The clinical characteristics and management of paediatric pelvic fractures: a changing landscape based on skeletal maturity

Paediatric pelvic fractures (PPFs) are relatively uncommon, with a reported incidence ranging from 2 to 7.5% [1‐4]. While the mechanism of injury varies, PPFs most commonly result from high-energy trauma, primarily due to motor vehicle accidents (MVAs) [5, 6]. Even in the category of MVAs, adult and paediatric pelvic fracture patterns tend to vary because of different mechanisms of injury (MOI).

Due to the forceful nature of these accidents, additional injuries often accompany pelvic fractures, with an average of five concomitant injuries being present [7]. In particular, in paediatric cases, injuries to the extremities [8] and genitourinary system [9] are common. The latter is attributed to the relative elasticity of the paediatric pelvis, which results in insufficient protection of the underlying viscera [10]. In contrast to adult cases, vascular haemorrhage in paediatric pelvic fractures is rare [8, 11]. Possible reasons include more effective vasoconstriction in the smaller vessels present in children, and the relative elasticity of the pelvic joints in paediatric patients, which makes single bone fractures more common—such fractures are associated with a decreased risk of haemorrhage from pelvic vessels [12, 13].

Reported mortality rates from pelvic injuries vary widely in this patient group, with reported rates being as low as 0.55% [7] to as high as 20% [14]. Chances of survival depend on the number and extent of other injuries sustained, with the presence of head and intrapelvic injuries increasing mortality risk [4, 15]. With regards to blunt trauma in children, central nervous system (CNS) injury is an important factor, and is associated with a tenfold increase in mortality risk compared to those without CNS injury [4]. However, there is a lack of data looking at mortality risk with particular reference to paediatric pelvic fractures, as opposed to general blunt trauma.

Overall, there is discordance in the literature on several findings reported on pelvic fractures in this age group. These include different findings on factors such as mortality rate, prevalence of fracture type, and the types of additional injuries associated with worse prognoses. The aim of this study is to perform a 5-years retrospective analysis of patients with paediatric pelvic fractures, looking at the factors discussed above, who were treated at a level one trauma centre. Factors assessed included general characteristics of the fracture, such as pelvic fracture type as confirmed by imaging, as well as additional injuries and mortality rate.

After receiving approval from our institutional review board, the Trauma Audit and Research Network database, radiology database, and the local department database were retrospectively reviewed between January 2015 and January 2020 for all children admitted with a pelvic fracture.

To be included, patients were under 18 years of age, acquired a pelvic fracture from a blunt-force mechanism, and had adequate pelvic imaging for review. Those with firearm injuries were excluded. Data obtained from the retrospective review of complete medical records and radiographs were divided into four categories:

MOI was classified as a motor vehicle (MV) versus MV, MV versus pedestrian, slipped, fall from height, and others. Associated orthopaedic injuries included all non-pelvic bony fractures excluding facial and skull fractures. Associated non-orthopaedic injuries included abdominal trauma (perforations, lacerations, ruptures of liver or spleen or bowel, abdominal aorta intima tear), chest trauma (pneumothorax, contusions, hemopneumothorax, lacerations), head injury (skull fracture, haematoma, intraventricular haemorrhage, pneumocephalus, cerebellum injury), and spinal injury (lamina fracture, facet dislocation, interspinous ligament laceration, transverse process fracture). The probability of survival (ps) was calculated using an online tool provided by the Trauma Audit and Research Network (TARN) [16]. A pelvic AIS ≥ 4 score was given for fractures with significant displacement with accompanying vascular damage and retroperitoneal haemorrhage.

All radiographs, including computerised tomography (CT) scans if available, were reviewed by two authors to classify pelvic fractures according to that proposed by Torode and Zeig [17]. CT scans were performed at the discretion of the consultant paediatric orthopaedic surgeon, for patients with clinical features of an unstable pelvis and persistent abdominal pain. The Torode classification divides pelvic fractures into four groups: (I) avulsion fractures, (II) iliac wing fractures, (III) simple ring fracture (stable), (IV) ring disruption (unstable) and pelvic fractures combined with acetabular fractures. Fracture stability was assessed by clinical and radiographical methodologies. Stable fractures had ≤ 2 mm fracture displacement on radiographs, were stable to pelvic compression, had no hip dislocations and no combined pelvic and acetabular fractures. The Torode classification was modified by Shore et al. who subdivided type III fractures into ‘A’ and ‘B’ [18]. The former (IIIa) is a stable anterior ring fracture, and the latter (IIIb) is a stable anterior and posterior ring fracture. The modified Torode PPF classification is predictive of morbidity and death in the multi-trauma setting [18], and was used in this study. Patients were divided into two groups based on skeletal maturity, which was assessed based on the status of the triradiate cartilage (open: immature; closed: mature).

