Abstract
Introduction
We investigated whether blood metal ions could effectively identify bilateral metal-on-metal hip patients at risk of adverse reactions to metal debris (ARMD).
Patients and methods
This single-centre, prospective study involved 235 patients (185 bilateral Birmingham Hip Resurfacings (BHRs) and 50 bilateral Corail-Pinnacles) undergoing whole blood metal ion sampling (mean time=6.8 years from latest implant to sampling). Patients were divided into ARMD (revised or ARMD on imaging; n=40) and non-ARMD groups (n=195). Metal ion parameters (cobalt; chromium; maximum cobalt or chromium; cobalt-chromium ratio) were compared between groups. Optimal metal ion thresholds for identifying ARMD patients were determined using receiver operating characteristic (ROC) analysis, which compares the performance of different tests using the area under the curve (AUC) (higher AUC=more discriminatory).
Results
All ion parameters were significantly higher (p<0.0001) in ARMD patients versus non-ARMD patients. Maximum cobalt or chromium produced the highest AUC for BHRs (91.0%) and Corail-Pinnacles (71.8%). For BHRs, maximum cobalt or chromium AUC was significantly greater than cobalt-chromium ratio AUC (p=0.019) but not compared to cobalt (p=0.574) or chromium (p=0.721). For Corail-Pinnacles, AUCs for all ion parameters were not significantly different (p>0.160). Optimal ion thresholds for identifying ARMD varied between implants (maximum cobalt or chromium: BHR=5.52µg/l; Corail-Pinnacle=4.04µg/l). Thresholds had good sensitivities (83.3%–90.0%) and specificities (65.0%–88.4%), high negative predictive values (96.3%–96.5%) and lower positive predictive values (39.1%–58.1%). Fixed USA (10µg/l) and UK (7µg/l) authority thresholds missed more hips with ARMD (11–14 hips missed; 4.7–6.0%) compared to Implant Specific Thresholds (6 hips missed; 2.6%).
Discussion
Bilateral BHR and Corail-Pinnacle patients with blood metal ions below Implant Specific Thresholds were at low-risk of ARMD. These thresholds could rationalise follow-up resources in asymptomatic patients.
Conclusion
Implant Specific Thresholds are preferable to fixed authority thresholds given that Implant Specific Thresholds were more effective for identifying patients at risk of ARMD requiring further investigation.
