9th March 2018, Volume 131 Number 1471

Mohit Bajaj, Giorgio Stefanutti, Haemish Crawford, Vipul Upadhyay

Trauma is a leading cause of mortality in the paediatric patient population.1 Pelvic fractures are uncommon in children, but can occur as a result of high-energy blunt trauma, such as vehicular injury. Pelvic fractures constitute between 0.3 and 4% of all paediatric injuries, with a mortality rate up to 25%.2 

The presence of a pelvic fracture is often a marker of severe trauma and should alert the clinician to actively exclude and treat associated life-threatening soft-tissue injuries. The severity of associated injuries and complications, especially those of the central nervous system, which accompany pelvic fracture, result in greater morbidity and mortality than the pelvic injury itself.3–5

Objectives

This study aims to review the experience with pelvic fractures at Starship Children’s Hospital and demonstrate its role as a marker of severe trauma. We compare the mechanisms of injury, clinical presentation, management and outcomes of children presenting with pelvic fractures in New Zealand with other studies published in English language literature.

Patients and methods

Study design

Starship Children’s Hospital provides the largest paediatric surgical and trauma service in New Zealand, with an average of 1,000 trauma admissions annually over the past decade. A retrospective review of children with pelvic fractures managed at our institution in the 20-year period between July 1995 and May 2015 was performed. Patients were identified from the hospital coding and trauma registry. The search identified 179 consecutive children admitted with a pelvic fracture.

Variables

Data fields collected included patient details, mechanisms of injury, investigations performed, length of hospital stay, management and complications. Data was also collected on Injury Severity Score (ISS),6 Glasgow coma scale (GCS),7 transfusion requirements and details of associated injuries (both orthopaedic and non-orthopaedic).

Associated orthopaedic injuries were defined as all bony fractures excluding skull, rib and facial fractures. Associated non-orthopaedic soft tissue and visceral injuries were sub-divided into major chest trauma (haemothorax, pneumothorax and lung contusion), closed head trauma (traumatic brain injury, intra-cerebral haemorrhage, cranial and all facial fractures), abdominal injuries (solid organ injury and viscous perforation) and urogenital trauma (bladder, urethral or vaginal injury).

Radiologist reports of all plain radiographs, including computed tomography scans when available, were used to classify each pelvic fracture according to the system proposed by Torode and Zieg (Figure 1).8,9 This classification, based exclusively on plain radiography, remains the best and most widely used system for describing paediatric pelvic fractures.10 It considers both anatomic and mechanical factors, relating mechanism of injury to severity of deformity. With the increased diagnostic use of cross-sectional imaging, modifications to the original Torode and Zieg classification have been proposed and shown to be predictive for significant morbidity and mortality in the setting of trauma.1

Figure 1: The Torode and Zieg classification of pelvic fractures. 

c 

Arrows show the direction of force. A, Type I: avulsion fractures. B, Type II: iliac wing fractures. C, Type III: simple ring fractures. D, Type IV: ring disruption fractures. Isolated acetabular fractures were added to the Type III group for completeness, as first proposed by Silber et al.10 Reproduced from Tachdjian's Pediatric Orthopedics9 with permission. 

Statistical methods

All collected data was recorded on an Excel 2010 spreadsheet (Microsoft Companies, Richmond, USA). The distribution of descriptive variables was expressed as median with inter-quartile range (IQR). All analyses were conducted using the Statistical Analysis System (SAS) software suite (North Carolina, USA). A p-value <0.05 was used to denote statistical significance.

The Chi-squared test was used to determine differences in proportion between independently distributed datasets. This included comparison of differences in patients’ sex, side of pelvic fracture and whether patient weight predicted likely mechanism of injury. A Spearman’s correlation test was used to measure the strength of the monotonic relationship between pelvic fracture complexity and the ISS. A Wilcoxon test was used to compare non-parametric matched samples, including the difference in ISS values between cases with mortality and the rest.

