The goal of pelvic fracture surgery is to restore the integrity and stability of the pelvic ring and allow patients to return to their preinjury functional level.1 Plain radiographs are important for identifying the overall injury pattern, fracture classification, and displacement. However, cross-sectional imaging in the form of computed tomography (CT) is essential to delineate posterior displacement and sacral morphology.2 The use of postoperative CT imaging is varied, with some surgeons using it routinely to assess reduction and metalwork position, others selectively in cases where there is doubt but in the majority it is not a standard practice.3 It is argued that routine use would incur an increase in cost and an increased exposure to radiation, particularly in a group of patients who may have had multiple preoperative scans including a preoperative CT, and several fluoroscopic images obtained during their procedures.
In our institute, we have been performing routine postoperative CT scans in all surgically treated pelvic fractures. The rationale behind this approach was to enable our team to accurately assess the quality of reduction, metalwork malposition, and for audit, teaching and research.
The aim of this study was to assess the value of routine use of CT scans after pelvic fracture surgery and to determine the sensitivity of conventional plain radiographs and intraoperative fluoroscopy in detecting the metalwork malposition.
PATIENTS AND METHODS
Standardized protocols in our trauma network dictate that all patients with polytrauma receive a trauma CT scan (CT scan from the head to the level of the hip joint) on arrival. If a pelvic fracture is identified, a standard pelvis anteroposterior (AP) radiograph is performed with the pelvic binder removed (unless there is uncontrolled hemorrhage). During the surgery, fluoroscopic screening is performed and appropriate images are saved at key stages. Computer-assisted or navigation systems were not used for sacroiliac (SI) screw insertion. After the surgery, all patients received a standard AP pelvic radiograph and a CT scan.
We retrospectively reviewed the radiological images of all patients with pelvic fractures that underwent surgical fixation during the 6-year period from January 2010 till December 2015. Patients who did not have a postoperative CT scan were excluded from the study. All the preoperative, peroperative, and postoperative radiological images were assessed independently by a pelvic fellow or specialist registrar trainee initially, then separately by one of the other pelvic trauma consultant surgeons not involved in the initial surgical care, to simulate a clinical setting. The pelvic ring injuries were classified according to Young and Burgess4 All assessors were blinded to the outcome of the postoperative scan imaging. Finally, the results of the postoperative CT scans were assessed by the same team with regard to the possible implant malpositioning.
Patients were divided into 2 groups: group A—patients whose fixation entailed the use of SI screws and group B—patients whose fixation did not entail the use of SI screws.
Patients were then subdivided, on the basis of the perioperative fluoroscopic images and postoperative plain radiographs, into 1 of 3 categories:
- Safe: When there is no suspicion of the metalwork malposition in all plain images.
- Suspicious: When there is some suspicion of implant malposition that cannot be confirmed or excluded on the peroperative fluoroscopic images and postoperative plain radiographs.
- Definite malposition: When the malpositioned metalwork was detected in the peroperative fluoroscopic images or plain radiographs.
Malposition was defined as penetration of the screw into the sacral foramina, sacral canal, or protrusion beyond bone into the soft tissue for more than 5 mm.
Data were coded and entered using SPSS version 24 software (Statistical Package for the Social Sciences). Data were summarized using mean, SD, median, minimum, and maximum in quantitative data and using frequency (count) and relative frequency (percentage) for categorical data. Standard diagnostic indices, including sensitivity and specificity, were calculated as described by Galen5. Comparisons between quantitative variables were performed using the nonparametric Mann–Whitney Test. For comparing categorical data, χ2 test was performed, and the Fisher exact test was used instead when the expected frequency was less than 5. A P value less than 0.05 was considered as statistically significant.
During the study period, 223 patients with pelvic fractures underwent surgical fixation in our institute. Twenty-five patients were excluded because there was no postoperative CT scan after the procedure, and the remaining 198 patients were included in this study.
The study included 136 men and 62 women, of mean age 44 years (range: 14–89), at the time of injury. Fracture classifications are shown in Table 1, with the majority being either APC2 or vertical shear. The mechanism of injury is shown in Table 2.
After excluding 2 cases of insufficiency fractures with late presentation, the mean time from injury to surgery was 4 days, (range 0–34), where 14 patients were operated upon on admission (emergency pelvic fixation).
