PEDIATRICS/BRIEF RESEARCH REPORT
Accuracy of Ultrasonography for Determining Successful Realignment of Pediatric Forearm Fractures Alexander Sasha Dubrovsky, MDCM, MSc, FRCPC*; Anna Kempinska, MD, FRCPC; Ilana Bank, MDCM, FRCPC, FAAP; Elise Mok, PhD *Corresponding Author. E-mail: [email protected], Twitter: @SashaDuby.
Study objective: The primary objective of this study is to assess the accuracy of point-of-care ultrasonography compared with blinded orthopedic assessment of fluoroscopy in determining successful realignment of pediatric forearm fractures. The secondary objective is to determine the rate of agreement of ultrasonography and fluoroscopy in real-time by the treating physician. Methods: A cross-sectional study was conducted in children younger than 18 years and presenting to an academic emergency department with forearm fractures requiring realignment of a single bone. Physicians performed closed reductions with ultrasonographic assessment of realignment until the best possible reduction was achieved. Fluoroscopy was then immediately performed and images were saved. A positive test result was defined as an inadequately reduced fracture on fluoroscopy by a blinded pediatric orthopedic surgeon (reference standard) and on ultrasonography (index test) and fluoroscopy in real-time by the treating physician. Results: One hundred patients were enrolled (median age 12.1 years; 74% male patients); the radius was involved in 98%, with 27% involving the growth plate. The sensitivity, specificity, positive predictive value, and negative predictive value were 50% (95% confidence interval [CI] 15.4% to 84.6%), 89.1% (95% CI 82.8% to 95.5%), 28.6% (95% CI 4.9% to 52.2%), and 95.3% (95% CI 90.9% to 99.8%), respectively. The treating physicians’ agreement rate of the real-time images was 98%. Conclusion: Point-of-care ultrasonography can help emergency physicians determine when pediatric forearm fractures have been adequately realigned, but inadequate reductions should be confirmed by other imaging modalities. [Ann Emerg Med. 2015;65:260-265.] Please see page 261 for the Editor’s Capsule Summary of this article. A podcast for this article is available at www.annemergmed.com. 0196-0644/$-see front matter Copyright © 2014 by the American College of Emergency Physicians. http://dx.doi.org/10.1016/j.annemergmed.2014.08.043
INTRODUCTION Background Fractures are common during childhood, with radius and ulna being among the most commonly injured bones. Pediatric emergency physicians reduce misaligned fractures, occasionally with fluoroscopy providing real-time assessment of fracture realignment before the termination of the procedure. Point-of-care ultrasonography is an alternative real-time imaging modality with the advantages of being increasingly available in emergency departments (EDs) and emitting no radiation. To date, preliminary pediatric studies have suggested ultrasonography to be a useful tool in determining need for reduction1-3 and confirming successful realignment of forearm fractures.1,2,4,5 However, the generalizability of the results has been limited by small sample sizes, heterogeneous fracture types, and a limited number of providers. To date, to our knowledge no study has compared ultrasonographic performance with the other real-time imaging modality (ie, fluoroscopy) that is available in some centers. 260 Annals of Emergency Medicine
Importance Although fluoroscopy provides real-time images that can confirm adequate fracture realignments, it requires fluoroscopic equipment and personnel, is not widely available, and exposes the patient and medical staff to small amounts of ionizing radiation. If point-of-care ultrasonography can accurately determine successful fracture realignment, it might improve the quality of care for children undergoing either blind or fluoroscopic-guided fracture reductions. Goals of This Investigation Our objective was to determine the test characteristics of point-of-care ultrasonography compared with blinded orthopedic assessment of fluoroscopy in assessing the adequacy of closed reduction of pediatric forearm fractures. The secondary objective was to establish the rate of agreement of ultrasonography and fluoroscopy in real-time by the treating pediatric emergency physician. Volume 65, no. 3 : March 2015
Dubrovsky et al
Ultrasonographically Guided Reduction of Pediatric Forearm Fractures
Editor’s Capsule Summary
What is already known on this topic Bedside ultrasonography may be useful in assessing forearm fractures. What question this study addressed This study reported test characteristics of emergency physician–performed ultrasonography for closed reduction adequacy of pediatric forearm fractures, compared with orthopedist review of fluoroscopic images. What this study adds to our knowledge In a study of 100 patients, the negative predictive value of ultrasonography for fracture malalignment was 95% (95% confidence interval [CI] 91% to 100%), but positive predictive value was only 29% (95% CI 5% to 52%); ultrasonography was useful in determining adequate fracture realignment, but not in identifying need for further reduction. How this is relevant to clinical practice If findings are confirmed in larger studies in a variety of settings, ultrasonography showing adequate alignment may obviate the need for postreduction radiography, assuming that casting will not alter the alignment.
