Article, Radiology

Indications and performance of pelvic radiography in patients with blunt trauma

Unlabelled imageAmerican Journal of Emergency Medicine (2012) 30, 1129-1133

Original Contribution

Indications and performance of Pelvic radiography in patients with blunt trauma?

James F. Holmes MD, MPHa,?, David H. Wisner MDb

aDepartment of Emergency Medicine, UC Davis School of Medicine, Sacramento, CA 95817-2282, USA

bDepartment of Surgery, UC Davis School of Medicine, Sacramento, CA 95817-2282, USA

Received 22 December 2010; revised 29 July 2011; accepted 2 August 2011

Abstract

Objectives: The objectives of this study are to validate a set of clinical variables to identify patients with pelvic fractures and to determine the sensitivity of anteroposterior (AP) pelvic radiographs in patients with Pelvic fractures.

Methods: We conducted a prospective observational cohort study of adults (N18 years) with blunt torso trauma evaluated with abdominal/pelvic computed tomography. Physicians providing care in the emergency department documented history and physical examination findings after initial evaluation. High-risk variables included any of the following: hypotension (systolic blood pressure b90 mm Hg), Glasgow Coma Scale score less than 14, pelvic bone tenderness, or instability. Pelvic fractures were present if the orthopedic faculty documented a fracture to the pubis, ilium, ischium, or sacrum.

Results: We enrolled 4737 patients, including 289 (6.1%; 95% confidence interval [CI], 5.4%-6.8%) with pelvic fractures. Of the 289 patients, 256 (88.6%; 95% CI, 84.3%-92.0%) had at least one of the high-risk variables identified. Initial plain AP radiographs identified 234 (81.0%; 95% CI, 76.0%- 85.3%) of 289 patients with pelvic fractures. The high-risk variables identified all 87 patients (100%; 95% CI, 96.6%-100%) undergoing surgery, whereas plain AP pelvic radiography identified a fracture in 83 patients (95.4%; 95% CI, 88.6%-98.7%) undergoing surgery.

Conclusion: Previously identified high-risk variables for pelvic fracture identify most but not all patients with pelvic fractures. However, these high-risk variables identify all patients undergoing surgery and may be implemented as screening criteria for pelvic radiography. Anteroposterior pelvic radiographs fail to demonstrate a fracture in a substantial number of patients with pelvic fracture including patients who undergo surgery.

(C) 2012

Introduction

Pelvic fractures are a frequent and potentially serious injury after blunt trauma. Early identification of theses injuries is imperative because Life-threatening hemorrhage

? Prior presentation: Society for Academic Emergency Medicine Annual Meeting, New Orleans, LA, May 2009.

* Corresponding author. Tel.: +1 916 734 1533; fax: +1 916 734 7950.

E-mail address: [email protected] (J.F. Holmes).

may occur and both temporary and operative stabilization may be needed. As such, advanced trauma life support (ATLS) recommends that injured patients undergo plain anteroposterior (AP) pelvic radiography immediately after the primary survey [1]. Routine pelvic radiography of all patients with blunt trauma, however, is likely inefficient and not indicated. Multiple studies attempt to identify indications for pelvic radiographic evaluation of injured patients [2-5]. These studies are limited either by their retrospective nature and/or small sample and use of different clinical criteria for

0735-6757/$ – see front matter (C) 2012 doi:10.1016/j.ajem.2011.08.001

pelvic radiography. A meta-analysis of these studies in- dicates that selective pelvic radiography based on high-risk criteria is feasible [6].

Furthermore, the appropriate radiographic screening of patients with blunt trauma at risk for possible pelvic fracture is unclear. Although initial plain AP pelvic radiography identifies most pelvic fractures, pelvic com- puted tomographic (CT) scanning is believed superior. In patients with planned abdominal/pelvic CT scanning as part of their emergency department (ED) evaluation, it is unclear if initial screening with plain pelvic radiography is beneficial.

The objective of this study is to validate previous iden- tified clinical variables (hypotension, Glasgow Coma Scale [GCS] score b14, and pelvic bone tenderness/instability) for pelvic radiography in patients with blunt torso trauma and to describe the test performance of plain AP pelvic radio- graphy. We hypothesize that the clinical decision rule identifies all patients with clinically important pelvic frac- tures. In addition, we hypothesize that plain AP pelvic radiography identifies all patients requiring definitive treatment for their pelvic fracture but that the initial AP pelvic radiograph adds little information in a patient also undergoing pelvic CT scanning.

Methods

We conducted a prospective observational cohort study at an urban level 1 trauma center. The study was approved by the study site’s institutional review board.

