Sacral fractures are difficult injuries to diagnose, evaluate, and treat. Several papers emphasize that the initial diagnosis is missed in a high percentage of patients. Delayed and missed diagnoses are attributed to the severity of associated injuries and the difficulties in obtaining clear radiographs of the sacrum.6,17 Most reports describe associated neurologic injuries and their long term sequelae. The natural history of sacral fractures is not well described.
Recent interest in the treatment of pelvic ring injuries has led to a better understanding of sacral fractures. Two large series document that sacral fractures are common components of pelvic ring injuries. Pohlemann et al17 diagnosed sacral fractures in 28% of patients with pelvic ring injuries (377 sacral fractures/1350 pelvic fractures); and Denis et al6 diagnosed sacral fractures in 30% of pelvic ring injuries (236 sacral fractures/776 pelvic fractures).
It is difficult to determine the best treatment of sacral fractures from a review of the literature. Most series are small or describe several techniques for internal fixation. Methods used for the stabilization of sacral fractures include: transiliac sacral bars; cobra plates, posterior transiliac reconstruction plates; local osteosynthesis with plates applied to the sacral lamina; and iliosacral screw fixation with either open or closed reductions.1,6,7,13,14,17,19
The indications for the internal fixation of sacral fractures are controversial. Most authorities agree, however, that the pelvis is unstable when posterior displacement exceeds or is equal to 1 cm.4,5,7,13,18,19,22 Because this is an objective measurement, patients with more than 1 cm of displacement were studied to determine the results of fixation by the technique of iliosacral screw fixation.
MATERIALS AND METHODS
A retrospective study of patients with sacral fractures that were displaced at least 1 cm was conducted. Between January 1, 1991 and October 1994, 30 patients with displaced sacral fractures underwent internal fixation. The procedures were performed by 4 different surgeons (PD, JG, DO, DT) at The Oregon Health Sciences University (Portland, OR), University Michigan Medical Center (Ann Arbor, MI), University of California at Davis Medical Center (Sacramento CA), and Hennepin County Medical Center (Minneapolis, MN).
There were 21 males and 9 females. The age at injury averaged 36 years (range, 15-70 years). The mechanism of injury included: motor vehicle accident in 11 patients; motorcycle accident in 5 patients; pedestrian accidents in 5 patients; falls in 3 patients; and various causes in 6 patients. Only 3 patients had isolated pelvic injuries; the remainder had other fractures and/or injuries to other organ systems. The average injury severity score was 22 points (range, 9-59 points). Two patients, ages 60 and 70, died at 7 months and after more than 2 years, respectively, from the time of their injuries. Both patients had multiple medical illnesses and died from causes other than their fractures. Table 2
A retrospective review of the medical records and radiographic studies was performed. Patients were reexamined and/or contacted by telephone to complete a patient oriented assessment (Iowa pelvic score).12 This scoring system asks patients their ability to perform activities of daily living, work status, degree of pain, perception of cosmetic change, and ability to walk (Table 1).
Maximum displacement of the anterior and posterior pelvic rings was measured in millimeters from preoperative and postoperative radiographs. Measurements were taken from the following pelvic projections: anteroposterior, inlet or caudad view, and outlet or cephalad view.21,22 The maximum displacement of the 3 projections was recorded for each patient.
Pelvic ring injuries were classified according to Bucholz3 and Tile21 as rotationally unstable and vertically stable (Tile B, Bucholz II), or rotationally and vertically unstable (Tile C, Bucholz III). There were 5 patients with bilateral posterior pelvic ring injuries. In addition to the displaced sacral fracture, the contralateral injury was a nondisplaced sacral fracture in 2 patients; a sacral fracture with less than 1 cm of displacement in 2 patients; and a fracture dislocation of the sacroiliac joint in 1 patient. Specific to the sacral fractures, there were 27 rotationally and vertically unstable injuries (Bucholz III, Tile C) and 3 rotational injuries (Bucholz II, Tile B2 and B3). Sacral fractures were classified as described by Denis et al6(Fig 1). The location of the sacral fracture was: Zone I (alar) in 2 patients; Zone II (transforaminal) in 26 patients; and Zone III (sacral canal) in 2 patients. Neither of the 2 Denis Zone III injuries were transverse sacral fractures. Twelve patients presented with neurologic injuries. These ranged in severity from radicular symptoms to motor loss.
