Unstable pelvic ring disruptions occur after a high energy mechanism and are commonly associated with multiple injuries. The acute treatment of these patients requires a multidisciplinary team approach including the trauma surgeon, orthopaedist, and interventional radiologist. Bleeding can occur from abdominal, thoracic, extremity, and pelvic sources. All bleeding sites must be found and treated expeditiously. Initial stabilization of the pelvic ring with an external fixator can help to reduce bleeding sites and encourage tamponade by closing and maintaining the pelvic volume. However, external fixation does not control posterior pelvic disruption well.13
Several studies have found internal fixation to be superior to both external fixation and conservative treatment in managing unstable pelvic ring disruptions11 (nonpublished data, Harvell J, Mears D: Late result of pelvic ring disruptions managed by external fixation, or combined methods. Presented at the 53rd annual meeting of the American Academy of Orthopedic Surgeons. New Orleans, 1986). Patients treated by internal fixation had less pain, better reductions, fewer losses of reduction, and fewer malunions.
Many types of internal fixation have been promoted in recent years. These include short plates, long plates, or double plating for the symphysis and sacral bars, cobra plates, anterior plates, or iliosacral screws for posterior disruptions.4 Yet despite the recent advances in fixation techniques, few reports critically analyze the reductions obtained.
This paper critically analyzes the reduction, union, and the incidence of early complications in a series of unstable pelvic ring injuries treated with a single methodologic approach.
MATERIALS AND METHODS
One hundred seven patients with unstable pelvic ring injuries treated by open reduction and internal fixation were reviewed. Instability was defined as the need for fixation to hold the pelvis in an appropriate position for union. In general, the indication for posterior fixation was an unstable posterior ring injury with displacement. The indication for anterior fixation was a symphysis diastasis.
The minimum followup time to be included was 6 months, and all patients were observed to union. All patients were operated on within 70 days of injury (average, 14 days). Eighty-seven of the patients were operated on within 21 days and 20 were operated on after 21 days from their injury. Ninety patients had associated injuries of which 63 required surgery (Table 1). There were 5 patients with open fractures and 12 patients with closed degloving injuries. Eighteen patients sustained neurologic injuries. There were 38 associated acetabular fractures that were all treated operatively.
Pelvic ring injuries were classified according to Bucholz3 as rotationally unstable and vertically stable (Type II) or rotationally and vertically unstable (Type III). There were 38 Bucholz Type II and 69 Bucholz Type III injuries. Any displacement in a posterior structure was recorded as an injury and classified by anatomic location as described by Letournel10(Table 2). This includes anterior opening of the sacroiliac joint and sacral fractures in Bucholz Type II injuries.
The anteroposterior (AP), 40 ° caudad (inlet), and 40 ° cephalad (outlet) radiographs of the pelvis were available for all patients. Displacements were measured to the nearest millimeter on all 3 views preoperatively. All displacements were recorded. Leg length discrepancy was measured as the difference in height of the femoral head from a line perpendicular to the long axis of the sacrum on the AP radiograph because this most closely represents the anatomic position of the pelvis during gait (Fig 1). These same measurements were repeated using the postoperative and 6-month followup radiographs.
Operative reports, inpatient charts, postoperative radiographs, and office notes were reviewed to determine the operative approach, blood loss, time of surgery, type of fixation used, and any complications that arose in the perioperative period.
Reductions were graded by the maximal displacement measured on the 3 views of the pelvis: excellent (≤ 4 mm), good (4-10 mm), fair (10-20 mm), and poor (> 20 mm). This scheme is somewhat arbitrary and may not have the same meaning for all injuries. For example, a 10-mm displacement of the ilium at union may be less symptomatic than a 10-mm displacement of the sacroiliac joint.
Thirty-one patients had only anterior fixation. Sixty-nine patients had unilateral posterior injuries fixed and 7 had bilateral posterior injuries fixed. This includes iliac fractures and fractures of the acetabulum with symphysis disruptions that were graded as Bucholz Type II. Thus, there were 83 posterior injuries fixed.
