вход по аккаунту


No evidence of an association between the interleukin-10 promoter GCC allele and primary Sjgren's syndromeComment on the article by Hulkkonen et al.

код для вставкиСкачать
Vol. 46, No. 6, June 2002, pp 1691–1698
© 2002, American College of Rheumatology
By electron microscopy, podocytes were found to be hypertrophic and showed retraction of their foot processes. In some
segments of the GBM, irregularly spaced subepithelial
electron-dense deposits accompanied by spike formation were
observed. Other segments displayed a broad and abnormal
GBM, incorporating many complexes that were less electron
dense, and resolving.
The patient was treated with diuretics and an
angiotensin-converting enzyme inhibitor, and anti-TNF␣ and
diclofenac treatment were withdrawn. The subsequent course
is depicted in Figure 1. The proteinuria remained difficult to
control for 3 months after cessation of the diclofenac and
anti-TNF␣ treatment, but eventually resolved. Interestingly,
symptoms of RA, which had been mild before and during the
phase that the nephrotic syndrome was active, relapsed in the
same week the nephrotic syndrome disappeared.
We then attempted to determine if drug treatment was
the cause of the nephrotic syndrome. Diclofenac was restarted;
however, proteinuria did not reappear until after anti-TNF␣
treatment was also reinstituted. The patient was subsequently
treated with prednisone (⬃60 mg daily) and cyclophosphamide
(100 mg daily). Within 1 week, the proteinuria had completely
disappeared. Cyclophosphamide was discontinued, and the
prednisone rapidly tapered to 10 mg daily. At this time point,
anti-TNF␣ was again reintroduced. The nephrotic syndrome
did not recur until the prednisone dosage was reduced to 5 mg
daily. Again, the patient responded to a higher dosage of
prednisone. Since December 1998, the patient has been
treated with anti-TNF␣ in combination with 10 mg prednisone
daily. He has continued to have an excellent clinical response,
and proteinuria has remained absent.
In our patient, a nephrotic syndrome developed during
treatment with adalimumab, a monoclonal antibody against
TNF␣ (3). Renal biopsy revealed a membranous glomerulopathy. The clinical course strongly incriminated anti-TNF␣,
since proteinuria disappeared after withdrawal of anti-TNF␣
treatment and reappeared upon rechallenge. Furthermore, we
were able to exclude the possibility that the other likely
candidate, the nonsteroidal antiinflammatory drug diclofenac,
was the cause of the nephrotic syndrome. To our knowledge,
this is the first report of a nephrotic syndrome as an adverse
effect of anti-TNF␣ treatment.
This may represent a case of membranous nephropathy caused by anti-TNF␣ treatment. We believe, however, that
the clinical data indicate an alternative explanation for the
nephrotic syndrome. The clinical picture in our patient mimicked the nephrosis that is associated with minimal-change
nephropathy (4). Proteinuria was highly selective, and there
was an immediate and brisk response to treatment with
steroids. In patients with established membranous nephropathy proteinuria disappears quite slowly if at all, even during
treatment with immunosuppressive agents (5). Among 34
patients with membranous nephropathy who were treated with
high-dose prednisone, we did not observe complete remission
of proteinuria within 3 months after the start of treatment in
any (6). Likewise, in patients with RA with membranous
nephropathy caused by either gold or D-penicillamine, treat-
DOI 10.1002/art.10344
Nephrotic syndrome associated with anti–tumor
necrosis factor ␣ therapy in a patient with rheumatoid
arthritis: comment on the article by Charles et al
To the Editor:
We read with interest the article by Charles et al
concerning the possible relationship between tumor necrosis
factor (TNF) blockade and the induction of anti–doublestranded DNA (anti-dsDNA) and systemic lupus erythematosus (SLE)–like syndromes (1). Herein we describe a patient
who developed a nephrotic syndrome as an adverse effect of
treatment with anti-TNF␣, without the presence of antidsDNA or signs of SLE.
