Adverse local tissue reactions (ALTRs) and elevated serum metal ion levels secondary to fretting and corrosion at head-neck junctions in modular THA designs have raised concern in recent years. ALTRs can lead to high levels of morbidity including pain, hip dysfunction, and, in the most devastating cases, major tissue loss resulting in reduced salvage options at revision surgery. Reporting of ALTRs as a result of THA fretting and corrosion has increased over the last decade with a clinical prevalence rate of 1% to 3% [4, 16, 22] with estimates as high as 4.7%  for specific components. It is unclear whether this is recognition of a previously underdiagnosed problem that may be unique to certain patient populations or if new implant designs are contributing to this increase [7, 17]. Over the past few years, many surgeons in the orthopaedic community have transitioned from using cobalt-chrome alloy (CoCr) femoral heads to the use of ceramic heads for hip replacement. This transition appears to be motivated by concerns regarding trunnion corrosion and the potential for ALTR . Although some institutions have reported ALTR with CoCr-on-polyethylene bearing surfaces , the prevalence remains uncertain, particularly because the vast majority of THAs done historically used CoCr alloy heads. Despite the fact that millions of hip replacements were previously performed with CoCr-on-polyethylene bearings, ALTR was not recognized until metal-on-metal bearings were reintroduced for hip replacement in the late 1990s .
The majority of studies concerning THA fretting and corrosion have been limited to analysis of revised THAs , although the implants considered in these studies were often removed for diagnoses other than ALTRs, including aseptic loosening, infection, instability, fracture, implant malpositioning, leg-length discrepancy, pain, polyethylene wear, or osteolysis [10, 20, 27]. A recent retrieval study observed less damage in a small cohort of THAs retrieved at autopsy than THAs from revision surgery . Overall, little is known about the condition of head-neck tapers in well-functioning THAs with respect to fretting and corrosion.
Corrosion at the head-neck junction has been shown to weaken the mechanical strength of femoral stems and taper locking strength, leading to fatigue fracture  and dissociation [2, 29]. Femoral head pull-off force is considered an indirect measurement of taper locking strength and has been positively associated with stem trunnion damage . Assessment of trunnion fretting and corrosion damage of retrieved implants continues to identify risk characteristics of THA designs and modular connections such as trunnion geometry and head material [13, 19, 20]. Depending on the metallurgy of the implant and environment, corrosion at the head-neck junction can take any of a number of manifestations such as etching, fretting, and pitting [8, 9, 18, 24]. These different forms of corrosion display different visual patterns, some of which require high magnification such as scanning electron microscopy (SEM) to discern [8, 9, 18].
Regarding a cohort of well-functioning autopsy-retrieved modular THAs, we asked: (1) Does trunnion geometry or femoral head material affect the pull-off force of the femoral head? (2) Is there a relationship between trunnion damage and length of implantation time, head diameter, and neck length? (3) Does visual damage scoring accurately determine the presence or absence of corrosion on CoCr trunnions?
Materials and Methods
Sixty-six well-functioning primary femoral stems and engaged femoral heads were retrieved at autopsy from 53 patients at Anderson Orthopaedic Research Institute from 1998 to 2014. Patients consented to participate in the postmortem implant retrieval program after their first postoperative visit. Coordination among the institution, next of kin, and the funeral home allowed for implant retrieval in the preparation room at the funeral home. Six femoral stems of five designs were excluded for low group size, and four femoral stems were excluded for insufficient head-stem clearance required for pull-off testing. Of the 56 THA implants included for analysis, 33 implants were retrieved from 25 men and 23 implants from 21 women (Table 1). The median patient age at the time of the index procedure was 69 years (range, 20-86 years); the median length of implantation (the time that the THA components had been implanted) was 10 years (range, 1-24 years).
Thirteen implants were alumina-on-polyethylene bearings, and 43 were metal-on-polyethylene bearings. All implant stems were one of three CoCr designs: AML®, Prodigy®, or Solution™ (DePuy Synthes, Warsaw, IN, USA) with either a 12/14 or a 14/16 trunnion. All three femoral stem designs are very similar and feature the same type of cylindrical distal stem geometry with a porous surface consisting of sintered beads for bone ingrowth. Femoral heads were either CoCr (n = 43) or alumina (n = 13) (Table 1).
