To The Editor:
I read with interest the article "Relationship Between Polyethylene
Wear and Osteolysis in Hips with a Second-Generation Porous-Coated Cementless
Cup After Seven Years of Follow-up" (2003;85:1095-9), by Orishimo et al.
However, the study methodology and the interpretation of the results are
questionable.
First, the study design analyzes retrospective radiographs, which
presumably were not performed as part of the current study and may not have
been sufficiently controlled to provide standardized projections. I am part of
a team involved in radiographic assessment of total hip replacements, and one
problem that we encountered when attempting to measure cup wear was the
apparent alteration in cup inclination due to minor differences in patient
positioning, which produced apparent differences in polyethylene wear. It
seems that, even with the most exact measuring equipment, if radiographs are
not standardized, then inconsistencies will occur.
Another drawback of the study is the potential bias during the measurements
of osteolysis. These observations appear to be an "eyeball"
technique; however, the authors refer to their study as
"quantitative." Couldn't the same computer software (that was used
to measure wear) be used to quantify the degree of osteolysis present? Was
there any effort to blind the observer from the presence of wear in the
radiographs?
Their interpretation of these results assumes that the increased production
of wear particles results in an increased prevalence of osteolysis through a
biological effect. However, the study does not have direct evidence for this
assumption. Could it be possible that loosening may be a direct result of
increased mechanical loading in these patients (who are of a younger age)?
Could osteolysis or loosening, therefore, be a marker of activity level?
Concerning the statistical analysis, it is incorrect to equate odds ratio
with relative risk. Only when a risk factor occurs rarely in a study
population does the odds ratio approach the relative risk ratio. In their
study, which had a commonly occurring risk factor, the odds ratio that they
calculated will overestimate the relative risk.
They conclude that patients with increased wear rates should be more
closely monitored. However, in practice, the resources in many hospitals
cannot justify the use of scanning equipment (and software), especially when
the study by Orishimo et al. could not prove any benefit over the traditional
methods of detecting early osteolysis with radiography in the outpatient
clinic, which is probably sufficient to activate necessary concern.
We appreciate Mr. Chamber's comments on our study and recognize his concern
regarding the methods and interpretation employed therein. First, we agree
that standardized radiographs are of paramount importance when it comes to the
accuracy of polyethylene wear measurements. Because of the high patient volume
at our institution, we routinely obtain long-term radiographic and wear data
on all of our patients. With this in mind, we have instituted strict
guidelines regarding patient positioning for the standardization of
radiographs. These guidelines have been in place for many years and have been
strictly followed by the same radiography technicians at our institution. As
reported by Moore et
al.1, our routine
clinical procedure includes an anteroposterior radiograph of the entire pelvis
made with the patient in the supine position and with both lower limbs in
maximum internal rotation. If the patient is unable to keep the lower limbs
internally rotated, positioning aids are placed around the feet to maintain
internal rotation while the radiograph is made. No other positioning devices
are used. For these radiographs, the tube-to-film distance is 1.016 m and the
x-ray beam is centered on the pubic symphysis. Our studies have shown no
difference in measurements of polyethylene wear on radiographs made with the
patient in the aforementioned supine position compared with those made with
the patient bearing weight, and good reproducibility has been
documented1.
Second, regarding the measurement of osteolysis, one author (A.M.C.)
independently reviewed each patient's series of radiographs and outlined the
areas of osteolysis with a pencil. This observer was blinded to the amount of
wear found in each hip. The same computer software used to measure
polyethylene wear (Hip Analysis Suite 4.0) also has the capability to measure
the two-dimensional radiographic area of osteolysis. However, because of the
limitations in assessing the three-dimensional size of an osteolytic lesion
with plain radiographs, we believed that the two-dimensional lesion area
calculated by the software would not accurately represent the entire
lesion2. Therefore,
in our paper, we chose to describe osteolysis simply in categorical
terms—either present or absent. Given the categorical data, we were able
to use logistic regression to obtain quantitative results regarding the
effects of continuous variables such as polyethylene wear and length of
follow-up on the prevalence of osteolysis.
Third, the biologic response to polyethylene particles has been well
documented in the
literature3-5.
In addition, the literature has demonstrated a positive correlation between
the amount of polyethylene particles generated and the prevalence of
osteolysis. That is, for a given group of patients, the prevalence of
osteolysis increases with increasing wear
rates6-9.