Subject eligibility

We enrolled adult patients, 18 years or older, with blunt torso trauma who underwent abdominal/pelvic CT imaging for the diagnosis of abdominal/pelvic injuries. We excluded all patients with penetrating trauma, pregnant patients, those presenting in cardiopulmonary arrest, and those patients with blunt torso trauma who did not undergo an abdominal/pelvic CT scan.

Data collection

Physical examination findings were recorded on a standardized data collection form by the treating physician (resident or faculty physician) before knowledge of the CT scan results. These historical and physical examination findings included the following: mechanism of injury, initial ED systolic blood pressure, GCS score, distracting painful injury, and pelvic bone examination findings (including tenderness, abrasions, and instability). Pelvic bone tender- ness was considered present for tenderness on palpation of Iliac crests, pubic symphysis, or sacroiliac joint. Pelvic bone instability was considered present for movement of the pelvic bones with lateral compression. Hypotenstion was

considered present for an initial ED systolic blood pressure less than 90 mm Hg.

The results of the initial plain AP pelvic radiographs as dictated by the faculty radiologist were also collected. Plain AP pelvis radiographs were considered abnormal if a pelvic fracture was definitively identified on the radiograph. Two physicians independently evaluated a convenience sample of 702 enrolled patients to assess Interobserver agreement of the physical examination findings of the pelvis.

Outcome measures

The outcome measures were (1) the presence of pelvic fracture (disease oriented outcome) and (2) pelvic fracture undergoing Surgical repair (patient-oriented outcome). Sub- jects were considered to have a pelvic fracture based upon the final clinical outcome by the treating orthopedic physi- cians. Pelvic fractures were not considered present unless confirmed by a treating, faculty Orthopedic surgeon.

Validation of previous indications for pelvic radiography

Patients were considered at risk for pelvic fracture for any of the following variables: hypotension (systolic blood pressure, b90 mm Hg), GCS score less than 14, pelvic bone tenderness, or instability. These high-risk variables were determined based upon numerous prior studies [2-7]. Addi- tional combinations from prior studies were also explored including increasing the GCS cutoff and adding distracting painful injury as a high-risk criterion.

Pelvic CT protocol

All patients underwent a combined abdominal and pelvic CT scan on a Lightspeed GE 16-slice CT scanner (GE, Milwaukee, WI). Computed tomographic scans were obtained with 5-mm cuts from the diaphragm through the pelvic bones, and bone windows were reviewed to diagnose all bony fractures.

Data analysis

We describe the study population with simple descriptive statistics and 95% confidence intervals (CI). Interrater reliability was measured with the ? statistic. Data analysis was performed with STATA version 11.0 (2009; College Station, TX).

Results

We prospectively enrolled 4737 patients with a mean age of 39.8 +- 17.1 years. This represents an enrollment rate of 81% of eligible patients. Of the enrolled, 289 (6.1%; 95% CI,

n

Prevalence (95% CI)

Motor vehicle collision

146/2441

6.0%

(5.1%-7.0%)

Fall

33/625

5.3%

(3.7%-7.3%)

Auto versus pedestrian/bike

53/507

10.4%

(7.9%-13.4%)

Motorcycle/motocross

26/323

8.0%

(5.3%-11.6%)

Assault

1/501

0.2%

(0%-1.1%)

Fall of bike

1/52

1.9%

(0%-10.3%)

Torso crush

12/49

24.5%

(13.3%-38.9%)

Other

17/239

7.1%

(4.2%-11.1%)

5.4%-6.8%) were diagnosed with a pelvic fracture. The mechanisms of injury and the prevalence of pelvic fracture by mechanism of injury for the 4737 patients are presented in Table 1. Pelvic fractures were most prevalent after Crush injuries to the torso, and automobile vs pedestrian/bicycle and were exceedingly rare in victims of assaults. Of the 289 patients with pelvic fractures, 87 (30%; 95% CI, 25%- 36%) were treated with operative repair. Eighteen patients (6.2%; 95% CI, 3.7%-9.7%) had angiographic embolization.

Table 1 Prevalence of pelvic fracture by mechanism of injury

Interrater reliability

There was moderate agreement between physicians for the variables pelvic tenderness (? = 0.50; 95% CI, 0.43-0.58) and pelvic abrasions (? = 0.40; 95% CI, 0.33-0.48). The presence of pelvic bone instability (? = 0.36; 95% CI, 0.28- 0.43) had only fair agreement [8].

