Hydroxyapatite coating (HA) in femoral stems was first used in THA in the late 1980s . Biologically, HA coating promised to promote better prosthesis bone ingrowth . The promise of improved prosthesis fixation and some early supportive clinical study findings [3, 14] ensured HA-coated stems gained appeal for THA. In 2011, 50% of the THAs recorded in national Scandinavian joint replacement registries had HA-coated stems . The widespread adoption of this technology has resulted in many HA-coated prostheses from multiple manufacturers now being available on the international market despite a lack of long-term clinical evidence supporting the use of such devices. At least three meta-analyses and several large registry studies have found no difference in clinical benefit [4, 6, 11] or risk of revision surgery [4, 5, 8-10, 16]. Furthermore, no large-scale prosthesis-specific investigation into the comparative performance of HA and non-HA coating has been undertaken.
The addition of HA coating is also usually associated with increased cost. In Australia the additional cost associated with HA coating is approximately AUD 400 (approximately USD 285) per femoral stem. The lack of evidence supporting the superiority of HA coating and the additional cost associated with its use led us to conduct our analysis.
The purpose of this study was to evaluate the risk of revision in patients with HA-coated stems compared with non-HA-coated stems in elective primary THA undertaken for osteoarthritis. We addressed this question by evaluating the risk of all-cause revision of any component, revision for aseptic loosening of any component, and revisions of any component resulting from prosthesis dislocation, fracture, and infection. These latter three causes of revision were also studied to determine if there were any unintended advantages or consequences associated with the use of HA coating on femoral stems. Although it is accepted that the intent of HA coating is to improve fixation, it is important to establish if HA coating alters revision rates for reasons other than aseptic loosening. The impact of surface finish and materials used on infection risk is an area of increasing interest and study. Fracture risk may also be potentially altered as a result of increased prosthesis thickness, differences in prosthesis migration, or bone response to the prosthesis. Finally, the rate of dislocation could potentially be impacted by the ability to optimize initial positioning or the degree of early or late migration.
To address these questions, our analyses were conducted using both an aggregated sample of primary THAs for osteoarthritis as well as subgroup analysis of five individual THA prosthesis combinations where it was possible to compare HA- and non-HA-coated femoral stems of the same design used with the same acetabular component.
(1) What is the difference in risk of revision of THAs with HA-coated femoral stems compared with non-HA-coated using an aggregate analysis? (2) What is the difference in risk of revision of THAs with HA-coated femoral stems compared with non-HA-coated for five prosthesis types that used the same femoral and acetabular component combination but where the femoral stem had both HA and non-HA coating options?
Materials and Methods
This is a registry study using data collected by the Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) between September 1, 1999, and December 31, 2014. Details on the AOANJRR data collection procedures, validation, and outcomes assessment are described in its annual reports . In brief, the AOANJRR covers the entire Australian population, which was estimated to be 22.7 million people in 2012. It collects data using paper-based forms from both public and private hospitals throughout the country and in a recent validation study reported almost complete capture of cases within the country (99%). Since its implementation it has registered > 312,000 primary conventional THAs . All data used in this study were obtained from the AOANJRR annual report files.
The overall study population was comprised of patients > 18 years old with cementless primary unilateral conventional THA for osteoarthritis (N = 147,645). The study cohorts used for the subgroup analyses included patients in the overall population who had component combinations where the same acetabular component was used and the femoral stems had a HA version and a non-HA version and at least 300 cases were done with each combination. These selection criteria limited the femoral stem and acetabular component combinations to five subgroups: VerSys/Trilogy (Zimmer, Warsaw, IN, USA), Mallory-Head/Mallory-Head (Biomet, Warsaw, IN, USA), SL-Plus/EP Fit-Plus (Smith & Nephew, Andover, MA, USA), Taperloc/Exceed (Biomet), and Taperloc/Mallory-Head (Biomet).
Whether the femoral stem was coated with HA or not was the main exposure of interest of this study. This was obtained from the intraoperative AOANJRR form device information and subsequent device attribute categorization conducted by the Registry.
