Article, Orthopedics

Performance of the 4-way range of motion test for radiographic injuries after blunt elbow trauma

a b s t r a c t

Objectives: acute elbow injuries are common in the acute care setting. A previous study observed that limited ac- tive range of motion (ROM) was highly sensitive for radiographic injuries after blunt trauma. Our aim was to val- idate these findings in patients >=5 years old with an acute (b 24 hours) nonpenetrating elbow injury.

Methods: This prospective study included a convenience sample of patients undergoing plain radiographs of an injured elbow in 3 emergency departments. Before imaging, treating clinicians completed a standardized data collection sheet including mechanism of injury and 4-way ROM findings (full extension, flexion to 90?, full pro- nation and supination). Radiographic interpretation by a staff radiologist was used to ascertain the presence of fracture or joint effusion.

Results: The median age of the 251 patients was 24 years. Ninety-two patients (36.7%) had active 4-way ROM, and 159 patients (63.3%) demonstrated limited ROM. Negative radiographs were present in 152 patients (60.6%), whereas 99 patients (39.4%) had abnormal radiographs: 75 with explicit fractures and 24 with only joint effusions. The 4-way ROM elbow test had a sensitivity of 0.99 (95% confidence interval [CI], 0.94-1.00), spec- ificity of 0.60 (95% CI, 0.52-0.68), positive predictive value of 0.62 (95% CI, 0.54-0.69), and negative predictive

value of 0.99 (95% CI, 0.94-1.00).

Conclusions: Active 4-way ROM test is 99% sensitive for all radiographic injures following blunt elbow trauma and 100% sensitive for injuries requiring surgical intervention. Caution should be used in relying on this test in the pediatric population until it is validated in a larger cohort.

(C) 2015

Introduction

Acute traumatic elbow injuries are common in primary care and emergency medicine, accounting for approximately 1% of emergency visits [1]. Radiographs are often obtained to distinguish fracture from soft tissue injury because the management of each injury can be sub- stantially different. Yet the prevalence of acute radiographic abnormal- ities among patients undergoing plain radiographs of the elbow can be as low as 25% [2]. If a set of physical examination findings could be iden- tified to predict which patients were at Very low risk of fracture, many of

? Sources of support: none.

?? Presentation: presented at the American College of Emergency Physicians Research Forum, Boston, MA, October, 2015.

* Corresponding author at: Department of Emergency Medicine, Kaiser Permanente Sacramento Medical Center, 2025 Morse Ave, Sacramento, CA 95825.

E-mail address: [email protected] (D.R. Vinson).

these radiographs could be safely avoided. This could save the patient the inconvenience, cost, and radiation risk of unnecessary radiographic evaluation. Foregoing needless radiographs could also shorten emer- gency department (ED) length of stay and facilitate Patient throughput. Many have sought to derive a clinical prediction rule to guide the se- lective use of radiographs in patients with acute nonpenetrating elbow trauma [1-12]. Shortcomings of these studies include absence of radio- graphs in all patients at the time of the initial evaluation [1,2,11], failure to apply the predictive test to both children and adults [4,8,11], and fail-

ure to include a community setting [1-6,8-11].

We performed a prior study of patients with acute elbow injuries that included only those undergoing radiographs, enrolled both adults and children, and involved both academic and community facilities [7]. After examining several potential predictors, we noted that preser- vation of active 4-way range of motion (extension, flexion to at least 90?, full supination and pronation) was 100% sensitive in predicting ab- normal radiographs [7]. However, our cohort was only of modest size and the confidence intervals (CIs) were relatively wide (0.92-1.0). We undertook this current study to prospectively validate those findings.

http://dx.doi.org/10.1016/j.ajem.2015.10.031

0735-6757/(C) 2015

Methods

Study design and setting

This prospective study with convenience sampling was undertaken at 3 medical centers in the greater Sacramento region between August 2012 and January 2015. Two are suburban community medical centers within Kaiser Permanente Northern California, a large integrated Health care delivery system. The third facility is an urban academic Level I trau- ma center with an emergency medicine residency training program. The census of each of the 3 EDs during 2013 was between 75,000 and 85,000. The study was approved by the respective institutional review boards with waiver of consent in the community medical centers and written informed consent of the patient or parent/guardian in the aca- demic medical center.