With few exceptions, all fractures and dislocations were opened and reduced under direct vision. A Pfannenstiel incision was used for anterior injuries and a vertical posterior approach lateral to the sacroiliac joint was used for dislocations and fracture dislocations of the sacroiliac joint unless an ilioinguinal approach was needed for an ipsilateral acetabular fracture.11 Eleven patients with sacroiliac dislocations and 4 with fracture dislocations were approached through an ilioinguinal incision. Seventeen sacroiliac dislocations and 30 fracture dislocations were approached posteriorly. All sacral fractures were approached posteriorly. Fractures through the ilium were fixed using an incision along the crest with an approach to the inner or outer table depending on the associated pathology. These approaches were previously described.11,12
Thirty-three patients had anterior and posterior fixation during the study period. Nine had the anterior injury approached first, 18 had the posterior injury approached first, and 6 had a simultaneous anterior and posterior exposure. The remaining 43 patients had posterior fixation without anterior fixation.
Because the reduction techniques developed during the course of the study, the methods herein described are those currently used and some cases in the study group were not treated with exacting protocol. The present protocol is to fix the posterior pelvic ring disruption first with the patient positioned prone. One exception to this is if the symphysis is dislocated and both innominate bones are intact (no fracture of the rami, acetabulum, or ilium on either side). In this special situation, the symphysis may be approached first. Reduction and fixation of the symphysis helps to decrease posterior displacement. The patient is then turned to a prone position for open reduction of posterior injury. If the anterior injury is rami fractures or a symphysis dislocation with rami fractures then the posterior injuries are addressed first.
Of note, different methods were often needed for the patients treated more than 21 days postinjury. Before the posterior injury can be addressed, the anterior pelvic ring disruption requires mobilization. Thus in these cases, an anterior incision is used first and the fractures of the rami or dislocation of the symphysis is mobilized. This is done by taking down the callus and scar interposed in the area of the injury. This is done as a first stage so that the partially healed anterior injury will not block the reduction of the posterior injury.
For most injuries, the following methods were used in obtaining the reduction. For the symphysis, a Weber clamp (Synthes, Inc, Paoli, PA) is placed anterior to the rectus muscles onto the body of the pubis bilaterally. If there is any anterior displacement then the point of the forceps is placed more anterior on that side to effect a reduction force. Thus, the forceps are placed such that once the reduction is achieved the points are at the same level (Fig 2). Cases with cephalad displacement of the hemipelvis are more difficult to reduce. The pelvic reduction forceps is used to aid in this reduction. One 4.5 screw is placed anteriorly on each side of the symphysis. On the side with the posterior displacement the screw is placed through a 4.5 gliding hole and anchored to the bone with a small plate and nut from within the pelvis (Fig 3). The Jungbluth clamp (pelvic reduction forceps; Synthes, Inc, Paoli, PA) can then be used to full mechanical advantage without the risk of screw pullout (Fig 4). Once the reduction is achieved, a 6-hole curved 3.5 reconstruction plate (Synthes, Paoli, PA) on the superior surface of the symphysis is used for fixation. A small amount of compression can be obtained by eccentrically drilling the medial holes. Double plating is used only in Type III injuries when it is not certain that posterior fixation is possible during the initial procedure, such as in a case of a patient undergoing emergent laparotomy.
Dislocations of the sacroiliac joint are reduced using angle jaw forceps designed by the senior author. One tip is placed around the sciatic notch onto the anterior portion of the sacral ala lateral to the foramen and the other is placed on the outer aspect of the ilium (Fig 5). This direction of placement allows a strong reduction vector perpendicular to the sacroiliac joint. The reduction is judged posteriorly by direct vision, anteriorly by palpating the joint through the notch, and by using triplane fluoroscopy. Cephalad displacement can be addressed using the Weber clamp (Fig 6). Manipulation of the femoral distracter placed between the posterior spines can also help to reduce cephalad displacement. Once the reduction is obtained, the joint is fixed with 2 6.5-mm short thread (16 mm) cancellous screws from the outer aspect of the ilium into the body of S1. A clear understanding of the AP, 40 ° cephalad, and 40 ° caudad views allow for safe placement of the screws. They must be superior to the S1 foramen and inferior to the superior margin of the ala on the 40 ° cephalad view, and between the neural canal and the anterior margin of the body on the 40 ° caudad view. In addition to fluoroscopic guidance, the surgeon must develop a feel for where the drill bit is as it passes through the joint and into the sacrum. An oscillating drill is used to decrease the chance of injury should the drill be misplaced. This drill also gives excellent proprioceptive feedback. The drill bit can be felt to pass through both cortices of the ilium and then the lateral cortex of the sacrum. After passing into the sacrum, the drill is advanced using a gentle pistoning action that enhances the surgeon's ability to delineate cancellous from cortical bone. No further cortex should be crossed. If a fourth cortex is encountered, drilling must cease because this indicates proximity to the S1 foramen, the anterior sacrum, the cephalad surface of the sacral ala, or the posterior neural canal. Drilling into any of these areas can result in significant complications. Washers should be used in older persons because the pull of the screw is often strong enough to pass through the ilium.
