The goals of this study were to find associations between anterior and posterior ring injuries, provide a descriptive comparison of pelvic ring disruptions as assessed by plain radiography, and compare the value of computed tomography (CT) over plain radiography in evaluating anterior and posterior structures.
A retrospective review of radiographic reports and records identified 142 patients with pubic ramus fractures as observed by plain radiography. A statistical analysis was performed to test the associations between anterior ring injury as assessed by plain radiography and posterior ring injury as assessed by CT. Forty-five point five percent of patients with bilateral ramus fractures and 42.0% of patients with dual-ramus fractures had concomitant sacral fractures not observed on plain radiographs. These occult sacral fractures were found in only 11.1% of patients with inferior ramus fractures.
The type of pubic injury on plain radiographs may be predictive of posterior ring injury, and therefore may help determine injury energy and severity, determine the need for further imaging studies, and help guide clinical management. Although CT is highly sensitive in identifying both anterior and posterior pubic ring injuries, elderly patients with simple fractures of a single pubic ramus are less likely to suffer from pelvic instability and thus may not benefit from CT.
Fracture classification systems are often used to characterize injury severity. The widely used Young and Burgess classification system for pelvic fractures is based on the direction of the injuring force and seeks to grade the magnitude of injury.1,2 This classification system divides pelvic injuries into categories of anterior-posterior compression, lateral compression, vertical shear, and combined patterns. The categories are further divided into subgroups, depending on the degree of bony and ligamentous injury. Lateral compression injuries, which are the most common, are classified into 3 grades: type I, type II, and type III. In particular, the type I lateral compression pelvic fracture is usually considered to be a stable fracture pattern, both in earlier descriptions by Young and Burgess1,2 and by more recent experts.3-6
However, mounting evidence suggests that this category of pelvic fracture is more diverse than originally thought and may include a variety of anatomic injuries. These include injuries from low-energy trauma that do not disturb the posterior ring, as well as high-energy injuries that lead to disruption of the sacrum and other posterior structures.7,8 A recent analysis of patents with pelvic fractures classified as type I lateral compression found injuries ranging from unaffected posterior structures to complete disruptions of the sacrum.7 In that study, more than half of lateral compression fractures were characterized by some degree of anterior sacral comminution, suggesting a higher degree of energy than had been previously expected.7
Although such data demonstrate the breadth and potential variety of posterior ring injuries in patients sustaining type I lateral compression pelvic fractures, the range of anterior ring disruptions has not yet been characterized. Furthermore, it is unknown whether specific anterior ring injuries could be representative of higher-energy trauma that may be associated with more severe disruptions of the posterior ring. To determine this relationship between anterior and posterior pelvic ring injuries, we analyzed plain radiographs. Plain radiography, which includes anteroposterior (AP), inlet, and outlet views, is the initial radiologic modality of choice for evaluating a suspected pelvic fracture. Although it is now recognized that its sensitivity in identifying posterior ring injuries is relatively low when compared with computed tomography (CT), we hypothesized that the findings from plain radiographs, particularly those regarding the status of the anterior pelvic ring, could still be useful in classifying injury severity.
The patients in our study were divided into fracture categories based on characteristics of the anterior ring injury as detected by plain radiography. Then, by analyzing the CT findings of those same patients, we sought to determine whether findings from plain radiographs would correlate with the extent of posterior ring injury and could thus potentially aid in classifying injury severity.
Materials and Methods
Patients from 2 radiology databases covering 3 affiliated institutions were retrospectively identified. We conducted a search of pelvis CT and radiography final reports using the search terms pubic, ramus, and fracture. We compiled a list of patients who met the following inclusion criteria: their primary indication for imaging was trauma; they were aged 18 years or older; their complete final pelvis CT and pelvis AP/inlet/outlet radiography reports were available; their studies were performed within the same 48-hour period; their studies were not postoperative; and at least 1 pubic ramus fracture was found on all studies. Patients who did not meet these criteria, as well as patients with malignant pelvic lesions and patients with old/healing fractures, were excluded from the analysis. A total of 149 patients were excluded from our analysis, resulting in 142 patients who met these criteria. These patients were all found to have pubic ramus fractures on plain radiography.
The final radiography and CT radiology reports of these patients were then compiled, and the following data were examined retrospectively: location (superior or inferior ramus); laterality (unilateral or bilateral); displacement (nondisplaced, minimally displaced, or displaced) of pubic ramus fractures; presence of pubic and sacroiliac joint diastasis; presence of iliac fractures; presence of sacral fractures; and presence of hip/femur fractures. Sacral fractures were further classified according to the system described by Denis et al9 (type I, fracture lateral to the sacral neuroforamina; type II, fracture enters the neuroforamina; type III, fracture enters the sacral body medial to the neuroforamina). Patient sex and age were also recorded.
