Pelvic fractures (PF) with concomitant injuries are on the rise due to an increase of high-energy trauma. Increase of the elderly population with age related comorbidities further complicates the management. Abdominal organ injuries are kindred with PF due to the proximity to pelvic bones. Presence of contrast blush (CB) on computed tomography in patients with PF is considered a sign of active bleeding, however, its clinical significance and association with outcomes is debatable.

To analyze polytrauma patients with PF with a focus on the geriatric population, co-injuries and the value of contrast blush.

This retrospective cohort study included 558 patients with PF admitted to level 1 trauma center (01/2017-01/2023). Analyzed variables included: Age, sex, mechanism of injury (MOI), injury severity score (ISS), Glasgow coma scale (GCS), abbreviated injury scale (AIS), co-injuries, transfusion requirements, pelvic angiography, embolization, laparotomy, orthopedic pelvic surgery, intensive care unit and hospital lengths of stay, discharge disposition and mortality. The study compared geriatric and non-geriatric patients, patients with and without CB and abdominal co-injuries. Propensity score matching was implemented in comparison groups.

PF comprised 4% of all trauma admissions. 89 patients had CB. 286 (52%) patients had concomitant injuries including 93 (17%) patients with abdominal co-injuries. Geriatric patients compared to non-geriatric had more falls as MOI, lower ISS and AIS pelvis, higher GCS, less abdominal co-injuries, similar CB and angio-embolization rates, less orthopedic pelvic surgeries, shorter lengths of stay and higher mortality. After propensity matching, orthopedic pelvic surgery rates remained lower (8% vs 19%, P < 0.001), hospital length of stay shorter, and mortality higher (13% vs 4%, P < 0.001) in geriatric patients. Out of 89 patients with CB, 45 (51%) were embolized. After propensity matching, patients with CB compared to without CB had more pelvic angiography (71% vs 12%, P < 0.001), higher embolization rates (64% vs 22%, P = 0.02) and comparable mortality.

Half of the patients with PF had concomitant co-injuries, including abdominal co-injuries in 17%. Similarly injured geriatric patients had higher mortality. Half of the patients with CB required an embolization.

The incidence of pelvic fractures (PF) has been reported between 17-37/100000 person-years[1,2]. Generally, patients with PF have comprised 2.9%-9.3% of all blunt trauma admissions to level I trauma centers[3-6]. PF with concomitant injuries are on the rise due to an increase of high-energy trauma. In multiple trauma patients, the frequency of PF can rise up to more than 25%, with a corresponding increase in mortality[7,8]. PF are considered a marker of severe injury and therefore require a thorough investigation for the presence of associated injuries[9]. Due to the anatomical proximity of abdominal organs to the pelvic bones, PF are often accompanied by abdominal and bladder injuries, with correspondingly higher mortality[6,10,11]. There is an existing controversy regarding the trauma burden of PF as an indicator of high injury severity or as being an independent risk factor for mortality, even nicknamed “the killing fracture”[4,12-14].

The number of geriatric patients is increasing and their lifestyle is becoming more active, which corresponds to an increase in trauma[15-18]. An increase of the elderly population with age related comorbidities further complicates the management of PF[15,17,19].

The presence of contrast blush (CB) on contrast computed tomography (CT) is considered a sign of active bleeding, however, its clinical implications and association with outcomes continue to be debated[20-24].

The aim of this study was to analyze the characteristics, management and outcomes in polytrauma patients with pelvic factures, with a focus on the geriatric population, the impact of abdominal co-injuries and the clinical value of contrast blush.

This Institutional Review Board approved, retrospective chart review study was granted a waiver of informed consent and included 558 adult (≥ 18 years old) blunt trauma patients with pelvic fractures (AO/Orthopedic Trauma Association fracture Type 61) admitted to an urban level 1 trauma center between January 1, 2017 and January 1, 2023[25,26]. The level 1 trauma center is located in an area with a predominantly geriatric population. Patients with open PF, periprosthetic fractures, pathologic fractures, PF after gunshot wounds and penetrating injuries, isolated acetabular fractures, pregnant women, prisoners and patients who died on arrival were excluded. Analyzed variables included: Age, sex, comorbidities, mechanism of injury (MOI), injury severity score (ISS), Glasgow coma scale (GCS), abbreviated injury scale (AIS) of pelvis, co-injuries, transfusion requirements, CT results, rates of pelvic angiography, embolization, laparotomies, orthopedic pelvic surgery, intensive care unit (ICU) and hospital lengths of stay, Do not resuscitate (DNR) orders, discharge disposition and mortality. Variables were identified via the International Classification of Diseases 9^th and 10^th edition and extracted from the electronic medical records of the patients. Geriatric age was defined as 65 years or older[27].