The method of operative stabilization was determined by each surgeon's assessment of the pelvic ring injury. The techniques used for the reduction and fixation of sacral fractures are described by Matta and Saucedo,13 Duwelius et al,7 and Routt et al.18 Open reduction was performed in 17 patients. Closed reduction and fluoroscopic insertion of iliosacral screws was performed for 13 patients; 6 had the initial reduction and fixation of a symphysis diastasis that improved the displacement of the sacral fractures so that a closed reduction was effected. Percutaneous fixation was guided by CT imaging for fixation in 3 patients.
The average delay to surgery was 7 days. This delay reflects a wide range in the time from injury to fixation (range, 0-20 days) and is a function of delays in the transfer to tertiary centers rather than the surgeon's preference for the timing of treatment. The delay to both open reduction and percutaneous fixation averaged 7 days. In the past 2 years of the series, however, there was a trend to perform percutaneous fixation on the first day after the injury.
All fractures were stabilized with iliosacral lag screws. Partially threaded stainless steel screws, 7-mm cannulated screws with 16 or 32-mm threads (Synthes, Paoli, PA), or Ti 7-mm cannulated screws (Ace, Los Angeles, CA) were used to gain interfragmentary compression. The surgeon tightened the screws to achieve compression across the sacral fracture. Transiliac plates were added to improve stability if the sacral fracture seemed unstable after the insertion of the iliosacral screw(s), or if the fracture plane across the sacral lamina opened when reduction clamps were removed. Transiliac plates were used in 6 patients.
Treatment after surgery was directed by the presence of other injuries. For patients who were physically able, walking with crutches or a walker with touch weightbearing only on the side of the sacral fracture was prescribed. Protected weightbearing was ordered for 10 to 12 weeks.
The preoperative displacement of the sacral fracture averaged 20 mm (range, 10-60 mm). The postoperative displacement averaged 6 mm (range, 0-14 mm). The preoperative displacement of the anterior ring (rami fractures or symphysis diastasis) averaged 26 mm (range, 10-80 mm). The postoperative displacement of the anterior ring averaged 7 mm (range, 1-25 mm).
In the 17 patients who underwent open reduction, the preoperative displacement averaged 24 mm and the postoperative displacement averaged 4 mm. In the 13 patients in whom percutaneous fixation was done, the preoperative displacement averaged 15 mm and the postoperative displacement averaged 5 mm.
All 30 fractures united. Three patients had partial loss of their operative reduction due to weightbearing. In the 3 patients the fractures united with approximately 1 cm of final displacement. All the patients (27 of 30) who remained nonweightbearing for 10 to 12 weeks maintained the reduction achieved at surgery.
Twenty-seven patients completed questionnaires (Iowa pelvic score), 2 died and 1 was lost to followup. The mean time to followup was 28 months (range, 12-57 months). The Iowa pelvic score averaged 79 points (range, 44-100 points, standard deviation 16 points). Fourteen patients (47%) had various manifestations of a neurologic injury; 12 of the patients completed questionnaires. The pelvic score for patients with a neurologic injury averaged 70 points (range, 48-95 points). The patient with the lowest score reported pain in the S1 distribution, had a normal motor examination, and was suspected of having a sympathetic dystrophy. The patient with a score of 95 points reported no pain and had weak dorsiflexion of the foot, but returned to heavy labor in a rural setting. For patients without neurologic injury, the pelvic score averaged 87 points (range, 48-100 points). The difference in the Iowa pelvic score for patients with a neurologic injury (mean, 70 points) and patients without a neurologic injury (86 points) was statistically significant (Student's t test; p = 0.008).
Sixteen of 30 patients returned to full time work, performing the same type of work as before their injuries. Six of the 16 patients had neurologic injuries. Six patients were unable to do any type of work and had been classified as totally disabled. Four of the patients had neurologic injuries and complained of pain. It was not possible to make an objective assessment comparing the 6 patients with nerve injuries who returned to work, with the 4 patients with nerve injuries who did not return to work. Thirteen of the 15 patients who did not have nerve injuries returned to full time work.
Iatrogenic neurologic injuries occurred in 2 patients. The first patient underwent open reduction and internal fixation of a disrupted symphysis pubis followed by prone positioning and open reduction of a Zone II sacral fracture. The other patient underwent closed reduction and percutaneous fixation in the supine position. Postoperative CT scans indicated successful positioning of the iliosacral screws and patent neural foramina. It is thought that the nerve injuries occurred during manipulation of the fractures. Intraoperative monitoring was not performed in either of the 2 cases. Both patients lost active dorsiflexion of the ipsilateral foot.