Comparisons between continuous variables were performed using the Student t test, with P<.05 considered statistically significant. For categorical variables, Fishers exact test was used.
Between January 2000 and May 2009, 142 patients with pubic rami fractures resulting from trauma were treated by our department and were included in our study. Of them, 44 (31.0%) were men and 98 (69.0%) were women (Table). Mean patient age was 69.27±20.6 years. Mean age of the men (59.94±23.7 years) was significantly lower than that of women (73.93±17.2 years) (P<.005).
Our goal was to characterize the injuries of the patients collected in the study and compare injuries sustained in the anterior pelvic ring with those identified in the posterior ring. First, we placed the patients into 1 of 5 distinct categories based on radiographic characteristics of anterior ring injurybilateral, nondisplaced, superior ramus, inferior ramus, or dual-rami (Figure 1). These categories would serve as the basis for our analysis. More patients had displaced fractures (88 patients; 62.0%) than nondisplaced fractures (32 patients; 22.5%). Of the patients with unilateral displaced fractures (Figure 1), dual-rami (superior and inferior) fractures were the most common (50 patients; 35.2%).
Figure 1: Association between anterior and posterior pelvic ring injuries. Pubic fractures were placed into 1 of 5 categories (bilateral, nondisplaced, superior ramus, inferior ramus, and dual ramus) as determined by plain radiographs. Posterior pelvic ring injuries were then identified using CT radiology reports (center column). Also identified was the number of occult posterior ring injuries that were identified by CT but missed by plain radiographs. The number of occult injuries is given before the number of total posterior ring injuries (center column). The percent of patients with sacral fractures (occult and total) was also determined (right column).
Concerning single-rami fractures, there were a higher number of patients in the superior ramus fracture category than in the inferior ramus category (20 patients [14.1%] and 18 patients [12.7%], respectively). Comminuted pubic rami fractures were found in 48 of 142 patients (33.8%).
We sought to determine whether certain categories of anterior ring injury were associated with a higher incidence or severity of posterior ring injuries. To do this, we examined different categories of anterior ring injuries and compared them to posterior ring injuries (Figure 1). We found that patients with bilateral anterior ring fractures were likely to have concomitant fractures of the sacrum (16 patients; 72.7%). In contrast, patients with inferior ramus fractures were significantly less likely to have a sacral fracture (4 patients; 22.2%; P<.03 compared to all fractures). Patients with nondisplaced anterior ring fractures were also less likely to have a sacral fracture, although results did not reach statistical significance (11 patients; 34.4%; P=.1 compared to all fractures).
We questioned whether these associations between anterior and posterior ring injuries could be due to differences in the sensitivity of plain radiography vs computed tomography. To determine this, we assessed for injuries identified on CT that had gone undetected by plain radiographs (Figure 2). Of all the fractures that were detected by CT and not by radiograph, sacral fractures were the most common: a total of 69 sacral fractures identified on CT had been missed by plain radiographs. The vast majority (58) of these fractures were Denis type I; 5 sacral fractures were Denis type II, and 1 fracture was Denis type III. Five fractures were unspecified. Of note, patients with Denis type II injuries (mean age 45±26 years) were significantly younger than the other patients in our study. Also of note, 80% of Denis type II patients belonged to the bilateral or dual-rami fracture category. Furthermore, of the 5 Denis type II fractures, 2 patients had sustained fractures of the fifth lumbar transverse process and 1 patient had sustained an avulsion fracture of the ischial spine. Interestingly, 80% of Denis type II patients had acquired pubic ramus fractures with some degree of comminution, which is consistent with Denis type II injuries being of higher energy than Denis type I. This would be expected given that the energy necessary to generate a sacral fracture could also result in increased comminution.
| || |
Figure 2: Demonstration of an occult sacral fracture. In the AP pelvis radiograph (A), a left superior pubic ramus fracture can be observed (red ellipse). The sacrum cannot be adequately evaluated due to overlying bowel gas. In the CT image (B), an occult sacral fracture was identified (red ellipse).
