A retrospective multi‐centre study of splenic volumes measured by CT following splenic artery angioembolisation for high‐grade blunt splenic injuries in adults

Splenic artery embolisation is a recognised modality in the management of high grade blunt splenic injury. The impact of embolisation on the spleen in terms of volume and function remains unclear. This results in a lack of clarity regarding post embolisation vaccination policy.


Introduction
This study attempts to provide further data to the literature on the effects of splenic angioembolisation on spleen volume and spleen function in the setting of high grade blunt splenic trauma in Australia. Management of high grade blunt splenic trauma has undergone an evolution over the past few decades; traditional treatment was either operative, resulting in a splenectomy, or conservative. Total splenectomy almost always results in absolute loss of splenic volume and function, placing the patient at risk of a life-threatening condition known as overwhelming post splenectomy infection (OPSI). Salvage of the spleen following high grade blunt splenic injury using angioembolisation or vigilant observation in hospital maintains a degree of functional splenic tissue, thus reducing the risk of OPSI. Splenic artery embolisation has emerged as the mainstay of splenic salvage in high grade blunt splenic trauma. 1,2 There is ongoing debate regarding the relationships between anatomical positioning of coils, splenic volumes, and resulting immune function. A previous small cohort study demonstrated a decreased splenic volume with distal embolisation and increased volume with proximal embolisation but has not commented on further associations with American Association of the Surgery of Trauma (AAST) grading and splenic function. 3 More recent data have shown no significant change in spleen volume over a median follow up of 1.9 years for the trauma cohort. 4 Longer term studies, out to 8 years, with predominantly proximal embolisation, have demonstrated preservation of immune function post splenic artery embolisation for trauma, with a corresponding decrease in volume, as measured by ultrasound. 5 Post embolisation volume may be of use for ongoing surveillance as while volume correlates poorly with splenic function, total splenectomy results in absolute loss of splenic function. This suggests that there is a threshold of volume loss, currently unknown, that impacts function. There are multiple tests for splenic function; however, the current recommendation of the Spleen Australia clinical guidelines, updated October 2021, is initial assessment of altered red cell morphology by screening for Howell-Jolly Bodies after at least 2 weeks post embolisation. 6 Other various tests described in the literature are carried out at 6 months and concentrate on immune function of the spleen. 7,8 None of these tests have been established as a clear surrogate marker of splenic function, and in the absence of this, the presence of Howell-Jolly Bodies has been put forward as the best alternative as a screen for splenic dysfunction. 9 For the purpose of this study, red cell morphology on peripheral smear was selected, as not only is this economical but also can be reproduced at various sites, and complies with the Spleen Australia guidelines.
The aims of the study were as follows: 1. Evaluate if embolisation of the splenic artery in the setting of high grade blunt splenic injury results in changes in six-month post-procedural splenic volumes. 2. Evaluate the impact of the initial injury as measured by AAST, the location of embolisation and presence of altered red cell morphology on splenic volumes, including evaluating any association between AAST and location of embolisation. 3. Assess for the presence of altered red cell morphology following embolisation in the setting of high grade blunt splenic injury with regard to AAST grading and location of embolisation.

Design and setting
A retrospective observational study was conducted across two tertiary referral trauma centres over a period of 24 months (January 2018-December 2019), following the implementation of a pathway for the management of blunt splenic injuries. 10 Both trauma centres have >800 major trauma admissions (Injury Severity Scale (ISS) >12), annually.

Participants
Participants were recruited to the study by way of informed consent. These were patients suffering from blunt traumatic injury to the abdomen, with CT confirmed splenic injuries of grade III or above according to the AAST guidelines, who had been resuscitated to haemodynamic stability. Patients under the age of 15 years and those with penetrating injuries were excluded. Once enrolled, participants followed the clinical pathway for the management of blunt splenic injuries as outlined in the Supporting Information. 10 Computed tomography (CT) All patients underwent a preliminary multiphase contrastenhanced CT with a 160-slice scanner. Industry standard protocols were used with vendor-specific doses and algorithms. The scans were reported by the on-call radiologist and graded as per AAST standards. Patients with contrast extravasation or vascular injury progressed to angiography. Extravasation was defined as contrast accumulation outside of a vessel on either the arterial or portal venous phase CT imaging. Vascular injury was defined as any evidence of a traumatic injury to a vessel, including pseudoaneurysm, AV fistula and abrupt vessel tapering. Patients within the inclusion criteria, but without contrast extravasation or evidence of vascular injury on initial CT, did not undergo immediate embolisation. These patients were admitted for observation with follow up inpatient surveillance scanning. CT imaging was conducted at 48 h for AAST IV and V injuries and 72 h for AAST III injuries. This utilised the same scanner as admission to allow for direct comparison.

