In this study we sought to identify thrombogenic risk factors and predictors among multiple injured patients with pelvic and/or acetabular fractures.

Available data regarding TE incidences among trauma patients are inconsistent. These numbers strongly depend on the respective trauma patient collective, including injury mechanisms as well as injury severity. In a retrospective study from the National Trauma Data Bank (NTDB) of the American College of Surgeons incidences for DVT and PE were 1.06% and 0.42%, respectively [9]. Paffrath et al. [5] documented venous TE rates of 1.8% among 7937 trauma patients with an ISS ≥ 9 and Lichte et al. [4] published overall TE rates of 2.8% among 40.846 polytrauma patients. In comparison, the incidence of thromboembolic events in our overall collective comprising multiple injured patients with ISS ≥ 16 and fractures of the pelvis and/or acetabulum was 4.9%, which states a considerable higher incidence.

To further appraise our higher rate of TE events in multiple injured patients one can compare TE rates in trauma patients with isolated pelvic / acetabular fractures. Patients with isolated fractures of the pelvis or acetabulum are also at high risk of developing a thromboembolic complication [3, 10,11,12,13] since these fractures are often associated with injuries to pelvic and sacral vessels and often require longer times of immobilization and bed rest [10, 11]. Moed et al. [12] retrospectively analyzed 13.589 patients with pelvic and acetabular fractures that were treated surgically and documented 113 (0.83%) TE events (DVT: 0.51%; PE: 0.21%; both: 0.12%). Kim [13] performed a scheduled CT-venography 7–14 days after trauma in 55 patients with fracture of the pelvis and 40 patients with acetabular fracture. In his study 32 (33.7%) of 95 patients developed a venous TE complication, PE was present in 9 cases (9.5%). In a similar study, Sen et al. found overall rates for TE events of 28.6% (DVT: 21.4%; PE: 17.9%). Some literature even reports rates for DVT of up to 61% and 10% for PE among traumatized patients with pelvis and/or acetabular fractures [3]. The major difference between their studies and ours is that we report on clinically suspected and then diagnostically confirmed TE complications, while the former stated studies explicitly looked for thromboembolism with scheduled radiographic exams. In light of this, TE events in patients with pelvic / acetabular fracture seem to be more common than it appears and could be a reason for the higher incidence of TE complications in our study in comparison to other investigations of severely injured patients.

It has been widely published that patients with pelvic trauma are at high risk to develop thromboembolic complications and e.g. age, prolonged immobilization and excessive blood transfusion are further contributing factors [14, 15]. Thromboprophylaxis is invariably recommended by guidelines worldwide in the treatment of severely injured patients or trauma patients taken care of in an ICU. In the absence of contraindications (e.g. active bleeding, relevant traumatic brain injury) early pharmacologic thromboprophylaxis is recommended over the use of mechanical prophylaxis (e.g. intermittent pneumatic compression). The routinely use of inferior vena cava filters for thromboprophylaxis is not recommended in the respective guidelines and the placement of this device is rather reserved for special circumstances [16,17,18,19,20]. Since the vast majority of patients included in our analysis were treated in German hospitals we can only assume a pharmacological treatment according to the German guideline [20], because details beyond the stated presence or absence of mechanical/pharmacological prophylaxis are not recorded in the TR-DGU. The adherence to current guidelines in our cohort is supported by the fact that 96.4% of patients suffering from TE received prior thromboprophylaxis treatment. This is in line with a recent survey on venous thromboembolism prophylaxis after pelvic and acetabular fractures where—although guidelines on thromboprophylaxis exist—no consensus was reported on the actual treatment practices [21].

Published data suggest that TE development is associated with increasing age. In a study by Fuchs et al. [22] age > 40 years was identified as an independent risk factor for thromboembolic complications. Similar, significantly higher numbers of TE events in patients aged > 60 years were reported in a study by Lichte et al. [4] Further studies confirmed the relationship between TE complications and increasing age [1, 9, 23, 24]. In line with the current literature we identified age > 65 years as an independent risk factor for TE complications in multiple injured patients with fractures of the pelvis and/or acetabulum.