Within each fracture category, we then calculated the percent of patients with occult sacral fractures, which were observed on CT but not on plain radiographs (Figure 2). An occult sacral fracture was observed on the CT scans of only 11.1% of patients in the inferior ramus fracture category (Figure 1); this category of patients was significantly less likely to have an occult sacral fracture on CT (P<.04, compared to all fractures). In contrast, 45.4% of patients in the bilateral fracture category and 42.0% in the dual-rami category were found to have occult sacral fractures on CT (Figure 1). Patients with nondisplaced pubic ramus fractures were the least likely to have occult posterior pelvic ring injuries. Fourteen of 32 (43.8%) patients in this category were found to have an isolated pubic ramus fracture on CT. This was not the case for the bilateral fracture category, in which a significantly higher percentage (18 of 22; 81.8%; P<.05 compared to nondisplaced fractures) of occult pelvic ring injuries were found, including iliac fractures, as well as widening of the sacroiliac joint (Figure 1). However, no single category of fractures was significantly more likely to have an isolated pubic ramus fracture when compared to all fractures. Interestingly, patients with inferior ramus fractures also seemed to have sacral fractures that were less severe than in patients with other fractures types. In patients with inferior ramus fractures, 50% (2 of 4) of the concomitant sacral fractures were nondisplaced. In contrast, only 12.5% (2 of 16) and 9.1% (1 of 11) of the sacral fractures were nondisplaced in patients with bilateral ramus and superior ramus fractures, respectively.
To compare the effectiveness of CT and plain radiography in evaluating pubic ramus fractures, we examined each fracture category and determined the percent correlation (Figure 3), which was defined as the percent of patients within a fracture category whose pelvis CT results were in agreement with pelvis radiographic results. Results from pelvis CT show that plain radiography correlates highly with CT in the case of inferior ramus fractures (83.3% correlation), bilateral rami fractures (77.3% correlation), and dual-rami fractures (84.0% correlation) (Figure 3). However, the percent correlation was significantly lower for superior ramus fractures (35.0% correlation; P<.004 compared to all fractures).
Figure 3: Pubic fractures evaluated by plain radiographs compared with CT. Pubic fractures were placed into 1 of 5 categories (bilateral, nondisplaced, superior ramus, inferior ramus, and dual ramus) as determined by plain radiographs (left column). The findings of CT radiology reports for those same patients were determined (center column). Percent correlation was calculated as the percent of patients within a category whose pelvis CT results were in agreement with pelvis radiographic results (right column).
Like all other pelvic fracture classification systems, the one pioneered by Young and Burgess is based on 3 plain radiographs of the pelvic ring.1 In early descriptions, Young et al1 reported that the vast majority of diagnoses can be made correctly by using radiographs. Indeed, plain radiographs, including AP, inlet, and outlet views, are usually the initial radiological study of choice for evaluating a suspected pelvic fracture. However, there is mounting evidence to suggest that CT significantly enhances both the sensitivity and specificity of detecting pelvic ring injuries (especially along the posterior ring structures) as compared to plain radiography alone.10-16 Although the system of Young and Burgess has been useful in understanding the mechanism of injury, fracture pattern, and ligamentous stability of pelvic fractures, it does not seem to adequately address posterior pelvic ring injuries, such as fractures of the sacrum, sacroiliac joint instability, and fractures of the ilium.
Because the classification system of Young and Burgess does not take into account the associations between anterior and posterior pelvic injury, it often fails to capture the full spectrum of pelvic injuries.7,8 Type I lateral compression pelvic fractures are typically characterized by low-energy fracture patterns, such as a stable pubic ramus fracture; however, some low-energy lateral compression fractures can also include associated injuries that are suggestive of higher-energy mechanistic subtypes that were not previously considered in the original classification. Indeed, it has been suggested that characteristics of sacral disruption, such as anterior sacral comminution and Denis type, may be useful for differentiating between different levels of injury severity.7
We sought to better characterize the extent of posterior pelvic ring disruption by identifying associations between anterior and posterior ring injuries. To our knowledge, no other study to date has successfully established these associations. Because anterior ring injuries are more easily evaluated during the initial radiographic screen following acute trauma, establishing associations between anterior and posterior ring injuries could allow the trauma surgeon to develop a better initial characterization of the patients pelvic injury.
Our findings agree with the general notion that the Young and Burgess type I lateral compression category represents a wide spectrum of injuries, including lower-energy and higher-energy trauma. This suggests that associations between anterior and posterior ring injuries may indeed be important for injury classification. Specifically, we found that patients with nondisplaced fractures were most likely to have isolated pubic ramus fractures without associated fractures or disruptions of the posterior ring. In addition, patients with inferior ramus fractures were less likely to have a sacral fracture, whether occult or nonoccult. Also, in patients with an inferior ramus fracture and a concomitant sacral fracture, the sacral injury was more likely to be nondisplaced. This suggests that patients with nondisplaced or inferior ramus fractures may be more likely to suffer from lower-energy trauma. In contrast, patients with a superior ramus, dual-rami, or bilateral rami fracture, as identified by plain radiography, were more likely to have a sacral fracture, which can be identified on CT scan.
