OPEN

FRACTURES

OF

REVIEW

PETER

B.

HANSON,

From

JOSEPH

Davis

OF

C.

Medical

THE

PELVIS

43 CASES

MILNE,

Center,

MICHAEL

W.

CHAPMAN

Sacramento

We reviewed 43 patients treated from 1984 to 1988 for open fractures of the pelvis. There were four Gustilo type I wounds, seven type II and 32 type HI; 22 fractures were stable and 21 unstable. The overall mortality was 30%; the average Injury Severity Score was 30, being 26 in the survivors and 40 in the fatal cases. We analysed the influence of a number of factors on the mortality rate. The most important were the Iss and the age in years, while the presence of a type III wound and instability of the fracture also had an influence. We describe two simple methods of assessment of the prognosis in individual cases, based on these factors. Open fractures of the pelvis usually result from high energy trauma and are associated with multiple injuries. They present a diagnostic and therapeutic challenge requiring an aggressive multi-disciplinary effort, since the quoted mortality rate is approximately 50% (Raffa and Christensen 1976 ; Maull, Sachatello and Ernst 1977; Rothenberger et al 1978a,b; Perry 1980; Kane 1984; Mears and Rubash 1986). These reports were from general surgical units so their focus is primarily on haemorrhage, sepsis, and the associated injuries ; there is little or no discussion of orthopaedic factors. Our study aimed to determine these factors in relation to the management and the prognosis of open fractures of the pelvis.

enteric organisms from an intra-abdominal injury. Each of these 43 cases was analysed retrospectively. Wounds were classified according to the criteria of Gustilo and Anderson (1976), with enterically contaminated

wounds

assessed,

PATIENTS

AND

METHODS

From 1984 to 1988, a total of 890 patients with pelvic ring disruptions were admitted to the Davis Medical Center, University of California. Of these, 43 (4.8%) were open injuries, classified as such where there was a soft-tissue defect communicating with the fracture or documented contamination of the retroperitoneum with

recorded

as type

III.

The

fractures

were

classified by the methods of both Pennal et al (1980) (Fig. 1) and Bucholz (1981) (see Figs 1 and 2). Fractures were considered to be unstable if either the anterior or the posterior sacroiliac ligaments or both were disrupted to give clinical instability. Associated injuries were recorded, and Injury Severity Scores(ISS)were calculated from the Abbreviated Injury Score (Civil and Schwab 1988; Copesetal 1988). The treatment of each fracture and wound was and

all

associated

operations

were

noted,

including colostomies, as was wound or systemic sepsis. The outcome was assessed primarily on mortality. Autopsy reports were available for all fatal cases ; the cause of death was derived from these and the in-patient records. The data were analysed using Student’s t-test and chi-squared analysis ; a p value of < 0.05 was considered significant. RESULTS

P. B. Hanson, MD, Resident J. C. Milne, MD, Resident M. W. Chapman, MD, Professor and Chairman University of California, Davis Medical Center, Department Orthopaedic Surgery, 2230 Stockton Boulevard, Sacramento, California 95817, USA. Correspondence

should

be sent

© 1991

to Dr M. W.

British Editorial Society ofBone 030l-620X/9l/2079 $2.00 J Bone Joint Surg [Br] 1991 ; 73-B :325-9.

VOL.

73-B, No. 2, MARCH

1991

and

Chapman. Joint

Surgery

of

Of the 43 patients studied, 31 were male and 12 were female. Their average age was 35 years (range six to 84). The most common cause of injury was a pedestrian being struck by a motor vehicle (17 patients). Other causes were accidents between motor vehicles (12), motorcycle accidents (5), gunshot wounds (4), and boating accidents (2). An

the cause

explosion,

a fall,

and

iatrogenic

injury

were

each

in one patient. 325

P. B. HANSON,

326

As would injuries,

seven

be expected,

involving

injuries

to the

abdomen

and

patients

there

there

all major

head,

23 of the were

organ

were

15 of the

genito-urinary

66 other

many

systems.

J. C. MILNE,

associated There

22 of the

system.

In the

musculoskeletal

rates for these and the Bucholz (1981) groups are shown in Table I. Of the five patients in whom haemorrhage was the major cause of death, four had Bucholz group III fractures, and one had a group II fracture. The severity of the wound varied widely from a small-calibre missile entry injury, to avulsion of both lower extremities. On the Gustilo scale there were four type I, seven type II, and 32 type III wounds. We did not

were

thorax,

43

injuries,

while 10 patients had 11 major spinal column injuries, and 10 had 12 major nerve disruptions. On average, there were 3.1 additional injuries per patient, only two having isolated open pelvic injuries. The average 155 was 30 (range 9 to 57). For survivors the average score was 26; it was 40 for those that died. Mortality rates are related to 155 in Table I.

