It’s more than size that matters: The role of glenoid concavity in shoulder instability with anterior bone loss

OBJECTIVES: The mechanism of concavity-compression is known to be a key factor for glenohumeral stability in the mid-range of motion. This stabilizing effect is impaired by traumatic bone loss at the anterior glenoid rim. Currently, a critical threshold based on the defect size is used as a decisive criterion for surgical treatment. However, recent studies using finite element method (FEM)-simulations indicate that glenoid concavity is essential for an assessment of remaining glenohumeral stability. To date, there is no biomechanical investigation involving glenoid concavity in combination with defect size. In this biomechanical study we focused on the interdependence between glenoid concavity, defect size and glenohumeral stability. We hypothesized that glenohumeral stability is mainly dependent on concavity and that the initial concavity affects the loss of stability caused by bony defects at the anterior glenoid rim. METHODS: A 6-degree-of-freedom industrial robot was utilized to determine the stability of 17 human cadaveric glenoids, depending on osteochondral concavity and anterior defect size. Load-and-shift tests were performed with artificial humeri equipped with a best-fit implant while joint positions and loads were captured. The Stability Ratio (SR), defined as the maximum tolerated anterior force related to a constant compression force, was determined for a compression of 50 N. In addition to a translation in 3 o’clock direction relative to a right scapula, a passive path dislocation was performed using compensatory translations to minimize superoinferior forces occurring during anterior translation. Defects were created in 2 mm steps parallel to the long axis of the glenoid until dislocation occurred self-acting and a 3D measuring arm was used for morphometric measurements as depicted in Figure 1. For statistical analysis, linear mixed-effects models were established to exploit the impacts of fixed effects (defect size and concavity gradient) as well as random effects (repeated measures and friction) on the SR. The influence of defect size on SR was analyzed for a translation in 3 o’clock by classifying the specimens into three groups of low (<25 %, n = 6), medium (25-35 %, n = 6) and high (>35 %, n = 5) initial concavity gradients. In addition, the Bony Shoulder Stability Ratio (BSSR), a characteristic based on glenoid depth and radius, was determined to evaluate its correlation with the measured SR and to find a suitable characteristic for the assessment of SR independent of defect size. RESULTS: For a translation in 3 o’clock, the linear model resulted in an intercept of 7.13 ± 1.57 (95 % CI [4.01, 10.24]), representing the SR for zero defect size and concavity gradient. The linear coefficient for the predictor concavity gradient averaged 1.05 ± 0.05 (95 % CI [0.96, 1.14]) corresponding to a rise of SR by 1.05 % with each percentage of concavity gradient. Both coefficients were significantly different from zero with p<0.001. The defect size had only an indirect impact on SR, as the linear coefficient of 0.03 ± 0.04 (95 % CI [-0.10, 0.05]) differed insignificantly from zero (p = 0.53). The entire model featured a determination coefficient of R² = 0.98 and a mean squared error (MSE) of 4.22 %. This relationship is diagramed in Figure 2. Using the defect size as an exclusive predictor reduced R² to 0.87 and increased MSE up to 25.72 %. The passive path translation started on average in 2:16 o’clock for the intact glenoid and shifted to 3:06 o’clock with increasing defect size. Though the model indicated a significant impact of concavity gradient as well as defect size on SR (p<0.001), the influence of defect size ( 0.18 ± 0.03, 95 % CI [ 0.24, -0.11])) was significantly smaller than the effect of concavity gradient (0.97 ± 0.04, 95 % CI [0.88, 1.05]). However, the linear model for the passive path resulted in R² = 0.97 and MSE = 5.5 %. Separate linear models for the three groups of low, medium and high initial concavity gradients indicated significant differences in the slope coefficients (low: -0.55 ± 0.05 (95 % CI [ 0.65, 0.45]); medium: 0.78 ± 0.04 (95 % CI [-0.87, -0.70]); high: -1.25 ± 0.06 (95 % CI [ 1.36, -1.13])). This represented a significant impact of the initial glenoid concavity on the loss of SR per defect size. Raw data points as well as the linear approximations are shown in Figure 3. The linear model with the BSSR as a predictor for the measured SR is depicted in Figure 4 indicating a highly linear correlation with R² = 0.98 and MSE = 3.4 % for the translation in 3 o’clock. CONCLUSIONS: The SR is significantly dependent on the glenoid concavity whereas the defect size has a negligible indirect impact, provided that both predictors are included in a linear model. Due to constitutional different glenoid shapes, the loss of SR per defect size is significantly dependent on the initial concavity gradient. However, the BSSR has proven to be a reliable predictor of glenohumeral stability independent of defect size. These findings demonstrate that concavity is a crucial factor in estimating residual SR and substantiate that defect size as the only critical threshold is an inappropriate decisive criterion in the treatment of shoulder instabilities with anterior glenoid bone loss.