Introduction

The majority of blunt splenic injuries are managed nonoperatively [1]. Patients who undergo splenectomy are at risk for a variety of postoperative complications, including pancreatitis, pancreatic fistula, thromboses, gastric wall necrosis, typical postoperative infections, and overwhelming post-splenectomy infection (OPSI) [2]. Nonoperative management fails in 4–14% of cases [1], especially in higher Grade injuries or those with contrast extravasation noted on imaging [3]. Angioembolization has emerged as an adjunct with the potential to increase the success rate of nonoperative management [4].

Angioembolization usually involves the main splenic artery (proximal embolization) and occasionally segmental splenic artery branches (distal embolization) [5]. Occlusion of the vasculature in either manner frequently results in areas of atrophy or infarction, and therefore, a decrease in functional splenic volume.

Post-splenectomy patients are known to be at increased risk for venous thromboembolism (VTE), and this is likely multifactorial [6, 7]. A reactive thrombocytosis frequently occurs post-splenectomy, which leads to platelet hyperaggregation, and likely contributes to a systemic hypercoagulable state [8, 9]. Hypercoagulability post-splenectomy has also been attributed to causes unrelated to platelet number or function, including increased plasma levels of fibrinogen, d-dimer, and other prothrombotic factors observed in post-splenectomy patients [7]. Similar to post-splenectomy, a reactive thrombocytosis has been demonstrated after angioembolization of the splenic vessels [10]. It is therefore plausible that some of the coagulation abnormalities complicating splenectomy may also occur after angioembolization of the spleen.

Trauma patients managed with splenectomy are at increased risk for VTE compared to patients managed nonoperatively due to their post-splenectomy hypercoagulable state, however, the potential for an analogous increased risk after angioembolization of splenic injuries has not been studied. The purpose of this study was to determine if patients who undergo angioembolization as an adjunct to nonoperative management of splenic injury are at increased risk for VTE in comparison to those who do not. In order to eliminate many variables making comparisons difficult, the study population was restricted to isolated blunt, Grades III-V splenic injuries.

Methods

A retrospective analysis was performed of the Trauma Quality Improvement Database (TQIP) from 2013 to 2016. All adult (>16 years) patients with significant blunt splenic trauma (AIS 3, 4, or 5) were queried (Fig. 1). Patients with other significant injuries (AIS 4 or 5), including head, chest, spine, upper and lower extremities, liver, kidney, or pancreas or with any untreatable injury (AIS 6) were excluded. Patients with no signs of life on arrival, transferred from another hospital, or who died within 72 h were excluded. In addition, those with missing information, such as age, ICU or hospital length of stay (LOS), procedures performed, hospital disposition, or type of VTE prophylaxis, were also excluded. Institutional Review Board approval was granted from the University of Southern California for this study.

Fig. 1
figure 1

Flowchart

Full size image

Those patients who were managed nonoperatively and received pharmacological thromboprophylaxis were then divided into two groups: those that underwent angioembolization and those that did not. Univariate analysis was performed to compare the two groups. Medians for continuous data were compared with the Mann–Whitney U test. Comparison between percentages of categorical variables was performed with the Fisher Exact or Pearson's chi-squared test. Primary outcomes for the study were the incidence of deep venous thrombosis (DVT), pulmonary embolism (PE), or any VTE. Secondary outcomes included mortality, ventilation days, ICU LOS, and hospital LOS.

Logistic regression was then performed with potentially associated variables in which p was less than 0.2 to identify the independent risk factors for the study outcomes. Subgroup analysis was also performed of only patients who were treated with unfractionated heparin (UH) for VTE prophylaxis, and of only patients who were treated with low molecular weight heparin (LMWH) for VTE prophylaxis. Additional subgroup analyses were performed of patients with AIS 3 splenic injuries and patients with AIS 4 or 5 splenic injuries. Correlation between variables was tested with multicollinearity analysis. The area under the ROC curve with 95% confidence interval was used to assess the accuracy of the test. Statistical significance was set as p<0.05. All statistical analysis was performed using SPSS for windows version 23.0 (SPSS Inc. Chicago, IL).

