Secondary Logo

Journal Logo

Techniques and Outcome in Pelvic Fractures

Outcome of Operatively Treated Unstable Posterior Pelvic Ring Disruptions

Tornetta, Paul*; Matta, Joel**

Editor(s): Tornetta, Paul

Author Information
Clinical Orthopaedics & Related Research: August 1996 - Volume 329 - Issue - p 186-193
  • Free


Unstable pelvic ring disruptions occur via high energy mechanisms and are associated with multiple concomitant injuries.12 Initial treatment usually includes external fixation for emergent skeletal stabilization in conjunction with other resuscitative efforts. Once patients are stabilized, definitive treatment of the pelvis is undertaken.

Treatment of these unstable injuries has evolved several times. Treatment began with bed rest augmented by traction or pelvic sling.4,5,15 This was followed by external fixation (with or without traction).3,7,11,13 External fixation controls the anterior pelvis fairly well, but yields poor control of unstable posterior injuries.11 (nonpublished data, Harvell J, Mears D: Late result of pelvic ring disruptions managed by external fixation, internal fixation, or combined methods. Presented at the 53rd annual meeting of the American Academy of Orthopedic Surgeons. New Orleans, 1986) Traction was added to aid in correcting the cephalad displacement of the hemipelvis.16,20 Most recently, open reduction and internal fixation has become the method of choice for stabilization of the pelvis.1,8,11,12,16,20,22

Long term followup of patients with unstable pelvic ring injuries has been reported for earlier methods of treatment by several authors.4,6,8,11,15,16,18,19,20,21 Extremely high rates of back pain, impaired gait, obliquity, sitting problems, and severe neurologic sequelae are common findings. In reviewing the literature, however, particular attention must be given to distinguish patterns of posterior pelvic injury. Three main categories exist: fractures of the sacrum, fracture dislocations of the sacroiliac joint (fractures of the ilium exiting through a portion of the sacroiliac joint), and pure sacroiliac dislocations. Most reports have concluded that poor results correlate with the degree and location of the posterior pelvic ring injury. Pure sacroiliac dislocations and sacral fractures result in higher rates of back pain than fractures through the ilium. Severe neurologic compromise with significant residual has been associated with vertically unstable injuries, particularly if there is a transforaminal sacral fracture.1,5,10,14,15,18

Interpreting the findings of these earlier reports is difficult because of the classification system used at the time of their writing. Until Bucholz's2 classic article differentiating stable from unstable posterior ring injuries, the location of the fracture or dislocation was reported, but little reference to stability was made. Additionally, Bucholz Type II injuries with symphyseal dislocation and anterior opening of the sacroiliac joint(s) were categorized as sacroiliac dislocations in many reports.4,15,18,19,21 These injuries are clearly different than Type III sacroiliac dislocations with disruption of the entire posterior ligamentous support system. One would expect a greater percentage of sequelae from Type III injuries. Thus, by including Type II patterns, early reports tended to underestimate the percentage of long term problems patients with Type III pelvic ring disruptions will have. To get an idea of the number of injuries that may be truly unstable posteriorly, one can look at the number of patients in these studies who had hemipelvic displacement. An example of this is a widely quoted study from Slatis and Huittinen18 of double vertical fractures of the pelvis. In a series of 163 patients, long term followup was reported for 65 patients. Seventeen percent of the patients had disabling low back pain, 32% had impaired gait, and 48% had neurologic sequelae. However, only 18% of the original patients had displacement of the hemipelvis. Thus, many of the patients (possibly a majority) might have had Type II injuries.

The present study was undertaken in an attempt to delineate the long term outcome of Type III posteriorly unstable pelvic ring injuries treated operatively.


In a series of 107 unstable pelvic ring injuries treated operatively, 69 patients had 76 unstable posterior pelvic ring injuries treated operatively.12 Forty-six of these patients with 48 unstable posterior pelvic ring injuries were available for examination and form the basis of this study.

All patients had Type III injuries and were treated with open reduction and internal fixation. Methods of reduction and fixation were described in an earlier report.12

Minimum followup to be included was 1 year. The average followup was 44 months (range, months 12-101 months) There were 11 sacral fractures, 24 fracture dislocations, and 13 pure sacroiliac dislocations followed for an average of 39, 50, and 47 months, respectively. The average age of the patients was 32, 40, and 29, respectively.

Associated ipsilateral acetabular fractures occurred in 13 patients (27% of the 48 posterior injuries) and were more common in conjunction with pure sacroiliac dislocations (8/13) compared with sacral fractures (1/11) or fracture dislocations (4/24). Significant neurologic injury was more common in patients sustaining sacroiliac dislocations (8/13) and sacral fractures (6/11) than in those with fracture dislocations (3/24). The overall rate of neurologic injury was 37.5%. This information is summarized in Table 1.