Fracture dislocations of the sacroiliac joint can be reduced by reconstructing the ilium anatomically. If the intact portion of ilium is large and firmly attached to the sacrum, fixation can be restricted to the ilium without placing iliosacral screws. If the fragment is small or the integrity of the posterior ligaments questionable then the ilium is fixed anatomically and 1 or 2 iliosacral screws are placed to further support the reduction.
Fractures of the sacrum are reduced under direct vision using the posterior spines to aid in manipulation. In a transforaminal fracture, the nerve roots are visualized through the fracture and palpated during the reduction using a finger through the notch. Loose pieces of bone abutting the nerve roots should be removed before reduction and fixation. The iliosacral screws are then placed as described and tightened maximally. Palpation of the neural foramen and nerve roots anteriorly can help to prevent neural impingement before it occurs. In some cases, the width of the sacrum is slightly narrowed, but there were no cases of new neurologic injuries postoperatively in any of the cases. Anatomic restoration of the foramen protects the roots from damage during compression of the fracture.
Fractures of the ilium are reduced using standard methods and fixed with 3.5-mm lag screws along the crest or with reconstruction plates.
The methods of fixation used for the 83 posterior injuries that were repaired are shown in Table 3.
Displacements and Reduction
Displacement was measured as the maximum point to point distance between the fragments or separated portions of the pelvis on each of the 3 views of the pelvis. The largest preoperative displacement measured on the 3 views of the pelvis were recorded. Displacements averaged 36 mm (range, 3-95 mm) anteriorly and 10 mm (range, 0-30 mm) posteriorly for Type II disruptions and 49 mm (range, 10-125 mm) anteriorly and 20 mm (range, 4-50 mm) posteriorly for Type III injuries. Ninety-nine patients had a leg length discrepancy that averaged 6 mm for Type II and 12 mm for Type III injuries. The largest displacement was seen on the 40 ° caudad view in 90% of the cases.
Postoperative displacements averaged 4 mm anteriorly and 2 mm posteriorly for Type II disruptions and 6 mm anteriorly and 4 mm posteriorly for Type III injuries (Fig 7). These values and those used to grade the adequacy of reduction excluded rami fractures that were not fixed because there was generally 5 to 10 mm of residual displacement of the fracture. This amount of displacement was considered to be clinically insignificant, but would artificially worsen the grading of the reduction. For the 99 patients who had leg length discrepancies preoperatively, the average postoperative measurement was decreased to 1 mm (range, 0-15 mm).
Using the grading system described there were 72 excellent, 30 good, 4 fair, and 1 poor reductions. However, there was a higher percentage of excellent reductions in patients operated on within 21 days (61 of 87 = 70%) compared with those operated on after 21 days (11/20 = 55%). However, these differences were not statistically significant. In both groups, 95% of the reductions were either excellent or good.
For patients who had both anterior and posterior injuries fixed, the order of fixation (front or back first) did not result in a statistically significant difference in the quality of the reductions.
Nonsurgical perioperative complications were uncommon. There were 2 urinary tract infections, 4 deep vein thromboses, and 1 nonfatal pulmonary embolus.
Operative complications included 1 femoral vein laceration, 1 bladder laceration, 4 broken drill bits, 3 wound hematomas, and 1 transient L5 neuropraxia.
Deep infections occurred in 3 patients. Two infections occurred posteriorly. An infection developed in 1 patient at the site of a posterior approach performed through an ipsilateral soft tissue degloving injury. The infection resolved after incision and drainage and treatment with antibiotics. Five of the other 11 patients who had soft tissue degloving injuries required repeat debridements but none became infected. The other deep posterior infection occurred in a patient with a Type III pelvic disruption who also sustained bilateral femur fractures, a contralateral both column acetabular fracture, rib fractures, a lumbosacral plexopathy, and a severe soft tissue injury to his lower abdomen. A deep infection developed at the posterior incision used to fix a fracture dislocation. The area was debrided twice and he received a 6 week course of antibiotics. Drainage stopped only after removal of the hardware.
One patient became infected anteriorly and healed with the hardware in place after local incision and drainage. The likely source of this infection was an abscess in the proximal medial thigh that eventually communicated with the symphysis. An infection developed at the operative site of the acetabular fracture that did not affect the pelvic fixation in 2 patients who had simultaneous open reduction and internal fixation of the acetabulum.