The test performance of the high-risk variables (hypoten- sion [systolic blood pressure b90 mm Hg], GCS score b14, pelvic bone instability, or pelvic bone tenderness) is presented in Table 2. The high-risk variables identified all 87 patients (100%; 95% CI, 96.6%-100%) undergoing pelvic bone surgery. Imploring a less restrictive GCS score (replacing GCS b14 with a GCS b15) had the following test characteristics: sensitivity, 266 (92.0%; 95% CI, 88.3%- 94.9%) of 289 and specificity, 3019 (67.8%; 95% CI, 66.5%- 69.2%) of 4448. Adding distracting painful injury to the original variables (hypotension, GCS score b14, pelvic bone instability/tenderness) had the following test characteristics: sensitivity, 274 (94.8%; 95% CI, 91.6%-97.1%) of 289 and

specificity, 2577 (57.9%; 95% CI, 56.5%-59.4%) of 4448.

Initial AP plain radiographs identified 234 (81.0%; 95% CI, 76.0%-85.3%) of 289 patients with pelvic fractures. The

Table 2 Performance of the high-risk clinical variables in the study

plain AP pelvic radiography identified a fracture in 83 patients (95.4%; 95% CI, 88.6%-98.7%) undergoing pelvic bone surgery. Descriptions of the 4 patients undergoing pelvic bone surgery but not identified by plain AP pelvic radiography are in Table 3.

Forty-eight patients (1.0%) were considered to have pelvic bone instability on physical examination. Eight (16.7%; 95% CI, 7.5%-30.2%) of these patients did not have a pelvic fracture. Twenty patients had open book-type pelvic fractures (rotationally unstable but vertically stable). Pelvic instability was only identified in 7 (35.0%; 95% CI, 15.3%-59.2%) of these.

Discussion

In this study, we demonstrate that clinical examination findings identify most patients with blunt torso trauma who have pelvic fractures. Patients with pelvic bone tenderness, hemodynamic instability, or decreased mental status (GCS scores b14) are at high risk and require pelvic bone radio- graphy. These criteria, though, fail to identify a small set of patients with pelvic fractures. None of those patients, how- ever, with pelvic fractures not identified by these high-risk criteria, required surgical therapy.

Furthermore, this study demonstrates that pelvic CT scan- ning is a superior diagnostic test to plain AP pelvic radiography. Plain AP pelvic radiography failed to identify 20% of patients with pelvic fractures, which is similar to prior smaller studies [2,9,10]. Computed tomographic superiority to Plain radiography for identification of traumatic injuries is known for other body locations as well (chest, cervical, thoracic, and lumbar spine) [11-14]. Not only does plain radiography fail to identify fractures but it also often fails to determine pelvic bone stability [15], although instability on plain pelvis radiography increases the likelihood of the need for angiographic embolization [16]. Most concerning, though, was the failure of plain AP pelvic radiography to identify all clinically important pelvic fractures.

Several small studies now suggest that plain AP pelvic radiography is unnecessary during the initial evaluation of the patient with trauma [17,18]. Our data support this con- cept. In those patients determined to require abdominal/pel- vic CT scanning, we suggest that avoidance of plain AP pelvic radiography as the pelvic CT component is sufficient

Any pelvic fracture

Pelvic fracture

undergoing surgery

Sensitivity

256/289

88.6%

(84.3%-92.0%)

87/87

100%

(96.6%-100%)

Specificity

3496/4448

78.6%

(78.3%-79.8%)

3529/4650

75.9%

(74.6%-77.1%)

Positive predictive value

256/1208

21.2%

(18.9%-23.6%)

87/1208

7.2%

(5.8%%-8.8%)

Negative predictive value

3496/3529

99.1%

(98.7%-99.4%)

3529/3529

100%

(99.9%-100%)

The tested clinical decision rule consisted of the following variables: hypotension (systolic blood pressure b90 mm Hg), Glasgow Coma Scale score less than 14, pelvic bone instability, or pelvic bone tenderness.

Age (y) Mechanism Injury Additional clinical findings Pelvic surgery

27 Fall 30 ft Acetabulum, ilium Liver and splenic lacerations Acetabulum repair

33 Fall N20 ft Sacrum Unstable pelvis on physical examination Open fracture, stabilization

36 MVC at high speed Acetabulum Prehospital hypotension Acetabulum repair

59 MVC at 60 mph Sacrum flail chest Sacral screw

MVC, motor vehicle collision.

for evaluating the presence or absence of a pelvic fracture. Initial AP pelvic radiography is indicated when immediate management decisions would be made based on the AP pelvic radiograph such as those patients with hemodynamic instability or concern for an unstable pelvic fracture. As an unstable pelvic fracture may require pelvic binder placement for hemor- rhage control, identification during the Initial resuscitation is important [19]. However, routine plain AP pelvic radiography per “trauma protocol” is not indicated for Hemodynamically stable patients with no concerns for unstable pelvic fractures. We would also advocate plain AP pelvic radiography in those patients meeting high-risk criteria but are not otherwise planned to undergo abdominal/pelvic CT scanning.

