Long-term follow-up of infection, malignancy, thromboembolism, and all-cause mortality risks after splenic artery embolization for blunt splenic injury: comparison with splenectomy and conservative management

Abstract

Background

Non-operative management, including splenic artery embolization, is preferred for blunt splenic injuries, but its long-term risks need further investigation. Long-term splenic functions were assessed in patients with blunt splenic injuries, hypothesizing that splenic artery embolization would preserve function and reduce long-term risks.

Methods

This retrospective cohort study used Taiwan’s National Health Insurance Research Database to analyse patients with blunt splenic injuries from 2004 to 2019. To balance baseline characteristics across the treatment groups, inverse probability of treatment weighting was used based on propensity scores. Outcomes of main interest included the incidence and cumulative infection, malignancy, thromboembolism, and all-cause mortality risks after one year.

Results

Altogether, 18 771 patients sustained blunt splenic injuries; 8195 were eligible for inclusion. The mean age was 38.4 years, with male predominance (70.5%). Outcomes were compared between splenectomy, splenic artery embolization, and conservative treatment groups. After applying a generalized boosted model with inverse probability of treatment weighting, patients who underwent splenic artery embolizations or conservative treatment had lower infection risks than those in the splenectomy group. The conservative treatment group had a lower malignancy risk than the splenectomy group. Patients with blunt splenic injuries who underwent splenic artery embolizations or conservative treatment had a lower thromboembolism risk than those who underwent splenectomies.

Conclusion

Long-term infection, malignancy, thromboembolism, and all-cause mortality risks were not significantly different between the splenic artery embolization and conservative treatment groups. In contrast, patients who underwent splenectomies had increased infection, malignancy, thromboembolism, and all-cause mortality risks. Clinicians must be familiar with the potential long-term complications associated with the different treatment modalities for splenic injuries and provide appropriate prophylactic measures.

Non-operative management for blunt splenic injuries, including splenic artery embolization, is preferred, but its long-term risks need further investigation. Long-term infection, malignancy, thromboembolism, and all-cause mortality risks were not significantly different between the splenic artery embolization and conservative treatment groups. In contrast, patients who underwent splenectomies had increased infection, malignancy, thromboembolism, and all-cause mortality risks.

Introduction

The spleen is one of the most commonly injured organs in blunt abdominal trauma^1–4. Prompt resuscitation and timely haemostasis are essential to save patients’ lives. For patients with unstable haemodynamics who do not respond to resuscitation and show evidence of intra-abdominal haemorrhage, urgent laparotomy is required^5–7. In such critical cases, splenectomy often becomes unavoidable once the splenic injury is confirmed during exploration^8,9. Conversely, for patients who remain stable after resuscitation, CT can be performed to identify active bleeding, with splenic artery embolization (SAE) serving as an adjunct for achieving haemostasis^5–7. The latter treatment algorithm constitutes non-operative management (NOM) of blunt splenic injury (BSI), which can be attempted in 74–88% of patients^10,11. NOM is currently the primary method for managing BSIs, with a reported 80–100% success rate^10,12–15.

The spleen is divided into two sections: the red pulp and the white pulp. The red pulp is abundant in macrophages and is crucial in removing pathogens, cellular debris, and ageing erythrocytes and platelets. The white pulp is a well-organized lymphoid region that initiates adaptive immune responses, featuring distinct areas for B and T cells, all encircled by the marginal zone^16–20. The marginal zone bridges innate and adaptive immunity and contains macrophages with pattern-recognition receptors that take up blood-borne pathogens. Marginal zone B cells can be activated by these macrophages or directly respond to pathogens, becoming antigen-presenting cells or immunoglobulin M (IgM)-producing plasma cells. Activated dendritic cells or marginal zone B cells enter the white pulp to initiate adaptive immune responses by activating T cells, which then assist B cells^16,17. This compartmental organization allows the spleen to mount complex adaptive immune responses and effectively clear pathogens from the blood.

