To The Editor:
The Editorial entitled "A Potential Concern in Total
Joint Arthroplasty: Systemic Dissemination of Wear Debris" (82-A:
455-456, April 2000), by Clark, calls attention to wear particles
that are produced whenever there is articular or rubbing motion
in or against a prosthesis. Together with the article by Urban et
al.1 and an Instructional Course Lecture
on periprosthetic osteolysis and particulate debris2, these papers imply increased concern,
beyond that justifiably related to periprosthetic osteolysis, about
particles of wear debris in patients managed with total joint replacement.
Is that increase in concern about the particles’ abscopal distribution
justified, over and above the concern about their local toxicity?
The local consequences, such as osteolysis, indeed
justify significant concern, and the many attempts to minimize articular
wear attest to that concern. Concern about the abscopal particles
should be based on evidence and not on theoretical grounds; this
letter will consider in detail how strong, or more precisely, how weak
that evidence is.
Lest this commentary be interpreted as adverse criticism of the
publications cited, I endorse the pioneering efforts of these authors
in the field and I also point out that their work includes all of
the multidisciplinary techniques presently available, some of which
may be described as being at the forefront of modern technology.
Without reservation, therefore, the three published pieces under
scrutiny deserve fulsome praise. The following comments should be
regarded as evaluative rather than critical.
First of all, consideration should be given briefly to assessments of
the concentrations of the index substances—the metals (cobalt,
chromium, molybdenum, titanium, etc.) and the polymers (polyethylene
and methylmethacrylate)—in blood and urine. The many articles
on this subject all allude to the absence of evidence that the index
substances cause systemic pathological changes. The article by Urban
et al. and the Instructional Course Lecture contain data on the
concentration of such substances in blood and urine, but apparently
no data on systemic toxicity exist.
The specific concern is about articular wear, as attested by
the innumerable attempts that have been made to minimize it. The report
by Urban et al. primarily addresses fretting wear, in which one
or more prosthetic components rub against bone. Although articular
wear (especially of the polyethylene) might also have been involved, the
evidence points strongly to fretting, so that, in a sense, the article
by Urban et al. is not relevant to the issue of articular wear.
Emphasis is needed on two vital points. All thirty-one patients from
whom the index tissues (aortic lymph nodes, spleen, and liver) were
obtained had had the arthroplasty done years before. Only one was
young (eighteen years old), and most were seventy years old or older.
There were two cases in which the submitted tissue came from living
subjects, and those tissues were studied specifically because there
were concerns about the effects of observed particles in the tissues.
One may assume that the autopsy material was obtained from patients
who had had relevant treatment (arthroplasty or follow-up) at the institution
mentioned, who subsequently died, and who had an autopsy. We can
ignore such details as whether the study group was consecutive,
how many patients did not have an autopsy, and whether
there were some cases in which the index tissues were inadequately sampled
at the time of the autopsy. Therefore, I assume that the study included
all cases in which one could expect to find pertinent data and that
there were no exclusions of questionably eligible cases.
It should be understood from these assumptions that there had to
be a large number (possibly thousands) of arthroplasties performed
at the institution from which twenty-nine of the thirty-one patients
were drawn. All twenty-nine patients died of diseases presumably
unrelated to the particle problem, and, with one exception, did
so at an advanced age. If toxicity from metal or polymer were to
be implicated in shortening a patient’s life span, one
would have to either ignore or take into account other major possibilities for
the shortening, such as the disease process that resulted in the
lesion at the involved joint(s), the drug treatment (if any), and any
comorbidity that existed (and in the age-group under consideration
there surely was comorbidity).
Finally, in none of the cases was osteolysis recorded.
Clark’s Editorial provides the logical pathophysiological explanation
of the waste-management sequence—why particles would migrate
to the index tissues. He postulates that the macrophages phagocytose the
particles produced locally and transport them to the waste-management
facilities, which for the lower extremities are primarily the aortic
lymph nodes and secondarily the spleen, liver, and perhaps the lungs.
He further postulates that the primary facility may be overwhelmed
so that secondary facilities are called upon.