Evaluation included repeat radiographs, a standard questionnaire, and physical examination. Factors evaluated included associated injuries, initial pelvic displacement, operative reduction, position at union, type of fixation, return to work, ambulation, neurologic sequelae, and pain during various activities (sitting, standing, walking, bending, lifting, and sexual functions).

Patients were determined to have pain with strenuous activities only if they had no limitations and had some posterior pelvic soreness after hours of hard work or sport. Pain with sitting was recorded for patients who could not sit for an 8-hour day without pain in the area of their posterior pelvis. If slight changes in position allowed for complete relief, then the patient was not included.

Numeric scoring for patients with pelvic fractures can be difficult. Many patients have associated injuries to other areas that can affect their function. The high rate of neurologic injury also skews functional assessments. Instead of rating each patient regarding multiple categories and summing to a final score, it was thought to be more valuable to report descriptive data for the patients. Thus, a patient with a persistent neurologic deficit who cannot work or ambulate well but has no pain attributable to a pelvic fracture can be distinguished from a patient with the same problems caused by posterior pelvic pain. In a scoring system, the patients, although manifesting different sequelae, could have the same functional score.9 By using descriptive information, the pelvic ring injury can be better evaluated with respect to pain, effect on daily activities, initial displacement, and reduction.


For all patients, 32 (67%) were able to return to their previous jobs without restriction. Of the others, 14.5% changed jobs because of pelvic pain, 8% changed jobs because of associated injuries, and 8% changed jobs because of neurologic problems. Only 5 patients could not work at all and in 3 patients this was from pelvic pain.

Neurologic injury was present in 35% of patients. Of the entire group, 13% had sexual problems, 13% had causalgia, and 27% had lower extremity weakness, 54% of whom required an ankle foot orthosis for ambulation.

Thirty patients (63%) reported no pain (24) or pain only with strenuous activity (6). One patient had severe pain that he thought interfered with every aspect of his life. The other patients had various amounts of pain with sitting and standing for long periods or with bending and lifting. They reported only minimal changes in their daily activities. No patient complained of anterior pain.

No patient complained of pain or ambulatory deficiency due to pelvic obliquity.

Thirty patients ambulated without restriction (63%). Six had a limp or required a cane because of posterior pelvic pain (12%). The other patients had impairment of their ambulation from neurologic deficit or from associated injuries.

In the following sections, these parameters will be reported separately for each type of posterior injury: fracture dislocation, sacral fracture, or pure sacroiliac dislocation. See Table 1 for a summary of results.

Fracture Dislocations

Nineteen of 24 patients were able to return to the same job without modification. Included in this group were a firefighter and a jockey. Of the 5 patients who could not return to the same job, 3 changed their jobs because of posterior pelvic pain and 2, because of an associated injury to the hip.

Eleven patients reported no pelvic pain and another 3 patients had pain only with strenuous activity. The other patients had pain with sitting or standing for long periods and with strenuous activity.

Fifteen patients ambulated normally. Of the 9 other patients, 2 had problems with ambulation attributable to pelvic pain and 7, from neurologic or associated injuries. Three patients had neurologic sequelae, 1 had decreased erections and 2 had lower extremity problems, 1 requiring an ankle foot orthosis.

Fourteen patients had an initial displacement of greater than or equal to 15 mm. Eight patients had an initial displacement of less than 15 mm. Four had pain and 4 did not. Twenty-one patients had an excellent reduction (≤ 4 mm) and 3 had a good reduction (≤ 10 mm). For the 3 patients with a good reduction, 1 had no pain, 1 had pain only with strenuous activity, and 1 had pain with long sitting.

Sacral Fractures

Seven of 11 patients with sacral fractures returned to work at the same job. Two changed their jobs because of pelvic pain and 2, because of neurologic injury.

Five patients had no pain and 2 had pain only with strenuous activity. The other patients had pain with long standing and sitting.

Nine patients ambulated without restriction and 2 had neurologic problems that interfered with ambulation. None had difficulty with ambulation from pelvic pain. Six patients had neurologic injuries. Three had decreased erection, 4 had lower extremity weakness, and 2 required an ankle foot orthosis. One patient had causalgic type pain.

Of the 9 patients who had greater than or equal to 15 mm of initial displacement, 5 had pain and 4 did not. Of the 2 patients with less than 15 mm of initial displacement, 1 had pain and 1 did not. Pain was reported by 5 of 7 patients with an excellent reduction, 0 of 3 with a good reduction, and 1 of 1 with a fair reduction.

