Thus far the ability to predict who will develop early failure following the insertion of a metal-on-metal (MoM) bearing has been very limited. Our objective was to assess the effect of smoking on failure rates in patients with MoM bearing, compared with patients with ceramic-on-polyethylene (CoP) bearing. From a prospective hospital-based registry we included all primary THAs operated upon between 1/2001 and 12/2011 with MoM or CoP bearings of the same cup design and head size (28 mm). We compared revision rates through 10/2013 classified by smoking status and type of bearing. We included 1,964 patients (median age 71, 57% women), 663 with MoM and 1,301 with CoP bearing. Mean follow-up was 6.9 years (range 1.8-12.8). Revisions were required for 56 THAs. In patients with MoM bearing the adjusted incidence rate of revision among ever-smokers was four times greater than among never-smokers (95% CI 1.4-10.9). Among those with CoP bearing, the rate ratio was only 1.3 (95% CI 0.6-2.5). We found a strong association between smoking and increased failure of MoM THAs. In contrast, the association was weak for patients with CoP bearing. Smoking might be a trigger or an effect amplifier for adverse reactions to metal debris from MoM bearings. © 2014 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 32:762-768, 2014.