Immunological and haematological disorders are anticipated following splenectomies. Bacterial infections, particularly caused by encapsulated pathogens, are well-recognized adverse events of splenectomies^21–23. Overwhelming post-splenectomy infection is a notorious potentially fatal condition²⁴. Furthermore, the risk of solid tumours and haematological malignancies is increased in patients who have undergone splenectomies^25,26. Lastly, the incidence of thromboembolic disorders is higher among patients who have undergone splenectomy^25,27–30. However, in the NOM era, the long-term effects of SAE on splenic function have not been thoroughly investigated. A systematic review reported preserved splenic immune function in 11 studies with various parameters³¹. However, evidence from the Nationwide Readmissions Database suggested that patients who underwent SAEs may experience an increased risk of infections and venous thromboembolisms (VTEs) within a one-year follow-up period^32,33. Current studies are limited to laboratory examinations of immunological markers or have insufficient follow-up periods, leading to ongoing controversy.

This study aimed to clarify the long-term effects of SAEs on splenic function in patients with BSIs. It was hypothesized that patients who underwent SAEs would retain preserved splenic function and exhibit comparable long-term infection, malignancy, thromboembolism, and all-cause mortality risks to patients who received conservative treatment.

Method

Data source

Taiwan’s National Health Insurance (NHI) system was launched in 1995. It covers over 99% of the population and 92% of medical institutions. Medical payments cover all types of diseases, including trauma. A peer-review process was used to justify the medical procedures. This study was conducted using the National Health Insurance Research Database (NHIRD), which records insurance reimbursement claims data. The NHIRD and its associated databases are currently administered and maintained by the Health and Welfare Data Science Centre (HWDC) as part of the National Health Informatics Project. More details about the NHIRD and its validations have been described elsewhere^34,35. Moreover, the NHI has a catastrophic illness programme that covers certain severe conditions, including cancer, major trauma with an Injury Severity Score (ISS) ≥16, autoimmune disorders, and haematological diseases. Patients who qualify receive comprehensive coverage with reduced out-of-pocket expenses. This programme requires a rigorous review process to determine patient eligibility and ensure diagnostic accuracy.

Study design

This study was approved by the Institutional Review Board of Chang Gung Memorial Hospital (No.: 202101615B0). The data used in the application span from 2000 to 2019. This retrospective cohort study included patients who sustained BSIs from 2004 to 2019 because the number of SAEs was low prior to 2004 (Fig. 1a). The exclusion criteria were missing demographics (that is sex and age) or age < 18 years. The study also excluded patients with concurrent liver injuries, kidney injuries, or pelvic fractures, as indicated by emergency room or hospitalization diagnoses, owing to a lack of detailed information regarding their transarterial embolization sites. Patients with a history of malignancy, cardiovascular events, immune deficiency, autoimmune disorders, or coagulation defects prior to the injury or within one year after discharge were ineligible. Patients using antiplatelet or anticoagulant medications before the injury or within a year after discharge were excluded. The criteria also excluded trauma-related deaths during the index admission and patients with less than 365 days of post-discharge follow-up. Patients eligible for analysis were categorized into splenectomy, SAE, or conservative treatment groups based on their treatment modalities, as indicated by the Taiwan NHI reimbursement codes. The ICD diagnostic codes from the ninth and tenth editions are summarized in Table S1.

Fig. 1.

Flowchart of the study

a Temporal trends in the treatment of blunt splenic injuries throughout the study period. b Inclusion and exclusion criteria for patient selection. SAE, splenic artery embolization.

Covariates

Basic demographic information, including patient age and sex, was recorded. Associated injuries during the index BSI admission, such as traumatic brain injury, spinal cord injury, injuries to the heart and lungs, pneumothorax and haemothorax, injuries to the gastrointestinal tract, femur fractures, and major trauma with an ISS ≥ 16, were recorded. Patient co-morbidities, including hypertension, ischaemic heart disease, diabetes mellitus, chronic obstructive pulmonary disease, cirrhosis, and chronic kidney disease, were also recorded. Co-morbidities were defined based on at least two outpatient diagnoses or any single inpatient diagnosis within the year preceding the index BSI admission. Additionally, bed confinement status or paralysis during the first year of follow-up after discharge was identified. The ICD codes used for the associated injuries and co-morbidities are provided in Table S1.

