PII: S1350-4533(97)00032-5

Med. Eng. Phys. Vol. 19, No. 8, pp. 711719, 1997

1997 IPEM Published by Elsevier Science Ltd
Printed in Great Britain
13504533/97 $17.00 + 0.00

The effect of surface topography of retrieved

femoral heads on the wear of UHMWPE sockets
R. M. Hall*, P. Siney, A. Unsworth* and B. M. Wroblewski
*Centre for Biomedical Engineering, School of Engineering, University of
Durham, Durham DH1 3LE, UK; Centre for Hip Surgery, Wrightington Hospital,
Appley Bridge, Wigan WN6 9EP, UK
Received 8 December 1996, accepted 17 July 1997

ABSTRACT
The study was undertaken to investigate the association, if any, between the surface roughness of 35 explanted
femoral heads and the clinical wear factors of the corresponding polyethylene sockets. The wear of the socket was
measured using a shadowgraph technique. The surface topography was investigated using a Rodenstock RM 600
non-contacting profilometer, and six parameters were used to characterise the roughness. Further, qualitative inspections of the femoral surface were undertaken using both a Joel JSM-IC-848 scanning electron microscope and a Zeiss
Axiotech microscope with a differential interference contrast facility. Two parameters were found to correlate with the
clinical wear factor, namely the skewness of the amplitude distribution function, Sk, and the arithmetic mean
roughness, Ra. Simple parameters describing the peak heights of the asperities were found not to have a significant
association with the clinical wear factor. The exponent in the power relationship between the arithmetic mean roughness and kclinical was found to be equal to 0.5 (SE: 0.2). This value is significantly smaller than that found in
laboratory experiments and may be due to the non-uniform nature of the roughening of the femoral head, three-body
wear and the effect of other clinical factors on the wear process. Further, extremely strong correlations were detected
between the differing roughness parameters, which would suggest that any attempt to deduce which one is the most
important in affecting the wear of the polymeric counterface is fraught with difficulty. However, further investigation
of those parameters, such as the reduced peak height or the material filled profile peak area, which may better describe
the effect of the counterface on the wear of the UHMWPE surface would appear to be prudent. 1997 IPEM
Published by Elsevier Science Ltd
Keywords: Wear, surface topography, UHMWPE, total hip replacement
Med. Eng. Phys., 1997, Vol. 19, 711719, December

1. INTRODUCTION
The wear of ultra-high molecular weight polyethylene (UHMWPE) components remains a key
issue in joint replacement surgery, especially of
the lower limb1,2. The wear process may lead to
component loosening and the eventual requirement for revision surgery3. Of those parameters
which may lead to excessive wear, changes in the
surface roughness of the femoral head are currently thought to be important and have been
investigated extensively in the laboratory49. As
well as affecting the volume of debris produced,
laboratory experimentation has demonstrated
that particle morphology is influenced by the
roughness of the counterface10,11. In particular,
rougher surfaces tend to produce more particles
with sub-micron dimensions. It is these subCorrespondence to: R. M. Hall, Academic Department of Orthopaedic Surgery, Clinical Sciences Building, St Jamess University Hospital, Beckett Street, Leeds LS9 7TF, UK. Tel: 0113 206 4186,
0113 233 2132; fax: 0113 206 6791.

micron particles which are considered to have the

greatest effect in eliciting an immune response
which eventually leads to bone resorption. Recent
clinical evidence of a wide range of particle sizes,
similar to those found for different counterface
roughnesses in vitro, has been reported, although
the observations were not correlated with the topography of the retrieved femoral heads11.
A number of studies have investigated the
changes in the surface topography of previously
implanted femoral components. In particular,
Isaac et al.12 studied the surface roughness of 35
Charnley heads in which the maximum value of
the arithmetic mean roughness (Ra) from a number of scans across each specimen was recorded.
A significant increase in the surface roughness,
relative to that found on new femoral heads, was
recorded. A further study by Isaac et al.13, whilst
reporting similar increases in surface roughness
for a larger cohort of joints, did not observe a
strong correlation between the rate of penetration, P/t, and the Ra value. This result con-

Wear and surface topography in retrieved hip prostheses: R. M. Hall et al.

