A systematic review of blunt abdominal aortic injury and analysis of predictors of death

At present, research on blunt abdominal aortic injury (BAAI) is limited, with the majority being case reports. Consequently, there is a significant knowledge gap concerning this condition. To address this, we conducted a systematic review by extensively searching major databases. We included all literature that provided individual (non-identifiable) data on BAAI patients, irrespective of the study design. Furthermore, we undertook regression analyses to identify predictors of death after BAAI. The search yielded 2099 results, leading to the inclusion of 102 case reports and one conference abstract. Using the Joanna Briggs Institute (JBI) checklist for assessment, all studies were deemed of medium to high quality. In total, 133 patients were included, with a median age of 34 years, and 73.7% being male. The predominant clinical manifestation was pain, reported in 65.6% of patients. The most frequently observed aortic lesion severity was grade A (intimal tear or intramural hematoma) at 46.9%, and the most common lesion location was zone III (infrarenal aorta) in 88.3% of cases. The overall mortality after BAAI was 15.3%. Multivariate regression analyses revealed the following predictors of death after BAAI: lower limb ischemia (relative risk [RR] = 7.137, 95% confidence interval [CI] 1.154 – 44.161), cardiopulmonary arrest (RR = 10.250, 95% CI 1.452 – 72.344), and injuries to body parts other than the abdomen and lumbar spine (RR = 2.593, 95% CI 1.189 – 5.655). In conclusion, this review provides a detailed quantitative summary of BAAI’s clinical manifestations, diagnosis, treatment, and prognosis, emphasizing its high mortality rate and identifying three critical variables as predictors of death.


INTRODUCTION
Blunt abdominal aortic injury (BAAI) is believed to be caused by an injury to the aorta due to both direct and indirect blunt biomechanical forces exerted on the abdominal aorta (AA).
The AA is tethered between the spinal column and the peritoneum, along with the abdominal viscera [1].BAAI is rare in both adults and children [2].According to previous studies, it accounts for only 4-6% of total aortic injuries and less than 1% of all blunt traumas [3][4][5].
Despite its rarity, BAAI is lethal.Previous research has highlighted a high but variable mortality rate, with figures ranging from 17% to 92% [5][6][7][8].In comparison to blunt thoracic aortic injury (BTAI), BAAI is significantly less common.This is believed to result partly from the relatively fixed position of the entire AA, in contrast to the thoracic aorta, which is fixed only at the ligamentum arteriosum and the base of the heart [9].Shalhub et al. proposed a classification of aortic lesion locations into three types to facilitate surgical exposure and repair [1].This classification method has been favored by many researchers [10,11].On the other hand, when dealing with BTAI or nonspecific blunt aortic injury (BAI), various grading criteria have been developed to assess the severity of aortic lesions based on the pathological changes in the injured aorta [12][13][14][15].Despite these advancements, limited research has explored the impact of aortic lesion location and severity on the prognosis of BAAI patients.The therapeutic approaches for BAAI mainly consisted of surgical intervention and conservative observation.Kondo et al. found no statistically significant difference in 24h mortality or overall hospital mortality rates among the different treatment modalities [16].However, Sheehan et al. reported that patients who underwent aortic surgery exhibited a statistically lower mortality rate than those who did not [17].There is limited research in this area.
The limited existing studies have not extensively addressed the core aspects of BAAI, leaving significant gaps in the understanding of this dangerous traumatic condition.To address these gaps, we conducted this systematic review.This review was registered in the International Prospective Register of Systematic Reviews (PROSPERO) (CRD42023408842) and aligns with the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework [18].

Search strategy
A systematic search was conducted in the Excerpta Medica Database, PubMed, Web of Science (WOS), and Cochrane Library on December 23, 2022.All terms synonymous with or similar to "abdominal aortic injury (AAI)" were searched to ensure no relevant studies were overlooked.The detailed search strategy can be found on Figure S1.