Outcomes

The outcomes of main interest in this study included the incidence and cumulative risks of newly diagnosed infections, malignancies, thromboembolism, and all-cause mortality rate beyond one year after the index admission. Target infections included septicaemia, pneumonia, and meningitis. All malignancies listed under the catastrophic illness programme were used as outcomes and further categorized into solid tumours and haematological malignancies. Medical expenses for cancer treatment are exempt if a patient holds a catastrophic illness certificate card for malignancies. The thromboembolic events of interest included acute myocardial infarction, pulmonary embolism, ischaemic stroke, embolism and thrombosis of the extremity arteries, portal vein thrombosis, deep vein thrombosis, and intestinal vascular insufficiency. Infectious and thromboembolic events were detected in the inpatient setting. The events in each category were aggregated into composite outcomes. Information regarding the causes, dates, and places of death can be linked to the Taiwan Death Registry database within the HWDC. The ICD diagnostic codes used for the outcomes are provided in Table S1. In exploratory studies like this, addressing multiplicity is not a primary concern because the aim is to identify potential associations and generate hypotheses rather than confirm prespecified hypotheses. This approach allows for a broader evaluation of outcomes without the stringent need for multiple testing corrections as in confirmatory trials.

The mean follow-up was 8.1 years (until death or 31 December 2019), with only 50% of the patients remaining at risk; therefore, the maximal follow-up period was set to 8 years. Patients were followed-up from the day of discharge until the outcome occurred, death, end of the database period (31 December 2019), or the eighth year, whichever occurred first.

Statistical analyses

To address potential confounding and balance baseline characteristics across the three study groups (splenectomy, SAE, and conservative groups), inverse probability of treatment weighting (IPTW) based on propensity scores was employed. Propensity scores were estimated using generalized boosted modelling (GBM) with 10 000 decision trees³⁶. Compared with conventional methods like logistic regression, the GBM using machine learning techniques avoids many assumptions, such as linearity between explanatory variables and study groups and the absence of multiway interactions (for example three-way interactions) among explanatory variables³⁶. Typically, setting the interaction depth to 3 and using 10 000 trees strikes an optimal balance between overfitting and bias reduction. All covariates were used to calculate the propensity scores (Table 1). Notably, the admission date for BSIs was included in the propensity score calculations to ensure equal potential follow-up durations across the three study groups. The balance of baseline characteristics among the three study groups before and after IPTW was evaluated using the maximum absolute standardized difference (MASD), with a value below 0.1 indicating negligible differences. Consequently, all outcome comparisons between the study groups were conducted using an IPTW-adjusted cohort. The Cox proportional hazards model was used to compare the risk of all-cause death among the three study groups. Owing to the all-cause mortality being the competing risk for all other outcomes (for example a new diagnosis of malignancy), the incidence of all other outcomes among the three study groups was compared using the Fine–Gray subdistribution hazard model. Using a subdistribution hazard model provides more accurate estimates of incidence for outcomes. The parameter estimates in the Fine and Gray model differ from those in the Cox model, as the Fine and Gray model focuses on the incidence or the likelihood of experiencing an outcome. The proportional hazards assumption was investigated using the Schoenfeld residuals approach. All correlation coefficients between Schoenfeld residuals and ranked survival time were not significant for all outcomes except pneumonia (data not shown).

Table 1.

Baseline demographics and clinical characteristics of patients with blunt splenic injuries following splenectomy, splenic artery embolization, or conservative management