Selection of participants

Patients 5 years and older with an acute (b 24 hours) nonpenetrating elbow injury were eligible for enrollment if the clinician felt that radio- graphs of the affected elbow were clinically indicated. No clinical policy or practice guideline was in place throughout the study period to specify which injuries warranted radiographs. This decision was entirely in the hands of the treating clinician. We excluded patients with a preexisting con- dition in the affected limb that limited mobility (eg, palsy), altered mental status, a significant distracting injury (in the clinician’s judgment), obvious elbow deformity, or neurovascular abnormality in the affected limb.

Study protocol and measurements

Before the commencement of the study, participating clinicians were instructed on how to perform the 4-way range-of-motion maneuvers and given explicit printed instructions on the standardized data collection instrument. Clinicians prospectively recorded the patient’s age, sex, and the mechanism of injury (free text). Before obtaining radiographs of the elbow, clinicians examined the patient in the seated or standing position with the injured arm at the patient’s side in the anatomical position, ex- tended at the elbow with the palm forward. We recorded the patient’s ability to actively extend to a full locked position (0?), to actively flex to at least 90?, and to actively pronate and supinate to full range of motion (180?) while flexed as close to 90? as possible. In individuals who were unable to flex to at least 90? or extend to locked position, a comparison between the degree of flexion and extension was made with the unin- jured arm. In such cases, inability to flex or extend equal to the unaffected limb was considered a positive (ie, abnormal) finding.

Any limitation of range of motion, in any of the 4 maneuvers, was considered a positive test result. Patients able to meet target end points in all 4 directions were considered to have preserved 4-way range of motion–a negative test result. A positive range-of-motion test result was our primary independent, or predictor, variable.

All enrolled individuals received at a minimum standard 3-view radio- graphs of the injured elbow, including lateral and anteroposterior views. The main outcome, or dependent measure, was a set of elbow radio- graphs with fracture or joint effusion (anterior sail sign, posterior fat pad sign, or both). Radiographs were interpreted by board-certified ra- diologists unaware of the study and blinded to the results of range-of- motion examination. Radiographic results were obtained through in- vestigator review of the Radiology reports. Individuals without obvious radiographic fracture but with the presence of a joint effusion were con-

sidered positive for occult fracture, a conservative assumption.

Association between physical examination and radiographic findings was performed to determine performance characteristics of our independent variable, impaired active 4-way range of motion, in

Table 1

age distribution of ED patients with acute nonpenetrating elbow injuries

Age range, y N = 251

n (%)

Pediatric (5-17 y) 111 (44.2)

5-12 69 (27.5)

13-17 42 (16.7)

Adult (>= 18 y) 140 (55.8)

18-64 116 (46.2)

65-89 24 (9.6)

Statistical analysis

Continuous variables are presented as medians with interquartile ranges, and categorical data are presented as the percentage of frequen- cy of occurrence (Microsoft Excel 2013 v. 15; Redmond, WA). Frequen- cies observed between groups were compared by means of Fisher exact test (GraphPad Software 2015; La Jolla, CA). Sensitivity and specificity of the 4-way range-of-motion test were calculated (Vassarstats.net; Poughkeepsie, NY). We included 95% exact CIs. A 2-tailed P value of less than .05 was considered to indicate statistical significance.

Results

A total of 251 individual patients were enrolled; 124 (49.4%) were fe- male. No patient was enrolled on more than 1 occasion. The age break- down is summarized in Table 1, with a median age of 24 years (interquartile range, 12-50 years; range, 5-89 years); nearly half of the co- hort was pediatric. The most common mechanism of injury was fall (n = 189; 75.3%); 20 patients (8.0%) injured their elbow in a motor vehicle ac- cident (automobile, or automobile vs. pedestrian, or motorcycle).

Ninety-two patients (36.7%) had preserved active 4-way range of motion on physical examination (a negative test result). One hundred fifty-nine patients (63.3%) demonstrated some limitation of active range of motion in 1 or more directions (a positive test result). One hun- dred fifty-two patients (60.6%) had negative radiographs. Ninety-nine patients (39.4% prevalence) had fracture or occult fracture on elbow ra- diographs. Of those, 75 had explicit fractures and 24 only had joint effu- sions. The relation between range-of-motion findings and radiographic findings is reported in Table 2. Patients with limited range of motion were more likely to have an elbow fracture or effusion than those with preserved range of motion (62% vs 1%; P b .001).

