Diagnostic values of tests for acromioclavicular joint pain.

Background: This prospective study was performed to determine which clinical and imaging tests were most helpful for diagnosing acromioclavicular joint pain.

Methods: Of 1037 patients with shoulder pain, 113 who mapped pain within an area bounded by the midpart of the clavicle and the deltoid insertion were eligible for inclusion in the study. Forty-two subjects agreed to participate, and four of them were lost to follow-up. Twenty clinical tests, radiography, bone-scanning, magnetic resonance imaging, and an acromioclavicular joint injection test were performed on all patients. The patients were divided into two groups according to whether they had a > or =50% decrease in pain following the acromioclavicular joint injection. Statistical analysis, including multivariate regression analysis, was performed in order to evaluate the diagnostic effectiveness of the various tests.

Results: Acromioclavicular joint pain was confirmed in twenty-eight of the thirty-eight patients. The most sensitive tests were examination for acromioclavicular tenderness (96% sensitivity), the Paxinos test (79%), magnetic resonance imaging (85%), and bone-scanning (82%), but these studies had low specificity. In the stepwise regression model, with the response to the injection used as the dependent variable, bone-scanning and the Paxinos test were the only independent variables retained. Patients with a positive Paxinos test as well as a positive bone scan had high post-test odds (55:1) and a 99% post-test probability of having pain due to pathological changes in the acromioclavicular joint. The likelihood ratio for patients with one negative test and one positive test was indeterminate (0.4:1). Patients with both a negative Paxinos test and a negative bone scan had a likelihood ratio of 0.03:1 for having acromioclavicular joint pain, which basically rules out the disorder.

Conclusions: The highly sensitive tests had low specificity, and the highly specific tests had low sensitivity. However, the combination of a positive Paxinos test and a positive bone scan predicted damage to the acromioclavicular joint as the cause of shoulder pain with a high degree of confidence.