firmed an earlier observation made by Atkinson

et al.14. This lack of association between P/t
and Ra is in contrast to studies undertaken in the
laboratory in which an extremely strong correlation was found to exist4,8. The lack of association
clinically may be due to a number of factors,
including the observation that the changes in topography across the femoral surface are not uniform, and additional wear due to third bodies.
Work previously undertaken on explanted femoral heads has observed multidirectional scratching, especially in areas on or adjacent to the presumed area of contact15,16. This qualitative
observation was confirmed quantitatively by Hall
et al.17 who observed no significant difference
between five roughness parameters when measured in either the anteriorposterior or medial
lateral directions.
A number of different devices can be used to
make a quantitative assessment of the surface topography of the explanted femoral heads18. The
most frequently used method of appraisal employs
the contacting profilometer in which a stylus is
drawn across the surface on which the texture is to
be investigated. A second set of devices comprises
optical profilometers, which can be further split
into groups such as interference and laser stylus
instruments. These types of device may be advantageous in situations where distortions of, or damage to, the surface may occur. The results gained
from these non-contacting instruments may deviate from those obtained from more conventional
contacting profilometers. These discrepancies are
dependent on the difference in the vertical and
horizontal resolutions of the instruments and the
precise nature of the surface under investigation.
Traditional stylus techniques are still considered
almost exclusively in the standards literature19.
It is the aim of this study to investigate the interrelationship between those topographical changes
that occur in vivo and the wear of the ultra-high
molecular weight polyethylene (UHMWPE)
socket. In particular, the association between the
clinical wear factor, kclinical, and six parameters
used to characterise the surface is analysed.
2. EXPERIMENTAL PROCEDURE
Thirty-five prostheses were acquired from revision
surgery at the Centre for Hip Surgery, Wrightington Hospital. The artificial joints were all of the
Charnley type (Thackray/DePuy International,
Leeds) comprising a single-piece stainless steel
stem and an all-UHMWPE socket (Figure 1). The
components are fixed in position with the aid of
bone cement. The mean period of implantation
was 12.0 (SD: 4.7) years while the mean age at
primary surgery was 54 (SD: 14) years. The mean
patient weight was 710 (SD: 150) N. Thirty-three
of the sockets were found to be loose at revision
surgery.
Quantitative assessment of the surface topography was achieved by using a Rodenstock RM 600
profilometer at the National Physical Laboratory
(NPL) which utilises the focus error detection
principle20,21. Twenty scans were taken from each

of the femoral heads in areas which were considered to be in contact with the socket and, in
general, this coincided with the most heavily
scratched region. Each profile consisted of a
1.4 mm evaluation length, le, with a cutoff of
0.2 mm and contained 770 sampling points.
Twenty-two surface parameters were calculated
once the profiles had been acquired. Six parameters were investigated further. Definitions of
these parameters and the reasons for their
inclusion in the study are presented in Table 1. In
particular, those parameters that have been used
in the assessment of wear in the orthopaedic
environment have been included. The excluded
parameters, including the mean peak-to-valley
height, Rz, and the mean spacing of the profile
irregularities, Sm, were thought not to have a
strong influence on the wear of the polymeric surface. The mean of each of the chosen parameters
across the twenty profiles was calculated.
Qualitative assessment of the femoral heads was
undertaken using both scanning electron (Joel
JSM-IC848) and optical (Zeiss Axiotech)
microscopy. A differential interference contrast
(DIC) facility22 was available on the optical microscope and was found to be particularly suitable for
observing small changes in the slope of features
on the surface of the head (e.g. marks left after
polishing). The objective lenses used were 5,
10, 20 and 50, which allowed overall magnifications of 50, 100, 200 and 500 times respectively. Features of interest were recorded on photographic film. Images on the SEM were recorded
using an image capture system.
The changes in the volume of the internal
socket bore, V, were recorded using the shadowgraphic technique which has been explained
extensively elsewhere23. In brief, a mould is taken
of the internal bore of the socket and projected
onto a screen at a given magnification. A profile
is recorded of the wear plane, and the centres of
the femoral head with respect to the worn and
unworn surfaces of the socket are recorded using
a circular template. The distance between the
centres of the circular template in these two positions is the penetration depth, P. The angle of
penetration, , is the angle between the direction
of penetration and the plane defining the open
face of the socket. The total change in the bore
volume, V, can then be calculated from the
modified form of the equation presented by Kabo
et al.24. Simulator investigations of the wear of
Charnley sockets have revealed a creep component to the volume change in the socket bore,
which is of the order of 30 mm3 25. This is small
compared to the mean volume change of approximately 550 mm3 for explanted Charnley prostheses and was therefore neglected in this analysis26. Thus the total volume change in the internal
bore of the socket was assumed to be attributable
to wear alone.
In keeping with the in vitro analysis, a clinical
wear factor, kclinical, was derived using the empirical wear equation:

V = kclinical Ldx
0

712

(1)

Wear and surface topography in retrieved hip prostheses: R. M. Hall et al.

