Humeral Stem Loosening is Not Always Prosthetic Joint Infection.

BACKGROUND: Aseptic humeral stem loosening is an uncommon complication of shoulder arthroplasty, and its presence has been considered a highly specific predictor of prosthetic joint infection (PJI). Literature on aseptic humeral stem loosening is sparse. The primary purpose of this study was to determine the rate of aseptic humeral loosening in revision shoulder arthroplasty. Secondarily, we sought to identify predictors of septic and aseptic humeral loosening.

METHODS: Our institutional revision shoulder arthroplasty database was reviewed. Inclusion criteria were patients with humeral stem loosening as indicated in the operative report. Patient demographics, index surgery and indication, revision surgery and indication, and operative data were recorded. Charts were manually reviewed, and PJI scores were calculated using the International Consensus Meeting (ICM) scoring criteria. ICM classifications of PJI "unlikely" and "possible" were considered aseptic, and PJI "probable" and "definite" were considered septic. Statistical analysis was performed to determine associations between the abovementioned variables and aseptic humeral loosening.

RESULTS: Forty-six patients with forty-eight shoulders were included in our analysis. Ten cases (21%) were classified as definite PJI, thirteen cases (27%) were classified as probable PJI, four cases (8%) were classified as possible PJI, and twenty-one cases (44%) were classified as PJI unlikely. On analysis of stem design, there were no statistically significant associations with aseptic loosening; although, a lack of proximal porous ingrowth surface trended toward higher rates of aseptic loosening in all patients and in ATSA and RTSA subgroup analyses. In the index RTSA subgroup, aseptic loosening was associated with female sex (p=0.005). Seventeen of 39 shoulders (44%) that underwent either index ATSA or RTSA demonstrated concomitant glenoid loosening. The absence of glenoid loosening was associated with aseptic humeral loosening in index ATSA and RTSA (p<0.001).

CONCLUSION: 52% of revision shoulder arthroplasty cases with humeral loosening performed at our institution were aseptic. There appear to be distinct demographic and radiographic factors that are more commonly associated with aseptic as opposed to septic humeral loosening. Our data demonstrate that demographic predictors of aseptic loosening of RTSA include female sex. The absence of glenoid component loosening is associated with aseptic humeral loosening and concomitant glenoid loosening is associated with septic humeral loosening. Understanding of these factors can guide the preoperative index of suspicion for PJI in the setting of humeral stem loosening.