There were 45 patients with active range-of-motion limitations iso- lated to only 1 of 4 directions. Eighteen of these had abnormal radio- graphs (Table 2). Fractures and effusions were more common in patients with range-of-motion limitations in 2 or more directions that those with limitations isolated to 1 direction (70% vs 40%; P b .001).

The 4-way range-of-motion test has a 99% sensitivity and a 60% specificity. The full performance characteristics of the 4-way range-of- motion test are reported in Table 3.

One patient had a false-negative range-of-motion test result; that is, he had preserved active range of motion yet was found to have traumat- ic radiographic abnormalities. This 7-year-old boy had tripped and suf- fered a ground-level fall. Radiographs in the ED revealed a nondisplaced

Table 2

Results of elbow range-of-motion examination and radiographic findings (N = 251)

Fracture or effusion

4-Way range of motion

n

No

n (% of row)

Yes

n (% of row)

Complete (test result negative)

92

91 (99)

1 (1)

Limited (test result positive)

159

61 (38)

98 (62)

predicting abnormal radiographs (positive predictive value). We also

In only 1 directiona

45

27 (60)

18 (40)

determined the utility of full 4-way range of motion (test negative) to

In 2 or more directions

114

34 (30)

80 (70)

predict normal radiographs (negative predictive value).

a Distribution: extension (n = 35), flexion (n = 1), pronation (n = 8), supination (n = 1).

Table 3

Performance characteristics of the elbow 4-way range-of-motion test

95% CI

for patients with preserved range of motion, 40% for those with limita- tion in 1 isolated direction, and 70% for patients with impediments in 2 or more directions.

Sensitivity 0.99 0.94-1.00

Specificity 0.60 0.52-0.68

Positive predictive value 0.62 0.54-0.69

Negative predictive value 0.99 0.93-1.00

Positive likelihood ratio 2.47 2.03-3.00

Negative likelihood ratio 0.02 0.00-0.12

supracondylar fracture of the distal humerus. He was initially treated with splint immobilization and a sling. At follow-up with orthopedics 2 days later, he was placed in a long arm cast, which was removed at 4 weeks. He had full return to preinjury function.

Limitations

This study used convenience sampling, which could have introduced selection bias. Though the range-of-motion examination was undertak- en before administration of ED analgesia, patients may have taken anal- gesia before arrival. We did not record this variable nor report the degree of pain at the time of examination. Also, the sample was relative- ly small, which accounts for the moderately wide CIs around our point estimates. However, it would require a much larger multicenter study to achieve a substantially larger sample size. We failed to measure interrater reliability among our clinicians, although the ? for elbow ex- amination findings in other studies has been relatively high [1,10,11]. Informal rechecks of early cases by the investigators revealed no dis- agreements, perhaps due to the simple nature of the examination and the thorough instructions and training we used. Lastly, our population included patients from academic and Community settings in Northern California and may not represent other Geographic regions or practice settings.

Discussion

This prospective study found that an abnormal 4-way range-of- motion test result was highly sensitive for the detection of abnormal ra- diographs in patients with acute nonpenetrating elbow injuries. Con- versely, nearly all patients (99%) with full range of motion lacked signs of effusion or fracture on plain radiographs, and no patient with a negative range-of-motion test result required surgical intervention (100% negative predictive value). The single patient in our study with a false-negative test result had a minor nondisplaced fracture that responded well to conservative Nonoperative management.

We found that the prevalence of abnormal radiographs increased significantly as the degree of range-of-motion limitation increased: 1%

In our earlier multicenter prospective study of elbow injuries, we ob- served that preservation of active 4-way range of motion was 100% sen- sitive in detecting abnormal radiographs [7]. The CIs (0.92-1.00) of that study overlap with the intervals for this current study (0.93-1.00).

We did not include ultrasonographic or magnetic resonance imaging as methods of fracture detection, as these are not commonly used in the community emergency medicine setting. Research using these other imaging modalities, however, has demonstrated the nonspecificity of fat pad signs on plain radiographs as indicators of occult fracture [13-15]. We were being conservative by including radiographically de- tected joint effusions in our outcome measure of abnormal radiographs because many of these do not ultimately have a cortical fracture [16]. We are not the only ones to include effusions as a radiographic outcome. Nearly all other studies that have correlated examination findings with radiographic abnormalities have done the same [1-11].