Two patients required fixation of additional components of their injury that were not fixed at the index procedure. One patient had a symphysis dislocation, a left ilium fracture, and opening of the right sacroiliac joint that measured 12 mm. The symphysis and left ilium were fixed during the index procedure. Persistent 10-mm displacement of the right sacroiliac joint was treated by open reduction and internal fixation 6 days later.
The second patient sustained bilateral rami fractures and a right sacral fracture. The patient was operated on more than 21 days after injury. During the procedure, bilateral rami fractures were mobilized surgically to allow for accurate reduction of the right sacrum. The sacrum was reduced and fixed. The rami reduced from 25 mm (right) and 10 mm (left) to 4 mm on each side after posterior stabilization and were not fixed. The displacement subsequently increased to 25 mm necessitating open reduction and fixation of both the rami 10 days after the initial procedure. In both patients the final reduction was maintained until healing.
Two patients had loss of fixation. One patient had bilateral rami fractures, an acetabular fracture, and 5 mm of sacroiliac widening. The 2 medial screws in the 6-hole anterior plate loosened and the plate was replaced with an 8-hole plate. There was no future loosening or displacement. The other patient was the only 1 in the series who lost posterior fixation and had redisplacement. He had left rami and bilateral sacral fractures. The sacral fractures were fixed with 1 iliosacral screw on each side and a posterior tension band. The preoperative displacement of the right side was 6 mm and the left side was 35 mm. The postoperative displacement of the right side was 4 mm and the left side was 9 mm. When the patient returned for his 1-month followup radiograph, it was evident that the reduction of the left hemipelvis was lost and that the displacement had increased. A rereduction and fixation with 2 iliosacral screws was performed and the pelvis healed uneventfully.
Five patients with only anterior fixation had a fatigue fracture of the plate. All occurred more than 8 weeks postoperatively and none lost their reduction. The initial displacement averaged 42 mm anteriorly and 12 mm posteriorly for these patients compared with 36 mm anteriorly and 10 mm posteriorly for all Type II injuries. This difference was not significant. However, only 1 of the patients had a reduction graded as excellent. The other 4 were graded as good (2) and fair (2). Thus, although the numbers are too small to be statistically significant, the quality of the reduction may be related to plate failure in Type II injuries. No patient who had anterior and posterior fixation had a failure of the anterior plate.
All other patients healed without any displacement or hardware failure. There were no instances of neural impingement caused by maximally compressing sacral fractures in this series. It must be noted, however, that Browner et al2 cautioned to avoid excessive tightening of the screws because this can cause neural impingement by overreduction of the fracture.
Nonoperative treatment of unstable pelvic ring disruptions requires prolonged immobility and yields poor results.7,11,14 External fixation is used in the acute phase after injury to add to bony stability during transport, decrease the pelvic volume, and allow patients to avoid recumbency.6 However, external fixation of the pelvis cannot adequately maintain the alignment of posterior injuries, is not as strong biomechanically as internal fixation, and risks pin tract infection.1,4
Internal fixation of the pelvis is significantly stronger than external fixation, even in Type II injuries.18 Kellam10 was able to relate the quality of the posterior element reduction to outcome. Van Gulik et al19 found that leg length discrepancy was an important factor in outcome. For these reasons, recent authors have recommended internal fixation of unstable pelvic ring disruptions2,5,9,11,16,20 (nonpublished data, Harvell J, Mears D; Late result of pelvic ring disruptions managed by external fixation, or combined methods. Presented at the 53rd annual meeting of the American Academy of Orthopedic Surgeons. New Orleans, 1986). Yet despite these recommendations, few authors have critically analyzed the reductions obtained in a standardized way and most do not include information about leg length discrepancy.
Kellam8 defined an acceptable reduction to be less than 10 mm posteriorly and less than 20 mm anteriorly, but did not describe the radiograph used to evaluate the reduction. Slatis and Karaharju17 graded their reductions only by posterior displacement as measured on an AP film; with excellent being less than 5 mm, good 5 to 10 mm, and poor greater than 10 mm. Semba et al15 evaluated 53 Malgaigne fractures and found that initial combined anterior and posterior displacement of greater than 10 mm leads to a high rate of severe low back pain, but they did not report on the reduction at all. Ward et al20 reported on 12 vertical shear injuries treated with open reduction and internal fixation but did not objectively assess the quality of the reductions. Goldstein et al5 reported on 15 patients treated with early open reduction and internal fixation but did not evaluate the reductions obtained. Browner et al2 reported no anatomic reductions in 7 sacroiliac dislocations, but only 1 case had greater than 10 mm displacement. No information was given regarding their method of measurement.2 Simpson et al16 obtained anatomic reductions using an anterior approach to the sacroiliac joint, but no specific measurements were given.
