Who will redislocate his/her shoulder? Predicting recurrent instability following a first traumatic anterior shoulder dislocation

Key points

1. 0136% (46/128) of participants had recurrent shoulder instability within 12 months
2. 02Six independent predictors were identified: bony Bankart lesion, age 16-25, dominant shoulder affected, no sling immobilisation, higher pain and disability (SPADI), and higher fear of reinjury (TSK-11)
3. 03Being immobilised in a sling was protective against recurrence (OR 0.28, 95% CI 0.09 to 0.86)
4. 04Bony Bankart lesion carried the highest individual risk (OR 6.04 in multivariate model)
5. 05This was the first prospective study to link fear of reinjury and self-reported pain with actual recurrence rates

How it was conducted

Design

Prospective cohort study

Participants

128 adults aged 16-40 years with a first-time traumatic anterior shoulder dislocation in New Zealand, recruited via Accident Compensation Corporation records May 2015 to February 2016

Follow-up

Telephone follow-up at 3, 6, 9, and 12 months, by research assistants blinded to baseline variables

Primary outcome

Recurrent shoulder instability (subluxation or dislocation) within 12 months

Analysis

Univariate logistic regression followed by multivariate backwards stepwise logistic regression to develop predictive equation

Baseline variables

Demographics, radiology findings (bony Bankart, Hill-Sachs, greater tuberosity fracture), immobilisation, physiotherapy, and validated questionnaires (SPADI, SAS, TSK-11, WOSI)

What they found

• Recurrence rate at 12 months: 35.9% (46/128); 50.0% (64/128) had no recurrence; 14.1% (18/128) lost to follow-up
• Bony Bankart lesion: univariate OR 3.65 (95% CI 1.05 to 12.70, p=0.04); multivariate OR 6.04 (95% CI 1.40 to 26.06, p=0.016)
• Age 16-25 vs other: multivariate OR 2.89 (95% CI 1.12 to 7.44, p=0.028)
• Dominant shoulder affected: multivariate OR 2.23 (95% CI 0.92 to 5.42, p=0.077)
• Immobilisation in sling: multivariate OR 0.28 (95% CI 0.09 to 0.86, p=0.026), i.e., protective
• SPADI total score: multivariate OR 1.034 per point (95% CI 1.003 to 1.066, p=0.031)
• TSK-11 fear of reinjury score: multivariate OR 1.14 per point (95% CI 0.994 to 1.313, p=0.061)
• Multivariate predictive equation: Risk = -4.73 + 1.06*(age 16-25) + 1.80*(bony Bankart) + 0.80*(dominant side) - 1.27*(immobilised) + 0.03*(SPADI total) + 0.13*(TSK-11 total)
• Hill-Sachs lesion was not significantly associated with recurrence: OR 1.45 (95% CI 0.63 to 3.35, p=0.38)
• Gender was not significantly associated with recurrence in multivariate analysis
• Highest recurrence rates were in contact/collision sports: rugby union, football, towed water sports, and skateboarding all had recurrence rates of 50% or above

Limitations

• Recurrent instability outcome was self-reported and not confirmed by radiograph, which may have introduced reporting bias and inflated psychosocial variable correlations
• The study was not sufficiently powered to examine some subgroup associations, such as greater tuberosity fractures and limb dominance effects independently
• Radiological reporting of bony lesion size and characteristics was inconsistent across radiologists, preventing more detailed analysis of Hill-Sachs and bony Bankart lesion severity
• The predictive tool has not yet been validated in an independent population, so its true predictive accuracy is unknown

Why it matters

For patients

If you are aged 16-25, had a bony Bankart fracture on imaging, dislocated your dominant shoulder, were not immobilised in a sling, and have high pain levels or fear of reinjury, you are at substantially higher risk of dislocating again and should discuss this with your clinician.

For clinicians

This six-factor multivariate tool enables individualised risk stratification at first presentation, supporting shared decision-making about conservative versus surgical management after a first traumatic anterior shoulder dislocation.

For readers

This is the first prospective study to confirm that psychosocial factors such as fear of reinjury and self-reported pain independently predict actual recurrence, broadening the picture beyond purely anatomical risk factors.