Figure 1 Explanted Charnley acetabular and femoral components.

Table 1 Surface roughness parameters used to describe the topography of the explanted femoral heads
Parameter

Definition

Arithmetic mean roughness, Ra

Reason for assessment

Ra =

Frequently used in assessing the surface

roughness of the counterface in the wear of
orthopaedic polymers. Cited in BS 7251 Part
4 with respect to the maximum roughness of
the surface of a femoral head

1
zdx
l
0

Root mean square roughness, Rq

Rq =

Often used as an alternative to Ra. The

parameter gives a greater weighting to the
larger deviations from the reference line

1 2
z dx
l
0

Peak to valley height, Rt

Distance between the highest peak and lowest May be used as a measure of counterface
valley within one evaluation length
imperfections, but does not distinguish
between peaks and valleys

Peak height, Rp

Distance between the mean line and the

highest peak within one evaluation length

Mean peak height, Rpm

Mean distance between the highest peak and Less prone to exceptional peaks than Rp. The
the mean line in five consecutive sampling
mean peak height can be used in providing
lengths
information on profile shape in conjunction
with the mean peak to valley height

Skewness, Sk

1
Sk =
(Rq)3

z3p(z)dz

where L is the load across the articulating surface

and x the sliding distance. Equation (1) was simplified by Atkinson et al.14, using motion and loading curves originally observed by Paul27 such that
the integral on the right-hand side was reduced to

I = Ldx = 2.376NMr

Counterface imperfections with exceptional

peaks (esp. scratches) above the mean line
have been shown to have the greatest effect
on the wear of the polymeric surface

(2)

where M is the weight of the patient, r the radius

Provides information on the shape of the

amplitude distribution function. The skewness
indicates whether or not there is a
disproportionate number of high peaks or
deep valleys

of the femoral head and N the number of cycles

undergone by the prosthesis during the period of
implantation. A value of N was calculated using
the formula presented by Wallbridge and Dowson28 in which N varies as a function of both age
and implantation period.
The interrelationship between the variables was
undertaken using Spearmans rank correlation
test. Initial analysis between each of the surface
roughness parameters and the clinical wear factor
was also performed using this non-parametric test,

713

Wear and surface topography in retrieved hip prostheses: R. M. Hall et al.

Table 2 Summary of those values deduced from penetration depth
measurements or clinical records
Parameter

Median value
(interquartile range)

Penetration depth (mm)

Wear volume (mm3)
Number of cycles
I (N m)
kclinical (mm3 [N m]1)

2.3 (0.73.4)
608 (274920)
1.69 107 (1.032.09 107)
3.05 108 (1.634.13 108)
1.51 106 (1.092.51 106)

but was supplemented with the regression of a

power function similar to that performed by Dowson et al.8. All the analyses were accomplished
using the STATA 4.0 statistical analysis package29.
3. RESULTS
A summary of the results for the shadowgraphic
measurements and the subsequent calculations of
the wear volumes are presented in Table 2. Also
presented are the calculations for median values
of the estimated number of cycles undergone by
the prostheses while in vivo, the parameter, I, in
Equation (2) and the clinical wear factor, kclinical.
Observations of the femoral heads indicated a
wide range of scratch densities (Figure 2) as well
as scratch sizes. On the heads where less damage
had occurred, original polishing marks were
observed which tended to be of a more unidirectional nature. Polishing pits were also revealed by
the DIC microscopy, which tended to have oriented tails in which the direction was correlated
with the polishing marks (Figure 3). These were
observed on new as well as retrieved specimens.
On two heads, raised features were observed
which were consistent with a preferentially polished surface in which grain boundaries protruded (Figure 4).
The SEM studies revealed greater detail,