There were 2 postoperative infections; both were successfully treated with operative debridements and antibiotics. One infection occurred after an open reduction. The other infection occurred after percutaneous fixation; this patient also had a colonic perforation that was thought to have seeded the fracture site despite the performance of a diverting colostomy.
Four patients were treated for deep venous thrombosis. One patient had a pulmonary embolus before surgery and received a vena caval filter. Two of the patients treated for deep venous thrombosis had equivocal duplex ultrasound studies.
Classification systems that describe pelvic ring injuries are used so that injuries of similar type and severity can be studied in a meaningful way. Tile,21 Tile and Pennal,22 Bucholz,3 and Burgess et al5 have proposed classifications based on the stability of the pelvic ring injury. In contrast to these systems, Letourne11 used a classification system based on the anatomic site of the injury. Although Letournel's system specifies sacral fractures as a distinct subset of pelvic ring injuries, it does not define the stability of the injury. In the classic article, Sacral Fractures: An Important Problem, Denis et al6 cited an additional 4 classification systems of sacral fractures and added the Denis classification.
The authors think that the presence of multiple classification systems to define sacral fractures and the difficulty in predicting which sacral fractures are unstable makes it difficult to define the indications for the operative fixation of sacral fractures. However, the combined use of Letournel's system that defines a sacral fracture and the use of the systems that describe pelvic instability provides a rationale for the operative treatment of displaced sacral fractures.
All the patients in this series, except 3, had Tile C or Bucholz III injuries, which are vertically and rotationally unstable. Two of the patients treated in this series had Tile B2 and B3 injuries (lateral compression injuries). The indications for surgery in these 2 patients were decompression of the S1 foramina in a patient with signs of S1 sciatica and an attempt to reduce a 1.4 cm deformity. In both cases attempts to reduce the fractures through a posterior approach were futile, secondary to the inherent stability of the impacted sacral bone. The findings in the 2 patients agrees with Tile's impression that these are inherently stable injuries that do not require fixation for stability.21 Alternative techniques are needed to address the neurologic problems caused by Tile B injuries to the sacrum.
In contrast, Tile C or Bucholz III injuries were found to have mobile fracture sites at the time of surgery. Without internal fixation, these patients would have been treated with several weeks of bed rest to prevent further displacement.
Neurologic injuries were present in 12 of 30 patients (40%) treated in this series. This is higher than the 28.4% incidence of neurologic injury reported by Denis et al6 for transforaminal sacral fractures. However, their study included all fractures that were transforaminal and might have included fractures with lesser magnitudes of displacement. Pohlemann et al17 also observed a 42.9% rate of neurologic injury in Tile C injuries that were transforaminal (Denis Zone II). The authors concur with Pohlemann et al that the rate of neurologic injury is related to the degree of pelvic instability and the specific fracture pattern.
The rationale for the use of interfragmentary compression is predicated on the anatomy of sacral fractures and the mechanical loading of the posterior pelvis. The presence of the sacral canal and the sacral foramina restrict the placement of fracture implants. Densitometry studies indicate that the body of the S1 vertebrae is 60% more dense than the surrounding bone and is the best site for secure screw fixation.20 Because many patients have multiple injuries, successful fixation requires that patients be mobilized to at least a sitting position for pulmonary care after surgery. A basic tenet of fracture fixation is the use of interfragmentary compression to achieve rigid fixation.15
The use of interfragmentary compression requires an accurate reduction of sacral fractures. Denis et al6 found in cadaver studies that the S1 and S2 nerve roots occupy ⅓ of the diameter of their sacral foramina. This implies that compression of malreduced fractures may result in overcompression of the foramina and cause iatrogenic neurologic injuries.6 It seems that the 2 neurologic injuries in this series resulted from maneuvers used to reduce the sacral fractures. In both cases, postoperative CT scans indicated that the foramen had not been overcompressed by use of interfragmentary compression. After reviewing this series the authors increasingly rely on the use of intraoperative neural monitoring.