Results

A total of 2643 patients with blunt splenic injury managed nonoperatively were reviewed. Patient demographics are displayed in Table 1. The majority of patients were male (65.1%) with a median age of 37 years. Patients who underwent angioembolization as an adjunct to nonoperative management were slightly older (43 years) than those who did not (36 years) and were twice as likely to present with signs of shock (HR>120 or SBP<90). They were also more likely (8.3%) to present with GCS<9 (5.4%). Comorbidities were similar among the two groups, with the exception of congestive heart failure, which was more common in the group that did not undergo angioembolization. The time to angioembolization was 3.49 h (2.6–4.9).

Table 1 Patient demographics
Full size table

On logistic regression, angioembolization was found to be an independent risk factor for both DVT (p<0.006) or any VTE event (p<0.006), with patients who underwent angioembolization being more than twice as likely to develop these complications (Table 2). Administration of LMWH for prophylaxis was relatively protective for DVT (p = 0.05) or VTE event (p = 0.04) in comparison to UH (Table 2).

Table 2 Potentially associated variables (a) VTE, (b) DVT, (c) PE (logistic regression)
Full size table

As anticipated, patients with Grade III splenic injuries made up a much larger proportion of the population managed without angioembolization (71.6%) than the population managed with the intervention (47.6%). For this reason, subgroup analyses were performed on patients with Grade III injuries and on patients with Grade IV or V injuries. Angioembolization remained an independent risk factor for DVT (p = 0.004) in the Grade IV–V injury group, and for VTE (p<0.01) in the Grade III injury group.

Overall, prophylaxis with LMWH (79%) was significantly more common than prophylaxis with UH (21%). A greater number of patients in the angioembolization group were given UH for pharmacological prophylaxis (26%) than the group without angioembolization (20.4%), therefore subgroup analyses were performed on patients who received UH and on patients who received LMWH. Angioembolization remained an independent risk factor for DVT (p = 0.001) among patients who received UH, and for both DVT (p = 0.04) and VTE (p = 0.04) among patients who received LMWH.

There was no significant difference between the angioembolization and no angioembolization groups with regard to the timing of administration of VTE prophylaxis. Half of the patients (48.5%) received their first dose of thromboprophylaxis within 48 h of admission to the hospital, and 51.5% received their initial dose after 48 h. On logistic regression, initiation of thromboprophylaxis greater than 48 h after admission was associated with increased VTE rates in comparison to early initiation (OR 1.75, p = 0.02).

Discussion

Nonoperative management of hemodynamically stable patients with splenic trauma has been the standard of care for several decades [11]. Angioembolization as an adjunct to nonoperative management began in the 1980s [12], and its use has increased over time [1]. Recent literature has reported complication rates of 14–29% after angioembolization, including bleeding/vascular injuries, fever, pleural effusion, ARDS, pancreatitis, contrast-induced acute kidney injury, coil migration, and infarction/necrosis/atrophy or abscess of the spleen [13, 14]. DVT, PE, and VTE rates after angioembolization have not been studied.

The role of angioembolization has not been precisely defined. Guidelines of The Eastern Association for the Surgery of Trauma recommend considering angioembolization for American Association for the Surgery of Trauma (AAST) injuries Grade III or greater, presence of a contrast blush, moderate hemoperitoneum, or evidence of ongoing splenic bleeding [15]. However, there is debate concerning the risk-benefit ratio of angioembolization in different cases. Some centers advocate for routine angioembolization of all high-Grade splenic injuries, even in the absence of other risk factors. Bhullar and colleagues performed a large retrospective study of over a thousand high Grade blunt injuries. They found a 0% failure rate of nonoperative management in the patients with Grade IV or V injury who underwent angioembolization despite the absence of contrast blush on CT scan, compared to 26% failure rate in those that did not. They concluded that angioembolization should be performed in hemodynamically stable adults after Grade IV or V splenic injury routinely [16]. Conversely, Chastang and colleagues performed a multicenter randomized trial of 91 patients who underwent nonoperative management with or without angioembolization, and concluded that angioembolization should not be routinely performed, due to increased morbidity [17]. A survey of AAST members found considerable variation in opinions regarding routine angioembolization of high-Grade injuries [18]. A recent meta-analysis evaluated angioembolization as an adjunct to nonoperative management and concluded that it has improved the success of nonoperative management for AAST Grade IV and V injuries, however it has demonstrated no benefit for Grades I, II, and III [19]. Furthermore, the authors concluded that morbidity is higher in patients who undergo angioembolization in comparison to those who are managed strictly nonoperatively [19]. Proper selection of patients for angioembolization as an adjunct to nonoperative management therefore remains an important area of clinical research.

