Acute management and outcome of multiple trauma patients with pelvic disruptions

Disruptions of the pelvic ring represent 2% to 3% of all fractures. The injury mechanisms most likely to cause a pelvic fracture are high-speed road traffic accidents and falls from high altitude. In multiple-trauma patients, the incidence of pelvic injuries is increased, and pelvic hemorrhage may result from bony bleeding or disruptions of the surrounding perivesical or presacral venous plexus, but also arterial pelvic vessels have been implicated as a significant factor leading to exsanguination and death [1‐3]. The reported mortality rates differ largely from 5% to 50% and are dependent not only on the type of pelvic-ring fracture but also on the severity of associated injuries involving the abdomen, chest, and central nervous system [3‐10]. Although, according to recent literature, a distinct decrease of mortality rate is found in patients with a combination of severe pelvic disruptions and hemodynamic instability or with so-called "complex pelvic injuries," it still remains unacceptably high [10, 11]. Complex pelvic injuries are defined as all pelvic fractures (acetabulum, pelvic ring, and sacrum) with pelvic soft-tissue injuries (that is, open fracture including Morel-Lavallée lesion, disruption of pelvic vessels including retroperitoneal hematoma, and urogenital and neurologic injuries directly related to the pelvic fracture) [1]. For these seriously injured patients, a multitude of scientific investigations and as many management guidelines exist [12‐20]. Most of the latter focus on the patient's time in the emergency department (ED) or trauma room, and examples of actual guidelines are the "Complex Pelvic Fracture Module," as part of the trauma algorithm by Tscherne and Pohlemann (Figure 1) and the ATLS Pelvic Fracture Algorithm (Advanced Trauma Life Support). Both pelvis-specific trauma algorithms recommend early hemorrhage control and massive fluid resuscitation [13, 17]. Thereby, hemorrhage control is achieved through mechanical immobilization of the pelvic ring by external counterpressure with the aid of a pelvic binder, a pelvic C-clamp, or an external fixator. In certain cases, ongoing pelvic hemorrhage may require pelvic packing and/or angioembolization. In contrast to the emergency department or trauma room well-established management guidelines, data on the prehospital management of pelvic fractures and its relation to outcome is rarely reported in the literature. This is rather regrettable because of the upcoming evidence that limiting the amount of fluids given by following a strategy of permissive hypotension during the initial resuscitation period may improve trauma outcomes [21‐23]. Although the German Pelvic Injury Register (PIR) represents the only nationwide database specifically focusing on pelvic trauma, unfortunately, a scarcity exists of records of the prehospital phase, and the trauma room and intensive care unit (ICU) data are fewer. Conversely, the TraumaRegister DGU (TR) includes all these missing data, but all included injuries, even the pelvic fractures, are coded by using the Abbreviated Injury Scale (AIS). A further difficulty is appearing from the fact that, in contrast to the Tile/OTA (Orthopaedic Trauma Association) classification that is used in the PIR, the AIS is not adopted in recent trauma algorithms nor is it in clinical use [24]. We report herein for the first time on matching these two anonymous trauma databases to create an intersection set that benefits from complementary data of each register, respectively. We focus on not-yet-described differences in fluid management for different Tile/OTA types of pelvic-ring fractures in the initial resuscitation period (that is, prehospital phase and time from arrival in the trauma room until admission to the ICU). We would also like to find out whether, apart from the "complex pelvic fractures," the mechanically unstable pelvic-ring fractures type B/C, according Tile's/OTA classification, form a distinct entity that must be considered separately within future pelvic-trauma algorithms.

