Similarly low blood metal ion levels at 10‐years follow‐up of total hip arthroplasties with Oxinium, CoCrMo, and stainless steel femoral heads. Data from a randomized clinical trial

Abstract The use of inert head materials such as ceramic heads has been proposed as a method of reducing wear and corrosion products from the articulating surfaces in total hip arthroplasty, as well as from the stem‐head taper connection. The aim of the present study was to compare the blood metal ion levels in patients with Oxinium and CoCrMo modular femoral heads, as well as monoblock stainless steel Charnley prostheses at 10 years postoperatively. The 150 patients with osteoarthritis of the hip joint included in a randomized clinical trial were grouped according to femoral head material. One group (n = 30) had received the Charnley monoblock stainless steel stem (DePuy, UK). The other patients (n = 120) received a Spectron EF CoCrMo stem with either a 28 mm CoCrMo or Oxinium modular head (Smith & Nephew, USA). After 10 years, 38 patients had withdrawn, 19 deceased, 7 revised due to aseptic loosening and 5 revised due to infection. The 81 patients with median age of 79 years (70–91) were available for whole blood metal ion analysis. The levels of Co, Cr, Ni and Zr in the blood were generally low with all the head materials (medians <0.3 micrograms/L) and no statistical difference between the groups were found (p = .2–.8). Based on the low blood metal ion values in our study groups, no indication of severe trunnion corrosion in patients with CoCrMo heads was observed, neither was there any beneficial reduction in metal ion exposure with the Oxinium femoral heads.


| INTRODUCTION
The occurrence of wear and corrosion at the trunnion-head interface (trunnionosis) has gained wider recognition as an important clinical matter in metal-on-metal 1-3 as well as metal-on-polyethylene total hip arthroplasty (THA). 4,5 Some suggest that trunnionosis accounts for up to 3% of all THA failures. 6,7 Apart from aseptic loosening and osteolysis from local release of metal debris, also systemic toxicity related to high cobalt ion levels with neurological 8 and cardiac 9,10 manifestations (termed cobaltism), has caused further concerns.
Oxidized zirconium (Oxinium™) femoral heads were introduced with the prospects of combining the strength of a metal head with the smoothness and hardness of ceramic. 11 Oxinium has a metal alloy core (zirconium-niobium) and an approximately 5 μm thick zirconia ceramic surface made by heat treatment of the surface. 12 Theoretically, the Oxinium femoral head and taper junction can reduce the release of metals due to the ceramic surface. One study showed that the use of Oxinium femoral head decreased the blood cobalt levels versus a cobalt-chromium-molybdenum alloy (CoCrMo) femoral head, albeit it did not reach statistical significance. 13 However, a retrieval study showed no difference in the corrosion score between Oxinium and CoCrMo femoral heads. 14 In this randomized clinical study, we wanted to evaluate the blood metal ion levels of Oxinium modular femoral heads at 10 years postoperatively and compare this with patients with the CoCrMo modular femoral head counterpart, and the monoblock stainless steel Charnley prosthesis. The null hypothesis was that blood metal ion levels with Oxinium femoral heads were equal to that of the metal heads. Of the 150 patients in the original study, 19 patients had deceased and 38 patients withdrew from the study and could F I G U R E 1 Participant flow diagram describing the original five study groups based on the different head-cup articulations and the final three groups based on head material (stainless steel, CoCrMo, and Oxinium). Exclusions after 10 years follow-up are sorted after revision causes (infection or aseptic loosening) and reasons for drop out (deceased or lost to follow-up). The gender distribution, age and Harris Hip Score (median and range) in the three study groups are specified therefore not provide a blood sample at 10 years follow-up (not able or willing to participate) ( Figure 1). Seven patients were revised before 10 years follow-up due to aseptic loosening of either the cup or stem or both. 4 of these cases was included in a retrieval project where explants were collected with periprosthetic tissue and whole blood samples. 17 Additionally, five hips were revised due to infection. None of the patients in the Charnley group were revised.

