i

Micro-grooved Surface Topography Does Not Influence Fretting Corrosion of

Tapers in THA: Classification and Retrieval Analysis

A Thesis

Submitted to the Faculty

of

Drexel University

By

Christina Marie Arnholt

In partial fulfillment of the requirements

For the degree

of

Masters of Science in Biomedical Engineering

June 2015

ii

Copy Right

iii

Acknowledgements

First I would like to thank my mom, Cindy Arnholt for always being supportive and

believing that I could accomplish anything. This wouldn’t have been possible without you.

Additionally I would like to thank my boyfriend of 6 years, Chris Ousey for always being

supportive and proud of me.

To Dr. Steven Kurtz I would like to say thank you for taking a chance on me. When

we first met I had done nothing notable and had no proof of what I could do. Knowing

that someone with as much experience believed in me gave me the strength to push myself

far beyond what I dreamed possible. Your far sighted style of thinking has given me a better

approach to work and life. I admire and hope to be like you someday.

To Daniel Macdonald thank you for being a good friend and mentor. Thank you for

answering the phone in the middle of the night and talking me off the ledge. Thank you for

proof reading all of my papers and being a sounding board to bounce ideas off of. I don’t

think this would have been possible without you.

To Michael Beeman thank you for being the other neurotic person in the lab. Thank

you for booking all of my flights, for answering all of my silly questions and for handling all

of the paper work. I will miss you.

To Richard Underwood thank you for being a good friend and mentor. Thank you

for joking with me, fighting with me and pushing me to be better. I respect your opinions

even if I don’t always say it. I look forward to working with you in the future.

iv

To Drexel Implant research center family I love you guys. Thank you for dealing

with me when I’m half asleep and talking incessantly, for making fun of me for drinking too

much coffee and for always listening. I will always feel like I was a part of something great

because of you guys.

v

Table of Contents

List of Figures .................................................................................................................................... viii

List of Tables ....................................................................................................................................... xi

Abstract ................................................................................................................................................ xii

1. Problem Statement .......................................................................................................................... 1

2. Specific Aims .................................................................................................................................... 3

1. Develop a method to characterize a broad range of machined smooth stems and

intentionally micro-grooved stems. ............................................................................................... 3

2. Using multivariate analysis of covariance to determine the associated factors with

fretting corrosion. ............................................................................................................................ 3

3. Introduction and Background ........................................................................................................ 5

3.1. Total Hip Arthroplasty ............................................................................................................ 5

3.2. Tapers ......................................................................................................................................... 6

3.2.1. Introduction of Tapers ..................................................................................................... 6

3.2.2. History of Tapers .............................................................................................................. 9

3.3. Clinical Implications ............................................................................................................... 10

3.4. Corrosion ................................................................................................................................. 11

3.4.1. Background ...................................................................................................................... 11

vi

3.4.2. Mechanically Assisted Crevice Corrosion .................................................................... 12

3.4.3. Semi Quantitative Scoring Method ............................................................................... 14

3.4.2. Taper Corrosion Implant Factors ................................................................................. 15

3.6. Surface Topography ............................................................................................................... 18

3.6.1 Introduction of Surface Topography ............................................................................ 18

3.6.2. Current Analysis of Surface Topography in THA ...................................................... 19

4. Aim 1: A Classification System for Surface Topography ......................................................... 22

4.1. Hypothesis ............................................................................................................................... 22

4.3. Design Block Diagram ........................................................................................................... 22

4.4. Materials and Methods ........................................................................................................... 23

4.4.1. Implant Information ....................................................................................................... 23

4.4.2. Microscopy and Visual Inspection ................................................................................ 24

4.4.3. White Light Interferometry ............................................................................................ 24

4.4.4. Stylus Profilometer .......................................................................................................... 24

4.4.5. Final Design Validations ................................................................................................ 25

4.5. Results ...................................................................................................................................... 26

5. Aim 2: Patient and Implant Factors Associated with Fretting Corrosion ............................. 36

5.1. Hypothesis ............................................................................................................................... 36

5.2.1. Clinical Information ........................................................................................................ 36

5.2.2. Design Variables .............................................................................................................. 37

vii

5.2.3. Fretting and Corrosion Evaluation ............................................................................... 38

5.2.4. Statistical Analysis ............................................................................................................ 40

5.3. Results ...................................................................................................................................... 40

7. Discussion and Conclusion .......................................................................................................... 45

References ........................................................................................................................................... 50

viii

List of Figures

Figure 1. Image of total hip arthroplasty components used in a standard procedure ............... 6

Figure 2. Description of Morse Taper Design and terminology ................................................... 7

Figure 3. Diagram of THA head neck taper junction ..................................................................... 8

Figure 4. Description of crevice corrosion reaction with stagnant fluid, reactive species and

released metallic ions ......................................................................................................................... 12

Figure 5. Example of imprinting within femoral head of a micro-grooved stem surface ....... 18

Figure 6. Three different components photographed using Nikon D80 (Chiyoda Tokyo) and

3D digital microscope (Keyence VHX-5000 series Digital Microscope). Although figure C is

clearly micro-grooved, A and B are subjectively classified using this method .......................... 28

Figure 7. Figure 4.5.2. Two different components analyzed using white light interferometry.

The components are clearly distinguishable between smooth and micro-grooved. Figure A

has a Wmax value of 1.73 um, while figure B has a value of 11.30. Although the filters

utilized during data collection made this method not appropriate ............................................. 29

Figure 8. Plot of PA values collected from 398 femoral stem components using a stylus

profilometer. It is clear that there is no separation describing two different surface

topographies. ....................................................................................................................................... 30

Figure 9. Plot of PA values for different geometries .................................................................... 30

Figure 10. Typical surface profiles (Left) and micrographs (Right) of A) un-patterned

smooth, B) low amplitude and wavelength patterned smooth surface and C) micro-grooves

from femoral neck taper surfaces. ................................................................................................... 32

ix

Figure 11. Amplitude values of known smooth and micro-grooved components. ................. 33

Figure 12. Wavelength values of known smooth and micro-grooved components ................ 34

Figure 13. ROC curve representing different cutoff values of amplitude (A) and wavelength

(W) ........................................................................................................................................................ 34

Figure 14. Reported area under the curve values and standard errors from ROC curve

analysis ................................................................................................................................................. 34

Figure 15. Known micro-grooved and smooth wavelength vs. amplitude plot with final

classification system marked with black line. ................................................................................. 35

Figure 16. Classification of 398 THA components....................................................................... 35

Figure 17. Design variables measured from the femoral stem. .................................................. 38

Figure 18. Exemplar components for each of the scores in the four point scoring method. 39

Figure 19. Figure 5.3.1. Stacked bar plot of the fretting corrosion damage score of femoral

heads with respect to surface topography classification. From the plots it is observed that the

smooth taper surface topography has a higher damage score potentially due to confounded

factors such as engagement length, implantation time and flexural rigidity .............................. 41

Figure 20. An example of a micro-grooved surface topography with a fretting corrosion

damage score of 4 (A) and a femoral head that has an imprinted corrosion pattern (B). ...... 42

Figure 21. Distribution of implantation time by fretting corrosion damage scores of the

femoral head taper. A significant positive correlation (p < 0.0001) was observed between

implantation time and fretting corrosion score of the femoral head in multivariate analysis of

variance. ............................................................................................................................................... 42

Figure 22. Distribution of apparent engagement length by fretting corrosion damage scores

of the femoral head taper. A significant negative correlation (p < 0.0001) was observed

x

between the apparent engagement length and fretting corrosion score of the femoral head in

multivariate analysis of variance. ...................................................................................................... 43

Figure 23. Distribution of flexural rigidity of the stem trunnion by fretting corrosion damage

scores of the femoral head taper. A significant negative correlation (p = 0.008) was observed

between the flexural rigidity and fretting corrosion score of the femoral head in multivariate

analysis of variance. ............................................................................................................................ 44

xi

List of Tables

Table 1. Scoring criteria for semi quantitative scoring method ................................................... 15

Table 2. Design matrix quantifying compliance of different measurement methods with the

design specifications listed in section 4.2. Maximum compliance was described with a score

of 5 and minimum compliance was described with a score of 0 ................................................ 27

xii

Abstract

Micro-grooved Surface Topography Does Not Influence Fretting Corrosion of

Tapers in THA: Classification and Retrieval Analysis

Christina Marie Arnholt

Dr. Steven Kurtz, Ph.D.

Surface topography of the femoral stem trunnion has been suggested as a factor in

fretting corrosion of the taper interface in total hip arthroplasty (THA). The purpose of this

study was to identify if femoral stem taper morphology was correlated with taper fretting

and corrosion. This was accomplished by first developing a method to characterize a broad

range of machined smooth stems and intentionally micro-grooved stems. Through

multivariate analysis of covariance, different factors were tested for correlation with fretting

corrosion. Finally a matched cohort was designed controlling all related factors to deduce

weather surface topography was related with fretting corrosion.