MATERIALS AND METHODS Study Design and Setting This was a cross-sectional study conducted from May 2012 to May 2014 in an academic pediatric ED. A convenience sample of eligible patients was prospectively recruited when a research assistant was present (5 days per week, 2 PM to 10 PM) and a trained study physician was available. Written informed assent and consent was obtained from participants and their guardians. Approval for this study was granted by the hospital’s institutional ethics review board. Selection of Participants Participating study physicians were emergency physicians who routinely performed forearm reductions. Before the study launch date, no physician was using ultrasonography to assess adequacy of reductions and most (88%) were beginners at basic point-of-care ultrasonographic skills (ie, attended an introductory course and performed fewer than 30 focused abdominal sonographic procedures in trauma examinations). Eligible patients were included if aged 0 to 18 years with acute (72 hours) forearm fractures for which the treating physician deemed a closed reduction of a single bone was necessary. Given the inherent technical difficulty in reducing fractures in which both bones need realignment, coupled with novice ultrasonographic providers, we excluded fractures in which both the radius and ulna required reductions. Patients were excluded if they had a Monteggia or Galeazzi fracture; intra-articular, open,
Figure 1. Study flow chart. A positive or negative ultrasonographic result was defined as an inadequately or adequately realigned fracture, respectively, as assessed by the treating physician. A positive or negative fluoroscopic result was defined as an inadequately or adequately realigned fracture, respectively, as determined by a blinded pediatric orthopedic surgeon. Volume 65, no. 3 : March 2015
Annals of Emergency Medicine 261
Dubrovsky et al
Ultrasonographically Guided Reduction of Pediatric Forearm Fractures or neurovascularly impaired fractures; or could not provide consent (polytraumatized; non-French or non-English speaking).
Table. Demographics and clinical characteristics (N¼100). Subject Characteristics
Percentage (%)*
†
Methods of Measurement Before the start of the study, study physicians attended a 1-hour educational simulation session with 4 components (Appendix E1, available online at http://www.annemergmed. com): (1) introductory video demonstrating the skills required to assess adequacy of reductions, (2) ultrasonographic visualization of one another’s forearms, (3) ultrasonographic visualization of fractures of different characteristics (ie, nondisplaced, angulated, and displaced) on fractured chicken bones embedded in gelatin,6 and (4) a mobile angulated fracture model (homemade hinge was applied to each end of a fractured chicken bone) that allowed real-time ultrasonographic visualization of fracture realignment. Indications for fractures requiring reductions and what constituted adequate realignment were based on local practice guidelines that stipulated that fractures need to be less than 15 degrees angulated in prepubertal children (aged 20 degrees 53% > 30 degrees 17% Angulation of distal fragment Volar 29% Dorsal 69% Displacement: Any 25% 50-99% 10% 100% (ie, non-overlapping) 5% Fluoroscopic radiation indices§ Time, median (IQR), seconds 6.7 (4.8-9) Dose, median (IQR), Mradcm2 544.3 (431.3-777.2) *Data are presented as proportion unless otherwise indicated. † IQR, interquartile range (25% to 75%) ‡ Location of fractures on the radius and ulna were defined as distal if the fracture was in the distal third of the bone, mid if in the middle third, as a Salter Harris 1 fracture if displaced at the level of the physis and as a Salter Harris 2 fracture if displaced at the physis with a metaphyseal extension of the fracture line. § Fluoroscopic radiation indices were based on the total radiation emitted; this included the two-views obtained immediately following the ultrasound and all additional images taken for continued fluoroscopic-guided fracture re-alignment and/ or following application of the cast.
Outcome Measures A positive case result was defined as an inadequately reduced fracture (ie, abnormal anatomic alignment present) on the imaging modality being assessed. The treating physician assessed fracture realignment on the ultrasonographic (index test) and the fluoroscopic images in real-time. A single pediatric orthopedic surgeon, blinded to the treating physicians’ assessments of the ultrasonographic and fluoroscopic images, assessed the fluoroscopic images and this was used as the reference standard. Volume 65, no. 3 : March 2015
Dubrovsky et al
Ultrasonographically Guided Reduction of Pediatric Forearm Fractures
Primary Data Analysis Descriptive statistics were used to report the characteristics of our subjects and test performance, including sensitivity, specificity, and positive and negative predictive values with 95% confidence intervals (CIs). The concordance of the treating physicians’ real-time assessment of ultrasonography and fluoroscopy was used to calculate the agreement rate. Data were analyzed with SPSS Statistics (version 22; IBM, Armonk, NY).
