Risk Factors for Anterior Glenohumeral Instability

OBJECTIVES: While anterior glenohumeral instability has been shown to be common in young athletes, the risk factors for injury are poorly understood. The elucidation of risk factors is critical to help develop prevention strategies. We hypothesized that specific modifiable and non-modifiable factors at baseline would be associated with the subsequent risk of injury in a cohort of young athletes. METHODS: We conducted a prospective cohort study in which 714 young athletes were followed from June 2006 through May 2010. Baseline assessments included a subjective history of instability, physical examination by a sports-trained orthopaedic surgeon, range-of-motion, strength with a hand-held dynamometer, and bilateral noncontrast shoulder MRI. A musculoskeletal radiologist measured glenoid version, glenoid height, glenoid width, glenoid index (height-to-width ratio), glenoid depth, rotator interval (RI) height, RI width, RI area, RI index, and the coracohumeral interval. Subjects were followed to document all acute anterior shoulder instability events during the 4 year follow-up period. The time to shoulder instability event during the follow-up period was the primary outcome of interest. Univariate and multivariable Cox proportional hazards regression models were used to analyze the data. RESULTS: We obtained complete data on 714 subjects. During our 4 year surveillance period, there were 38 anterior instability events documented. While controlling for covariates, significant risk factors of physical exam were: apprehension sign HR=2.96 (1.48, 5.90, p=0.002) and relocation sign HR=4.83 (1.75, 13.33, p=0.002). Baseline range-of-motion and strength measures were not associated with subsequent injury. Significant anatomic risk factors on MRI measurement were glenoid index HR=8.12 (1.07, 61.72) p=0.043 and the coracohumeral interval HR=1.20 (1.08, 1.34, p=0.001). CONCLUSION: This prospective cohort study revealed significant risk factors for shoulder instability in this high-risk population. While modifiable risk factors such as strength and range-of-motion were not associated with subsequent instability, some non-modifiable risk factors were. That the exam findings of apprehension and relocation were significant while controlling for prior history of injury suggests that patients may be unaware of prior instability episodes. The anatomic variables of significance are also not surprising - tall and thin glenoids were at higher risk compared to short and wide glenoids; and the risk of instability increased by 20% for every 1mm increase in coracohumeral distance.