Study selection
The AA was defined as the section of the aorta extending from the diaphragm to where it bifurcates into the right and left common iliac arteries.The study subjects were human individuals diagnosed with BAAI, resulting from blunt or non-piercing external forces.All brief details from the identified articles were imported into Endnote X9 to facilitate the removal of duplicates and to conduct an initial brief review.Subsequently, the full texts of potential available articles that passed the preliminary screening were downloaded for a more thorough review to identify those articles which were included in the study.
Initially, all articles pertaining to AAI that were retrieved from the primary search were considered.In further search, any literature providing individual data on BAAI patients, regardless of the study design, was eligible for inclusion, even if there were missing values for certain variables.The exclusion criteria were as follows: (1) subjects with concomitant thoracic aortic injury; (2) subjects with pre-existing dilative aortic conditions (such as aortic aneurysm) or aortic injury; (3) potential duplicate subjects; (4) articles which were published in languages other than English.If a particular article reported on multiple BAAI cases and only some met the selection criteria, only those specific cases were included in this review.
Two authors, ML and CW, independently performed the data selection process.
Discrepancies between their selections were addressed and resolved through consensus.

Data extraction
An extensive set of variables related to BAAIs was established for detailed data extraction.
The definitions and descriptions of each variable can be found on Table S1.Specifically, "death after BAAI" was defined as any death resulting from the initial traumatic event, excluding other causes such as malignant cancer.Aortic lesion locations were categorized into three zones based on Shalhub et al.'s methodology [1]: zone I corresponding to the zone extending from the diaphragmatic hiatus to the superior mesenteric artery (SMA); zone II corresponding to the zone encompassing the SMA and the renal arteries; and zone III corresponding to the zone spanning from the infrarenal aorta to the aortic bifurcation.
Furthermore, the severity of aortic injuries was classified into five grades building on the basis of Rabin et al.'s method [15]: grade A corresponding to intimal tear or intramural hematoma; grade B corresponding to small pseudoaneurysm (less than 50% circumference); grade C corresponding to large pseudoaneurysm (more than 50% circumference); grade D corresponding to intraluminal truncation; and grade E corresponding to rupture.Data extraction was performed by two independent authors, HT and HZ.If a variable's value or classification wasn't explicitly reported in an included article based on our definitions, these authors would independently interpret the original text to determine it.Any queries and discrepancis in interpretation were addressed, and consensus was achieved through further discussion.

Quality assessment
The methodological quality of the original studies was evaluated using the Joanna Briggs Institute (JBI) checklist for case reports [19].This assessment tool includes eight items, which evaluate whether the descriptions of the patient's clinical characteristics, diagnosis, treatment, outcomes, and other relevant aspects in each report are comprehensive and clear.
Each item earns a score of 1 point for responses "yes" or "not applicable" and 0 points for responses "no" or "not clear".In instances were a single study reported on multiple cases separately, each case was evaluated separately, and the overall score of the study was determined by the lowest score among its individual cases.
Based on theirs scores, all included studies were classified in terms of methodological quality as "low" (0 -3 points), "medium" (4 -6 points), or "high" (7 -8 points).Three independent authors (ML, ZG and LG) each carried out the quality assessments for all included articles.
In cases of differing opinions, the lower score was taken into consideration.

Ethical statement
Given that this review exclusively utilizes previously published literature and does not contain any personally identifiable information, neither ethical approval nor consent to participate were required, which is in line with prevailing local regulations and policies.Furthermore, our study adhered to the principles outlined in the Declaration of Helsinki (as revised in 2013).

