The Longer The Decision-TAE Time, The Higher The Risk of Mortality: A Retrospective Study Of Trans-Arterial Embolization In Pelvic Fracture


 Early administration of hemostasis strategies, such as transcatheter arterial embolization (TAE), is critical in pelvic injury patients because they often experience hemorrhagic shock and other fatal injuries. We investigated the influence of delays in TAE administration on mortality. Patients admitted to the Advanced Critical Care Center at Gifu University with pelvic injury between January 2008 and December 2019 who underwent acute TAE were retrospectively enrolled. The time from when the doctor decided to administer TAE to the start of TAE (needling time) was defined as “decision-TAE time.” We included 162 patients. The median decision-TAE time was 59.5 min. Twenty-five patients died. Kaplan-Meier curves for overall survival were compared, with a significant difference observed between the patients above and below the median cutoff value for decision-TAE time (p=0.02). The age and sex-adjusted multivariable Cox proportional hazards regression model revealed that the longer the decision-TAE time, the higher the risk of mortality (p=0.01). The interaction between TAE duration (procedure time) and decision-TAE time was statistically significant (p=0.044), indicating that TAE duration modified the effect of decision-TAE time on mortality. Decision-TAE time may play a key role in establishing resuscitation in pelvic fracture patients, and efforts to shorten this time should be pursued.


Introduction
Pelvic injury is often associated with hemorrhagic shock and other fatal injuries, and mortality rates remain high, particularly in patients with hemodynamic instability 1 . The unstable hemodynamics and hemorrhage-related mortality rate may be as high as 40%, and overall mortality in these patients may be over 10-32%, even if they are hospitalized in a level 1 trauma center [2][3][4][5] . Early diagnosis of the bleeding source and rapid hemostasis strategies are critical for these patients. In the emergency department (ED), the treatment options used to achieve hemodynamic stabilization in patients with pelvic injury include transcatheter arterial embolization (TAE) and pre-peritoneal packing (PPP) 1,6,7 .
The literature contains many reports on the relative advantages of TAE and PPP [8][9][10][11] . As indicated in these reports, TAE is a less invasive procedure and has become widely accepted as a safe and e cacious substitute for direct surgical intervention. On the other hand, considerable delays in performing embolization and the lack of readily available experts in angiography have been highlighted 8,10 . Some researchers have reported that the mortality rates in patients treated with TAE range from 16-50% 12,13 and this represents a higher mortality rate compared to patients treated with PPP 10 .
Following on from these background studies, reports suggest that earlier administration of TAE results in lower mortality rates 4,14 . In these reports, the so-called "door-to-angioembolization time" should be shorter for better outcomes. In reality, the time course of patients with pelvic injury varies according to patient status. For some patients, there may be time to perform a computed tomography (CT) scan before TAE because their vital signs are relatively stable, and for other patients, this might not be possible. The true effectiveness of shortening the delay before administration of TAE can be con rmed using "decision-TAE time," which represents the time from the decision to administer TAE to the actual administration of TAE.
In this study, we aimed to investigate how "decision-TAE time" in uences mortality in patients admitted to hospital with pelvic trauma.

Patient demographics
A total of 611 patients with pelvic fractures were included in this study. A owchart of the inclusion process is shown in Fig. 3. Table 1  C3 in ve patients (3.1%). There was an unknown fracture type in one patient (0.6%) because of the lack of CT scan, sacral fractures in two patients (1.2%), and acetabular fractures in 15 patients (9.3%). The median decision-TAE time was 59.5 [IQR: 40-86.5] min. The DT group (door-TAE group; patients who did not undergo a CT scan before TAE) comprised 47 patients (29.0%) and the CT group comprised 115 patients (71.0%). The indication for TAE was contrast extravasation on the CT scan in 136 patient (84.0%), massive hematoma on the CT scan in 13 patients (8.0%), and unstable hemodynamics in 13 patients (8.0%).