especially in the horizontal plane, of the scratch

morphologies. The scratch widths tended to vary
in size but most were less than 2m. Build-up of
material either side of the scratches was also
observed, although this was often interrupted by
further damage to the surface (Figure 5). In
addition to the scratches, more localised indentations were also observed, but with a topography
unlike that of the polishing pits.
As well as studying the surfaces of the femoral
heads, a sample of the worn acetabular regions
was also investigated using SEM microscopy. Most
areas showed evidence of abrasive wear, in terms
of scratches, as well as adhesive features. Material
was also found to be embedded in the UHMWPE
which may contribute to the scratching of the femoral head or may contribute to three-body wear
processes when released. On a qualitative level,
the amount of embedded material varied appreciably.
The median values and interquartile ranges of
the six surface roughness parameters are
presented in Table 3. Quantitative assessment indicated strong correlations between the different
surface amplitude parameters (Table 4). As one
would expect, the strongest correlation of all was
between Ra and Rq (Figure 6). However, only poor
correlations were observed between the skewness
of the amplitude distribution function and the
amplitude parameters. Initial assessment of the
correlation between the clinical wear factor and
each of these parameters indicated only marginally significant associations between either the
skewness (r = 0.374, P = 0.027) or the arithmetic
mean roughness (r = 0.284, P = 0.099) and kclinical
(Figure 7 and Table 5). Using the empirical model,
first presented by Dowson et al.8 in analysing such
data, the only significant association appeared to
be between Ra and the clinical wear factor (Table
5). Here, the exponent was found to have a value
of 0.5 (SE: 0.2). However, the presence of one or

Figure 2 DIC micrograph image of a heavily scratched area on the surface of an explanted head.

714

Wear and surface topography in retrieved hip prostheses: R. M. Hall et al.

Figure 3 DIC micrograph of polishing pits observed on a portion of the heads. Note the tails eminating from the pits and that their direction
appears to be oriented with the polishing marks.

Figure 4 DIC micrograph of raised features observed on two explanted femoral heads which may be due to the grain structure of the stainless steel.

two outliers may affect the overall significance of

the result.
4. DISCUSSION
As observed in other studies of retrieved femoral
head prostheses, multidirectional scratches
existed across much of the surface, but with varying densities. The microscopic examination
revealed that, while lightly damaged surfaces have
isolated scratches, this was not generally the case
in the regions where the greatest portion of the
contact was assumed to exist. It was observed that
as the scratch density increased then so did the
severity, in terms of damage, of individual

scratches, and this was confirmed by the linear

tendency between both the single amplitude parameters (Rt or Rp) and the arithmetic mean roughness, Ra. This has important implications for modelling of the femoral surface in laboratory
experiments. Until recently it has been usual to
relate the laboratory wear factor, k, to the uniform
surface roughness as measured by an Ra value.
Using pin-on-plate apparatus with distilled water
as the lubricant, Dowson et al.8 were the first to
show the following empirical relationship between
k and Ra:
k = 4 105R1.2
(3)
a
Analysing the data presented by Weightman

715

Wear and surface topography in retrieved hip prostheses: R. M. Hall et al.

Figure 5 SEM micrograph of the material build-up either side of a scratch. It was noted that the build-up was not uniform along the length
of the scratch.

Table 3 Summary of those parameters gained from the measurement of surface texture
Parameter

Median value
(interquartile range)

Arithmetic mean roughness, Ra (nm)

Root mean square roughness, Rq (nm)
Peak-to-valley height, Rt (nm)
Peak height, Rp (nm)
Mean peak height, Rpm
Skewness, Sk

62 (4180)
96 (75145)
1080 (8201550)
500 (400690)
240(200390)
0.28(1.340.38)

and Light4 in which serum rather than water was

used as the lubricant, the exponent was of the
order of 1.4. However, later experiments by Dowson et al.6 illustrated that the wear factor of the
softer polymeric surface could be inflated dramatically in the presence of prominent individual
scratches on the metallic counterface. This was
especially true if these scratches were oriented
perpendicular to the direction of articulation.
Further, it was noted that it was the material buildup on either side of the scratches that was the
major cause of this increased wear. Single indents
appeared to raise the wear factor only marginally.
Fisher et al.9 recently confirmed this finding,
although the effects of scratches with heights of
material pile-up of the order of 1m were even

Figure 6 Relationship between the arithmetic mean surface roughness, Ra, and the root mean square roughness, Rq.