The question then arises: why should these sequences be a cause
for concern? Should one worry that the primary mechanism for removal
of implanted particles is overwhelmed and that secondary waste-management
facilities then come into play? Worry is warranted only if there
are important pathological lesions involving the facilities. However, the
evidence that Urban et al. addresses also bears on the question
of what the concentrations of implant particles were. In Table III,
the particles are labeled "metal-like" and "polyethylene-like," signifying
that microscopic observation could not prove that the particles
were metal or polyethylene; one has to assume, for the purposes
of the study, that they were. A troubling finding here is that,
in one patient (Case 15), the highest concentration of particles
was detected but no chemical evidence of metal was found.
If we assume that patients with the highest concentrations of particles
would be most likely to show pathological changes that justify "concern" and
also that the observed particles indicate higher concentrations
of particles too small to be detected, we note that five patients
(omitting Case 15 mentioned above) had an important concentration
of particles in the lymph nodes. All but one of these five patients
had had at least one revision of the original arthroplasty. Does
this mean that concern about the particles should focus on cases
in which revision arthroplasty, not primary arthroplasty, was done?
These five patients also were evaluated with regard to the concentrations
of particles in the spleen and liver. The data are confusing, to
say the least. If one establishes the concentration of fifty particles
per high-power field as the threshold of importance (as we have
done in defining the group of five patients mentioned above), none of
these patients had an important concentration of particles in the
liver, whereas two had an important concentration in the spleen. These
concentrations must be evaluated in the context of concentrations
in the controls (no arthroplasty), all of whom showed some particles,
often ten to nineteen per high-power field, and in one case, fifty
to 499 per high-power field.
We turn now to the pathological findings. Setting aside the microscopic
observation of particles in macrophages in all three index tissues,
with occasional foreign-body giant cells, it should be emphasized that
only in the lymph nodes was there chemical evidence
that the particles consisted of metal or polyethylene, and only
in the lymph nodes were other important pathological findings noted.
One would expect to encounter a few or even many particle-bearing
macrophages in some patients, especially those who have had revision arthroplasty.
Of greater importance were the observations that seventeen patients
had granulomatous inflammation and ten had fibrosis. One assumes
that these two groups involved only patients who had had multiple operations
and, perhaps, that they involved mostly patients who had had loosening
of the prosthesis. In "several" patients (all
of whom had had revisions) the nodes stained black, which was attributed
to large numbers of metal particles. Perhaps it is important that,
of the ten cases in which revision arthroplasty was done, all but
one involved the insertion of prosthetic components that contained titanium.
It is well known that titanium oxides (which are part of the surface
film that protects the titanium-bearing alloy from progressive corrosion
but are brittle and therefore not suitable for the articulating
surface of a prosthesis) are black or brown.
None of the four important pathological findings (granuloma,
fibrosis, necrosis, or gross replacement of parenchyma) were evident
in either the spleen or the liver, except in one patient who had
a comorbidity (systemic granulomatous disease).
What conclusions can be drawn from the evidence described in the
three publications cited above? When it comes to risk, an argument
for the existence of important risks—demonstrable pathological
changes attributable to disseminated particles—is supportable
only when the patient has had a revision arthroplasty and not just
a primary arthroplasty, either unilaterally or bilaterally. In the study
by Urban et al., all twenty-one patients in whom only primary arthroplasty
had been done had collections of particle-bearing macrophages that represented
physiological rather than pathophysiological reactions per
se. None of the histological abnormalities represented
a threat to life or even to continuation of function of the viscera
(spleen, liver) or tissues (lymph nodes), with the possible exception
of lymph-node necrosis in one patient. In that case, multiple revisions
of a total hip replacement had been performed so that during the years
between operations and before death, there was ample opportunity
for increasing toxicity to develop.