From a multi-institutional retrieval collection of over 3,000 THAs, 398 stems paired

with CoCr (ASTM-F75) alloy heads were collected as part of a multi-center, IRB-approved

retrieval program. Stems were fabricated from CoCr (ASTM-F75) or Titanium (Ti6V4Al)

alloys and were used in metal on polyethylene (M-PE) bearing total hip devices. Other stem,

femoral head alloys, and different bearing combinations were removed from this study. The

devices had a single location of modularity at the head neck junction.

In order to classify the surface topography into two groups visual inspection,

microscopy, white light interferometer (Zygo New View 5000) and a stylus profilometer

xiii

(Talyrond 585, Taylor Hobson, UK) were used. The final method employed a stylus

profilometer with which linear profiles were measured using a diamond tip stylus capturing a

10 mm line trace. Commercial software (Ultra, Taylor Hobson, UK) was used to analyze a 1

mm representative as-manufactured region. Three parameters were calculated from the

profiles: average surface roughness, amplitude and wavelength of micro-grooves (if any).

Surface observations led to a classification system in which a surface had to contain a

periodic pattern, a wavelength > 140 μm and an amplitude of > 5 μm to be considered

micro-grooved. Fifty percent (200/398) of the femoral stem taper surfaces were classified as

smooth tapered stems. The remaining 50% (198/398) femoral stem taper surfaces were

classified as micro-grooved tapered stems.

A semi-quantitative scoring method was used to classify the stem-head pairs into

four groups based on quantity of damage on the female and male taper surfaces [1, 2]. For

this study the femoral head taper was characterized. With this scoring system, a score of 1 is

assigned when the damage is considered minimal and corresponds to fretting damage occurring

on less than 10% of the surface with no pronounced evidence of corrosion. A score of a 2

indicates mild damage where more than 10% of the surface has fretting damage or there is

corrosion attack confined to small areas. A score of a 3 reflects moderate damage where more

than 30% of the surface has fretting damage or localized corrosion attack. A score of 4

corresponds to severe damage over the majority of the taper (>50%) with abundant corrosion

debris. The femoral head taper was independently evaluated by three trained investigators

(C.A., G.B.H, and D.W.M.). Any differences between investigators damage scores were

resolved in a conference, resulting in a final damage score for both the femoral head and

stem.

xiv

Using multivariate analysis of covariance we found implantation time (p<0.0001),

apparent engagement length (p<0.0001), flexural rigidity (p=0.008) and head size (p=0.01)

were significant factors in fretting corrosion head damage scores. Surface topography (i.e.

smooth or micro-grooved, p=0.78), surface wavelength (p=0.60), and surface amplitude

(p=0.23) were not associated with femoral head fretting corrosion damage score.

Overall, the results of this study do not support trunnion surface morphology as a

contributing factor to fretting and corrosion damage at the modular head-neck interface.

This was deduced through an initial classification system, followed by statistical analysis,

which showed correlated factors and finally a controlled matched cohort that showed no

correlation between surface topography and fretting corrosion. This study was limited by

the use of a semi quantitative scoring method. Future work would include the use of

quantified volume of metal released from each cohort.

xv

1

1. Problem Statement

Total hip arthroplasty is a successful surgery that allows increased patient

functionality and loss of pain. Although it is a successful surgery, within the United States

the rates of both primary and revision surgeries have increased [3]. Because of this,

researchers have been trying to identify implant and patient factors associated with failure.

One failure mode involves mechanically assisted taper corrosion, which has previously been

linked to a multitude of factors including femoral head size, flexural rigidity and implantation

time [2].

Recently, surface topography has been suggested as a factor in fretting corrosion,

because of the recent observation of “micro-grooved” surface topography [4-7]. It is

believed that these surfaces were created to limit the stress concentration created from a

mismatch of cone geometry tolerances specifically for ceramic heads [8]. These surfaces

have been observed as large and smaller grooves with varying wavelengths and amplitudes

dependent on manufacturer and design [5, 7]. Femoral stems with these machined surfaces

can be paired with both ceramic and femoral heads [7]. It has been observed in several

studies that “imprinting” from the stem surface topography can occur on the inside of metal

femoral heads [8-12]. Imprinting is explained as in vivo corrosion damage [8, 11].

Currently, a standard method for classifying surface morphology does not exist.

Previous research has suggested using a classification system consisting of two groups:

threaded and non-threaded [13]. Additional studies have looked at the effects of surface

topography on fretting corrosion, however these studies are limited by observing the femoral

stem [14] which has been shown to not be the main cause of material loss. Other studies are

2

limited by number of manufacturers leading to confounded factors such as apparent

engagement length and surface topography [15, 16].

In order to accurately distinguish if there is a correlation between surface topography

and fretting corrosion, an adequately sized cohort classified for surface topography with

controlled implant, patient and clinical factors is needed.

3

2. Specific Aims

1. Develop a method to characterize a broad range of machined smooth stems and

intentionally micro-grooved stems.

To understand the potential impact of manufactured micro-grooves on the fretting

corrosion behavior of modular total hip arthroplasty (THA) components, a

classification system is needed to identify which femoral neck tapers have micro-

grooves. Additionally large variations in design dimensions exist between

manufactures, making it necessary to create a method of classification for all designs

and manufactures to deducing the “type” of trunnion surface. This methodology is

necessary for creating a standardization and comparability between studies of THA

trunnion topography. Developing a highly repeatable classification method was

crucial to the success of this study.

2. Using multivariate analysis of covariance to determine the associated factors with fretting

corrosion.

The use of step-forward multivariate linear modeling utilizing an ordinal logistic fit

allowed for the analysis of ordinal fretting corrosion scores. This was used to

evaluate the role of surface morphology in the context of other variables known to

influence corrosion damage. Through this statistical analysis, it will be possible to

4

deduce whether fretting corrosion is influenced by surface morphology and other

variables.

5

3. Introduction and Background

3.1. Total Hip Arthroplasty

The use of modern total hip arthroplasty (THA) to combat patient ailments such as

osteoarthritis, osteonecrosis, dysplasia and inflammatory arthritis began between the 1960s

and 1970s when Sr. John Charnley introduced the Charnley hip [3, 17]. Surgery is commonly

indicated for chronic pain and osteoarthritis. This approach is considered a secondary

solution to less invasive techniques such as activity modification, weight loss or even intra-

articular injections [3]. This surgery is clinically successful with reports of survivorship over

93% survivorship after eight years [18]. The surgery accounts for a 2.7 billion dollar annual

burden with 284,000 primary surgeries and 45,000 revision surgeries each year [19].

THA is a surgery in which an artificial hip is used to replace a damaged native hip

structure. The native hip is composed of a ball and socket joint. The femoral head is located

at the superior aspect of the femur and its articulating surface is covered with cartilage. This

fits into a concave cavity also covered with cartilage called the acetabulum. During surgery

the femoral head is removed and the femoral shaft is reamed creating a canal, in which the

femoral stem is implanted [17]. The acetabulum is commonly replaced with a metal cup

called the acetabular shell and an ultra-high molecular weight polyethylene (UHMWPE) liner

[17], Figure 1.

6

Figure 1. Image of total hip arthroplasty components used in a standard procedure

A single modular taper junction is incorporated in most THA components between

the femoral head and stem. Modularity allows for interchangeability of femoral heads[20],

and an easier insertion of the femoral stem[20, 21]. Modularity also allows the combination

of different materials and different sized femoral heads, which can be decided on a per

patient basis [21]. Additional modular junctions have been incorporated into modern THA

designs to assist with the femoral neck angle and limb length discrepancies [22], while these

allow for a more successful surgery in the short term, they may result in increased corrosion

in the long term [1].

3.2. Tapers

3.2.1. Introduction of Tapers

The Morse taper, invented in 1864 by Stephen A. Morse, was originally used to

connect rotating machinery components. [23]. Machinery handbooks define this style of

taper as “self-handling”, and indicate its common uses such as drilling and machine lathing

7

[24]. The term self-handling means that the taper can be impacted and maintain the juncture

without any additional design features such as locking pins. This strong association originates

from the friction created between mating components due to close contact resulting from

smaller taper angle [25]. In contrast, self-releasing styled tapers, which have a much larger

taper angle and decreased contact between mating surfaces require an additional locking

mechanism to secure the fit. Through the control of length and diameters with a constant

tapering slope, machinists are able to control the amount of contact a taper has with its

mating surface. Machining tapers are described by taper per foot, which fluctuates according

to the taper number [25]. For example a 1 foot taper with a 1/3 inch diameter on one end

and 2/3 inch diameter on the other end, would be described as being tapered 1/3 inch per

foot, Figure 2.

