To date, little has been reported about the effect of arthroscopy
on the long-term course of a total joint replacement. A chance observation
during arthroscopy of a total knee replacement suggested a possible
adverse effect and led to this study. While performing an arthroscopic
lysis of adhesions in a knee with a total replacement, we observed
a linear alteration in the smoothness of the femoral component very
close to where we had passed the arthroscopic cannula. Therefore,
we decided to determine how easily a prosthetic surface could be
affected by the arthroscope and to determine the type of damage
that could be created. We also wanted to outline some of the parameters,
such as the force necessary to produce the alterations, the roughness
produced, and the relative risks of plastic and stainless-steel cannulae.
The first of two hypotheses was that the stainless-steel cannula
would produce surface damage to a femoral component of a total knee
replacement when forces typically encountered during arthroscopy
were applied to the interface between the cannula and the component.
The second hypothesis was that the use of a plastic cannula would
reduce or eliminate this risk of damage to the femoral component.
Specimens and Materials
Two "virgin" implant-grade cobalt-chromium total knee replacement
femoral prostheses (Advantim Posterior Stabilized Femoral Implant,
lot 116A047200, and Advantim Total Condylar Femoral Implant, lot
068A082517; Wright Medical Technology, Arlington, Tennessee) were
tested. Forces were applied to one femoral component with a "virgin"
stainless-steel arthroscopic cannula (lot 22940101; Arthrex, Naples,
Florida) with an inner diameter of 5.45 mm and an outer diameter
of 6.35 mm, and forces were applied to the other component with
a disposable plastic cannula (Disposable Arthroscopic Cannula, lot
327338; Smith and Nephew, Memphis, Tennessee) with an inner diameter
of 6.5 mm, an outer diameter with threads of 8.96 mm, and an outer
diameter of 8.00 mm. At all times, the femoral components were handled
with gloves to limit surface contamination and damage. Prior to
testing and imaging, each component and cannula was cleaned thoroughly.
The components were scrubbed with a soft nylon brush under running
tap water and then soaked in a 2% detergent solution (Alquinox liquid)
in water and ultrasound for ten minutes. Then they were rinsed in
deionized water in an ultrasonic bath for ten minutes. Finally, the
components were air blow-dried, soaked in 100% ethanol for ten minutes
in the ultrasonic bath, and placed in a desiccator overnight.
Mechanical Testing
The femoral components were held in a specially designed chamber
during mechanical testing to allow for their submersion in normal
saline solution to simulate the arthroscopic environment. In addition,
the mounting and testing system allowed for an accurate method of
positioning for controlling the location of the forces applied during
the loading procedure. This setup also provided alignment of the
cannula and the femoral component such that the applied forces were
perpendicular to the interface (normal force) between the condylar
and cannular surfaces. The femoral components were positioned in
the chamber such that their condylar surfaces were facing up with
the cannula crossing the surface in the anterior-posterior direction.
At a predetermined location, a constant compressive normal force
was applied to the cannula with use of a servohydraulic testing
machine (model 8521; Instron, Canton, Massachusetts) while the cannula was
moved manually across the condylar surfaces in a linear cycle. Lacking
clinical references for the range of forces encountered during arthroscopy,
we considered a force of about 20 lb (89 N) to be well above a feasible
maximum compressive normal force between the cannular and condylar
surfaces. Therefore, twelve load steps from 4 to 100 N (4, 8, 16,
25, 35, 45, 55, 65, 75, 85, 95, and 100 N) were applied for one
cycle.
Scanning Electron Microscopy
The femoral condylar surfaces to be imaged were first hand-marked
(following cleaning) with a grid of 2 by 2-mm squares with use of
a marking pen to assist in component orientation and surface characterization.
The grid was lettered from A to Q in the medial-lateral direction
and numbered from 1 to 15 in the anterior-posterior direction. In
addition, a scanning-electron-microscopy finder grid (catalog number
16060; Ted Pella, Redding, California) was placed on the intercondylar
surface to further assist in component orientation under scanning
electron microscopy. Prior to and after each mechanical testing
session, the femoral components were examined with a scanning electron
microscope (model XL 30 FEG; Philips Electronics, Mahwah, New Jersey).
Three different types of scanning electron microscopy images were
obtained: (1) a secondary electron image-a standard scanning electron microscopy
image showing all of the details of the component surface structure;
(2) backscatter mode with atomic number contrast-to show variations
in the atomic number of the surface material (that is, cobalt-chromium
versus stainless steel, the contaminant); and (3) backscatter mode
with topographical contrast-to give a true determination of height variation
in the component surface present initially and the damage resulting
from mechanical testing.
