Poor overall survivorship of large-diameter head metal-on-metal (MoM) THA has been reported by several national registries [2, 11, 17]. It has been estimated that > 1 million MoM hip arthroplasties were performed before widespread concerns about metal debris-related local reactions (adverse reaction to metal debris [ARMD]) . Although the overall revision rates of large-diameter head MoM THAs were higher than expected, there is a large number of patients with large-diameter head MoM THA still in situ and these patients require regular followup. ARMD is the leading cause of revision surgery among patients with large-diameter head MoM THA . However, the natural course and prevalence of ARMD and revision surgery for large-diameter head MoM THA over the longer term are not known, and the followup recommendations are mostly directive [4, 5]. Prevalence of ARMD and risk of revision surgery may also vary among various large-diameter head MoM THA designs [2, 17].
In early reports based on data from the Finnish Arthroplasty Register (Finnish Registry) , short-term survivorship of large-diameter head MoM THA and hip resurfacing arthroplasty (HRA) was comparable with that of conventional cemented metal-on-polyethylene THA at 2 years followup [6, 9]. However, the longer term, overall survivorship of HRA has been inferior compared with conventional THA . Survivorship at 3.6 years for large-diameter head MoM THA was also inferior compared with conventional THA based on combined Nordic data . Longer term survivorship of large-diameter head MoM THA based on Finnish Registry data has not been available to date. Other large, high-quality national registries also show high revision rates for MoM THA [1, 11]. However, the majority of these implants are still in situ and we need longer term followup data to better understand the natural history of MoM THA failures and to optimize the followup of these patients.
The aims of this study were to (1) compare the 10-year competing risk survivorship of large-diameter head MoM THA with the survivorship of conventional THA in the Finnish Register; (2) report the large-diameter head MoM THA survival at the manufacturer/brand level; and (3) identify the most common reasons for revision of large-diameter head MoM THA in the Finnish Register.
Materials and Methods
The Finnish Registry has collected information on total joint arthroplasties since 1980 . Orthopaedic units are obliged to provide all information essential for maintenance of the register to the Finnish National Institute for Health and Welfare. Dates of death and emigration are obtained from Statistics Finland. Data completeness in primary THA in the Finnish Registry is > 95%. Finland is a small country where registries and the healthcare system are publicly funded with practically no loss to followup. If a patient dies or emigrates, they are censored from the registry at that point and this information is updated regularly. Since May 19, 2014, all hip and knee data have been recorded electronically based on bar code reading . Before 2014, indications for revision did not include a diagnosis specifically related to metal-related pathology. As a result, indications such as “aseptic loosening” or “other reason” may have been used to indicate such a diagnosis. However, in 2014, ARMD was added as a reason for revision in the Finnish Arthroplasty Register data content revision.
For the purpose of this study, a large-diameter head MoM THA was described as a diameter of ≥ 38 mm. Data from the six most commonly used cementless large-diameter head MoM THA brands in Finland between years 2004 and 2013 were included (Table 1). All data were gathered from the Finnish Registry database. The end of followup was December 31, 2015. There were 5166 large-diameter head MoM THAs in our study group, of which 89% were performed for primary osteoarthritis. The conventional THA control group consisted of the two most frequently used cementless devices (3060 Vision/Bimetric, Biomet, Warsaw, IN, USA; and 2106 ABG II/ABG II, Stryker, Mahwah, NJ, USA) with metal-on-polyethylene or ceramic-on-ceramic bearing surfaces implanted between 2002 and 2013 (n = 5166) (Table 1). These two conventional cup/stem combinations resulted in a large enough control group to perform analyses while avoiding the possible bias of including a large number of different implants in the analyses. The mean age of the patients in the large-diameter head MoM THA group was 65.1 years and 64.6 years in the control group. Forty-six percent of patients in both groups were men.
The study group was formed by selecting all pairs of large-diameter head MoM and reference THA protheses with the same age group (< 49, 50-54, 55-59, 60-64, 65-69, 70-74, and 75+ years), sex, diagnosis (osteoarthritis, other), and hospital yearly operation count (< 100 operations yearly, ≥ 100 operations yearly). A total of 5166 matched pairs were found. Revision for any reason was considered as the failure endpoint of followup (Table 2). Implant survival (the proportion not revised) was calculated from the corresponding cumulative incidence function adjusted for patient death as a competing event for revision.