Pure Sacroiliac Dislocations

Six of 13 patients returned to the same job. Two patients changed jobs primarily because of posterior pelvic pain, 2 because of associated injuries, 2 primarily because of neurologic injuries, and 1 because of associated psychiatric problems.

Eight patients had no pain and 1 had pain only with strenuous activity. The other 4 patients had pain with standing or sitting for long periods.

Six patients ambulated normally. The other 7 had neurologic injuries that interfered with gait. The ability to ambulate in 4 of the 7 patients was also affected by posterior pelvic pain. Eight patients had a neurologic problem. Two patients had decreased erections and 7 had weakness. Four of the 7 required an ankle foot orthosis. Three patients had pain consistent with causalgia.

Seven of 9 patients with greater than or equal to 15 mm of initial displacement had pain and 2 did not. Three of the 4 patients with less than 15 mm initial displacement had pain and 1 did not. Four of the 8 patients with an excellent reduction, 1 of 13 with a good reduction, and 0 of 2 with a fair reduction (10-20 mm) had pain.


The long term outcome of operatively treated displaced posterior ring injuries is to a large degree unknown. Several previous authors have reported high rates of posterior pain, impaired gait, pelvic obliquity, and inability to return to gainful employment.1,4,7,8,15,16,18,20,21 However, in the larger series, the treatment used was bedrest, traction, or pelvic slings.

Holdsworth4 reported on 50 pelvic ring disruptions treated with 12 weeks in a pelvic sling with or without traction. By using Bucholz's classification, Type II and Type III injuries were included. Holdsworth found that sacroiliac dislocations had poorer outcomes than sacral or iliac fractures with only 44% compared with 87% of patients returning to work.

Raf15 reported on 101 patients with double vertical fractures. Of these, the anterior displacement exceeded 5 mm in only 66 cases (66%). Twenty eight of the 66 (42%) had no craniocaudad displacement. Thus, only 38 (38%) of the series had vertically unstable fractures (Type III). Treatment of these included traction and some operative fixation of symphyseal widening. Followup was reported for 66 patients. Fractures through the ilium had a better prognosis for pain than sacral fractures or sacroiliac joint injuries. Back pain was a complaint in 29% of patients with iliac fractures, 38% of those with sacroiliac injuries, and 63% after sacral fractures. This was disabling in 6%, 14%, and 29% respectively. Again, in this series sacroiliac injuries included Type II injuries with anterior opening but intact posterior ligaments. Likewise, injuries to the sacrum that would be considered stable by today's standards were also included.

One of the most widely quoted series of pelvic fractures is from Slatis and Huittinen.18 The study was comprised of 163 patients. One hundred thirteen patients were treated with recumbency and a light pelvic sling for 6 weeks. Only 30 patients (18%) had displacement of the hemipelvis, indicating a Type III injury. Most of these patients had injuries through the sacroiliac joint as opposed to the remainder of the patients who sustained primarily sacral fractures. This, in conjunction with the fact that rami fractures outnumbered symphyseal separation by 5 to 1, indicates that most of the injuries occurred via lateral compression. Sixty-five patients were observed for 1 to 7 years. Major complaints included pelvic obliquity and impaired gait in 21 patients (32%), lumbosacral nerve deficit in 31 (48%), and disabling low back pain in 11 (17%).

Semba and coworkers16 observed 30 of 53 patients for 2 to 12 years. All pelvic ring injuries were defined as Malgaigne fractures. However, less than ½ of these injuries had more than 5 mm vertical translation and only ½ had displacement greater than 10 mm. Treatment was varied throughout the series, but was essentially nonoperative, with the exception of 2 cases. Only 36% of the patients observed were asymptomatic, 31% had impaired gait, and 26% complained of severe low back pain on the side of the injury. Greater than 10 mm combined anterior and posterior initial displacement correlated with pain. No patient with less than this amount of initial displacement had severe low back pain. Initial combined displacement of 10 mm essentially means 5 mm of hemipelvic displacement. A high percentage of the injuries were probably unstable posteriorly (Type III). Thus, it can be inferred that Type III injuries correlated with posterior pain whereas Type II injuries did not have long term pain.