Relationship between mortality and decision-TAE time
The median decision-TAE time was 59.5 (IQR: 40-86.5). Twenty-ve patients died, and the mortality rate was 15.4%. The Kaplan-Meier curves for overall survival were compared, and a statistically signi cant difference was observed between the patients above and below the median cutoff value for decision-TAE time (p = 0.02) (Fig. 4). The multivariable Cox proportional hazards regression model adjusted for age and sex revealed that the longer the decision-TAE time, the higher the risk of mortality (hazard ratio [HR] for IQR: 40-86.5: 1.5, 95% con dence interval [CI]: 1.1-2.06, p = 0.01). The TAE duration (procedure time) did not show a statistically signi cant in uence on mortality (HR for IQR: 39-75: 1.12, 95%CI: 0.67-1.89, p = 0.667) ( Table 2).
The interaction between TAE duration and decision-TAE time was statistically signi cant (p = 0.044), indicating that TAE duration modi ed the effect of decision-TAE time on mortality (Fig. 5).
Age, gender, and parameters that could in uence the decision-TAE time on arrival were aggregated in every IQR and statistically tested (Table 6).

Discussion
The primary nding of this research was that a longer decision-TAE time resulted in a higher risk of mortality. Moreover, although the actual TAE duration did not have a statistically signi cant in uence on decision-TAE time, the interaction between TAE duration and decision-TAE time was statistically signi cant, indicating that TAE duration modi ed the effect of decision-TAE time on mortality.
In this study, we excluded 10 patients who underwent TAE after PPP because this may have in uenced the effects of TAE on hemostasis. In contrast, we included patients who underwent TAE before PPP to reveal the effects of "primary" TAE. As a result, ve patients ( Some reports have suggested the importance of early TAE for improving mortality 4,13,14,18 . In clinical settings, there are many variations in the circumstances surrounding patient delivery to the ER and the condition of the patient upon delivery. These variables include the presence of associated injuries and the severity of said injuries and differences in vital signs. Moreover, they may or may not have been transferred from another hospital and they may have received previous treatment by prehospital medical professionals. Physicians must decide upon a treatment plan for these patients taking into consideration these factors. Based on these considerations, the true effectiveness of shortening the delay to TAE administration can be con rmed by analyzing the time from when the decision to administer TAE is made to when the TAE is performed and its effect on outcomes.
Several studies have reported that the time to angioembolization is longer than time to PPP, which may be in part due to the availability of orthopedic surgeons compared to interventional radiologists 7,10,19 , and TAE may be delayed at night or on weekends based on reports on other catheter-based interventions 14,20,21 . In our institution, interventional radiologists and the equipment required for TAE are available 24 h a day, 365 days a year. Therefore, the availability of staff and/or equipment was not an issue in the present study. The overall decision-TAE time was 59.5 min, even after performing the other resuscitation procedures. Although PPP still may have advantages in terms of an earlier start time compared to TAE 7,10 , most patients with pelvic fracture, even if they are unstable, can be managed with primary TAE strategies.
In our facility, we aim to nish TAE, including treatment of other bleeding injuries, within 60 min. Although we could not directly clarify the relationship between TAE duration and mortality, our intervention analysis showed that TAE duration modi ed the effect of decision-TAE time on mortality. From these results, the combination of a longer decision-TAE time and shorter TAE duration could have a negative impact on mortality. This indicates that when there was a delay before TAE (often due to hemodynamic instability), the interventional radiology specialists must have nished TAE in a shorter time, and this resulted in a higher risk of mortality. In these situations, physicians may decide to change to PPP for faster hemostasis.
We could not con rm the factors that in uenced decision-TAE time, except for hospital transfer in the DT group (Table 6). Based on these results, the expected parameters that could in uence the severity of the patient's condition, such as ISS, vital signs, and even associated injuries, were not related to decision-TAE time. In addition, transferred patients underwent faster TAE. These results might indicate that faster CT scanning can reduce the decision-TAE time, and therefore, the development of faster imaging strategies is essential. Recently, there have been reports on the effectiveness of hybrid emergency room systems 22,23 , hybrid operation rooms 24 , and mobile angiography systems 25 for trauma patients. These systems consist of an angiography-computed tomography (CT) machine in a trauma resuscitation room and have the potential to provide new evidence in this eld.
This study had several limitations. First, the performance of the CT scan depended on the patient's mode of admission. This meant that we could not determine the severity of the patient's condition based on the de nitions of decision-TAE time could be established, and the results would be changed. In fact, it is di cult to retrospectively ascertain the exact time when the decision to administer TAE is made, and we believe that the de nition of decision-TAE time warrants further discussion.
In conclusion, overall survival was statistically different between the patients above and below the median cutoff value for decision-TAE time, and the longer the decision-TAE time, the higher the risk of mortality. Our results suggest that decision-TAE time may play a key role in establishing resuscitation in pelvic fracture patients, and efforts to shorten the time should be pursued. there was enough information to make a decision, additional examinations were skipped and the patient was delivered directly for TAE. The patients were treated according to the algorithm shown in Fig. 1.