Each head-neck junction was disassembled on a uniaxial load frame (MTS, Eden Prairie, MN, USA) at a rate of 0.05 mm/s to determine the pull-off force per ASTM F2009-00 . After pull-off, the mating surfaces of the femoral head taper and the stem trunnion were cleaned gently with 41% isopropyl alcohol to remove any extraneous debris. The taper and trunnion surfaces were then visually examined by two independent observers (CNK, AW) using an optical stereomicroscope with x 6 to x 10 magnification (Wild Type 376788, Heerbruug, Switzerland) to assess the extent and severity of damage. When scores differed, graders regraded simultaneously and reached a consensus (this occurred in three cases). Both taper and trunnion surfaces were scored for fretting and corrosion in eight quadrants with proximal and distal regions subdivided into posterior, anterior, superior, and inferior regions. Surfaces were visually scored using the Goldberg scoring system  with a modification of a minimum score of 0 corresponding to the absence of damage and maximum score of 3 corresponding to severe fretting or corrosion. Each implant was scored as a single unit with a maximum total score of 96 (two surfaces x eight regions x two modes of damage x maximum score of 3).
To corroborate visual corrosion scores and to determine if fretting and corrosion were evident at higher magnifications on trunnions that showed little or no damage with stereomicrosocopy, three stem trunnions were selected for SEM analysis: the trunnion with the largest corrosion score and two trunnions that were free of damage; the two latter trunnions were matched to the most corroded trunnion on the basis of trunnion type, stem design, and head material. The length of implantation for the two trunnions free of damage, 8 years each, was shorter than the damaged trunnion, 11 years. The SEM images were collected using a Zeiss Supra® 55VP SEM (Carl Zeiss Microscopy GmbH, Jena, Germany). Secondary electron mode SEM images were qualitatively assessed to identify modes of corrosion such as intergranular corrosion, etching, or pitting. Apparent anomalies (eg, cracks, pitting, grain boundaries) on the surface were identified as regions of interest from SEM scans and chosen for energy dispersive x-ray analysis (EDAX) to determine the elemental composition of the surface material. Large concentrations of oxygen (oxides) and molybdenum (Mo) were considered indicative of corrosion products [8, 28].
Owing to their nonparametric distribution, pull-off force is specified as median (range). Kruskal-Wallis one-way analysis of variance on ranks was used to compare pull-off force and damage score between implants of different trunnion designs and head materials with Dunn’s method for pairwise, multiple, post hoc comparison. We used Spearman’s correlation to determine if there was a relationship between damage scores and length of time of implantation, head diameter, and neck length. Significance was set at p < 0.05. All statistical analyses were performed using Sigma Plot 12 (Systat Software, Inc, San Jose, CA, USA).
The median pull-off force for the groups of alumina-12/14 and alumina-14/16 couples were 3127 (2320-6992) N and 2670 (1095-7919) N, respectively, and 2255 (1332-5939) N and 2812 (1655-4246) N for groups of CoCr-12/14 and CoCr-14/16 couples, respectively (Fig. 1). With the numbers available, no difference was found in the pull-off forces among the four implant groups based on femoral head material and trunnion geometry (p = 0.132). The pull-off force for the implant with the highest damage score was 2474 N, but this force was not different than the median pull-off force of 2519 N for the other 55 specimens (p = 0.980). The pull-off force ranged from 1095 N to 7919 N among the 56 THAs with a median value of 2497 N.
A positive correlation was found between damage score and length of implantation (ρ = 0.543, p < 0.001). No correlation between damage score and either head diameter or neck length was found (ρ = -0.012, p = 0.930 and ρ < 0.001, p = 0.995, respectively). Of the 56 implants studied, 39 had a score of 0, 16 demonstrated damage scores of 7 or lower, and only one specimen demonstrated a higher score of 46 (Fig. 2). Of the 39 implants that showed no evidence of fretting or corrosion, 29 implants had CoCr heads (67% of the 43 CoCr heads in the cohort) and 10 implants had alumina heads (77% of the 13 alumina heads in the cohort). The median modified Goldberg damage score was 0 (range, 0-46) for the entire cohort (maximum possible score of 96). A difference in damage score was found among the four head material-trunnion geometry combinations (p = 0.006) (Fig. 2). However, as a result of variable group sizes, no differences were found after pairwise multiple comparison procedures.
