INTRODUCTION: Assessment of radiolucent lines (RLL) is the main component of the radiographic analysis of the glenoid component. The purpose of this study is to compare the radiographic results of two glenoid preparation techniques by analyzing periglenoid radiolucencies.

MATERIEL AND METHODS: The series consists of seventy-two shoulder arthroplasties with primary osteoarthritis. Shoulders were divided into two groups based on glenoid preparation technique:

Group 1: Thirty-seven shoulders operated on between 1991 and 1995 with flat back, polyethylene glenoid implants cemented after “curettage” of the keel slot.

Group 2: Thirty-five shoulders operated on between 1997 and 1999 with flat back, polyethylene glenoid implants cemented after cancellous bone compaction of the keel slot.

At least three of the following four fluoroscopically positioned, postoperative AP radiographs were analyzed: immediate postoperative, between the 3rd and 6th postoperative months, at one year postoperative and at two years postoperative. The immediate and the two year radiograph were required for study inclusion. The radiolucent line score (RLLS) was calculated using the technique of Molé, involving the summation of radiolucencies in each of six specified zones. The RLLS was compared between the two groups.

RESULTS: On the immediate postoperative radiographs the average of the total RLL score of the 9 analyzes was 2.39 in Group 1 and 1.67 in Group 2 (p=0.042). There was a statistically significant association between the glenoid preparation technique and the incidence of radiolucency around the keel as well (p=0.001). There was no significant difference in radiolucency behind the faceplate between the two groups (Group 1: 1.54 and Group 2: 1.41; p=0.394). On the 2-year postoperative radiographs the average RLL score of the 9 analyzes were 6.44 in the Group1 (4.05 under the tray, and 2.39 around the keel), and 4.19 in Group2 (p=0.0005) (2.86 under the tray, and 1.33 around the keel). The radiolucency around the keel and behind the faceplate (p=0.0005) was significantly more important (p=0.001) in the “curettage” glenoid preparation population. A significantly higher degree of progression of the total RLL score (p=0.002) and of the radiolucency behind the faceplate (p=0.001) was observed in the “curettage” glenoid preparation group.

DISCUSSION/CONCLUSION: Preparation of the glenoid component keel slot with cancellous bone compaction is radiographically superior to the “curettage” technique with regard to periglenoid radiolucen-cies. Although new techniques of glenoid preparation may help to decrease the rate of RLL, this study shows that even with better technique, the RLL are evolutive and may appear after few years in initially perfectly implanted glenoid.