This is the first study to demonstrate an increased risk for VTE associated with splenic angioembolization. Though increased risk has been previously established post-splenectomy for trauma, the use of angioembolization as an adjunct is a relatively new treatment option. In 2016, Wernick and colleagues evaluated hematologic parameters post-splenic trauma in patients who underwent splenectomy, splenic angioembolization, or observation. They confirmed a thrombocytosis in patients post-splenic angioembolization comparable to patients who underwent splenectomy, and in contrast to normal platelet values in those managed with observation alone [20]. This suggests that some of the same physiologic processes that drive hypercoagulability post-splenectomy may be significant after angioembolization.

This study evaluated a large population of patients with isolated blunt splenic injury managed nonoperatively and found a statistically significant increase in DVT and VTE rates associated with splenic angioembolization. This difference remained evident in subgroup analyses of Grades III, IV, or V injury. The difference also remained true despite use of UH or LMWH for VTE prophylaxis. Increased rates of PE were not demonstrated; however, the incidence was very low in the population overall.

The most important limitation of this study is its retrospective nature, and the inherent potential for selection bias. However, the TQIP database is a high-quality data repository maintained by the American College of Surgeons (ACS). Due to the importance of VTE in trauma populations, the ACS Committee on Trauma added specific variables to the TQIP database to evaluate the timing and type of VTE prophylaxis administered to every trauma patient. Inclusion of these variables has resulted in mandatory attention to VTE by all contributing centers with both data entry and receipt of benchmarked data. Use of a national database to study VTE is important because though it is much more common in trauma patients, VTE is still a relatively rare occurrence overall. In order to eliminate potential confounders, trauma patients with relatively isolated injuries must be considered to evaluate risk factors for VTE. Isolated injuries present rarely among a diverse, polytrauma population, which practically precludes prospective study of VTE contributors.

Patients who underwent splenic angioembolization tended to be older, less hemodynamically stable, and with higher AAST Grade injuries. A higher percentage of patients who underwent angioembolization received UH for initial VTE prophylaxis (as opposed to LMWH). Logistic regression was used to compare the two groups, and subgroup analyses were performed of individual AAST Grades, and of patients who initially received UH and those who initially received LMWH. These methods were employed to eliminate potential confounders. Splenic angioembolization was identified as an independent risk factor associated with VTE, however, a causal relationship is not able to be established retrospectively. A well designed, prospective trial is needed to confirm that splenic angioembolization confers an increased risk for VTE.

Major complications following splenic artery angioembolization are not significantly different if a proximal versus distal artery technique is utilized [5]. However, distal splenic artery angioembolization has been associated with a higher rate of splenic infarction not requiring splenectomy. In addition, proximal embolization is a shorter duration procedure. Due to these differences, it is possible that the two techniques are not associated with an equivalent risk of VTE. The TQIP database does not include the specific technique, so this study was not able to evaluate a potential difference between techniques with regard to incidence of VTE. However, in the majority of trauma cases, the proximal technique is utilized, so even if the technique had been reported in the TQIP database, the number of distal angioembolizations would likely have been too low to identify a statistically significant difference in VTE rate. Similarly, angioembolization method and materials vary across centers, and this could not be evaluated using the TQIP database. Variations of VTE rate with a specific technique may be an area for future study.

The results of this study underscore the importance of maintaining a high level of suspicion for VTE after blunt splenic injury, as well as the necessity for early initiation of pharmacologic VTE prophylaxis. This may include surveillance for VTE, when appropriate. Early VTE prophylaxis (within 48 h) is safe after blunt splenic injury and decreases rates of VTE [21, 22]. In this analysis, early prophylaxis was identified as a protective factor against the development of VTE (p = 0.02). Consistent with previous studies, this study found prophylaxis with LMWH to be more protective than UH for the development of both DVT (p = 0.05) and VTE (p = 0.04) [23].

Angioembolization is an important adjunct for splenic preservation after trauma. It is imperative that clinicians understand the potential risks for VTE associated with the procedure and promptly institute appropriate pharmacologic prophylaxis.