The main outcomes of interest in this study were all-cause revision of any component and all-cause revision for stem-only and acetabular component-only revisions. Revision was defined as any reoperation where a prosthesis was replaced. Secondary outcomes of interest included revision (any component, stem, and acetabular component) for aseptic loosening, revision (any component) for prosthesis dislocation, revision (any component) for fracture, and revision (any component) for infection. Age, sex, bearing surface, femoral head size, and HA coating of the acetabular component were evaluated as possible confounders of the HA stem coating and risk of revision relation. The bearing surface of the device construct was categorized into ceramic-on-ceramic, ceramic-on-highly crosslinked polyethylene, metal-on-noncrosslinked polyethylene, metal-on-highly crosslinked polyethylene, ceramicized metal-on-crosslinked polyethylene, and other. Head size was categorized into < 32 mm, 32 mm, 36 mm, and ≥ 36 mm.
Of the 147,645 THAs identified, 119,673 (81.1%) had HA-coated stems. Overall, 57.0% (N = 68,237) of patients with HA-coated stems were 65 years old or older and 50.6% (N = 60,539) were female, whereas 59.7% (N = 16,713) of patients without HA-coated stems were 65 years or older and 51.0% (N = 14,271) were female. The most common bearing surface used in HA-coated stem THAs was ceramic-on-ceramic (39.1% [N = 46,846]), whereas in cases without HA-coated stems, it was metal-on-crosslinked polyethylene (41.0% [N = 11,458]) (Table 1). In the overall cohort, there were 164 different femoral stems used, which led to 966 different cup-stem combinations. The most commonly used stems in this cohort of patients were DePuy’s (Raynham, MA, USA) Corail (18.7% [N = 27,604]), Smith & Nephew’s Synergy (7.5% [N = 11,079]), Medacta’s (Castel San Pietro, Switzerland) Quadra-H (6.3% [N = 9337]), Stryker’s (Kalamazoo, MI, USA) Accolade I (6.3% [N = 9308]), and Stryker’s Secur-Fit (5.9% [N = 8650]). For the subgroups evaluated, where the same construct was available in an HA- and non-HA-coated version, the numbers in the groups ranged from highest in the VerSys/Trilogy (N = 3924) group to the Mallory-Head/Mallory-Head (N = 2538), SL-Plus/EP Fit-Plus (N = 2028), and Taperloc/Exceed (N = 1668) and lowest in the Taperloc/Mallory-Head (N = 1240) group (Table 2). There were 5015 (3.4% of 147,645) revisions during the study period (Table 3) for the overall cohort and 173 (0.54%) in the VerSys/Trilogy group, 112 (0.58%) in the Mallory-Head/Mallory-Head group, 90 (0.66%) in the SL-Plus/EP Fit-Plus group, 38 (0.82%) in the Taperloc/Exceed group, and 47 (0.68%) in the Taperloc/Mallory-Head group (Table 4).
Statistical analysis frequencies, proportions, means, and SDs were used to describe the patient population. All analyses were conducted separately for the entire population (Question 1) and the five prosthesis combination subgroups (Question 2): (1) VerSys/Trilogy; (2) Mallory-Head/Mallory-Head; (3) SL-Plus/EP FIT-Plus; (4) Taperloc/Exceed; and (5) Taperloc/Mallory-Head. This subgroup analysis was conducted to investigate whether an interaction (that is, effect modification) between prosthesis combinations and HA coating existed. Kaplan-Meier survival curves were used to visualize cumulative revision rates in the patients included in the study and Cox regression models to evaluate the HA stem coating’s association with time to revision. Proportional hazard assumptions were checked using log-log graphs (log of the negative log of survival versus log of survival time). When proportional hazard assumptions were not met, time-dependent exposure variables were created and are specified in the final presentation of the estimates. Cutoff periods for time-dependent variables were extrapolated from the survival curves. Models attempted to account for clustering of observations by hospital using the robust sandwich estimate of Lin and Wei for the covariance matrix . Crude (data not shown) and adjusted revision estimates are presented. Final models were adjusted for age, sex, bearing surface, and HA coating of the acetabular component. Head size was not included in the final models because it was determined not to be a confounder as a result of similar head size distribution in all groups analyzed. Effect modification by age (< 65 versus ≥ 65 years old), sex, and HA coating of the acetabular component was also evaluated. Risk estimates were not modified by these strata data and therefore data are not shown. Hazard ratios (HRs), 95% confidence intervals (CIs), and Wald chi-square p values are reported. Collinearity between included variables was investigated and tolerance values > 0.10 were used as the cutoff period. Alpha = 0.05 was used as the threshold for statistical significance and all tests conducted were two-sided in the analyses conducted. The data were analyzed using SAS software, Version 9.3 (SAS Institute Inc, Cary, NC, USA).