The vast majority of sacral fractures in our study were Denis type I. Denis type III fractures, which are typically seen in higher-energy trauma, were not observed, suggesting that the injuries found in our series were characterized by lower-energy trauma. This is also supported by the observation that a majority (66.2%) of patients did not have comminuted fractures. Interestingly, all Denis type II patients were significantly below the mean age of our patient population and had comminuted fractures or dual-rami fractures. These features are all suggestive of injuries sustained via a higher-energy mechanism than Denis type I. This suggests that a subset of younger patients with comminuted or dual ramus fractures on plain radiographs are more likely to suffer from posterior ring injuries, while elderly patients whose plain radiographs show isolated noncomminuted single-ramus fractures are less likely to have an unstable pelvis with posterior ring involvement.
In addition to providing a better characterization of pelvic injuries for classification purposes, the associations established between anterior and posterior ring injuries could also aid in clinical diagnostic management. This is particularly true in the case of occult sacral fractures, which may be missed by initial radiographic examinations only to be found on subsequent CT evaluation (Figure 2). Previous studies have found that CT is up to 98% sensitive in identifying posterior pelvic ring disruptions.10-16 We found that CT was far more sensitive than plain radiography in identifying fractures of the posterior pelvic ring (Figure 1). Thus, we are left with the task of identifying those patients in which an occult sacral fracture is most likely, thus necessitating evaluation by CT. Our data established a clear link between the type of anterior ring injuries and the incidence of occult sacral fractures (Figure 1). When combined with observations regarding Denis type and patient age, our findings suggest that some patients may benefit significantly from CT evaluation, while for others a CT scan may not significantly alter clinical management. Our findings suggest that in the case of an elderly patient with an isolated noncomminuted inferior pubic ramus fracture, a CT scan may be of lower diagnostic use.
Although our study focused on establishing associations between anterior and posterior ring injuries, it would be interesting to further investigate the relationship between these associations and clinical outcomes. Sembler et al17 found that nonsurgical treatment with immediate weight bearing is sufficient for patients with unilateral lateral compression sacral fractures of <10 mm displacement. Our study further confirms the notion that at least in a subset of type I lateral compression patients, CT scans may not be warranted, and nonoperative treatment followed by observation for healing may be sufficient. However, CT is also a tool that could aid in the discernment of sacral displacement, thus helping the clinician to determine the appropriate weight bearing status.
This study has several limitations. The patients in our sample were examined retrospectively, and our sample size was limited. Additionally, details regarding the specific mechanism of injury and clinical outcomes were not available for our sample; this information would have provided a broader data set, which could have helped to cement our conclusions. Furthermore, this study was based on official radiology reports, which are generated by radiologists of various backgrounds, and are thus prone to interobserver variability. It is conceivable that a plain radiograph interpreted as an isolated pubic ramus fracture (type I lateral compression injury) by 1 observer may also include a posterior iliac fracture as seen by another observer, thus altering the injury classification to type II lateral compression. Finally, although the patients in our sample were demographically diverse (Table), the findings would be more generalizable if data from a multicenter analysis had been available.
Our findings agree with the idea that the Young and Burgess type I lateral compression category of pelvic fractures represents a wide spectrum of injury. In addition, the type of pubic injury on plain radiographs may be predictive of posterior ring injury, and therefore may help determine injury energy and severity and the need for further imaging studies, and help guide the clinical management of the fracture. This may highlight the importance of determining associations between anterior and posterior ring injuries as we seek to refine the classification of pelvic ring fractures. Our data also confirmed that CT is highly sensitive in identifying both anterior and posterior pubic ring injuries, although its advantage over plain radiography varies by anatomic region. However, elderly patients with a simple fracture of a single pubic ramus are less likely to suffer from pelvic instability and may not benefit from CT. There is evidence to suggest that for these patients, nonoperative treatment with immediate weight bearing and close follow-up may be sufficient.17 In analyzing patients within the context of radiographic fracture categories, our study mimics the clinical situation, where a patient is given an initial evaluation by plain pelvis radiographs prior to or in lieu of a CT scan.
We discovered clear associations between anterior pubic fractures and disruptions of the posterior ring. We found that patients with bilateral and dual ramus fractures were most likely to have concomitant fractures of the sacrum. These patients were also most likely to have an occult sacral fracture that was detected by CT but not by plain radiographs. Meanwhile, patients with inferior ramus fractures were least likely to suffer from sacral fractures, whether occult or nonoccult.