Figure 2. Description of Morse Taper Design and terminology

8

The use of taper junction became widely used in orthopedics with the introduction

of ceramic femoral heads [23]. The original method for attaching a ceramic femoral head to

a metallic stem was to glue and then screw the components together which caused ceramic

failure through fracture [23]. In 1974 Professor Mittelmeier used the Morse taper to facilitate

the connection, decreasing the rate of head fractures [23]. The term Morse taper is loosely

used to describe the cone shape and self-handling capability of orthopedic tapers. However,

general naming techniques such as 11/13 or 12/14 describe the small and large diameters of

THA tapers. These tapers are used today in orthopedics (total knee arthroplasty, total hip

arthroplasty, and shoulder arthroplasty), dentistry, and still in machining tools. A taper

consists of a protruding cone (male component) fitting into a tapered concave shape (female

component). The two components are assembled intraoperatively. This process is

commonly called impaction, in which a hammer is used to secure the femoral head onto the

femoral stem, Figure 3.

Figure 3. Diagram of THA head neck taper junction

9

3.2.2. History of Tapers

In the 1990s femoral tapers were used throughout THA in multiple bearing

combinations such as metal on polyethylene and metal on metal. Taper corrosion was noted

at the head neck junction and attributed to particle generation, implant fracture and mixed

alloys [20, 22, 26-29]. The use of taper connections was continued however, because it was

determined that the problems originated from poorly designed components and were

potentially not the cause of the failures in THA [30].

In the 2000s large head metal on metal devices were widely used in THA [31]. These

devices were used because they had lower rates of dislocation and the metallic bearing was

predicted to have lower wear. However, the clinical experience differed. These systems had

reported taper fretting corrosion resulting in metallic debris release and biological reactions

[31-35]. It was speculated that the larger head size allowed for increased torque at taper and

bearing surfaces leading to increased wear [16].

In the 2010s metal on metal devices continued to have debris release leading to

biological reactions [32, 36, 37]. Metal on polyethylene bearing couples exhibited blackened

taper trunnions and elicited reactions to metallic debris similar to metal on metal devices [38,

39]. These findings supported future research into other implant and patient factors that

may be associated with fretting corrosion at the head neck taper junction.

10

3.3. Clinical Implications

As previously stated, fretting corrosion at the head neck junction has the possibility

of releasing metallic debris and corrosion byproducts. It is important to note that the metal

release is not biologically inert. The size of metal particles has been associated with the type

of response elicited by metallic debris. Meyer et al. showed in vivo periprosthetic tissues have

stored metal debris, eliciting a foreign body reaction and necrosis [36]. Through staining for

CD68, a macrophage marker, and CD3, a lymphocyte marker there were sightings of both

reactions although the former was more abundant [36]. These findings were supported by

Howie et al. who observed large numbers of macrophages around cementless metal on metal

tissues while also noting minimal giant cell and lymphocyte activity [40]. Jacobs et al.

reported the association of larger particles with giant cells or necrotic regions, while smaller

particles were found within macrophages. Mathiesen et al. reported patients with

hypersensitivity experience a lymphocyte dominated reaction. Macrophages and giant cells

were also reported having phagocytized wear particles [29].

The reaction is similar to a biological reaction to bacteria in which macrophages and

other foreign body cells migrate to the metallic particles and phagocytize them [37, 41]. The

reaction differs because the metallic particles remain after the cells die, which then repeats

the process escalating it to an elevated inflammatory response [41]. Cells are hypothesized to

die due to toxic levels of cobalt after the removal of the passive layer within the lysosome of

the cell [37]. An inflammatory mass can develop in response to the cytokines further

continued by the repeating of the process [37].

11

Metal release can lead to adverse local tissue reactions [38, 42] which can manifest as

pseudo tumors or hypersensitivity [43, 44]. Although it was originally observed in metal-on-

metal bearing surfaces [36, 37], these phenomena can also occur in metal-on-polyethylene

bearings [38, 39]. It is sometimes difficult to diagnose these reactions because patients often

complain about pain or recurrent instability [38, 42, 45]. Revision surgeries reveal black

deposits at the head/neck junction associated with discoloration of the capsule,

granulomatous mass or effusions [38, 42]. Before revision surgery this can be determined

using serum ion levels observing for elevated cobalt levels and metal artifact reduction

sequences (MARS) MRI observing for large masses [38, 42, 44, 46, 47]. Sectioning of the

periprosthetic tissue can also describe the reactions present around the device. Campbell et

al. proposed an aseptic lymphocytic vasculitis-associated lesions (ALVAL) score which

describes examination of synovial lining integrity, inflammatory cell infiltrates, and tissue

organization [48].

3.4. Corrosion

3.4.1. Background

Corrosion is described as an environmental affect causing the loss of material

properties or mass over time [49]. It can affect multiple industries and has raised concern

from a safety and economic perspective [49]. Corrosion impacts multiple materials, one

group being metals. Metallic components are subject to nine basic forms of corrosion:

uniform corrosion, intergranular corrosion, galvanic corrosion, crevice corrosion, pitting,

erosion corrosion, stress corrosion cracking, biological corrosion and selective leaching [49].

12

3.4.2. Mechanically Assisted Crevice Corrosion

Although older theories have included galvanic corrosion [28, 50, 51], it is currently

well accepted that mechanically assisted crevice corrosion is the mechanism of corrosion at

the head neck junction. The theory states: “Mechanical loading of prosthesis can result in

fretting of the modular taper interface, fracture of the passive oxide films, repassivation, and

crevice corrosion” [52].

Figure 4. Description of crevice corrosion reaction with stagnant fluid, reactive species and released metallic ions

The theory of mechanically assisted crevice corrosion is comprised of two

components: fretting and corrosion. Fretting is described as the small oscillatory movement

between two solid surfaces in contact [53]. In orthopedics fretting occurs during the human

walking cycle and is known to abrade the protective oxide film allowing for increased

exposure of virgin metal [2, 52]. Crevice corrosion is a geometry dependent (cannot exceed

3mm) type of corrosion that relies on stagnant fluids and reactive species (oxygen,

13

hydroxide, chloride) [49]. The local oxygen present in the fluid reacts with metal ions in

solution creating a concentrated cell that can react through the bulk material [49]. Released

metallic ions react with chlorine and hydroxide to make metal hydroxides and hydrochloric

acid, Figure 4 [49]. The increased acidity decreases the metals passivity creating an

environment supporting further corrosion [49].

The materials traditionally used in THA are known to form passive films making

them highly corrosion resistant [50]. Titanium (Ti-6Al-4V) is commonly used for femoral

hip stems and has excellent biocompatibility. The material has a low elastic modulus (110

GPa) and high fatigue resistance due to an almost perfect stoichiometric oxide allowing less

defects [50]. Cobalt chrome (Co-Cr-Mo), is also used in THA generally for the femoral head

or stem. It has high wear resistance and a higher elastic modulus compared with titanium

(200 GPa).

Within total hip arthroplasty the head neck junction has an initial tolerance creating a

narrow crevice with fluid penetration that can be exacerbated by irregularities within the

device [50]. The onset of walking introduces shear stresses into the junction that are large

enough to fracture the oxide film followed by spontaneous re-passivation [50]. The synovial

fluid contains oxygen and chloride species that can penetrate the crevice through the

minimal allotted fluid flow [28, 50]. As previously described this results in heightened

hydroxide and hydrochloric acid concentration [49, 52, 54]. The result of the chemical

species leads to a lower fluid pH in the crevice creating an active metal within the taper

junction. A reaction occurs within the inside and outside of the crevice allowing the

corrosion mechanism to continue [50]. As the corrosion continues the minimal amount of

14

fluid flow allows debris to escape the crevice and replenishes the reactive chloride and

oxygen species. This process repeats allowing for metal corrosion and wear.

3.4.3. Semi Quantitative Scoring Method

A historical study produced in 2002 by Goldberg et al. quantified fretting corrosion

damage of modular connections in 231 THA [2]. This study produced a methodology that

would score damage based on percent of area affected inside the femoral head and on the

surface of the femoral stem, Table 1 [2]. It was an easily repeatable method that has been

used in multiple research studies [1, 7]. This method was modified to include different

implant types and surfaces (Total knee arthroplasty [55], and acetabular shell and femoral

neck surfaces in metal on metal THA [1]). It was also modified to create a better description

of the wear specifically in regards to the phenomena of imprinting [1]. The accuracy of this

method has been proven through comparison of measured material loss and component

scores [56].

15

Table 1. Scoring criteria for semi quantitative scoring method

Damage Score Criteria

Minimal 1 Fretting on < 10% surface and no corrosion

damage

Mild 2 Fretting on > 10% surface and/ or corrosion attack

confined to one or more small areas

Moderate 3 Fretting >30% and/ or aggressive local corrosion

attack with corrosion debris

Severe 4 Damage over major (>50%) of mating surface with

sever corrosion attack and abundant corrosion

debris

3.4.2. Taper Corrosion Implant Factors

The distribution of forces throughout the taper are affected by different design

parameters and can affect the likelihood of fretting corrosion. Tapers have an angle

describing the slope of the line connecting two different diameter circles. This angle can

affect the wear and fit of the taper [57]. Bishop et al. observed wear patterns within the

femoral head which they described as asymmetric (non-homogenous) and axisymmetric

(homogenous) [8]. The angle mismatch between the male and female taper allows non-

16

continuous contact of the taper could create an opened and closed end of the taper. This

potentially allows a toggling motion with greater motion at the open end that has been

explained through the turning moments resulting from joint force vectors [8, 16]. This

results in non-homogenous wear, which has also been confirmed through finite element

(FE) analysis that explains double the amount of wear on one side through off axis loading

[57]. This type of wear was correlated to decreased head offset and angle differences [8].