The patient was diagnosed in 1982 as having rheumatoid factor–positive, antinuclear antibody (ANA)–negative
erosive rheumatoid arthritis (RA). Shortly after diagnosis he
was treated with D-penicillamine and then with parenteral
aurothioglucose. This was stopped in 1984 because of proteinuria (to 1.5 gm/liter). After cessation of this therapy, the
proteinuria completely disappeared. Disease-modifying antirheumatic drug (DMARD) therapy was continued, this time
with methotrexate, followed again by aurothioglucose. Proteinuria did not develop during this second episode of gold
therapy, nor did it occur during the next several years. Thereafter he was treated with a succession of DMARDs, without
clinical efficacy.
In May 1997 the patient was recruited for a placebocontrolled phase I trial of a fully human monoclonal antibody
against TNF␣ (adalimumab; Knoll/BASF, Ludwigshafen, Germany). Treatment with anti-TNF␣ was started at a dosage of
1.0 mg/kg biweekly, which was later increased to 3.0 mg/kg
biweekly. Concomitant medication consisted of prednisone (10
mg once daily), diclofenac, alternating calcium and etidronic
acid, omeprazole, and acetaminophen as needed. Clinical
response was excellent: the Disease Activity Score (a combined
disease activity measure for RA [2]) decreased from 5.89 to
2.41, and the C-reactive protein level decreased from 143
mg/liter to 24 mg/liter. The treatment was well tolerated.
Prednisone was gradually tapered to 2.5 mg/day.
In April 1998, the patient developed rapidly progressive pitting edema of both legs, and weight gain. Physical
examination revealed bilateral leg edema and evidence of
pleural and peritoneal fluid. Laboratory investigations revealed the following: serum creatinine 82 ␮moles/liter, serum
albumin 15 gm/liter, ANA negative, and normal levels of
complement components C3 and C4. Proteinuria was present
at an average of 16.7 gm/24 hours, with a selectivity index of
0.07 (highly selective).
A renal biopsy was performed. On light microscopy of
silver-stained sections, all glomeruli demonstrated segmental
irregularities of the glomerular basement membrane (GBM),
suggestive of a membranous glomerulopathy. Congo red staining did not reveal amyloid. With immunofluorescence, fine
granular deposits, particularly of IgG and C3, were seen along
the capillary wall in a characteristic membranous pattern. No
AA amyloid deposits were found using a monoclonal antibody.
Figure 1. Clinical course of the patient (proteinuria over time). Arrows indicate injections of anti–tumor necrosis factor ␣ (anti-TNF␣). A clear
relationship between the level of proteinuria, administration of anti-TNF␣, and reduction of the prednisone dosage is seen.
ment with prednisone apparently does not influence the duration of proteinuria (7).
The finding of subepithelial deposits certainly does not
prove that proteinuria is caused by a membranous nephropathy. It is well established that in patients with membranous
nephropathy the deposits can remain visible after remission
has occurred, even for many years (8). In our patient, many of
the deposits seemed to be old, incorporated in the GBM and
dissolving, suggesting that they may have been caused by the
previous gold or diclofenac therapy.
In conclusion, as shown in our patient, TNF blockade–
related renal adverse effects can occur in the absence of
anti-dsDNA or SLE. Careful monitoring of patients receiving
TNF-blocking agents is therefore warranted. Also, the nephrotic syndrome in our patient was highly steroid sensitive,
but also steroid dependent. Prednisone at a daily dosage of 10
mg was able to prevent the recurrence of proteinuria. Since a
considerable number of patients who receive anti-TNF␣ are
concurrently being treated with prednisone, the potential of
anti-TNF␣ to induce proteinuria may be masked. It remains to
be established whether this side effect will appear more
frequently if anti-TNF␣ treatment is instituted in patients who
do not receive steroids.