Others have studied the correlation between active 4-way elbow range-of-motion and radiographic findings, with varying sensitivities (from 0.813 to 0.983) and specificities (from 0.248 to 0.963) (Table 4) [5,7-9]. These studies differed significantly in their clinical setting, the prevalence of radiographic abnormalities, and the educational training of the clinician examiners. The more complicating issue may be the wide age range of participants (Table 4). Studies with pediatric patients often find these clinical evaluations less sensitive and have reported lower negative predictive values [1,8,11]. Indeed, in our study, the only patient with an abnormal radiograph that had a negative elbow range-of-motion test result was 7 years old. Much larger or more fo- cused studies of pediatric patients may be needed to address the age- dependent performance variation of these tests.

None of the false-negative cases in these 4 published series required operative intervention [5,7-9]. These results concur with our own. But the treatment provided to the false-negative cases may well have averted operative intervention. For example, had our child with a supracondylar fracture failed to receive immobilization and close outpa- tient follow-up, his nondisplaced fracture may have displaced, in which case operative intervention could well have been required. This hypo- thetical situation reminds us that clinical decision rules are meant to serve as an adjunct to clinical judgment and not a replacement for it. Caution should be used in relying on this test in the pediatric population until it is validated in a larger population.

A test can serve as a reliable basis for universally foregoing radio- graphs only if it meets these stringent criteria: (1) a sensitivity of 99% or greater, (2) a lower 95% confidence limit not less than 97%, and

(3) minimal clinical consequences of a false-negative result [2,8]. If

that is indeed the threshold, no study of active 4-way elbow range of

Table 4

Prospective studies of preserved active 4-way range of motion to predict radiographic findings among ED patients with acute elbow injurya

First author (y)

Setting

(no. of facilities)

Ages (y)

Abnormal radiographs/ cases n/N

(% prevalence)

Clinicians

Radiographs at index visit

n (%)

Outcome

Sensitivity (95% CI)b

Specificity (95% CI)b

Lennon

Academic (1)

2-96

178/335 (53.1)

nurse practitioners,

335 (100.0)

Fracture, dislocation

0.983 (0.948-0.996)

0.248 (0.185-0.325)

(2007) [5]

residents, and

or effusion at

Darracqc

Academic (1),

5-60

53/113 (46.9)

faculty

Residents and

113 (100.0)

index visit

Fracture or effusion

1.00 (0.916-1.0)

0.967 (0.875-0.994)

(2008) [7]

community (2)

faculty

at index visit

Baker

Academic (1)

3-16

106/177 (59.9)

Residents and

177 (100.0)

Fracture, dislocation,

0.934 (0.864-0.971)

0.338 (0.233-0.461)

(2011) [8]

faculty

or effusion at

index visit

Amiri

Academic (2)

5-87

15/102 (14.7)

Residents

102 (100.0)

Fracture or effusion

0.867 (0.584-0.977)

0.920 (0.836-0.964)

(2012) [9]

at index visit

a Some data included here were missing from the sources and were obtained from the corresponding authors.

b For uniformity sake, we calculated all CIs using the Wilson procedure with a correction for continuity.

c Tenderness at 1 of 4 points was included in the study but did not improve sensitivity when added to 4-way range of motion alone, which is why we present this study here as a range- of-motion study.

Table 5

Prospective studies of full active extension and combination tests to predict radiographic findings among ED patients with acute elbow injurya

First author (y)

Setting

(no. of facilities)

Ages (y)

Abnormal radiographs/ cases n/N

(% prevalence)

Clinicians

Radiographs at index visit n (%)

Outcome

Sensitivity (95% CI)b

Specificity (95% CI)b

Test: full active extension

Hawksworth

Not reported (1)

1-85

54/100 (54.0)

Triage nurses

100 (100.0)

Fracture or effusion

0.907 (0.789-0.965)

0.587 (0.433-0.727)

(1991) [3]

at index visit

Docherty

Academic (1)

16-76

54/110 (49.1)

Residents and

110 (100.0)

Fracture, dislocation

0.974 (0.846-0.999)