Statistical analysis was performed using IBM SPSS Statistics version 27. Differences between the skeletally mature and immature cohorts were analysed. Categorical data were analysed by the chi-squared test or Fischer’s exact test if an expected value was below five. The Mann–Whitney U test was used for non-parametric continuous variables. Univariable analysis was performed to analyse the risk factors for outcomes of interest, from which the odds ratio of the outcome occurring for a given risk factor, 95% confidence interval (CI), and p value were obtained. Statistical significance was defined as p ≤ 0.05, and shown in bold in the results tables.

Pelvic fractures are usually a consequence of high-energy trauma and are uncommon amongst the paediatric population. In this single-centre study over a 5-years period, PPFs corresponded to 2.8% of our paediatric trauma admissions. The literature only provides a vague indication of the incidence of PPFs [20]. Our incidence rate is lower than a recent study over a 14-years time period which suggested 3.5% [5] but greater than an earlier study from a large paediatric hospital which suggested a rate of 10 PPFs per year [21]. The actual rate is likely to be higher since one study reported that 93% of post-mortem assessments of paediatric patients that died from blunt trauma had pelvic polyfractures [22]. The main findings of this study include the predominance of motor vehicle accidents as an MOI, male gender, and non-pelvic skeletal fractures being the most commonly encountered associated injury. Compared to their skeletally mature counterparts, skeletally immature children were more likely to be hit by a motor vehicle as a pedestrian, be intubated, acquire Torode type II and rami fractures, and receive chest and head injuries. Skeletally immature children were also less likely to acquire Torode type IV fractures, receive surgical treatment for their pelvic injuries, and had a faster time to FWB. Furthermore, the mortality rate, GCS score, and average probability of survival did not significantly differ depending on skeletal maturity or severity of fracture.

Parameters that can assess skeletal maturity include the status of the triradiate cartilage, Risser classification, greater trochanteric apophysis closure, and proximal femoral apophysis closure. Since this is a study on paediatric pelvic fracture, a parameter based on the pelvis was deemed most appropriate. The Risser classification has limitations that complicates accurate staging, such as failing to take into account anomalous ossification of the iliac apophysis [23]. This can be overcome by bending anteroposterior (AP) pelvic films, but may not be suitable for the younger paediatric cohort [24]. Recent studies have suggested dividing Risser stage 0 into two groups based on triradiate cartilage status [25]. Hence, we deemed an assessment of the triradiate cartilage on radiographs to be the most reliable for determining skeletal maturity. The age at which the triradiate cartilage closes is clinically important because the mechanical properties of the pelvis changes, namely a loss in elasticity with skeletal maturity [18]. There is a general consensus amongst studies that this is 14 for males and 12 for females [26]. The transition point could distinguish paediatric from adult patients, given the different treatment choices and outcomes between the two cohorts. Skeletally mature patients were more likely to sustain Torode type IV fractures (p = 0.018) and less likely to acquire type II fractures (p = 0.047). Importantly, skeletally mature patients were more likely to receive operative treatment for both pelvic (p = 0.036) and non-pelvic injuries (p = 0.652). Karunakar et al. suggested that skeletally mature patients with unstable pelvic and acetabular fractures should be managed in the same manner as adults, namely by operative fixation [27]. Even in immature patients, Karunakar et al. showed successful operative management, but the clinical indications are less clear, perhaps warranted for pelvic fractures that distort the skeletally immature pelvis [27]. The conservative treatment of fifteen skeletally immature patients with unstable pelvic fractures by McDonald et al. led to limb length discrepancy, triradiate and sacroiliac growth arrests [28]. Indeed, patients in our cohort with type IV fractures were more likely to receive operative treatment (p = 0.261). However, reports of pelvic growth arrest and bony remodelling after the operative treatment of skeletally immature patients [5], and the loss of pelvic elasticity in females affecting future pregnancies may warrant future research into the optimal treatment modality.

MVAs accounted for the majority of injuries in our cohort (76.9%), agreeing with earlier studies [6, 29, 30]. Studies on adult pelvic fractures have also reported MVAs to be the most common MOI [26, 31]. Despite the high proportion of MVAs, none resulted in mortality in our cohort. Early studies have suggested a mortality rate as high as 25% [32], a figure higher than the 17.5% mortality rate reported for adult patients [33, 34], and suggested a significant number suffered from pelvic haemorrhage which mirrors the situation in adults. Yet our cohort and other recent studies disprove that. Adult pelvic fractures usually involve an ‘open book’ injury type created by an AP-directed force, leading to increased pelvic volume and fatal pelvic haemorrhage [26]. On the contrary, especially with the younger skeletally immature cohort who are more likely to be pedestrian struck (p < 0.001), a lateral compression force is experienced which is less likely to lead to pelvic haemorrhage [35]. Nearly half our cohort (47.7%) suffered from rami fractures, which was seen more frequently in the skeletally immature group (p = 0.037), indicating a lateral compression type of injury. Saglam et al. reported a ten-fold greater incidence of pedestrian crashes than MV crashes alone [36]. Furthermore, paediatric blood vessels have decreased vascular stiffness, increased endothelial function, and greater vasoactive properties, allowing a more robust vasoconstrictive response compared to their adult counterparts [37].