The authors critically analyzed preoperative and postoperative AP, 40 ° caudad, and 40 ° cephalad views of 107 patients with unstable pelvic ring injuries to assess the reductions and the ability to maintain them. Reductions were graded based on the largest displacement measured on the 3 views. This is an extremely important part of the evaluation. Most commonly, maximal displacements were seen on the 40 ° caudad view. It was also found that this view was particularly helpful in delineating displacements of the posterior pelvic ring and rotational deformities about an axis parallel to the xray beam. The 40 ° cephalad view was helpful in visualizing sacral fractures and rotation in the sagittal plane.
Using the operative techniques described reductions were obtained within 4 mm in 67% and within 10 mm in 95% of cases. In cases done within 3 weeks of injury, 70% were reduced to within 4 mm, whereas in those done after this time frame, only 55% were within 4 mm. Leg length discrepancy was also decreased to less than 10 mm in 102 of the 107 patients. This is considered to be an acceptable level to avoid sitting problems.19
Several authors have voiced concerns about infection in the open treatment of pelvic ring injuries.4,9,18 Kellam et al9 reported a 25% rate of infection in patients who had sustained a crush injury. Goldstein et al5 had an 18% infection rate in fractures fixed early and 27% in those fixed late. However, other authors have reported much lower infection rates.2,11,20
There were only 3 deep infections in the present series (2.8%) and it is thought that open reduction of the pelvic ring need not be considered a high risk procedure. However, care must be taken to recognize severe soft tissue injury and treat it appropriately. The authors routinely perform an incision, drainage, and debridement of closed subcutaneous degloving injuries. These injuries represent a separation of the fat and subcutaneous tissue from the deep fascia. This space then becomes filled with a hematoma and variable amounts of fat necrosis. Following incision and drainage, the wound is left open, packed daily, and allowed to close secondarily.
The 2 patients who required further fixation highlight the need for close interval radiographs when not all components of a patient's pelvic ring disruption are fixed. Displacement can be easily treated if it is detected early. Five patients had a late fatigue fracture of anterior plates without loss of reduction. Although this finding implies motion at the symphysis, because no patient lost their reduction, these plate failures seem to be clinically inconsequential. It is interesting to note that only 1 of these failures occurred after an excellent reduction. The possibility exists that a perfect reduction may prevent some motion through the symphysis.
The 1 loss of posterior fixation in the series occurred in a patient whose bilateral sacral fractures were treated with only 1 iliosacral screw on each side. The authors now use 2 screws whenever possible.
A recent biomechanical study showed that iliosacral screws allowed less lateral sacroiliac motion than anteriorly placed plates or sacral bars (nonpublished data, Schopfer A, Hearn T, DiAngelo D, Tile M: Biomechanical comparison of fixation methods of vertically unstable pelvic ring with sacroiliac dislocation. Presented at the 8th annual meeting of the Orthopedic Trauma Association. Minneapolis, MN 1992). The authors prefer screw fixation of the sacroiliac joint because it is stronger, and the pull of the screw effects a reduction force on the joint because of its placement in an interfragmentary compression mode. Also, the main stress of the posterior pelvis during weightbearing is in vertical shear, and iliosacral screws are supported against this stress by the entire length of the screw within the sacrum. Sixteen-millimeter thread length screws are used so that this stress is borne almost exclusively by the thick shank of the screw. Additionally, longer screws provide better fixation because there is more resistance to toggle. The authors prefer to place solid rather than cannulated screws across the sacroiliac joint because the oscillating drill provides excellent proprioception and the 3.2-drill bit is rigid, allowing for more directional control than thinner cannulated screw guide wires. Without the proprioceptive feedback described earlier, screw placement is completely reliant on intraoperative radiologic interpretation. There is a greater risk of misplacement under these circumstances. Additionally, thin guide wires can deform to follow curved surfaces. This can cause difficulty in seating the screws and weaken the screws via a bending moment.