These findings were supported by stochastic FE simulation where angular mismatch, center

offset and patient mass were correlated with change in contact mechanics [58].

However Kocagoz et al. observed taper angle clearance, described as the difference

between femoral head taper angle and femoral stem taper angle, and reported no correlation

with observed fretting corrosion score [9]. This study observed a matched cohort of ceramic

and CoCr components highlighting both distal and proximal engagement, featuring

previously described wear patterns. However the distribution of taper angle clearance was

similar between all scored cohorts.

It has been speculated that the femoral head size or diameter is linked with fretting

corrosion. This has been explained through increased moment arm and torque, specifically

occurring in metal on metal hips [16, 59, 60]. Axisymmetric wear was hypothesized to be

related with axial rotation due to friction, supported by increased revision rates of metal on

metal bearing systems with larger diameter heads [8]. The moments were speculated to be

large enough to rotate the head around the stem taper in combination with toggling effects

of axial loading, allowing the pattern to encompass the entire stem surface [8]. The

hypothesis of frictional torque being associated with wear has been disputed due to the lack

of correlation with surface wear [16]. Additionally, it was hypothesized that frictional torque

17

could initially destabilize the femoral head and screw the femoral head down onto the

femoral stem creating the observed wear pattern without damaging the bearing surface [16].

It was hypothesized that the magnitude of impaction used to assemble the femoral

head and stems could affect the success of components. Nassif et al. utilized an analytical

model to demonstrate that radial stresses at the junction were uniform after a large

impaction applied through the center of the trunnion [59]. If the initial impact force was too

low or offset it could create asymmetrical radial stresses affecting the wear of the internal

taper [59]. Although it is difficult to regulate, studies show a linear relationship between

impaction force and disassembly strength [21]. Multiple impactions aligned with the taper

axis elicited the most secure fit [21], essentially capitalizing on the self-handling quality of the

taper. Additionally it was shown that assembly under dry conditions allowed for a minimal

fretting corrosion [21].

Additional factors associated with fretting corrosion include flexural rigidity,

trunnion length and trunnion thickness. Decreased flexural rigidity has been correlated with

decreased fretting corrosion [2, 45]. This term involves the elastic modulus and the moment

of inertia through the trunnion, which includes the neck diameter [2]. Trunnion length and

thickness have also been linked in the literature with fretting corrosion. Brown et al. showed

through laboratory studies that shorter neck length and increased trunnion thickness were

correlated with fretting corrosion [30]. This was supported in the literature by laboratory

and retrieval studies [16, 30, 59]. Additionally it was shown that alloy combination was not

associated with fretting corrosion [30]. These findings were contradicted by FE models in

the literature [58].

18

3.6. Surface Topography

3.6.1 Introduction of Surface Topography

Continuous sightings of imprinting,[1, 7-9, 11, 16, 59] or transfer of topography

from the femoral stem to the femoral head have lead researchers to question the impact of

femoral stem surface topography on fretting and corrosion,

Figure 5. The introduction of micro-grooved surfaces occurred to limit the local stress

concentrations due to a mismatch of cone geometry tolerances specially for ceramic heads

[8]. This was supported by Kocagoz et al. who reported through a matched cohort that

ceramic femoral heads exclusively had a proximal contact within the femoral head [9]. This

was done to prevent increased hoop stresses in the femoral head which allow crack

propagation [9, 61]. Femoral stems with these machined surfaces can be paired with both

ceramic and cobalt chromium (CoCr ASTM-F75 or ASTM-F90, here after referred to as

CoCr) alloy femoral heads.

19

Figure 5. Example of imprinting within femoral head of a micro-grooved stem surface

3.6.2. Current Analysis of Surface Topography in THA

Researchers have hypothesized that stem taper surface finish may influence taper

corrosion due to the introduction of crevices through the structure of topography. Kop et al.

conducted a study of THA with multiple points of modularity containing micro-grooved

surfaces [4]. The findings reported that devices composed of CoCr with micro-grooved

surfaces contained more corrosion. The opposite trend was observed with Ti alloyed

components [4]. An in vitro study conducted by Panagiotidou et al. noticed contradicting

trends where the passive film of micro-grooved stems was breached, leading to increased

corrosion [5]. These results are contradictory and require further analysis.

Currently, a standard method for classifying taper surface morphology is not available.

Previous research has suggested using a classification system consisting of two groups:

threaded and non-threaded [13]. A 3D optical profilometer used in conjunction with an

20

interference microscope was employed to measure Sa (defined as the arithmetic deviation of

the surface) and Sz (defined as the difference in the height between 5 peak heights and 5

valleys) parameters of 11 never implanted stems from 5 different manufacturers[13].

Additionally the thread height, the pitch, skewness (Ssk), and kurtosis (Sku) parameters were

also measured. [13]. The measurements were taken at the proximal, mid-point and distal

segment [13]. The authors reported features of the designs including variable micro-groove

heights on a case-by-case basis for all of the designs included within the study. An implied

classification was also suggested on a case-by-case basis. Seven stem designs were classified

as micro-grooved designs based on the thread heights and presented 3D surface topography

images: Profemur, Synergy, Summit, Silent, Secur-fit Max, Corail and Tri-Lock [13]. The

ABGII, Taper-lock, Accolade, and SROM were classified as non-micro-grooved[13].

Currently, it is unknown what affect (if any) the groove amplitude and wavelength would

have on fretting corrosion in vivo. Additionally, it is not clear if the classification system can

be adapted to retrieved implants with severe corrosion.

Additional studies have involved a single manufacture when observing metal on metal

hips. DePuy Synthes (Warsaw, Indiana) Corail and SROM designs were observed in each

study [15, 16]. The Corail stem contained a micro-grooved surface topography and the

SROM contained a smooth surface. A classification was not utilized in the study by Langton

et al. however the study speculated a relationship between surface topography and fretting

corrosion. They claimed a potential increase of wear due to contact stresses resulting from

plastic deformation of the taper surface[16]. A related study used Contour GT-K 3D optical

profilometer (Bruker, UK) to measure three stems for height and pitch of the machined

grooves [15]. An objective lens of 53 was used to scan the surface with a back scan of 80

21

mm and length of 50 mm [15]. The methodology used was not expanded to other stem

designs or manufactures. The study could not conclude any information regarding the

relationship between topography and fretting corrosion due to the confounded factors.

Another study conducted by Ha et al. created a system of characterizing machining

height and spacing [14]. This was done by femoral stem surface measurements and femoral

head bore measurements. The former was done using white light interferometry (NewView

6k, Zygo, Middlefield, CT) and the latter was accomplished using profilometry (Surftest

SJ201P, Mitutoyo, Aurora, IL) [14]. The system observed fretting corrosion scores from the

semi-quantitative scoring method [2]. The study concluded that machining height and

implantation time predicted the fretting corrosion scores observed [14]. A limitation of this

study was that the material loss from the femoral stem was studied and used as a predictor.

Our previous research has shown that the majority of the material loss occurs in the femoral

head with negligible loss form the femoral stem.

The existence of multiple techniques with no clearly defined repeatable method of

determining surface topography makes it difficult to compare results from different studies

and truly discern the impact of surface topography on fretting corrosion.

22

4. Aim 1: A Classification System for Surface Topography

4.1. Hypothesis

In this study, we asked: can a classification system be developed to differentiate

between intentionally micro-grooved and smooth femoral neck taper topographies? It was

hypothesized that upon visual inspection a distinct difference would be observed between

surface topographies rendering additional measurements unnecessary.

4.2. Design Specifications

Historically femoral stem dimensions have been shown to differ between

manufactures and designs. Because of this, it is crucial that the proposed classification

system is design and manufacture independent. Additionally the method should be accurate

and repeatable. It should not employ any manipulation on the measurements and show a

clear difference between smooth and micro-grooved topographies. The machine utilized in

the method should be easily accessible by different research groups across the country and

would ideally not take an excessive amount of time to complete measurements as the

number of components being classified may exceed 100 components.