In a Cox regression model, MoM group implant revision hazard ratios with 95% confidence intervals were estimated with age group, sex, diagnosis, and hospital volume (yearly operation count) as confounding factors in the model.
Revisions were linked to the primary operation through the personal identification number. The survival endpoint was defined as revision when either one of the components or the entire implant was removed or exchanged (including isolated liner exchanges). Revision for any reason served as an endpoint. Patients with large-diameter head MoM THA or conventional THA devices who died or emigrated during the followup period (until December 31, 2015) were censored at that point.
The proportional hazards assumption of the Cox models was checked by inspecting the corresponding log-log graphs. For Cox analyses comparing the large-diameter head MoM THA brands and the reference THA group, we divided the total followup time into three periods (first year, second and third years combined, and fourth year onward), because the proportional hazards assumption was not fulfilled for the total followup.
Death of the patient and revision are competing risks in registry studies. Mortality of the large-diameter head MoM THA group as a whole was 9.7% and that of the conventional THA control group 17.7%. Therefore, we used competing risk survivorship analysis instead of Kaplan-Meier survivorship.
Inclusion of bilateral cases in a survival analysis violates the basic assumption that all cases are independent. However, several reports have shown that the effect of including bilateral cases in studies of hip and knee prosthesis survival, as done in our study, is negligible [8, 14]. The Wald test was used to test the estimated hazard ratios. Differences between groups were considered to be statistically significant if the p values were < 0.05 in a two-tailed test.
The 10-year competing risk survival rate of 83% (95% confidence interval [CI], 82%-84%) for the large-diameter head MoM group was lower than the 92% survival rate (95% CI, 91%-93%) of the conventional THA control group. The large-diameter head MoM group had a lower revision hazard ratio (HR) during the first postoperative year compared with the conventional THA control group (HR, 0.60; p < 0.001). From the fourth year onward, the large-diameter head MoM group had a higher revision HR compared with the conventional THA control group (HR, 4.88; p < 0.001) (Fig. 1).
ReCap/Bimetric (Biomet), Durom, MMC/CLS, MLTaper (Zimmer, Warsaw, IN), and M2a38/Bimetric (Biomet) large-diameter head MoM designs were associated with a lower revision HR than the conventional THA control group during the first postoperative year (HR, 0.57, p = 0.006; HR, 0.44, p = 0.008; HR, 0.52, p = 0.013, respectively). During the second and third postoperative years, the ASR/Summit and Corail large-diameter head MoM design was associated with a higher revision risk than the conventional THA control group (HR, 3.64; p < 0.001) and from the fourth postoperative year onward, all of the large-diameter head MoM THA devices were associated with a higher risk of revision than the conventional reference THA (HR, 23.57 for ASR THA; HR, 5.53 for Conserve THA [Wright Medical Technology, Arlington, TN, USA); HR, 4.24 for BHR THA; HR, 4.16 for Durom, MMC/CLS, MLTaper THA; HR, 2.47 for M2a38/Bimetric THA; HR, 2.14 for ReCap/Bimetric THA; p value for all < 0.001) (Table 3).
The 10-year competing risk survivorship rates among the six large-diameter head MoM THA designs analyzed varied from a high of 88% (95% CI, 86%-90%) for the ReCap/Bimetric and 85% (95% CI, 84%-87%) for the M2a38/Bimetric to a low of 46% (95% CI, 41%-51%) for the recalled ASR with either the Summit® or Corail® stem (Fig. 2; Table 4). The 8-year competing risk survivorship of the Conserve Plus/Profemur was 85% (95% CI, 82%-88%) (Table 4).
Before the Finnish Registry data content revision in May 2014, the main reason for revision of large-diameter head MoM THAs was aseptic loosening of both components (42%), whereas conventional THAs were most often revised as a result of dislocation (25%). Unspecified reasons for revision ("other") or missing data were recorded for 27% of the large-diameter head MoM THA revisions and for 11% of the conventional THA revisions (Table 5). Since that time, the main reason for revision of large-diameter head MoM THA has been ARMD (69.2%), whereas conventional THAs were most often revised as a result of periprosthetic fracture (18.3%) and dislocation (16.7%) (Table 6).
Despite the high revision rates of large-diameter head MoM THAs, many patients still have these implants in situ. Reporting results from high-quality registries provides additional information to the orthopaedic community as they continue to follow this population. In our study, the 10-year competing risk survivorship of large-diameter head MoM THA in Finland was 83%. According to the current National Institute for Health and Care Excellence (NICE) recommendations, the target revision rate of THA should be < 5% at 10 years . None of the large-diameter head MoM THAs in the current study achieved this goal. Based on our results, strict followup recommendations for MoM THA should be maintained over the longer term.