Alfons A. den Broeder, MD
Karel J. M. Assmann, MD, PhD
Piet L. C. M. van Riel, MD, PhD
Jack F. M. Wetzels, MD, PhD
University Medical Center Nijmegen
Nijmegen, The Netherlands
1. Charles PJ, Smeenk RJT, de Jong J, Feldmann M, Maini RN.
Assessment of antibodies to double-stranded DNA induced in
rheumatoid arthritis patients following treatment with infliximab, a
monoclonal antibody to tumor necrosis factor ␣: findings in openlabel and randomized placebo-controlled trials. Arthritis Rheum
2. Van der Heijde DMFM, van’t Hof MA, van Riel PLCM, Theunisse
LM, Lubberts EW, van Leeuwen MA, et al. Judging disease activity
in clinical practice in rheumatoid arthritis: first step in the development of a disease activity score. Ann Rheum Dis 1990;49:916–20.
3. Van de Putte LBA, van Riel PLCM, den Broeder A, Sander O, Rau
R, Binder C, et al. A single dose placebo controlled phase I study of
the fully human anti-TNF antibody D2E7 in patients with rheumatoid arthritis [abstract]. Arthritis Rheum 1998;41 Suppl 9:S57.
4. Glassock RJ, Cohen AH, Adler SG. In: Brenner BM, editor.
Primary glomerular diseases: Brenner and Rector’s the kidney. 5th
ed. Philadelphia: WB Saunders; 1996. p. 1392–7.
5. Ponticelli C, Zuchelli P, Imbasciati E, Cagnoli L, Pozzi C, Passerini
P, et al. Controlled trial of methylprednisolone and chlorambucil in
idiopathic membranous nephropathy. N Engl J Med 1984;310:
6. Wetzels JFM, Hoitsma AJ, Koene RAP. Analysis of short-term
alternate-day prednisone therapy in patients with membranous
glomerulonephritis [thesis]. Nijmegen (The Netherlands): University of Nijmegen; 1989.
7. Hall CL, Fothergill NJ, Blackwell MM, Harrison PR, Mackenzie
JC, MacIver AG, et al. The natural course of gold nephropathy:
long term study of 21 patients. BMJ 1987;295:745–8.
8. Tornroth T, Honkanen E, Petterson E. The evolution of membranous glomerulonephritis reconsidered: new insights from a study on
relapsing disease. Clin Nephrol 1987;28:107–17.
DOI 10.1002/art.10350
To the Editor:
Dr. den Broeder and colleagues make a strong case for
the association between the onset of nephrotic syndrome and
administration of adalimumab, a fully human anti-TNF␣–
specific antibody. A previous, similar episode of nephropathy
during treatment with D-penicillamine and gold (both well
described as causes of membranous nephropathy) in their
patient is intriguing. It raises the possibility of a genetic
predisposition to this renal disorder. However, to the best of
our knowledge the mechanism of drug-induced nephrotic
syndrome, associated with membranous nephropathy, is not
understood. In SLE, membranous nephropathy is not normally
associated with anti-dsDNA antibodies (Venables PJ, Yi T,
Woodrow DF, Moss J, Maini RN. Relationship of precipitating antibodies to soluble cellular antigens and histologically
defined renal lesions in systemic lupus erythematosus. Ann
Rheum Dis 1983;42:17–22), and lack of it in this patient is,
therefore, not unexpected.
Nevertheless, it would be interesting to establish
whether anti-dsDNA antibodies (as an example of an autoimmune reaction) are induced by adalimumab, as they are by
infliximab, in RA and Crohn’s disease. Although infliximab
and adalimumab are both anti-TNF␣–specific monoclonal
antibodies, there are likely to be immunochemical and immunologic differences, and these may ultimately translate into
differences in their safety profiles. If anti-dsDNA antibodies
are induced by adalimumab, the evidence for a class effect in
the induction of autoimmunity, at least, would be strengthened. It would also be of interest to obtain data on the number
of patients who have been exposed to adalimumab, so the
frequency of nephropathy as an adverse event can begin to be
ascertained. For example, according to the Centocor database
as of September 2001, 4 cases of a lupus syndrome in 1,897
patients exposed (0.2%), none with renal or neurologic involvement, were reported in clinical trials of infliximab. It is,
therefore, difficult to conclude that routine monitoring for
nephropathy would be indicated in infliximab-treated patients.