In a review of patients treated at many institutions with varying methods including bed rest, traction, and external fixation, Tile20 reported a 36% incidence of pain. This number increased to 60% for vertically unstable injuries (Type III). He thought that pain was related to malreduction of the sacroiliac joint. Following this, Kellam7 reported results in patients treated with external fixation. He found that long term pain was related to the reduction for Type II and Type III injuries. For Type II injuries, greater than 2 cm displacement resulted in an 80% incidence of pain requiring analgesia as opposed to 0% with an acceptable reduction. For patients with Type III posterior disruptions, reduction within 1 cm yielded a 50% incidence of pain whereas malreduction was associated with a 77% incidence of pain. Overall, only 31% of patients with Type III disruptions had normal function.

Several authors have more recently reviewed smaller series of unstable injuries treated with open reduction and internal fixation, with or without external fixation. Browner et al1 treated 15 patients with emergent external fixation and delayed internal fixation. Twelve of the patients were observed for 1 to 4 years. Six of 12 (50%) went back to work. Those who did not return to work were divided between those with associated injuries and those who did not return to work because of their pelvic injury. Three of 12 (33%) had significant posterior pelvic pain. Of note, in patients who had pure sacroiliac dislocations, 6 of 7 were reduced to within 10 mm and 1 was not. There were no anatomic reductions. Browner et al postulated that this was in part due to the delay in performing the reduction. The difficulty of late reduction of the posterior complex was also reported by Matta and Tornetta12 and Matta and Saucedo.11 Simpson et al17 reviewed 16 patients with unstable sacroiliac dislocations treated with open reduction and internal fixation via an anterior approach and plating or stapling. This technique yielded excellent reductions, but long term followup of the patients with respect to function and pain were not reported.

Functional results for a series of 11 patients with 12 vertically unstable fractures were reported by Ward et al.22 Treatment was by internal fixation of the posterior injury augmented by internal or external fixation of the anterior injury. Nine patients walked without difficulty and only 2 had significant posterior pain. Ten patients were observed for more than 1 year. Ward et al felt that an accurate reduction was paramount in obtaining a good functional result.

This series of 48 unstable posterior pelvic injuries all treated with open reduction and internal fixation is large with respect to the currently available literature. There were no nonunions in the series. The approach to the operative reduction and methods of fixation have been previously reported.12

Two-thirds of the patients returned to their previous jobs without restriction and an additional 16% did not return because of associated injuries. Sixty-three percent had no pain or pain only with strenuous activity. Likewise, 63% of the patients ambulated normally and an additional 25% did not ambulate normally because of associated injuries. These results are comparable with or superior to the series reviewed above. This is especially significant considering that those series probably included Type II and Type III injuries.

Neurologic injury was found in 35% of the patients in the current study. This is similar to the incidences reported by other authors.5,6,10,14,15,17,23 Partial recovery of neurologic deficits have been reported by Patterson and Morton,14 Weis,23 and Majeed.9,10 Full recovery has not been observed, however. Matta and Saucedo made the observation that the course of the nervous injury is not affected by a satisfactory reduction. They thought that the ultimate prognosis of the nervous injury is determined at the time of injury. The current series confirms this finding. Only 1 patient had complete nervous system recovery, and this was an isolated problem with incontinence. Several other patients had improvement, but not complete resolution. Causalgia was a significant problem for 6 (13%) of the patients. Neurologic injury was found to be more common in patients with sacral fractures and pure sacroiliac dislocations than in patients with fracture dislocations. Although it has been postulated that some posterior pelvic pain may be from injury to the nerve roots innervating the sacroiliac joint, clinical nerve injury did not correlate with posterior pelvic pain. Of the 17 patients with clinically apparent neurologic injury, 10 had pain and 7 did not. This does not preclude a subclinical nerve injury as part of the cause of posterior pain.

The effect of reduction on pain, ambulation, and work related outcome is difficult to assess in the current series. Forty-five of 48 posterior injuries were reduced to within 10 mm of anatomic. Seventy-five percent of the patients were reduced to within 4 mm of anatomic. Other authors have used 10 mm as an acceptable reduction. With this in mind, the authors did not have enough patients with poor reductions to evaluate this factor. However, reduction to within 4 mm did not result in less posterior pain than reduction between 4 to 10 mm (Table 2).

Semba et al16 correlated initial displacement with long term pain. The authors compared the patients with greater than or equal to 15 mm of initial posterior displacement with the patients who had less than 15 mm initial displacement (Table 2). For the patients with greater than or equal to 15 mm initial displacement, 11 patients had pain and 23 did not; whereas for patients with initial displacement less than 15 mm, 7 had pain and 7 did not. Thus, initial displacement did not correlate with pain.

Van Gulik et al21 thought that more than 1 cm of leg length inequality created problems during sitting. None of the patients in the current study had greater than 1 cm of leg length inequality, so this premise could not be examined. The authors think that correction of leg length is extremely important, however.

