OBJECTIVETo investigate normal bony anatomy of the glenoid rim, to measure inner glenoid rim angle and outer glenoid rim angle, and the angles for successful anchor insertion for arthroscopic labral repairs.

METHODSTwelve unpaired isolated human glenoids (6 right, 6 left) without any evidence of trauma were for studying. The glenoid specimens were scanned using 320-slice CT (Aquilion ONE), then reconstruction glenoid to a three dimensional model using materialise's interactive medical image control system (Mimics) and to obtain cross-sectional images in 6 different planes, mark the right glenoid rim with clockwise tag, the left with counterclockwise tag. Inner glenoid rim angle marked as angle α and outer glenoid rim angle marked as angle β were measured from the cross-sectional images of the glenoids at 8 positions: 2-, 3-, 4-, 5-, 6-, 7-, 8- and 9-o'clock positions. Glenoid morphology was noted for each position. Using 12 mm as radius, measured the minimum insertion angle of anchor, marked as angle γ. Normal distribution of the data was confirmed with Kolmogorov-Smirnov test. Paired t-test was performed to detect differences in the angles between two locations. Two independent samples t-test was performed to detect differences in the angles between same location of left and right. Analysis of variance (ANOVA) was performed to detect differences in the angles between right and left, and different locations of the glenoid rim.

RESULTSThe smallest α was at the 4-o'clock position (right 50° ± 6°, left 52° ± 9°), significant difference were seen when compared with the 6-o'clock position (t = 10.466, P = 0.000) or the 5-o'clock position (t = 3.754, P = 0.003), no significant difference exist between 4-o'clock position and 3-o'clock posion (t = 0.926, P = 0.374). The smallest β was at the 3-o'clock position (right 50° ± 6°, left 53° ± 10°), significant difference were seen when compared with the 6-o'clock position (t = 9.862, P = 0.000) or the 5-o'clock position (t = 3.634, P = 0.003), no significant difference exist between 4-o'clock position and 3-o'clock posion (t = 0.697, P = 0.501). Asymmetric morphology of the glenoid was noted with an almost straight line extending medially from the rim at the 3-o'clock position, whereas a concave morthology was noted at the 9-o'clock position. Similary at the 4- and 5-o'clock position, the scapular bony surface did not curve toward the base as markedly as it did at the corresponding posterior 8- and 7-o'clock position. Angle γ from the 3-o'clock to the 9-o'clock were 25° ± 4°, 54° ± 6°, 83° ± 4°, 119° ± 2° at right side, 23° ± 4°, 57° ± 4°, 89° ± 7°, 119° ± 4° at left side. No significant difference of any angle at the same position was noted between left and right (α:t = 0.283-1.785, P > 0.05;β:t = 0.369-2.067, P > 0.05;γ:t = 0.145-0.492, P > 0.05).

CONCLUSIONSThe available bone mass for the anchor insertion is found to vary depending on the position of the glenoid rim. The smallest inner and outer glenoid rim angle are at the 4- and 3-o'clock position. The minimum insertion angles of anchor differ at different position. Both rim angle and glenoid morthology for each position must be considered when selecting the ideal anchor insertion angle for Bankart repair. Meanwhile, minimum insertion angle of anchor should also be considered before anchor insertion.