• Keijo T Mäkelä, Tuomo Visuri, Pekka Pulkkinen, Antti Eskelinen, Ville Remes, Petri Virolainen, Mika Junnila, and Eero Pukkala.
• Department of Orthopaedics and Traumatology, Surgical Hospital, Turku University Hospital, Turku.
• Acta Orthop. 2014 Feb 1; 85 (1): 32-8.

Background And PurposeMetal-on-metal hip implants have been widely used, especially in the USA, Australia, England and Wales, and Finland. We assessed risk of death and updated data on the risk of cancer related to metal-on-metal hip replacements.Patients And MethodsA cohort of 10,728 metal-on-metal hip replacement patients and a reference cohort of 18,235 conventional total hip replacement patients were extracted from the Finnish Arthroplasty Register for the years 2001-2010. Data on incident cancer cases and causes of death until 2011 were obtained from the Finnish Cancer Registry and Statistics Finland. The relative risk of cancer and death were expressed as standardized incidence ratio (SIR) and standardized mortality ratio (SMR). SIR/SIR ratios and SMR/SMR ratios, and Poisson regression were used to compare the cancer risk and the risk of death between cohorts.ResultsThe overall risk of cancer in the metal-on-metal cohort was not higher than that in the non-metal-on-metal cohort (RR = 0.91, 95% CI: 0.82-1.02). The risk of soft-tissue sarcoma and basalioma in the metal-on-metal cohort was higher than in the non-metal-on-metal cohort (SIR/SIR ratio = 2.6, CI: 1.02-6.4 for soft-tissue sarcoma; SIR/SIR ratio = 1.3, CI: 1.1-1.5 for basalioma). The overall risk of death in the metal-on-metal cohort was less than that in the non-metal-on-metal cohort (RR = 0.78, CI: 0.69-0.88).InterpretationThe overall risk of cancer or risk of death because of cancer is not increased after metal-on-metal hip replacement. The well-patient effect and selection bias contribute substantially to the findings concerning mortality. Arthrocobaltism does not increase mortality in patients with metal-on-metal hip implants in the short term. However, metal-on-metal hip implants should not be considered safe until data with longer follow-up time are available.

30 keywords...

hide…