Traumatic Brain Injury Associated Radiological Deep Venous Thrombosis Incidence and Significance

Deep Venous Thrombosis (DVT) is common following trauma and patients with TBI are at increased risk. In studies of TBI that screened for DVT in the proximal lower limb veins, the incidence is between 14 and 18%. Pharmacological thromboprophylaxis (PT) may be effective in reducing the incidence of DVT following TBI. However, there is a reluctance to commence it due to concerns about the risk of intracranial haematoma expansion and contusion enlargement. PT is therefore often delayed, exposing patients to potential risks and complications of DVT. Recent systematic reviews could not answer the question of when to commence PT as no high quality randomised controlled trials addressing this question have been conducted. The latest edition of the Brain Trauma Foundation Guidelines could not make a recommendation on this issue . A Delphi exercise involving the investigator's network of United Kingdom(UK) Eurotherm3235 study centres found that the leading concern of specialists in UK neurocritical care, that they felt required further study, was when to commence PT following moderate to severe TBI.

Paradoxically however, despite these concerns over the use of PT in this population of patients, evidence of harm associated with the confirmed presence of proximal DVT is not yet convincing. This has additional significance as without clear evidence of harm associated with DVT complicating TBI, trials to address the questions of safety and efficacy of earlier use of PT are difficult to design and unlikely to attract the necessary funding.

From a search of a trauma registry of patients at risk of DVT who underwent venous screening, in TBI patients the occurrence of DVT was associated with increased ventilator days (Mean (SD), 18.0 (14.9) vs. 9.3 (7.9) days for DVT vs. no DVT, p=0.014) and ICU length of stay (23.0 (14.5) vs. 13.1 (9.8) days for DVT vs. no DVT, p=0.006). Similar results were seen in a post hoc analysis of the EPO TBI trial. This was a randomised controlled trial (RCT) of the effect of erythropoietin (EPO) verses placebo on outcome following TBI. Repeated lower limb screening for the presence of DVT was conducted as part of the protocol. Both ventilator days (Median (IQR) 11 (6-17) vs. 8 (4-14) days for DVT vs. no DVT, p<0.001) and ICU length of stay (17 (11-22) vs. 12 (6-19) days DVT vs. no DVT, p<0.001). Good neurological outcome at 6 months was significantly less frequent in those that developed DVT (45.4 vs. 57.9% DVT vs. no DVT, p=0.01). Whilst these findings are clinically important they are the result of retrospective and post hoc analysis. They therefore require to be prospectively validated if they are to be accepted.

One possible explanation for an increase in ventilator dependency and ICU length of stay could be the occurrence of undiagnosed pulmonary emboli causing acute lung injury (ALI). The occurrence of ventilation perfusion mismatching, deterioration in arterial partial pressure: fraction inspired oxygen ratio (PaO2:FiO2) and radiological changes on chest x-ray, being interpreted as evidence of pulmonary infection. This is supported by local audit data from June to October 2017 which found that 37% of brain injured patients had a clinical diagnosis of ventilator associated pneumonia (VAP) and were treated with antibiotics but lacked subsequent microbiological confirmation. It is also possible that small pulmonary emboli might predispose to pulmonary infection, explaining why the incidence of ventilator associated pneumonia (VAP) remains high in this population (40%, in the same local audit) despite the adoption of measures which have lowered VAP rates in other critically ill patient groups.

Lung Brain "cross-talk" could link evidence of ALI to a negative impact on neurological outcome in these patients. A disproportionate increase in mortality in patients with TBI and ALI has been reported. This was greatly in excess of deaths attributable to pulmonary dysfunction. Similarly, in animal models, ALI was found to be associated with cognitive impairment, and biomarker/histological evidence of neuronal injury.