вход по аккаунту


Liver Enzyme Levels in Arthritis Patients Treated with Long-Term Bolus Methotrexate.

код для вставкиСкачать
Potential interpretational errors in the use of the
polymerase chain reaction to detect Yersinia in clinical
To the Editor:
I read with interest the article by Viitanen et al, in
which the authors conclude that, in Yersinia-associated
reactive arthritis, “whole bacteria d o not enter the affected
joints’’ (Viitanen A-M, Arstila TP, Lahesmaa R, Granfors
K, Skurnik M, Toivanen P: Application of the polymerase
chain reaction and immunofluorescence techniques to the
detection of bacteria in Yersinia-triggered reactive arthritis.
Arthritis Rheum 34:8%%, 1991). In my opinion, the data
presented cannot support this important conclusion.
In order to interpret a negative result from the
polymerase chain reaction (PCR), as was done in their study,
meticulous attention to controls and standards is essential,
so that the minimum number of organisms that might have
been detected, had they been present, can be determined.
Their report contains several important sources of error in
this regard, any one of which could have resulted in falsenegative results. First, no evidence is presented that would
demonstrate that the clinical specimens of interest, principally, those from synovial fluid, would have yielded detectable Yersinia DNA if the organisms had been present. This is
important for at least two reasons. One is that the mild
conditions used to extract DNA from the specimens might
have failed, at least in part, to extract DNA from whole
organisms present in these specimens. (The only cellbacteria mixtures that were extracted appear to have been of
peripheral blood leukocytes. These may have important
differences from those in synovial fluid specimens, which,
for example, contain large amounts of polysaccharide that
might affect the extraction.) Second, there might well have
been inhibitors of the PCR present which prevented amplification of Yersinia DNA, had any been extracted. One such
inhibitor was in fact present in the specimens, namely,
heparin. The heparin would probably not have been eliminated by the routine DNA purification techniques employed
(Beutler E, Gelbart T, Kuhl W: Interference of heparin with
the polymerase chain reaction. Biotechniques 9: 166, 1990).
Other inhibitors are not as well defined, but are certainly
well recognized to occur in clinical specimens.
An entirely separate source of error derives from the
method used to standardize the PCR by defining the number
of organisms whose extracted DNA was detectable. This
was accomplished by using colony counts from an overnight
culture of bacteria. The PCR detects all bacterial DNA, not
just that contained in viable organisms. Colony counts detect
only the latter. An overnight culture generally contains
organisms grown past log phase, and thus contains an
undetermined number of nonviable organisms and probably
extracellular bacterial DNA as well. The consequence of
using this approach to standardize the PCR would be to
overstate to an unknown extent the sensitivity of the
method. The fact that in these studies, the target was not
genomic DNA, but rather, bacterial plasmid DNA raises
additional uncertainties, some of which were addressed by
the authors.
The PCR is an extraordinarily powerful technique
Arthritis and Rheumatism, Vol. 35, No. 1 (January 1992)
with potential for providing reasonably definitive answers to
questions such as that asked here. As with many such
techniques, however, its conceptual simplicity must not
induce complacency regarding its practical complexity.
Charles R. Steinman, MD
State University of New York
Stony Brook, NY
To the Editor:
Dr. Steinman bases much of his argument on the fact
that peripheral blood cells, but not synovial fluid cells, were
used as controls, and that synovial fluid would possibly have
contained inhibitors of both DNA extraction and the polymerase chain reaction (PCR). First, Ficoll-Isopaque gradient
centrifugation was used to separate mononuclear cells and
granulocytes from both the peripheral blood and the synovial
fluid samples, as stated in our report. The cells were washed
at least once before processing for PCR. This procedure
should remove any accompanying polysaccharides from the
synovial fluid cells.
With regard to the heparin treatment, both the peripheral blood and synovial fluid samples were treated, and
in this respect, cell samples from neither source should be
expected to behave differently in the PCR. Indeed, when
used to construct positive samples for the preliminary sensitivity testing, synovial fluid cells showed no evidence of
inhibitors of DNA extraction or of the PCR. Peripheral blood
cells, however, occasionally had inhibitory effects on the
PCR, and this was probably due to minor erythrocyte
contamination. The dialysis step described in the paper was
found to be necessary in order to remove the inhibitory
substances. This, of course, should have been explained in
the article.
Concerning the possibly overstated sensitivity of our
PCR, Dr. Steinman’s arguments themselves are certainly
justified. They are not, however, cause for changing the
interpretation of the results. Indeed, in the immunofluorescence-positive synovial fluid samples analyzed by the PCR,
the percentage of positively staining cells was -1%. At this
level of positivity, bacterial DNA should easily have been
detected by our PCR.
Mikael Skurnik, PhD
Anna-Man Viitanen, MD
Paavo Toivanen, MD
Turku University
Turku, Finland
Liver enzyme levels in arthritis patients treated with
long-term bolus methotrexate
To the Editor:
: The study by Fries et al(1) regarding levels of serum
glutamic oxaloacetic transaminase (SGOT) and serum glutamic pyruvic transaminase (SGPT) in patients with rheumatoid arthritis (RA) is of special interest to us. It prompted us
to review the levels of SGOT, SGPT, and gamma glutamyl
transpeptidase (GGTP) in 40 patients with arthritis (RA,
Table 1. Liver enzyme levels in arthritis patients treated with bolus
methotrexate (MTX) for more than 10 years, with or without
hydroxychloroquine (HCQ)
that these drug interactions are ultimately not clinically
Julio Aponte, MD, FACP
Mary Petrelli, MB, BS, FRCPath
Neal von Dawson, MD
MetroHealth Medical Center
at Case Western Reserve University
School of Medicine
Cleveland, OH
Mean f SEM
% abnormal
Mean f SEM
% abnormal
Mean 2 SEM
% abnormal
(n = 28)
(n = 12)
27.1 f 1.53
39.8 f 12.7
20.3 f 1.0
23.8 k 4.1
25.4 2 3.4
24.7 f 7.1
* Values are IUM1. SGOT = serum glutamic oxaloacetic transaminase; SGPT = serum glutamic pyruvic transaminase; GGTP =
gamma glutamyl transpeptidase.
