Aims. Metal and ceramic humeral head bearing surfaces are available choices in anatomical shoulder arthroplasties. Wear studies have shown superior performance of ceramic heads, however comparison of clinical outcomes according to bearing surface in total shoulder arthroplasty (TSA) and hemiarthroplasty (HA) is limited. This study aimed to compare the rates of revision and reoperation following metal and ceramic humeral head TSA and HA using data from the National Joint Registry (NJR), which collects data from England, Wales, Northern Ireland, Isle of Man and the States of Guernsey. Methods. NJR shoulder arthroplasty records were linked to Hospital Episode Statistics and the National Mortality Register. TSA and HA performed for osteoarthritis (OA) in patients with an intact rotator cuff were included. Metal and ceramic humeral head prostheses were matched within separate TSA and HA groups using propensity scores based on 12 and 11 characteristics, respectively. The primary outcome was time to first revision and the secondary outcome was non-revision reoperation. Results. A total of 4,799 TSAs (3,578 metal, 1,221 ceramic) and 1,363 HAs (1,020 metal, 343 ceramic) were included. The rate of revision was higher for metal compared with ceramic TSA, hazard ratio (HR) 3.31 (95% confidence interval (CI) 1.67 to 6.58). At eight years, prosthesis survival for ceramic TSA was 98.7% (95% CI 97.3 to 99.4) compared with 96.4% (95% CI 95.2 to 97.3) for metal TSA. The majority of revision TSAs were for cuff insufficiency or instability/dislocation. There was no significant difference in the revision rate for ceramic compared with metal head HA (HR 1.33 (95% CI 0.76 to 2.34)). For ceramic HA, eight-year