| Blood metal ion analysis
Whole blood was drawn from a vein in the forearm using an intravenous catheter (venesection) in our hospital (Becton Dickinson Venflon Pro™, 1.5 Â 45 mm, Helsingborg, Sweden). After venipuncture, the steel needle was removed, and the blood (5-10 ml) was allowed to flow directly into a 15 ml metal free polypropylene tube (VWR International, CS500, Radnor, PA, US). The blood was stored at minus 20 C until further handling. The concentration of chromium, cobalt, nickel and zirconium in whole blood was determined by Inductively  CoCrMo modular femoral head, retrieved from study case n 819, with abrasive marks on the stem-cement interface, mainly on the posterior-distal-medial side. The collared Spectron EF stem was fabricated with a matte distal surface (Ra 0.7 μm) and a grit blasted roughened surface (Ra 7.3 μm) on the proximal third of the stem. The neck and collar are highly polished. 15 The stem used in this RCT was supplied with three tantalum markers, one of which was attached to a cone in the neck, the second in the medial side of the collar and the third on the distal tip of the stem. (C) CoCrMo 28 mm femoral head (left) and Oxinium 28 mm femoral head (right). Spectron EF stem taper size 12/14, standard offset (in the middle). (D) Reflection All-Poly UHMWPE cup.

| Statistics
versus patients with the study prosthesis only (pooled).
Clinical outcome was evaluated with the Harris Hip Score (HHS).
Median group HHS scores are reported preoperatively and at 10 years ( Figure 1) and study groups statistically compared with Kruskal-Wallis one-way analysis of variance (two-sided). Probability (p) values of <.05 denote statistical significance.

| RESULTS
Patients in the stainless steel, CoCrMo and Oxinium femoral head groups demonstrated significantly higher HHS scores 10 years postoperatively compared to preoperative scores, and there were no significant differences in clinical function between the three groups (p = .69) (Figure 1).
The median blood metal ion levels for Cr, Co, Zr, and Ni were generally low in all study groups (Table 1 and Figure 3). Patients with the modular CoCrMo head had approximately twice as high median Cr blood concentration (0.29 μg/L) than the Oxinium group (0.17 μg/ L) and the nonmodular stainless steel group (0.14 μg/L). The median Co level for the CoCrMo head group (0.15 μg/L) was almost twice the median level (0.08 μg/L) for the stainless steel group, but similar to the Oxinium group (0.15 μg/L). The median zirconium level was highest in the Oxinium group (0.2 μg/L), actually 4 times higher than in the stainless steel group. Still, we were not able to find statistical differences in any of the blood metal ion levels between the three head groups (p = .2-.8) ( Table 1). The different acetabular cup materials (conventional or highly crosslinked polyethylene) did not influence the metal levels (p = .2-.9, five groups).
Seven patients were revised due to aseptic loosening of either the cup or stem or both. None of the patients in the Charnley group were revised. See table 2 for revision cause and blood metal levels.
One of the Spectron cases with CoCrMo head that was revised after