Angiography
Formal angiography or digital subtraction angiography (DSA) was performed with one of two single plane Philips AluraClarityâ floor mounted units. Embolisation coils were selected according to the anatomy of the vessel and operator preference. Proximal splenic artery embolisation was defined as any embolisation of the artery proximal to the bifurcation of the splenic artery. Initial attempt was always for selective/distal embolisation. This was the preference due to departmental experience of altered red cell morphology following proximal splenic artery embolisation. However, if this was unsuccessful or deemed not appropriate by the operator, non-selective/ proximal embolisation was utilised. Cook, Hilal, Tornado and Nester coils, as well as Balt Spirale coils and Concerto detachable coils, were deployed. In proximal embolisation cases, a microvascular plug (MVP) (Medtronicâ) was sometimes deployed before the coil depending on the operator's preference. Microvascular plugs or any form of temporary haemostatic measures were never utilised in isolation.

Splenic volume calculation
Splenic volumetric calculations were obtained using the 1 mm data sets. Calculations were conducted on the initial imaging, 48 h imaging, and the 6 month follow up scans. All calculations were performed by the same personnel with the use of Philips IntelliSpace Portal 8.0 software. Volumetric calculations were obtained by tracing the viable enhancing splenic architecture, excluding nonenhancing lacerations, subcapsular hematomas and infarction. Three-dimensional volumetric representation of the splenic architecture was saved with associated volume calculations. Example in Figure 1.

Assessment of altered red cell morphology
Altered red cell morphology was determined by assessment of the presence of Howell-Jolley bodies on a blood smear by a certified Pathologist. Day 14 was chosen as per the Spleen Australia online guidelines. This was reassessed on a repeat blood smear at 6 months. Patients with evidence of altered red cell morphology at Day 14 were vaccinated as per the Spleen Australia guidelines.

Data analysis
Statistical analyses were performed using IBM SPSS Statistics [Version 24; SPSS Inc., Armonk, NY, USA, IBM Corp]. Statistical significance was set at a P value of equal or less than 0.05. Standard descriptive statistics were generated. Data normality was confirmed with a Shapiro-Wilk Test. Comparison of pre-and post embolisation volumes was performed using a T-test. Volume subgroup analysis was performed with separate-one way ANOVA testing comparing splenic volumes (preembolisation and percentage change) with location of embolisation (proximal, distal or combined) and grading of initial injury as defined by AAST grading (III, IV or V). Analysis of initial volume by altered red cell morphology was conducted using a T-test. Assessment of association between location and embolisation, as well as comparing the presence of altered red cell morphology with both AAST grading, and location of embolisation were performed using Fisher-Freeman-Halton exact tests.

Results
Results are summarised in Table 1 and Figure 2. Vascular extravasation was identified on initial CT in 30 (91%) who all progressed to angiography and angioembolisation. Of these patients, 3 (10%) who underwent Twenty of the 30 patients who underwent successful splenic artery embolisation were further investigated at 6 months with pre-and post-contrast enhanced computed tomography scans. Ten patients were lost to follow up. Their initial volume data were included in the analysis. The data demonstrated a statistically significant mean volume change from 308.05 to 263.91 cm 3 with a mean volume loss of 44.14 cm 3 (P = 0.038). There was a mean volume increase of 8.91% for the 6 patients undergoing proximal (non-selective) splenic artery embolisation, a mean volume decrease of 9.49% for the 21 patients undergoing distal splenic artery embolisation, and a 58.15% decrease for patients undergoing combined proximal and distal splenic artery embolisation. Results of the subgroup analyses are displayed in Table 2. No statistical significance was found with regard to analysis of the location of embolisation with initial volume (P = 0.466) or percentage volume change (P = 0.170).