However, continued careful documentation and reporting of
adverse events by physicians prescribing biologic targeted
therapies is an important ongoing contribution to the assessment of the risk-benefit of these effective new agents.
R. N. Maini, MBBChir, FRCP, FMedSci
P. Charles, FIMLS
Kennedy Institute of Rheumatology Division
Imperial College of Science, Technology and Medicine
London, UK
DOI 10.1002/art.10281
Sample sizes estimated in clinical trials using either a
composite index (ASAS response criteria) or single
outcome variables in ankylosing spondylitis
To the Editor:
Several domains (e.g., pain, functional impairment,
inflammation) should be evaluated together in order to determine the symptomatic effect of treatments in ankylosing
spondylitis (AS). Most tools evaluating these domains measure
continuous variables (e.g., 100-mm visual analog scale [VAS]
for pain). The results of trials using such tools are usually
reported as mean improvement experienced by the patients.
To facilitate interpretation, the advantages of reporting results
on an individual basis (percentage of responders) have been
emphasized. For this purpose, the continuous variable (e.g.,
change in the pain VAS) must be converted to a dichotomous
variable (e.g., improvement in pain: yes/no).
The smallest detectable difference (SDD) is one of the
techniques defining a cutoff that permits changing a continuous variable into a dichotomous variable. Such a cutoff (defined using the SDD technique) was recently reported for
several outcome variables in AS (1). The Assessments in
Ankylosing Spondylitis (ASAS) Working Group also recently
described the use of a composite index comprising a set of
criteria that takes into account several domains and that allows
the results to be presented on an individual basis (2). Our
objective in the present study was to calculate the estimated
sample size needed in future clinical trials by using either the
ASAS response criteria or each individual outcome variable,
considering the percentage of patients with improvement
greater than the SDD. For this purpose, and as the first step,
we calculated the percentage of responders in previously
reported trials (3–7).
Thereafter, we devised a trial evaluating the effect of a
nonsteroidal antiinflammatory drug (NSAID), using a randomized, placebo-controlled, parallel design. A 2-sided test
was used, with ␣ ⫽ 5% and 1 ⫺ ␤ ⫽ 80%. The minimal
relevant difference expected was the difference in percentages
of responders between placebo-treated and NSAID-treated
patients (Table 1), calculated from previous reported AS trials
(3–7). The outcome variables used are shown in Table 1.
The smallest sample size needed per group (n ⫽ 58)
was observed for the ASAS response criteria (Table 1).
Required sample sizes using single outcome variables, except
VAS spinal pain, were 1.5–3-fold that estimated using the
ASAS response criteria.
The ASAS response criteria represent a multiple
domain–based measure that includes patient overall assess-
Table 1. Estimated sample sizes per group, using the Assessments in
Ankylosing Spondylitis (ASAS) response criteria and smallest detectable differences (SDDs) in outcome variables in AS*
% responders observed
Variable (ref.)
ASAS response criteria (2)
VAS spinal pain, 0–100
VAS spinal pain at night,
Spinal pain, FDA guidelines,
0–16 (10)
Duration of morning stiffness, 54.7
BASFI, 0–100 (8)
Dougados functional index,
0–36 (11)
VAS disease activity,
BASDAI, 0–100 (9)
sample size
* Responders were defined as patients in either the placebo group or
the nonsteroidal antiinflammatory drug (NSAID) group who fulfilled
the ASAS criteria or had improvement greater than the SDD in
previously reported trials (see refs. 3–7). NA ⫽ not applicable; VAS ⫽
visual analog scale; FDA ⫽ US Food and Drug Administration;
BASFI ⫽ Bath Ankylosing Spondylitis Functional Index; BASDAI ⫽
Bath Ankylosing Spondylitis Disease Activity Index.
† These results have been previously reported (see ref. 1).