There are several conclusions that can be made from this series of patients. Return to work, pain, and ambulation are better in this series than in other series previously reported; however, statistical comparisons were not done. This is most likely because of the reduction and stabilization of the posterior pelvic ring. These findings are despite the fact that the injuries in this series are all Type III and the earlier reports include many Type II patterns. However, adequate reduction did not positively influence neurologic injury.

As would be expected, there was a higher percentage of excellent reductions for the fracture dislocation group than in the sacral fracture or pure sacroiliac dislocation groups. This is because there is intact posterior ilium to guide the reduction. However, reduction to within 10 mm (as determined by the greatest measurement on the standard three views of the pelvis) seems to be adequate for functional results. This may be because the pelvis exhibits only structural properties with weightbearing and sitting as opposed to joints, in which arthritis may ensue in a short time if there is not an anatomic reduction. However, arthritis may in fact occur in these patients with longer followup. For this reason, the authors think that every effort should be made to obtain an anatomic reduction of all posteriorly unstable pelvic ring disruptions. The results of this series show that a significant percentage of patients with severe posterior pelvic ring disruptions can expect good return to function and gainful employment. The presence of neurologic and associated injuries can have profoundly deleterious effects on the outcomes of patients with pelvic ring injuries.


1. Browner B, Cole D, Graham M, et al: Delayed posterior internal fixation of unstable pelvic fractures. J Trauma 27:998-1006, 1987.
2. Bucholz R: The pathologic anatomy of Malgaigne fracture-dislocations of the pelvis. J Bone Joint Surg 63A:400-404, 1981.
3. Gylling S, Ward R, Holcroft J, et al: Immediate external fixation of unstable pelvic fractures. Am J Surg 150:721-724, 1985.
4. Holdsworth F: Dislocation and fracture-dislocation of the pelvis. J Bone Joint Surg 30B:461-466, 1948.
5. Huittinen VM: Lumbosacral Nerve Injury in Fracture of the Pelvis. Acta Chir Scand 138(Suppl):1-44, 1972.
6. Huittinen VM, Slatis P: Nerve injury in double vertical fractures. Acta Chir Scand 138:571-575, 1972.
7. Kellam J: The role of external fixation in pelvic disruptions. Clin Orthop 241:66-82, 1989.
8. Kellam J, McMurtry R, Paley D, et al: The unstable pelvic fracture: Operative treatment. Orthop Clin North Am 18:25-41, 1987.
9. Majeed SA: Grading the outcome of pelvic fractures. J Bone Joint Surg 71B:304-306, 1989.
10. Majeed SA: Neurologic deficits in major pelvic injuries. Clin Orthop 282:222-228, 1992.
11. Matta J, Saucedo T: Internal fixation of pelvic ring fractures. Clin Orthop 242:83-97, 1989.
12. Matta JM, Tornetta III P: Internal fixation of unstable pelvic ring injuries. Clin Orthop 329:129-140, 1996.
13. Mears D, Fu F: Modern concepts of external skeletal fixation of the pelvis. Clin Orthop 151:65-72, 1980.
14. Patterson FP, Morton KS: Neurological complications of fractures and dislocations of the pelvis. J Trauma 12:1013-1023, 1973.
15. Raf L: Double vertical fractures of the pelvis. Acta Chir Scand 131:298-305, 1966.
16. Semba R, Yasukawa K, Gustilo R: Critical analysis of results of 53 Malgaigne fractures of the pelvis. J Trauma 23:535-537, 1983.
17. Simpson L, Waddell J, Leighton R, et al: Anterior approach and stabilization of the disrupted sacroiliac joint. J Trauma 27:1332-1339, 1987.
18. Slatis P, Huittinen VM: Double vertical fractures of the pelvis. Acta Chir Scand 138:799-807, 1972.
19. Slatis P, Karaharju E: External fixation of unstable pelvic fractures: Experience in 22 patients treated with a trapezoid compression frame. Clin Orthop 151:73-80, 1980.
20. Tile M: Pelvic ring fractures: Should they be fixed? J Bone Joint Surg 70B:1-12, 1988.
21. Van Gulik T, Raaymakers E, Broekhuizen A, et al: Complications and late theraputic results of conservatively managed unstable pelvic ring disruptions. Neth J Surg 39:175-178, 1987.
22. Ward E, Tomasin J, Vander Griend R: Open reduction and internal fixation of verticalshear pelvic fractures. J Trauma 27:291-295, 1987.
23. Weis EB: Subtle neurological injuries in pelvic fractures. J Trauma 24:983-985, 1984.

Section Description



© Lippincott-Raven Publishers.