	Total (n = 8195)	Splenectomy (n = 2533)	SAE (n = 1193)	Conservative (n = 4469)	MASD
Demographics
Age (years), mean(s.d.)	38.4(16.4)	39.7(16.9)	37.5(16.4)	38.0(16.1)	0.13
Age group (years)
<30	3288 (40.1)	938 (37.0)	502 (42.1)	1848 (41.4)	0.10
30–49	2778 (33.9)	874 (34.5)	392 (32.9)	1512 (33.8)	0.03
50–64	1524 (18.6)	494 (19.5)	216 (18.1)	814 (18.2)	0.04
≥65	605 (7.4)	227 (9.0)	83 (7.0)	295 (6.6)	0.09
Male	5779 (70.5)	1831 (72.3)	837 (70.2)	3111 (69.6)	0.06
Female	2416 (29.5)	702 (27.7)	356 (29.8)	1358 (30.4)
Associated injuries
Traumatic brain injury	99 (1.2)	30 (1.2)	16 (1.3)	53 (1.2)	0.01
Spinal cord injury	45 (0.6)	16 (0.6)	5 (0.4)	24 (0.5)	0.03
Injury to heart and lung	711 (8.7)	188 (7.4)	137 (11.5)	386 (8.6)	0.14
Pneumothorax or haemothorax	1761 (21.5)	536 (21.2)	296 (24.8)	929 (20.8)	0.10
Injury to gastrointestinal tract	443 (5.4)	279 (11.0)	19 (1.6)	145 (3.2)	0.42
Femur fracture	332 (4.1)	126 (5.0)	44 (3.7)	162 (3.6)	0.07
Major trauma*	1628 (19.9)	715 (28.2)	396 (33.2)	517 (11.6)	0.54
Co-morbidities
Hypertension	863 (10.5)	264 (10.4)	132 (11.1)	467 (10.5)	0.02
Ischaemic heart disease	116 (1.4)	49 (1.9)	17 (1.4)	50 (1.1)	0.07
Diabetes mellitus	117 (1.4)	21 (0.8)	37 (3.1)	59 (1.3)	0.19
COPD	126 (1.5)	53 (2.1)	13 (1.1)	60 (1.3)	0.08
Cirrhosis	231 (2.8)	108 (4.3)	31 (2.6)	92 (2.1)	0.13
Chronic kidney disease	148 (1.8)	46 (1.8)	23 (1.9)	79 (1.8)	0.01
Bed confinement status	61 (0.7)	28 (1.1)	8 (0.7)	25 (0.6)	0.06

Data are n (%) unless otherwise stated. COPD, chronic obstructive pulmonary disease; MASD maximum absolute standardized difference; SAE, splenic arterial embolization. *Injury Severity Score ≥16.

Because of the low incidence rate (<30 events per 10 000 person-years; Table S2) of individual thromboembolic outcomes, group comparisons were conducted only for composite thromboembolic outcomes. A two-tailed P <0.05 was considered statistically significant. All statistical analyses were performed using R (version 4.2.1; R Foundation for Statistical Computing) and SAS (version 9.4; SAS Institute, Inc.).

Results

Temporal trend in treatment of blunt splenic injuries

Figure 1a illustrates the temporal trend in the treatment of BSIs, showing a shift from splenectomies to SAEs over time. In Taiwan, SAE emerged in 2004 and has become a mainstay of treatment for BSIs with active bleeding and stable haemodynamics. During the study period (2004–2019), 18 771 patients sustained BSIs. After applying the exclusion criteria, 8195 patients were analysed. Of these, 2533 (30.9%) underwent splenectomies, 1193 (14.6%) underwent SAEs, and 4469 (54.5%) received conservative treatment (Fig. 1b).

Baseline characteristics

Table 1 presents the demographic characteristics of the study participants. The patients were relatively young, with a mean age of 38.4 ± 16.4 years, and were predominantly male (70.5%). Patients who underwent splenectomies experienced more gastrointestinal tract injuries than those in the SAE and conservative treatment groups (11.0% versus 1.6% versus 3.2% respectively; MASD = 0.42). Additionally, the SAE group had a higher proportion of patients with major trauma (ISS ≥ 16) than the splenectomy and conservative treatment groups (33.2% versus 28.2% versus 11.6% respectively; MASD = 0.54). The underlying co-morbidities of the three study groups were comparable. After applying IPTW, the three study groups were found to be comparable, with MASD values less than 0.1.

Infectious outcomes

Regarding the infection risk, patients who underwent SAEs (subdistribution hazard ratio (SHR) 0.72, 95% c.i. 0.65 to 0.81) or conservative treatment (SHR 0.79, 95% c.i. 0.71 to 0.88) had a lower risk than the splenectomy group (Fig. 2a). Similar results were noted for septicaemia and pneumonia. There was no statistically significant difference in the overall infection risk between the SAE and conservative treatment groups. However, the conservative treatment group had a higher septicaemia risk (SHR 1.29, 95% c.i. 1.10 to 1.51) but a lower meningitis risk (SHR 0.31, 95% c.i. 0.14 to 0.60) than the SAE group (Table 2).

Fig. 2.