Because of the retrospective nature of this study we adopted the previously reported methodology of characterizing the severity of PF by AIS for body region rather than classifying PF according to the fracture geometry or vector of force[6,8,9,28]. Previous studies have shown no relationship between the morphology of the PF, categorized by AO or the Young and Burgess classifications, and mortality[29-31]. The flow chart of the study is presented in Figure 1.

Open in New Tab   Full Size Figure   Download Figure

Figure 1  Study flowchart.

Patients with PF comprised 4.1% (558/13655) of all trauma admissions to our center. The mean age of the study cohort was 62 years (range 18-99), half of the patients were geriatric and all had an equal sex distribution. Motor vehicle collision accounted for almost two thirds of MOI and falls for about one third. Mean ISS was 16, mean GSC was 14, with half of the cohort having concomitant injuries. Contrast enhanced CT was done in 448 (80.3%) of patients. Overall mortality was 7.2 %. The general characteristics of all patients with PF are presented in Table 1.

The overall comparison between geriatric and non-geriatric patients is presented in Table 2. Non-geriatric patients had more high impact MOI, higher injury severity, with lower GCS than geriatric patients. Geriatric patients had less CT with contrast and less orthopedic pelvic surgeries, with shorter lengths of stay, and higher mortality. A comparison between geriatric and non-geriatric patients after propensity score matching is presented in Table 3 and Figure 2. After propensity score matching, the pelvic injury characteristics were comparable, but the rates of CT with contrast and pelvic surgeries remained lower in the geriatric group. The hospital lengths of stay were significantly shorter, and DNR orders, as well as mortality remained higher among geriatric patients.

Open in New Tab   Full Size Figure   Download Figure

Figure 2  Mortality in all comparison groups.

Concomitant abdominal injuries were diagnosed in 93 (16.7%) of patients with PF. A general comparison of patients with PF with abdominal co-injuries vs without is presented in Table 4. Patients with abdominal co-injuries were younger, predominantly after a motor vehicle collision, were more severely injured, required more blood transfusions, had CB seen on CT more often, had higher rates of angiography, pelvic surgery, laparotomy, however, had similar rates of embolization, longer hospitalization, and significantly higher mortality. The results of a comparison between patients with and without abdominal co-injuries after propensity score matching are shown in Table 5. After propensity score matching, patients with abdominal co-injuries still had higher ISS, rates of blood transfusions, rates of laparotomy, had CB more often, higher rates of pelvic angiography, but similar rates of embolization and pelvic surgery, however, longer ICU and hospital stay. Mortality continued to be more than three times higher in patients with concomitant abdominal injuries. Distribution of different intra-abdominal injuries is presented in Figures 3 and 4.

Open in New Tab   Full Size Figure   Download Figure

Figure 3  Abdominal co-injuries in 93 patients with pelvic fractures.

Open in New Tab   Full Size Figure   Download Figure

Figure 4  Solid organ and visceral injuries in patients with pelvic fractures and abdominal co-injuries.

An overall comparison between patients with CB vs without CB is presented in Table 6. Patients with CB had more severe pelvic injury, higher rates of angioembolization, more frequent abdominal co-injuries, more pelvic surgeries, longer hospitalization and higher mortality. After propensity score matching, similarly injured patients with CB compared to without CB had more blood transfusions, higher rates of angioembolization and more orthopedic pelvic surgeries; however, lengths of stay and mortality were not statistically different (Table 7). A distribution of propensity matched patients with CB vs without CB by the Young and Burgess classification showed that patients with CB had more often (45.2% vs 20.5%, P = 0.002) unstable fracture patterns (Anteroposterior Compression-2, Anteroposterior Compression-3, Lateral Compression-2, Lateral Compression-3, Vertical Shear).

The incidence of PF in our study is similar to the most frequently reported occurrence, which is around 4%[4,5,9].

In our study, geriatric patients comprised half of the population and were twice the age of non-geriatric subjects, with a correspondingly higher rate of comorbidities, including the use of anticoagulation. As expected, geriatric patients had more low energy trauma, which resulted in lower PF severity, higher GCS and less abdominal co-injuries. Consequently, in the unmatched geriatric patients contrast enhanced CT was done less often, however, the transfusion requirements, as well as the rates of CB, angiography, and embolization were comparable. Even after propensity score matching, which ensured similar pelvic injury severity, we found similar rates of blood transfusion and angioembolization in both age groups, which points toward the same strategic approach, namely that bleeding control should be achieved irrespective of other factors.

Other authors also advocate for early blood transfusions and for a liberal administration of angioembolization in older patients given their physiological changes (reduced physiologic reserve due to many co-existing diseases), anatomical changes (non-constricting atherosclerotic vessels and periosteum that is not densely adhered to the bone) and the prevalence of comorbidities, such as anticoagulation or antiplatelet therapy[32-34].

