Physical examination tests in the acute phase of shoulder injuries with negative radiographs: a diagnostic accuracy study

Key points

1. 01Inability to abduct above 90 degrees had sensitivity 84%, specificity 71%, and diagnostic odds ratio 12.9 for detecting full-thickness tears
2. 02Small finger test for external rotation strength had specificity 86% and diagnostic odds ratio 12.4
3. 03Combining inability to abduct above 90 degrees with external rotation weakness raised sensitivity to 91-93% and diagnostic odds ratio above 22
4. 0438% of 120 patients with acute shoulder injury and negative x-rays had a full-thickness rotator cuff tear or occult avulsion fracture
5. 05Tests validated by non-specialist physicians in an emergency and primary care setting, increasing real-world applicability

How it was conducted

Design

Prospective phase III diagnostic accuracy study (STARD-compliant)

Setting

Combined primary care walk-in clinic and secondary care orthopaedic emergency department, Oslo University Hospital

Participants

120 consecutive patients aged 40 years or older with acute shoulder injury and negative plain x-rays, enrolled within 21 days of injury

Index tests

13 physical examination tests performed by non-specialist physicians blinded to imaging results

Reference standard

Ultrasound (with MRI in 44% of cases for confirmation or where ultrasound was indeterminate), blinded to physical examination results

Target condition

Acute full-thickness rotator cuff tear and/or occult avulsion fracture at tendon insertion

What they found

• Prevalence of target condition (full-thickness tear and/or occult fracture): 38% (n=46; 38 tears, 8 occult avulsion fractures)
• Inability to abduct above 90 degrees: sensitivity 84% (95% CI 69-93), specificity 71% (95% CI 59-82), DOR 12.9 (95% CI 4.8-34.2)
• Resisted abduction pain: sensitivity 91% (95% CI 78-97), specificity 55% (95% CI 43-66), DOR 12.1 (95% CI 3.9-37.4)
• Abduction strength test: sensitivity 74% (95% CI 59-86), specificity 77% (95% CI 65-86), DOR 9.6 (95% CI 4.0-23.0)
• Small finger test (external rotation strength) for supraspinatus: sensitivity 66% (95% CI 51-80), specificity 86% (95% CI 77-93), DOR 12.4 (95% CI 5.0-30.8)
• External rotation conventional strength test for supraspinatus: sensitivity 70% (95% CI 55-83), specificity 79% (95% CI 68-88), DOR 9.2 (95% CI 3.9-21.8)
• Combination of inability to abduct above 90 degrees and small finger test: sensitivity 91% (95% CI 78-97), specificity 69% (95% CI 57-79), DOR 22.2 (95% CI 7.1-69.3)
• Combination of inability to abduct above 90 degrees and external rotation strength: sensitivity 93% (95% CI 81-99), specificity 66% (95% CI 54-77), DOR 26.8 (95% CI 7.5-95.1)
• External rotation lag sign: sensitivity 16% (95% CI 5-33) and specificity 97% (95% CI 89-100) for supraspinatus tears; too low sensitivity to recommend for screening
• ROC curve AUC for reduced abduction predicting supraspinatus tear: 0.83
• Median age of patients with rotator cuff tear was 67 years vs. 51 years in those without (p<0.001)

Limitations

• Results apply only to patients aged 40 years or older seen within 3 weeks of injury; generalizability to other populations or time windows is unknown
• More than half of eligible patients were excluded, mainly due to pre-existing neck or shoulder pain, limiting external validity
• Confidence intervals for test accuracy metrics were wide, suggesting a larger sample size would have been preferable
• Subscapularis tear evaluation was not possible due to low case numbers (only 2 isolated subscapularis tears)

Why it matters

For patients

Patients who have had a shoulder injury but normal x-rays can be more accurately assessed for hidden tendon tears using two simple arm movement tests, helping ensure they get imaging or appropriate follow-up when needed.

For clinicians

Non-specialist clinicians in emergency or primary care settings can use inability to abduct above 90 degrees combined with external rotation weakness as a reliable screening tool to select patients for ultrasound or MRI after acute shoulder injury with negative x-rays.

For readers

This study provides the first diagnostic accuracy data for physical examination tests in non-referred, acutely injured shoulder patients, with ultrasound as blinded reference standard and good methodological quality, filling a gap noted in prior systematic reviews.