Despite the mortality rate in our cohort (1.5%) being lower than that in the recent literature [5, 6, 11, 30, 31, 35, 37], our mean ISS score (28.9) and proportion of Torode type IV fractures (18.5%) was not significantly different from recent studies [5, 11, 18, 35]. Nevertheless, Shaath et al. had 44.9% type IV fractures but only a 6% death rate [30]. The rate of paediatric unstable pelvic fractures (type IV) is lower than the adult population [3, 38], which could be due to the increased likelihood of the adult pelvis to break in multiple locations like a ‘pretzel’ [29]. Compared to bone and surrounding ligaments, the intrinsic weakness of the epiphyseal and apophyseal cartilage in children increases the chance of an avulsion fracture through the growth plate, which could lead to limb length discrepancy [39]. They have a high capacity for energy absorption and when not displaced, the damage can be hard to pick up on radiographs, with some suggesting that a minimum follow-up time of one year is needed for diagnosis [40].

Whilst PPFs per se are not necessarily life-threatening, they are indicative of serious trauma. With the greater flexibility of sacroiliac joints and malleability of the paediatric skeleton [35], greater forces are absorbed for a given pelvic fracture compared to adults. Hence, PPFs are often accompanied by concomitant injuries which can lead to mortality or long-term morbidity if not identified and treated promptly [41]. Mortality in PPF patients can be associated with many factors, such as decreased GCS, higher ISS scores, longer time spent in ICU, and a higher head AIS score. The only patient that died in our study suffered a traumatic brain injury. Nevertheless, no relationship between Torode grade and mortality was found (p = 0.718), agreeing with previous studies [5, 31, 33]. Other than additional skeletal injuries, chest and head injuries were the most prevalent, identified in 38.5% and 36.9%, respectively. This is in line with earlier studies that noted an increased involvement of the upper trunk, especially with more severe injuries [6, 31, 42, 43]. Chest injuries however occur less frequently than in the adult population, perhaps due to the greater compliance of a child’s chest wall [44].

Studies have pointed out that patients with intra-abdominal injuries, especially hollow viscus perforations, are sometimes missed, with difficulty to interpret signs in a traumatised child [3, 45]. This is concerning given that the clinical course in patients with concomitant pelvic/abdominal injuries is significantly longer [46], and the poorly developed abdominal wall musculature in children leads to a higher chance of injury than adults [44]. Our cohort had a 16.9% incidence rate of intra-abdominal injuries, which did not significantly differ between the skeletally mature and immature cohorts. Bond et al. reported a 20% rate, which could be due to their exclusion of Torode type I avulsion fractures [47]. Risk factor analysis for associated abdominal injuries revealed an important relationship with pelvis AIS score ≥ 4 (OR 5.3; 95% CI 1.3–22.6; p = 0.023), which agrees with the findings of Demetriades et al. [3]. Regarding non-pelvic skeletal injuries, lower limb fractures predominated (40%), with the incidence of femoral fractures equivalent to the incidence of upper limb fractures (24.6%). Intriguingly, the rate of femoral fractures is higher in the skeletally immature cohort and in those with a higher Torode grade. This could be due to the predominant MOI of skeletally immature patients, which is being pedestrian struck.

Our centre is a specialised adult and paediatric trauma centre, with the resources to prioritise treatment of critical trauma at risk of mortality. The main limitation of this study is the retrospective design and low patient numbers. The scarcity of these cases emphasises the need for multicentre studies or the establishment of national networks that deal with paediatric pelvic trauma management. The inclusion of patients 18 years or under was based on the World Health Organisation’s cut-off age for paediatric patients, despite evidence suggesting that there cannot be a single cut-off point for when a paediatric patient becomes an adult patient, but rather a gradual transition depending on multiple factors, such as skeletal maturity. The follow-up times varied widely, meaning that no conclusions can be drawn from this study about long-term results. This is important to investigate in future studies, especially in cohorts with a large group of skeletally immature patients, since the pelvis can malunite during growth [48]. Furthermore, studies could also incorporate a health economics analysis, to provide a more efficient specialist paediatric pelvic trauma service in tertiary centres.

PPFs are a reliable hallmark of severe trauma in children and our study suggests the importance of early identification of associated trauma such as abdominal injuries which can be overlooked. A clinically useful classification system such as the modified Torode classification should be used to differentiate patients according to the severity of trauma, guide clinician’s understanding of the disease process, and aid treatment decisions. Patients with closed triradiate cartilage can be treated like adults, as they have similar MOIs, need for operative management, and fracture patterns. Those with open triradiate cartilage are unique, with significant differences in terms of MOI, management strategy, fracture severity and patterns. With improved intensive care and trauma management, mortality from pelvic fractures can be kept at a low level.