Visual assessment of the three selected trunnions at x 10 magnification noted original machining lines created during fabrication visible on all three trunnions with a large band of disruption visible on the highly damaged trunnion (Fig. 3A). SEM images of the trunnions showed surface quality differences between the highly and the two minimally damaged implants (Fig. 3B). The presence of corrosion products such as oxides or Mo-rich zones [8, 9, 28] on the highly damaged trunnion was verified by EDAX, which showed large concentration peaks of oxygen and Mo (Fig. 3C). Evidence of corrosion products was absent from minimally damaged trunnions (Fig. 3C). The pattern of corrosion on the highly damaged trunnion was identified as intergranular corrosion (Fig. 4).
Concerns regarding the risk of ALTR with fretting and corrosion at the head-neck junction motivate investigations of retrieved THA. However, a full understanding of the prevalence of fretting and corrosion is lacking because little is known about the condition of head-neck tapers in well-functioning THAs. In our cohort of 56 THAs retrieved at autopsy, no differences were found in pull-off forces as a result of trunnion geometry or femoral head material. Evidence of major fretting and corrosion damage was rare with 70% of the implants showing no signs of damage. The remaining cohort demonstrated mild fretting or corrosion with no correlation between damage and neck lengths (0-15.5 mm) or head diameters (28-36 mm). A positive correlation was found between damage score and length of implantation. Only one of the 56 head-neck junctions showed visual evidence of severe damage resulting from corrosion. The presence of corrosion products was confirmed with semiquantitative data by SEM imaging and EDAX as well as the absence of corrosion in two trunnions, each with a damage score of 0.
This study has several limitations, including those inherent to a retrospective study design. The patients included in this study may not be representative of all populations receiving THAs, particularly patients who are diagnosed for revision. All THAs studied incorporated CoCr-CoCr or alumina-CoCr head-neck material couples, which likely reduce the amount of fretting or corrosion that would be observed compared with other dissimilar material couples, which have been shown to be more prone to fretting and corrosion [13, 27]. The study population also included only two trunnion geometries from a single manufacturer, which reflected the historic preferences of the institution at which all of the components had been implanted. Thus, we were unable to examine trunnions of smaller sizes, different geometries, or different materials that may be more susceptible to fretting and/or corrosion [14, 26, 27]. Because fretting and corrosion in retrievals correlate with various trunnion design, surface, and material factors [7, 14, 19, 26], the interpretation of our results is limited to implants of similar design and not to the multitude of currently used implants. Although pull-off force can be regarded as a measure of mechanical integrity for modular head-neck assemblies, the impaction forces for the specimens in this study are unknown, because they were not quantified at the time of surgery. Finally, although this is the largest published cohort of autopsy-retrieved well-functioning THAs to our knowledge, it still represents a relatively small cohort. Although the incidence of ALTR is low, the sample size of this cohort may not be sufficient to identify the true incidence of fretting and corrosion damage in well-functioning THAs. Similarly, only a small subset was tested with further imaging to confirm visual damage scores.
The pull-off force is considered a measure for taper strength and has been shown to vary with impaction force and head-neck taper surface properties [11, 21, 25]. A recent study by Higgs et al.  found a positive correlation between pull-off force and trunnion damage in a cohort of THAs retrieved at revision surgery (N = 93) and autopsy (N = 16). The median pull-off force for the autopsy-retrieved cohort (median = 2.5; interquartile range [IQR] = 1.7 kN) is comparable to that presented in the current study (median = 2.5; IQR = 1.1 kN). The study also demonstrated correlations between head size, trunnion angle, and taper grooves and pull-off force for the total cohort . These relationships may be better observed in their total cohort of 109 THAs with six different manufacturers represented.
