The number of splenic embolizations has been steadily increasing in frequency at our institution, and similarly reflected in data from other institutions including from different countries (Roy et al. 2018). There are many factors that may contribute to this. From the perspective of our network, our hospital is the largest Level 1 center in the Victorian State Trauma system which has preprogrammed responses built in to patient referral and treatment processes and is supported by efficient ground and air ambulance including a 24-h helipad. This has decreased mortality and also time to definitive hemorrhage control (Cameron et al. 2008). On arrival to the trauma center, a multidisciplinary trauma team approach to damage control resuscitation has decreased the time to stabilization, access to CT, and definitive hemorrhage control. This may allow for endovascular rather than emergency operative intervention (Matsumoto et al. 2015). In addition, a dedicated interventional radiology on-call roster allows for rapid activation of services. These factors support a median time to treatment of 0 days. The increasing trend also reflects the increasing abundance of supportive data towards the role of SAE in supporting NOM in blunt trauma including recent systematic reviews (Moore et al. 1989; Schnüriger et al. 2011).
The demographic in this study with a mean age of 40 and predominantly males, is in keeping with the expected demographic of road traffic trauma (Ferrah et al. 2019). The median injury severity score of 22 also reflects the complexity of high-energy multi-system trauma at our institution (Ferrah et al. 2019). The demographics in both the overall cohort and for those who experienced a complication were similarly matched.
The overall rate of complications in this study was low at 5.6% and occurred at a median time of 2.1 days after the treatment. In addition, only 3% of patients proceeded to splenectomy at a median time of 4 days which is low, and in keeping with the rate shown in prior literature (Moore et al. 1989; Schnüriger et al. 2011; Davies and Wells 2019; Hughes et al. 2017). Of the 7 patients requiring splenectomy after embolization, 1 patient was a 30-year-old male who presented after a motor vehicle accident with multiple injuries (ISS 38) including grade V splenic injury, hollow viscus injury, and pelvic bleeding on CT scan. The patient was emergently transferred to the angiography suite before the operating theatre, where pelvic embolization was performed, and a decision was made between the interventional radiologist and trauma surgeon to embolize the spleen proximally to facilitate safe transfer to the operating suite where splenectomy was performed as a pre-planned procedure. While this has been included in the analysis as a splenectomy after embolization, in the opinion of the authors it does not reflect a complication of the embolization itself.
With the update to the AAST classification in 2018 to incorporate splenic artery vascular injury, 82% are now considered high-grade (AAST IV or V). This compares to 126 high-grade injuries (54%) if our patients were graded per the previous 1994 AAST classification. The grade V cohort showed a significantly higher rate of vascular injury than for the lower grade groups which is expected given the severity of injury to meet the grade V criteria. The rate of splenectomy in the grade V cohort of 8.5% is also low and comparable to literature including the systematic review of Rong et al. which included 10 studies and 876 patients (Moore et al. 1989). It is encouraging to see that of all patients with grade IV splenic injury or lower, only 2 of 150 patients (1.3%) required splenectomy after SAE.
Our practice to strongly consider angiography in patients who have AAST grade III injury and three or more quadrant hemoperitoneum (41 patients, 17.6%) is acknowledged to be a controversial decision. However, as shown in this study there was a significant increase in the number of vascular injuries identified at angiography compared to those identified at CT in patients who were deemed to have AAST III injury based on their trauma CT scan. It is for this reason that considering angiography in the grade III cohort is likely to identify patients who are at increased risk of re-bleed. The absolute risk however is difficult to quantify without a prospective and randomized design, which may not be feasible. The effect of these results is to further sub stratify this patient group beyond the AAST classification where it is possible that occult vascular injuries are the reason for those who do subsequently re-bleed with a grade III injury. In addition, it can be argued that the low rate of complications in the AAST III cohort of 0% and preservation of splenic function (Lukies et al. 2020; Schimmer et al. 2016) further warrants such consideration for treatment. The presence of vascular injury in the grade III cohort from this study also supports the recent changes to AAST to acknowledge the importance of vascular injury in their classification system. In addition, with a median ISS of 22 and as high as 59 in the grade III cohort, a diminished risk of re-bleed after embolization is often appreciated in complex multi-trauma patients where the cause of abnormal or changing physiology can be hard to pinpoint.
In the whole cohort, the choice of proximal embolic location (75.9%) and pushable fibred coils (84.9%) reflects a decision at our institution to provide a cost-effective solution for our patients in an institution that is government-funded (Yip et al. n.d.). In many cases, proximal embolization also offers a rapid option to provide treatment even with significant anatomic tortuosity, anatomic variation, vasospasm, and/or underlying vascular disease which may inhibit or prolong attempts at distal embolization. Such choices are at the discretion of the treating interventional radiologist at the time of embolization even though a range of catheters and embolic materials are always in stock.
It is encouraging to see that there is no significant difference in the rate of complication or splenectomy based on the choice of embolic material or the embolic location in this study. The analysis of Rong et al. suggested a higher rate of complications with the use of gelatin sponge and this has also been shown in other studies (Moore et al. 1989; Abada and Golzarian 2007). However, as the number of cases using gelatin sponge in our study was extremely low, it is not possible to correlate with their findings reliably. While it is unusual to see zero complications in the distal embolization cohort, the small number of patients in this cohort (35, 15.1%) which was heavily biased towards proximal embolization may account for this.
The authors acknowledge a number of limitations with this study. This is a retrospective audit and as such lacks patient-level clinical data in which the trauma setting it is arguable most vital. As such, we have used AAST and ISS to risk-stratify patients. As a single-center study, our practice reflects the internal structure of our trauma and interventional radiology teams and is also impacted by the nearby position of emergency/trauma and imaging services. This may not necessarily reflect how other hospitals are structurally designed. We also acknowledge that AAST grade was determined at initial CT and while treatment was performed based on a combination of CT and angiographic findings, the grade was not changed after angiography for the purposes of this analysis. We acknowledge the heterogeneity in the way embolizations were performed in our cohort and that there has been grouping of individual patients regardless of the fine detail of splenic parenchymal injury variation. The choice of embolic and choice of embolic location is also heavily skewed to proximal embolization with pushable coils, leaving smaller numbers for the other groups. This will reduce the statistical reliability of analysis. In addition, all embolics are different even for pushable coils and the effectiveness between 0.018″, 0.035″, and different types or lengths of coils in each size was not measured. Nevertheless, our results indicate that proximal splenic embolization is a robust strategy that can be successful despite technical differences and this is supported in previous systematic reviews which also suffer from similar heterogeneity in data (Moore et al. 1989; Schnüriger et al. 2011).
In conclusion, the SPLEEN-IN study shows that treatment of intermediate-high grade blunt force traumatic splenic injuries using SAE results in a low overall rate of complication and splenic salvage in 97% of patients, providing a safe and effective adjunct to NOM in these patients. Splenic salvage after SAE increases to 98.7% of patients with a grade IV injury or lower. These results can be achieved with lower cost pushable coils and proximal splenic embolization in most patients.
The SPLEEN-IN study also shows that embolization of grade III injuries in certain patients is safe, and that proceeding to angiography in this cohort can identify vascular injuries initially occult on CT. The results support the recent changes to the AAST classification where the importance of splenic vascular injury is now acknowledged.