The cumulative revision probability at 5 years postsurgery ranged from 1.9 (1.2-2.8) in Mallory-Head/Mallory-Head to 4.8 (3.5-5.6) for SL-Plus/EP FIT-Plus in non-HA stems and from 2.4 (1.5-5.6) in Taperloc/Mallory-Head to 3.7 (2.7-4.9) in Mallory-Head/Mallory-Head for HA-coated stems (Table 5). We found a lower risk of any component revision for any reason after 6 months (HR, 0.83; 95% CI, 0.72-0.95; p = 0.007) (Fig. 1) and for loosening at any time (HR, 0.73; 95% CI, 0.61-0.88; p < 0.001) in THAs with HA-coated stems compared with those without HA-coated stems when adjusting for the confounders of age, sex, bearing surface, and cup HA coating (Table 6). We also noted a lower risk of stem-only revisions for any reason (HR, 0.70; 95% CI, 0.57-0.85; p < 0.001) and stem-only revisions for loosening (HR, 0.62; 95% CI, 0.44-0.88; p = 0.006) within 1.5 years and after 1.5 years (HR, 0.44; 95% CI, 0.32-0.61; p < 0.001) in THAs with HA-coated stems compared with those without HA-coated stems (Table 6).
In our analyses of groups with the same cup and HA-coated and noncoated stem versions, we found that the effect of HA coating on revision risk varied by prosthesis design. Only the SL-Plus/EP Fit-Plus group showed a lower risk of any-cause revision between 6 months and 3 years (HR, 0.32; 95% CI, 0.15-0.70; p = 0.024), for any component revision for loosening between 0 and 7 years (HR, 0.17; 95% CI, 0.06-0.44; p < 0.001) as well as for stem revisions for any reason (HR, 0.23; 95% CI, 0.10-0.55; p < 0.001) and stem revision for loosening (HR, 0.07; 95% CI, 0.02-0.33; p < 0.001) for THAs with HA stems compared with non-HA stems. However, a higher risk of revision was observed for all-cause revisions in the SL-Plus/EP Fit-Plus group in the first 6 months after surgery (HR, 2.88; 95% CI, 1.15-7.21; p = 0.026) for the HA-coated stem group (Table 6; Fig. 2). Additionally, in the VerSys/Trilogy subgroup, a lower risk of revision for dislocation (HR, 0.41; 95% CI, 0.18-0.93; p = 0.033) as well as a higher risk of fracture (HR, 4.26; 95% CI, 1.86-9.74; p < 0.001) were observed in THAs with HA stems compared with non-HA stems (Table 5; Fig. 3). In the Mallory-Head/Mallory-Head subgroup, only the risk of stem-only revisions for any reason was significantly different between the groups with HA-coated stems having a higher risk (HR, 5.30; 95% CI, 1.30-21.5; p = 0.020) (Fig. 4). In the Taperloc/Mallory-Head group, there was a higher risk of stem-only revisions for loosening in the HA-coated stemmed THAs than those without HA-coated stems (HR, 2.53; 95% CI, 1.55-4.13; p < 0.001) (Fig. 5). No differences in revision risks for HA-coated compared with noncoated stems were observed in the Taperloc/Exceed subgroup (Fig. 6).
To date, no evidence of superior clinical and revision outcomes has been shown with the use of HA-coated stem components and no large-scale prosthesis-specific investigation into HA coating has been carried out. In this registry analysis of primary elective uncemented THAs, there was a varying risk of revision associated with the use of HA-coated femoral stems and that variation depended on the specific stem/acetabular component combination used. Of the five stem/acetabular component combinations that had both HA and non-HA versions of the femoral stem analyzed in this study, only one had a lower risk of revision associated with HA coating of the femoral stem. Understanding that HA coating of the femoral stem is not universally beneficial and that the risk of revision varies by prosthesis design is important when considering the optimal prosthesis choice.