more dramatic. However, the explanted femoral

surface fits neither of these descriptions since it is
not a uniformly rough surface, although it may
tend to this situation if the head is severely damaged. Neither is a single prominent scratch often
observed in the presumed contact region,
although one could surmise that this could be the
case for a brief period early in the joints life. In
fact, the topography of the femoral head lies
somewhere between these two extremes and the
effects of the topography of the counterface on

Table 4 Matrix displaying the correlations (Spearmans rank) between the different surface roughness parameters
Ra

1.0000

Rq

0.9351
(0.000)
0.7218
(0.000)
0.7344
(0.000)
0.8311
(0.000)
0.0644
(0.705)
Ra

Rt
Rp
Rpm
Sk

716

1.0000
0.8127
(0.000)
0.8234
(0.000)
0.8727
(0.000)
0.1248
(0.461)
Rq

1.0000
0.8184
(0.000)
0.8184
(0.000)
0.1293
(0.446)
Rt

1.0000
0.8838
(0.000)
0.1347
(0.429)
Rp

1.0000
0.0253
(0.882)
Rpm

1.0000
Sk

Wear and surface topography in retrieved hip prostheses: R. M. Hall et al.

Figure 7 Dependency of the clinical wear factor, kclinical, on the

mean skewness, Sk, recorded for each femoral head.

the softer polymeric material are a weighted integration of the effects of all the individual scratches
with some sort of orientational parameter
attached. The polishing pits observed on the less
damaged areas of the head appeared not to have
any lip around their edges, and therefore it could
be assumed they may not have any significant
affect on the wear of the UHMWPE socket.
Isaac et al.12 investigated retrieved acetabular
components and noted that over fifty per cent had
cement embedded within the articulating surface.
Further, it was demonstrated that bone cement
had a detrimental effect on the roughness of the
metallic counterface. There is a high probability
that the material embedded in the surfaces of
sockets investigated in this study was bone cement.
Bone may also ingress into the surface of the polymeric bearing and cortical bone, in particular, has
been shown to be of sufficient hardness to scratch
the surface of a stainless steel counterface30.
In terms of correlations between the clinical
wear factor and the roughness parameter, only
two were observed to have any significant association. The power function fit of k against Ra produced an exponent which was significantly different from zero but considerably lower than the
values gained from the studies by Dowson et al.8
or Weightman and Light4. This difference in
exponent may be due to a number of factors,
including the previously mentioned observation
that the surface roughness of the explanted heads
is not uniform. Further, the lubricating regimes
may be very different between simple wear
machines and the in vivo environment. The other
amplitude roughness parameters produced no
exponent values that were significantly different
from zero. The marginally significant association

between the skewness and the clinical wear factor

is not too surprising. A negatively skewed surface,
that is one with a few exceptional valleys but no
prominent peaks, is often considered to be a good
bearing surface in that it will reduce the abrasive
wear of the counterface relative to one with a positively skewed amplitude distribution function31.
The poor correlation between the peak-to-valley
height and the clinical wear factor may be
explained by the fact that the parameter does not
distinguish between valleys and peaks. Further,
this roughness parameter is heavily influenced by
very high individual asperities or deep valleys
which in themselves do not have a great influence
on the wear process across the whole counterface,
as indicated by Dowson et al.6. The weak association between Rp or Rpm and the clinical wear factor seems at first to be problematical since it is
the material above the mean line, especially that
formed by scratches, that has the greatest effect
on the wear of the UHMWPE. However, these
parameters, as with the Rt value, are affected by
singular asperities and do not reflect the spatial
extent of the pile-up formed around the side of
the indentations (Figure 8). These peak height
parameters may have some significance in laboratory experiments in which there are a few prominent scratches of uniform height. However, in a
clinical situation the build-up of material varies
between scratches and even along a single scratch.
In addition, these laboratory tests normally have
scratches which are well in excess of the height
changes on the surrounding surface and therefore have an unusually high and disproportionate
effect on the wear of the opposing polymeric surface. On an explanted head, however, a number
of scratches may exist in close proximity and,
whilst the largest scratch may have the greatest
effect, the role of the others cannot be ignored.
Further, the loci mapped out by a feature on the
femoral head with a detrimental effect on the

Figure 8 Schematic diagram illustrating the inability of the peak

height parameters, Rp and Rpm, to distinguish between single
asperities and material build-up adjacent to scratches. It is the latter
which have the greatest effect on the wear of the UHMWPE surface.