4.3. Design Block Diagram

23

Femoral Stem Tapers: 398 Components 5 Major Manufactures 79 Femoral Stem Designs

Manufacture and design independent binary classification system depicting surface topography

4.4. Materials and Methods

4.4.1. Implant Information

Approximately 3,100 THA systems were collected between 2000 and 2014 during

routine revision surgeries as part of an IRB-approved, multi-institutional, orthopedic implant

retrieval program. Head/stem pairs were selected based on the following inclusion criteria: a

metal-on-polyethylene bearing system, a CoCr alloy femoral head paired either with a

titanium-6aluminum-4vanadium (Ti6V4Al, here after referred to as Ti alloy) alloy or a CoCr

alloy stem, both the head and femoral stem were implanted/explanted at the same time, and

both retrieved components were available for analysis. We excluded devices with any type of

modular stem or with a femoral head incorporating a taper adapter sleeve. We also excluded

femoral stems that were not fabricated from CoCr or Ti alloy (e.g., femoral stems made from

TMZF were excluded).

The final study cohort included the head-stem pairs from 398 total hip devices,

which were produced by 5 major manufacturers (in alphabetic order): Biomet (Warsaw,

Visual Inspection 3D Digital Microscopy White Light Interferometry Stylus Profilometer

24

Indiana; n=30 of 398); DePuy Synthes (Warsaw, Indiana; n=45 of 398); Smith and Nephew

(Memphis, Tennessee; n= 14 of 398); Stryker (Mahwah, New Jersey; n= 84 of 398); Zimmer

(Warsaw, Indiana; n=219 of 398), as well as other manufacturers (n= 6 of 398). There were

different stem designs within manufacturer cohorts and the stems were manufactured from

CoCr alloy (57%; n=227 of 398) and Ti alloy (43%; n=171 of 398). The femoral heads

ranged in size from 22 to 44mm.

4.4.2. Microscopy and Visual Inspection

Components were analyzed using a 3D microscope (Keyence VHX-5000 series

Digital Microscope) and visual inspection. Features such as notable patterned ridges, macro

appearance of grooves and overall texture were observed.

4.4.3. White Light Interferometry

A pilot study of 68 retrieved femoral neck tapers surfaces were measured with a

white light interferometer (Zygo New View 5000). Each component was scanned using a

prefabricated application controlling the settings of the machine. The Rmax (maximum

peak-to-valley roughness), Wmax (maximum height of the waviness) [62] and number of

points collected were reported.

4.4.4. Stylus Profilometer

25

The horizontal straightness function of a Talyrond 585 (Taylor Hobson, UK)

roundness machine1 or stylus profilometer was used to measure 10mm long profiles on an

unworn region of the femoral stem tapers. A conispherical diamond tip stylus with 5μm

radius was used [63].

Unworn regions were selected by visual inspection of the surface. Regions were

avoided that appeared to have been engaged with the femoral head, had areas featuring

discoloration, and regions with clear evidence of damage generated by the surgeon during

removal or impaction.

Surface profiles were analyzed using the Ultra commercial surface analysis software

(Taylor Hobson, UK). From each profile, a representative as-manufactured region of 1mm

in length was selected to standardize results. The as-manufactured region was selected based

on uniformity and trace amplitude. Regions of plastic deformation or noticeable impaction

were excluded. The amplitude and wavelength of the micro-grooves was calculated from the

representative 1mm portion of the profiles. Filters and corrections were purposefully not

utilized to avoid data manipulation or distortion of the long wavelength surface topography,

such as the micro-grooves. The amplitude was calculated as the vertical distance between the

peak and valley of a representative groove. The wavelength was calculated by taking the

average distance between 5 peaks.

4.4.5. Final Design Validations

In order to validate the final chosen method of analysis, a true positive and negative

group was constructed through literature search and manufacture information. Munir et al.

1 The Talyrond 585 has a radial straightness unit with ceramic datum, which allows the radial arm to take surface roughness profiles.

26

published that known smooth (Taperloc (n=2)) and micro-grooved (Tri-Lock (n=12), Corail

(n=5), Synergy (n=3) and Summit (n=1))[13]. Additionally Stryker orthopedics (n=24) does

not make any micro-grooved components, which is partially confirmed by the lack of

ceramic femoral heads. These components were blindly analyzed and the data was evaluated

with a binary logistic analysis and an ROC curve. The binary logistic analysis is testing that

the observations will fall into one of two categories based on one or more the tested

variables. The assumptions of this analysis include: dichotomous dependent variable,

categorical nominal independent variable, and mutually exclusive exhaustive categories. The

ROC curve is the illustration of the performance of a binary classifier system with varying

thresholds. The plot represents all positives against false positives. The area under the curve

shows the accuracy of the classification method.

4.5. Results

After attempting to measure each component using five different methods of

measurement a design matrix was completed evaluating the compliance with the necessary

specifications, Table 2.

27

Table 2. Design matrix quantifying compliance of different measurement methods with the design specifications listed in section 4.2. Maximum compliance was described with a score of 5 and minimum compliance was described with a score of 0

Visual

Inspection

3D Digital

Microscope

White Light

Interferometry

Stylus

Profilometer

PA values

Stylus Profilometer

Wavelength and

Amplitude

Accuracy 1 1 3 5 5

Repeatability 1 1 4 5 5

Design

Independent

5 5 3 5 5

Manufacture

Independent

5 5 3 5 5

Unaltered

measurement

(i.e. no use of

filters)

5 5 1 5 5

Clear

Difference

between

Topographies

1 1 3 2 5

Visual inspection of the components was given one of the lowest compliance ratings

with a total score of 18 out of 30. This method was completely subjective making it not

accurate or repeatable. A machine was not employed making allowing for unaltered data,

however it did not produce consistent results. The use of a 3D digital microscope had the

same compliance rating of 18. It shared many of the features of the visual inspection

28

method, lack of accuracy and repeatability, Figure 6. However the machine itself is not

readily available to research labs and the technique of taking pictures adds an element of

time to the measurement.

Figure 6. Three different components photographed using Nikon D80 (Chiyoda Tokyo) and 3D digital microscope (Keyence VHX-5000 series Digital Microscope). Although figure C is clearly micro-grooved, A and B are subjectively classified using this method

The white light interferometer had the lowest compliance rating of 17. This was

because it utilized filters to increase the number of data points collected which also lead to a

depleted accuracy and repeatability score, Figure 7. Additionally the machine had trouble

measuring different geometries and alloy compositions of a specific manufacture making it

not design or manufacture independent. It was able to present a difference between

topographies that was not fully trusted due to the previously described pitfalls.

29

Figure 7. Figure 4.5.2. Two different components analyzed using white light interferometry. The components are clearly distinguishable between smooth and micro-grooved. Figure A has a Wmax value of 1.73 um, while figure B has a value of 11.30. Although the filters utilized during data collection made this method not appropriate

The last two methods involve the profiles measured using stylus profilometer, PA

values and wavelength and amplitude dimensions. The PA values shared the benefit

accuracy, repeatability, design and manufacture independence and unaltered data and had a

score of 27. Where these two methods diverge is in the distinctive difference between

surface topographies. Figure 8 shows a plot of PA values collected from 398 femoral

components with no clear differentiation indicating two different groups. The PA value is

30

described as 𝑃𝐴 =1

𝐿∫ |𝑦(𝑥)| 𝑑𝑥𝐿

0 which indicates that it is the average of roughness values.

Because of this feature multiple geometries can have the same reported measurement, Figure

9.

Figure 8. Plot of PA values collected from 398 femoral stem components using a stylus profilometer. It is clear that there is no separation describing two different surface topographies.

Figure 9. Plot of PA values for different geometries

31

The use of x and z descriptive parameters to define the topography was then

explored. In order to employ this method additional structuring was necessary. Initial review

of femoral stem surface profiles revealed that there were profiles with distinct periodic

structures and others without. The non-periodic profiles had random surface topography

and were deemed smooth tapered stems (Figure 10 A). The periodic structured surfaces were

separated into two groups based on trends observed from the amplitude and wavelengths

measured from femoral stem surface profiles. In order to determine the cut off values for

amplitude and wavelength values a validation was performed using known true positive and

negative values. Binary logistic analysis was performed testing wavelength values (ranging

from 80µm to 160µm varying by 20µm) and amplitude values (ranging from 3µm to 6µm

varying by 1µm). These values were selected by observing differences between known

smooth and micro-grooved amplitude, Figure 11 and wavelength values, Figure 12. The

results of the ROC curve showed that the methods chosen were accurate, Figure 13. The

area under the curve reported ten perfect combinations of wavelength and amplitude, Figure

14. However, after further analysis of the data, Figure 15, there was an optimal predictor of

wavelength and amplitude. The validation study and initial observations, revealed that in

order to be considered micro-grooved, the surface topography of the femoral stem taper

must have a periodic pattern, a wavelength >140μm, and an amplitude >5μm. If surface

topography did not meet any of the requirements, they were considered smooth, Figure 16.

32

Figure 10. Typical surface profiles (Left) and micrographs (Right) of A) un-patterned smooth, B) low amplitude and wavelength patterned smooth surface and C) micro-grooves from femoral neck taper surfaces.

33

Figure 11. Amplitude values of known smooth and micro-grooved components.

