Can elbow-extension test be used as an alternative to radiographs in primary care?

Abstract

Objective: To evaluate the ability of elbow extension, with the patient in a supine position, as a diagnostic test of an insignificant injury, with the purpose of avoiding unnecessary radiographs. Methods: Seventy patients suffering from an acute elbow injury were examined at the accident and emergency department. Inability to fully actively extend the elbow in a supine position was defined as a positive diagnostic test. Radiographs were interpreted by a consultant radiologist, blinded to all clinical examination results. Sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios along with their 95% confidence intervals were calculated for the elbow-extension test. Results: Forty out of 70 patients had a positive test. Elbow fracture or dislocation was identified radiographically in 22 patients with positive test (sensitivity 92%). Two out of 30 (with negative test) had a hairline radial head fracture, which was found on radiographs (specificity 61%). Conclusion: Elbow extension as a diagnostic test in a primary care setting can predict severe elbow injuries and can be safely used in practices with no radiology facilities.

• Elbow extension
• elbow-extension test
• elbow injury
• general practice
• radiographs

Introduction

In recent years there has been much effort to produce guidelines to decrease excessive use of radiological procedures, which are time consuming, expensive, and harmful to patients, especially children, including the effects of ionizing radiation [1]. Consequently, international organizations and institutions have issued concrete guidelines on the use of conventional radiology for diagnostic purposes, such as the Ottawa Ankle and Knee Rules and the Canadian Cervical Spine Rules [2–4].

Elbow injuries are commonly seen in the accident and emergency (A&E) department, and there is ongoing debate concerning the effectiveness of conventional radiology in outpatient departments. McGinley et al. suggested that non-displaced fractures occur commonly following impact injuries and are difficult to diagnose with plain film imaging [5]. The posterior fat pad sign, first described by Norell in 1954, is strongly associated with a high incidence of occult fractures [6], although some studies have disputed this association [7]. Although magnetic resonance imaging (MRI) reveals a variety of bone and soft-tissue injuries not detectable on plain radiographs, this additional information has little bearing on treatment or clinical outcome [8]. Furthermore, MRI is an expensive and not always available facility, with known difficulties in performing the procedure on children. According to the literature, patients able to fully extend the elbow while in a supine position are highly unlikely to have significant injury that needs radiological control [9].

This discussion has strong relevance to the current situation in Greece where a conventional radiology department is not available in all rural primary healthcare centres, hundreds of which serve remote areas without any diagnostic facilities. Under these circumstances, general practitioners/family physicians working in remote areas need to distinguish between cases that are not indicated for further observation and those that should be referred to hospital care for further investigation and treatment. This study explores whether a simple clinical test in patients with acute elbow injury is an accurate indicator of elbow injury for use in the primary care setting.

Materials and methods

Setting

The study was conducted at the A&E department of a district general hospital in Greece from November 2005 to April 2006. Seventy patients (nine of whom were <16 years old) with a mean age of 36 years (from 10 to 75 years), suffering from an acute elbow injury of less than 6 hours were included in the study. Mechanisms of injury included mainly road traffic accidents (RTAs), falls, and sport injuries. Severely injured patients and patients with known musculoskeletal disorders were excluded.

Sample size estimation

Estimation was based on the desired likelihood ratio (LR) confidence interval. Sensitivity of the diagnostic test was assumed to be at least 80%, specificity 73%, and positive likelihood ratio 2.96 [10]. The following formula, which represents the likelihood ratio of the 95% confidence interval calculated for any level of test result x, was used when calculating the sample size estimation:

where LRx=LR+, p1 = sensitivity, p2 = 1 – specificity, n1 = number of required patients with fracture, and n2 = number of patients without fracture. The equation was solved for n1 = n2, and it was found that n = 73.4.

Diagnostic tools and methods

The examining orthopaedic surgeon performed a clinical evaluation according to established standards [11]. The mechanism of injury was assessed, and important bony landmarks, including the radial head, were palpated. The range of movement for flexion, extension, supination, and pronation was estimated. Finally, an elbow-extension test and a comparison examination on the unaffected side were performed. Inability to fully extend the injured elbow when in a supine position was regarded as a positive sign [12]. Radiological evaluation of the elbow was performed on all patients, and included anteroposterior and lateral views. Radiographs were interpreted by the same consultant radiologist for all patients, blinded to all clinical findings. All patients had a follow-up with conventional radiography at 2 weeks to ascertain the presence of periosteal reaction or sclerosis, indicating a missed fracture.

Ethics

The study was approved by the ethics committee of the General Panarcadic Hospital of Tripolis, Greece. All patients gave informed consent for participation in the survey.

Analysis of data

To estimate the accuracy of the diagnostic test, sensitivity (the proportion of patients with the disease that have a positive test result, out of the total number of patients with the disease) and specificity (the proportion of healthy patients that have a negative test result, out of the total number of patients not having the disease) were measured. Positive and negative predictive values as well as positive (LH + ) and negative likelihood (LH − ) ratios along with their 95% confidence intervals were calculated for the elbow-extension test [13]. The formula for measuring LH+ was sensitivity/(1 – specificity), while LH− was (1 – sensitivity)/specificity.

Results

Forty out of seventy patients had a positive test. An obvious fracture or dislocation was found in 22 out of 40 patients with a positive test (55%). From the remaining 30 patients with a negative test, only two had actually sustained a fracture. Radiographs showed an undisplaced hairline radial head fracture in both cases. Sensitivity was found to be 92% and specificity 61% (Table I). The negative predictive value (NPV) was 93% (95% CI 83–98.6%), which implies the sensitivity of excluding a fracture, while the positive predictive value (PPV) was 55% (95% CI 50–75.6%). LH+ was 2.36 (95% CI 1.73–3.9), and LH− 0.13 (95% CI 0.32–0.06).