RESULTS Characteristics of Study Subjects One hundred patients were enrolled (Figure 1), with their clinical information presented in the Table. Participating study physicians (n¼16) were pediatric (n¼12) or general (n¼4) emergency physicians with a median number of years in independent practice of 7.5 (interquartile range 2.3 to 15 years) and among whom 69% reported performing 1 to 5 reductions per month; 31%, 6 to 10 per month. Main Results The ultrasonography performed by study physicians compared with the reference standard and agreement with their assessment of fluoroscopy is shown in Figure 2. Physicians performed a median number of 4 ultrasonographically guided reductions; 78% ranked its use as easy and 5% as difficult, and 78% indicated that ultrasonography was helpful with the reduction procedure. Test performance revealed a sensitivity and specificity of 50% (95% CI 15.4% to 84.6%) and 89.1% (95% CI 82.8% to 95.5%), respectively. The negative predictive value was excellent, at 95.3% (95% CI 90.9% to 99.8%), albeit the positive predictive value was 28.6% (95% CI 4.9% to 52.2%), indicating that ultrasonography was best at identifying when fractures were adequately realigned. As seen in Figure 2A, the results were similar in a sensitivity analysis in which subjects treated by the principal investigator (n¼30) were excluded. To assess the possible association of provider experience with ultrasonography, subanalysis was conducted in which physicians who performed less than or equal to 5 and less than or equal to 10 fracture reductions were compared with those who performed greater than 5 and greater than 10 fracture reductions (ie, “experienced”); more experienced providers were more specific at determining an adequately reduced fracture (Figure 2A). Furthermore, the top 2 physicians (ie, together performed nearly half of the reductions with ultrasonography) performed with perfect sensitivity compared with those who performed at less than 10, who performed with poor sensitivity. The agreement rate between real-time assessment of ultrasonography and fluoroscopy by the treating physician was 98%; one physician had 2 cases in which the ultrasonography was assessed as inadequate and fluoroscopy adequate, and there were no cases in which the treating physician’s ultrasonography was assessed as adequate but fluoroscopy inadequate (Figure 2B). Volume 65, no. 3 : March 2015
LIMITATIONS Our study’s main limitation was the inclusion of a convenience sample of patients enrolled when the research assistant and a study physician were present. A potential source of bias was introduced into the study because the principal investigator (A.S.D.) performed 30% of the ultrasonographically guided reductions. The results remained similar when these subjects were removed from the analysis. Although the ultrasonographic skill required to assess fracture reduction was easily learned, reproducibility of the test was not evaluated. All of the providers were novice at this ultrasonographic skill, and with a similar training program our results may be generalizable to other pediatric EDs. The small number of inadequately reduced fractures limited the precision with which the test characteristics could be determined. Another limitation is the subjective nature of the reference test; there are no definitive guidelines on what constitutes an age-appropriate adequate reduction, albeit the blinded pediatric orthopedic surgeon’s assessment of adequacy represented our standard of care from which follow-up decisions were made. Moreover, we obtained the fluoroscopic images immediately after the ultrasonography as our reference standard, not images obtained post–cast application, to directly compare adequacy of the ultrasonographic images before any further manipulation or cast application. It is possible that fracture alignment changed after application of the appropriate cast. Last, our results apply only to patients for whom a single-bone fracture requires realignment because we excluded those in whom both the radius and ulna required reduction.
DISCUSSION Our results indicate that point-of-care ultrasonography performed by emergency physicians can help determine when fractures have been successfully realigned, with a high negative predictive value. Given the increasing availability of ultrasonography in EDs, along with the simple and safe technique that allows real-time visualization, ultrasonographic confirmation of forearm fracture realignment should be encouraged in place of blind (ie, no real-time imaging) reductions and as a possible alternative to fluoroscopy. To our knowledge, this is the first pediatric study to assess the diagnostic accuracy of ultrasonography compared with fluoroscopy, as well as the first study in a pediatric ED that used a large group (n¼16) of novice providers. Chinnock et al7 reported the diagnostic accuracy of ultrasonography in an ED treating both children and adults; the test characteristics would be strikingly similar to ours had their analysis used a similar case definition (ie, positive result defined as inadequately realigned). One drawback of studies to date is the subjective reference standards used as opposed to patient-focused outcomes (for example, the need for reintervention). Pediatric forearm fractures remodel well despite residual malalignments. In our study, follow-up radiographs of 3 of the 4 false-negative cases revealed adequate fracture alignments in the appropriate casts and no need for reinterventions. Similarly, among fractures correctly identified as inadequately reduced, one had a residual 28-degree angulation, albeit active remodeling was observed in follow-up. Annals of Emergency Medicine 263
Ultrasonographically Guided Reduction of Pediatric Forearm Fractures
264 Annals of Emergency Medicine
Dubrovsky et al
Volume 65, no. 3 : March 2015
Dubrovsky et al
Ultrasonographically Guided Reduction of Pediatric Forearm Fractures
Figure 2. A, Diagnostic results for each point-of-care ultrasonographic study performed by the treating physician in the pediatric ED compared with the reference standard of the blinded pediatric orthopedic surgeon’s assessment of fluoroscopic images. B, Agreement between each of the treating physician’s own assessments of his or her ultrasonographic and fluoroscopic images for fracture realignments. Each block represents a unique patient enrolled per study physician (x axis), with the number of ultrasonographic scans performed by each physician presented from the first ultrasonography performed at the bottom and arranged vertically in chronologic order. Blocks are color coded as per the inset. A true positive and true negative were an inadequately and adequately realigned fracture, respectively, on both ultrasonography and fluoroscopy; a false positive was an inadequately realigned fracture on ultrasonography but adequate on fluoroscopy, whereas a false negative was an adequately realigned fracture on ultrasonography but inadequately aligned on fluoroscopy. *Test characteristics of ultrasonography versus blinded pediatric orthopedic surgeon’s assessment of fluoroscopy. †Test characteristics with principal investigator’s cases excluded. ‡ Test characteristics among physicians who performed less than or equal to 5 versus greater than 5 fracture reductions using ultrasonography. §Test characteristics amongst those who performed less than or equal to 10 versus greater than 10 fracture reductions using ultrasonography.
