This study is the first to evaluate and explore the co-occurrence of VTE outcomes and trauma care characteristics in an Irish trauma population. Majority of studies investigating VTE in trauma patients originate outside of Europe with few reporting VTE incidence in the European trauma population. During the 7-year period, VTE incidence in our trauma population was 0.0036 cases/people-year (95% CI 0.0 to 3.69). Globally, the estimated rates of VTE in trauma patients range from 0.39 to 11.20% [22, 24]. This variation is mirrored in both DVT rates (0.59 to 31.90%) and PE rates (0.32 to 6.8%) [25, 26]. These differences are largely due to patient demographics, VTE surveillance strategies, reporting of VTE outcomes, and institutional VTE thromboprophylaxis protocols [27].

Despite our low VTE incidence, we were able to parse the heterogeneity of DVT and PE outcomes and associated clinical management, and explore possible links between identified sub-groups, VTE timing, and individual clinical characteristics. Two groups had a high probability of PE and low probability of DVT and two had a high probability of DVT but a low to moderate probability of PE. Additionally, we have found that most PE events occurred independently of DVT with less than 10% of patients developing both concurrent DVT and PE.

In our trauma population, we had a DVT rate of 0.001% per annum with roughly 65% occurring proximally. These DVT were diagnosed in patients with clinical signs suggestive of DVT. None was detected as part of routine surveillance. Evidence has shown that surveillance bias tends to identify more distal DVT, does not reduce the risk of PE or fatal PE, and leads to unnecessary anticoagulation due to false positive finding [12, 28]. Distal DVT do have the same prognostic significance as proximal DVT due to lower risk of embolisation, with studies suggesting that 80% resolve spontaneously [29].

In our study, groups iii, nil risk trauma patients requiring critical care, and iv, moderate risk uncomplicated trauma patients, had a high likelihood of developing DVT only. Both groups underwent surgery within 48 h of admission (more than the half of all patients). This finding is partly consistent with data suggesting a significant association between DVT and surgery in trauma patients, with the risk of thrombosis starting perioperatively [19, 30, 31]. Across groups iii and iv, spinal injury and intracranial haemorrhage occurred in over 50% of patients. Neurosurgery and spinal surgery remain absolute indications for withholding thromboprophylaxis, with the optimal timing to restart postoperatively ranging between 24 and 72 h [32]. Delayed initiation of prophylactic LMWH 12 to 24 h postoperatively has been shown to result in suboptimal antithrombotic effectiveness [33]. Reiff and colleagues [34] demonstrated a three- to fourfold increase in DVT risk in patients with TBI. DVT has been found to occur in one-third of moderate and severe patients with isolated head injuries [35]. Knudson et al. [30] reported patients with major head injury to be more at risk for DVT (OR, 1.34) than for PE (OR, 0.87). Patients with traumatic spinal injuries have an increase VTE risk which is often outweighed by the risk of epidural haematoma expansion [12, 13]. Groups iii and iv reflect a potential description of trauma patients with high probability to develop DVT who may benefit from targeted DVT surveillance and prevention.

In our setting, we reported a PE incidence of 0.002% per annum. In their review, Shuster et al. [8] determined the rates of post trauma PE to be variable and dependent on study design, inclusion, and diagnostic criteria. Differences in PE rates in trauma patients have been attributed to advances in CT technology, 24-h availability of CT scanners, and liberal use of immediate imaging by trauma centres [30, 36]. In a UK major trauma centre study, Glover et al. [37] found a PE rate of 4.6% in their trauma population, with most PEs occurring after 72 h of admission. This is similar to our finding of late PE accounting for 62% of all PE events.

In our most critically unwell trauma patients (group i — Moderate-high risk trauma requiring critical care), nearly two-thirds of all PE events occurred within 48 h of admission and in the absence of DVT. Benns et al. found a significant number of PEs occurring early in the hospital course, with no prior DVT, and concluded that PE occurring early is likely secondary to biochemical mechanisms outside the conventional explanation of distal clot embolisation [11]. In investigating predictors of early versus late PE in trauma patients, Velmahos et al. [10] found that most PE patients did not have evidence of prior DVT and suggested that certain PE may occur de novo within the lungs. Early PE may potentially arise due to local injury and TIC, with later PE a result of embolisation of distal thrombus caused by prolonged immobility and inadequate anticoagulation [8]. Brakenridge et al. [16] found that half of all PEs in their trauma population occurred within the first 4 days after injury, and that patients with severe head injuries were more likely to experience late PE. Gambhir et al. [38] found proximal DVT as the largest risk factors for late PE. This is a crucial finding to this study that in our population, the most critically unwell trauma patients are more likely to develop early PE without any preceding indication of a DVT. In an already vulnerable group, this can significantly compound recovery.

What this study has enabled us to observe, is that, in trauma patients, DVT and PE can occur both concurrently and independently of each other. As opposed to evaluating the contribution of each risk factor individually, we have considered VTE risk as multidimensionally determined, identifying their patterns of co-occurrence with trauma management variables, and exploring their links to types of injury and patients’ characteristic. This is important, as VTE risk in trauma patients has been associated with both individual patient characteristics (age, biological sex, history of VTE, etc.) and interventions carried out as part of their acute trauma resuscitation (femoral venous line insertions, mechanical ventilation, major operative repair, etc.). Therefore, the authors acknowledge that the cause of VTE in trauma patients is multifactorial.

Identifying trauma patients with higher DVT and PE risk can facilitate earlier discussions on thromboprophylaxis strategies, potential surveillance screening, and the use of inferior vena cava filters. This approach to risk, as opposed to a generalisable standard, can enable clinicians to understand the risk hierarchy of PE and DVT and focus DVT and PE prevention practice in accordance with injury and management characteristics.

This study is not without limitations. Due to the retrospective nature, we were limited to a small sample which affected the statistical power of the conducted comparisons and the generalisability of our findings. As part of our methodology, this study utilised the TARN database and inpatient chart records for data collection. Missing data (due to incomplete patient capturing) in the TARN database limited our sample. Due to the lack of a national electronic health record system in Ireland, it is possible that patients may have represented post discharge to hospitals outside of the NIMIS PACS network, resulting in their omission from the VTE ( +) group. In addition, the low rates of routine post-mortem examination may underestimate the incidence of VTE. Lastly, although our hospital has a local VTE guideline, decisions on VTE prophylaxis in trauma patients are at the admitting physician’s discretion. This may have affected the VTE outcomes in our population.

Across major trauma centres, including the UK, decision-making with respect to the timing of VTE prophylaxis in trauma patients is often made by multidisciplinary teams on a case-by-case basis [39]. Individual patient characteristics continue to guide the safe selection and timing of prophylaxis, with daily VTE and bleeding risk assessment recommended as the standard of care [39, 40]. What this study highlights is the need for trauma-specific, VTE prevention guidelines. Understanding that trauma patients are often complex, a support tool can help clinicians set early and safe target times to prophylaxis. The centralised availability of information addressing various injury patterns can assist with the standardisation of VTE prevention practice in Ireland. The national VTE program (NTVEP) is best positioned to provide the clinical and administrative support to supervise the development of such a guideline.