Article, Orthopedics

Do all patients with shoulder dislocations need prereduction x-rays?

Original Contribution

Do all patients with Shoulder dislocations need prereduction x-rays??

Joseph Orloski MD a,?, Barnet Eskin MD, PhD a,b,

Paul C. Allegra MD c, John R. Allegra MD, PhD a,b

aMorristown Memorial Residency in Emergency Medicine, Morristown Memorial Hospital, Morristown, NJ 07960, USA

bEmergency Medicine Associates of New Jersey Research Foundation, Livingston, NJ 07039, USA

cSt. Cloud Hospital, Emergency Department, St. Cloud, MN 56303, USA

Received 21 October 2009; revised 9 January 2010; accepted 11 January 2010

Abstract

Objective: Some would advocate against routine use of prereduction x-rays for shoulder dislocations. Our objective was to examine the percent of dislocations that also had fractures as a function of age to determine whether there are some decades of life with a sufficiently low risk of fracture to avoid routine prereduction x-rays.

Methods: This was a retrospective cohort study in 19 New Jersey and New York emergency departments. We used The International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9) codes to identify patients with shoulder dislocations and, among these, those who also had fractures. A priori, we chose to group the patients by decade of life. For patients with dislocations in each age group, we calculated the percent that also had fractures.

Results: Of the total 5 408 837 visits in the database, there were 7209 patients with dislocations, of which 465 (6.5%) had fractures. We found the lowest percent of fractures in the 2nd and 3rd decades of life (0.7% [95% confidence interval, 0.3%-1.2%] and 0.8% [0.4%-1.3%]). In the 4th and 5th decades of life, the percents were 2.6% (1.7%-3.5%) and 4.6% (3.2-6.0%), respectively, and they steadily increased to 19% or greater in the 8th to 10th decades of life. Omitting the prereduction x-rays for patients in the 2nd and 3rd decades of life would reduce the total number of prereduction x-rays by 40%.

Conclusion: In the second and third decades of life, less than 1% of patients with shoulder dislocations also had fractures. Because these patients have a sufficiently low risk of fracture, routine prereduction x- rays may not be necessary for them.

(C) 2011

Introduction

? Prior presentations: Society for Academic Emergency Medicine Annual Meeting, Poster Presentation, Washington DC, May 31, 2008.

* Corresponding author. Emergency Medicine Residency and Fellow-

ships, Morristown Memorial Hospital, Morristown, NJ 07960, USA. Fax: +1 973 290 7202.

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

Shoulder dislocations are the most frequently dislocated joint seen in the emergency department (ED) [1]. Although typical care of a shoulder dislocation consists of both prereduction and postreduction radiographs [1,2], multiple studies have questioned the utility of this practice [3-10]. It increasescost, timetodisposition[7], andthetimetoreduction, which may worsen Muscle spasm and make reduction more

0735-6757/$ - see front matter (C) 2011 doi:10.1016/j.ajem.2010.01.005

difficult [10]. Because of a small but potentially serious risk of failing to identify a persistent dislocation after reduction [5,9], omitting prereduction x-rays may be preferable to omitting Postreduction x-rays [11]. Although some have argued that fracture-dislocations can be successfully reduced in the ED [12], medicolegal concerns may make the identification of fractures important before manipulation.

When emergency physicians report that they are certain of the diagnosis of shoulder dislocation, they are correct in 98% to 100% of cases [7,8]. Although most shoulder dislocations can be safely reduced in the ED, a coexisting fracture may prevent the emergency physician from reducing the dislocation and require orthopedic consultation [1]. A previous study derived a clinical decision rule seeking to predict which patients with dislocations have clinically significant fractures; however, the study excluded patients younger than 18 years and also mandated x-rays on all patients with first-time dislocations regardless of low clinical suspicion of fracture-dislocation [3].

Our study attempts to determine whether age may be used to predict the likelihood of fracture-dislocation by examining the occurrence of fractures among patients with shoulder dislocations across all decades of life in a large database. We hypothesized that some decades of life may have a sufficiently low incidence of fracture associated with shoulder dislocations so that prereduction radiographs may be omitted.

