Aims. The aim of this study was to determine whether the rates of revision for metal-on-metal (MoM) total hip arthroplasties (THAs) with Pinnacle components varied according to the year of the initial operation, and compare these with the rates of revision for other designs of MoM THA. Patients and Methods. Data from the National Joint Registry for England and Wales included 36 mm MoM THAs with Pinnacle acetabular components which were undertaken between 2003 and 2012 with follow-up for at least five years (n = 10 776) and a control group of other MoM THAs (n = 13 817). The effect of the year of the primary operation on all-cause rates of revision was assessed using Cox regression and interrupted time-series analysis. Results. For MoM THAs involving Pinnacle components, those undertaken between 2007 and 2012 had higher rates of revision compared with those undertaken between 2004 and 2006 (hazard ratio (HR) 2.01; 95% confidence interval (CI) 1.57 to 2.57; p < 0.001). For THAs undertaken during and after 2007, the number of revisions per 1000 implant-years at risk significantly increased by 5.20 (95% CI 0.52 to 9.89; p = 0.033) compared with those undertaken before this time. In the control group, THAs undertaken between 2007 and 2012 also had higher rates of revision (HR 1.77; 95% CI 1.49 to 2.10; p < 0.001), with revisions per 1000 implant-years for those undertaken during and after 2007 significantly increasing by 6.13 (95% CI 1.42 to 10.83; p = 0.016). Conclusion. The five-year revision rates were significantly increased for all primary MoM THAs undertaken from 2007 onwards. Contrary to recent reports, this finding was not specific to those involving Pinnacle acetabular components and may be explained by increased surveillance and recent lowering of the threshold for revision. Cite this article: Bone Joint J 2018;100-B:33–41