Discussion
This study aimed to assess for changes in splenic volume between an initial CT and 6 months post intervention. It also aimed to assess the impact of location of embolisation and initial AAST grade on both splenic volumes and on altered red cell morphology post embolisation. There was a statistically significant change in splenic volumes at 6 months following embolisation for blunt splenic trauma (P = 0.038). With respect to the volume change, coil location within the splenic artery was associated with positive volume gains if coiled proximally and volume loss if coiled distally. Whilst this was not statistically significant with respect to initial volume (P = 0.466) or percentage volume change (P = 0.170), this is consistent with splenic volumes published in the literature following proximal and distal splenic artery embolisation. 3 Other studies have shown a mixed picture. One study had no significant change in splenic volume with a mean follow up of 1.9 years for their trauma cohort. 4 Another followed a cohort of 17 patients out to 7.7 years and showed preserved spleen function as measured by IgM memory B Cells, in spite of an overall reduction in splenic volume. 11 Both these cohorts predominantly underwent proximal embolisation. 4,11 This study differed with 73.3% of patients undergoing only distal embolisation. A further difference was that the clinical pathway utilised in this study allows for observation of high-grade splenic injuries, 10 whereas other published studies have relied on protocols based on routine angioembolisation. 12 This inherently results in a cohort that differs from previous publications and allows this data to add to the literature.
In considering the outcomes of splenic volume and function following splenic artery embolisation in trauma, it is important to ascertain the impact of both the initial trauma as well as the angioembolisation procedure. The data demonstrated that the AAST grading was significantly associated with the initial splenic volume (P = 0.004), percentage volume change (P = <0.001) and with the presence of altered red cell morphology (P = 0.039). It is noted that AAST grading can be considered as a function of perfused spleen, and therefore, the method of calculating volume of enhancing spleen would be expected to have a high concordance with the AAST grading. This does not account for the association between AAST and altered red cell morphology. There was a statistical association between altered red cell morphology and the initial splenic volume (P = 0.044). This supports the hypothesis that at least one of the drivers of altered red cell morphology is the initial trauma. Qualitative splenic function was assessed by examining for the presence of altered red cell morphology. Quantitative tests for splenic function, such as IgM Memory B cells, are gaining research validation, but they remain expensive. Qualitative assessment of altered red cell morphology is in the current guidelines from Spleen Australia following splenic artery embolisation, and on this basis was utilised for this study. 13 Two of the three cases with altered red cell morphology had no evidence of further altered red cell morphology at the 6 month assessment, with the third case being lost to follow up. This occurred despite volume loss from the initial trauma scan. As no patient who attended follow up had altered red cell morphology, no analysis could be conducted on the change in volume of patients with persisting altered red cell morphology. This suggests a global transient insult to the spleen. It is unclear whether this is from the trauma, the angioembolisation or a combination.
Currently, there is a lack of consensus regarding the role of vaccination following splenic artery embolisation. 14 The current Spleen Australia guidelines recommend vaccination post splenic artery embolisation. 13 The results raise the question as to whether blanket policies for vaccination and/or antimicrobial prophylaxis, triggered by the detection of altered red cell morphology in the first few months post splenic artery embolisation in trauma, are overtreating. This has been further supported by recent publications. 5,15 Whilst the decrease in volume with distal embolisation was statistically significant, the decrease in volume was not associated with altered red cell morphology. There were no cases of distal embolisation resulting in altered red cell morphology. Of our cohort, the three patients who had altered red cell morphology on their peripheral blood smear, all had coils placed proximally, with one having both proximal and distal coiling. All patients had high grade injuries with two AAST grade IV and one AAST V splenic injury. Notably all patients with distal embolisation had no evidence of altered red cell morphology. There was a significant association between altered red cell morphology and both the location of embolisation (P = 0.019) and AAST grading (P = 0.039). There was no significant association between the location of embolisation and AAST grading (P = 0.077). We concluded that, in our cohort, injuries requiring proximal splenic artery embolisation are more likely to result in altered red cell morphology with preserved splenic volume. One of the major limitations of this study was the sample size. This is a common issue for splenic artery angioembolisation studies, which typically have low sample size. This resulted in a smaller number of proximal and combined embolisation cases. This may have resulted in a bias in the data. However, 87.5% of proximal and combined cases completed the 6-month follow up. Small sample size can be combatted by methods such as utilising a multi-centre approach or a longitudinal approach. Each of these solutions has its own limitations. A multi-centre approach increases the heterogeneity at multiple levels: the patient demographics will change; and different cohorts of interventionalists broaden the range of approaches taken. Similarly, whilst a longitudinal approach increases sample numbers, progression at an institution, as well as progression in techniques and equipment, provides additional confounding variables. This study attempted to minimise these aspects by using two similar sites and recruiting over a 2-year period.
Another limitation to the study was the loss to follow up. This is one of the issues of a large trauma catchment area, where patients are often brought from other geographical locations to the Level 1 trauma centre. This is an inherent issue for trauma studies and is difficult to control. Careful engagement with local medical practitioners and increased telehealth services may help to combat this.
In conclusion, this retrospective observational study demonstrated multiple key factors associated with preprocedural and 6-month post-procedural splenic volumes following splenic artery angioembolisation after blunt splenic trauma. There was a statistically significant decrease in splenic volume post embolisation. Further sub analyses demonstrated significant associations for the impact of AAST grades on initial volume and volume change. Altered red cell morphology was significantly associated with AAST grading, initial splenic volume, and location of embolisation. It is proposed that the lack of significant association between the location of embolisation and AAST indicates that both are independent factors with respect to the presence of altered red cell morphology. Notably, in the cases who attended follow up, altered red morphology detected at 2 weeks was transient and had resolved at 6 months. There was no statistically significant association between the choice of embolisation location and splenic volume changes. Overall average splenic volume decreases with distal splenic artery angioembolisation for trauma, which is consistent with previously published data.
With respect to splenic function, only 10% of cases had evidence of splenic dysfunction with the presence of altered red cell morphology, and this was transient two thirds of the time. This does not support blanket vaccination policies for patients undergoing splenic artery embolisation following blunt splenic trauma. Further research is required to ascertain the most appropriate patient population to treat with respect to vaccination and antimicrobial prophylaxis.