0.694 (0.573-0.795)

(2002) [4]

faculty

or effusion at

index visit

Lamprakis

Community (1)

10-75

24/70 (34.3)

Orthopedic

70 (100.0)

Fracture, dislocation

0.917 (0.715-0.985)

0.609 (0.454-0.745)

(2007) [6]

surgical faculty

or effusion at index

visit and evidence

of bony healing

Appelboam

Academic (5)

3-94

538/1740 (30.9)

Nurse practitioners

1319 (75.8)c

at 2 wk

Fracture or effusion

0.968 (0.949-0.981)

0.464 (0.436-0.492)

(2008) [2]

and physicians

at index visit or

at 7-10 d

Test: combination of full active extension with palpation of tenderness at one of several points (+-bruising)

Arundel

Academic (1)

16-96d

114/348 (32.8)

Nurse practitioners,

Not reported

Fracture or effusion

Not reported

Not reported

(2014) [1]

residents and

at index visit or

physicians

at 7-10 d

Jie

Academic (2)

3-92

348/587 (59.3)

Nurse practitioners,

587 (100.0)

Fracture or effusion

0.983 (0.961-0.993)

0.075 (0.046-0.118)

(2014) [10]

residents and

at index visit

faculty

Dubrovsky

Academic (1)

0-18

183/332 (55.1)

Senior medical

324 (97.6)

Fracture or effusion

0.961 (0.920-0.983)

0.174 (0.119-0.247)

(2014) [11]

residents and faculty

at index visit or

at 7-10 d

a Some data included here were missing from the sources and were obtained from the corresponding authors.

b For uniformity sake, we calculated all CIs using the Wilson procedure with a correction for continuity.

c Adults with full extension were by protocol not to undergo radiographs at index visit but were to be reassessed at 7-10 days by telephone call, and those who met Explicit criteria were then asked to return for radiographs. Children with full extension at index visit could receive radiographs at the discretion of the treating clinician.

d The larger study included Younger children, as well. We report here only the data from the adult substudy because their test had high sensitivity only in the adult population.

motion fully meets the mark, partly because of sample size constraints driv- ing wide CIs (Table 4) [5,7-9]. The same is true for studies that used exten- sion only as their predictive test (Table 5) [2-4,6]. Likewise, most studies that combined extension with points of tenderness were also insufficiently sensitive (Table 5) [10,11]. A recent single-center British study may have hit that mark if its results can be validated [1]. They derived a novel predictive rule, the East Riding Elbow Rule, among children and adults with acute elbow injuries. Their unique combination approach (extension plus tender- ness at 1 of 3 points and the presence of bruising) was found to be 100% sensitive (95% CI, 97%-100%) among their 348 adults but did not perform as well among children [1]. However, the achievement in sensitivity came at a significant cost in specificity: 24% (95% CI, 19%-30%).

The reality is that application of all tests in clinical decision rules involves thresholds that are a trade-off between sensitivity and specificity. And not all radiographic studies are of equal importance clinically. Although a sensi- tivity of 99% may be necessary for ruling out Cervical spine injuries, detect- ing elbow injuries that do not require active intervention may allow a slightly more relaxed threshold. Otherwise, it is unlikely that any clinical test could ever meet such a high sensitivity threshold without sacrificing specificity to the point of uselessness. Rather than expecting these predic- tive tests to provide reliable binary outcomes (obtain or forego radio- graphs), one can integrate the information they provide in the context of pretest probabilities. Although a negative active 4-way range-of-motion test result may not absolutely exclude an abnormal elbow radiograph, espe- cially in children, it does substantially lower the probability of such a find- ing. Preservation of complete active range of motion may still influence the decision making for radiographs, especially when radiography is not readily available and close follow-up is ensured. Moreover, the cumulative weight of evidence suggests that patients with preserved active range of motion are unlikely to require operative intervention, although the role of immobilization in these cases has not been elucidated.

The application of the elbow range-of-motion test in our clinical set- ting can be illustrated as follows, assuming a prevalence of abnormal ra- diographs of 40% and performance of the elbow range-of-motion test as

we described (Table 3). Applying the test to 250 patients who would have previously undergone plain radiographs, 92 (37%) would test neg- ative. If radiographs were foregone (or deferred) in this group, 1 patient with a radiographic abnormality would have been initially missed. Using the 4-way range-of-motion test as the basis for selected radio- graphic ordering would need to be linked to close outpatient follow- up for patients who failed to improve in short order.