Statistical analysis
Continuous variables were expressed as either "mean ± standard deviation" or "median with interquartile range (IQR)".Categorical variables were presented as "number with percentage".Statistical analyses were conducted using the Stata version 16.0 software (StataCorp., College Station, TX, USA).All hypothesis tests were two-sided, and a P value of < 0.05 was considered statistically significant.The Mann Whitney U test [20] was employed to compare age distributions of BAAI patients across sexes.The constituent ratios for all samples in categorical variables were also calculated separately.The details of patients who died after BAAI were also described.
For assessing potential predictors, the binary "death after BAAI" outcome variable was set.
All variables, excluding those who were deemed unrelated like "diagnostic method", were individually incorporated into a binomial family generalized linear model (GLM) (link: log; standard error type: default; optimization method: maximum likelihood estimation) for univariate regression analyses, with relative risk (RR) and 95% confidence interval (CI) as metrics [21].To achieve a probable positive result and retain maximum detailed categorization, we tried multiple combinations of core treatments, incorporating each set of derived variables into the model individually.An adjusted multivariate regression analysis was initially attempted using the above GLM for variables with P < 0.20, but convergence wasn't achieved and no results were derived.Subsequently, a robust Poisson family GLM (link: log; standard error type: robust; optimization method: maximum likelihood estimation) was used for a multivariate regression analysis of these variables.Only variables with a P value < 0.05 and an absolute RR value < 1000 were accepted and considered as predictors.
To assess the final adopted regression model's fitting ability we recorded the values of the Akaike information criterion (AIC) [22] and log pseudolikelihood [23].Furthermore, subgroup analyses were conducted using Fisher's exact test [24] on patients with the calculated predictors to explore the influence of different treatment modalities on mortality within these specific patients.All statistical analyses were conducted by ML.

Characteristics of studies and patients
Our systematic search initially identified 2,099 articles.After removing 1,123 duplicates, we screened the titles and abstracts of the remaining studies, excluding additional 766 articles.Subsequently, full-text assessment led to the final inclusion of 103 articles  (Figure S2).
These articles comprised 102 case reports and one conference abstract.Among them, 14 articles (13.6%) reported on multiple cases, ranging from two to seven cases.In terms of quality assessment, 91 studies were deemed high-quality studies (88.3%), 12 studies were deemed medium-quality ones (11.6%), with none classified as low-quality studies.These articles were published from 1961 to 2021, with 83 (80.6%) published after 1990.They originated from 25 countries, with the USA contributing the most (48, 46.6%).One case involved an injury abroad with subsequent treatment in the USA later and was reported by American doctors [31].We assumed that the rest of the patients were injured and, if treated, received their care in their respective reporting countries.The authors of one article published in 2018 were from Serbia and Montenegro, respectively, [98], and the reporting country of this article was considered to be one rather than two.In total, 133 BAAI patients were included, with a median age of 34 years (IQR 17-54), and an age range of 1 to 89 years.The detailed characteristics of the included studies and patients can be found in Table S2.Among the patients, 35 were females (26.3%), who were generally older than males (P = 0.012) (Figure S3).