Previous studies have highlighted the influence of certain underlying medical conditions on the development of thromboembolic complications. Especially pre-existing cardiac, respiratory or musculoskeletal disorders place trauma patients at higher risk for complications, i.e. thromboembolism. In the study by Paffrath et al. [5], underlying medical disorders such as diabetes, renal failure, malignancies, congenital or acquired coagulation disorders were independently associated with thromboembolic events. Our findings support the current evidence. Patients with relevant pre-existing medical disorders as defined by American Anesthesiologist`s ASA score of 3 or 4 were more likely to develop a TE complication in our study. Possibly, establishing risk scores and identifying trauma patients at higher risk for TE development could lead to a more individual and patient-tailored antithrombotic therapy.

In our descriptive analysis higher numbers of TE complications were associated with increasing injury severity (ISS) and injuries to the trunk (thorax, abdomen, spine). Our multivariate logistic regression model revealed (severe) abdominal injury (AISAbdomen ≥ 3) as an independent risk factor for the occurrence of thromboembolism. The relevance of overall injury severity as a risk factor for TE complications has been widely studied [5, 25,26,27,28]. In contrast, sufficient evidence on different injury patterns or injury localizations associated with higher thrombogenic risk among multiple injured patients is missing. Data suggesting higher thromboembolic risk exists for thoracic, spinal, abdominal, pelvic and lower extremity trauma as well as penetrating injuries in polytrauma patients [4, 9, 25, 28,29,30,31,32,33]. Our data suggest higher TE rates associated with injuries to the thorax, abdomen and spine in multiple injured patients with pelvic or acetabular injury. However, in our collective, multivariate logistic regression found that severity of pelvic or acetabular injury is not an independent risk factor for TE complications in multiple injured patients with pelvic or acetabular injury. This is true despite our significantly higher rates of TE complications in patients suffering from more than serious acetabular injury (AISAcetabulum ≥ 3) or more then severe pelvic injuries (AISPelvis ≥ 4). One reason and possible explanation might be that these patients also suffer from relevant hemorrhage, coagulopathy, which already proved to be an independent risk factors for TE complications in our study cohort. Also, supposedly, patients with trauma to the trunk are more likely to suffer from TE events because of higher numbers of hemorrhage in that cohort, increasing number of operative procedures and longer time of immobilization [4, 9, 25, 28,29,30,31,32,33]. As hemorrhage, extensive surgery and sepsis are common complications in patients with abdominal injury, these patients pose an intensive thrombogenic risk [1]. As expected, neither in our univariate nor the multivariate analysis moderate traumatic brain injury had an impact on thromboembolic complications. These finding are in line with the current literature [34].

Multiple traumatized patients with pelvic/acetabular fractures often suffer from internal or external bleeding and hemorrhagic shock [5, 10, 35]. Hence, resuscitation often requires transfusion of high amounts of blood products [4, 5]. In the present study either the presence of hemorrhagic shock or the need for mass transfusion (≥ 10 PRBCs) significantly increased the thromboembolic risk.

Differences concerning thromboembolic complications with regard to pelvic and acetabulum fracture type or severity are little investigated in the literature. Our univariate analysis revealed higher TE rates with increasing fracture complexity of pelvic/acetabular injuries. In contrast, multivariate analysis did not identify pelvic and acetabular fractures as independent risk factors for TE development. Bagaria [36] observed that injuries to posterior structures in type B and C pelvic fractures are associated with kinking of iliac and femoral vessels thus, supposedly, promoting DVT and PE. In a study by Kim et al. [13], vertical sheer type injuries according to the Young-Burgess classification were associated with significantly higher chance for thromboembolism when compared to anterior compression type and lateral compression type pelvic injury. When the relationship between thromboembolism and acetabular fracture was examined, dominantly posterior localized fracture types according to Judet-Letournel were associated with higher TE rates. Similar results were reported by Sen et al. [11] with higher TE numbers related to increasing fracture instability and complexity of pelvic and acetabular fractures. Our results are in line with this literature showing higher thromboembolic risk with increasing injury severity of pelvic and acetabular fractures (according to AISPelvis and AISAcetabulum) as well as fracture instability of pelvic injuries (according to Tile/AO/OTA-classification). As shown, especially pelvic fractures with AISPelvis ≥ 4/Tile C as well as severe acetabular fractures with AISAcetabulum = 3 represent a group of fractures in patients with mainly high-energy injury mechanism, highly unstable fracture situation, high percentage of hemorrhages, high numbers of required operative procedures and prolonged time of immobilization. Multiple injured patients with high-energy injury mechanism, injuries to predominantly posterior pelvic and acetabular structures, relevant blood loss and the need for repetitive operative treatment are at high risk for the development of thromboembolic complications.

