Modeling highly crosslinked polyethylene vs. non–highly crosslinked polyethylene glenoid revision rates for anatomic shoulder arthroplasty in osteoarthritis including differing polyethylene glenoid fixation designs

Abstract

Background

We compared anatomic total shoulder arthroplasty (aTSA) for osteoarthritis to both highly crosslinked polyethylene (XLPE) and non–highly crosslinked polyethylene (non-XLPE) to determine the rate of revision for multiple patient and implant characteristics, modeling the effect of variation in glenoid fixation design (glenoid component type).

Methods

Data from a large national arthroplasty registry were analyzed for the period April 16, 2004, to December 31, 2022. The study population included all primary aTSA (stemmed and stemless shoulder arthroplasty) procedures with a primary diagnosis of osteoarthritis and performed using prostheses in current use. These procedures were grouped into 2 cohorts: all polyethylene-bearing glenoid components with either XLPE or non-XLPE. The cumulative percent revision was determined using Kaplan–Meier estimates of survivorship and hazard ratios from Cox proportional hazard models adjusted for age, sex, humeral head size, humeral fixation, type of primary (total stemmed or stemless anatomic), glenoid component type (modular and nonmodular metal-backed glenoid, cemented polyethylene glenoid, and polyethylene glenoid with modified central peg), and surgeon volume (after 2008). Possible interactions were examined. A subanalysis from January 1, 2017, captured the additional patient demographics of American Society of Anesthesiologists score, body mass index, and glenoid morphology.

Results

Of 11,003 aTSA procedures, the cumulative percent revision at 14 years for all XLPE glenoids (n = 3865) was 5.8% (95% confidence interval [CI] 3.9, 8.7), and 18.7% (95% CI 16.6, 21.0) for non-XLPE (n = 7138). XLPE had a lower rate of revision from 2 years (non-XLPE vs. XLPE 2 years + hazard ratio = 1.66, (95% CI 1.09, 2.53), P = .018) adjusting for age, sex, humeral head size, type of primary, humeral stem fixation, and glenoid component type. Overall, glenoid component type and polyethylene type were strongly associated (P < .001 and P = .021, respectively) with all-cause aTSA revision rates. The difference between non-XLPE and XLPE is observed across all polyethylene glenoid types. When considering procedures performed between 2017 and 2022, XLPE vs. non-XLPE rates of revision were not significantly different with extended adjustment at subanalysis, but loosening did not predominate for non-XLPE until year 6 of follow-up.

Conclusion

Both the glenoid design and the type of polyethylene predict the revision rate for aTSA. However, the relative rates of revision between glenoid designs did not differ with polyethylene type. While the polyethylene type was not associated with aTSA revision rates in a more contemporary analysis, the incidence of loosening in non-XLPE prostheses combinations was higher from 6 years may explain this.