1. Fries JF, Singh G, Lenert L, Furst DE: Aspirin, hydroxychloro-
quine, and hepatic enzyme abnormalities with methotrexate in
rheumatoid arthritis. Arthritis Rheum 33: 1611-1619, 1990
2. Aponte J, Petrelli M: Histopathologic findings in the liver of
rheumatoid arthritis patients treated with long-term bolus methotrexate. Arthritis Rheum 31:1457-1464, 1988
3. Petrelli M, Aponte J: Absence of significant hepatotoxicity in
arthritis patients on long-term methotrexate treatment (abstract).
Hepatology 10:746, 1989
systemic lupus erythematosus, and Reiter’s syndrome) who
had been receiving bolus methotrexate (MTX) either with or
without hydroxychloroquine (HCQ) for more than 10 years.
Liver pathologic findings in most of these patients have been
reported previously (2,3). Only the enzyme levels at the time
of biopsy or the most recently measured enzyme levels from
patients who did not undergo biopsy were evaluated. Most
of the patients were receiving concomitant nonsteroidal
antiinflammatory drugs. Statistical comparisons were performed using the t-distribution and Fisher’s exact test.
The mean f SEM liver enzyme levels in the patients
are shown in Table 1. In contrast to the findings of Fries et
al, we found no statistically significant difference in enzyme
levels between the 2 groups (MTX with HCQ versus MTX
without HCQ) (the lack of a significant difference was
perhaps due to small sample size). However, qualitatively,
the direction of difference was the same as that reported by
Fries ztnd colleagues, i.e., lower SGOT and SGPT levels in
the patients who received HCQ. The magnitude of the
difference in mean enzyme levels between groups in our
study was similar to that reported by Fries et a1 for SGOT
(difference of 12.7 IU/dl in our study, versus 11.1 IU/dl in
theirs), but was smaller for SGPT (3.5 IU/dl in our study,
versus 13.1 IU/dl in theirs). In addition, in our 32 patients
who had liver biopsies, there was no difference in histologic
grading between the HCQ-treated and the non-HCQ-treated
patients. Neither liver enzyme levels nor treatment with
HCQ correlated with the degree of anisonucleosis, fatty
change, necrosis, inflammation, or fibrosis.
The mode of administrationof MTX is not mentioned
in the article by Fries et al. We believe that the mode of
MTX administration can affect its hepatotoxicity. MTX
given in oral bolus form, with all tablets taken at once, is
associated with less hepatotoxicity. The clinical significance
of the variability of enzyme levels is unclear since SGOT,
SGPT, and to a lesser degree, GGTP are produced by other
organs as well as the liver. We concur with Fries et a1 that
“elevations in transaminase levels with MTX therapy are
not consistently related to subsequent hepatic fibrosis, and
To the Editor:
We are pleased that Dr. Aponte and colleagues were
able to confirm our finding of the “protective” effect of
hydroxychloroquine (HCQ) upon liver enzyme elevation in
methotrexate (MTXttreated patients (1). As they point out,
the statistical power of their study was low due to small
sample size, but the direction and magnitude of the findings
are the same as those that we reported.
However, the full quotation from our article, in
contrast to that noted in their letter, was “Zt might be argued
that elevations. . . .” We went on to indicate that we believe
the association between liver enzyme elevation and subsequent fibrosis is likely to be real, and that there may be
important clinical benefits from reduction of “transaminitis,” in terms of both prevention of fibrosis and reassurance
of the womed patient.
Unfortunately, neither our data nor theirs permit a
rigorous approach to the critical question, i.e., whether
improvement in enzyme levels can be extrapolated to improvement in hepatic histology. The best available data are
those of Kremer et al (2), which suggest that repeatedly
abnormal findings on liver function tests along the course of
MTX treatment are associated with greater degrees of hepatic fibrosis. Cross-sectional studies such as those reported
by Aponte et al, using only the serum glutamic oxaloacetic
transaminase and serum glutamic pyruvic transaminase levels at the time of biopsy, not unexpectedly, fail to show any
association. In order to rigorously determine the relationship
between enzyme abnormalities and subsequent fibrosis, and
then the role of HCQ (if any) in reducing fibrosis, a large
prospective longitudinal study, using duration variables such
as “years taking MTX” and “years taking HCQ,” and with
multiple hepatic enzyme determinations throughout the
course, will be required. Unfortunately, the data sets which,
to our knowledge, have the best information of this kind are
from studies whose protocols excluded HCQ treatment
We believe, on balance, it is more likely than not that
prevention of repeated hepatic injury (as measured by en-
Без категории
Размер файла
238 Кб
terms, level, patients, enzymes, long, arthritis, methotrexate, live, treated, bolus
Пожаловаться на содержимое документа