proximal DVT was detected in 477 (12.4%) patients, and distal DVT was isolated in 342 (8.9%) patients. In the remaining 3064 patients there were 23 venous thromboembolic events (0.75%, 95% CI: 0.50-1.12) during the 3-month follow-up. The specificities of DD, age-adjusted DD, and PTP-adjusted DD in patients without high PTP levels were 47% (95% CI: 45-49), 61% (95% CI: 59-62), and 67% (95% CI: 65-68), respectively. The negative predictive value (NPV) was 96% (95% CI: 95-97) for all diagnostic strategies. When only proximal DVTs were considered, the NPV increased to 99% (95% CI: 98-99). US was avoided in 37% (95% CI: 36-38) of patients with a fixed cut-off DD, 48% (95% CI: 47-50) with age-adjusted DD, and 52% (95% CI: 51-54) with PTP-adjusted DD. The failure rate for all DVTs of DD, age-adjusted DD, and PTP-adjusted DD was 2.0% (95% CI: 1.6-2.5), 2.7% (95% CI: 2.2-3.2), and 2.5% (95% CI: 2.1-3.0), respectively. Compared with the standard DD cut-off, both age-adjusted and PTP-adjusted DD reduced the proportion of patients who required US at the cost of a small increase in failure rate.

Massive DVT from the proximal IVC to the pedal vein: Our approach using aspiration mechanical thrombectomy and open surgery thrombectomy

Deep venous thrombosis might present in an acute condition requiring early thrombus removal. Several endovascular and surgical approaches are available with a short treatment time and minimal complications compared to pharmacotherapies. However, due to a lack of evidence, these are not the first treatment choice for deep vein thrombosis. Our case report showed a successful multimodality treatment for an acute-on-chronic massive deep vein thrombosis from the inferior vena cava to the pedal vein. A 47-year-old with chief complaints of cold, significant swelling, and severe pain in her left leg was diagnosed with deep vein thrombosis through Doppler ultrasound and contrast-enhanced computed tomography. The patient received aspiration mechanical thrombectomy with the "kissing catheter" technique, adjunctive stent, percutaneous transluminal angioplasty, and open surgical thrombectomy by Fogarty catheter without recurrence and complication.

Patients with venous thromboembolism after spontaneous intracerebral hemorrhage: a review

Background

Patients with spontaneous intracerebral hemorrhage (ICH) have a higher risk of venous thromboembolism (VTE) and in-hospital VTE is independently associated with poor outcomes for this patient population.

Methods

A comprehensive literature search about patients with VTE after spontaneous ICH was conducted using databases MEDLINE and PubMed. We searched for the following terms and other related terms (in US and UK spelling) to identify relevant studies: intracerebral hemorrhage, ICH, intraparenchymal hemorrhage, IPH, venous thromboembolism, VTE, deep vein thrombosis, DVT, pulmonary embolism, and PE. The search was restricted to human subjects and limited to articles published in English. Abstracts were screened and data from potentially relevant articles was analyzed.

Results

The prophylaxis and treatment of VTE are of vital importance for patients with spontaneous ICH. Prophylaxis measures can be mainly categorized into mechanical prophylaxis and chemoprophylaxis. Treatment strategies include anticoagulation, vena cava filter, systemic thrombolytic therapy, catheter-based thrombus removal, and surgical embolectomy. We briefly summarized the state of knowledge regarding the prophylaxis measures and treatment strategies of VTE after spontaneous ICH in this review, especially on chemoprophylaxis and anticoagulation therapy. Early mechanical prophylaxis, especially with intermittent pneumatic compression, is recommended by recent guidelines for patients with spontaneous ICH. While decision-making on chemoprophylaxis and anticoagulation therapy evokes debate among clinicians, because of the concern that anticoagulants may increase the risk of recurrent ICH and hematoma expansion. Uncertainty still exists regarding optimal anticoagulants, the timing of initiation, and dosage.

Conclusion

Based on current evidence, we deem that initiating chemoprophylaxis with UFH/LMWH within 24–48 h of ICH onset could be safe; anticoagulation therapy should depend on individual clinical condition; the role of NOACs in this patient population could be promising.