Previous studies have implicated several factors contributing to fretting and corrosion at head-neck junctions, including trunnion geometry, head-trunnion material couple, femoral head diameter, neck length, force of head impaction at the time of surgery, and length of implantation [15, 23]. In a multicenter retrieval study including 231 modular femoral components, Goldberg et al.  demonstrated that similar-alloy couples conferred an advantage over mixed-alloy couples with respect to reducing fretting and corrosion. Even with this advantage, 28% of similar-alloy heads and 15% of similar-alloy necks demonstrated evidence of moderate or severe corrosion. Kurtz et al.  found that ceramic-metal material couples demonstrated less severe fretting and corrosion than metal-metal couples in a matched cohort of 100 retrieved modular femoral implants, although evidence of fretting and corrosion was present on most implants, regardless of material couple. Similar results were reported by other investigators [15, 26]. In our cohort of well-functioning THAs with different head-neck couples, 30% of the CoCr-CoCr alloy couples and 23% of the alumina-CoCr couples showed mild evidence of fretting or corrosion; no difference was found among head materials with the available population. Based on a matched cohort analysis of 23 retrieved implants with 28-mm diameter CoCr heads and 23 retrieved implants with 32-mm diameter CoCr heads, Del Balso et al.  found that 32-mm heads were associated with more severe fretting, but not corrosion. Dyrkacz et al.  investigated the difference between 36-mm heads (15 retrievals) and 28-mm heads (59 retrievals), finding that 36-mm heads demonstrated more corrosion than 28-mm heads. Conversely, in a previous study from our laboratory, the role of head diameter was investigated in a large cohort of 154 retrieved implants with head diameters ranging from 22 to 44 mm and found to be unrelated to fretting and corrosion scores . In our current study, we also found no correlation between head diameter and damage scores. Increased head size would impart larger bending moments to the modular connection, thus potentially increasing fretting and subsequent corrosion. The disagreement among studies suggests that other factors are also affecting fretting and corrosion damage at head-neck junctions. For example, both in vitro and retrieval studies have indicated that longer neck lengths confer a higher risk for fretting and corrosion. In our well-functioning postmortem THA cohort, no correlation was found between neck length and damage score, although the single specimen with severe corrosion had a 15.5-mm neck length, the longest in our study population. Many retrieval studies have indicated that increased implantation length correlates with increased damage from fretting or corrosion [3, 5, 27]. Although damage scores remained low, this study also demonstrated a slight increase in fretting and corrosion score with implantation length.
We can compare results from the current study of autopsy retrievals with our previously published results from damage scoring of fretting and corrosion of head-neck junctions on retrieved femoral components from revision THAs (Fig. 5) . This comparison underscores the low levels of damage seen on the autopsy retrievals and raises questions about whether the observations from implant retrievals obtained at the time of revision can be generalized to well-functioning THAs. The low incidence and severity of fretting and corrosion damage observed in the presented study, affecting correlation with THA component properties, may be reflective of well-functioning THAs . The results from this study suggest that patients with well-functioning hip replacements using polyethylene bearing surfaces with ceramic or CoCr heads appear to be at low risk for trunnion corrosion for the specific type of CoCr alloy stems and trunnion geometries that we analyzed. Lastly, we would encourage continued analysis of postmortem retrievals from willing patients and their families because it provides additional perspective beyond the findings from revision-retrieved components.
Of the well-functioning THA cohort, one implant demonstrated severe corrosion. The corroded implant was a 28-mm CoCr skirted head implanted on a 12/14 trunnion with a +15.5-mm neck length. The patient was initially trialed with a +5 neck length intraoperatively, but the hip was found to be unstable, so a longer neck was used. Without pathologic specimens for histologic analysis, ALTR cannot be confirmed or denied, although no evidence of femoral osteolysis was found. Further image analysis of this trunnion with SEM/EDAX demonstrated both presence of corrosion products and identification of corrosion manifestation. Oxides and Mo-rich zones have been found on the surface of corroded implants [8, 9]. In mechanically assisted crevice corrosion, fracture of the passive oxide film exposes the underlying metal to further corrosion. The presence of oxides as identified by EDAX may be fractures of oxide film. Mo-rich zones have been shown to be isolated during corrosion at the grain boundary regions during intergranular corrosion . Published examples of the manifestations of corrosion were used to characterize the corrosion found on the damaged trunnion as intergranular corrosion . SEM and EDAX results for the trunnions without damage corroborate the visual damage scoring of the absence of corrosion. From these results, we believe it is reasonable for future studies relying on visual damage scoring to use further imaging on only a subset of the examined cohort.