Results

Participants

Over the study period, 179 patients were treated at Starship Hospital with a pelvic fracture. There were 116 boys (65%) and 63 girls (35%), this difference in sex being significant (Chi-square test p<0.0001). The median age of the children was eight years (IQR, 5–12 years).

At presentation, 62 children (34.6%) had a depressed level of consciousness with a GCS less than 15. Thirty-four children (19%) had a severe traumatic brain injury with a GCS less than 9. The median ISS was 9 (IQR 4–22).

The median hospital stay was five days (IQR 2–12). Sixty-four patients (35.8%) required admission to the intensive care unit, with their stay lasting a median of three days (IQR 1–6). Forty-four children (24.5%) required a packed red blood cell transfusion during their stay, with a median requirement of two units (IQR 2–5 units). All children that required a transfusion had one or more associated injuries.

Mechanism of injury

The most common mechanism of injury was a pedestrian struck by a motor vehicle in 46% (n=83) of cases. Passengers in motor vehicle accidents accounted for 23% (n=41) of presentations and this was the second most common injury mechanism, followed by falls in 13% (n=24) and bicycle-motor vehicle collisions in 10% (n=17). Eight children were injured playing sport (4.5%). A further six (3%) had miscellaneous mechanisms of injury, including two cases of non-accidental injury.

In the pedestrian struck by motor vehicle group, nine children were injured by a motor vehicle in a residential driveway. In children 10 years or younger, this mechanism accounted for 54% of cases (64/117), reducing to only 30% beyond the first decade of life (19/62). This difference was statistically significant (Chi Square test, p<0.0021).

Twenty-two (26%) of the 83 children struck by motor vehicles had bilateral pelvic injuries. Additionally, there was a predominance of right-sided pelvic injuries (41%, n=34) versus left-sided injuries (32%, n=27), but this did not achieve statistical significance (Chi-square test Test p=0.0523).

The mechanisms involved leading to pelvic injuries are remarkably consistent across the world, with motor vehicle-related injuries being most common in all studies4,10,12–17 (Table 1). We further compared the mechanism of injury prevalence between left-hand drive and right-hand drive countries. Interestingly, we note that pedestrian-motor vehicle injuries occurred more frequently in left hand driven countries, while passengers injured in motor vehicles were more frequent in right-hand driven countries.

Table 1: Prevalence of mechanisms of injury reported in the literature (%).

c

Fracture classification

Almost half of the fractures (47%, n=84) in the study were of the Torode and Zeig Type III classification (single ring fractures or isolated acetabular fractures). Type IV fractures (ring disruption fractures) were the second most common at 32% (n=57).

The median injury severity score (ISS) for each fracture type was determined; for Type I fractures it was 4 (IQR 4–9), the median ISS for Type II fractures was 9 (IQR 4–20), 6 (IQR 4-17.5) for Type III fractures and 13 (IQR 9–29) for Type IV fractures (Table 2). An increase in the complexity of the fracture type in the Torode and Zeig classification correlated with an increase in the ISS (Spearman correlation coefficient = 0.625, p=0.0006).

Table 2: Injury Severity Score (ISS) of pelvic fracture types and associated injuries.

Torode/Zieg # type

I

II

III

IV

IV

Sum

Median ISS (IQR)

4.0 (4–9)

9.0 (4–20)

6.0 (4–17.5)

13.0 (9–29)

13.0 (9–29)

-

Number of patients

5

33

84

57

57

179

Closed head injury

-

9

24

22

22

55 (31%)

Thoracic injury

-

12

22

25

25

59 (33%)

Abdominal injury

-

10

15

22

22

47 (26%)

Genitourinary injury

-

-

6

9

9

15 (8%)

Other orthopaedic fractures

-

9

35

31

31

75 (42%)

Patients with associated injuries (%)

-

57.5

63

88

88

122 (68%) 

Management

Associated injuries occurred in 68% (n=122) of patients. Other orthopaedic fractures (42%, n=75) and thoracic injuries (33%, n=59) were most common, followed by injuries to the head (31%, n=55), abdomen (26%, n=47) and lastly the urogenital region (8%, n=15). This cohort suffered fewer cranial injuries compared to other studies, but had higher rates of orthopaedic, thoracic and abdominal injuries4,10,14–17 (Table 3).