| DISCUSSION
In the present study of primary THAs we found no difference in blood ion levels with the use of Oxinium femoral heads compared to CoCrMo or stainless steel heads at 10 years follow-up. To our knowledge this is the first RCT who compares metal ion release from these three femoral head materials.
All the Co and Cr values were well below the British Medicines and Healthcare Products Regulatory Agency threshold value of 7 μg/ T A B L E 1 Cr, Co, Zr and Ni concentrations in the three groups based on head material (stainless steel, CoCrMo and Oxinium), represented with median, mean and maximum (max) values (microgram/L). Statistical difference between the groups is calculated by Kruskal-Wallis one-way test It is also comparable with preoperative serum metal concentrations to another study, where they found mean Cr and Co concentrations of THAs had more necrosis and lymphocytic infiltration in their tissues than patients with hip resurfacings, indicating that trunnion wear debris produced more cytotoxic and immunogenic reactions than bearing wear debris. 21 Especially the cobalt species has been under suspicion for its adverse effects, such as cardiac toxicity. 9 However, recent data from the National Joint Registry and NHS English hospital inpatient episodes about CoCr-containing THAs, did not find increased risk of all-cause mortality, or clinically meaningful heart outcomes, cancer, or neurodegenerative disorders into the second decade post-implantation. 22 One study has suggested that by using a ceramic femoral head, CoCrMo tribocorrosion may be mitigated. 23 Two other studies also concluded that ceramic head confers an advantage in trunnion fretting and corrosion. 14,24 Mid-term in vivo wear results (2-and 5-year follow-up based on radiostereometric analysis [RSA]) have not shown wear benefits of Oxinium femoral heads over conventional CoCrMo femoral heads. 11,16 There are reports that Oxinium femoral heads are susceptible to damage upon hip dislocation with subsequent increased polyethylene wear. 25 A recent case report showed catastrophic failure of an Oxinium-coated total knee prosthesis (Smith and Nephew) that resulted in metallosis with extra-articular extravasation. 26 In a systematic review and meta-analysis, Oxinium femoral heads did not lead to lower polyethylene wear rate or higher survival rate, when compared with CoCrMo femoral heads in patients treated with THA. 27 In a prospective study comparing whole blood Cr, Co and Ti ion concentrations in young, active patients (N = 36) undergoing primary THA with uncemented TiAlV stems and CoCrMo-alloy (32 mm), ceramic (32-40 mm), or Oxinium (32 and 36 mm) femoral heads, no statistically significant differences were found between the three cohorts, between 1 and 5 years follow-up (p = .3-.5). 13 The evidence of any benefits with the use of Oxinium femoral heads is therefore still sparse and further investigation regarding any benefit with Oxinium femoral heads is therefore needed. That said, clinical data from the Australian and UK registry indicate excellent results for Oxinium used in combination with well-performing cup and stems. 28,29 Furthermore, it is now well established that highly cross-linked polyethylene acetabular cups and liners significantly reduces wear compared to conventional polyethylene. 30 This has allowed for the use of larger head sizes to improve prosthetic stability. However, larger femoral head sizes may increase stress at the modular headneck junction, increasing the effective horizontal lever arm, potentially leading to fretting corrosion and metal debris with the use of cobaltalloy femoral heads. 24,31 In this study only small head sizes were used.
Thus, we cannot deduct our results to larger head sizes. However, in a recent study studying metal ion levels with modular 28 mm, 36 mm and 40 mm CoCrMo femoral heads, no difference in cobalt and chromium levels were found with increasing femoral head sizes after 10 years follow-up. 32 The median cobalt and chromium levels for the 28 mm head group were 0.66 μg/L and 0.68 μg/L, respectively. Similarly, no difference was found in cobalt and chromium blood levels in a randomized study comparing 32 mm, 36 mm and 40 mm modular CoCrMo heads at 1-and 2-years follow-up. 33 Neck taper flexural rigidity also affect trunnion corrosion and one study showed that lower stem flexural rigidity predicted stem fretting and corrosion damage in a ceramic head cohort, but not in the metal head cohort. 23 Care should therefore be taken to generalize the results to different stem designs and materials.
T A B L E 2 Patients that were revised before 10 years follow-up due to aseptic loosening of either the cup or stem or both. Three of these cases was included in a retrieval project and whole blood samples collected and the concentration (microgram/L) of Cr, Co, Ni, and Zr measured (study no 819, 831, and 848)  (Table 2). This is most likely attributed to fretting wear of the loose stem against the abrasive ZrO 2 containing cement mantle ( Figure 2B).
Our results highlight the importance of retrieval centers for understanding cause and effect relationships in THA. 17 Detecting differences in blood ion levels between groups is demanding due to small detection limits. 35 In the present study a single laboratory, with a highly sensitive methodology and adequate detection limits, analyzed all blood samples. This allows more subtle differences between groups to be detected.
Our study patients were quite old, so our results may not apply to younger and more active patients, who are increasingly becoming eligible for the THA procedure. The mechanical demands at the head-trunnion in a younger population interface may be higher. Thus, any mechanically driven corrosion and rise in blood ion levels may be underestimated in the population of this study.

| CONCLUSION
In this study of THAs with small femoral heads, no indication of severe trunnion corrosion was found. Although we found no significant difference in blood ion levels with the use of Oxinium or metal femoral heads, further investigation of periprosthetic tissues and retrieved implants may give insight to any differences in wear and corrosion mechanism at the trunnion-head interface between CoCrMo and Oxinium femoral heads.

FUNDING INFORMATION
The study was supported financially by Western Norway Regional Health Authority, Smith & Nephew Norway and Ortomedic AS. None of the sources of funding were involved in the planning or analysis of results.