In summary, we demonstrated a low incidence of fretting and corrosion in a cohort of 56 well-functioning CoCr-CoCr or alumina-CoCr head-neck couples retrieved at autopsy after up to almost 25 years of implantation. Despite the variations in the pull-off forces among the 56 postmortem-retrieved specimens included in this study, only one specimen demonstrated extensive evidence of corrosion. Our results suggest that the mechanical integrity of the modular head-neck junction, namely the use of large (12/14 and 14/16) trunnions combined with stiff materials (CoCr alloy and alumina) to form the connection, likely played an important role in limiting taper corrosion. A similar study of well-functioning THAs of different materials (eg, titanium alloy stems with CoCr or alumina heads) and smaller taper sizes would be required to fully support this conclusion.
We thank James F. McDonald III for assistance with the identification, preparation, shipping, and tracking of the specimens used for this study.
1. ASTM F2009-00. Standard Test Method for Determining the Axial Disassembly Force of Taper Connections of Modular Prostheses. West Conshohocken, PA, USA: ASTM International; 2011.
2. Banerjee S, Cherian JJ, Bono JV, Kurtz SM, Geesink R, Meneghini RM, Delanois RE, Mont MA. Gross trunnion failure after primary total hip arthroplasty. J Arthroplasty. 2015;30:641–648.
3. Chuen Tan S, Lau AC, Del Balso C, Howard JL, Lanting BA, Teeter MG. Tribocorrosion: ceramic and oxidized zirconium vs cobalt-chromium heads in total hip arthroplasty. J Arthroplasty. 2016;31:2064–2071.
4. Cooper HJ, Della Valle CJ, Berger R a, Tetreault M, Paprosky WG, Sporer SM, Jacobs JJ. Corrosion at the head-neck taper as a cause for adverse local tissue reactions after total hip arthroplasty. J Bone Joint Surg Am. 2012;94:1655–1661.
5. Del Balso C, Teeter MG, Tan SC, Howard JL, Lanting BA. Trunnionosis: does head size affect fretting and corrosion in total hip arthroplasty? J Arthroplasty. 2016;31:2332–2336.
6. Dyrkacz RMRR, Brandt J-MM, Ojo OA, Turgeon TR, Wyss UP. The influence of head size on corrosion and fretting behaviour at the head-neck interface of artificial hip joints. J Arthroplasty. 2013;28:1036–1040.
7. Esposito C, Wright T, Goodman S, Berry D. What is the trouble with trunnions? Clin Orthop Relat Res. 2014;472:3652–3658.
8. Gilbert JL, Buckley CA, Jacobs JJ. In vivo corrosion of modular hip prosthesis components in mixed and similar metal combinations. The effect of crevice, stress, motion, and alloy coupling. J Biomed Mater Res. 1993;27:1533–1544.
9. Gilbert JL, Buckley CA, Jacobs JJ, Bertin KC, Zernich MR. Intergranular corrosion-fatigue failure of cobalt-alloy femoral stems: a failure analysis of two implants. J Bone Joint Surg Am. 1994;76:110–115.
10. Goldberg JR, Gilbert JL, Jacobs JJ, Bauer TW, Paprosky W, Leurgans S. A multicenter retrieval study of the taper interfaces of modular hip prostheses. Clin Orthop Relat Res. 2002;401:149–161.
11. Hallab NJ, Messina C, Skipor A, Jacobs JJ. Differences in the fretting corrosion of metal–metal and ceramic–metal modular junctions of total hip replacements. J Orthop Res. 2004;22:250–259.
12. Higgs G, MacDonald DW, Lowell J, Padayatil A, Mihalko WM, Siskey RL, Gilbert JL, Rimnac C, Kurtz SM. Is corrosion a threat to the strength of the taper connection in femoral components of total hip replacements? Corrosion. 2017;73:1538–1543.
13. Higgs GB, Hanzlik JA, MacDonald DW, Gilbert JL, Rimnac CM, Kurtz SM; Implant Research Center Writing Commitee. Is increased modularity associated with increased fretting and corrosion damage in metal-on-metal total hip arthroplasty devices? A retrieval study. J Arthroplasty. 2013;28:2–6.