After exclusion of 18 patients with unfavorable prognosis due to severe traumatic brain injury (AIS head >4), 402 multiple-trauma patients with pelvic disruptions were under investigation. The distribution of the pelvic disruptions revealed, in the majority, pelvic-ring fractures divided into 19.7% type A (n = 79), 29.4% type B (n = 118), and 36.6% type C (n = 147). In addition, 14.3% (n = 58) isolated acetabular fractures and/or fractures of the sacrum were found. Table 1 gives the main characteristics of the investigated patients. No significant differences were noted between the fracture groups in the ratio of blunt injury, age, gender, GCS on scene, time from accident to hospital admission, and time in the trauma room. In contrast, the ISS, the ratio of ISS ≥16, as well as the New-ISS and PTS, revealed significantly more severely injured patients with pelvic-ring fractures types B/C compared with type A. On AIS breakdown, more severe (AIS ≥3) head injuries were related to types A and B. Conversely, thoracic trauma did not have predominance in any of the three types, but types B and C were related to more abdominal and extremity injuries. Complex pelvic injuries were identified in 18.9% (n = 76), and the distribution showed a distinct shift to the rotationally and/or translationally unstable pelvic-ring fractures types B/C (that is, 7.9% type A (n = 6), 19.7% type B (n = 15), 67.1% type C (n = 51), and 5.3% isolated acetabular fractures and/or fractures of the sacrum (n = 4). With increasing pelvic-ring instability, the incidence of complex pelvic injuries increased to 7.6% for type A, 12.7% for type B, and 34.7% for type C. The overall mortality was 7.5% (n = 30). Of the pelvic-ring subgroups, type C fractures showed the highest mortality, but this was not statistically significant. In the non-complex-injury group, a 5.5% mortality rate was found compared with 15.8% of the patients with complex pelvic injuries (Figure 2). Of the 30 patients who died, 53% had sustained type C; 27.7%, type B; 13.3%, type A; and 6.7%, isolated acetabular and/or sacral fractures. Therefore, approximately 80% of the patients who died sustained a type B/C pelvic ring fracture.

By matching for the first time the German Pelvic Injury Register with the TraumaRegister DGU, statements are feasible about the recent practice of initial fluid management for different Tile/OTA types of pelvic-ring fractures as well as about the patient's posttraumatic course, including ICU data. Because of the innovative idea of interrelating two trauma registers, a stand-alone technical-notes manuscript with detailed descriptions of the matching process and its weakness in method will be published separately. Nevertheless, the literature contains a multitude of studies about the management of fluid resuscitation of trauma patients, and even a Cochrane review exists [21]. Some studies focus on the prehospital setting [23, 29‐31], some on the emergency department [3, 5, 13, 17, 20], and some on the time in the operating theatre [22]. Starting with the innovative resuscitation study of Bickell et al. [29] from 1994, their "immediate resuscitation group" received means of 870 ± 667 ml Ringers acetate before arrival at the hospital and 1,608 ± 1,201 ml in the ED. Today, 15 years later, the standard regimen in fluid resuscitation in our study has doubled [29]. Certainly, we must keep in mind that Bickell et al. investigated penetrating torso injuries and not multiple-trauma patients with pelvic disruptions. Looking for similar populations, Gruen et al. [13] studied 312 patients with pelvic-ring fractures, type B/C with shock, receiving 5,900 ± 4,000 ml of crystalloids given in the ED. This amount of fluid is even more than we saw in the selective group of type C fractures (3,587 ± 2,565 ml). In contrast, the ratio of our patients in shock on hospital admission was only about 20%, explaining the less fluid given by our study. More recent studies described the administration of 4,271 ± 2,428 ml and 2,750 ml of fluids given in the ED. Thereby, Verbeek et al. [17] also analyzed hemodynamically unstable pelvic fractures, and Giannoudis et al. [3] even represented the median value of the nonsurviving group [3, 17]. In 2011, Hussmann et al. [23] showed that increasingly, preclinical volume led to a slight elevation of lethality as well as of transfused packed red blood cells concentrates (PRBCs) in multiply injured patients after severe abdominal and pelvic trauma, and recommended for both a moderate prehospital volume replacement.