Table 3: Associated injuries in pelvic fractures comparison (%).

Authors

Year of publication

Location

Closed head

Thorax

Abdominal

Genito-urinary

Other orthopaedic #

Current

2015

Auckland (NZ)

31

33

26

8

42

Banerjee17

2009

London
(UK)

12

34

14

-

13

Chia4

2004

Sydney (Australia)

44

27

17

17

42

Grisoni14

2002

Ohio
(USA)

26

7

14

4

49

Silber10

2001

Philadelphia (USA)

39

20

17

9

54

Rieger16

1997

Munster (Germany)

48

10

19

24

27

Lane-O’Kelly15

1995

Dublin
(Ireland)

38

16

4

16

24 

Management of the pelvic fractures was conservative in 93% of patients, with only 13 cases requiring surgical intervention. Nearly all fractures that required surgical management were either Type III or IV. There was one Type I fracture which required operative intervention for fixation of a significantly displaced avulsed segment of the left anterior superior iliac spine (ASIS).

Four patients required external fixation and stabilisation (all Type IV fractures with pelvic instability), seven required an open reduction and internal fixation while two patients required manipulation and reduction of the hip under anaesthetic. One patient required an initial external fixation and subsequent open reduction and internal fixation of anterior pubic symphysis fracture.

Operative procedures for associated injuries were much more common and were required in 38% (n=68) of cases. These included nine emergency laparotomies, four diagnostic laparoscopies, two thoracotomies and 13 neurosurgical procedures (craniotomy, craniectomy, evacuation of SDH and placement of EVD). Two children in the current study’s cohort suffered a transection of an iliac vessel and both underwent successful embolisation. The most commonly performed surgical intervention was open reduction and internal fixation of other orthopaedic long bone fractures (19 procedures). 

Mortality

Eleven children died in our study (6.1%). Eight children were struck by motor vehicles and two were passengers in a motor vehicle accident. One death was as a result of multiple injuries secondary to non-accidental trauma suffered at home.

The mean ISS in the 11 children who died was 36.5 (range 17–59), whereas the remainder had a mean ISS of 14.8 (range 4–59). The difference in ISS in these groups was statistically significant (Wilcoxon Two-Sample test, p<0.0001).

All fatalities had significant associated injuries to other organ systems and were of Type II, III and IV pelvic fracture classification (Table 4). Among these, nine were directly attributable to severe head injuries while two patients died of a combination of severe injuries to the head, chest, abdomen and pelvis. There were no deaths directly attributable to the pelvic fracture.

Table 4: Summary of deaths (n=11).

Child

Age

(years)

Pelvic fracture type

Admission GCS

ISS

Cause of death

1

1

II

4

45

Severe head injury

2

2

II

3

17

Abdominal (multiple liver lacerations) + severe head injury + pelvis (fractured left ilium)

3

3

III

3

30

Severe head injury

4

4

III

3

43

Severe head injury

5

8

III

3

38

Severe head injury

6

10

IV

8

27

Severe head injury

7

11

IV

3

59

Severe head injury

8

12

III

3

43

Severe head injury

9

13

III

8

29

Severe head injury

10

14

III

4

41

Abdominal (splenic injury) + thoracic (bilateral haemothorax) + pelvic injury (ring disruption fractures)

11

14

IV

3

29

Severe head injury 

Discussion

Paediatric pelvic ring injuries constitute 0.3–4% of all paediatric injuries.1 With the exception of avulsion injuries, they occur secondary to high-energy blunt force trauma. Fifty to 80% of pelvic fractures in children are caused in road traffic crashes, mostly when the child is struck by a motor vehicle.4,10,14–17 In our study, 79% of all injuries were caused by motor vehicles, with 46% of the children overall being pedestrians struck by motor vehicles. Significantly, children under 10 years of age were more frequently struck by motor vehicles compared to older children.