Although this study was not designed to assess the test performance among users of differing experience levels, a trend toward improved specificity among more experienced providers was noted. The sensitivity also greatly improved among the 2 providers with the greatest experience. Experienced providers might have been able to more precisely differentiate the periosteum (which can appear angulated) from the cortex, or use 3 views more consistently than 2, leading to a more accurate assessment of adequacy. Although future studies are needed to determine the optimal number of procedures to attain competency, it can be conjectured that accuracy may be maximized once physicians perform more than 10 reductions with ultrasonography. In summary, point-of-care ultrasonography can accurately assess the successful realignment of fractures of the forearm in children, but inadequate reductions should be confirmed by other imaging modalities. Ultrasonography provides clinicians the ability to safely guide fracture realignment without requiring the equipment or personnel needed for fluoroscopy, and it does not expose the patient or the staff to any radiation. Future randomized studies are needed to determine whether this approach improves procedure time, number of attempts or successful reductions, and patient and parental satisfaction, as well as whether it can aid in the reduction of fractures in which both bones need realignment. The authors acknowledge Rami Rezk, BSc, (research assistant) and Thierry Benaroch, MD, MSc, (from the Department of Pediatric Orthopedic Surgery) for their help and dedication to this project.
Supervising editor: Kelly D. Young, MD, MS Author affiliations: From the Division of Pediatric Emergency Medicine (Dubrovsky, Bank) and the Clinical Research Centre (Mok), Montreal Children’s Hospital-McGill University Health Center, Montreal, Quebec, Canada; the Faculty of Medicine (Dubrovsky, Bank), and Centre for Medical Education and Arnold and Blema Steinberg Simulation Centre (Bank), McGill University, Montreal, Quebec, Canada; and the Division of Pediatric Emergency Medicine, Hospital for Sick Children and Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada (Kempinska). Author contributions: ASD and AK conceived the study, designed the trial, and obtained research funding. ASD supervised the Volume 65, no. 3 : March 2015
conduct of the trial and data collection, as well as managed the data, including quality control. EM contributed to the design of the study and analysis of the data. ASD drafted the article, and all authors contributed substantially to its revision. ASD takes responsibility for the paper as a whole. Funding and support: By Annals policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article as per ICMJE conflict of interest guidelines (see www.icmje.org). The authors have stated that no such relationships exist. Funding for the research assistant was provided by the Department of Pediatric, Montreal Children’s Hospital through an internal research fellowship competition. The lead author is a member of the Research Institute of the McGill University Health Centre, which is supported by the Fonds de Recherche du Québec–Santé. Publication dates: Received for publication May 3, 2014. Revisions received July 2, 2014; July 28, 2014; and August 11, 2014. Accepted for publication August 26, 2014. Available online October 16, 2014. Presented at the Canadian Association of Emergency Physicians annual conference, May 2014, Ottawa, Canada; and the Canadian Pediatric Society annual conference, June 2014, Montreal, Canada.
REFERENCES 1. Chen L, Kim Y, Moore CL. Diagnosis and guided reduction of forearm fractures in children using bedside ultrasound. Pediatr Emerg Care. 2007;23:528-531. 2. Patel DD, Blumberg SM, Crain EF. The utility of bedside ultrasonography in identifying fractures and guiding fracture reduction in children. Pediatr Emerg Care. 2009;25:221-225. 3. Barata I, Spencer R, Suppiah A, et al. Emergency ultrasound in the detection of pediatric long-bone fractures. Pediatr Emerg Care. 2012;28:1154-1157. 4. Durston W, Swartzentruber R. Ultrasound guided reduction of pediatric forearm fractures in the ED. Am J Emerg Med. 2000;18: 72-77. 5. Wong CE, Ang AS, Ng KC. Ultrasound as an aid for reduction of paediatric forearm fractures. Int J Emerg Med. 2008;1:267-271. 6. Heiner JD, McArthur TJ. A simulation model for the ultrasound diagnosis of long-bone fractures. Simul Healthc. 2009;4: 228-231. 7. Chinnock B, Khaletskiy A, Kuo K, et al. Ultrasound-guided reduction of distal radius fractures. J Emerg Med. 2011;40:308-312.