There are a number of limitations to this study. This is an observational study and therefore only associations and not causation between the exposure of interest and the outcomes can be inferred from our findings. Also, the Registry does not have clinical and radiologic postoperative assessment of the procedures and therefore only the outcome of revision was evaluated, limiting the inferences of this study to only this outcome. Another limitation is the possible underestimation of the effect of HA coating on the risk of revision in the overall cohort. Because the combined outcome of multiple components was evaluated in the overall assessment of HA coating, it is possible that poorly performing stems, regardless of whether HA-coated or not, may dilute or alter the effects observed. It is also possible that other factors (that is, residual confounding) such as other patient characteristics, femoral stem attributes, or even postoperative rehabilitation variables such as postoperative activity, which are not included in our analysis, could possibly impact the analysis. We attempted to minimize the impact of the known confounders by either restricting the sample to remove certain confounders or including all remaining covariates in our multivariable analysis. A subgroup analysis was also undertaken comparing HA and non-HA versions of a single femoral stem/acetabular component combination. The cohort size for these analyses had a minimum number restriction of 300 procedures to ensure that a sufficiently large sample size was available to conduct the planned analysis. Consequently, a number of femoral stems with HA coating were not evaluated in this subanalysis. Furthermore, even with the large numbers of this study, we recognize that some of the subgroup analyses conducted had fewer events and the estimations for some combinations have wider CIs. The impact of this is that slight differences in the risk of revision in these subgroups may not be detected because of limited power.
Our study strengths include the prospective and complete data capture of patients undergoing THAs in Australia, the large number of prostheses evaluated, the ability to isolate the effect of HA coating in our analysis, and evaluation of several revision types and indications. No selection bias, which can typically be introduced by insurance coverage status, hospital setting, surgeon training, or experience level, or even patient self-selection, is present in this national, fully captured cohort of patients. Additionally, our findings are generalizable to all surgeons and patients undergoing THAs in Australia. As a result of the predefined data capture model, definitions, and the prospective data collection of the registry used for this study, we can also assure high data integrity of the information presented. We were also able to evaluate the general effect of HA coating in a large number of prostheses and not just on those used in high frequency, which could be an indication for good performance and therefore a self-selection bias issue. The size of the sample also enabled the subgroup analysis to be undertaken. Restricting the analysis to the same cup-stem constructs removed the effect that cup design could have on the performance of the stem. Furthermore, restricting the subgroup analysis to specific femoral stem designs with HA and non-HA versions enabled isolating the effect of HA coating for those stems and allowed the assessment of whether the effect of HA coating varied depending on the type of femoral stem. Finally, also as a result of the large numbers of THAs in our cohort, it was possible to evaluate different revision types and indications for revision allowing us to identify and understand in greater detail the association of HA coating and certain outcomes.
Using the entire cohort in the AOANJRR that fit our study criteria, we compared the risk of revision by HA coating versus non-HA coating in femoral stems. This was done to compare the overall risk estimates in our cohort with both our own subgroup analysis, which focuses on specific prostheses that have a HA-coated version and non-HA-coated version, and compare it with existing reports in which this higher level comparison has been made. Our general findings from this overall analysis, which likely does not properly estimate the effect of HA coating on the risk of revision surgery because of the inclusion of prostheses with and without this property, found that HA stem coating is associated with a lower risk of all-component revision for any reason, stem revision for any reason, and stem revision for loosening. These findings differ from other clinical studies looking at stem survivorship [4, 6] and the Scandinavian registry studies that have also evaluated this question [8-10, 16]. In Gandhi et al.’s  review of nine clinical studies with 1764 patients, the risk of revision for aseptic loosening in nine HA-coated femoral stem designs was not different compared with non-HA-coated stems. Similarly, in Chen et al.’s  meta-analysis of 21 studies, 16 of which had survival information on 3935 patients, the authors also reported no difference in stem survivorship. In agreement with these clinical studies, a collaboration study among Sweden, Norway, Denmark, and Finland with 116,069 THAs reported that the risk of stem revision was also not significantly different in patients with HA-coated stems compared with those without HA coating . This was also consistent with previous reports from these countries individually [5, 9, 10, 16]. The different numbers of prostheses evaluated and the specific prostheses included in our study compared with other studies are likely the major reasons for this difference in risk of revision surgery associated with HA stem coating found in these studies. Hailer et al.’s analysis, which included 22 stems, had the largest number of stems evaluated. However, Bimetric (Biomet) femoral stems, which were not used in our cohort, made up 37% of the total number of cases evaluated by Hailer et al. . These are high-performing stems, with versions with HA coating and without HA coating, that have been used with good results in Denmark, Sweden, and Finland since the early 1990s. The second most common type of stem in the Hailer et al. cohort was Corail (DePuy), which was the most commonly used stem in our general study as well, and it is only available with a HA coating. In addition to thus including two high-performing HA-coated stems, Hailer et al. included only stems that had been used at least 500 times in an attempt to limit selection bias from poor-performing stems uncommonly used. However, at the same time, this limited their ability to evaluate HA coating in other stems rarely used.