In our analysis, we noted that pedestrian-motor vehicle injuries occurred more frequently in left-hand driven countries while passengers injured in motor vehicles were more frequent in right-hand driven countries. While many variables would certainly contribute to this finding, it is nonetheless an important facet to consider in the local and global development of preventative measures.

The implications of pedestrian-motor vehicle accidents are significant and can give a characteristic pattern of injuries, described by Waddell and Drucker.18 Children are usually struck from the side when running onto the road or cycling in the near lane. Initial impact from the car bumper causes lateral compression injuries to the femur/pelvis and the trunk on one side, following which the child is thrown, striking the contra-lateral side of head on the ground.18 The fracture patterns seen in this study support this finding, as does the distribution of the associated injuries.

The greater cartilaginous volume and bony plasticity of the paediatric pelvis provides an increased capacity for energy absorption. Additionally, the symphysis pubis and sacroiliac joints also have increased elasticity and flexibility.10,11 Together, these factors result in considerable force being required to disrupt the integrity of the pelvic ring and cause a fracture. Soft tissue trauma to the perineum and buttocks should alert the treating physicians to potential underlying pelvic fracture.

As considerable force is required to fracture the pelvis, it is important to recognise that significant trauma has also occurred to other physiological systems; associated injuries should be expected and investigated. With a compliant chest wall and lack of abdominal musculature, the underlying viscera and soft tissues in children are less well protected and are more easily injured.10 Within our study, associated injuries involving the chest wall, abdominal organs, head injuries and the genito-urinary system occurred in 68% of patients. In the setting of significant blunt trauma even a ‘minor’ paediatric pelvic fracture should be seen as a ‘red flag’ indicating a potentially serious systemic injury.

In a recent systematic review of the literature incorporating 25 studies of pelvic fracture management since 1966 (sample sizes n=13–1,190), the average mortality rate was 6.4%. In addition, there was no significant change in the mortality rates of children with pelvic fracture noted during the last 30 years.1 The mortality rate from the current study is 6.1% (n=11). All of the 11 children who died in the current cohort had a head injury and this was the direct cause of death in nine. Snyder et al demonstrated that children with associated CNS injury have a 10-fold increase in mortality than children without CNS injury.19 Associated closed head injuries were the primary cause of mortality in multiple paediatric studies from around the world.4,10,17

Unlike adult trauma patients, fatal exsanguination is uncommon and a rare cause of death. Paediatric vasculature is smaller and vessels have a more effective vasoconstriction response, are less friable and have not undergone atherosclerotic change as in adults.3,20 Children typically sustain lateral compression injuries.21 This configuration does not cause increase of the pelvic volume and is thus mechanically less likely to precipitate haemorrhage. Both these factors likely account for the reduced incidence of fatal exsanguination in children compared to adults. In our series, only two patients required angiographic embolisation of ruptured iliac vessels. Other paediatric studies have reported a similarly low incidence of vascular intervention.22 

Treatment of paediatric pelvic fracture is primarily non-operative.5,15 Only 7% (n=13) of the children in this cohort were managed operatively. In this study, majority of pelvic fractures were managed with bed rest and occasionally with traction and then mobilisation. Traction was used until stability was achieved. The younger patients were sometimes treated in a hip spica. It is important that the abdomen has been “cleared” before the application of spica as further assessment is near impossible. Occasionally an external fixator was used when the injured child was haemodynamically unstable and source of bleeding was not clear. The remodeling potential of the immature skeleton,23 coupled with the high success rates noted in patient studies,5,15,24 are proposed as the rationale for this conservative approach.