Methods

We performed a retrospective cohort study in a computerized billing database of consecutive visits seen by emergency physicians in 19 New Jersey and New York EDs from January 1, 1996, to November, 20, 2007. In these EDs, an electronic medical record including a final diagnosis was generated by the treating emergency physicians for each of the patients seen. Trained coders reviewed all completed records and assigned The International Classification of Diseases, 9th Revision, Clinical Modification (ICD9) codes based on the final diagnosis. The annual audits of the coding process by an outside firm have consistently met or exceeded preset criteria. Furthermore, virtually all ED charts have ICD9 codes assigned (in the last yearly audit, only 0.07% of all ED charts were missed). Where the diagnosis was not explicit, the coders reviewed the entire medical record including the emergency physician and/or radiologist interpretation of the x-rays. The study population consisted of all patients with shoulder dislocations identified by the following ICD9 codes: 831.00, 831.01, 831.02, and 831.03. Fractures were identified by the following ICD9 codes: 811.00, 811.01, 811.02, 811.03, 811.09, 811.10, 811.11,

811.12, 811.13, 811.19, 812.00, 812.01, 812.02, 812.03,

812.09, 812.10, 812.20, 812.21, 812.30, and 812.31. A

priori, we chose to group the patients by decade of life. We

calculated the percentage of patients with shoulder disloca- tions who also had shoulder fractures in each decade of life. We also calculated the odds ratio of fracture versus no fracture in patients older than 40 years versus those aged 18 to 40 years for comparison with a previous study. Data were analyzed using Excel (Microsoft, Redmond, WA). We obtained institutional review board approval for this study.

Results

Of the 5 408 837 visits in the database, 7209 patients had dislocations, of which 34% were female. Fig. 1 shows the percentage of patients with dislocations in each decade of life. The median age for dislocations was 35 years (interquartile range, 23-56). As illustrated in Fig. 1, patients in the second, third, and fourth decades of life constituted 17%, 23%, and 17% of all the patients with dislocations. The number of patients with fractures was 465 (6.5% of patients with dislocations). All fractures identified were humeral; none were scapular. Fig. 2 shows the percentage of patients with fracture-dislocations across each decade of life. We found the lowest percentage of fracture-dislocations in the second and third decades of life with 9 of 1233 (0.7%; 95%

confidence interval, 0.3%-1.2%) and 14 of 1682 (0.8%; 0.4%-1.3%) patients, respectively. Above the third decade of life, the percentage of fracture-dislocations steadily increases with each decade. When comparing patients with ages older than 40 years to those between 18 and 40 years old, the odds ratio for fracture versus no fracture was 11.4 (8.2-15.9).

Discussion

Several studies have established the ability of emergency physicians to accurately diagnose shoulder dislocations without prereduction x-rays [3,10]; however, they have not addressed the risk of missing a fracture-dislocations [7,8]. Our study shows that patients in the second to third decades

Fig. 1 The percentage of the total shoulder dislocations of patients in each decade of life.

Fig. 2 The percentage of fracture-dislocations among patients with shoulder dislocations in each decade of life with 95% confidence intervals.

of life are unlikely to have fracture-dislocations, so prereduction x-rays may be safely omitted. Of course, this does not apply when the clinical suspicion for fracture is high. For these patients, omitting all prereduction x-rays would miss 0.8% of fractures regardless of mechanism, history, or suspicion of fracture, and this would lead to a 40% decrease in prereduction x-ray utilization. If patients in the fourth decade of life were also included, this would decrease prereduction x-ray utilization by 57%, but the rate of Missed fractures would increase slightly to 1.3%. This can be compared with the potential of Emond et al for reducing prereduction x-rays by 20% using the three predictive factors, while missing 3% of clinically significant lesions. We would expect that in clinical practice, physicians would continue to order prereduction x-rays on all patients with high-energy traumatic mechanisms of dislocation, or patients otherwise believed to be of High clinical suspicion for fracture, which would further decrease the rate of missed clinically significant fracture-dislocations.

We found an overall incidence of fractures in patients with dislocations of 6.5%. This is lower than the incidence found in the 2 studies of Emond et al, 25.5% and 18%. However, these results cannot be directly compared because the studies of Emond et al excluded patients younger than 18 years. Our study found a relative risk for fracture versus no fracture of 6.8 (5.5-8.0) for patients older than 40 years compared with those aged 18 to 40 years. We calculated the value of Emond et al [3] for this to be 2.6 (1.9-3.1) [13], although they compared presence versus absence of clinically significant fractures and excluded patients younger than 18 years.