Orthopedic pelvic surgery was done significantly less often in our geriatric patients. Similarly, Henry et al[32] also found that patients older than 55 underwent operative fracture fixation less often. In a study by Matityahu et al[15] pelvic procedures were performed in 9.4% of geriatric patients (≥ 65 years old) compared to 17.3% in the adult population (18-64 years old), with a gradual increase in pelvic surgeries in patients with more severe PF (AIS ≥ 3) in both age groups.

Overall, in our patients with PF, orthopedic pelvic surgeries were performed in 13.3%, which is comparable to 12.8% reported by Matityahu et al[15]. Lustenberger et al[35] described 21% of orthopedic pelvic surgeries, however, their patients were younger and had higher ISS than our patients. Gabbe et al[28] reported that 32% of patients with PF received neither external fixation nor open reduction and internal fixation, and O’Sullivan et al[29] reported that 48%of patients with PF were treated non-operatively.

Time to pelvic fixation and the duration are modifiable risk factors that affect complications, mortality and long term functional results[36,37]. Therefore, the most beneficial timing and duration of orthopedic interventions in patients with PF continues to be debated[37-39]. The time to definitive fixation has decreased over the last decade, particularly in the hemodynamically unstable patients and in polytrauma cohorts with PF[40]. In our study, the mean time from admission to an orthopedic surgery was 75 h.

We found that the hospital lengths of stay were shorter in our geriatric patients. A shorter hospital stay of elderly patients with PF was also reported by Matityahu et al[15], however, the same difference was not observed in the study by Keil et al[19]. Moreover, younger patients in our study were discharged to home four times more often than the geriatric patients, and this tendency was also observed by Keil et al[19].

Mortality in our study reached 9.6% in the unmatched and 13% in the matched geriatric cohort, which was twice higher than in the non-geriatric patients, and it was primarily due to discharge to hospice and higher frequency of the DNR orders. Higher mortality in the elderly patients, between 12% and 56%, was previously reported by several authors, however geriatric age was defined differently, from > 55 to ≥ 65, and the inclusion criteria varied significantly, from including only severe trauma to only patients with angiography[18,19,29,33,41]. Matityahu et al[15] also reported a higher mortality in the elderly, and particularly in octogenarians with PF, however their patients were not propensity matched by the severity of trauma.

Studies utilizing propensity score matching to analyze the outcomes of PF in the elderly patients are rare[12,16,19]. Our findings support the conclusions of Keil et al[19] who found a significantly higher mortality among propensity matched geriatric than non-geriatric patients with PF after high-energy trauma. It must be noted that in their study the in-hospital mortality resulting from the withdrawal of care was similar in geriatric and non-geriatric patients, while in our study, the discharge to hospice and DNR were significantly higher in geriatric patients.

In other propensity score match studies, Gogna et al[16] did not find an increase in mortality in severely injured geriatric polytrauma patients with PF, but Andrich et al[42] confirmed a higher mortality in subjects with PF aged 60 years or older, however, in both of the studies the propensity match comparison was done to patients without PF.

The conclusion from our findings is that after propensity score matching by the severity of pelvic trauma, and having comparable rates of CB and angioembolization, the orthopedic pelvic surgery rates were lower and mortality among geriatric patients remained significantly higher.

The anatomical proximity of the abdominal organs residing in or near the pelvic cavity inevitably leads to a close association of pelvic bone fractures with abdominal injuries and in turn mandates a thorough examination for the latter. Concomitant abdominal co-injuries in our study were diagnosed in 16.7% of all patients with PF. The reported incidence of intra-abdominal co-injuries in patients with PF ranges from 11% to 20.3% and up to 51%[6,43-46]. Our data are similar to the findings of 16.5% reported by Demetriades et al[6], 16% by Murr et al[10], 19% by Lunsjo et al[31], and 15% by Emerman et al[47]. In a study by Kwon et al[11], abdominal solid organ injures comprised 17.4% of all patients with PF who also had a poorer prognosis. Fu et al[48] emphasized that PF present diagnostic difficulties in detection of intra-abdominal injuries. Other studies demonstrated a notably higher incidence of associated intra-abdominal injuries in patients with severe PF (AIS > 4)[6]. Several authors underline that Malgaigne fractures and bilateral pubic rami fractures have a high incidence of associated intra-abdominal injuries[10,49]. Among our patients, 20% had AIS pelvis score of 4 and above, indicating a severe pelvic fracture. The most commonly injured abdominal structures were the liver, mesentery and spleen. In both, the matched and unmatched analyses of patients with abdominal co-injuries, the transfusion requirements, the rates of laparotomies, ICU and hospital lengths of stay were all higher in patients with concomitant abdominal injuries. The rates of angiography were higher in the abdominal injury groups, but the rates of pelvic interventions, such as embolization and pelvic surgery were similar.