Some long-term follow-up studies have however reported significant residual morbidity with this approach, due primarily to low back pain and leg length discrepancies. Operative stabilisation of pelvic ring has started to play an increasing role in the management of pelvic fractures over the past decade, with the treatment aims being anatomical reduction and maintenance of a symmetrical pelvis.25 Currently, there are no established guidelines for operative management as evidenced by the wide range of rates of operative interventions from 0.6 to 30% in the literature.1

Limitations of this study include its retrospective nature, both in aspects of data collection and in that the results are compared against similar retrospective reviews in the literature. Additionally, long-term morbidity outcomes and Quality of Life scores were not examined. It is certainly something that will be a useful addition to the analysis and would be best collected prospectively to reduce bias.

Conclusion

We have demonstrated that pelvic fractures are an important marker of severe trauma. Patterns of paediatric pelvic fractures reported by studies around the world are very similar, with motor vehicle accidents the highest occurring mechanism of injury. When evaluating and managing pelvic fractures, a multi-disciplinary approach needs to be employed to prioritise the management of pelvic fracture and associated injuries. 

Summary

We examined recent smoking trends among doctors and nurses in New Zealand using recent census data. We found that smoking had declined steadily and by 2013 only 2% of male and female doctors and 9% of male and 8% of female nurses were regular cigarette smokers. Smoking was more common among Māori doctors (7%) and nurses (19%), and also among psychiatric nurses. The findings suggest that New Zealand doctors had achieved the Smokefree 2025 goal of minimal (<5%) smoking prevalence and all nurses except psychiatric nurses were on track to do so. Targeted workplace smoking cessation support could be used to reduce smoking among key occupational groups such as Māori nurses.

Abstract

Aim

Pelvic fractures constitute between 0.3% and 4% of all paediatric injuries, with a mortality rate up to 25%. This study aims to review the experience with pelvic fractures at Starship Children’s Hospital and demonstrate its role as a marker of severe trauma.

Method

A retrospective review of children with pelvic fractures managed at our institution in the 20-year period between July 1995 and May 2015 was performed. The search identified 179 consecutive children admitted with a pelvic fracture. Data fields collected included patient details, mechanisms of injury, investigations performed, length of hospital stay, management and complications. Data was also collected on Injury Severity Score (ISS), Glasgow coma scale (GCS), transfusion requirements and details of associated injuries (both orthopaedic and non-orthopaedic).

Results

Median age was eight years (IQR 5-12 years) with 65% boys. The median Injury Severity Score (ISS) was 9 (IQR 4-22). Pedestrian-motor vehicle injuries were most common at 46% of cases, followed by passengers injured in motor vehicle accidents accounting for 23% (n=41). Associated injuries were present in 68% (n=122) of patients, with other orthopaedic fractures (42%, n=75) and thoracic injuries (33%, n=59) most common. Management of pelvic fractures was primarily non-operative, with only 7% (n=13) requiring operative intervention. In comparison, operative procedures for associated injuries were much more common and were required in 38% (n=68) of cases.

Conclusion

Pelvic fractures represent an important marker for severe trauma. Patterns of paediatric pelvic fractures reported by other studies around the world are very similar. Understanding the patterns in which pelvic fractures and their associated injuries occur and the outcome of treatment is fundamental to the establishment of effective preventative, diagnostic and therapeutic interventions.

Author Information

Mohit Bajaj, Paediatric Surgical Registrar, Starship Children’s Hospital, Auckland;
Giorgio Stefanutti, Paediatric Surgical Registrar, Starship Children’s Hospital, Auckland;
Haemish Crawford, Paediatric Orthopaedic Surgeon, Starship Children’s Hospital, Auckland;
Vipul Upadhyay, Paediatric Surgeon, Starship Children’s Hospital, Auckland.

Acknowledgements

Rong Hu (Research Statistical Consultant – Auckland District Health Board) for assistance with statistical analysis. 

Correspondence

Dr Mohit Bajaj, Paediatric Surgery Department, Starship Children’s Hospital, 2 Park Road, Grafton, Auckland 1023.