In our subgroup analysis, which only included stems manufactured with and without HA coating options, we restricted our comparison to two exact prosthesis constructs in an attempt to remove the possible confounding introduced by different prosthesis designs and constructs. We found that HA coating performed differently depending on the device type. In the comparison with the non-HA-coated version of the stem/cup combination SL-Plus/EP Fit-Plus (Smith & Nephew), the benefits of HA coating for revision (including for any revision, revision for loosening, stem-only all-cause revisions, and stem-only for aseptic loosening) were obvious, but a shorter early risk of revision for any reason was observed in the components with HA coating (0-6 months postsurgery). This earlier revision risk does not seem associated with increased revision of the stem (as a result of the low number of events) during that timeframe, but we are unable to determine this with certainty because of the smaller sample size once we evaluated the cause-specific revisions. A prior small clinical study (n = 126) had shown favorable osseointegration through radiologic assessment in SL-Plus (Smith & Nephew) stems with HA coating at 5 years, which is supportive of our observations of improved outcomes with these stems . The largest subgroup in our analysis, the VerSys/Trilogy (Zimmer) combination, had a higher risk of revision for fractures when HA coating was used. A 2012 case-series report from Taiwan  reported on a greater than expected incidence of fracture on fully bead-coated VerSys (Zimmer) stems, which, although not specific to HA-coated stems, could be indicative of a stem design femoral fit or technical issue that could lead to this complication. In the case of the Mallory-Head/Mallory-Head (Biomet) combination, the risk of stem revision was actually higher in the HA-coated stem version than in the non-HA-coated versions. This is in contrast with the findings of two smaller clinical studies (n = 54 and n = 61) that found these Mallory-Head (Biomet) stems to have similar performance in regard to stem revision [2, 18]. However, it is possible these two studies were underpowered to detect this difference. In the Taperloc/Exceed (Biomet) combination, we did not find a difference in revision risk in HA-coated stems and those without coating, but in the Taperloc/Mallory-Head combination, we found a higher risk of stem-only revisions for loosening. A small cohort study has reported no difference in risk of all-component or stem revision between coating status of Taperloc (Biomet) stems . None of the HA-coated stems evaluated in this subgroup analysis were evaluated by the registry studies previously mentioned and therefore no comparison to our estimates can be made [10, 16].
In this nationally representative sample of THAs, we found that HA coating in femoral stems is associated with an overall lower risk of revision for any reason, lower risk of revision for aseptic loosening, and lower risk of stem revisions. However, when prosthesis-specific analysis was undertaken, only one of the THA combinations had a lower rate of revision when the femoral stem was HA-coated. This suggests that there is an effect of HA, which may be related to the design of the prosthesis and/or perhaps the manufacturing and application of the HA. Surgeons should be aware if the addition of HA to their choice of stem is beneficial in view of the associated increased cost and the variable effect of HA coating with specific femoral stem designs.
We thank the hospitals, orthopaedic surgeons, and patients whose contributions made this work possible.
1. Australian Orthopaedic Association National Joint Replacement Registry. Annual report. 2014. Available at: https://aoanjrr.dmac.adelaide.edu.au/documents/10180/172286/Annual%20Report%202014
. Accessed June 1, 2015.