Group A included 161 patients, and group B included 37 patients. In group A, the mean number of screws used was 1.7 (median 1, range 1–6) per patient in whom 148 (92%) were deemed safe, 10 were suspicious (6%), and 3 (2%) showed definite malposition. When the CT scans were assessed, 78% of the fractures which were believed to be safe on plain radiographs were confirmed to be safe on CT scans, whereas 22% (33 patients) showed a malpositioned metalwork. The malpositioned metalwork was related to the SI screws in all but 1 patient who had long anterior pubic screws and developed dyspareunia later. The others were either screws protruding into the sacral foramina (8 cases) (Fig. 1), screws protruding into the sacral canal in 11 cases, or SI screws protruding into the true pelvis in 13 cases (Figs. 2, 3). Only 2 of the patients were immediately symptomatic from the malpositioned metalwork (Table 3). However, 7 (4%) of the patients whose plain radiographs were deemed safe eventually required a revision surgery because of the metalwork malposition. Of the 10 patients who had inconclusive plain radiographs, 9 were found to have metalwork malposition on CT, whereas 1 was safe. Three of these patients eventually required a revision surgery.
When the fluoroscopic images and radiographs of group B were assessed, all of them were deemed safe (no suspicion of metalwork malposition) in relation to the plate and/or screw implants used. When the CT images of that group were assessed, 2 cases were identified with anterior screws protruding more than 5 mm from the pubic bone into the soft tissues, none of which required a revision surgery.
Overall in this series, 10 patients (6%) required a revision surgery because of metalwork malposition. Seven cases (4%) had metalwork malposition that was not picked up on the standard routine perioperative plain radiographs and was only picked up on CT films.
If we assume that CT scans are 100% sensitive and specific in detecting metalwork malposition and if both the suspicious and definitive malposition groups are considered as positive indicators of SI screw malposition on radiographs and/or fluoroscopy, then the use of perioperative fluoroscopic images and postoperative AP radiographs shows a sensitivity of 27% in detecting metalwork malposition and a specificity of 99%. In the 13 patients defined with suspicious (10 patients) or definite malpositioning (3 patients) on the plain imaging, only 3 required a revision surgery after further assessment with CT scan.
Increasing the number of SI screws used in fixation significantly increased the risk of metalwork malposition (P = 0.006), and the risk or reoperation due to malposition (P = 0.002). However, emergency SI screws conducted on day zero were not significantly correlated with a higher risk of screw malposition or need for a revision surgery (P = 0.113 and 0.189 respectively), although the numbers were small (14 patients).
There is no current unified postfixation imaging protocol for pelvic fractures. Our unit's standardized approach evolved over 20 years, and this study aimed to review the protocol in view of the increasing concerns of radiation exposure. Previous literature has shown that the use of SI screws led to the highest incidence of malposition, requiring a potential revision surgery.6
In this series, when an SI screw was not used, the postoperative CT scan did not add any value. We were only able to detect 2 patients (out of a cohort of 37 patients) with anterior screws protruding into the soft tissues, which were of no clinical significance and none of the patients required any further intervention. In the other group involving posterior SI screw fixation, 148 patients were believed to have safe metalwork placement on perioperative fluoroscopic and radiographic assessment. However, the CT showed that 33 patients (22%) of that group had a malpositioned metalwork, of which 7 (4%) required a revision surgery. The sensitivity of plain radiographs in detecting metalwork malposition was much lower than we had expected, showing only a sensitivity on 27% with a specificity of 99%. The high specificity demonstrated that if any malposition of metalwork is suspected on fluoroscopy or radiographs, there is a high chance of corresponding findings on the CT scan.
The rate of malpositioning of SI screws has been reported to be high in the literature. Tonetti et al7 reported extraosseous trajectories in 19.8% on their series involving 120 patients, in which all cases were assessed by postoperative CT scans. Pishnamaz et al8 reported a malposition of 12.4% in a series of 137 screws in 102 patients, also assessed using postoperative CT scans. They also found a significantly higher rate of malposition when 2 unilateral screws were used versus 1 unilateral screw. On the other hand, van den Bosch et al reported approximately 7% malpositioning in their series of 88 patients. However, they conducted postoperative CT scans in only 53 of the 88 patients.9 Grossterlinden et al reported a malposition rate of 15% using fluoroscopic guidance, and only 3% when navigation was used. Malposition was influenced by surgeon experience, degree of dislocation of the SI joint, and the number of screws inserted.10 In our study, the SI screws were all inserted using fluoroscopy screening and none with computer-assisted or CT-guided insertion techniques that have been shown in a meta-analysis6 to reduce the risk of implant malpositioning and potentially alter the need for postoperative screening.