Correspondence Email

mohitb@adhb.govt.nz

Competing Interests

Nil.

References

  1. Gansslen A, Heidari N, Weinberg AM. Fractures of the pelvis in children: a review of the literature. Eur J Orthop Surg Traumatol, 2013; 23(8):847–61.
  2. Mazurek AJ, Epidemiology of paediatric injury. J Accid Emerg Med, 1994; 11(1):9–16.
  3. Bond SJ, Gotschall CS, Eichelberger MR, Predictors of abdominal injury in children with pelvic fracture. J Trauma, 1991; 31(8):1169–73.
  4. Chia JP, et al. Pelvic fractures and associated injuries in children. J Trauma, 2004; 56(1):83–8.
  5. Musemeche CA, et al. Selective management of pediatric pelvic fractures: a conservative approach. J Pediatr Surg, 1987; 22(6):538–40.
  6. Baker SP, et al. The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma, 1974; 14(3):187–96.
  7. Teasdale G, Jennett B, Assessment of coma and impaired consciousness. A practical scale. Lancet, 1974; 2(7872):81–4.
  8. Torode I, Zieg D. Pelvic fractures in children. J Pediatr Orthop, 1985; 5(1): 76–84.
  9. Herring JA. Tachdjian’s Pediatric Orthopaedics: From the Texas Scottish Rite Hospital for Children, 2014, Elsiever Saunders.
  10. Silber JS, et al. Analysis of the cause, classification, and associated injuries of 166 consecutive pediatric pelvic fractures. J Pediatr Orthop, 2001; 21(4):446–50.
  11. Shore BJ, et al. Pediatric pelvic fracture: a modification of a preexisting classification. J Pediatr Orthop, 2012; 32(2):162–8.
  12. Upperman JS, et al., Early functional outcome in children with pelvic fractures. J Pediatr Surg, 2000; 35(6):1002–5.
  13. Nabaweesi R, et al. Prehospital predictors of risk for pelvic fractures in pediatric trauma patients. Pediatr Surg Int, 2008; 24(9)1053–6.
  14. Grisoni N, et al. Pelvic fractures in a pediatric level I trauma center. J Orthop Trauma, 2002. 16(7):458–63.
  15. Lane-O’Kelly A, Fogarty E, Dowling F. The pelvic fracture in childhood: a report supporting nonoperative management. Injury, 1995; 26(5):327–9.
  16. Rieger H, Brug E. Fractures of the pelvis in children. Clin Orthop Relat Res, 1997; (336):226–39.
  17. Banerjee S, Barry MJ, Paterson JM, Paediatric pelvic fractures: 10 years experience in a trauma centre. Injury, 2009; 40(4):410–3.
  18. Waddell JP, Drucker WR, Occult injuries in pedestrian accidents. J Trauma, 1971; 11(10):844–52.
  19. Snyder CL, et al. Blunt trauma in adults and children: a comparative analysis. J Trauma, 1990; 30(10):1239–45.
  20. Ismail N, et al. Death from pelvic fracture: children are different. J Pediatr Surg, 1996; 31(1):82–5.
  21. Leonard M, et al. Paediatric pelvic ring fractures and associated injuries. Injury, 2011; 42(10):1027–30.
  22. Demetriades D, et al. Pelvic fractures in pediatric and adult trauma patients: are they different injuries? J Trauma, 2003; 54(6):1146–51; discussion 1151.
  23. Turgut A, et al. Demographic Characteristics of Paediatric Pelvic Fractures: 10-Years’ Experience of Single Paediatric Orthopaedics Clinic. Eurasian J Med, 2015; 47(2):130–4.
  24. Nierenberg G, et al. Pelvic fractures in children: a follow up in 20 children treated conservatively. Journal of Pediatric Orthopaedics, 1992; 1(2).
  25. Schwarz N, et al. Long-term results of unstable pelvic ring fractures in children. Injury, 1998; 29(6):431–3.