Classification

of pelvic

Fig.

OrthopaedicsandRelated

Fig. Classification

fractures:

a) lateral

compression,

la

Pennal

Fig. a)group

classification,

wounds

c) vertical

shear.

into

A,

B or C categories,

had internal contamination The mortality for wounds was 14%, as compared to

Reprinted

with

Fig.

I, b)group

2b

Fig.

II, c)group

III.

Reprinted

On the classification of Pennal et al (1980) there 16 anterior compression fractures, eight due to lateral compression, and six to vertical shear. Thirteen of the fractures did not fit into this classification ; some which appeared radiographically to be due to vertical shear were known to be caused by anterior compression which had produced vertical instability. Using the (1981)

III

lb

were

Bucholz

type

but 14 ofthe type III wounds only, and 18 were external. with internal contamination

C/inical

permission,

lc

et al (1980).

2a fractures:

sub-classify

compression,

Fig.

Research

ofpelvic

b) anterior

M. W. CHAPMAN

12 were

group

I injuries,

eight were group II, and 13 group III. By definition, all these groups have ‘double breaks’ of the pelvic ring; 10 of our patients had stable fractures which did not fit the classification. We also grouped the fractures into stable and unstable injuries from the clinical findings. Mortality

with

permission

J BoneJoint

2c

Surg

[Am]

1981 ; 63-B

:401-2.

50% for those placed in type III on the basis of external wound size ; this difference is not statistically significant. Thirty patients had colostomies either for bowel injuries or to produce faecal diversion to assist wound care. All patients with type III wounds who survived had required a colostomy, and all those in whom sepsis played a role in mortality had had colostomies. In two patients, the association of small external wounds with obvious fractures of the pelvis was missed. One had a type I posterior wound and an acetabular fracture. No special wound care had been given before transfer to us one week after injury for fracture management. Extensive debridement was required and THE

JOURNAL

OF BONE

AND

JOINT

SURGERY

OPEN

sepsis

avoided. involving The wounds or fracture

The second case had perineal the anus and a Bucholz group III had been closed primarily with no fixation. One week later, when

was

lacerations fracture. colostomy transferred sepsis and

to our required

FRACTURES

institution, multiple

the patient debridement.

had

wound

OF THE

327

PELVIS

or drivers and motorcyclists together had only a 12% mortality. There was no significant difference in ISS between these groups, but 53% of the pedestrians had suffered Bucholz group III pelvic fractures. None of the four patients with gunshot injuries died.

DISCUSSION

The managementofthe soft-tissue and bony injuries involved both general and orthopaedic surgeons. All wounds were treated vigorously, using irrigation and debridement as necessary. All patients received broad-

Pelvic fractures have received much attention in the literature (Patterson and Morton 1973 ; Trunkey et al 1974; Rothenberger et al 1978a,b; McMurtry et al 1980; Gilliland et al 1982; Gylling et al 1985 ; Cryer et al 1988), but most of these reports have considered only closed injuries. Studies of open fractures are mainly in the general surgical literature (Raffa and Christensen 1976; Maull et al 1977; Rothenberger et al l978a; Perry 1980; Richardson et al 1982). Raffa and Christensen (1976) reported 26 patients with open pelvic fractures. Of the 16 suffering high energy trauma and with unstable fractures, eight died, seven of them because of pelvic sepsis. There was a 25% mortality rate after immediate colostomy as against 58% in those having a delayed colostomy or no colostomy. Rothenberger et al (1978a,b) and Perry (1980) reported a series of 31 patients with open pelvic fractures. There was a 42% mortality, with sepsis and haemorrhage each accounting for 40% of the deaths. There was a high proportion of vascular injuries : 27%, as compared with 0.3% in closed fractures, and patients who died with open fractures had three times the average blood loss of those dying after closed fractures. Immediate colostomy was advised, since none of the patients so treated died of

spectrum

sepsis.

Table I. The influence following open fractures

ofvarious of the

factors pelvis

on mortality

Mortality Number patients

Injury Severity Score(ISS)

%

7 18 9

0 2 4

11

44

>40

9

7

78

0 1 2

10 12 8

1 3 3

10 25 37

3

13

6

46

Stable Unstable

22 21

4 9

43

40

34 9

6t 7t

18 78

0 to 20 21to30 3lto4O

BuchoLzgroup

Fracture

Ageinyears

of

Number

Factor

*

single break in the pelvic ring

t

difference

significant

bacteriocidal

0

18

at p