Unstable pelvic fractures: a retrospective analysis

Injury Vol. 26, No. 2. pp. 81-85. 1995 Copyright 0 1995 Elsevier Science Ltd Printed in Great Britain. All rights reserved 0020.1383/95 $lO.OO+O.OO

Unstable

pelvic

fractures:

a retrospective

analysis

I. H. P. A. A. van Veen, A. A. M. van Leeuwen, T. van Popta, P. A. van Luyt, P. J. Bode’ and A. B. van Vugt Department of General Surgery and Traumatology

and Department

of Radiology,

University

Hospital of Leiden,

Leiden, The Netherlands

Thirty-nine patients with unstable pelvic fractures were analysed relrospectiuely. The mean age of the group was 41 years (range 15-77). Of these cases 35 had sustained high energy trauma. The mean Hospital Trauma Index-Injury Severity Score of the population was 32 (16-66). Nine cases were kaemodynamically unstable on admission. The type of unstable pelvic fracture was classified according lo Tile, Sixteen patients had a type B fracture and 23 had a vertical instability (type C) fracture. In two patients, an open fracture was seen. Directly associated injuries were diagnosed in I 1 putienls,of which eight showed damage of the urogenital system, fhree of the rectum and three of the peripheral nerve system. In seven cases the fracture was treated non-operatively; in the remaining 32 patients the pelvic ring was stabilized operatively. Additional therapy for kypovolaemic shock due lo pelvic bleeding was necessay in six cases. The overall mortality in this series was 13 per cent. Early and aggressive resuscitation and standardized treatment in well-equipped and staffed injury centres is mandatory in these severely traumatized patients to achieve optimal results and to minimize Ihe risk of fatal outcome.

Injury, Vol. 26,

81-85,

1995

Introduction Unstable pelvic ring fractures are notorious for their high morbidity and mortality rates. These fractures are almost exclusively caused by high energy trauma. Road traffic accidents (RTA) are responsible in the majority of cases. Other mechanismsof injury are falls from a great height and local compression by high forces’-4. Becauseof the mechanismsinvolved, a high percentage of the patients are multiply injured. Instability occurs in 13-17 per cent of all pelvic ring fractures2,3.An open injury in this region with a connection between the fracture and the skin, rectum or vagina occurs in 3.5-4.5 per cent of the unstable fractures’,z*4*5. In this study we retrospectively analysed 39 patients with an unstable pelvic ring fracture. The severity of injury (Trauma score and Hospital Trauma Index - Injury Severity Score (HTS-ISS), type of fracture6, method of treatment, morbidity and mortality rates are reported.

Materials

and methods

From January 1988 till January 1993, 39 patients were admitted to the Department of Traumatology of the

Table I. Classification of pelvic ring fracturesaccordingto Tile6 Type A

Type

B

Type C

Stable Al, Fractures of the pelvis not involving A2, Stable, minimally displaced fractures Rotationally unstable, vertically stable Bl , Open book Bl .I, Diastasis i 2.5 cm B1.2, Diastasis > 2.5 cm B2, Lateral compression: ipsilateral 83, Lateral compression: contralateral

the ring of the ring

(bucket

handle)

Rotationally and vertically unstable Cl, Unilateral C2, Bilateral C3, Associated with an acetabular fracture

University Hospital of Leiden with an unstable pelvic fracture. To be included in this study the pelvic fracture had to conform to Tile’s definition6 of instability as summarized in Table I. The fractures were defined asunstable when the radiographs showed a disruption of both the anterior and posterior part of the pelvic ring (type B or C fractures). Pathological fractures were excluded. The mean age of this population was 41 years (range 15--77), 28 were male and II female. In 90 per cent (N= 35) of the cases,high energy trauma was involved, 24 of thesepatients had an RTA, five fell from a height and six were compressed in the pelvic region by an industrial accident. Thirty patients presented first at the emergency department and nine were referred to our institute, mostly more than 24 h after the initial injury.