Cumulative event rates

Cumulative event rates for a composite infectious diseases, b newly diagnosed malignancies, c composite thromboembolism outcomes, and d all-cause mortality among patients who underwent various treatments for blunt splenic injuries in the IPTW-adjusted cohort. SAE, splenic artery embolization; SHR, subdistribution hazard ratio.

Table 2.

Long-term follow-up outcomes of patients with blunt splenic injuries: a comparison of splenectomy, splenic artery embolization, and conservative management in the IPTW-adjusted cohort

	HR/SHR (95% c.i.)
Outcome	SAE versus splenectomy	Conservative versus splenectomy	Conservative versus SAE
Infectious disease
Composite outcome*	0.72 (0.65–0.81)†	0.79 (0.71–0.88)†	1.09 (0.97–1.22)
Septicaemia	0.65 (0.56–0.76)†	0.84 (0.73–0.97)†	1.29 (1.10–1.51)†
Meningitis	1.23 (0.68–2.23)	0.38 (0.17–0.87)†	0.31 (0.14–0.69)†
Pneumonia	0.82 (0.71–0.93)†	0.83 (0.73–0.94)†	1.02 (0.89–1.17)
Malignancy
Any type of malignancy	0.89 (0.74–1.08)	0.83 (0.69–0.99)†	0.92 (0.76–1.12)
Haematological	1.44 (0.60–3.45)	1.16 (0.47–2.86)	0.81 (0.35–1.86)
Solid	0.87 (0.72–1.06)	0.82 (0.68–0.99)†	0.94 (0.77–1.15)
Thromboembolism disease
Composite outcome‡	0.68 (0.56–0.82)†	0.68 (0.56–0.83)†	1.01 (0.82–1.25)
Ischaemic stroke	1.02 (0.80–1.31)	0.75 (0.58–0.96)†	0.73 (0.56–0.94)†
All-cause death	0.86 (0.75–0.98)†	0.92 (0.81–1.04)	1.07 (0.93–1.22)

IPTW, inverse probability of treatment weighting; SAE, splenic arterial embolization; SHR, subdistribution hazard ratio. *Anyone with septicaemia, meningitis, and pneumonia. †P-value <0.05. ‡Anyone with acute myocardial infarction, venous thromboembolism, ischaemic stroke, embolism and thrombosis of arteries of the extremities, portal vein thrombosis and intestinal vascular insufficiency.

Malignancy outcomes

Regarding malignancy, the conservative treatment group had a lower risk than the splenectomy group (SHR 0.83, 95% c.i. 0.69 to 0.99; Fig. 2b). The SAE group exhibited a numerically lower risk of malignancy than the splenectomy group for all types of cancers (SHR 0.89, 95% c.i. 0.74 to 1.08). Additionally, the SAE and conservative treatment groups showed a numerically higher risk for haematological malignancies (Table 2).

Thromboembolism events

The incidence of individual thromboembolism event components is detailed in Table S2. Patients with BSIs who underwent SAEs (SHR 0.68, 95% c.i. 0.56 to 0.82) or conservative treatment (SHR 0.68, 95% c.i. 0.56 to 0.83) had a lower thromboembolism risk than those who underwent splenectomies (Fig. 2c). The risk of composite thromboembolism events did not differ significantly between the SAE and conservative treatment groups. Regarding ischaemic stroke, the conservative treatment group had a lower risk than the SAE (SHR 0.73, 95% c.i. 0.56 to 0.94) or splenectomy (SHR 0.75, 95% c.i. 0.58 to 0.96) groups (Table 2).

All-cause mortality

Lastly, patients with BSIs who underwent SAEs had a lower risk of all-cause mortality than those with splenectomies (hazard ratio 0.86, 95% c.i. 0.75 to 0.98) and not statistically different from those with conservative treatment (hazard ratio 1.07, 95% c.i. 0.93 to 1.22) (Fig. 2d; Table 2).

Discussion

The current study demonstrated that the long-term infection, malignancy, thromboembolism, and all-cause mortality risks in patients who underwent SAEs for blunt splenic injury were not significantly different from those in patients who received conservative treatment. Furthermore, patients with BSIs who underwent NOM, regardless of whether they underwent SAEs, showed a lower adverse outcome risk than those who underwent splenectomies. This provides real-world evidence that SAEs do not significantly alter splenic function and that NOM offers long-term benefits for patients with BSIs.