Annals of Emergency Medicine 265
Accuracy of Ultrasonography for Determining Successful Realignment of Pediatric Forearm Fractures Alexander Sasha Dubrovsky, MDCM, MSc, FRCPC*; Anna Kempinska, MD, FRCPC; Ilana Bank, MDCM, FRCPC, FAAP; Elise Mok, PhD *Corresponding Author. E-mail: [email protected], Twitter: @SashaDuby.
Study objective: The primary objective of this study is to assess the accuracy of point-of-care ultrasonography compared with blinded orthopedic assessment of fluoroscopy in determining successful realignment of pediatric forearm fractures. The secondary objective is to determine the rate of agreement of ultrasonography and fluoroscopy in real-time by the treating physician. Methods: A cross-sectional study was conducted in children younger than 18 years and presenting to an academic emergency department with forearm fractures requiring realignment of a single bone. Physicians performed closed reductions with ultrasonographic assessment of realignment until the best possible reduction was achieved. Fluoroscopy was then immediately performed and images were saved. A positive test result was defined as an inadequately reduced fracture on fluoroscopy by a blinded pediatric orthopedic surgeon (reference standard) and on ultrasonography (index test) and fluoroscopy in real-time by the treating physician. Results: One hundred patients were enrolled (median age 12.1 years; 74% male patients); the radius was involved in 98%, with 27% involving the growth plate. The sensitivity, specificity, positive predictive value, and negative predictive value were 50% (95% confidence interval [CI] 15.4% to 84.6%), 89.1% (95% CI 82.8% to 95.5%), 28.6% (95% CI 4.9% to 52.2%), and 95.3% (95% CI 90.9% to 99.8%), respectively. The treating physicians’ agreement rate of the real-time images was 98%. Conclusion: Point-of-care ultrasonography can help emergency physicians determine when pediatric forearm fractures have been adequately realigned, but inadequate reductions should be confirmed by other imaging modalities. [Ann Emerg Med. 2015;65:260-265.] Please see page 261 for the Editor’s Capsule Summary of this article. A podcast for this article is available at www.annemergmed.com. 0196-0644/$-see front matter Copyright © 2014 by the American College of Emergency Physicians. http://dx.doi.org/10.1016/j.annemergmed.2014.08.043
INTRODUCTION Background Fractures are common during childhood, with radius and ulna being among the most commonly injured bones. Pediatric emergency physicians reduce misaligned fractures, occasionally with fluoroscopy providing real-time assessment of fracture realignment before the termination of the procedure. Point-of-care ultrasonography is an alternative real-time imaging modality with the advantages of being increasingly available in emergency departments (EDs) and emitting no radiation. To date, preliminary pediatric studies have suggested ultrasonography to be a useful tool in determining need for reduction1-3 and confirming successful realignment of forearm fractures.1,2,4,5 However, the generalizability of the results has been limited by small sample sizes, heterogeneous fracture types, and a limited number of providers. To date, to our knowledge no study has compared ultrasonographic performance with the other real-time imaging modality (ie, fluoroscopy) that is available in some centers. 260 Annals of Emergency Medicine
Importance Although fluoroscopy provides real-time images that can confirm adequate fracture realignments, it requires fluoroscopic equipment and personnel, is not widely available, and exposes the patient and medical staff to small amounts of ionizing radiation. If point-of-care ultrasonography can accurately determine successful fracture realignment, it might improve the quality of care for children undergoing either blind or fluoroscopic-guided fracture reductions. Goals of This Investigation Our objective was to determine the test characteristics of point-of-care ultrasonography compared with blinded orthopedic assessment of fluoroscopy in assessing the adequacy of closed reduction of pediatric forearm fractures. The secondary objective was to establish the rate of agreement of ultrasonography and fluoroscopy in real-time by the treating pediatric emergency physician. Volume 65, no. 3 : March 2015
Dubrovsky et al
Ultrasonographically Guided Reduction of Pediatric Forearm Fractures
Editor’s Capsule Summary
What is already known on this topic Bedside ultrasonography may be useful in assessing forearm fractures. What question this study addressed This study reported test characteristics of emergency physician–performed ultrasonography for closed reduction adequacy of pediatric forearm fractures, compared with orthopedist review of fluoroscopic images. What this study adds to our knowledge In a study of 100 patients, the negative predictive value of ultrasonography for fracture malalignment was 95% (95% confidence interval [CI] 91% to 100%), but positive predictive value was only 29% (95% CI 5% to 52%); ultrasonography was useful in determining adequate fracture realignment, but not in identifying need for further reduction. How this is relevant to clinical practice If findings are confirmed in larger studies in a variety of settings, ultrasonography showing adequate alignment may obviate the need for postreduction radiography, assuming that casting will not alter the alignment.