34

Figure 12. Wavelength values of known smooth and micro-grooved components

Figure 13. ROC curve representing different cutoff values of amplitude (A) and wavelength (W)

Figure 14. Reported area under the curve values and standard errors from ROC curve analysis

35

Figure 15. Known micro-grooved and smooth wavelength vs. amplitude plot with final classification system marked with black line.

Figure 16. Classification of 398 THA components

36

5. Aim 2: Patient and Implant Factors Associated with Fretting Corrosion

5.1. Hypothesis

In this study we asked: can statistical analysis of a large cohort of femoral stem

components be used to predict factors associated with fretting corrosion and determine

whether surface topography is an associated factor. It was hypothesized that implantation

time, flexural rigidity and stem alloy would be associated factors. Surface topography was

hypothesized to not be a correlated factor.

5.2. Materials and Methods

5.2.1. Clinical Information

Clinical information and operative reports were collected from each patient including

age, gender, revision reason, and implantation time. The average patient age at the time of

implantation was 60 ± 14 years. Forty-eight percent (191 of 398) of the patients were female.

The components were implanted on average (±SD) for 6 ± 7 years (range 0 to 32y). Sixty-

five percent (257 of 398) of all components were implanted during primary surgeries and the

37

head and the stem were implanted during the same surgery. The components were

predominantly revised for loosening (n=154), infection (n=130), peri-prosthetic fracture (29)

and polyethylene wear (n=26). We studied the medical records for evidence of metal-related

pathology. In only one patient was there a notation of evidence of metallosis at the time of

revision surgery.

5.2.2. Design Variables

Because taper corrosion damage is known to be multifactorial, we identified design

variables for our analysis to complement our understanding of the potential role of stem

taper morphology. Thus, in addition to surface topography, the femoral stem taper length

and diameter were measured, and the flexural rigidity was calculated. It has been reported in

the literature that increased flexural rigidity is correlated with lower head and femoral neck

taper fretting corrosion damage scores after THA [2]. The stem femoral neck taper diameter

was measured at the distal point of engagement with the femoral head when apparent, or the

distal end of the femoral stem taper for designs where the entire taper is engaged, Figure 17.

The flexural rigidity was calculated by multiplying the 2nd moment of area by the elastic

modulus (𝐸 ∗ (𝜋 ∗(𝑁𝐷)4

64)) [2]. The alloy composition was determined using X-ray

florescence (Niton XL3t GOLDD+; Thermo Scientific, Waltham, Massachusetts). The

apparent length of engagement was measured from the proximal edge of the femoral stem

taper to the point of femoral head engagement and the femoral stem taper length was

measured as the distance from the proximal to the distal edges of the femoral stem taper.

38

Figure 17. Design variables measured from the femoral stem.

5.2.3. Fretting and Corrosion Evaluation

The femoral head and stem components were received in both assembled and

disassembled states. Assembled components were disassembled using a femoral head

extractor. The components were photographed in their as-received condition and if

applicable after disassembly. Devices were cleaned in two consecutive 20-minute soaks in a

1:10 ratio of detergent (Discide®; AliMed, Dedham, Massachusetts) to water. After

decontamination they were placed in an ultra sonication bath twice for 30 seconds in water.

After cleaning, the junctions were visually inspected for evidence of fretting corrosion

damage

39

Fretting corrosion damage at the head taper of the head/neck junction was

characterized using a modified semi-quantitative score adapted from the Goldberg method

[1, 2]. For this study the femoral head female taper was characterized, Figure 18. In this

scoring system, a score of 1 is assigned when the damage is considered minimal and

corresponds to fretting damage occurring on less than 10% of the surface with no pronounced

evidence of corrosion. A score of a 2 indicates mild damage where either more than 10% of the

surface has fretting damage or there is corrosion attack confined to small areas. A score of a 3

reflects moderate damage where more than 30% of the surface has fretting damage or localized

corrosion attack. A score of 4 corresponds to severe damage over the majority of the taper

(>50%) with abundant corrosion debris. Irregular, acute artifacts on the surface were

considered iatrogenic damage and excluded from the taper damage assessment. The femoral

head taper was independently evaluated by three trained investigators (C.A., G.B.H, and

D.W.M.). Any differences between investigators damage scores were resolved in a

conference, resulting in a final damage score for both the femoral head and stem.

Figure 18. Exemplar components for each of the scores in the four point scoring method.

40

5.2.4. Statistical Analysis

The fretting corrosion scores in this study are ordinal in nature. Therefore, we used

step-forward multivariate linear modeling utilizing ordinal logistic fitting to evaluate the role

of surface morphology in the context of other variables known to influence corrosion

damage. Specifically, we included the following factors in the step-forward analysis: stem

surface finish (i.e., smooth vs. micro-grooved), as well as metal alloy coupling, flexural

rigidity, trunnion length, engagement length, head size, head offset, femoral stem taper

length, and implantation time. For all tests, the alpha level was set to 0.05. Statistical analyses

were performed using commercial statistical software (JMP 11.0, Cary , North Carolina).

5.3. Results

Mild to severe damage (fretting corrosion head damage score ≥ 2) was observed in

294 (74%, Figure 19) of the analyzed femoral head taper. Mild to severe damage (fretting

corrosion head damage score ≥2) was observed in 70% of femoral heads that were used with

micro-grooved stem tapers (n = 140 of 198, Figure 20), and in 77% of femoral heads that

were used with smooth stem tapers (n = 154 of 200). Severe damage (fretting corrosion head

damage score = 4) was observed in 15% of the analyzed femoral head tapers (n = 60 of

398). Severe damage was observed in 9% of femoral heads that were used with micro-

41

grooved stem tapers (17/198) and in 22% of femoral heads that were used with smooth

stem tapers (43/200).

Using multivariate analysis of covariance ordinal logistic fitting, implantation time

(p<0.0001; Figure 21), apparent engagement length (p<0.0001; Figure 22), flexural rigidity (p

= 0.008; Figure 23), and head size (p=0.01) were significantly associated with increased

fretting corrosion damage of the femoral head. However, these factors were only able to

explain approximately 19% of the variation in fretting corrosion damage scores. Surface

topography (i.e. smooth or micro-grooved, p=0.78), surface finish wavelength (p = 0.60),

surface finish amplitude (p = 0.23), femoral stem taper length (p=0.66), and alloy coupling

(p=0.95 were not associated with fretting corrosion damage scores.

Figure 19. Figure 5.3.1. Stacked bar plot of the fretting corrosion damage score of femoral heads with respect to surface topography classification. From the plots it is observed that the smooth taper surface topography has a higher damage score potentially due to confounded factors such as engagement length, implantation time and flexural rigidity

42

Figure 20. An example of a micro-grooved surface topography with a fretting corrosion damage score of 4 (A) and a femoral head that has an imprinted corrosion pattern (B).

Figure 21. Distribution of implantation time by fretting corrosion damage scores of the femoral head taper. A significant positive correlation (p < 0.0001) was observed between implantation time and fretting corrosion score of the femoral head in multivariate analysis of variance.

43

Figure 22. Distribution of apparent engagement length by fretting corrosion damage scores of the femoral head taper. A significant negative correlation (p < 0.0001) was observed between the apparent engagement length and fretting corrosion score of the femoral head in multivariate analysis of variance.

44

Figure 23. Distribution of flexural rigidity of the stem trunnion by fretting corrosion damage scores of the femoral head taper. A significant negative correlation (p = 0.008) was observed between the flexural rigidity and fretting corrosion score of the femoral head in multivariate analysis of variance.

45

7. Discussion and Conclusion

This retrieval study classified and assessed the surface topography of 398 explanted

femoral stem tapers of 79 different designs from 5 major manufacturers. The final

classification method was validated using binary logistics and ROC curve. Statistical analysis

was used to identify predictive factors for fretting corrosion and determine if surface

topography was a predictive factor. The results of this study do not support the hypothesis

that trunnion surface morphology is a contributing factor to fretting and corrosion damage

at the modular head-neck interface.

The strengths of this study lie in the control of known factors and the quantity of

components, designs and manufactures assessed. Initially 398 components were assessed

using a multitude of methods to definitively characterize the surface as micro-grooved or

smooth. Initial methods such as macro inspection and digital 3D microscopy were not

complex enough to bear the intricate differences between topography. They were subject to

analysis and not repeatable. White light interferometry presented possible data manipulation

and a lack of ability to process certain geometries and alloys. Additionally this method gave

a limited field of view (0.719 × 0.719 mm). Our previous research found that the surface of

the stems is highly anisotropic with axial symmetry, and therefore the surface topography is

best characterized using a 2D surface profile measured along the whole length of the stem

taper. In the current study, the use of a diamond tipped stylus profilometer allowed reliable

measurements from the surface of the explanted stem tapers and overcame the problems of

analyst subjectivity, repeatability, data dropout and limited field of view associated with other

methods.