Our study design has some limitations. The design of our study did not allow us to collect data regarding the circumstances of fracture-dislocations including mecha- nism, frequency of Recurrent dislocations, and baseline x- ray utilization rates. Because each chart was identified by 1 trained coder, we do not have information to calculate a ? for intraobserver reliability. Furthermore, because our retrospective study relies on ICD9 codes based partly on

the final diagnosis assigned to be an emergency physician, shoulder fractures that were not listed as a final diagnosis may not have been included in the study. Although this would artificially lower the percentage of fracture-disloca- tions, an emergency physician would be unlikely to omit the diagnosis of a clinically significant fracture in the medical records. We did not differentiate between clinically significant and insignificant fractures. As noted above, not getting prereduction x-rays on patients in the second and third decades of life would mean that 23 (0.8%) of 2915 patients would have reductions attempted when concomi- tant fractures were present. However, if we restrict fractures to those that are clinically significant and if, for example, 24% of the fractures are not clinically significant as was found in the study of Emond et al, then the actual rate of reductions attempted when clinically significant fractures are present would be only 0.6%. Furthermore, it is possible that if only clinically significant fractures are considered, this rate in other decades of life (eg, the fourth decade) would be sufficiently low so that those patients could also forego prereduction x-rays. This would further lessen the number of prereduction x-rays that would need to be done. Therefore, distinguishing between clinically insignificant and significant fractures would likely make the clinical impact of our findings even more impressive than what we present here.

Our study shows that in the second and third decades of life, less than 1% of patients with shoulder dislocations also had fractures, suggesting that for these ages, routine prereduction x-rays may not be necessary unless otherwise warranted by high clinical suspicion, but should be considered for other age groups. Omitting prereduction x- rays in patients in the second and third decades of life would reduce prereduction x-ray utilization by 40%.

References

  1. Uehara DT, Rudzinski JP. Injuries to the shoulder complex and humerus. In: Tintinalli JE, Kelen GD, Stapczynski JS, Ma OJ, Cline DM, editors. Tintinalli’s emergency medicine: a comprehensive study guide. 6th ed. Chapel Hill (NC): McGraw-Hill; 2004. p. 1695-702.
  2. Chong M, Karataglis D, Learmonth D. Survey of the management of acute traumatic first-time Anterior shoulder dislocation among trauma clinicians in the UK. Ann R Coll Surg Engl 2006;88:454-8.
  3. Emond M, Le Sage N, Lavoie A, et al. Clinical factors predicting fractures associated with an anterior shoulder dislocation. Acad Emerg Med 2004;11:853-8.
  4. Shuster M, Abu-Laban RB, Boyd J, et al. Prospective evaluation of a guideline for the selective elimination of pre-reduction radiographs in clinically obvious anterior shoulder dislocation. CJEM 2002;4:257-62.
  5. Harvey RA, Trabulsy ME, Roe L. Are postreduction anteroposterior and scapulary views useful in anterior shoulder dislocations? Am J Emerg Med 1992;10:149-51.
  6. Kahn JH, Mehta SD. The role of post-reduction radiographs after shoulder dislocation. J Emerg Med 2007;33:169-73.
  7. Hendey GW, Chally MK, Stewart VB. Selective radiography in 100 patients with suspected shoulder dislocation. J Emerg Med 2006;31: 23-8.
  8. Hendey GW. Necessity of radiographs in the emergency department management of shoulder dislocations. Ann Emerg Med 2000;36: 108-13.
  9. Hendey GW, Kinlaw K. Clinically significant abnormalities in postreduction radiographs after anterior shoulder dislocation. Ann Emerg Med 1996;28:399-402.
  10. Shuster M, Abu-Laban RB, Boyd J. Prereduction radiographs in clinically evident anterior shoulder dislocation. Am J Emerg Med 1999;17:653-8.
  11. Shuster M, Abu-Laban RB, Boyd J. The role of radiographs in anterior shoulder dislocation. [letter]J Emerg Med 2008 [Epub ahead of print].
  12. Perron AD, Ingerski MS, Brady WJ, et al. Acute complications associated with shoulder dislocation at an academic emergency department. J Emerg Med 2003;24(2):141-5.
  13. Zhang J, Yu K. What’s the risk? A method of correcting the odds ratio in cohort studies of common outcomes. JAMA 1998;280: 1690-1.