Reported mortality is significantly higher in the presence of intra-abdominal injuries and depending on studied cohorts, usually ranges between 12% and 29%[6,11,48]. Mortality in our patients with abdominal co-injuries was three times higher, between 16%-18%, than in patients without abdominal co-injuries, in both the matched and unmatched comparisons. Kwon et al[11] similarly reported more than three times higher mortality in patients with abdominal solid organ injuries compared to patients with isolated PF (3.1% vs 11.6%). Mortality among our patients with PF with concomitant abdominal injuries was equally distributed between in-hospital mortality and discharge to hospice. Other authors emphasized that mortality is more often related to co-injuries than the PF per se[6,9,11,50]. Eastridge et al[43] also suggested that abdominal co-injury affects patient outcomes more substantially compared to other co-injuries. It was also reported that the odds ratio for mortality (approximately 2.0) after trauma associated with a PF is similar to that posed by an abdominal injury[51].

The clinical value of CB in patients with PF, including the incidence, correlation with angioembolization and association with outcomes continues to be a point of controversy. Some authors question the clinical significance of CB and report no association between CB and embolization rates and no correlation with the length of hospitalization or mortality[20-22,52]. Others support the diagnostic value of CB in patients with pelvic trauma[23,24]. A recent meta-analysis found that contrast extravasation (CE) seen on CT has an acceptable diagnostic accuracy (high sensitivity and specificity) and indicates a severe hemorrhage in patients with PF[53].

In patients with PF who had contrast CT, CB was reported in 5.7%, with 56.9% undergoing angiography and 45.1% undergoing embolization with no significant difference in the outcomes[20]. Juern et al[54] reported that 15% had CE and 37% of CE patients underwent angiography and 23% required embolization. In our study, CB was present in 20% of patients with PF who had contrast CT, and two thirds of them (75.3%) received a pelvic angiogram, however, only half of the patients with CB (45 out of 89, 50.6%) were embolized. A study by Do et al[22] showed that CB on CT is a poor predictor of active bleeding and CE is not a sufficient indicator for pelvic angiography, as the authors reported that 10.3% of patients with PF underwent angiography following pelvic trauma, but CE on CT was seen in 66% of the angiography patients. In a study by Kuo et al[55], 37.3% of patients with pelvic fractures who had a contrast CT underwent angioembolization, however, only 23.4% had CE on CT. In a study by Lustenberger et al[35] 13% of all patients with PF had CE on CT and 9% received angioembolization. These observations demonstrate that not all embolized patients had CE, and vice versa, not all CE required embolization.

Patients with PF and contrast blush did not receive embolization if during the angiography there were no signs of active bleeding and patients were hemodynamically stable. These patients, with CB present on CT but without embolization, were managed with blood transfusions and fluid boluses combined with a stabilization of the pelvic fracture by external fixation or by a binder application. Additionally, in several of these patients CB was described by the radiologist as “small” or “faint”. In addition, a few patients with concomitant abdominal injuries required an urgent laparotomy and did not undergo an angioembolization. For patients without the contrast blush on CT the main reasons for angiography were concerns related to the hemodynamic status and particularly the degree and duration of hypotension and the efficacy of fluid resuscitation, especially in cases without other obvious sources of bleeding. Considerations for prophylactic embolization included: Cases of pseudoaneurysm, artery spasm with temporary hemostasis, truncated branch indicating possible vessel transection. Utilization of newer CT scanners with a better resolution, updated protocols with improved imaging algorithms, ability to provide an objective size and volume of the extravasation add to the accuracy of indications for angioembolization.

In our study, after propensity score matching by pelvic injury severity, age and abdominal co-injuries, there was no difference in the length of hospitalization or mortality related to the presence of CB. We conclude that while CB indeed prompts angioembolization, the presence of CB per se is not associated with worse outcomes.

In our study, stratification by Young and Burgess classification revealed that after propensity score matching patients with CB had unstable fracture patterns more often.

Pelvic fractures occurred in 4% of blunt trauma patients admitted to our trauma center, with half of the patients having concomitant injuries. Similarly injured geriatric patients, having comparable rates of CB and angioembolization, had lower rates of orthopedic pelvic surgeries, but higher mortality than non-geriatric patients. Abdominal co-injuries occurred in one sixth of patients with pelvic fractures, and were associated with a higher mortality. Half of patients with contrast blush required an embolization and in similarly injured patients, the presence of CB was not associated with longer hospitalization or a higher mortality.