Copyright © 2007 by The Journal of Bone and Joint Surgery, Inc.
Commentary & Perspective
Commentary & Perspective by
Thomas W. Bauer, MD, PhD*,
Cleveland Clinic Foundation, Cleveland, Ohio
Posted July 2007
The erythrocyte sedimentation rate (often called the ESR or the
"Sed Rate") is one of the oldest laboratory tests still in use. It expresses
the rate at which cells sediment in blood that has been anticoagulated with sodium
citrate or ethylenediaminetetraacetic acid (EDTA). The erythrocyte
sedimentation rate reflects the relative density of the cells with respect to
the plasma, and it is influenced by, among other things, the ability of the
cells to form rouleaux. The surface attraction between erythrocytes is in turn affected
by the proportions and types of proteins in the plasma. For example, increased
fibrinogen and alpha, beta, and gamma globulins decrease the negative charges that
normally keep red cells apart. This causes increased rouleaux and more rapid
sedimentation1. Plasma viscosity, hematocrit, and cell shape also
influence the sedimentation rate. The test is quite nonspecific, and an increased
erythrocyte sedimentation rate can be seen in inflammatory bowel disease, rheumatoid
arthritis and other inflammatory arthropathies, cardiac disease, cancer and
lymphoma, gout, hepatitis, cirrhosis, systemic lupus erythematosus, and
pregnancy, among other conditions. Blood that has been anticoagulated with
heparin, either in vivo or in vitro, also may have an elevated erythrocyte
sedimentation rate.
In 1930, Tillett and Francis identified a protein that
precipitated with the C-polysaccharide antigen of the pneumococcus bacterium2.
C-reactive protein is now known to be an "acute phase reactant" that is produced
in the liver and has a biologic activity similar to immunoglobulins in that it
binds membranes of microorganisms and cancer cells. In so doing, it opsonizes those
membranes and activates complement to help initiate an inflammatory reaction
prior to the production of specific antibodies. Similar to the erythrocyte
sedimentation rate, the C-reactive protein level is elevated in a number of
inflammatory disorders, including inflammatory bowel disease, inflammatory
arthropathies, lymphoma, and infections. It increases and normalizes more
rapidly than the erythrocyte sedimentation rate does, and a persistent, mild
increase in the C-reactive protein level is associated with an increased risk
of coronary artery disease and stroke.
While these two laboratory tests are quite nonspecific, they
are rather sensitive, so many previous studies have documented their utility as
screening tests for diagnosing orthopaedic infections3-5. In the
July 2007 issue of The Journal, Greidanus
and coauthors describe a prospective study intended to quantify how good these
tests really are in diagnosing periprosthetic infections.
From 1997 to 2001, the authors measured the serum C-reactive
protein level and the erythrocyte sedimentation rate in all patients who were
scheduled to undergo revision knee arthroplasty at one center. Each consenting
patient then underwent aspiration of the affected knee, and that aspirated
fluid specimen was divided into three samples for microbiologic culture. Each
patient also had three cultures obtained from the knee at the time that the implants
were removed. Knees that had at least two positive fluid cultures or two
positive operative cultures were considered to be infected; knees that had
fewer positive cultures were considered not infected. The sensitivity and
specificity of each test was then calculated with use of the culture results as
the so-called gold standard. In addition, the authors used receiver-operating-characteristic
curve analysis to identify the optimal cutoff points (i.e., threshold values)
for distinguishing positive from negative results in their hospital. The
results confirmed that these are both good screening tests, and that used
together they are even better, but there are several important caveats to be considered
when looking at the results of this study.
First, one might argue that the use of two positive culture
results to define infection is an imperfect "gold standard." But of more
importance is the very nature of the test population itself. Of the original
201 patients, twenty-six were excluded because they had an underlying disease
or condition known to be associated with elevated erythrocyte sedimentation
rate and/or C-reactive protein level. Some of these patients had rheumatoid
arthritis, but we do not know the exact reasons for excluding the other
patients. This is important, because anyone attempting to reproduce these
results would need to use the same exclusion criteria. Nineteen patients who
presented for second-stage revision for a known knee infection were also
excluded, and eleven patients who were receiving antibiotics were excluded from
the primary data set but were analyzed separately.