One may speculate about the "potential risk" not
only from disseminated particles (as well-demonstrated in the article
and Editorial cited above) but also from those solubilized in blood (and
therefore in all tissues, including the brain). One may argue that
the evidence described by Urban et al. is the tip of the iceberg
in that it pertained to only thirty-one out of the million or more
arthroplasties performed in this country and represented only patients
who had an autopsy, whatever the cause of death. The iceberg should include
a multitude of patients in whom the potential for particle formation
has persisted for more than the few years that it persisted in the
elderly subjects studied by Urban et al. For the moment, the concern
may be appropriate for patients who have had repeated arthroplasties, but
the orthopaedic surgeon in practice should not extrapolate that
concern to patients who are about to undergo primary total joint
replacement. To do so would amount to scaring patients unduly.
C.R. Clark replies:
Dr. Cohen has raised an important question: is the increased
concern about the abscopal distribution of wear particles justified,
over and above the concern about their local toxicity? In response,
I suggest that one needs to carefully review my Editorial, including
the title. The key word is "potential." I believed
then, as I do now, that raising a concern should provide a stimulus
for further study. Patients need to be followed long-term, and I
believe that it is better to be vigilant, particularly as the number
of total hip and knee arthroplasties being performed in this country is
increasing and we are performing these operations on yet younger
patients. Dr. Cohen correctly points out the paucity of information
on this subject. Raising such awareness allows others to be vigilant
and to share their observations. Indeed, in the scientific process,
it is important not only for one individual to make an observation
but also for this observation to be repeated by others.
One particular issue that Dr. Cohen raises relates to the source of
wear debris. There is an important difference between articular
wear and fretting wear, and a great deal of effort has been expended
to evaluate the former. The point, however, is that all wear debris
is a potential concern, especially if there is systemic dissemination.
Two other issues raised by Dr. Cohen are the relatively small number
of cases and the older age of the patients in the study by Urban
et al. I agree that the overall number of cases (thirty-one) is
extremely small when one looks at the more than one million arthroplasties
that have been performed in this country; however, one must appreciate
the fact that the source of this information was autopsy specimens,
which were obtained in twenty-nine of their thirty-one index cases1. This is actually quite a large series
of specimens obtained in this manner. In addition, it should be
noted that autopsy specimens are more likely to be obtained from
older patients. Furthermore, since the age of patients undergoing
arthroplasty is now decreasing, it is not surprising that
patients treated in the past would have a greater mean
age. This, however, further underscores the importance of exploring
this issue as we are performing total joint replacements in younger patients.
Dr. Cohen comments on my postulates regarding phagocytosis by
the macrocytes and the "overwhelming of the system." I
based these postulates on the study by Urban et al. as well as other
studies3 cited in my Editorial.
I do believe that Dr. Cohen has raised an important question: what
is meant by "overwhelm"? The answer is not known
and the question is, in and of itself, another stimulus for further work.
I agree that one should worry if an important pathological lesion
is indeed present. However, this may not be determined unless one carefully
observes patients over time. I believe that there is a potential
concern and, with the increasing numbers of total joint replacements
being performed and the decreasing age of patients undergoing these procedures,
vigilance is warranted. It would be terrific if a pathological lesion
of consequence does not result from this "overwhelming
of the system."
An additional point made by Dr. Cohen relates to the preponderance
of findings in patients with at least one revision of the original
arthroplasty. It should be remembered that as the age of the patient
at the time of arthroplasty decreases (which also implies a longer
remaining life span), the risk of undergoing one or more revision arthroplasties
in the future increases. Another point raised by Dr. Cohen relates
to the presence of pathological findings in a patient with an important comorbidity
(in this case, systemic granulomatous disease). Perhaps patients
with such comorbidities should be the focus of our observations.
Dr. Cohen points out that none of the pathological findings appeared
to represent a threat to life; however, in patients undergoing total
hip and knee arthroplasty, it is often the quality of life that
is the most important concern.
The purpose of my Editorial was not at all to unduly scare patients.
I believe that as responsible physicians we need to be vigilant
regarding the long-term effects of our operative interventions.
As we are operating on younger patients, one or more revisions become
a reality, and we do not know the long-term effects thirty, forty, and
fifty years postoperatively. I believe it is appropriate to raise the
issue, to carefully observe these patients, and to report findings
as they are determined. Indeed, such observations are the cornerstone
of scientific advancement, which is so important to any medical discipline.