Upper Extremity Deep Vein Thrombosis: Incidence, Risk Factors, and Effectiveness of Chemoprophylaxis.

OBJECTIVES: Upper extremity deep vein thrombosis (UEDVT) is associated with pulmonary embolism and other complications, but there are no recommendations for UEDVT prophylaxis. The purpose of this study was to establish incidence and risk factors for UEDVT and to determine efficacy of pharmacologic prophylaxis for UEDVT prevention. METHODS: For this retrospective cohort study, we identified medical patients aged 18 years and older admitted to 13 Cleveland Clinic hospitals from January 2011 to December 2019. Patients with venous thromboembolism (VTE) on admission, length of stay <1 day, and who received therapeutic anticoagulation were excluded. The potential risk factors included demographics, comorbidities, and medical procedures. Comorbidities were identified via International Classification of Diseases codes, (ICD9 and ICD10), procedures from flowsheets, and prophylaxis from medications administered in the electronic medical record. DVT events were identified by a combination of International Classification of Diseases codes and confirmed by chart review. We performed multivariable logistic regression to identify independent risk factors and the association between VTE prophylaxis and UEDVT. The model's C statistic was obtained using 1000 bootstrap runs. RESULTS: Of 194,809 patients, 496 (0.25% of cohort, 36.8% of all VTE) developed UEDVT by 14 days. In the logistic regression model (bias-corrected C statistic 0.87), 11 risk factors predicted UEDVT, the strongest being peripherally inserted central catheter (odds ratio [OR] 4.62, 95% confidence interval [CI] 3.81-5.60) and central venous catheter (OR 3.57, 95% CI 2.91-4.37). The predicted risk among individuals ranged from 0.02% to 23.4%. Prophylaxis was negatively associated with the development of UEDVT (OR 0.72, 95% CI 0.60-0.87). CONCLUSIONS: UEDVT is rare but some patients are high risk. Therefore, UEDVT risk factors should be added to VTE risk assessment models, and patients at high risk for UEDVT should receive chemoprophylaxis.

Value of psychosocial evaluation for left ventricular assist device candidates.

OBJECTIVE: Left ventricular assist devices require a psychosocial assessment to determine candidacy despite limited data correlating with outcome. Our objective is to determine whether the Stanford Integrated Psychosocial Assessment for Transplant, a tool validated for transplant and widely used by left ventricular assist device programs, predicts left ventricular assist device program hospital readmissions and death. METHODS: We performed a retrospective analysis of adults at the Cleveland Clinic with Stanford Integrated Psychosocial Assessment for Transplant scores before primary left ventricular assist device program implantation from April 1, 2013, to December 31, 2018. The primary outcome was unplanned hospital readmissions censored at death, transplantation, and transfer of care. The secondary outcome was death. RESULTS: There were 263 patients in the left ventricular assist device program with a median (Q1, Q3) Stanford Integrated Psychosocial Assessment for Transplant score of 16 (8, 28). During a median follow-up 1.2 years, 56 died, 65 underwent transplantation, and 21 had transferred care. There were 640 unplanned hospital readmissions among 250 patients with at least 1 outpatient visit at our center. In a multivariable analysis, Stanford Integrated Psychosocial Assessment for Transplant components but not total Stanford Integrated Psychosocial Assessment for Transplant score was associated with readmissions. Psychopathology (Stanford Integrated Psychosocial Assessment for Transplant C-IX) was associated with hemocompatibility (coefficient 0.21 ± standard error 0.11, P = .040) and cardiac (0.15 ± 0.065, P = .02) readmissions. Patient readiness was associated with noncardiac (Stanford Integrated Psychosocial Assessment for Transplant A-III, 0.24 ± 0.099, P = .016) and cardiac (Stanford Integrated Psychosocial Assessment for Transplant A-low total, 0.037 ± 0.014, P = .007) readmissions. Poor living environment (Stanford Integrated Psychosocial Assessment for Transplant B-VIII) was associated with device-related readmissions (0.83 ± 0.34, P = .014). Death was associated with organic psychopathology or neurocognitive impairment (Stanford Integrated Psychosocial Assessment for Transplant C-X, 0.59 ± 0.21, P = .006). CONCLUSIONS: Total Stanford Integrated Psychosocial Assessment for Transplant score was not associated with left ventricular assist device program readmission or mortality. However, we identified certain Stanford Integrated Psychosocial Assessment for Transplant components that were associated with outcome and could be used to create a left ventricular assist device program specific psychosocial tool.