46

Previous studies looking at the surface topography of taper junctions have proposed

using a plethora of surface roughness parameters [13], metrologists have warned against the

“parameter rash” of sophisticated surface roughness parameters whose functional and

tribological significance is not understood [64]. We therefore chose to characterize the

micro-grooves using amplitude and wavelength rather than the less tangible parameters

available with software packages, whose significance is currently unclear.

Analysis has been conducted on the effect of stem taper surface topography and in

vivo fretting and corrosion damage [4-7]. In the laboratory, researchers have cited increased

fretting corrosion in micro-grooved stem tapers for cobalt chrome alloy stems, but found

the opposite effect for titanium alloy stems [4]. Additional studies assessing surface

topography have looked at an implant collection with limited manufactures or at the material

loss of the wrong component. This presents limitations by confounding associated factors

such as apparent length of engagement and surface topography leading them to conclude

that topography is a predictor of fretting corrosion damage [15]. Other studies have chosen

to observe the femoral stem as a predictor of fretting corrosion [14], while previous research

has shown that the femoral head bears the bulk of material loss. This could also lead to

incorrect conclusions about the impact of surface topography on fretting corrosion.

It is the opinion of this researcher that different patient and implant factors have a

larger impact on implant survivorship. Surface topography may not be correlated with

increased fretting corrosion because of the amount of contact area present in smooth and

micro-grooved stems. Either the micro-grooves may have a summation of forces along

multiple contacting grooves that is equivalent to the contact area of a smooth stem or a local

deformation may occur in micro-grooved stems where the contact patch is equivalent to a

smooth stem contact patch.

47

Additional factors may be associated with fretting corrosion. One of these factors is

the seating of the femoral head on the femoral stem [8, 16]. The presence of different wear

patterns observed in metallic femoral components [9] identifies a lack of correlation with

impaction site (proximal or distal) but rather firm seating. Head size speculated to create a

larger moment arm around the taper in combination with toggling has an affect about the

entire taper creating uniform wear in some areas of the taper supported by the axisymmetric

wear observed in this study [16, 59, 60]. The trunnion length and flexural rigidity also affect

the connection between the taper and the flexibility of the taper itself. Increased contact in a

loose seating with a flexible surface can affect the passive film of the material and the

fretting allowing further corrosion of the components.

If attempting to design around the problem of corrosion different aspects of the

component design have to be considered. To stop crevice corrosion the main method is to

prevent crevices which cannot be done due to the crevice at the interior of the femoral head

[49, 54]. However, another description of crevice corrosion is dissolution of the anode

through electron exchange within the material itself creating a gradient. Separation of the

metallic components with a polymer of some type would prevent the passage of electrons.

It can also prevent destructive fretting which can fracture the passive film, initiating

corrosion. Additionally increased contact surface area would decrease the amount of fretting

within the femoral head and lead to less fretting corrosion. This could be done by increasing

the trunnion diameter, the compaction force or decreasing the taper angle.

There were several limitations in this study. First, the method used to assess fretting

corrosion was semi-quantitative and does not provide a measure of material loss from the

taper junction. Previous studies have established that there is a correlation between visual

scoring methods and the volume of metal debris released at these junctions, but for head

48

tapers with a score of 3 or 4, there is a wide range of volume of material loss [9, 56].

However, use of this semi-quantitative method allowed for comparisons with prior research

in the literature that also utilized this method [1, 7]. Another limitation is that this study was

limited to explanted components; inspection of never-used implants could provide

additional insight into the as-manufactured surface topography [13]. However, we were able

to analyze regions that did not have gross evidence of damage, corrosion or deformation

that appeared to be representative of the as-manufactured surface. In addition, this study was

confined to monoblock stem designs incorporating a metal-on-polyethylene (M-PE) bearing

and, thus, the findings of the study should not be extrapolated to hip systems incorporating

multiple modularity or metal-on-metal bearing systems. However, the devices included in

this study are representative of M-PE total hip replacement devices currently used in the

United States.

In summary, this study developed and provided evidence to support a contact stylus

measurement technique for distinguishing surface topography. Observations of a periodic

pattern with amplitudes greater than 5μm and wavelengths longer than 140μm represented a

micro-grooved stem taper. Statistical analysis was used to determine predictive factors within

a 398 component cohort and predict the significance of surface topography on fretting

corrosion at the head neck junction. The current retrieval analysis does not support the

hypothesis that surface morphology is associated with fretting corrosion in metal on

polyethylene bearing systems, when studied in the context of other design and clinical

factors.

The findings of this study support the interchangeable use of micro-grooved stems

with ceramic and CoCr femoral heads. Clinically these findings are important as the

49

modularity of femoral stem was introduced to allow for interchangeable femoral head

materials.

50

References

[1] Higgs GB, Hanzlik JA, MacDonald DW, Gilbert JL, Rimnac CM, Kurtz SM. Is Increased Modularity Associated With Increased Fretting and Corrosion Damage in Metal-On-Metal Total Hip Arthroplasty Devices?: A Retrieval Study. The Journal of arthroplasty. 2013;28:2-6. [2] Goldberg JR, Gilbert JL, Jacobs JJ, Bauer TW, Paprosky W, Leurgans S. A multicenter retrieval study of the taper interfaces of modular hip prostheses. Clinical orthopaedics and related research. 2002;401:149-61. [3] Pivec R, Johnson AJ, Mears SC, Mont MA. Hip arthroplasty. The Lancet. 2012;380:1768-77. [4] Kop A, Keogh C, Swarts E. Proximal component modularity in THA—at what cost?: an implant retrieval study. Clinical Orthopaedics and Related Research®. 2012;470:1885-94.

[5] Panagiotidou A, Meswania J, Hua J, Muirhead‐Allwood S, Hart A, Blunn G. Enhanced wear and corrosion in modular tapers in total hip replacement is associated with the contact area and surface topography. Journal of Orthopaedic Research. 2013;31:2032-9. [6] Esposito CI, Wright TM, Goodman SB, Berry DJ. What is the Trouble With Trunnions? Clinical Orthopaedics and Related Research®. 2014:1-7. [7] Kurtz SM, Kocagöz SB, Hanzlik JA, Underwood RJ, Gilbert JL, MacDonald DW, et al. Do ceramic femoral heads reduce taper fretting corrosion in hip arthroplasty? A retrieval study. Clinical Orthopaedics and Related Research®. 2013;471:3270-82. [8] Bishop N, Witt F, Pourzal R, Fischer A, Rütschi M, Michel M, et al. Wear patterns of

taper connections in retrieved large diameter metal‐on‐metal bearings. Journal of Orthopaedic Research. 2013;31:1116-22. [9] Kocagöz SB, Underwood RJ, Sivan S, Gilbert JL, MacDonald DW, Day JS, et al. Does taper angle clearance influence fretting and corrosion damage at the head–stem interface? A matched cohort retrieval study. Seminars in arthroplasty: Elsevier; 2013. p. 246-54. [10] Hexter A, Panagiotidou A, Blunn G, Skinner J, Hart A. Trunnion design and femoral head diameter increase corrosion at the taper interface in retrieved large-diameter metal-on-metal total hip arthroplasty. International Journal of Surgery. 2012;10:S55. [11] Matthies AK, Racasan R, Bills P, Blunt L, Cro S, Panagiotidou A, et al. Material loss at

the taper junction of retrieved large head metal‐on‐metal total hip replacements. Journal of Orthopaedic Research. 2013;31:1677-85. [12] Underwood RJ, Kocagoz SB, Smith R, Sayles RS, Siskey R, Kurtz SM, et al. A Protocol to Assess the Wear of Head/Neck Taper Junctions in Large Head Metal-on-Metal (LHMoM) Hips. ASTM STP1560: ASTM International, West Conshohocken, PA; 2013. [13] Munir S, Walter WL, Walsh WR. Variations in the trunnion surface topography between different commercially available hip replacement stems. Journal of Orthopaedic Research. 2014:n/a-n/a. [14] Ha NQ, Pourzal R, Hall DJ, Urban RM, Levine BR, Lundberg HJ. Surface Topography and Damage Grading of Modern Total Hip Arthroplasty Modular Head-Neck Junctions. In: Goldring MB, editor. Orthopaedic Research Society. Los Vegas Nevada2015. [15] Hothi HS, Whittaker RK, Meswania JM, Blunn GW, Skinner JA, Hart AJ. Influence of stem type on material loss at the metal-on-metal pinnacle taper junction. Proceedings of the