In summary, despite the well-known problem of comparability of all studies in the field of multiple-trauma patients, the large fluid volumes of our study confirm the actuality of traditional pelvis-specific trauma algorithms [14, 18]. Low-volume resuscitation seems not yet to be accepted in practice in managing this special patient entity. Regarding the bleeding risk for different Tile/OTA types of pelvic-ring fractures, our fluid volumes in the prehospital phase suggest that type B and C fractures are different from type A fractures, whereas in the hospital setting type A and B are rather similar and type C is different. A possible explanation could be the volume triggered and synchronized changes of the shock ratio with no significant differences in the pre-hospital phase that turns after ED admission into significant more type C fractures with haemorrhagic shock. Notably, the type C group was related to the presence of a higher percentage of severe concomitant injuries (AIS ≥3) that can contribute to increased hemorrhage (that is, abdomen and extremities). Conversely, the hemorrhagic shock cannot be entirely attributed to the pelvic fracture alone, as has been stated by other authors [8‐10, 25, 32, 33]. A potential bias in treatment might exist once the diagnosis is known, as type C fractures are considered a more severe injury, and the patient receives more fluids than the fluids being administered in response to resuscitation. Nevertheless, bleeding and resuscitation reflect a dynamic process. Looking at the total initial resuscitation period, the infusion volume as well as the transfused PRBCs of the type B and C fractures are different from type A fractures. In accord with others, in our study, increased pelvic-ring instability resulted in increased injury severity (that is, ISS, New-ISS, and PTS), and thus elevated mortality [2, 5]. The highest mortality in the group of the complex pelvic injuries can be explained by its dominance of type B/C fractures, with significantly higher overall trauma load and injury severity [1, 2, 8, 10]. Unquestionably, the pelvic fracture is only one variable among many (that is, age, shock, head injury, abdominal or chest injury, and extremity injury, which contribute to mortality risk [8‐10, 25, 34, 35]). Regarding the ICU data, the overall rates of 5.2% of sepsis and 25.8% of MODS we found are quite similar to those reported in the literature [2, 20, 23, 32]. The same is true for the ventilated days and the length of ICU stay; both variants are shorter compared with those reported in older literature [2, 13, 17, 19, 23]. Analogous to the mortality, the outstanding position of the type C fractures in sepsis, ventilation, and ICU stay is beyond doubt influenced by the concomitant abdominal and extremity injuries with more pronounced shock in the initial resuscitation phase, as well as the higher overall trauma load and injury severity compared with the other subtypes [8‐10, 25, 32, 33]. Notably, multiple-trauma patients are at high risk for developing MODS and/or sepsis [34‐36]. This also applies for the type A fractures that appeared in about three fourths of patients with an ISS ≥16. The higher rate of sepsis and MODS of type A in comparison to type B fractures might be the consequence of the exclusion of the prognostic unfavorable AIS head > 4 population before starting the analysis meanwhile, almost all of the AIS head > 4 population had type B/C fractures. No doubt certain limitations exist in our study. First, weakness is found in the method of interrelating two trauma registers. In-depth analysis of that will be presented in the future in a separate article. In brief, in the technical notes manuscript, the degree of data validation of twofold documented records (that is, ISS, systolic blood pressure, and hemoglobin on ED arrival, as well as mortality will be analyzed). Second, the matching of PIR and TR data resulted in a respectable study population of 402 multiple-trauma patients with pelvic disruptions. The majority of the hospitals that contributed to the two data registers were Level I trauma centers with additional specialization in pelvic trauma. Thus, the experience in managing pelvic injuries is biased and might explain the low mortality rate of 7.5% compared with that in the literature [5, 7, 9, 17, 23]. Another point is the incompleteness of the assessed parameters reflected by the differing numbers of patients for different parameters. Although data-entry errors rates for unknown or missing information are familiar problems in medical registries, the error rates in our study did not differ from the range of 19% to 76% out of the literature [37].

The present study confirms the actuality of traditional trauma algorithms with initial massive fluid resuscitation in the recent therapy of multiple-trauma patients with pelvic disruptions. Low-volume resuscitation seems not yet accepted in practice in managing this special patient entity. Mechanically unstable pelvic-ring fractures type B/C (according to Tile/OTA classification) form a distinct entity that must be considered in future trauma algorithms.