Conclusions

In summary, this is one of the largest prospective studies to explicitly evaluate the performance of a 4-way range-of-motion test to rule out sig- nificant injury of the elbow after blunt trauma. We found that the test had very high (99%) sensitivity for radiographic injuries and 100% negative predictive value for excluding injuries requiring surgical intervention. It appears that the elbow range-of-motion test provides a useful combina- tion of high sensitivity balanced with reasonable specificity to help guide radiograph ordering in cases of blunt elbow trauma. The range-of- motion test, however, may not perform as well in pediatric patients.

Acknowledgments

We would like to thank our ED clinicians who took the time to enroll patients amidst their pressing clinical care duties.

References

  1. Arundel D, Williams P, Townend W. Deriving the East Riding Elbow Rule (ER2): a maximally sensitive Decision tool for elbow injury. Emerg Med J 2014;31:380-3.
  2. Appelboam A, Reuben AD, Benger JR, Beech F, Dutson J, Haig S, et al. Elbow extension test to rule out elbow fracture: multicentre, prospective validation and observational study of diagnostic accuracy in adults and children. BMJ 2008;337:a2428.
  3. Hawksworth CR, Freeland P. Inability to fully extend the injured elbow: an indicator of significant injury. Arch Emerg Med 1991;8:253-6.
  4. Docherty MA, Schwab RA, Ma OJ. Can elbow extension be used as a test of clinically

    significant injury? South Med J 2002;95:539-41.

    Lennon RI, Riyat MS, Hilliam R, Anathkrishnan G, Alderson G. Can a normal range of elbow movement predict a normal elbow x ray? Emerg Med J 2007;24:86-8.

  5. Lamprakis A, Vlasis K, Siampou E, Grammatikopoulos I, Lionis C. Can elbow- extension test be used as an alternative to radiographs in primary care? Eur J Gen Pract 2007;13:221-4.
  6. Darracq MA, Vinson DR, Panacek EA. Preservation of active range of motion after acute elbow trauma predicts absence of elbow fracture. Am J Emerg Med 2008;26:779-82.
  7. Baker M, Borland M. Range of elbow movement as a predictor of bony injury in chil- dren. Emerg Med J 2011;28:666-9.
  8. Amiri H, Shams Vahdati S, Fekri S, Zadegan SA, Shokoohi H, Rahimi-Movaghar V. Does preservation of active range of motion after acute elbow injury rule out the need for radiography? Ulus Travma Acil Cerrahi Derg 2012;18:479-82.
  9. Jie KE, van Dam LF, Verhagen TF, Hammacher ER. Extension test and ossal point ten- derness cannot accurately exclude significant injury in acute elbow trauma. Ann Emerg Med 2014;64:74-8.
  10. Dubrovsky AS, Mok E, Lau SY, Al Humaidan M. point tenderness at 1 of 5 locations and limited elbow extension identify significant injury in children with acute elbow trauma: a study of diagnostic accuracy. Am J Emerg Med 2015;33:229-33.
  11. Joshi N, Lira A, Mehta N, Paladino L, Sinert R. Diagnostic accuracy of history, physical examination, and bedside ultrasound for diagnosis of extremity fractures in the emergency department: a systematic review. Acad Emerg Med 2013;20:1-15.
  12. Jie KE, van Dam LF, Hammacher ER. Isolated fat pad sign in acute elbow injury: is it clinically relevant? Eur J Emerg Med 2015 [Epub ahead of print].
  13. Al-Aubaidi Z, Torfing T. The role of fat pad sign in diagnosing occult elbow fractures in the pediatric patient: a prospective magnetic resonance Imaging study. J Pediatr Orthop B 2012;21:514-9.
  14. Zuazo I, Bonnefoy O, Tauzin C, Borocco A, Lippa A, Legrand M, et al. Acute elbow trauma in children: role of ultrasonography. Pediatr Radiol 2008;38:982-8.
  15. Donnelly LF, Klostermeier TT, Klosterman LA. Traumatic elbow effusions in pediatric patients: are Occult fractures the rule? AJR Am J Roentgenol 1998;171:243-5.