Table I.  Results of the use of elbow extension as a diagnostic test.

	Fracture or dislocation
	Present	Absent	Totals
Diagnostic test result
Positive	22	18	40
Negative	2	28	30
Totals	24	46	70

Sensitivity = 92% (95% CI 83–95.6), specificity = 61% (95% CI 52–75.6), PPV = 55% (95% CI 50–75.6), NPV = 93% (95% CI 83–98.6), LH + = 2.36 (95% CI 1.73–3.9), LH − = 0.13 (95% CI 0.32–0.06).

There were nine patients aged less than 16 years. Five had a positive test, and two of them had a fracture. None of the four patients with a negative test sustained an obvious fracture. Subgroup analysis of the children showed a sensitivity of 100% and a specificity of 57% (Table II).

Table II.  Results of the use of elbow extension as a diagnostic test in the subgroup of children.

	Fracture or dislocation
	Present	Absent	Totals
Diagnostic test result
Positive	2	3	5
Negative	0	4	4
Totals	2	7	9

Sensitivity = 100%, specificity = 57%.

Discussion

According to the new European definition, general practice/family medicine is an academic and scientific discipline. It supports the efficient use of healthcare resources through coordinated care, and manages simultaneously both acute and chronic health problems [14].

This study is highly relevant to the core characteristics of general practice/family medicine by attempting to measure the diagnostic accuracy of a simple and inexpensive diagnostic tool. According to our survey results, the elbow-extension test reached a high sensitivity, being in accordance with other published reports [9], [12]. Docherty et al. reported a sensitivity of 97.3% and specificity of 69.4%, while Hawksworth & Freeland's findings were 90.7% and 69.5%, respectively. A rather interesting study by Lennon et al. concluded that emergent elbow radiographs are not required in patients with normal extension, supination, and flexion [15].

It is the authors’ suggestion that patients with a positive test should definitely be referred to hospital for radiological assessment and appropriate treatment. Additional views, ultrasonography, CT, or MRI scans are some examples of helpful diagnostic adjuncts in hospitals, which primary care in rural areas in Greece cannot provide.

As far as patients with a negative test are concerned, they can be managed with a broad arm sling and analgesia as required. A follow-up appointment for clinical assessment should be arranged at 1 week. If symptoms have subsided, patients can be safely discharged. In case symptoms persist, referral to hospital can be made for radiological control; some of these patients may indeed have a fracture, but not a displaced or comminuted one. The type of fracture will need simple conservative measures and, as a result, radiographs will not differentiate initial treatment or outcome.

The present study demonstrates that the test failed to suggest a fracture in two cases which actually involved the radial head. Nevertheless, these fractures were Mason type I and could be safely treated conservatively with a broad arm sling. Non-displaced or minimally displaced radial head fractures with stable configuration and no bony block to motion can be treated with early mobilization [16]. For diagnostic and medicolegal reasons, a more careful clinical examination by eliciting tenderness while palpating the radial head may possibly reduce this incidence.

The elbow-extension test is easily performed by A&E doctors or general practitioners serving in the front line of remote rural areas with no radiology facilities. This test facilitates the screening of patients that should be referred to a hospital, reducing unnecessary radiographs by 40% with significant potential savings. The test can save valuable time during busy on-call periods and is cost effective, while protecting patients, especially children, from unnecessary harmful radiation. It is an A&E doctor's or general practitioner's confidence that may convince the patient, especially when waiting time in a radiology department can be avoided.

This study comes at a time when the debate on the effectiveness of general practice and primary care reform is active in Greece, and when concrete proposals have been published in the literature [17], [18]. However, there are certain concerns regarding the setting and population of the study. The setting in an A&E department and not a rural practice may have an impact on the target population that might differ significantly from the studied population. Sample size estimation was based on the assumption that the number of patients with fracture was equal to the number of those without; this was not achieved in our study. Safer results may be obtained by a larger population of 120 subjects, when the ratio of non-diseased to diseased subjects is 2.0. In this way, a type II error can be avoided.

Conclusions

In conclusion, despite the above-mentioned methodological constraints, the ability to fully extend the elbow while in a supine position is highly indicative of insignificant injury. Patients with a negative test can be treated conservatively with 1-week follow-up, while those with a positive test should be referred to hospital for radiological control. The inclusion of children in further studies could provide a population where the elbow extension test is more sensitive and leads to safer results.

Acknowledgements

The authors thank Mr A. Doulalas for his contribution to the statistical analysis, Mr A. Ziagos for radiological interpretation, Mr A. P. Fortis, the Clinical Director of the Second Orthopaedic Department, for permission and guidance during the data collection and, finally, Mr K. Lamprakis for technical assistance. Conflict of interest: The authors declare that they have no competing interests. Authors’ contributions: A.L. conceived the study design, contributed in data collection, carried out the analysis, and co-wrote the final manuscript; K.V. conceived the study design, contributed in data collection, carried out the analysis, and co-wrote the final manuscript; E.S. participated in data collection and interpretation; I.G. contributed in data interpretation and co-wrote the final manuscript; C.L. contributed to data analysis, formed the layout of the manuscript, wrote the final draft, and contributed substantially in re-drafting the manuscript. All authors approved the final manuscript.