Biomarkers and Clinical Characteristics Of Patients With and Without DVT In The Intensive Care Unit (ICU)

Within the first weeks of ICU admission, 10-30% of medical/surgical patients develop venous thromboembolism (VTE), and up to 60% of trauma patients. VTE has significant morbidity including longer mechanical ventilation, ICU and hospital stay. The Caprini VTE risk assessment model (Ann Surg, 2010; 251[2]:344-50) is based on clinical factors such as age and comorbidities. Previous data from our lab has demonstrated the role of soluble P-selectin (sPsel), D-dimer, CRP, von Willebrand factor activity (VWF), and ADAMTS13 in the diagnosis of DVT. The aim of this study is to determine levels of biomarkers and clinical characteristics in relation to Caprini risk level in patients both with and without DVT upon admission to the ICU.

We performed a prospective cohort study at the University of Michigan with patients from the surgical, neurological, trauma and medical ICUs. Inclusion criteria: age > 18 and at risk for DVT (i.e. not on therapeutic anticoagulation) when admitted to the ICU. At baseline, clinical information including demographics and comorbidities, Wells’ and APACHE score was collected, blood was drawn for biomarkers and upper and lower extremity duplex ultrasound (DVU) was performed. Blood samples and DVU were repeated every 7 days. Subjects were recruited within 24 hrs of ICU admission and stratified into low/moderate (Group 1), high (Group 2), and highest (Group 3) risk groups according to the Caprini model; Group 4 were DVT positive on screening ultrasound. Patients were followed until DVT positivity, floor transfer, or 35 days. All patients received VTE prophylaxis and had imaging performed as clinically indicated. ELISAs were performed for sPsel, D-dimer, CRP, and VWF activity, and slightly modified FRETS-VWF73 assay for ADAMTS-13 activity. ANOVA and t-test were used to compare continuous, and Chi-square categorical, variables between the groups. Spearman correlation coefficients were performed to assess the relationship between Caprini score and other variables.

From 12/08-10/12, 145 subjects were recruited. The age range was 20-88 years. DVT positive patients on screening ultrasound were significantly older, more likely to be male, and have a personal history of VTE. There was no difference in BMI or family history of VTE.

DVT positive subjects (Group 4) had a significantly higher baseline mean D-dimer, sPsel, day 1 Wells’ and APACHE score compared to the lowest risk group (Group 1), and a significantly lower ADAMTS13 level (see Table). Baseline D-dimer, Wells’ and APACHE scores rose with increasing Caprini risk group and were also significantly higher in Group 4 vs. Groups 2 and 3. There was a trend towards increasing baseline sPsel and declining ADAMTS13 level with increasing Caprini score. Increasing baseline CRP and VWF was also significantly correlated with increasing Caprini score.

Eight patients (6.2%) developed a new DVT within the 35 day observation period (average day 10); 6/8 (75%) were in the highest risk group (Group 3). There was a trend towards higher baseline D-dimer, CRP, Wells’ and APACHE score in subjects who developed DVT vs. stayed negative. In patients who developed DVT, these markers did not significantly increase over time with serial sampling. There was a significantly higher BMI (mean, SE) 35.9 ± 3.3 vs. 29.4 ± 0.6 kg/m², p<0.01, in those who turned positive vs. stayed negative.

Patients with DVT on presentation to the ICU have higher levels of biomarkers and more severe clinical characteristics including higher Wells’ and APACHE scores, than patients without DVT, with variations related to Caprini risk. For the 6.2% of patients who were initially negative and went on to develop DVT, there was a tendency to increased biomarker levels and more severe clinical characteristics at baseline but not on serial testing.

No relevant conflicts of interest to declare.