Research utilizing the National Readmission Database indicated that patients with SAEs experienced a higher infection risk at one-year follow-up than those who received conservative treatment^33,37. However, infections occurring within the first year of the index admission may be related to the trauma or interventions and are not necessarily indicative of splenic function. Numerous studies have assessed serum markers related to immune function, with most reporting that splenic function remained intact after more than one year of follow-up^38–44. Furthermore, Bhatia et al. reported that splenic volume on CT did not change significantly following SAE over an average period of two years⁴⁵. These findings align with the current observation that the SAE group does not have the same long-term infection risk associated with splenectomies.

However, the association between splenectomies and cancer remains unclear. Kristinsson et al. reported an increased risk of cancer among veterans who had splenectomies in the United States²⁵. Similarly, an analysis of Taiwan’s NHIRD revealed comparable results, regardless of the splenectomy indications²⁶. Nevertheless, three studies found no significant increase in the cancer risk among patients who underwent splenectomies for trauma^46–48. No studies have reported the cancer risk in patients who underwent SAEs. The current study demonstrated a significantly reduced cancer risk between the conservative treatment and splenectomy groups. Although a similar trend was observed in patients who underwent SAEs, the difference was not statistically significant. One possible explanation for these findings is the relatively young population of the current study, which may require a longer follow-up period to detect differences in cancer development. Additionally, a higher risk of haematological malignancies was observed in the SAE and conservative treatment groups. It is hypothesized that this could be attributed to the low incidence of haematological malignancies, which might increase the risk owing to the small number of target outcomes. Nevertheless, the occurrence of malignancies may be influenced by multiple factors that warrant further investigation.

Patients who undergo splenectomies face increased risks of infections and certain cancers due to the spleen's critical role in the immune system. The spleen is essential for filtering blood, producing antibodies, and generating immune responses against encapsulated bacteria²⁵. Following splenectomy, there is a notable decline in memory B cells and marginal zone macrophages, which are vital for responding to these pathogens, leading to impaired antibody production and increased susceptibility to overwhelming post-splenectomy infection²⁴. The pathophysiology of the heightened infection risk is further complicated by immunological changes such as diminished T-cell function and altered cytokine profiles, which can result in a systemic inflammatory response. Additionally, splenectomized patients exhibit an elevated risk of developing certain malignancies, potentially due to chronic inflammatory states that promote tumorigenesis through dysregulated cytokine signalling and immune evasion mechanisms^24,48.

Undergoing a splenectomy has been associated with a hypercoagulable state six weeks post-surgery, characterized by thrombocytosis, elevated fibrinogen levels, and rapid fibrin cross-linking, as observed on thromboelastography⁴⁹. A similar observation was noted in patients with SAEs, where thrombocytosis and elevated fibrinogen levels were detected within 7 days⁵⁰. Lewis et al. reported that SAE was an independent risk factor for VTE events during the index admission for BSIs, using data from the ACS Trauma Quality Improvement Program (ACS TQIP) database⁵¹. Mulder et al. analysed the National Readmission Database and identified that SAE was an independent risk factor for VTE. However, the risk decreased over time, from an odds ratio of 2.04 (1.51–3.38) during the index admission to 1.29 (1.01–1.64) one year later, compared to the conservative treatment group³². In conjunction with the current study’s findings, it is proposed that SAE leads to a temporary alteration in patients’ coagulation profiles, resulting in an increased thromboembolic event risk, which is expected to diminish over time as patients recover. In the long term, the SAE group did not have the same thromboembolism risk as the splenectomy group. Concerns have arisen owing to the high incidence of ischaemic strokes in patients undergoing SAEs, which requires clinician attention.

Patients who have undergone splenectomies have been reported to face an increased mortality risk when experiencing pneumonia, sepsis, pulmonary embolism, coronary artery disease, or cirrhosis^25,46. The current study revealed comparable results and further disclosed that patients who underwent SAEs did not have the same elevated mortality risk. An immunocompromised and hypercoagulable state after splenectomy increases the subsequent mortality risk. In contrast, patients who undergo SAEs retain adequate splenic function, which may protect them from the associated mortality risks.