Table 5 Degree of association between the surface roughness parameters and the clinical wear factor measured by the Spearmans rank
correlation coefficient and the value of the exponent in the empirical model used by Dowson et al.8 (the P-values are given in parentheses)
Roughness parameter

Spearmans rank correlation coefficient

Ra
Rq
Rt
Rp
Rpm
Sk

0.284
0.248
0.004
0.130
0.211
0.374

(P
(P
(P
(P
(P
(P

=
=
=
=
=
=

0.099)
0.151)
0.981)
0.457)
0.225)
0.027)

Exponent
0.54,
0.40,
0.07,
0.26,
0.47,

SE
SE
SE
SE
SE

0.23
0.26
0.28
0.29
0.30

(P
(P
(P
(P
(P

=
=
=
=
=

0.02)
0.13)
0.79)
0.37)
0.13)

717

Wear and surface topography in retrieved hip prostheses: R. M. Hall et al.

wear of the UHMWPE will probably not cover the

zone of contact between the bearing surfaces32. In
this situation, accelerated wear will occur only in
the zones which intersect with these loci while
adjacent areas will wear at a lower rate. Thus a
situation will arise in which the load is supported
by the lower rate zone only and the effect of the
highly damaged region will be negated, at least
intermittently.
It seems likely, however, that damage to the
femoral head does indeed cause excessive wear of
the polyethylene socket. Wroblewski et al.33 noted
that the surface roughness of four femoral heads,
which were coupled with sockets of low wear and
long implant periods, were relatively undamaged.
The problem lies in choosing a single or a range
of roughness parameters which may adequately
describe the relationship between topography and
wear. Multiple linear regression may be used, with
relatively uncorrelated independent variables, in
an attempt to identify correlations that may have
been masked by variations in the other explanatory parameters. Indeed, this approach was
applied here using Sk and Ra but, while the initial
results indicated an improvement in the strength
of the correlations between these variables and
the clinical wear factor, further investigations
revealed problems with data specification,
especially with respect to the arithmetic mean
roughness. Another possibility is to investigate
additional sets of parameters which may describe
more clearly the surface topography of the counterface and its effect on the wear of the socket.
An alternative may be those parameters that are
derived from the bearing area curve, such as the
reduced peak height or the material filled profile
peak area31. These parameters are related to the
asperities above the core roughness and may give
a better description of the material pile-up on
either side of the scratches. They will also give
some weighting to the length of the scratch, provided a three-dimensional image of the surface is
acquired. As a consequence this will overcome the
problems associated with the peak height parameters, Rpm and Rp, used here.
A further complication that can arise from the
measurement of the surface topography is the
problem of sampling. It is important to know
which results from a particular set of profiles or
areas are the most representative. In this study,
the mean values of the twenty profiles for each
femoral head, all gained from the presumed
region of contact, were used, whilst in the study
by Isaac et al.13 the maximum Ra from a series of
traces was utilised. Use of the maximum Ra value
is not a robust measure of the overall roughness
and may in itself be unduly influenced by exceptional values. As mentioned previously, the path
mapped out by the trace corresponding to the
maximum Ra value will most probably not cover
the whole contact zone. Therefore, it will not
exert such a large influence as might be expected
from the extrapolation of results gained in simple
screening tests in which the whole face of a polymeric pin comes in contact with a perpendicular
scratch on the metal counterface. It would be pru-