MATERIALS AND METHODS Study Design and Setting This was a cross-sectional study conducted from May 2012 to May 2014 in an academic pediatric ED. A convenience sample of eligible patients was prospectively recruited when a research assistant was present (5 days per week, 2 PM to 10 PM) and a trained study physician was available. Written informed assent and consent was obtained from participants and their guardians. Approval for this study was granted by the hospital’s institutional ethics review board. Selection of Participants Participating study physicians were emergency physicians who routinely performed forearm reductions. Before the study launch date, no physician was using ultrasonography to assess adequacy of reductions and most (88%) were beginners at basic point-of-care ultrasonographic skills (ie, attended an introductory course and performed fewer than 30 focused abdominal sonographic procedures in trauma examinations). Eligible patients were included if aged 0 to 18 years with acute (72 hours) forearm fractures for which the treating physician deemed a closed reduction of a single bone was necessary. Given the inherent technical difficulty in reducing fractures in which both bones need realignment, coupled with novice ultrasonographic providers, we excluded fractures in which both the radius and ulna required reductions. Patients were excluded if they had a Monteggia or Galeazzi fracture; intra-articular, open,
Figure 1. Study flow chart. A positive or negative ultrasonographic result was defined as an inadequately or adequately realigned fracture, respectively, as assessed by the treating physician. A positive or negative fluoroscopic result was defined as an inadequately or adequately realigned fracture, respectively, as determined by a blinded pediatric orthopedic surgeon. Volume 65, no. 3 : March 2015
Annals of Emergency Medicine 261
Dubrovsky et al
Ultrasonographically Guided Reduction of Pediatric Forearm Fractures or neurovascularly impaired fractures; or could not provide consent (polytraumatized; non-French or non-English speaking).
Table. Demographics and clinical characteristics (N¼100). Subject Characteristics
Percentage (%)*
†
Methods of Measurement Before the start of the study, study physicians attended a 1-hour educational simulation session with 4 components (Appendix E1, available online at http://www.annemergmed. com): (1) introductory video demonstrating the skills required to assess adequacy of reductions, (2) ultrasonographic visualization of one another’s forearms, (3) ultrasonographic visualization of fractures of different characteristics (ie, nondisplaced, angulated, and displaced) on fractured chicken bones embedded in gelatin,6 and (4) a mobile angulated fracture model (homemade hinge was applied to each end of a fractured chicken bone) that allowed real-time ultrasonographic visualization of fracture realignment. Indications for fractures requiring reductions and what constituted adequate realignment were based on local practice guidelines that stipulated that fractures need to be less than 15 degrees angulated in prepubertal children (aged 20 degrees 53% > 30 degrees 17% Angulation of distal fragment Volar 29% Dorsal 69% Displacement: Any 25% 50-99% 10% 100% (ie, non-overlapping) 5% Fluoroscopic radiation indices§ Time, median (IQR), seconds 6.7 (4.8-9) Dose, median (IQR), Mradcm2 544.3 (431.3-777.2) *Data are presented as proportion unless otherwise indicated. † IQR, interquartile range (25% to 75%) ‡ Location of fractures on the radius and ulna were defined as distal if the fracture was in the distal third of the bone, mid if in the middle third, as a Salter Harris 1 fracture if displaced at the level of the physis and as a Salter Harris 2 fracture if displaced at the physis with a metaphyseal extension of the fracture line. § Fluoroscopic radiation indices were based on the total radiation emitted; this included the two-views obtained immediately following the ultrasound and all additional images taken for continued fluoroscopic-guided fracture re-alignment and/ or following application of the cast.
Outcome Measures A positive case result was defined as an inadequately reduced fracture (ie, abnormal anatomic alignment present) on the imaging modality being assessed. The treating physician assessed fracture realignment on the ultrasonographic (index test) and the fluoroscopic images in real-time. A single pediatric orthopedic surgeon, blinded to the treating physicians’ assessments of the ultrasonographic and fluoroscopic images, assessed the fluoroscopic images and this was used as the reference standard. Volume 65, no. 3 : March 2015
Dubrovsky et al
Ultrasonographically Guided Reduction of Pediatric Forearm Fractures
Primary Data Analysis Descriptive statistics were used to report the characteristics of our subjects and test performance, including sensitivity, specificity, and positive and negative predictive values with 95% confidence intervals (CIs). The concordance of the treating physicians’ real-time assessment of ultrasonography and fluoroscopy was used to calculate the agreement rate. Data were analyzed with SPSS Statistics (version 22; IBM, Armonk, NY).
