Study setting
In the Pusan National University Hospital Regional Trauma Center, there are more than 900–1,000 severe trauma-related admissions annually (Injury Severity Score [ISS] ≥ 16), of which 200–250 patients present with pelvic fracture. At our institution, three interventional radiologists and the equipment required for TAE are available 24 h a day, 7 days a week [16]. Thus, the time from arrival to angiography can be less than 2 h. Patients with pelvic fractures without extrapelvic injuries requiring emergency treatment are treated according to the pelvic fracture management algorithm (Figure 1). Indication for TAE is intrapelvic contrast extravasation or hematoma in a computed tomography scan or a transient responder with hemodynamic instability (HI) associated with pelvic fractures. If needed, TAE is also conducted after pelvic packing or any damage control operation or procedures (Figure 1).
Study population
We retrospectively reviewed data from the medical records and included a total of 1,017 patients with pelvic fracture admitted to the trauma resuscitation unit at our Trauma Center between November 1, 2015 and December 31, 2019. Pelvic injuries almost always accompany injuries to other organ systems. Considering only isolated pelvic injuries would not be realistic; thus, polytraumatic patients with pelvic bone fracture were included in this study. We excluded patients declared dead-on-arrival or discharged or transferred from a trauma resuscitation unit within 24 h or patients who did not undergo TAE (n = 702) or with unclear medical records (n = 7). Patients who underwent angiography more than 12 h after admission were excluded (n = 4), as they likely had delayed presentation of the indications for TAE or had prolonged periods of time with operative treatment of multiple injuries. We further excluded patients with an Abbreviated Injury Scale (AIS) score for pelvic ring fracture ≤ 2 (n = 100). The final study population included 204 TAE patients (Figure 2).
Available data included age, sex, mechanism of injury, vital sign on arrival, transfusion with packed red blood cells (pRBCs) within 4 h and 24 h of arrival, AIS, ISS, Glasgow Coma Scale score (GCS), Revised Trauma Score (RTS), shock index, Trauma-related ISS (TRISS), massive transfusion within initial 24 h of arrival, hospital length of stay, intensive care unit (ICU) stay, and survival status at 7 days, 28 days, and discharge. Massive transfusion was defined as the replacement by transfusion of 10 units of red blood cells in 24 h.
Definitions and outcome measures
We defined DTE time as the time from the arrival at hospital to the first application of embolic agents such as polyvinyl alcohol, Gelfoam, coils, and so forth to pelvic arteries. We defined door-to-angiography (DTA) time as that from the arrival at the hospital to the beginning of angiography, injury-to-embolization time as that from the onset of injury to the first application of embolic agents, and injury-to-door time as that from the injury onset to the arrival at hospital (Figure 3). Severe pelvic fracture was defined as a pelvic fracture with an AIS for pelvic ring fracture ≥ 3. The shock index was defined as heart rate (beat/min)/systolic blood pressure (SBP; mmHg). HI was defined as shock index ≥ 0.9. Daytime was defined as 8:30 a.m. through 5:30 p.m., and the weekend was defined as 5:31 p.m. Friday through 08:29 a.m. Monday.
The primary outcomes were mortality at day 7 and day 28. Secondary outcomes included overall mortality (in-hospital mortality), pRBC transfusion amounts during the initial 24 h, ICU-free days to day 28, and hospital-free days to day 90. ICU-free days to day 28 were calculated as 28 minus the number of days or part-days in the ICU. All patients who died before the day 28 follow-up were counted as having zero ICU-free days, on the basis that they should be counted as having the worst possible outcome. Hospital-free days to day 90 are a composite of in-hospital death and hospital length of stay, defined as the number of days alive and out of the hospital between the index visit to the trauma resuscitation unit and 90 days later. Patients who died during the index hospitalization and those hospitalized for more than 90 days were classified as having zero hospital-free days. For patients discharged alive before day 90, the number of hospital-free days was calculated as 90 minus the length of stay.
We divided the patients into two groups according to their DTE time (≤150 min vs. >150 min) to assess the effects of that on clinical outcomes. We arbitrarily set the cut-off point (150 min) at the median of the DTE time.
Statistical analyses
We present continuous variables as median and interquartile ranges and categorical variables as numbers and percentages. We compared categorical variables using the chi-square test when appropriate; otherwise, we used Fisher’s exact test. We compared continuous variables with a Wilcoxon rank-sum test on the basis of the distribution. Multivariate binomial logistic regression analyses were performed in a stepwise fashion, evaluating the effects on mortality of age, SBP, lactic acid, base excess, ISS, GCS, RTS, pRBC transfusion amounts during the initial 24 h, DTE time, and injury-to-embolization time. In addition to comparing survival at 7 days and 28 days between DTE time and mortality, Kaplan–Meier plots of survival curves up to 28 days for each group were drawn and their differences were assessed using the log-rank test. We used a multivariate Cox proportional-hazards model to estimate the hazard ratio of DTE time for mortality at day 28 by adjusting for compounding factors. We performed multivariate linear regression analyses to estimate the impact of DTE time on ICU-free days to day 28, hospital-free days to day 90, and 24 h pRBC transfusion requirement. A value of p < 0.05 was declared to be statistically significant. The Statistical Package for the Social Sciences (Version 20.0, SPSS, Inc., Chicago, IL, USA) and STATA software (Version 14.2, Stata Corp., College Station, TX, USA) were used to analyze the data.