2. Camazzola D, Hammond T, Gandhi R, Davey JR. A randomized trial of hydroxyapatite-coated femoral stems in total hip arthroplasty: a 13-year follow-up. J Arthroplasty. 2009;24:33-37.
3. Chambers B, St Clair SF, Froimson MI. Hydroxyapatite-coated tapered cementless femoral components in total hip arthroplasty. J Arthroplasty.
4. Chen YL, Lin T, Liu A, Shi MM, Hu B, Shi ZL, Yan SG. Does hydroxyapatite coating have no advantage over porous coating in primary total hip arthroplasty? A meta-analysis. J Orthop Surg Res.
5. Eskelinen A, Remes V, Helenius I, Pulkkinen P, Nevalainen J, Paavolainen P. Total hip arthroplasty for primary osteoarthrosis in younger patients in the Finnish arthroplasty register. 4,661 primary replacements followed for 0-22 years. Acta Orthop.
6. Gandhi R, Davey JR, Mahomed NN. Hydroxyapatite coated femoral stems in primary total hip arthroplasty: a meta-analysis. J Arthroplasty. 2009;24:38-42.
7. GeesinkRG. Osteoconductive coatings for total joint arthroplasty. Clin Orthop Relat Res.
8. Hailer NP, Lazarinis S, Makela KT, Eskelinen A, Fenstad AM, Hallan G, Havelin L, Overgaard S, Pedersen AB, Mehnert F, Karrholm J. Hydroxyapatite coating does not improve uncemented stem survival after total hip arthroplasty! Acta Orthop. 2015;86:18-25.
9. Hallan G, Lie SA, Furnes O, Engesaeter LB, Vollset SE, Havelin LI. Medium- and long-term performance of 11,516 uncemented primary femoral stems from the Norwegian arthroplasty register. J Bone Joint Surg Br. 2007;89:1574-1580.
10. Lazarinis S, Karrholm J, Hailer NP. Effects of hydroxyapatite coating on survival of an uncemented femoral stem. A Swedish Hip Arthroplasty Register study on 4,772 hips. Acta Orthop. 2011;82:399-404.
11. Li S, Huang B, Chen Y, Gao H, Fan Q, Zhao J, Su W. Hydroxyapatite-coated femoral stems in primary total hip arthroplasty: a meta-analysis of randomized controlled trials. Int J Surg. 2013;11:477-482.
12. Lin DY, Wei LJ. The robust inference for the Cox proportional hazard model. J Am Stat Assoc. 1989;84:1074-1078.
13. Lu HC, Lin CL, Chang CW, Lai KA. Fracture of VerSys fully bead-coated long femoral stems: report on four fractures in 41 hips. Kaohsiung J Med Sci. 2012;28:345-349.
14. Oosterbos CJ, Rahmy AI, Tonino AJ, Witpeerd W. High survival rate of hydroxyapatite-coated hip prostheses: 100 consecutive hips followed for 10 years. Acta Orthop Scand. 2004;75:127-133.
15. Parvizi J, Sharkey PF, Hozack WJ, Orzoco F, Bissett GA, Rothman RH. Prospective matched-pair analysis of hydroxyapatite-coated and uncoated femoral stems in total hip arthroplasty. A concise follow-up of a previous report. J Bone Joint Surg Am. 2004;86:783-786.
16. Paulsen A, Pedersen AB, Johnsen SP, Riis A, Lucht U, Overgaard S. Effect of hydroxyapatite coating on risk of revision after primary total hip arthroplasty in younger patients: findings from the Danish Hip Arthroplasty Registry. Acta Orthop. 2007;78:622-628.
17. Steens W, Schneeberger AG, Skripitz R, Fennema P, Goetze C. Bone remodeling in proximal HA-coated versus uncoated cementless SL-Plus® femoral components: a 5-year follow-up study. Arch Orthop Trauma Surg. 2010;130:921-926.
18. Yee AJ, Kreder HK, Bookman I, Davey JR. A randomized trial of hydroxyapatite coated prostheses in total hip arthroplasty. Clin Orthop Relat Res. 1999;366:120-132.