RESULTS Characteristics of Study Subjects One hundred patients were enrolled (Figure 1), with their clinical information presented in the Table. Participating study physicians (n¼16) were pediatric (n¼12) or general (n¼4) emergency physicians with a median number of years in independent practice of 7.5 (interquartile range 2.3 to 15 years) and among whom 69% reported performing 1 to 5 reductions per month; 31%, 6 to 10 per month. Main Results The ultrasonography performed by study physicians compared with the reference standard and agreement with their assessment of fluoroscopy is shown in Figure 2. Physicians performed a median number of 4 ultrasonographically guided reductions; 78% ranked its use as easy and 5% as difficult, and 78% indicated that ultrasonography was helpful with the reduction procedure. Test performance revealed a sensitivity and specificity of 50% (95% CI 15.4% to 84.6%) and 89.1% (95% CI 82.8% to 95.5%), respectively. The negative predictive value was excellent, at 95.3% (95% CI 90.9% to 99.8%), albeit the positive predictive value was 28.6% (95% CI 4.9% to 52.2%), indicating that ultrasonography was best at identifying when fractures were adequately realigned. As seen in Figure 2A, the results were similar in a sensitivity analysis in which subjects treated by the principal investigator (n¼30) were excluded. To assess the possible association of provider experience with ultrasonography, subanalysis was conducted in which physicians who performed less than or equal to 5 and less than or equal to 10 fracture reductions were compared with those who performed greater than 5 and greater than 10 fracture reductions (ie, “experienced”); more experienced providers were more specific at determining an adequately reduced fracture (Figure 2A). Furthermore, the top 2 physicians (ie, together performed nearly half of the reductions with ultrasonography) performed with perfect sensitivity compared with those who performed at less than 10, who performed with poor sensitivity. The agreement rate between real-time assessment of ultrasonography and fluoroscopy by the treating physician was 98%; one physician had 2 cases in which the ultrasonography was assessed as inadequate and fluoroscopy adequate, and there were no cases in which the treating physician’s ultrasonography was assessed as adequate but fluoroscopy inadequate (Figure 2B). Volume 65, no. 3 : March 2015
LIMITATIONS Our study’s main limitation was the inclusion of a convenience sample of patients enrolled when the research assistant and a study physician were present. A potential source of bias was introduced into the study because the principal investigator (A.S.D.) performed 30% of the ultrasonographically guided reductions. The results remained similar when these subjects were removed from the analysis. Although the ultrasonographic skill required to assess fracture reduction was easily learned, reproducibility of the test was not evaluated. All of the providers were novice at this ultrasonographic skill, and with a similar training program our results may be generalizable to other pediatric EDs. The small number of inadequately reduced fractures limited the precision with which the test characteristics could be determined. Another limitation is the subjective nature of the reference test; there are no definitive guidelines on what constitutes an age-appropriate adequate reduction, albeit the blinded pediatric orthopedic surgeon’s assessment of adequacy represented our standard of care from which follow-up decisions were made. Moreover, we obtained the fluoroscopic images immediately after the ultrasonography as our reference standard, not images obtained post–cast application, to directly compare adequacy of the ultrasonographic images before any further manipulation or cast application. It is possible that fracture alignment changed after application of the appropriate cast. Last, our results apply only to patients for whom a single-bone fracture requires realignment because we excluded those in whom both the radius and ulna required reduction.
DISCUSSION Our results indicate that point-of-care ultrasonography performed by emergency physicians can help determine when fractures have been successfully realigned, with a high negative predictive value. Given the increasing availability of ultrasonography in EDs, along with the simple and safe technique that allows real-time visualization, ultrasonographic confirmation of forearm fracture realignment should be encouraged in place of blind (ie, no real-time imaging) reductions and as a possible alternative to fluoroscopy. To our knowledge, this is the first pediatric study to assess the diagnostic accuracy of ultrasonography compared with fluoroscopy, as well as the first study in a pediatric ED that used a large group (n¼16) of novice providers. Chinnock et al7 reported the diagnostic accuracy of ultrasonography in an ED treating both children and adults; the test characteristics would be strikingly similar to ours had their analysis used a similar case definition (ie, positive result defined as inadequately realigned). One drawback of studies to date is the subjective reference standards used as opposed to patient-focused outcomes (for example, the need for reintervention). Pediatric forearm fractures remodel well despite residual malalignments. In our study, follow-up radiographs of 3 of the 4 false-negative cases revealed adequate fracture alignments in the appropriate casts and no need for reinterventions. Similarly, among fractures correctly identified as inadequately reduced, one had a residual 28-degree angulation, albeit active remodeling was observed in follow-up. Annals of Emergency Medicine 263
Ultrasonographically Guided Reduction of Pediatric Forearm Fractures
264 Annals of Emergency Medicine
Dubrovsky et al
Volume 65, no. 3 : March 2015
Dubrovsky et al
Ultrasonographically Guided Reduction of Pediatric Forearm Fractures
Figure 2. A, Diagnostic results for each point-of-care ultrasonographic study performed by the treating physician in the pediatric ED compared with the reference standard of the blinded pediatric orthopedic surgeon’s assessment of fluoroscopic images. B, Agreement between each of the treating physician’s own assessments of his or her ultrasonographic and fluoroscopic images for fracture realignments. Each block represents a unique patient enrolled per study physician (x axis), with the number of ultrasonographic scans performed by each physician presented from the first ultrasonography performed at the bottom and arranged vertically in chronologic order. Blocks are color coded as per the inset. A true positive and true negative were an inadequately and adequately realigned fracture, respectively, on both ultrasonography and fluoroscopy; a false positive was an inadequately realigned fracture on ultrasonography but adequate on fluoroscopy, whereas a false negative was an adequately realigned fracture on ultrasonography but inadequately aligned on fluoroscopy. *Test characteristics of ultrasonography versus blinded pediatric orthopedic surgeon’s assessment of fluoroscopy. †Test characteristics with principal investigator’s cases excluded. ‡ Test characteristics among physicians who performed less than or equal to 5 versus greater than 5 fracture reductions using ultrasonography. §Test characteristics amongst those who performed less than or equal to 10 versus greater than 10 fracture reductions using ultrasonography.