Clinical manifestations, diagnoses, treatments, and outcomes
The predominant cause of BAAI in patients was "direct strike (non-seat belt)", such as collisions, falls, gas shocks, and more, accounting for 53.4% of cases.The various clinical presentations were classified into four categories, with pain emerging as the most common symptom (65.5%).Notably, 5.0% of patients presented with cardiopulmonary arrest and 36.4% with shock.Using the first day after trauma as a reference point, 11.8% of patients experienced delayed manifestations, whereas 4.2% showed no acute manifestations.The distribution of patients across various clinical manifestation variables are shown on Figure 1.
Computed tomography (CT) was the most commonly used diagnostic tool, utilized in 52.6% patients.Three patients were diagnosed with BAAI after substantial delays of 7 months, 8 years, and 9 years after trauma, respectively, due to different reasons [41, 43, and 58].Using the established classification systems, 88.3% of patients had an aortic lesion located in zone III, with 47.0% being of grade A severity.Thrombosis at the aortic lesion site was reported in 28.8% of cases, and 78.3% of cases were reported to have aortic degeneration.In addition, 66.9% of patients had concurrent injuries to other abdominal organs, 25.6% to the lumbar spine, and 23.3% to other body parts.Concomitant injuries to as many as seven organs or tissues after trauma were reported in one patient [102].Among all abdominal organ injuries, the gastrointestinal tract was most frequently affected, accounting for 49.6% of all BAAI patients.The distribution of diagnostic variables is depicted in Figure 2.
The most common treatment modality for BAAI patients was primary open surgery (OS), accounting for 55.2% of cases, followed by primary endovascular therapy (EVT) at 24.0%, with other modalities making up the remainder.One patient initially underwent percutaneous aortic stent implantation after an unsuccessful conservative observation.However during the procedure, it was discovered that the stent had migrated to the suprarenal aorta.
Consequently, the patient underwent an open aorto-aortic bypass and made a successful recovery [91].This patient was subsequently categorized as having undergone a secondary OS as the core treatment modality.Among the 114 patients who underwent surgical treatment, the decision of their treatment approach (OS or percutaneous EVT) was not statistically associated with the presence of neither injuries to abdominal organs, lumbar spine, or other body parts (P = 0.144; P = 0.240; and P = 0.486, respectively).Similarly, the incidence of gastrointestinal injuries, the most common abdominal trauma, didn't significantly sway the choice between OS and EVT (P > 0.999).Of the 114 patients who underwent surgery, 60 patients had gastrointestinal injuries, and 29 of them underwent OS rather than EVT (P > 0.999).Regarding the timing of core treatments after injury, it was found that most patients (64.5%) received treatment rapidly, within three days after injury.
One patient underwent immediate OS upon admission [43], while another patient received treatment in the ninth year after the injury due to a missed diagnosis [44].The longest follow-up duration reported was ten years [105].Post-treatment, 27.1% of patients experienced various non-fatal adverse events, with acute renal failure being the most common at 47.8%.Some patients reported up to four distinct adverse events following OS, excluding instances of multiple organ dysfunction syndrome [43,45].Residual chronic lower limb ischemia and neurological dysfunction were observed as post-treatment sequelae in 5.0% and 10.0% of patients, respectively.A total of 19 patients (15.3%) died after BAAI.They were classified based on the time of death (Table 1).The distribution of treatment modality and outcome variables are shown in Figure 3.Among the 53 cases with available data who had gastrointestinal injury and underwent surgical procedures, three died after BAAI in the OS group and none in the EVT group, which did not suggest a statistically significant difference (P = 0.543).

Death after blunt abdominal aortic injury
Due to failed convergence of the binomial regression model after the selections of variables through univariate analyses, a robust Poisson GLM was employed for analysis, which derived several variables with P < 0.05.However, the absolute RR values of some of these variables, including all the multi-categorical ones, were exaggeratedly exceeding 1000, with some even reaching trillions.To pinpoint variables that destabilized the model, we utilized multiple stepwise regression approaches, such as forward and backward regression.After excluding all multi-categorical variables, a stable Poisson regression model emerged, encompassing 86 observations.The model had an AIC value of 1.018 and a log pseudolikelihood value of -34.776.We collected all the variables in this model and tried the binomial regression again, but the binomial regression still could not achieve convergence.Thus, we ultimately utilized the Poisson regression model, and three statistically significant predictors (or risk factors) of death after BAAI were identified, which were lower limb ischemia (RR = 7.137, 95% CI 1.154-44.161;P = 0.035), cardiopulmonary arrest (RR = 10.250,95% CI 1.452-72.344;P = 0.020), and injuries to other body parts (RR = 2.593, 95% CI 1.189-5.655;P = 0.017).A detailed breakdown of the analysis is presented in Table 2.The adopted model results are visualized in a forest plot (Figure S4).Upon conducting subgroup analyses, it was found that BAAI patients with either lower limb ischemia or injuries to other body parts had lower mortality rates with both primary and secondary operations compared to those who received conservative observation (P = 0.020 and P = 0.035, respectively).Conversely, for patients with cardiopulmonary arrest, mortality rates did not differ significantly across different treatment modalities (P > 0.999).Detailed findings are shown in Table S3.

Significance of this study
This review elucidated the multifaceted characteristics of BAAIs and identified risk factors associated with death.Such insights could help clinicians to more effectively recognize this rare yet fatal traumatic condition, and to provide timely and accurate treatment for high-risk patients.

