Of all skeletal injuries, approximately 3% consist of pelvic fractures1). Especially in high energy trauma (HET), disruption of the pelvic ring occurs in approximately 13-17% of all pelvic fractures2). High forces are required to disrupt the pelvic ring and the association with other significant injuries is therefore high3). Mortality is reported to be as high as 10-16% in this patient group4). When there is an open fracture, mortality as frequent as 50% has been reported5). Significant hemorrhage is an often encountered complication seen in unstable pelvic ring injuries678). Because early exsanguination is the major cause of death in these patients, early identification of the bleeding source and rapid stabilization of the fracture are therefore critical in their survival9).

The exact treatment pathway of hemodynamically unstable patients with an unstable pelvic ring fracture remains controversial and is often dictated by available hospital facilities. Mortality and functional outcome of pelvic ring injuries are determined by several factors, especially concomitant injuries. The major factors influencing chronic disablities are urogenital and neurological injuries1011).

This study was conducted to evaluate the patient profile, fracture characteristics, treatment and outcome of a large patient group with various pelvic ring injuries in our clinic.

A total of 537 patients with a pelvic fracture were admitted to our hospital (Radboud University Medical Centre in Nijmegen, the Netherlands) in the period January 1, 2004 to January 1, 2015. Our hospital is a level 1 trauma center and a supra-regional referral center for pelvic and acetabular fractures in the Netherlands. Also, it is the main referral center for the Dutch Helicopter Emergency Medical Service (HEMS) for the middle, east and south part of the Netherlands. After the approval from Institutional Review Board of Radboud University Medical Centre, all data were analyzed from the electronic patient charts and the trauma database. Data analyzed included patient demographics, date, cause and type of the accident, HEMS involvement, transfusion requirement, associated injuries, abbreviated injury scale for each body region, injury severity score (ISS), treatment in the emergency room (ER), length of intensive care unit (ICU) and hospital stay, complications and mortality. The pelvic fractures were described according to Tile's classification12). In our center, each patient with a pelvic fracture is routinely evaluated with a computed tomography scan. The specific fracture sites were also described. If operative therapy was performed, type of fixation, secondary operations and operation time were recorded.

Data of patients with an isolated acetabular fracture were excluded as well as data of patients who were referred for non-, or malunions of pelvic fractures.

During this 10-year period no significant changes in management or diagnostic work-up of pelvic fractures were made.

Five hundred and thiry-seven patients had suffered pelvic fractures. There were 328 (61.1%) men and 209 (38.9%) woman. Mean age was 44 years (standard deviation, 18; range, 2-100 years). Age didn't differ significantly between men and women. There were 45 pediatric patients in the range from 2 to 17 years. The HEMS was involved on the place of accident in 200 patients (37.2%).

The patient data describe a consecutive patient series over a period of 10 years in a level 1 trauma center in the Netherlands and describes our experiances with pelvic fractures.

As shown by other authors114), pelvic fractures are predominantly seen in men less than 50 years of age and are involved in a HET of which most of them in a motor vehicle accident. These findings are similar to our patient group.

An explanation for the majority of HEMS involvement after 2006 can be explained by the fact that operating hours were extended to 24 hours a day and the criteria for HEMS involvement were broadened.

Tile A fractures are thought to be relatively benign because the stability of the pelvic ring remains intact and the risk of severe hemorrhage from the venous plexus is considered low15). However, in our group of patients with a Tile A fracture, almost 80% had concomitant injuries and had a relatively high ISS with a mean of 22. Also, almost 40% was hemodynamically unstable. Possible explanations can be that most patients with a simple fracture of the pubic ramus are not admitted and can be treated in the out-patient clinic with analgesics and physical therapy and therefore are not included in our analysis. This can bias our results. Also, in patients with an isolated iliac wing fractures, high energy impact is necessary to cause this type of fracture and a lot of concomitant injuries are seen in these patients. Therefore, this group of patients have similar patient profiles and outcomes as patients with unstable fractures have.

In the patient group with a Tile B-type fracture, also a high ISS was seen. Prehospitally, in only 31 patients (14.7%) a pelvic stabilization device was used. This might be related to the limited use of pelvic compression devices by the emergency medical team, unclear findings in physical examination or insufficient documentation of the paramedics or the ER department. Experience nowadays in our clinic is that in patients with the slightest suspicion of a pelvic fracture, a pelvic stabilization device is used. This subject is currently studied in a new prospective case series.

Early total care was done in 50.8% of the patients with an intrinsic unstable fracture with a shock class of 2 or lower. Most of these patients were treated with symphyseal plating alone. In hemodynamically unstable patients, damage control principles were followed; either an external fixator was placed or a symphyseal plate was used. When the patient was stabilized, a secondary procedure was performed with placing of SI screws. Plating of the symphysis is a fast and easy procedure and if the patient and the fracture allow it, we prefer this over the external fixator. The advantage of this approach is direct and often better reduction of the anterior ring without the complications of external fixation devices, such as pintract infections or non-adequate placement of pins. The posterior pelvis was often stabilized in a second procedure.

The patients with the Tile B2 type fracture were treated conservatively, because we believe that this type of fracture is intrinsic stable. However, one patient with a B2 fracture who could not mobilize due to extensive pain in the SI joint were stabilized later with percutaneous SI screws, with good result.

The majority of our patients had a Tile C fracture. Ninety percent of patients had concomitant injuries; therefore, the mean ISS was high. As expected the rate of major complications and mortality was the highest in the type C group. However, the number of urogenital and neurologic complications was considerably low, especially considering the amount of complex pelvic fracture patients included in this study.

Most complications seen were due to late effects of hemorrhage (systemic inflammatory response syndrome) and due to thoracic trauma (acute respiratory distress syndrome, pneumonia). Patients who had an open fracture suffered mostly from infectious complications.

The number of complications reported here is suspected to be higher than documented. This is especially true for the minor complications like electrolyte disturbances and urinary tract infections. Major complications requiring operative therapy or ICU admittance are well documented and are therefore believed to be accurate.

Mortality in our total group was 13.6%, which is acceptable in comparison with other studies141116). For example, Rommens and Hessmann16) report a mortality rate of 5% in the group with Tile B fractures and 14.8% in Tile C fractures. In this group also a lot of concomitant injuries were seen. Death due to hemorrhage shock was seen in 9 patients (40%) compared with 10 patients (52.6%) in our group. Mortality over the past 10 years didn't change. However, this study can be the basis for evaluating our protocol in order to reduce mortality even further.

Approximately 47% of all patients could be discharged to their homes with physical therapy and no further need for clinical rehabilitation. Outcome was therefore considered generally good. However, no functional outcome scores were measured during follow-up. Information regarding walking, pain and return to former jobs could only be found in a select number of patient-charts, so no conclusions can be made of this retrospective data. Evaluation of functional outcome with the 36-Item Short Form Survey (SF-36) and the Majeed questionnaire17) is currently being studied in a prospective trial.

This study, conducted in a trauma center in the Netherlands which is dedicated to the treatment of pelvic ring injuries, shows a high incidence of pelvic fractures. To our knowledge, this is the largest reported single centre experience in the Netherlands. Patients with a Tile C fracture are more severely injured and require more resources. Mortality doesn't differ between the 3 fracture types according to Tile. This is suggestive that mortality is caused by a variety of other injuries such as neurotrauma. The overall outcome is considered good with almost half of the patients can be discharged to their homes.

This data is an important tool to evaluate our treatment protocol and thereby improving outcome.