Table 3 Characteristics of the patients with pelvic fractures undergoing surgery and with a normal plain AP pelvis radiograph

Uncertainty exists regarding the use of examining the pelvic bones for stability [1,20-22]. Our data further question the use of this examination. Identification of pelvic bone instability warrants placement of a pelvic binder to prevent further hemorrhage [23]. In this study, the positive predictive value of pelvic bone instability for any pelvic fracture was only 83%, and only 35% of patients with open-book pelvic fractures had pelvic instability identified on examination. In addition, pelvic bone instability was never a sole high-risk criteria for pelvic radiography (all had other variables). Finally, the interrater reliability of pelvic bone instability was only fair. This low ? value, however, may be artificially decreased because of the rarity of pelvic instability. Reliability of the ? values decreases, as the variable being studied is rare [8]. Even such, the variable does not appear replicable by different physicians.

Studies evaluating the performance of clinical variables for detecting pelvic fractures are limited by the necessity to collect data during the initial evaluation while the patient is in spinal precautions (supine and nonambulatory). Such studies have not tested the performance of ambulation for detecting pelvic fractures. It is highly unlikely that patients with clinically significant pelvic fractures (who are report- edly alert, stable, and negative for pelvic bone tenderness) would be cleared and would ambulate without pelvic bone pain. Obviously, before patient clearance, careful ambulation testing is mandatory, and inability or difficulty ambulating requires further diagnostic evaluation.

Prior studies have used a variety of mental status cutoffs (GCS scores from 10 to 14) to consider a patient at risk for pelvic fracture such that pelvic radiography is warranted. In this large, prospective study, altering the GCS score cutoff from GCS less than 14 to a GCS less than 15 resulted in an

improved sensitivity of the criteria but at a cost of worsened specificity. Based on our data, we would advocate for a GCS cutoff of less than 14 as an indicator for pelvic radiography (in those patients whose mechanism of injury place them at risk for pelvic fractures). In addition, some investigators have considered distracting painful injury as an indication for pelvic radiography [6]. Adding distracting injury to the high- risk criteria increases sensitivity to nearly 95%, but a substantial sacrifice in specificity occurs.

There are certain limitations with this study. The study was conducted at a single, level 1 trauma center. The results may not be generalizable to all centers. Although clinicians were requested to complete the data collection forms before CT scanning, it was not ethically possible to have the clinician complete the data form before knowledge of the plain AP pelvic radiography results. Thus, the findings on the data collection form may be biased as the physicians may have known the results of the plain AP pelvic radiograph. We would anticipate that this would bias the study in favor of the clinical decision rule (ie, falsely elevated sensitivity). As the decision rule did not identify all patients with pelvic fracture, we believe that the likely impact of this limitation is minimal. Subjects were enrolled based on undergoing abdominal/pelvic CT scanning. Although all subjects with pelvic fractures would undergo abdominal/pelvic CT scanning and be enrolled, patients evaluated solely with plain AP pelvic radio- graphy were not. Thus, we are unable to estimate reduction in plain radiography that would be realized if clinicians used these variables to determine pelvic radiography. Finally, we allowed resident and faculty physicians to complete data forms. These physicians may have different clinical skills and abilities to assess for the high-risk variables.

In conclusion, high-risk clinical variables for determining the need for pelvic radiography in patients with blunt trauma identified all patients requiring surgical therapy. In stable patients planned to undergo abdominal/pelvic CT scanning, initial plain pelvic radiography is not necessary.