Characteristics of blunt abdominal aortic injuries
Most of the studies regarding BAAI are case reports, underscoring the rarity of the condition.
Among the 133 included BAAI patients, the median age for males was significantly lower than that for females (32 years vs 45 years; P = 0.012).This discrepancy might be attributed to the inclination of men in adolescence or youth towards high-intensity activities, such as driving motor vehicles, playing football, and boxing.
A direct blunt external force of sufficient magnitude can cause BAAI, and this type of trauma accounted for the majority (53.4%) of causes in this review.However, a significant portion of patients (43.1%) did not experience a direct crash, instead they were restricted by the seat belt and faced injuries during sudden deceleration, thereby developing BAAI.Such injuries are believed to result from the combined compressive effect on the AA by the anteriorly located abdominal viscera and the posteriorly located lumbar spine [99,128].Meanwhile, we also found 4 patients (3.4%) who developed BAAI without suffering any direct external force applied to the abdomen, attributing their injuries to back hyperextension which caused excessive AA traction [82,83,88,100].Additionally, regardless of the injury type, damage to the lumbar spine may promote the development of BAAI due to the increased local force on the AA [46,47,82].The presented review showed that the clinical manifestations of BAAI patients are diverse.Severe cases may present with shock or cardiopulmonary arrest, but pain, lower limb ischemia, and neurological dysfunction were representative.Among them, the proportion of patients with reported pain in the abdomen, chest, back, or lumbar flakes reached 65.5%, suggesting that pain may be the most common clinical symptom of BAAI patients.In addition to pain and lower limb ischemia due to reduced blood flow, 12.6% of patients exhibited neurological dysfunctions at different degrees, including paraplegia, hypoesthesia, asynodia and so on.These complications may arise from the direct injury to the lumbosacral spinal cord or plexus, or the damage of Adamkiewicz artery, which originates from the AA and supplies blood directly to the lumbosacral spinal cord [129].
The majority (52.6%) of BAAIs were diagnosed using CT.This underscores the significance of CT as the most important imaging modality for the diagnosis of BAAI, which is also in line with the findings of many researchers [10,130].Up to 88.3% of patients exhibited aortic lesions in zone III, spanning from the infrarenal aorta to the aortic bifurcation.In contrast, only 15 cases presented with injuries located in zone I (above the SMA) and zone II (encompassing SMA and renal arteries).This could be attributed to the infrarenal AA being less protected compared to the suprarenal AA [131].However, following the regression analyses, there was not enough evidence to prove that different injury locations affected the mortality rate of BAAI.Since the severity grading criteria for aortic lesion are not uniform [10,15,132], after comparing them we adopted Rabin et al.'s method specific to BAAI [15].Yet, under this grading, intimal tears, categorized as the mildest grade (grade A), left large aortic dissections (AD), with or without thrombosis leading to luminal obstructions, unaccounted for.Consequently, we introduced this severe form of AD as grade D, positioned just below grade E (rupture) in terms of severity, based on this standard.Even after this inclusion, the proportion of BAAI patients of the mildest grade (grade A) remained the highest (47.0%), indicating that most BAAIs result in mild AA damage.Aortic degenerative pathology was present in 78.3% of patients, while 25.6% reported lumbar spine fractures, with relevant reports.These two conditions, based on the compression pathogenic mechanism theory of the AA, may increase the susceptibility of trauma patients to BAAI, a hypothesis in line with the ideas of several studies [18,131].
The prevailing treatment preference for BAAI among most clinicians is primary OS, accounting for 55.2%.When considering primary operations, including primary EVT, this proportion even reaches 79.2%.These statistics indicate a limited application of conservative treatments in BAAI.However, it's evident that not all BAAI patients encounter identical risks.Shalhub et al. believed that the decision to operate and the choice of surgical modality should be chosen dependent on the location and severity of the aortic injury [1,6].Moreover, findings from this review suggest that the choice of surgical approach by clinicians isn't influenced by the presence of injuries to other organs or tissues, especially abdominal ones.
We believe that this tendency can be partly attributed to the established practices and preferences of individual medical centers and clinicians.
Generally, the underlying causes of some unwanted or unexpected events in BAAIs could not be solely explained as traumatic or iatrogenic.Consequently, we collectively referred to these negative events, which occurred after treatment (including conservative observation), as "adverse events" (excluding death).We observed that the incidence of these adverse events stood at 27.1%, aligning closely with findings from a prior study [17].Excluding deaths unrelated to BAAI (such as those due to malignancies years later), the mortality after BAAI in this presented review was 15.3% (n = 19).The immediate cause of death among these patients was mostly haemorrhage resulting from the AA injury, even in cases that had received treatment.