De nition of parameters
Emergency physicians decided to administer TAE when: 1) the CT scan indicated massive hemorrhage from pelvic injury, or 2) the patient was hemodynamically unstable and did not undergo a CT scan, or was transferred from another hospital after the CT scan. We analyzed the time from the decision to administer TAE to the start of TAE (needling time) and called this time the "decision-TAE time." When CT scan indicated massive hemorrhage, "decision-TAE time" was de ned as the time from starting the CT to TAE (CT-TAE group: CT group). When the patient did not undergo a CT scan or was transferred after the CT scan, "decision-TAE time" was de ned as the time from arrival at the ED to the administration of TAE (door-TAE group: DT group) (Fig. 2). Demographic and biological data on admission were collected from medical records. The ISS and the Abbreviated Injury Score (AIS) by body area (head, chest, abdomen, pelvis, and extremities) were calculated for each patient. Outcomes The primary outcome of this study was the time from the end of the TAE to death. There were 10 secondary outcomes, including parameters associated with TAE (decision-TAE time, number of arteries involved in TAE, localizations, embolic materials, time of TAE administration, and number of secondary TAEs), number of patients requiring PPP after TAE, surgical management for pelvic fractures, hospital length of stay, and reasons for death.

Statistical analysis
Baseline characteristics of the patients and the outcome variables were expressed as median and interquartile range (IQR) for continuous variables, and counts and percentages for categorical variables.
In the primary analysis, a Cox proportional hazards regression analysis was performed to con rm the effect of decision-TAE time on the time from the end of TAE to death. The Cox proportional hazards model was adjusted for age and sex to avoid confounding by the patients' baseline characteristics. To avoid over tting, the number of covariates had to be limited to two 26 . Kaplan-Meier estimation was conducted to estimate the cumulative survival rate for each group divided by the median of the decision-TAE time. The difference in the cumulative survival rate between the two groups was con rmed using the log-rank test.
The effects of decision-TAE time on the time from the end of TAE to death were analyzed using a model similar to that used for the primary analysis. In addition, an interaction term (decision-TAE time * TAE duration) was incorporated into the Cox proportional hazards model to test whether the effect of decision-TAE time on mortality was modi ed with the inclusion of TAE duration. The hazard ratio (HR) of the 75th percentile for the 25th percentile of decision-TAE time or TAE duration with the 95% con dence interval was reported in each Cox proportional hazards analysis. The predicted HR values were obtained from the model, and interaction terms were included. Parameters that could in uence the decision-TAE time on arrival were summarized for each group by dividing decision-TAE time into quartiles, and comparisons between groups were conducted using a Fisher's exact test for categorical variables and a Kruskal-Wallis test for continuous variables. Imputation was not used for missing data because no data were missing for the primary outcome. A p-value (two-sided) < 0.05 was considered statistically signi cant. There was no adjustment for multiple comparisons because all analyses were performed as exploratory. All statistical analyses were performed using R version 4.0.3 (https://www.r-project.org/).

Declarations
Funding: None Acknowledgments: We would like to thank Editage (www.editage.jp/ <https://www.editage.jp/>) for English language editing of this paper.      Table 5 Outcomes of the pelvic fracture patients who received acute angioembolization for pelvic injury (N=162   Figure 1 Treatment algorithm for pelvic injury CT, computed tomography TAE, transarterial catheter embolization  The Kaplan-Meier curves of overall survival The Kaplan-Meier curves of overall survival are compared, and a statistically signi cant difference is observed between the patients above and below the median cutoff value for decision-TAE time (p=0.02).