718

dent to suggest that the measurements should be

taken from the areas which appear to have the
greatest damage and that a sufficient number of
profiles has to be taken in order to get a representative sample. Further investigations need to be
undertaken to assess the minimum sample size
required both in terms of the number of evaluation lengths undertaken and the total joint
count. A standard protocol also needs to be
designed to assist in the comparison of results
between different research groups.
A further cause of the relatively poor correlation between roughness and wear is the variations in certain clinical parameters that are not
taken into account in the calculation of kclinical. In
particular, the activity rates of patients for any
given age were considered to be constant whereas
clinical evidence has shown it to vary by as much
as a factor of ten34. Other parameters not considered are the original diagnosis for primary surgery that may subsequently affect gait and activity.
5. CONCLUSIONS
The correlations between a selection of parameters characterising the surface texture and the
clinical wear factor were, in general, poor with
only two, the skewness, Sk, and the arithmetic
mean roughness, Ra, warranting further investigation. This is in contrast to the in vitro situation
where the correlation between Ra and k tends to
be stronger and the effect of the former on the
latter more pronounced. Factors accounting for
the difference between the clinical and laboratory
situation are the variability in the surface roughness on explanted femoral heads, the influence of
other clinical variables (e.g. patient activity),
three-body wear from cement and bone debris,
and that scratches may not affect the whole contact area. The use of laboratory wear tests, whilst
providing extremely useful information on the
fundamental tribology of orthopaedic materials,
may not accurately reflect the in vivo situation
even in the narrow confines of the surface topography of the bearing surface. In terms of specifying which parameters should be used in
assessing the surface texture of femoral heads in
relation to the wear of the socket, considerable
work still needs to be undertaken. In particular,
further investigation of the parameters derived
from the bearing area curve, for example the
reduced peak height and material filled profile
peak area, is urgently required. These parameters
are more representative of the peak topography
than the simpler parameters, Rp and Rpm, which
were found to be unsuitable in describing the
effects of texture on the wear of the UHMWPE.
ACKNOWLEDGEMENTS
The research has been funded solely by the
Arthritis and Rheumatism Council for Research
(ARC) under grant No. U0505 and the authors
wish to record their appreciation.

Wear and surface topography in retrieved hip prostheses: R. M. Hall et al.

REFERENCES
1. Howie, D. W., Haynes, D. R., Rogers, S. D., McGee, M.
A. and Pearcy, M. J., The response to particulate debris.
Orthopedic Clinics of North America, 1993, 24, 571581.
2. Harris, W. H., Osteolysis and particle disease in hipreplacementa review. Acta Orthopaedica Scandinavica,
1994, 65, 113123.
3. Wroblewski, B. M., Cementless versus cemented total hiparthroplastya scientific controversy. Orthopedic Clinics of
North America, 1993, 24, 591597.
4. Weightman, B. and Light, D., The effect of the surface
finish of alumina and stainless-steel on the wear rate of
UHMW polyethylene. Biomaterials, 1986, 7, 2024.
5. McKellop, H. and Clarke, I., Wear of artificial joint
materials in laboratory tests. Acta Orthopaedica Scandinavica, 1988, 59, 349352.
6. Dowson, D., Taheri, S. and Wallbridge, N. C., The role
of counterface imperfections in the wear of polyethylene.
Wear, 1987, 119, 277293.
7. Cooper, J. R., Dowson, D. and Fisher, J., The effect of
transfer film and surface roughness on the wear of lubricated ultra-high molecular weight polyethylene. Clin.
Mater., 1993, 14, 295302.
8. Dowson, D., El-Haby Diab, M.M., Gillis, B.J. and Atkinson, J.R., The influence of counterface topography on
the wear of ultra-high molecular weight polyethylene
under wet and dry conditions. In Proc. Int. Symp. on Polymer Wear and its Control, St Louis, 1984, publ. American
Chemical Society, New York, pp. 226232.
9. Fisher, J., Firkins, P., Reeves, E. A., Hailey, J. L. and Isaac,
G. H., The influence of scratches to metallic counterfaces
on the wear of ultra-high molecular weight polyethylene.
Proc. Inst. Mech. Engrs, 1995, 209H, 263264.
10. Cooper, J. R., Dowson, D. and Fisher, J., Macroscopic and
microscopic wear mechanisms in ultra-high-molecularweight polyethylene. Wear, 1993, 162, 378384.
11. Hailey, J. L., Ingham, E., Stone, M., Wroblewski, B. M.
and Fisher, J., Ultra-high molecular weight polyethylene
wear debris generated in vivo and in laboratory tests; the
influence of counterface roughness. Proc. Inst. Mech.
Engrs, 1996, 210H, 310.
12. Isaac, G. H., Atkinson, J. R., Dowson, D. and Wroblewski,
B. M., The role of cement and roughening upon the penetration rate of femoral heads into acetabular cups of
Charnley prostheses. J. Bone Joint Surg. [Br.], 1986, 68,
496.
13. Isaac, G. H., Wroblewski, B. M., Atkinson, J. R. and Dowson, D., A tribological study of retrieved hip prostheses.
Clin. Orthop., 1992, 115125.
14. Atkinson, J. R., Dowson, D., Isaac, J. H. and Wroblewski,
B. M., Laboratory wear tests and clinical observations of
the penetration of femoral heads into acetabular cups
in total replacement hip joints. 3. The measurement of
internal volume changes in explanted Charnley sockets
after 216 years in vivo and the determination of wear
factors. Wear, 1985, 104, 225244.
15. Wroblewski, B. M., Revision Surgery in Total Hip Arthroplasty. Springer-Verlag, London, 1990.
16. Jasty, M., Bragdon, C. R., Lee, K., Hanson, A. and Harris,
W. H., Surface damage to cobaltchrome femoral-head
prostheses. J. Bone Joint Surg. [Br.], 1994, 76B, 7377.
17. Hall, R. M., Unsworth, A., Siney, P. and Wroblewski, B.