Having used both anterior and posterior approaches and fixation for sacroiliac joint dislocation, the authors can make several subjective statements regarding these techniques. Posterior reduction and fixation is preferred whenever possible. The posterior approach allows superior placement of clamps to effect a reduction, allows for visualization of the back of the joint combined with assessment of the anterior joint by palpation, and neurologic injury is less likely. Anterior reduction and plating of the sacroiliac joint has several potential problems. The strength of the fixation is weaker than sacroiliac screws, particularly when stressed in the cephalad direction. Visualization of the cephalad portion of the joint is good, but the posterior joint may open when the plates are placed. More traction is placed on the L5 and S1 nerve roots by retractors and the L5 root is extremely close to the operative field and fixation. Finally, anterior plate fixation must not be used when there is any injury to the sacral ala. Loss of reduction will occur by pullout of the sacral side fixation. However, the anterior approach is often best if there is an associated acetabular fracture, and the incision is more cosmetic.
This study supports the use of a single 6-hole curved 3.5 reconstruction plate for anterior fixation of Type II injuries or in Type III injuries when combined with posterior fixation. This series includes 43 patients who had posterior fixation without anterior fixation. This highlights the stable nature of rami fractures. Most rami fractures occur through a lateral compression injury, leaving the inguinal ligament intact. This ligament helps to support the reduction of the rami once the posterior portion of the injury is reduced and fixed. Additionally, because of the excellent soft tissue sleeve, the rami heal quickly. Thus, only in severely displaced rami fractures should fixation be considered.
These preferred methods of stabilization are not always the most practical. Stabilization must be tailored to the individual fracture pattern. In attempting to reduce and fix these complex injuries the surgeon must be familiar with all techniques and able to apply them confidently. Malunion and nonunion are not problems when properly done open reduction and internal fixation is done initially. If one is unaccustomed to operating in the region of the pelvis, then external fixation is an excellent option until the patient can be transferred to a surgeon specializing in these injuries.
1. Bosch U, Pohlemann T, Tscherne H: Primary management of pelvic injuries. Orthopade 21:385-392, 1992.
2. Browner B, Cole D, Graham M, et al: Delayed posterior internal fixation of unstable pelvic fractures. J Trauma 27:998-1006, 1987.
3. Bucholz R: The pathologic anatomy of Malgaigne fracture-dislocations of the pelvis. J Bone Joint Surg 63A:400-404, 1981.
4. Failinger M, McGanity P: Current concepts review: Unstable fractures of the pelvic ring. J Bone Joint Surg 74A:781-791, 1992.
5. Goldstein A, Phillips T, Sclafani S, et al: Early open reduction and internal fixation of the disrupted pelvic ring. J Trauma 26:325-333, 1986.
6. Gylling S, Ward R, Holcroft J, et al: Immediate external fixation of unstable pelvic fractures. Am J Surg 150:721-724, 1985.
7. Holdsworth F: Dislocation and fracture-dislocation of the pelvis. J Bone Joint Surg 30B:461-466, 1948.
8. Kellam J: The role of external fixation in pelvic disruptions. Clin Orthop 241:66-82, 1989.
9. Kellam J, McMurtry R, Paley D, et al: The unstable pelvic fracture: Operative treatment. Orthop Clin North Am 18:25-41, 1987.
10. Letournel E: Fractures of the Acetabulum. New York, Springer 1981.
11. Matta J, Saucedo T: Internal fixation of pelvic ring fractures. Clin Orthop 242:83-97, 1989.
12. Matta JM: Surgical Approaches to Fractures of the Acetabulum and Pelvis. JM Matta, MD, Inc, Los Angeles, CA 1989.
13. Mears D, Fu F: Modern concepts of external skeletal fixation of the pelvis. Clin Orthop 151:65-72, 1980.
14. Raf L: Double vertical fractures of the pelvis. Acta Chir Scand 131:298-305, 1966.
15. Semba, R, Yasukawa, K, Gustilo, R: Critical analysis of results of 53 Malgaigne fractures of the pelvis. J Trauma 23:535-537, 1983.
16. Simpson L, Waddell J, Leighton R, et al: Anterior approach and stabilization of the disrupted sacroiliac joint. J Trauma 27:1332-1339, 1987.
17. 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.
18. Tile M: Pelvic ring fractures: Should they be fixed? J Bone Joint Surg 70B:1-12, 1988.
19. 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.
20. Ward E, Tomasin J, Vander Griend R: Open reduction and internal fixation of vertical shear pelvic fractures. J Trauma 27:291-295, 1987.