Results Of the 30 cases who presented primarily, 15 had an optimal Trauma score (TS= 16) at presentation, 10 had a score from 13 to 15, and five (16 per cent) suffered serious signs of vital instability with a TS _ 12

No ARDS/MOF ARDS/MOF Survival Mortality

18 0 18 0

6 0 5 1

4 4 8 0

3 4

Total

18

6

8

7

(N=39)

;

Statistics: 2 x 2 table, Fisher’s exact test. O-4 versus ~4 units of blood related to ARDS/MOF: P 4 units of blood related to mortality: PC 0.05.

from 9cm to 2.5 cm, which was qualified as acceptable in this case. The overall mortality in this study was 13 per cent (N= 5). Four of those five died within the first 24 h due to cerebral injury (N= I) or hypovolaemic shock (N= 3). One patient died 33 days after injury due to multiple organ failure (MOF). The direct associationof the mortality from hypovolaemic shock with the pelvic fracture was noticed in two of the five fatal cases. In the present series,the mean ISS was 32. This value came to 53 for the non-survivors and 29 for the survivors. The mortality rate was 66 per cent in the group with an ES above 50 and 3 per cent below this level (adjusted x2 test: P< 0.01). No survivors were found among the patients with an ISS above 57 (N= 2). Figtlre I shows the frequency distribution of the ISS-agesum linked with the mortality in the present series.The mean age of the non-survivors came to 52 compared with 39 years for the survivors (Student’s t-test: P< 0.05). In 14 patients the ISS-age sum exceeded 80. The mortality rate in this group was merely 27 per cent versus 4 per cent in the remaining 25 cases.(Adjusted x2 test: P< 0.1, NS). No survival was seenin values of ISS-I- age above 100. Of the 34 survivors, 65 per cent suffered major morbidity. Persistent bleeding necessitating secondary intervention was registered once. Deep wound infection was seen five times and adult respiratory distress syndrome (ARDS) was noticed in five patients. One patient developed MOF with a fatal course as mentioned above. Excluding four caseswith early mortality, 29 patients who did not develop ARDS/MOF had a meanHTI-ISS of 29.0f 12.0 (range 16-47), versus those with ARDS/ MOF: 40.3 f 14.1 (range 25-66). Although there was a tendency towards higher scoresin the presenceof ARDS/ MOF this could not reach the level of significance in this small series(P < 0.1). In this series, 18 patients did not receive any blood transfusion. In the remaining 2 1 casesa wide range, of from two to 47 units of blood, was registered with a mean rate of 6.7 units, seven patients needing 12 units of blood or more (TableIV). Mortality was closely related to the need for transfusion. Only one patient with a transfusionneed of 4 units or lessdied in this series,owing to cerebral damage. ARDS and MOF were exclusively observed in caseswith 5 or more units of blood transfused (N= 15), which was significantly more than in the remaining patients (Fisher’s exact test, P< 0.01).

Discussion An unstable pelvic ring fracture is often seenin multiply-

Figure 1. ISS-agesum relatedto mortality in 39 caseswith an

unstablepelvic ring fracture. n Freq.Value survivors; 0 Freq. Value non-survivors.