In addition to the images, energy dispersive x-ray analysis was
performed in selected areas to identify variations found with the
backscatter mode with atomic number contrast.
Measurements of Surface Roughness
At each test location where damage was observed with scanning
electron microscopy, the surface roughness was measured with scanning
white-light interferometry with a ten times Mirau objective and a
0.72 by 0.54-mm field of view (NewView 200; Zygo, Middlefield, Connecticut).
Regions that were adjacent to the test sites but that did not show
damage were also measured, as baseline controls. For both femoral
components, twelve test and twelve control measurements were performed,
resulting in twenty-four measurements for each component. For each
of the twenty-four measurement sites, seven surface-roughness parameters
were determined: Ra-the average surface roughness, or average deviation,
of all points from a plane fit to the surface; rms-the root-mean-square
average of the measured height deviations, which is an alternative to
Ra that gives greater importance to features farther from the mean
plane; Rz-the ten-point, or average, absolute value of the five
highest peaks and the five lowest valleys over a selected line profile
(also known as the ISO 10-point height parameter); PV-the maximum
peak-to-valley height over the area or length of evaluation; Rtm-the
average peak-to-valley roughness determined by the difference between
the highest peak and lowest valley within multiple samples in the
evaluation area; H-the Swedish height, which is defined as roughness between
two predefined reference lines, a parameter that is less sensitive
than Rtm; and Rsk-the skewness of the data (distribution of peaks
and valleys), with a positive value meaning that the roughness is
above the surface (more peaks) and a negative value indicating that
the roughness is below the surface (more valleys)1,2.
For each site, the parameters were calculated from the entire
field of view (area) and the average of ten different line profiles
across the field of view1,2.
Data Analysis
The scanning electron microscopy images were subjectively evaluated,
and areas of deposition and scratching were identified for the two
components. Box plots were made from the surface-roughness data.
In addition, t tests were performed to compare the surface roughness
between the test and control regions. Correlation and regression
analysis was done to determine possible relationships between the
applied load and the surface roughness of the component.
Stainless-Steel Cannula
Gross Observations
Visible alterations to the surface of the cobalt-chromium femoral
component were produced. Disruption of the component grid markings
made determination of the precise locations of the surface alterations
easy. To the unaided eye, these marks appeared as scratches, resulting
in a slight loss of the component's mirror finish.
Scanning Electron Microscopy
Under scanning electron microscopy, the true character of the
damage was observed. Even at the lowest applied forces of 4 and
8 N, and at the lowest magnification used (twenty-five times), the
cobalt-chromium surface was altered (Figs. 1-AFigs. 1-A and 1-B1-B). The topographical
contrast backscatter images revealed that the majority of the new
surface alterations were not scratches into the cobalt-chromium surface,
but rather were the deposition of material on the surface of the
component (Fig. 1-AFig.
1-A). The atomic number contrast revealed that a substance with
atomic numbers different from those of cobalt-chromium had been
applied to the surface (Fig. 1-BFig. 1-B). Energy dispersive x-ray
analysis confirmed this new substance to be consistent with stainless
steel3.
The tracks of stainless steel on the component surface varied
greatly in size and character. The average width ranged from 265
to 763 mm, with a mean of 522 mm and a standard deviation of 127
mm. Tracts ranged in length from 4.4 to 9.2 mm, with a mean of 5.82
mm and a standard deviation of 1.53 mm. Some areas of deposit appeared
as a single path, while other test areas consisted of two separate
paths next to one another. These stainless-steel paths varied greatly
in gross appearance (Fig. 2Fig. 2). We did not measure the
thickness of the stainless-steel deposits.
In addition to the application of stainless steel to the surface
of the component, some scratches of the cobalt-chromium surface
were also observed, at loads of as little as 8 N (Fig. 1-AFig. 1-A).
These scratches varied in width, length, and orientation to the
moving cannula. Some were parallel to the motion of the cannula,
while others ran obliquely or almost perpendicular. Some scratches had
a modest buildup of cobalt-chromium at the edges and contained stainless-steel
deposits in their depths (Figs. 1-AFigs. 1-A and 1-B1-B).
Surface Roughness
The surface roughness of the test regions, given by Ra, showed
a correlation with the applied load (r = 0.6005) (Fig. 3Fig. 3), while
the surface roughness of the baseline control regions did not show
a correlation with the applied load. All of the roughness parameters except
for Rsk correlated with load. However, when the baseline control
and test regions were compared, all seven parameters demonstrated
significant changes (p < 0.05) (Fig. 4Fig. 4). Since the measured surface
roughness is of the deposited stainless steel and not of the cobalt-chromium
surface, the data suggest that, as the applied load increased, the
surface of the deposited stainless steel became rougher.