References

  1. American College of Surgeons. Advanced trauma life support. Chicago, IL: ACS; 2008.
  2. Duane TM, Tan BB, Golay D, Cole Jr FJ, Weireter Jr LJ, Britt LD. Blunt trauma and the role of routine pelvic radiographs: a Prospective analysis. J Trauma 2002;53:463-8.
  3. Gonzalez RP, Fried PQ, Bukhalo M. The utility of clinical examination in screening for pelvic fractures in blunt trauma. J Am Coll Surg 2002;194:121-5.
  4. Salvino CK, Esposito TJ, Smith D, Dries D, Marshall W, Flisak M, et al. Routine pelvic x-ray studies in awake blunt trauma patients: a sensible policy? J Trauma 1992;33:413-6.
  5. Yugueros P, Sarmiento JM, Garcia AF, Ferrada R. Unnecessary use of pelvic x-ray in blunt trauma. J Trauma 1995;39:722-5.
  6. Sauerland S, Bouillon B, Rixen D, Raum MR, Koy T, Neugebauer EA. The reliability of clinical examination in detecting pelvic fractures in blunt trauma patients: a meta-analysis. Arch Orthop Trauma Surg 2004;124:123-8.
  7. Civil ID, Ross SE, Botehlo G, Schwab CW. Routine pelvic radio- graphy in severe blunt trauma: is it necessary? Ann Emerg Med 1988;17:488-90.
  8. Koch GG, Landis JR, Freeman JL, Freeman Jr DH, Lehnen RC. A general methodology for the analysis of experiments with repeated measurement of categorical data. Biometrics 1977;33:133-58.
  9. Hilty MP, Behrendt I, Benneker LM, Martinolli L, Stoupis C, Buggy DJ, et al. Pelvic radiography in ATLS algorithms: a diminishing role? World J Emerg Surg 2008;3:11.
  10. Lucey BC, Stuhlfaut JW, Hochberg AR, Varghese JC, Soto JA. Evaluation of blunt abdominal trauma using PACS-based 2D and 3D MDCT reformations of the lumbar spine and pelvis. AJR Am J Roentgenol 2005;185:1435-40.
  11. Berry GE, Adams S, Harris MB, Boles CA, McKernan MG, Collinson F, et al. Are plain radiographs of the spine necessary during evaluation after blunt trauma? accuracy of screening torso computed tomography in thoracic/lumbar spine fracture diagnosis. J Trauma 2005;59:1410-3 [discussion 1413].
  12. Holmes JF, Akkinepalli R. Computed tomography versus plain radio- graphy to screen for cervical spine injury: a meta-analysis. J Trauma 2005;58:902-5.
  13. Sampson MA, Colquhoun KB, Hennessy NL. Computed tomography whole body imaging in multi-trauma: 7 years experience. Clin Radiol 2006;61:365-9.
  14. Trupka A, Waydhas C, Hallfeldt KK, Nast-Kolb D, Pfeifer KJ, Schweiberer L. Value of thoracic computed tomography in the first assessment of severely injured patients with blunt chest trauma: results of a prospective study. J Trauma 1997;43:405-11 [discussion 411-2].
  15. Their ME, Bensch FV, Koskinen SK, Handolin L, Kiuru MJ. Diag- nostic value of pelvic radiography in the initial trauma series in blunt trauma. Eur Radiol 2005;15:1533-7.
  16. Fu CY, Wu SC, Chen RJ, Wang YC, Chung PK, Yeh CC, et al. Evaluation of pelvic fracture stability and the need for angioemboliza- tion: pelvic instabilities on plain film have an increased probability of requiring angioembolization. Am J Emerg Med 2009;27:792-6.
  17. Chmelova J, Mrazkova D, Dzupa V, Baca V, Grill R, Pleva L. The role of plain radiography in pelvic trauma in the era of advanced computed tomography. Acta Chir Orthop Traumatol Cech 2006;73:394-9.
  18. Kessel B, Sevi R, Jeroukhimov I, Kalganov A, Khashan T, Ashkenazi I, et al. Is routine portable pelvic X-ray in stable Multiple trauma patients always justified in a high technology era? Injury 2007;38:559-63.
  19. Obaid AK, Barleben A, Porral D, Lush S, Cinat M. Utility of plain film pelvic radiographs in blunt trauma patients in the emergency depart- ment. Am Surg 2006;72:951-4.
  20. Olson SA, Pollak AN. Assessment of pelvic ring stability after injury. Indications for surgical stabilization. Clin Orthop Relat Res 1996:15-27.
  21. Shlamovitz GZ, Mower WR, Bergman J, Chuang KR, Crisp J, Hardy D, et al. How (un)useful is the pelvic ring stability examination in diagnosing mechanically unstable pelvic fractures in blunt trauma patients? J Trauma 2009;66:815-20.
  22. Weingart SD. Blunt trauma patients require a pelvic stability exami- nation. J Emerg Med 2010.
  23. Scalea TM, Stein DM, O’Toole RV. Trauma. New York: McGraw- Hill Professional; 2008.

Leave a Reply

Your email address will not be published. Required fields are marked *