14. Higgs GB, MacDonald DW, Gilbert JL, Rimnac CM, Kurtz SM, Chen AF, Klein GR, Hamlin BR, Lee G-CC, Mont MA, Cates HE, Malkani AL, Kraay MJ; Implant Research Center Wrting Committee. Does taper size have an effect on taper damage in retrieved metal-on-polyethylene total hip devices? J Arthroplasty. 2016;31:277–281.
15. Hussenbocus S, Kosuge D, Solomon LB, Howie DW, Oskouei RH. Head-neck taper corrosion in hip arthroplasty. Biomed Res Int. 2015;758123.
16. Hussey DK, McGrory BJ. Ten-year cross-sectional study of mechanically assisted crevice corrosion in 1352 consecutive patients with metal-on-polyethylene total hip arthroplasty. J Arthroplasty. 2017;32:2546–2551.
17. Jacobs JJ. Corrosion at the head-neck junction: why is this happening now? J Arthroplasty. 2016;31:1378–1380.
18. Jacobs JJ, Gilbert JL, Urban RM. Current concepts review: corrosion of metal orthopaedic implants. J Bone Joint Surg Am. 1998;80:1554.
19. Kao YJ, Koch CN, Wright TM, Padgett DE. Flexural rigidity, taper angle, and contact length affect fretting of the femoral stem trunnion in total hip arthroplasty. J Arthroplasty. 2016;31:254–258.
20. Kurtz SM, Kocagöz SB, Hanzlik JA, Underwood RJ, Gilbert JL, MacDonald DW, Lee GC, Mont MA, Kraay MJ, Klein GR, Parvizi J, Rimnac CM. Do ceramic femoral heads reduce taper fretting corrosion in hip arthroplasty? A retrieval study. Clin Orthop Relat Res. 2013;471:3270–3282.
21. Lavernia CJ, Baerga L, Barrack RL, Tozakoglou E, Cook SD, Lata L, Rossi MD. The effects of blood and fat on morse taper disassembly forces. Am J Orthop. 2009;38:187–190.
22. McGrory BJ, MacKenzie J, Babikian G. A high prevalence of corrosion at the head–neck taper with contemporary Zimmer con-cemented femoral hip components. J Arthroplasty. 2015;30:1265–1268.
23. Mistry JB, Chughtai M, Elmallah RK, Diedrich A, Le S, Thomas M, Mont MA. Trunnionosis in total hip arthroplasty: a review. J Orthop Traumatol. 2016;17:1–6.
24. Panigrahi P, Liao Y, Mathew MT, Fischer A, Wimmer MA, Jacobs JJ, Marks LD. Intergranular pitting corrosion of CoCrMo biomedical implant alloy. J Biomed Mater Res B Appl Biomater. 2014;102:850–859.
25. Scholl L, Schmidig G, Faizan A, TenHuisen K, Nevelos J. Evaluation of surgical impaction technique and how it affects locking strength of the head-stem taper junction. Proc Inst Mech Eng H. 2016;230;661–667.
26. Tan SC, Teeter MG, Del Balso C, Howard JL, Lanting BA. Effect of taper design on trunnionosis in metal on polyethylene total hip arthroplasty. J Arthroplasty. 2015;30:1269–1272.
27. Triantafyllopoulos GK, Elpers ME, Burket JC, Esposito CI, Padgett DE, Wright TM. Otto Aufranc Award: Large heads do not increase damage at the head-neck taper of metal-on-polyethylene total hip arthroplasties. Clin Orthop Relat Res. 2016;474:330–338.
28. Urban RM, Jacobs JJ, Gilbert JL, Galante JO. Migration of corrosion products from modular hip prostheses. J Bone Joint Surg Am. 1994;76:1345–1359.
29. Urish KL, Hamlin BR, Plakseychuk AY, Levison TJ, Higgs GB, Kurtz SM, DiGioia AM. Trunnion failure of the recalled low friction ion treatment cobalt chromium alloy femoral head. J Arthroplasty. 2017;32:2857–2863.
30. Willert HG, Buchhorn GH, Fayyazi A, Flury R, Windler M, Köster G, Lohmann CH. Metal-on-metal bearings and hypersensitivity in patients with artificial hip joints: a clinical and histomorphological study. J Bone Joint Surg Am. 2005;87:28–36.