Plastic Cannula
Gross Observations
As the plastic cannula passed over the component surface, it
merely rubbed off regions of the marked grid. Gross alterations
to the mirror finish of the component were not noted.
Scanning Electron Microscopy
Alterations of the cobalt-chromium surface due to the plastic
cannula appeared to be minimal. Even at the highest applied loads
(95 and 100 N), scanning electron microscopy revealed no track deposits. Other
than rubbing off the ink grid markings, the plastic cannula did
not seem to leave behind any surface changes (Fig. 5Fig. 5).
Surface Roughness
No significant correlations were found between the surface roughness
and the load for either the control or the test regions. Comparison
of the surface roughness between the control and test regions showed
no significant changes with regard to any of the seven parameters
(Fig. 4Fig.
4).
Cannular Changes
While not a planned part of this study, we also observed changes
in the plastic cannula after completion of the test. Scratches and
gouges were produced in the cannula. In addition, the surface threads
were severely damaged in some regions and simply compressed in other
regions. Damage to the cannula occurred even with the small force
of 4 N.
Total knee replacement is a highly successful procedure. From
1990 to 1995, the American Academy of Orthopaedic Surgeons reported
an increase in the number of total knee replacements performed in
the United States from about 129,000 to 216,0004.
While the vast majority of total knee replacements yield satisfactory
pain relief and the rate of prosthetic survival is 90% at ten to
fifteen years, there are a number of postoperative aseptic complications
that result in continued discomfort and limit the range of motion5. Arthroscopic evaluation and treatment
can be used to resolve many of these complications by removal of
loose bodies and resection of adhesions and scar tissue resulting
from total knee replacement6-8.
Care must be taken not to scratch the metal components during arthroscopy,
since alterations in the surface character of cobalt-chromium have
adverse effects on the wear of total joint replacements7,9.
The potential for one metal to transfer to another has been well
known and understood for many years10.
More recently, a clinical case of metal transfer from an acetabular
shell to a ceramic femoral head has been reported11.
Damage to and degradation of the articulating surfaces of a total
knee replacement have been associated with release of wear debris12,13. There is a correlation between
surface roughness of cobalt-chromium femoral components and polyethylene
wear of the tibial component12,14.
In addition, studies have shown extensive foreign-body giant-cell
reactions to polyethylene particles and synovial membrane reactions
to loose cobalt-chromium particles15,16.
Our study demonstrated that use of a stainless-steel cannula,
with forces as low as 4 N (approximately 1 lb), altered a cobalt-chromium
surface by depositing stainless steel on it. Although some minor scratching
was observed (Figs. 1-AFigs. 1-A and 1-B1-B), it was infrequent;
deposition of stainless steel from the cannula was the main observable
surface alteration. The resulting surface roughness increased with
increasing loads. In contrast, no observable change to the cobalt-chromium
surface was found after use of the plastic cannula. The plastic
cannula showed extensive abrasion and surface damage even at low
forces (4 and 8 N). Such damage to the plastic cannula could create
plastic wear debris, possibly causing a problem with third-body wear.
The results of this study support a preference for using plastic
cannulae rather than metal cannulae to avoid permanent changes in
the femoral surface. A plastic shell or cover for the arthroscope
might also be suggested by the results of this study. However, if
a plastic cannula or a plastic sleeve is used, copious irrigation
and evacuation at the end of the surgery are necessary to remove
plastic particles and decrease the risk of third-body wear.
Although we demonstrated the metallic transfer from a stainless-steel
cannula to a cobalt-chromium femoral component, the long-term effect
of this deposition was outside the scope of this study. Wear-testing
of femoral components with similar stainless-steel deposits would
be necessary to ascertain the potential cost with regard to increased wear
and reduction in the life expectancy of the total knee replacement.
The influence of the amount and orientation of the deposition, the
alteration in the surface roughness of the cobalt-chromium surface,
and the location of the surface damage on the wear of the prosthesis
also need to be investigated to fully assess the impact of arthroscopic
evaluation and treatment on the success and longevity of total knee
replacement.
Note: The authors express their thanks and gratitude to Krassimir
Bozhilov, Manager of the Analytical Electron Microscopy Facility
at the University of California, Riverside, for his assistance with
the scanning electron microscopy and energy dispersive x-ray analysis;
to Wright Medical and Team Surgical for their contribution of the
specimens used in the study; and to Zygo Corporation for its assistance
and cooperation in the surface-roughness measurements. The MacPherson
Society of Loma Linda University School of Medicine financially supported
this work.