We acknowledge that our study has a number of limitations. First, as is generally true in registry-based studies, we were unable to include comorbidity data in our analysis. This might have explained some of the variation in revision rates. Although some data show that young males are at a higher risk of revision , possibly as a result of their activity level, we do not think that the lack of comorbidity data has biased our results. To minimize the effect of sex and age, we have controlled for these in our analyses. Second, we were not able to assess any patient-reported outcome measures and it is possible that some of the patients are symptomatic although they have not undergone revision. Furthermore, the Finnish Registry does not include information on radiographs, MRI results, or data on metal ion levels. Therefore, we were not able to estimate the true ARMD prevalence, which might be higher than that based on revision rates alone. Additionally, revision threshold may vary between our study and the control group. It is possible that, especially after the publicity surrounding ARMD, and the ASR recall, the revision threshold for implants with metal bearings might have been lower for unspecified pain, for example, than that for patients with traditional THA. Furthermore, we were not able to analyze the possible effect of corrosion and fretting of the head-neck junction characterized as trunnionosis. This has been speculated to be one possible reason for lower revision and ARMD rates for small-diameter head MoM THA compared with large-diameter head MoM THA. However, the minimum head size in our study was 38 mm and therefore we do not think that this has biased our results.
Our 10-year competing risk survivorship of 83% for the large-diameter head MoM THA group is in line with the findings of the cumulative revision rate for all large-diameter head MoM THAs of 19% at 10 years in England and Wales and 19% in Australia [2, 12]. In a previous report based on Finnish Registry data from 2002 to 2009, we concluded that large-diameter head MoM THAs had survivorship comparable to that of conventional THAs at 2.4 years followup . It is now evident that this conclusion is not valid at longer followup. This is an important message to clinicians because our results show that strict followup protocols should be maintained even at longer followup. A large number of patients still have MoM THA implants in situ and our job should be to recognize possible ARMD signs and symptoms at the earliest stage possible. Followup guidelines need to be updated regularly to best achieve this goal.
One earlier study based on combined Nordic registry data reported higher revision rates for large-diameter head MoM THA compared with conventional THA . In that study, the revision risk difference between large-diameter head MoM and conventional THA diminished after excluding ASR devices from the MoM group. We were not able to replicate this in our material, in which all six large-diameter head MoM THA implant brands had individually inferior survival compared with the conventional THA control group. Even the best performing large-diameter head MoM THA in our analysis (ReCap/Bimetric) had a > 2.5-fold revision risk from 4 years onward compared with the control group and its up to 10-year survivorship was only 88%. One possible explanation for this difference is the mean followup time in the study by Varnum et al. was only 3.6 years for the MoM THA group compared with 7.1 years in the current study . Furthermore, we included only “true” large-diameter head MoM THAs with a head size of ≥ 38 mm, whereas the previous study based on combined Nordic data included also MoM THAs with smaller head sizes . At the individual implant level, there were some small differences in implant survival compared with the Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) and the National Joint Registry of England and Wales (NJR) (Table 7). Implant survival in our material was slightly inferior for all MoM THA devices compared with AOANJRR and NJR data. Possible explanations for this are slight variation in cup stem combinations and that our material included only very large-diameter heads (≥ 38 mm). In addition, some of the difference between revision rates in the Finnish Registry and the NJR might be explained by the lower early revision completeness in the NJR while the registry was still evolving in the early 2000s. This might have led to missing some of the revision cases . Furthermore, we used competing risk survivorship in our analyses instead of Kaplan-Meier survivorship, which may also have slightly influenced our results.
Before the Finnish Registry data content revision, the large-diameter head MoM THA group had a higher percentage of revisions as a result of aseptic loosening and “other reasons” than the reference THA group (51% versus 25% and 27% versus 11%, respectively). When metal bearing-related problems first started to appear, the underlying cause was initially unclear and there was no specific option to record revision as a result of ARMD in our registry. This was unique to the Finnish Registry as a result of the older data content. This makes the reason for revision analysis in the large-diameter head MoM THA group difficult when revisions resulting from metal reactions might have earlier been recorded under different revision indications. After the Finnish Registry data content revision, 69% of large-diameter head MoM THA revisions were recorded as performed for ARMD, which is in accordance with data from the NJR in which ARMD is the most common reason for revision in MoM THA .