Interestingly, while CT is generally superior to plain radiographs in terms of sensitivity when examining type I lateral compression pelvic fractures, we found that plain radiography correlates with CT in identifying inferior ramus, bilateral rami, and dual rami fractures. This was not seen with superior ramus fractures, which were poorly correlated. Of fractures that are detected by CT but missed by radiographs, sacral fractures were the most common, with the majority being Denis type I. Denis type II injuries often included features such as comminution, which may be associated with higher energy than Denis type I. Based on our findings, we recommend a CT scan for certain fracture patterns, especially in younger and active patients.
- Young JW, Burgess AR, Brumback RJ, Poka A. Pelvic fractures: value of plain radiography in early assessment and management. Radiology. 1986; 160(2):445-451.
- Young JW, Burgess AR, Brumback RJ, Poka A. Lateral compression fractures of the pelvis: the importance of plain radiographs in the diagnosis and surgical management. Skeletal Radiol. 1986; 15(2):103-109.
- Burgess AR, Eastridge BJ, Young JW, et al. Pelvic ring disruptions: effective classification system and treatment protocols. J Trauma. 1990; 30(7):848-856.
- Routt ML Jr, Simonian PT, Ballmer F. A rational approach to pelvic trauma. Resuscitation and early definitive stabilization. Clin Orthop Relat Res. 1995; (318):61-74.
- Miranda MA, Riemer BL, Butterfield SL, Burke CJ III. Pelvic ring injuries. A long term functional outcome study. Clin Orthop Relat Res. 1996; (329):152-159.
- Bellabarba C, Ricci WM, Bolhofner BR. Distraction external fixation in lateral compression pelvic fractures. J Orthop Trauma. 2006; 20(1 Suppl):S7-14.
- Lefaivre KA, Padalecki JR, Starr AJ. What constitutes a Young and Burgess lateral compression-I (OTA 61-B2) pelvic ring disruption? A description of computed tomography-based fracture anatomy and associated injuries. J Orthop Trauma. 2009; 23(1):16-21.
- Khoury A, Kreder H, Skrinskas T, Hardisty M, Tile M, Whyne CM. Lateral compression fracture of the pelvis represents a heterogeneous group of complex 3D patterns of displacement. Injury. 2008; 39(8):893-902.
- Denis F, Davis S, Comfort T. Sacral fractures: an important problem. Retrospective analysis of 236 cases. Clin Orthop Relat Res. 1988; (227):67-81.
- Rommens PM, Vanderschot PM, Broos PL. Conventional radiography and CT examination of pelvic ring fractures. A comparative study of 90 patients. Unfallchirurg. 1992; 95(8):387-392.
- Albrechtsen J, Hede J, Jurik AG. Pelvic fractures. Assessment by conventional radiography and CT. Acta Radiol. 1994; 35(5):420-425.
- Schapira D, Militeanu D, Israel O, Scharf Y. Insufficiency fractures of the pubic ramus. Semin Arthritis Rheum. 1996; 25(6):373-382.
- Their ME, Bensch FV, Koskinen SK, Handolin L, Kiuru MJ. Diagnostic value of pelvic radiography in the initial trauma series in blunt trauma. Eur Radiol. 2005; 15(8):1533-1537.
- Berg EE, Chebuhar C, Bell RM. Pelvic trauma imaging: a blinded comparison of computed tomography and roentgenograms. J Trauma. 1996; 41(6):994-998.
- Guillamondegui OD, Pryor JP, Gracias VH, Gupta R, Reilly PM, Schwab CW. Pelvic radiography in blunt trauma resuscitation: a diminishing role. J Trauma. 2002; 53(6):1043-1047.
- Cabarrus MC, Ambekar A, Lu Y, Link TM. MRI and CT of insufficiency fractures of the pelvis and the proximal femur. AJR Am J Roentgenol. 2008; 191(4):995-1001.
- Sembler G, Lien J, Tornetta P III. Nonoperative immediate weightbearing of minimally displaced lateral compression sacral fractures does not result in displacement. Paper presented at: Orthopaedic Trauma Association 25th Annual Meeting; October 7-10, 2009; San Diego, California.
Mr Lin and Drs Min, Christoforou, and Tejwani are from the Department of Orthopedic Surgery, NYU Hospital for Joint Diseases, New York University School of Medicine, New York, New York.
Mr Lin and Drs Min, Christoforou, and Tejwani have no relevant financial relationships to disclose.
Correspondence should be addressed to: Nirmal C. Tejwani, MD, Department of Orthopedic Surgery, NYU Hospital for Joint Diseases, 550 First Ave, NBV21W37, New York, NY 10016 (firstname.lastname@example.org)