Predictors of death after blunt abdominal aortic injury
The analysis results indicate an elevated risk of death after BAAI in patients presenting with lower limb ischemia, cardiopulmonary arrest, or injuries to other body parts.We believe that only severe AA injuries, such as those of grade D/E, are sufficient to cause lower extremity ischemic symptoms.Thus, the presence of lower limb ischemia can serve as a direct indicator of the ischemic severity resulting from aortic injury.Acute limb ischemia has long been associated with many serious consequences, including death, as substantiated by previous studies 133, 134.The preliminary multivariable model did suggest that the severity of aortic disease (a multiple categorical variable) had a direct impact on the mortality (P < 0.001), but it was not adopted due to its unusual RR values (over 1000).On the other hand, both cardiopulmonary arrest and injuries to other body parts signify the severity of trauma to the whole body, not limited to the AA.In essence, concurrent injuries could potentially aggravate the impact of an AA injury on the body or even lead to death.It is very difficult pinpointing the specific initial cause of death in a patient with multiple injuries.Several variables were also excluded from our model due to their excessively high RR values which exceeded 1000.
For instance, the mortality rates associated with zone I and II injuries, which are more difficult to access or control through OS, were higher compared to those of zone III injuries.Moreover, secondary EVT following unsuccessful conservative observation was associated with increased mortality compared to consistent conservative treatment.These findings, despite not being statistically valuable, should not be easily negated given their theoretical plausibility.The unusual RR values are likely a result of the small sample size.We anticipate future research with larger sample sizes to provide more clarity on these matters.Subgroup analysis results suggest that for BAAI patients with either lower limb ischemia or injuries to other body parts, both primary and secondary operative treatments (including OS and EVT) can reduce the mortality risk compared to conservative observation (P < 0.05).This underscores the importance of surgical intervention for high-risk BAAI patients.However, for BAAI patients presenting with cardiopulmonary arrest, the differences in protective effects between the three treatment modalities were not evident.Given the small sample size (n = 2) for this subgroup, it's inconclusive to state that surgical interventions do not play a role in reducing the risk of death for such BAAI patients.

Limitations
This review also has certain limitations.Firstly, due to the absence of uniform standards across the references, variable descriptions were occasionally vague, potentially affecting the accuracy of data extraction and subsequent analysis.Secondly, cases with favorable outcomes may be more readily reported and published, while those with less satisfactory outcomes might go unreported for various reasons.This could potentially lead to an underestimation of the true mortality rate, introducing publication bias.

CONCLUSION
BAAI represents a lethal injury with diverse characteristics and a minumum mortality rate of 15.3%.Factors such as lower limb ischemia, cardiopulmonary arrest, and injuries to body parts beyond the abdomen and lumbar spine increase the mortality risk associated with BAAI.Surgical intervention, whether through OS or EVT, can reduce the mortality in BAAI patients exhibiting lower limb ischemia or injuries to other body parts, even when used as a salvage measure following unsuccessful conservative observation.

Direct strike
Including motor vehicle crash (driver or passenger), fall, hit, airflow impact, or squeeze.

Seat belt
Seat belt was tied when trauma.

Back hyperextension
Back hyperextension without external forces.

Pain
Pain in abdomen, chest, back, or lumbar flanks.
Lower limb ischemia Including pain, weakness, cold, pulselessness, or pallor of lower extremity due to ischemia.
Other manifestations Other manifestations due to trauma not covered by the previous two variables, including gastrointestinal symptoms, consciousness disorders, vertigo, headache, weakness, hyperhidrosis, dyspnea, etc.

Delayed manifestations
Manifestations that occurred more than 1d after trauma.
No immediate manifestations No manifestations were observed within 1d following trauma.