The revision rate after SI screw malpositioning shows variation from one study to another. Approximately, one-third of patients with a malpositioned metalwork in the current series required a revision surgery, with an overall 7% rate of revision surgery due to malposition in the entire cohort. Zwingmann et al11 reported a rather high (68%) malposition rate using fluoroscopic-guided SI screw insertion with a 19% rate of reoperation in a cohort of 131 screws in 81 patients They graded the malpositions into 4 grades (0–3) according to the amount of cortical perforation. Most of their malpositions were grade 1 or 2. None of the patients had postoperative neurological deficits.11 Van den Bosch reported a revision rate of 8% (7 revisions in a cohort of 88 patients). All their patients had a neurological deficit from the malpositioned metalwork, which resolved after the revision surgery.9
A revision surgery after metalwork malposition can be a difficult decision, considerations being patient symptoms, the fracture characteristics, the significance or site of metalwork malposition, and adequacy of fracture fixation.
From this study, it is apparent that routine postoperative CT scans for patients without SI screws have added no clinical value unless there are unexplained symptoms. In the group fixed with SI screws, the sensitivity of perioperative fluoroscopy and postoperative plain x-rays in detecting implant malposition was 27%. Accordingly, we plan to continue routine postoperative CT scanning in this group of patients. To minimize radiation exposure to the patients, a protocol is now in place with the radiology department to selectively scan the sacral area only (from L5 to the inferior part of the SI joint), as radiation exposure is directly proportional to the area scanned. This reduces the radiation dose by approximately 50% compared with scanning the whole pelvis.
To the best of our knowledge, this is the first study to look at the clinical value of routine postoperative CT scans after pelvic fracture fixation. The study reports data from 198 patients over a 6-year period. All peri- and post-operative images were reassessed by 2 independent assessors, and we did not rely solely on the immediate postoperative documentation to detect malposition. The weakness of the study is the retrospective design. As such, symptoms related to malposition are only recorded if documented in the medical records. Other factors that could lead to malpositioning, such as poor intraoperative views (for example due to patient size or excessive bowel gas), and sacral morphology, were not assessed.
The sensitivity of perioperative fluoroscopy and postoperative plain radiographs in detecting metalwork malposition in patients whose pelvic fixation was performed with fluoroscopy-guided insertion of SI screws was 27%. We recommend the use of routine targeted postoperative CT scans in patients fixed with SI screws. In this series, routine scanning of patients who did not have SI screws added no significant clinical value.
1. Tile M, Helfet D, Kellam J. Fractures of the Pelvis and Acetabulum. 3rd ed Philadelphia, PA: Lippincott Williams & Wilkins; 2003.
2. Jenkinson R, Tile M, Rubenstein J. Defining the injury: assessment and principles of management of pelvic ring fractures. In: Tile M, Helfet D, Kellam J, et al, eds. Fractures of the Pelvis and Acetabulum. Davos, Switzerland: AO Foundation; 2015:83–98.
3. Archdeacon MT, Dailey SK. Efficacy of routine postoperative CT scan after open reduction and internal fixation of the acetabulum. J Orthop Trauma. 2015;29:354–358.
4. Burgess AR, Eastridge BJ, Young JW, et al. Pelvic ring disruptions: effective classification system and treatment protocols. J Trauma. 1990;30:848–856.
5. Galen RS. Predictive value and efficiency of laboratory testing. Pediatr Clin North Am. 1980;27:861–869.
6. Zwingmann J, Hauschild O, Bode G, et al. Malposition and revision rates of different imaging modalities for percutaneous iliosacral screw fixation following pelvic fractures: a systematic review and meta-analysis. Arch Orthop Trauma Surg. 2013;133:1257–1265.
7. Tonetti J, van Overschelde J, Sadok B, et al. Percutaneous ilio-sacral screw insertion. Fluoroscopic techniques. Orthop Traumatol Surg Res. 2013;99:965–972.
8. Pishnamaz M, Dienstknecht T, Hoppe B, et al. Assessment of pelvic injuries treated with ilio-sacral screws: injury severity and accuracy of screw positioning. Int Orthop. 2016;40:1495–1501.
9. van den Bosch EW, Van Zwienen CM, van Vugt AB. Fluoroscopic positioning of sacroiliac screws in 88 patients. J Trauma. 2002;53:44–48.
10. Grossterlinden L, Rueger J, Catala-Lehnen P, et al. Factors influencing the accuracy of iliosacral screw placement in trauma patients. Int Orthop. 2011;35:1391–1396.
11. Zwingmann J, Konrad G, Mehlhorn AT, et al. Percutaneous iliosacral screw insertion: malpositioning and revision rate of screws with regards to application technique (navigated vs. Conventional). J Trauma. 2010;69:1501–1506.