Of the remaining 151 knees, forty-five (approximately 30%)
were diagnosed as infected. This extremely high prevalence of infection accurately
reflects the referred patient base of the authors, but it is much higher than
that seen by most orthopaedic surgeons and has important implications
concerning the predictive value of the test results for C-reactive protein
levels and erythrocyte sedimentation rates. A laboratory test with a relatively
high frequency of false-positive results may seem like a good test when it is
used in a population with a high prevalence of disease; however, if that same
test is applied to a population with a low prevalence of disease, a high
proportion of the results may be false-positives.
For example, with use of the cutoff values for erythrocyte
sedimentation rate that the authors calculated from the receiver-operating-characteristic
curves (sensitivity = 93%, specificity = 83%, prevalence = 29.8%), the negative
predictive value is 96% and the positive predictive value is 71%. As these
results are based in part on the exclusion criteria of this study, it is
difficult to extrapolate how accurate these tests would be when applied to a
population with a lower prevalence of infection, but if we lower the prevalence
to, for example, 2%, the negative predictive value increases a little to 99.8%
but the positive predictive value drops sharply from 70% to 10%.
Similarly, changing the prevalence of infection from 30% to
2% slightly increases the negative predictive value of the C-reactive protein test
results from 95% to 99% but markedly decreases the positive predictive value from
73% to 12%.
What these additional calculations emphasize is an
observation made by the authors as well as by previous investigators6:
the erythrocyte sedimentation rate and the C-reactive protein level are both good
screening tests, not confirmatory tests. An elevated erythrocyte sedimentation
rate or C-reactive protein level in a patient who has had knee arthroplasty and
who has knee pain but who does not have an underlying inflammatory or
neoplastic disorder should raise suspicion of infection and should prompt
further investigation, especially aspiration and culture of knee fluid. On the
other hand, both the erythrocyte sedimentation rate test and the C-reactive
protein level test are associated with too many false-positives to be considered confirmatory
tests, especially when used in the usual orthopaedic practice that has a low
prevalence of infections.
Finally, the cutoff points calculated by the authors from
their receiver-operating-characteristic curves may be helpful in optimizing the
accuracy of their tests, but it is useful to remember that these tests give
continuous, not binary results, and that following the changes in erythrocyte
sedimentation rate and C-reactive protein values over time for any given
patient can also provide helpful information.
*The author did not receive any outside funding or grants in
support of his research for or preparation of this work. Neither he nor a
member of his immediate family received payments or other benefits or a
commitment or agreement to provide such benefits from a commercial entity. No
commercial entity paid or directed, or agreed to pay or direct, any benefits to
any research fund, foundation, division, center, clinical practice, or other
charitable or nonprofit organization with which the author, or a member of his
immediate family, is affiliated or associated.
References
1. McPherson RA, Pincus MR, editors. Henry's Clinical Diagnosis and Management by Laboratory Methods. 21st ed. Philadelphia: Saunders; 2006. p 465.
2. Tillett WS, Francis T Jr. Serological reactions in pneumonia with non-protein somatic fraction of pneumococcus. J Exp Med. 1930;52:561-71.
3. Carlsson AS. Erythrocyte sedimentation rate in infected and non-infected total hip arthroplasties. Acta Orthop Scand. 1978;49:287-90.
4. White J, Kelly M, Dunsmuir R. C-reactive protein level after total hip and total knee replacement. J Bone Joint Surg Br. 1998;80:909-11.
5. Spangehl MJ, Masri BA, O'Connell JX, Duncan CP. Prospective analysis of preoperative and intraoperative investigations for the diagnosis of infection at the sites of two hundred and two revision total hip arthroplasties. J Bone Joint Surg Am. 1999;81:672-83.
6. Parvizi J, Ghanem E, Menashe S, Barrack RL, Bauer TW. Periprosthetic infection: what are the diagnostic challenges? J Bone Joint Surg Am. 2006;88:138-47.
|