injured patients, These patients need a multidisciplinary approach and should be treated initially in well-equipped and staffed injury centres. Stabilization of vital parameters is the first goal, and a standardized trauma protocol, for diagnostic policy aswell as for surgical treatment, should be followed routinely. On physical examination in the emergency room, after vital parameters and resuscitation, staff should focus on directly associated injuries such as urogenital trauma, rectoperineal damage, open fracture of the pelvis and neurovascular lesions of the lower extremities. Careful rectoperineal inspection and digital rectal examination are mandatory. In addition to the routine aspectsof chest radiographs in blunt thoracic injury, special attention must be paid in thoracic radiography towards traumatic diaphragmatic hernia, which is mostly seenon the left side and can occur in compression of the pelvic ring and abdomens.Associated intraabdominal injuries should be evaluated by emergency abdominal ultrasound to diagnose or exclude intra-abdominal bleeding7. In caseof intra-abdominal blood loss on the abdominal ultrasound combined with persistent shock, an emergency laparotomy in the first step to be followed. Stabilization of the pelvic ring is essentialtoo in the acute treatment of these exsanguinating patients. The pelvic injury itself is often associatedwith massive blood loss. Volume reduction of the minor pelvic cavity by fracture reduction and stabilization can restore the natural tamponade effect and thereby limit blood 10ss~,~. Stabilization can be achieved by external and/or internal fixation. External fixation is a quick method but does not result in optimal stability. However, the patient can be mobilized early with results in a reduction of respiratory complications. To optimise stability, external fixation can be combined with ORIF ventrally. If an emergency laparotomy is indicated, it is advisable to apply internal fixation ventrally in the samesession.This policy can be followed in caseof rectal lesions,necessitatinglaparotomy for diverging colostomy and in urogenital lesions,which requires an extraperitoneal approach to the bladder and urethra. Different methods of internal fixation are available. Ventrally applied plate-osteosynthesis can be used in case of symphyseolysis, fracture of the rami ossispubis and iliac bone and dislocation of the sacra-iliacal joint. Posterior plate osteosynthesis can be useful in fractures of the posterior aspect of the iliac bone. In SI-dislocationfractures posterior screw-fixation is a useful alternative. In

84

Injury: International Journal of the Care of the Injured Vol. 26, No. 2, 1995

Staudar& Trauma Series:

Pelvic Rb Fracture

Pelvic AP-view and Abdominal UltraSound (AUS)

pzsg*,,,

I rectum uremladder

neurology skin/perineum

m

AUS

A&

bloodI

blood ++ I

LAF’AROT0MY specificsurgical treatment (+ ventralplate osteosynthcsis)

YES

I

-

Inlet-view Outlet-view 64la-view*) (Oburator-view*)

&NO/

ANGIOGRAF’HY+ selective embolisation

DEFINITIVE FRACTURE CARE

Figure 2. Diagnosticand therapeuticstrategy in patientswith multipleinjuriesor suspectedpelvic ring fractures. sacralfractures the use of sacralbars is advocated6. In our opinion, internal fixation of the posterior complex should only be carried out in experienced hands and only after complete evaluation of the pelvic ring by CAT scan. As mentioned before, 82 per cent of our patients have been treated operatively. In this series,type BI injuries were fixated without exception. B2 and B3 fractures in general can be treated non-operatively, becausethe pelvis has the elasticity to restore to a near normal positior?. In two of four patients, operative treatment was carried out. Type Cl and CZ fractures were almost exclusively stabilized, in four casesa combination of external and internal fixation was used. In type C3 fractures, ORIF of the acetabulum was used in four cases,three of them were

combined with internal fixation of the pelvic ring. In four of the 13 caseswith minimal displacement, the treatment of choice was non-operative. In the study of Hesp et a1.l on patients with unstable pelvic fractures, the group of patients with an ISSabove SO had a mortality rate over 50 per cent. The mean ISS of the non-survivors was 53, no patient with a score above 57 survived. The results of the present seriesare similar to this report. Hanson et al9 describeda relationship between the sum of age and ISSand the mortality rates in patients with open pelvic fractures. A mortality rate of 80 per cent was reachedwhen this ISS-age-sumexceeded 70 points, below this border it reached to 4 per cent. An explanation for the

