Distributed Analysis of Hip Implants Using Six National and Regional Registries: Comparing Metal-on-Metal with Metal-on-Highly Cross-Linked Polyethylene Bearings in Cementless Total Hip Arthroplasty in Young Patients
Ove Furnes, MD, PhD; Elizabeth Paxton, MA; Guy Cafri, PhD, MStat; Stephen Graves, MBBS, DPhil, FRACS, FAOrthA; Barbara Bordini, BSc; Thomas Comfort, MD; Moises Coll Rivas, MD; Samprit Banerjee, PhD; Art Sedrakyan, MD, PhD
Background: The regulation of medical devices has attracted controversy recently because of problems related to metal-on-metal hip implants. There is growing evidence that metal-on-metal implants fail early and cause local and systemic complications. However, the failure associated with metal-on-metal head size is not consistently documented and needs to be communicated to patients and surgeons. The purpose of this study is to compare implant survival of metal on metal with that of metal on highly cross-linked polyethylene.
Methods: Using a distributed health data network, primary total hip arthroplasties were identified from six national and regional total joint arthroplasty registries (2001 to 2010). Inclusion criteria were patient age of forty-five to sixty-four years, cementless total hip arthroplasties, primary osteoarthritis diagnosis, and exclusion of the well-known outlier implant ASR (articular surface replacement). The primary outcome was revision for any reason. A meta-analysis of survival probabilities was performed with use of a fixed-effects model. Metal-on-metal implants with a large head size of >36 mm were compared with metal-on-highly cross-linked polyethylene implants.
Results: Metal-on-metal implants with a large head size of >36 mm were used in 5172 hips and metal-on-highly cross-linked polyethylene implants were used in 14,372 hips. Metal-on-metal total hip replacements with a large head size of >36 mm had an increased risk of revision compared with metal-on-highly cross-linked polyethylene total hip replacements with more than two years of follow-up, with no difference during the first two years after implantation. The results of the hazard ratios (and 95% confidence intervals) from the multivariable model at various durations of follow-up were 0.95 (0.74 to 1.23) at zero to two years (p = 0.698), 1.42 (1.16 to 1.75) at more than two years to four years (p = 0.001), 1.78 (1.45 to 2.19) at more than four years to six years (p < 0.001), and 2.15 (1.63 to 2.83) at more than six years to seven years (p < 0.001).
Conclusions: We conducted a comparison of large-head-size, metal-on-metal implants and metal-on-highly cross-linked polyethylene implants in younger patients with uncemented fixation. We found consistent and strong evidence worldwide that large-head-size, metal-on-metal implants were associated with increased risk of revision after two years compared with metal-on-highly cross-linked polyethylene implants, with the effect becoming more pronounced over time.