51

Institution of Mechanical Engineers, Part H: Journal of Engineering in Medicine. 2015;229:91-7. [16] Langton D, Sidaginamale R, Lord J, Nargol A, Joyce T. Taper junction failure in large-diameter metal-on-metal bearings. Bone and Joint Research. 2012;1:56-63. [17] Siopack J, Jergesen H. Total hip arthroplasty. Western journal of medicine. 1995;162:243. [18] Kurtz SM, Ong KL, Schmier J, Mowat F, Saleh K, Dybvik E, et al. Future clinical and economic impact of revision total hip and knee arthroplasty. The Journal of Bone & Joint Surgery. 2007;89:144-51. [19] Kurtz SM, Ong KL, Lau E, Bozic KJ. Impact of the Economic Downturn on Total Joint Replacement Demand in the United States. The Journal of Bone & Joint Surgery. 2014;96:624-30. [20] Chmell MJ, Rispler D, Poss R. The impact of modularity in total hip arthroplasty. Clinical orthopaedics and related research. 1995;319:77-84. [21] Pennock AT, Schmidt AH, Bourgeault CA. Morse-type tapers: factors that may influence taper strength during total hip arthroplasty. The Journal of arthroplasty. 2002;17:773-8. [22] Hozack WJ, Mesa JJ, Rothman RH. Head—neck modularity for total hip arthroplasty: Is It Necessary? The Journal of arthroplasty. 1996;11:397-9. [23] Hernigou P, Queinnec S, Lachaniette CHF. One hundred and fifty years of history of the Morse taper: from Stephen A. Morse in 1864 to complications related to modularity in hip arthroplasty. International orthopaedics. 2013;37:2081-8. [24] Ramalingam K. Hand Book Of Mechanical Engineering Terms: New Age International; 1996. [25] Oberg E, Jones FD, Horton HL, Ryffel HH. Machinery's Handbook (29th Edition) & Guide to Machinery's Handbook. Industrial Press. [26] Bobyn JD, Tanzer M, Krygier JJ, Dujovne AR, Brooks CE. Concerns with modularity in total hip arthroplasty. Clinical orthopaedics and related research. 1994;298:27-36. [27] Salvati EA, Lieberman JR, Huk OL, Evans BG. Complications of femoral and acetabular modularity. Clinical orthopaedics and related research. 1995;319:85-93. [28] Collire JP, Surprenant VA, Jensen RE, Mayor MB. Corrosion at the interface of cobalt-alloy heads on titanium-alloy stems. Clinical orthopaedics and related research. 1991;271:305. [29] Mathiesen E, Lindgren JU, Blomgren G, Reinholt FP. Corrosion of modular hip prostheses. Journal of Bone & Joint Surgery, British Volume. 1991;73:569-75. [30] Brown S, Flemming C, Kawalec J, Placko H, Vassaux C, Merritt K, et al. Fretting corrosion accelerates crevice corrosion of modular hip tapers. Journal of Applied Biomaterials. 1995;6:19-26. [31] Kim RH, Dennis DA, Carothers JT. Metal-on-metal total hip arthroplasty. The Journal of arthroplasty. 2008;23:44-6. e1. [32] Bozic KJ, Browne J, Dangles CJ, Manner PA, Yates AJ, Weber KL, et al. Modern metal-on-metal hip implants. Journal of the American Academy of Orthopaedic Surgeons. 2012;20:402-6. [33] Cuckler JM, Moore KD, Lombardi AV, McPherson E, Emerson R. Large versus small femoral heads in metal-on-metal total hip arthroplasty. The Journal of arthroplasty. 2004;19:41-4. [34] Park Y-S, Moon Y-W, Lim S-J, Yang J-M, Ahn G, Choi Y-L. Early osteolysis following second-generation metal-on-metal hip replacement. The Journal of Bone & Joint Surgery. 2005;87:1515-21.

52

[35] MacDonald S. Metal-on-metal total hip arthroplasty: the concerns. Clinical orthopaedics and related research. 2004;429:86-93. [36] Meyer H, Mueller T, Goldau G, Chamaon K, Ruetschi M, Lohmann CH. Corrosion at the cone/taper interface leads to failure of large-diameter metal-on-metal total hip arthroplasties. Clinical Orthopaedics and Related Research®. 2012;470:3101-8. [37] Gill HS, Grammatopoulos G, Adshead S, Tsialogiannis E, Tsiridis E. Molecular and immune toxicity of CoCr nanoparticles in MoM hip arthroplasty. Trends in molecular medicine. 2012;18:145-55. [38] Cooper HJ, Della Valle CJ, Berger RA, Tetreault M, Paprosky WG, Sporer SM, et al. Corrosion at the head-neck taper as a cause for adverse local tissue reactions after total hip arthroplasty. The Journal of bone and joint surgery American volume. 2012;94:1655-61. [39] Cook RB, Bolland BJ, Wharton JA, Tilley S, Latham JM, Wood RJ. Pseudotumour formation due to tribocorrosion at the taper interface of large diameter metal on polymer modular total hip replacements. The Journal of arthroplasty. 2013;28:1430-6. [40] Howie DW. Tissue response in relation to type of wear particles around failed hip arthroplasties. The Journal of arthroplasty. 1990;5:337-48. [41] John KRS. Corrosion Effects on the Biocompatibility of Metallic Materials and Implants. 2006. [42] Jacobs J, Cooper H, Urban R, Wixson R, Della Valle C. What do we know about taper corrosion in total hip arthroplasty? The Journal of arthroplasty. 2014;29:668-9. [43] Akbar M, Fraser AR, Graham GJ, Brewer JM, Grant MH. Acute inflammatory response to cobalt chromium orthopaedic wear debris in a rodent air-pouch model. Journal of The Royal Society Interface. 2012:rsif20120006. [44] Shulman RM, Zywiel MG, Gandhi R, Davey JR, Salonen DC. Trunnionosis: the latest culprit in adverse reactions to metal debris following hip arthroplasty. Skeletal radiology. 2014:1-8. [45] Fricka KB, Ho H, Peace WJ, Engh CA. Metal-on-metal local tissue reaction is associated with corrosion of the head taper junction. The Journal of arthroplasty. 2012;27:26-31. e1. [46] Berry DJ, Abdel MP, Callaghan JJ, Group CR. What are the current clinical issues in wear and tribocorrosion? Clinical Orthopaedics and Related Research®. 2014;472:3659-64. [47] Kwon Y-M. Cross-sectional imaging in evaluation of soft tissue reactions secondary to metal debris. The Journal of arthroplasty. 2014;29:653-6. [48] Campbell P, Ebramzadeh E, Nelson S, Takamura K, De Smet K, Amstutz HC. Histological features of pseudotumor-like tissues from metal-on-metal hips. Clinical Orthopaedics and Related Research®. 2010;468:2321-7. [49] Schweitzer PA. Fundamentals of corrosion: mechanisms, causes, and preventative methods: CRC Press; 2009. [50] 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;27:1533-44. [51] Cook SD, Barrack RL, Baffes GC, Clemow AJ, Serekian P, Dong N, et al. Wear and corrosion of modular interfaces in total hip replacements. Clinical orthopaedics and related research. 1994;298:80-8. [52] Jacobs JJ, Gilbert JL, Urban RM. Current Concepts Review-Corrosion of Metal Orthopaedic Implants*. The Journal of Bone & Joint Surgery. 1998;80:268-82. [53] Briscoe B. Tribology—Friction and wear of engineering materials: IM Hutchings. Elsevier; 1992.

53

[54] Schweitzer PA. Fundamentals of metallic corrosion: atmospheric and media corrosion of metals: CRC press; 2006. [55] Arnholt C, Underwood R, MacDonald D, Higgs G, Chen A, Klein G, et al. Micro-grooved Surface Topography Does Not Influence Fretting Corrosion of Tapers in THA: Classification and Retrieval Analysis. Conshohocken PA: American Society of Testing Materials (ASTM). 2015. [56] Kocagoz SB, Underwood RJ, MacDonald DW, Higgs GB, Day JS, Siskey RL. Does visual inspection of the taper head / stem junctions in metal-on-metal hips adequately characterize material loss from corrosion and wear. San Antonio, TX, USA: Orthopedic research society. 2013. [57] Shareef N, Levine D. Effect of manufacturing tolerances on the micromotion at the Morse taper interface in modular hip implants using the finite element technique. Biomaterials. 1996;17:623-30. [58] Donaldson FE, Coburn JC, Siegel KL. Total hip arthroplasty head–neck contact mechanics: A stochastic investigation of key parameters. Journal of biomechanics. 2014;47:1634-41. [59] Nassif NA, Nawabi DH, Stoner K, Elpers M, Wright T, Padgett DE. Taper design affects failure of large-head metal-on-metal total hip replacements. Clinical Orthopaedics and Related Research®. 2014;472:564-71. [60] Dyrkacz RM, Brandt J-M, 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. The Journal of arthroplasty. 2013;28:1036-40. [61] Huet R, Sakona A, Kurtz SM. Strength and reliability of alumina ceramic femoral heads: review of design, testing, and retrieval analysis. Journal of the mechanical behavior of biomedical materials. 2011;4:476-83. [62] Inc. Z. Surface Texture Parameters. In: Zygo, editor.2013. [63] Talyrond 565/585: A revolutionary concept in automated roundness inspection. In: Hobson T, editor.2013. [64] Whitehouse D. The parameter rash—is there a cure? Wear. 1982;83:75-8.

54