van Veen et al.: Unstable

pelvic fractures

contrast with our results,is that Hanson’sreport dealt only with open fractures (overall mortality 30 per cent), while the present seriescontained fewer open injuries. The main causes of death in patients with unstable pelvic fractures are exsanguination, brain damage and sepsis.In this seriesthe mortality rate reachedup to 13 per cent which is consistent with the recent literature’~2~4~5. To reduce morbidity and mortality rates in patients with an unstable pelvic ring fracture it is important to follow a standard treatment protocol, with an aggressive approachg,‘O.In compound injuries, early open reduction and stableinternal fixation with primary managementof all direct associated injuries is advocated’l. A schedule for diagnostic and therapeutic policy is proposed in Figure2. Stabilization of respiration and circulation is the first step in emergency treatment, including emergency laparotomy3,‘. If manipulation of the pelvis establishesclinical instability6, or an AP view of the pelvis shows an unstable pelvic ring fracture, further examination should be focused on the diagnosisof possible direct associatedinjuries. The performance of a rectal and perineal inspection in this case is indispensable.Abdominal ultrasonography must be part of the standard protoco17. Depending on the circulatory stability of the patient, analysis can be completed. The radiological check up should include an inlet and outlet view of the pelvis and ala, and obturator views in casesof acetabular involvement. A retrograde urethrogram has to be done in all cases with lack of spontaneous miction before transurethral catheterization can be carried out safely. Analysis must be completed by computed tomography of the pelvis for definitive planning of fracture treatment. In caseof hypovolaemic shock time-consuming procedures like a CT-scan are impossible and immediate stabilization of the pelvic ring should be carried out to reduce transfusion needs’. External fixation forms a useful altemafive in the acute phase, especially when analysis of the posterior complex cannot be carried out adequately. In case of persistent shock without the need for laparotomy, selective angiography and embolization is an alternative to achieve haemostasis. It should include bilateral visualization of the internal iliac artery. Especially in casestreated with closed reduction and stabilization of the pelvic ring by external fixation in acute cases,radiological intervention can be successfulwith low risks of infection. Additional operative treatment of pelvic haemorrhage by packing of the pelvic cavity can be done easily in cases where ORIF is performed or an (emergency-)laparotomy is indicated.

85

Antibiotic prophylaxis, adequate surgical wound care including primary treatment of associatedpelvic injuries will result in lessinfection and septic complications. With this aggressive approach lower morbidity and mortality rates can be obtainedl’,“.

References I HespWL, van der Werken C, KeunenRWM et al. Unstable fractures and dislocationsof the pelvic ring - results of treatment in relation to the severity of injury. Nefh J Stlrg 1985; 37: 148.

2 Leung KS, Chien P, Shen WY and So WS. Operative treatmentof unstablepelvic fractures.Injury 1992; 23: 31. 3 Mucha P and Fame11MB. Analysis of pelvic fracture management.

J Trawza 1984; 24,379.

4 Wild JJ,HansonGW andTulles HS. Unstablefracturesof the pelvis treated by external fixation. ] Bonejoint Surg 1982; 64A: 1010. 5 LatenserBA, Gentile110LM, Tarver AA, Thalgott JS and

Batdorf JW. Improved outcome with early fixation of skeletallyunstablepelvic fractures.J T~uuma1991; 31: 28. 6 Tile M. Pelvicring fractures:shouldthey befixed?J Bone]oinf Surg 1988; 7OB: 1. 7 Bode PJ, Niezen A, van Vugt AB and Schipper J. Abdominal ultrasound as a reliable indicator for conclusive laparotomy in blunt abdominal trauma. J Trauma1993; 34: 27. 8 Van Vugt AB and Schoots FJ. Acute diaphragmatic rupture due to blunt trauma: a retrospective analysis. ] Trauma 1989; 29: 683. 9 HansonPB,Milne JCand ChapmanMW. Open fracturesof the pelvis.J BoneJoint Surg 1991; 73B: 325. 10 Richardson JD, Harty J, Amin M and Flint LM. Open pelvic fractures.] Trauma 1982; 22: 533. 11 LeenenLPH, van der WerkenChr, SchootsF and Goris RJA. Internal fixation of open unstable pelvic fractures. ] Trauma 1993; 35: 220. 12 Moreno C, Moore EE, Rosenberger A and Cleveland HC.

Haemorrhageassociatedwith majorpelvic fracture:a multispecialtychallenge.] Trauma 1986; 26: 987. Paper accepted 4 August 1994.

Requests for reprints should be addressed to: A. B. van Vugt

MD,

PHD,

University Hospital Leiden,Departmentof Traumatology,PO BOX 9600, 2300 RC Leiden, The Netherlands.