Given this information, it is very important to distinguish between the causes
of osteolysis and the causes of polyethylene wear. Factors such as
polyethylene type, implant design, and patient activity have been known to
influence the amount of particles generated. However, the development of
osteolysis is influenced by the presence of the polyethylene particles
themselves and the individual's sensitivity to wear debris. Therefore, while
we found an association between wear rates and osteolysis, the prevalence of
osteolysis cannot be used as an indicator of activity level because activity
level influences the rate of particle production and not the response to those
particles. Furthermore, in our study, we only assessed stable implants and,
therefore, can only report on the relationship between wear and osteolysis
with stable implants. No statements about osteolysis and loosening can be
derived from our data.
Fourth, regarding the assessment of relative risk, we were previously
unaware that the odds ratio can only be equated with relative risk when an
outcome is relatively rare. After further review of the statistical nuances,
we found that the quantification of what constitutes "rare" is up
for debate; however, Zhang and Yu suggest a figure of
<10%10. When the
outcome occurs in a population more frequently than 10%, Zhang and Yu offer a
correction for the odds ratio to provide a measure that more accurately
represents the concept of relative risk for that population. Given the
prevalence of osteolysis in our population of 41% and an odds ratio of 4 for
wear rate, the adjusted relative risk (calculated with use of the formula
derived by Zhang and Yu) is approximately 2. This indicates that for every 0.1
mm/year increase in wear rate, a person's risk for the development of
osteolysis increases twofold. Using the Zhang formula for volumetric wear, we
find that for every increase in volumetric wear of 40
mm3/year, a person's
risk for the development of osteolysis increases 1.5-fold. These relative risk
ratios are approximately one-half of those described in the paper. We thank
Mr. Chambers for bringing this rare assumption rule to our attention.
Finally, we acknowledge the fact that scanning equipment and software for
quantifying polyethylene wear is not available at many institutions. However,
cases of rapid wear are often detected (although not quantified) by the
trained orthopaedic eye. For those patients who present with increased wear,
we recommend close radiographic monitoring for the presence and future
development of osteolysis. Several radiographic views (anteroposterior pelvic,
anteroposterior femoral, 45° iliac oblique, and obturator foramen
radiographs) should be thoroughly assessed to detect osteolysis at early
stages, keeping in mind the limitations and merits of radiography for the
detection of pelvic osteolysis.
Moore KD, Barrack RL, Sychterz CJ,
Sawhney J, Yang AM, Engh CA. The effect of weight-bearing on the
radiographic measurement of the position of the femoral head after total hip
arthroplasty. J Bone Joint Surg Am.2000;82:
62-9.8262
2000
[PubMed][CrossRef]
Claus AM, Engh CA Jr, Sychterz CJ,
Xenos JS, Orishimo KIT, Engh CA Sr. Radiographic definition of pelvic
osteolysis following total hip arthroplasty. J Bone Joint Surg
Am.2003;85:
1519-26.851519
2003
Fisher J, Bell J, Barbour PS, Tipper
JL, Matthews JB, Besong AA, Stone MH, Ingham E. A novel method for the
prediction of functional biological activity of polyethylene wear debris.
Proc Inst Mech Eng [H].2001;215:
127-32.215127
2001
[PubMed]
Green TR, Fisher J, Matthews JB,
Stone MH, Ingham E. Effect of size and dose on bone resorption activity of
macrophages by in vitro clinically relevant ultrahigh molecular weight
polyethylene particles. J Biomed Mater Res.2000;53:
490-7.53490
2000
[PubMed][CrossRef]
Sethi RK, Neavyn MJ, Rubash HE,
Shanbhag AS. Macrophage response to cross-linked and conventional UHMWPE.
Biomaterials.2003;
24: 2561-73.242561
2003
[PubMed][CrossRef]
Barrack RL, Castro FP Jr,
Szuszczewicz ES, Schmalzried TP. Analysis of retrieved uncemented
porous-coated acetabular components in patients with and without pelvic
osteolysis. Orthopedics.2002;25:
1373-8.251373
2002
[PubMed]
Harris WH. Wear and
periprosthetic osteolysis: the problem. Clin Orthop.2001;393:
66-70.39366
2001
[PubMed][CrossRef]
Oparaugo PC, Clarke IC, Malchau H,
Herberts P. Correlation of wear debris induced osteolysis and revision
with volumetric wear rates of polyethylene: a survey of 8 reports in the
literature. Acta Orthop Scand.2001;72:
22-8.7222
2001
[PubMed][CrossRef]
Dowd JE, Sychterz CJ, Young AM, Engh
CA. Characterization of long-term femoral-head-penetration rates.
Association with and prediction of osteolysis. J Bone Joint Surg
Am.2000; 82:
1102-7.821102
2000
Zhang J, Yu KF. What's the
relative risk? A method of correcting the odds ratio in cohort studies of
common outcomes. JAMA.1990;
280: 1690-1.2801690
1990
[CrossRef]