The ideal technique for reducing sacral fractures would achieve anatomic reduction and stable fixation without complications. The number of percutaneous procedures performed in this series is too small to determine if closed reduction and percutaneous fixation is preferable to open reduction.18 Open reductions were performed for fractures with greater displacement (average, 24 mm) than fractures treated by closed techniques (15 mm). This suggests that there was a selection bias for proceeding with an open reduction when treating fractures with greater displacement based on the anatomic study by Denis et al.6 After reviewing this series, the authors think that transforaminal fractures (Denis Zone II) with displacement of greater than 1 cm should be treated with open reduction to ensure accurate alignment of the sacral foramina.
This review of 30 patients with displaced sacral fractures suggests that open reduction and iliosacral screw fixation of displaced fractures leads to better reduction of the fracture site than closed reduction and percutaneous fixation. The authors think that an accurate reduction is important to achieve the best results. It seems that the presence of a neurologic injury is the single most important predictor of compromised outcome in patients with displaced fractures.
1. Albert MJ, Miller ME, MacNaughton M, Hutton WC: Posterior fixation using a transiliac 4.5 reconstruction plate. J Orthop Trauma 7:226-232, 1993.
2. Bonnin JG: Sacral fractures and injuries to the cauda equina. J Bone Joint Surg 27:113-127, 1945.
3. Bucholz RW: The pathological anatomy of Malgaigne fracture-dislocations of the pelvis. J Bone Joint Surg 63A:400-404, 1981.
4. Bucholz RW, Peters P: Assessment of Pelvic Stability. In Bassett III F (ed). Instructional Course Lecturers. Vol 37. St Louis, CV Mosby 119-127, 1988.
5. Burgess AR, Eastridge BJ, Young JWR, et al: Pelvic ring disruptions: Effective classification system and treatment protocols. J Trauma 30:848-856, 1990.
6. Denis F, Davis S, Comfort T: Sacral fractures: An important problem. Retrospective analysis of 236 cases. Clin Orthop 227:67-81, 1988.
7. Duwelius PJ, VanAllen M, Bray TJ, Nelson D: Computed tomography-guided fixation of unstable posterior pelvic ring disruptions. J Orthop Trauma 6:420-426, 1992.
8. Goodell CL: Neurological deficits associated with pelvic fractures. J Neurosurg 24:837-842, 1966.
9. Huittinen V-M: Lumbosacral nerve injury in fracture of the pelvis. A postmortem radiographic and pathoanatomical study. Acta Chir Scand 429(Suppl):1-43, 1972.
10. Huittinen V-M, Slatis P: Nerve injury in double vertical pelvic fractures. Acta Chir Scand 138:571-575, 1972.
11. Letournel E: Pelvic Fractures. Injury 10:145-148, 1978.
12. Martin JG, Nepola JV, Marsh JL: The treatment of unstable pelvic injuries with external fixation. Orthop Trans 18:1053, 1994.
13. Matta JM, Saucedo T: Internal fixation of pelvic ring fractures. Clin Orthop 242:83-97, 1989.
14. Mears DC, Capito CP, Deleeuw H: Posterior Pelvic Disruptions Managed by the Use of the Double Cobra Plate. In Bassett III F (ed). Instructional Course Lectures. Vol 37. St Louis, CV Mosby 143-150, 1988.
15. Muller ME, Allgower M, Schneider R, Willenegger H: Principles of Surgical Stabilization. In Manual of Internal Fixation. Ed 3. New York, Springer 485-500, 1991.
16. Patterson FP, Morton KS: Neurological complications of fractures and dislocations of the pelvis. J Trauma 12:1013-1023, 1972.
17. Pohlemann T, Gansslen A, Tscherne H: The problem of the sacral fractures. Clinical analysis of 377 cases. Orthopade 21:400-412, 1992.
18. Routt ML, Kregor PJ, Simonian PT, Mayo KA: Early results of percutaneous iliosacral screws placed with the patient in the supine position. J Orthop Trauma 9:28-34, 1995.
19. Semba RT, Yasukawa K, Gustilo RB: Critical analysis of the results of 53 Malgaigne fractures of the pelvis. J Trauma 23:535-537, 1983.
20. Smith SA, Abitol JJ, Carlson GD, et al: The effects of depth penetration, screw orientation and bone density on sacral screw fixation. Spine 18:1006-1010, 1993.
21. Tile M: Pelvic ring fractures: Should they be fixed? J Bone Joint Surg 70B:1-12, 1988.
22. Tile M, Pennal GF: Pelvic disruption: Principles of management. Clin Orthop 151:56-64, 1980.