In conclusion, all commonly used large-diameter head MoM THAs in Finland had high revision rates at mean 10 years followup. All large-diameter head MoM THAs should thus be followed up in strict protocols based on recommendations by national bodies. Based on our results, these followup recommendations should be maintained even at longer term followup. In the future, followup guidelines should be reevaluated when even longer term followup data are available.
1. AOANJRR. AOANJRR Supplementary Report 2014. Metal on Metal Bearing Surface Total Conventional Hip Arthroplasty. Available at: https://aoanjrr.sahmri.com/annual-reports-2014
. Accessed October 29, 2017.
2. AOANJRR. Annual Report 2016. 2016. Available at: https://aoanjrr.sahmri.com/annual-reports-2016
. Accessed October 29, 2017.
3. FAR. The Finnish Arthroplasty Register (FAR). Available at: www.thl.fi/far
. Accessed October 29, 2017.
4. FDA. General Recommendations for Orthopaedic Surgeons After Metal-on-metal Hip Replacement Surgery (Follow-up). Available at: https://www.fda.gov/MedicalDevices/ProductsandMedicalProcedures/ImplantsandProsthetics/MetalonMetalHipImplants/default.htm
. Accessed October 29, 2017.
5. Finnish Arthroplasty Association. Finnish Arthroplasty Associations MoM Follow Up Recommendations2014: http://www.suomenartroplastiayhdistys.fi/index.php?page=1050&lang=1
.October 29, 2017.
6. Junnila M, Kostensalo I, Virolainen P, Remes V, Matilainen M, Vahlberg T, Pulkkinen P, Eskelinen A, Itälä A, Mäkelä K. Hip resurfacing arthroplasty versus large diameter head metal-on-metal total hip arthroplasty: comparison of three designs from the Finnish Arthroplasty Register. Scand J Surg. 2014;103:54–59.
7. Kwon YM, Lombardi AV, Jacobs JJ, Fehring TK, Lewis CG, Cabanela ME. Risk stratification algorithm for management of patients with metal-on-metal hip arthroplasty: consensus statement of the American Association of Hip and Knee Surgeons, the American Academy of Orthopaedic Surgeons, and the Hip Society. J Bone Joint Surg Am. 2014;96:e4.
8. Lie SA, Engesæter LB, Havelin LI, Gjessing HK, Vollset SE. Dependency issues in survival analyses of 55,782 primary hip replacements from 47,355 patients. Stat Med. 2004;23:3227–3240.
9. Mokka J, Mäkelä KT, Virolainen P, Remes V, Pulkkinen P, Eskelinen A. Cementless total hip arthroplasty with large diameter metal-on-metal heads: short-term survivorship of 8059 hips from the Finnish Arthroplasty Register. Scand J Surg. 2013;102:117–123.
10. NICE. NICE Guidelines for Total Hip Arthroplasty: https://www.nice.org.uk/guidance/ta304/chapter/3-The-technologies#the-technologies
. Accessed October 29, 2017.
13. Paavolainen P, Hämäläinen M, Mustonen H, Slätis P. Registration of arthroplasties in Finland. A nationwide prospective project. Acta Orthop Scand. 1991;241:27–30.
14. Robertsson O, Ranstam J. No bias of ignored bilaterality when analysing the revision risk of knee prostheses: analysis of a population based sample of 44,590 patients with 55,298 knee prostheses from the national Swedish Knee Arthroplasty Register. BMC Musculoskelet Disord. 2003;5:1.
15. Seppänen M, Karvonen M, Virolainen P, Remes V, Pulkkinen P, Eskelinen A, Liukas A, Mäkelä KT. Poor 10-year survivorship of hip resurfacing arthroplasty. Acta Orthop. 2016;87:554.
16. Sibanda N, Copley LP, Lewsey JD, Borroff M, Gregg P, MacGregor AJ, Pickford M, Porter ML, Tucker K, Van Der Meulen JH. Revision rates after primary hip and knee replacement in England between 2003 and 2006. PLoS Med. 2008;5:1398–1407.
17. Varnum C, Pedersen AB, Mäkelä K, Eskelinen A, Havelin LI, Furnes O, Kärrholm J, Garellick G, Overgaard S. Increased risk of revision of cementless stemmed total hip arthroplasty with metal-on-metal bearings. Acta Orthop. 2015;3674:1–8.