Given the infection, malignancy, thromboembolism, and mortality risk associated with splenectomies in patients with BSIs, splenic preservation should be prioritized whenever clinically feasible. Once a splenectomy is performed as a life-saving measure, patients should be thoroughly informed of its associated risks. Clinicians should also ensure patients receive appropriate vaccinations, thromboprophylaxis, and cancer surveillance^25,52. If patients who have undergone splenectomies develop infections or thromboembolisms, they must be aware of the increased mortality risk associated with these conditions and provide proper care. Routine vaccination and cancer surveillance may not be necessary for patients with BSIs who underwent SAEs. A practice management guideline from the Eastern Association for the Surgery of Trauma recommends against routine vaccination following SAE⁵³. Based on the current study, long-term thromboembolic prophylaxis is not necessary for patients with BSIs following SAE.

It should be emphasized that this study provides valuable insights into long-term outcomes following treatments, which can help clinicians prevent and manage complications effectively. Treatment for BSIs should continue to be guided by clinical scenarios in line with current management guidelines rather than being influenced by potential adverse events. Ultimately, prioritizing life-saving measures remains the biggest concern, regardless of long-term complications. The effort to avoid splenectomy in a critically ill patient to reduce long-term risks should never be considered.

This study had several limitations. First, it was not a prospective randomized clinical trial, meaning the results can only be interpreted as an association rather than a causation. Second, as a database study, detailed information may not have been available for analysis. For example, the specific site of transarterial embolization could not be assessed, requiring creation of a cohort of patients with isolated splenic injuries by excluding other potential embolization sites. Moreover, it is important to note that pneumococcal vaccination is not covered by NHI for splenectomized patients, which prevented the vaccination status of the study population being assessed and may have impacted the results. Furthermore, it would be beneficial to compare outcomes between splenectomized patients who have received vaccinations and those who have not. Additionally, efforts were made to address potential confounding factors related to the main outcomes of interest, including immune disorders, coagulation defects, patients’ basic demographics, associated injuries, and co-morbidities^23,54–56. However, other confounding factors such as family history, nutritional status, physical activity, smoking, alcohol consumption, and exposure to radiation or chemicals were not available in the database. Third, the follow-up period was too short to adequately detect cancer development or assess its effects over time. In the future, a prospective study with a follow-up period of at least 20 years is essential to thoroughly clarify the long-term outcomes of SAE, along with detailed records of the embolization technique, patient vaccination status, and cancer risks. Additionally, scheduling check-ups that include immunological serum markers and coagulation profiles could provide valuable physiological insights into the actual splenic function reserve.

In patients with BSIs, the long-term infection, malignancy, thromboembolism, and all-cause mortality risks were not significantly different between those who underwent SAEs and those who received conservative treatment. In contrast, patients who have undergone splenectomies are at an increased risk of infection, thromboembolism, and all-cause mortality. Clinicians must be familiar with the potential long-term complications associated with the different treatment modalities for splenic injuries and provide appropriate prophylactic measures.

Funding

Jen-Fu Huang received Grant No. MOST 111-2314-B-182A-019 from the Ministry of Science and Technology, Taiwan, R.O.C.

Disclosure

The authors declare that they have no relevant or significant financial interests related to the research presented in this paper.

Supplementary material

Supplementary material is available at BJS Open online.

Data availability

The study utilized Taiwan’s National Health Insurance Research Database, which is accessible through an application process.

Author contributions

Jenfu Huang (Conceptualization, Funding acquisition, Writing—original draft), Ling-Wei Kuo (Data curation, Resources), Chih-Po Hsu (Formal analysis, Validation), Chi-Tung Cheng (Investigation, Methodology, Software, Supervision, Validation), Sheng-Yu Chan (Data curation, Methodology), Pei-Hua Li (Investigation, Visualization), Szu-An Chen (Formal analysis, Resources, Software), Chia-Cheng Wang (Project administration, Validation), Yu-San Tee (Formal analysis), Chun Hsiang Ouyang (Conceptualization, Writing—review & editing), Chien-Hung Liao (Conceptualization, Validation, Writing—review & editing), and Chih-Yuan Fu (Conceptualization, Supervision)