18.

19.
20.
21.

22.
23.

24.
25.

26.
27.
28.
29.
30.

31.
32.

33.

34.

M., The surface topography of retrieved femoral heads.

J. Mater. Sci.: Materials in Medicine, 1996, 7, 739744.
Wong, W. P., Mainsah, E., Sullivan, P. J. and Stout, K. J.,
On instruments and measurement techniques of three
dimensional surface topography. In Three Dimensional
Surface Topography; Measurement, Interpretation and Applications, Vol. 1, ed. K. J. Stout. Penton Press, London,
1994.
British Standards Institution, British Standard BS 7251,
Orthopaedic joint prostheses. Part 4: Specification for
bearing surfaces of hip joint prostheses.
Braat, J. J. M., Read out of optical discs. In Principles of
Optical Disc Systems, 3rd edn, ed. G. Bouwhuis. Adam
Hilger, Bristol, 1987.
Visscher, M. and Struik, K. G., Optical profilometry and
its application to mechanically inaccessible surfaces. Part
1: Principles of focus error detection. Precis. Eng., 1994,
16, 192198.
Bousfield, B., Surface Preparation and Surface Microscopy of
Materials. John Wiley & Sons, Chichester, 1994.
Hall, R. M., Unsworth, A., Craig, P. S., Hardaker, C.,
Siney, P. S. and Wroblewski, B. M., Measurement of wear
in retrieved acetabular sockets. Proc. Inst. Mech. Engrs,
1995, 209H, 233242.
Kabo, J. M., Gebhard, J. S., Loren, G. and Amstutz, H.
C., In vivo wear of polyethylene acetabular components.
J. Bone Joint Surg. [Br.], 1993, 75B, 254258.
Elloy, M., Simulator testing of joint prostheses: The need
for realistic simulator testing. In Proc. SERC/IMechE
Annual Expert Meeting: Failure of Joint Prostheses, Bury St
Edmunds, 1993, publ. Mech. Engrs Pub. Ltd, pp. 7982.
Hall, R. M., Unsworth, A., Siney, P. and Wroblewski, B.
M., Wear in retrieved Charnley acetabular sockets. Proc.
Inst. Mech. Engrs, 1996, 210, 197207.
Paul, J. P., Forces transmitted by joints of the human
body. Proc. Inst. Mech. Engrs, 1967, 181, 815.
Wallbridge, N. C. and Dowson, D., The walking activity
of patients with artificial joints. Engineering in Medicine,
1982, 11, 9596.
StataCorp, Stata Statistical Software. Release 4.0. Stata Corporation, College Station, TX.
Caravia, L., Dowson, D., Fisher, J. and Jobbins, B., The
influence of bone and bone cement debris on the counterface roughness in sliding wear tests of ultra-high molecular weight polyethylene on stainless steel. Proc. Inst.
Mech. Engrs, 1990, 204H, 6570.
Sander, M., A Practical Guide to the Assessment of Surface
Texture. Feinpruf Perthen GmbH, Gottingen, 1991.
Bragdon, C. R., OConnor, D. O., Lowenstein, J. D., Jasty,
M. and Syniuta, W. D., The importance of multidirectional motion on the wear of polyethylene. Proc. Inst.
Mech. Engrs, 1996, 210H, 157165.
Wroblewski, B. M., McCullagh, P. J. and Siney, P. D.,
Quality of the surface finish of the head of the femoral
component and the wear rate of the socket in long term
results of Charnley low-friction arthroplasty. Proc. Inst.
Mech. Engrs, 1992, 206H, 181183.
Feller, J. A., Kay, P. R., Hodgkinson, J. P. and Wroblewski,
B. M., Activity and socket wear in the Charnley low-friction arthroplasty. J. Arthroplasty, 1994, 9, 341345.

719