Aortic lesion location
Zone I The zone from the diaphragmatic hiatus to the SMA.

Zone II
The zone encompassing the SMA and the renal arteries.

Zone III
The zone spanning from the infrarenal aorta to the aortic bifurcation.

Aortic lesion severity
Grade A Intimal tear or intramural hematoma.
Grade D Intraluminal truncation.
Grade E Rupture.

Aortic degeneration
Referred degenerative pathologies such as severe aortic calcification or atherosclerosis reported by imaging studies, pathological examination, or autopsy.

Injuries to organs in abdomen
Injuries to other vital organs or tissues in the abdomen such as the abdominal wall, pelvis, diaphragm, omentum, stomach, intestines and mesentery, liver, pancreas, spleen, kidney, etc.
Injuries to lumbar spine All injuries to lumbar spine which were referred, including fracture or deformation.
Injuries to other body parts Injuries to other body parts not covered by the previous two variables, including the head, thorax, thoracic spine, heart, lung, ribs, clavicles, or limbs.

Core treatment
There are five core treatment modalities.Both surgery and intervention were defined as "operation".

Primary surgery
The aorta was repaired, replaced, or bypassed under direct vision through incisional approach proactively after diagnosis of BAAI.

Secondary surgery
After the diagnosis of BAAI, the patient underwent initial conservative observation and then was switched to surgery passively.
Primary endovascular therapy Endovascular repair of aorta with covered stent under the monitoring of aortography was performed through percutaneous puncture approach proactively after diagnosis of BAAI.
Secondary endovascular therapy After the diagnosis of BAAI, the patient underwent initial conservative observation and then was switched to intervention passively.

Conservative observation
Conservative treatment was implemented throughout while monitoring for changes.

Adverse events
Refers to any unexpected adverse events other than death following core treatment, such as respiratory diseases, renal failure, heart diseases, infection, lower limb ischemia, neurological dysfunction, peptic ulcer, osteofascial compartment syndrome, poor incision healing, etc.

FIGURE 1 .
FIGURE 1.The distribution of patients across each of the clinical manifestation variable.The numbers of patients were labeled within the figure.

FIGURE 2 .
FIGURE 2. The distribution of patients across each of the diagnostic variable.The numbers of patients were labeled within the figure.CT: Computed tomography; PA: Percutaneous aortogram; SO: Surgical operation; DUS: Doppler ultrasound; MR: Magnetic resonance.

FIGURE 3 .
FIGURE 3. The distribution of patients across each of the treatment modality and outcome variable.The numbers of patients were labeled within the figure.PS: Primary surgery; PEVT: Primary endovascular therapy; CO: Conservative observation; SEVT: Secondary endovascular therapy; SS: Secondary surgery; BAAI: Blunt abdominal aortic injury.

FIGURE S2 .
FIGURE S2.PRISMA flow chart, showcasing the step-by-step process of study identification and inclusion for our systematic review.PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analyses; WOS: Web of Science.

TABLE 2 . Univariate and multivariate analysis exploring variables associated with death after BAAI Variable Univariate analysis Multivariate analysis 1 † Multivariate analysis 2 ‡ Num of obs Relative risk (95% CI) P value Relative risk (95% CI) P value Relative risk (95% CI) P value
All variables with a P value < 0.20 from the univariate analysis (whose P values are bolded in the table under the univariate analysis) were included in this model; ‡ All variables with a P value < 0.20 from the univariate analysis and a relative risk value < 1000 from multivariate analysis 1 were included in this model.The bolded P values under both multivariate analysis models represent statistically significant values (P < 0.05).BAAI: Blunt abdominal aortic injury; Num of obs: Number of observations; CI: Confidence interval.
table details an extensive set of variables related to BAAIs established for data extraction.Other variables included in the study were: age, sex, cardiopulmonary arrest, diagnostic method, timing to core treatment after injury, residual lower limb ischemia, and residual neurological dysfunction.BAAI: Blunt abdominal aortic injury; SMA: Superior mesenteric artery.