Although this study was not designed to assess the test performance among users of differing experience levels, a trend toward improved specificity among more experienced providers was noted. The sensitivity also greatly improved among the 2 providers with the greatest experience. Experienced providers might have been able to more precisely differentiate the periosteum (which can appear angulated) from the cortex, or use 3 views more consistently than 2, leading to a more accurate assessment of adequacy. Although future studies are needed to determine the optimal number of procedures to attain competency, it can be conjectured that accuracy may be maximized once physicians perform more than 10 reductions with ultrasonography. In summary, point-of-care ultrasonography can accurately assess the successful realignment of fractures of the forearm in children, but inadequate reductions should be confirmed by other imaging modalities. Ultrasonography provides clinicians the ability to safely guide fracture realignment without requiring the equipment or personnel needed for fluoroscopy, and it does not expose the patient or the staff to any radiation. Future randomized studies are needed to determine whether this approach improves procedure time, number of attempts or successful reductions, and patient and parental satisfaction, as well as whether it can aid in the reduction of fractures in which both bones need realignment. The authors acknowledge Rami Rezk, BSc, (research assistant) and Thierry Benaroch, MD, MSc, (from the Department of Pediatric Orthopedic Surgery) for their help and dedication to this project.
Supervising editor: Kelly D. Young, MD, MS Author affiliations: From the Division of Pediatric Emergency Medicine (Dubrovsky, Bank) and the Clinical Research Centre (Mok), Montreal Children’s Hospital-McGill University Health Center, Montreal, Quebec, Canada; the Faculty of Medicine (Dubrovsky, Bank), and Centre for Medical Education and Arnold and Blema Steinberg Simulation Centre (Bank), McGill University, Montreal, Quebec, Canada; and the Division of Pediatric Emergency Medicine, Hospital for Sick Children and Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada (Kempinska). Author contributions: ASD and AK conceived the study, designed the trial, and obtained research funding. ASD supervised the Volume 65, no. 3 : March 2015
conduct of the trial and data collection, as well as managed the data, including quality control. EM contributed to the design of the study and analysis of the data. ASD drafted the article, and all authors contributed substantially to its revision. ASD takes responsibility for the paper as a whole. Funding and support: By Annals policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article as per ICMJE conflict of interest guidelines (see www.icmje.org). The authors have stated that no such relationships exist. Funding for the research assistant was provided by the Department of Pediatric, Montreal Children’s Hospital through an internal research fellowship competition. The lead author is a member of the Research Institute of the McGill University Health Centre, which is supported by the Fonds de Recherche du Québec–Santé. Publication dates: Received for publication May 3, 2014. Revisions received July 2, 2014; July 28, 2014; and August 11, 2014. Accepted for publication August 26, 2014. Available online October 16, 2014. Presented at the Canadian Association of Emergency Physicians annual conference, May 2014, Ottawa, Canada; and the Canadian Pediatric Society annual conference, June 2014, Montreal, Canada.
REFERENCES 1. Chen L, Kim Y, Moore CL. Diagnosis and guided reduction of forearm fractures in children using bedside ultrasound. Pediatr Emerg Care. 2007;23:528-531. 2. Patel DD, Blumberg SM, Crain EF. The utility of bedside ultrasonography in identifying fractures and guiding fracture reduction in children. Pediatr Emerg Care. 2009;25:221-225. 3. Barata I, Spencer R, Suppiah A, et al. Emergency ultrasound in the detection of pediatric long-bone fractures. Pediatr Emerg Care. 2012;28:1154-1157. 4. Durston W, Swartzentruber R. Ultrasound guided reduction of pediatric forearm fractures in the ED. Am J Emerg Med. 2000;18: 72-77. 5. Wong CE, Ang AS, Ng KC. Ultrasound as an aid for reduction of paediatric forearm fractures. Int J Emerg Med. 2008;1:267-271. 6. Heiner JD, McArthur TJ. A simulation model for the ultrasound diagnosis of long-bone fractures. Simul Healthc. 2009;4: 228-231. 7. Chinnock B, Khaletskiy A, Kuo K, et al. Ultrasound-guided reduction of distal radius fractures. J Emerg Med. 2011;40:308-312.
Annals of Emergency Medicine 265