Interestingly, in our multivariate logistic regression analysis injury severity of pelvic or acetabular fracture was no independent risk factor for TE complication. Our explanation for this finding is that acetabular/pelvic injury increases TE risk as shown above when our multiple injury patient collective with mandatory pelvic/acetabular fractures has been compared to published research with overall multiple injured patient studies. In multiple injured patients with acetabular or pelvic injuries the severities of these injuries seem to have no further impact on thromboembolic risk. Importantly, in this context, more relevance seems to lay on risk factors like hemorrhage, mass transfusion, sepsis, abdominal injury etc. that have an imminent influence to exaggerate TE development in these patients. Further research might suggest adapted thrombosis prophylaxis or the establishment of screening examinations in this group of patients. This is somewhat supported by Lowe et al. who found relevant incidences of venous thromboembolism in pelvic and lower extremity trauma despite adherence to modern venous thromboembolism prophylaxis protocols [37].

The probability of developing a thromboembolic event increases with the number of operative procedures among trauma patients. The present study suggests an increased risk for the development of a TE complication among multiple injured patients with pelvic and acetabular fractures with increasing numbers of surgical procedures performed in the course of treatment. When ≥ 10 operations were required, patients were under significantly higher risk to suffer from thromboembolic complications. Pathophysiological mechanisms that drive prothrombotic factors in patients with multiple surgeries remain unclear. Evidence exists in trauma patients that, on a molecular level, a procoagulatory metabolic state (acute traumatic coagulopathy) caused by the initial trauma is intensified by repetitive surgical trauma [13, 38,39,40]. Furthermore, patients requiring multiple operative procedures might have longer times of immobilization, mechanical ventilation and lCU-length of stay that are also associated with TE development [4, 5, 22, 24, 32].

Thromboembolic events among multiple injured patients are a devastating complication during recovery. When compared to Non-TE patients, patients suffering from thromboembolic complications had significantly higher mortality and morbidity in our investigation. Besides the acute injury, complications during the course of treatment promote the development of thromboembolism. This is especially true for septic complications. Our data suggest sepsis to be a major prothrombotic factor in polytraumatized patients with 26.5% compared to 8.7% of patients suffering from sepsis in the TE-group and Non-TE-group, respectively. Multivariate regression analysis revealed sepsis as an independent risk factor for development of thromboembolism. In a study by Paffrath et al. [5], presence of sepsis almost tripled the chance to suffer from TE complications. Other authors report similar results identifying septic complications as key factor for thromboembolic events [41, 42].

The current study has several limitations. Clinically inapparent thromboembolism and thromboembolism diagnosed after hospital discharge were not part of the documentation of the TraumaRegister DGU®. Therefore documented incidences represent only clinically relevant TE complications. Supposedly, the incidence of thromboembolic events in our study population might be underestimated. Furthermore, no information can be given regarding advantages or disadvantages of different mechanism or substances of thromboprophylaxis as only the presence or absence of mechanical/pharmacological prophylaxis is recorded. This is why we are not able to comment on the type or substance administered for pharmacological prophylaxis or the mechanical device used. Additionally, we cannot report on efforts made to evaluate coagulability (e.g. thrombelastography) or to ascertain thromboprophylaxis (e.g. anti-Xa monitoring), since these parameters were not documented in the TR-DGU. On the other hand, measuring anti-Xa or thrombelastography for thromboembolic risk estimation is not part of the routine [20]. Types of acetabular fractures are only differentiated in either “none”, “closed” (AISAcetabulum = 2) or “open” (AISAcetabulum = 3). The high variability of fracture morphology as described in the classification of acetabular fractures by Judet and Letournel is therefore not documented in the TraumaRegister DGU® so that our reported results concerning acetabular fractures have to be interpreted accordingly.