Pellegrini-Guegnault B, et al. Ximoprofen in ankylosing spondylitis: a double blind placebo controlled dose ranging study. Scand
J Rheumatol 1994;23:243–8.
Dougados M, Gueguen A, Nakache J-P, Velicitat P, Veys EM,
Zeidler H, et al. Ankylosing spondylitis: what is the optimum duration
of a clinical study? A one year versus a 6 weeks non-steroidal
anti-inflammatory drug trial. J Rheumatol 1999;38:235–44.
Dougados M, Béhier J-M, Jolchine I, Calin A, van der Heijde D,
Olivieri I, et al. Efficacy of celecoxib, a cyclooxygenase 2-specific
inhibitor, in the treatment of ankylosing spondylitis: a six-week
controlled study with comparison against placebo and against a
conventional nonsteroidal antiinflammatory drug. Arthritis
Rheum 2001;44:180–5.
Calin A, Garrett S, Whitelock H, Kennedy LG, O’Hea J, Mallorie
P, et al. A new approach to defining functional ability in ankylosing spondylitis: the development of the Bath Ankylosing Spondylitis Functional Index. J Rheumatol 1994;21:2281–5.
Garrett S, Jenkinson T, Kennedy LG, Whitelock H, Gaisford P,
Calin A. A new approach to defining disease status in ankylosing
spondylitis: the Bath Ankylosing Spondylitis Disease Activity
Index. J Rheumatol 1994;21:2286–91.
Guidelines for the clinical evaluation of anti-inflammatory and
anti-rheumatic drugs (adults and children). Washington, DC: US
Department of Health and Human Services, Public Health Service, Food and Drug Administration; 1988:38–40.
Dougados M, Gueguen A, Nakache JP, Nguyen M, Amor B.
Evaluation of a functional index for patients with ankylosing
spondylitis. J Rheumatol 1990;17:1254–5.
DOI 10.1002/art.10238
ment, pain, Bath Ankylosing Spondylitis Functional Index, (8)
and either Bath Ankylosing Spondylitis Disease Activity Index
(9) or duration of morning stiffness (2). It is valid, easy to use,
and can avoid contradictory results that may occur in trials that
use several single outcome variables. This study suggests that it
also allows reduction in the required sample size in AS clinical
trials compared with those using single outcome variables.
Guy-Robert Auleley, MD, PhD
Maxime Dougados, MD
René Descartes University
and Hôpital Cochin, Assistance
Publique Hôpitaux de Paris
Paris, France
Gabrial Baron
Hôpital Cochin
Paris, France
1. Benbouazza K, Auleley GR, Collantes E, Hajjaj-Hassouni N, van
der Heijde D, Dougados M. Evaluation of the smallest detectable
difference (SDD) in symptomatic outcome variables in ankylosing
spondylitis [abstract]. Ann Rheum Dis 2000;59 Suppl 1:56.
2. Anderson JJ, Baron G, van der Heijde D, Felson DT, Dougados
M. Ankylosing Spondylitis Assessment Group preliminary definition of short-term improvement in ankylosing spondylitis. Arthritis
Rheum 2001;44:1876–86.
3. Dougados M, Gueguen A, Nakache JP, Nguyen M, Mery C, Amor
B. Evaluation of a functional index and an articular index in
ankylosing spondylitis. J Rheumatol 1988;15:302–7.
4. Dougados M, Caporal R, Doury R, Thiese A, Pattin S, Laffez B,
et al. A double blind crossover placebo controlled trial of ximoprofen in AS. J Rheumatol 1989;16:1767–9.
5. Dougados M, Nguyen M, Caporal R, Legeais J, Bouxin-Sauzet A,
No evidence of an association between the
interleukin-10 promoter GCC allele and primary
Sjögren’s syndrome: comment on the article by
Hulkkonen et al
To the Editor:
Primary Sjögren’s syndrome (SS) is a systemic autoimmune disease predominantly affecting women, in which the
primary targets of autoimmune destruction are the salivary
and lacrimal glands. Autoantibodies against ribonucleoproteins (SSA/Ro and SSB/La) are common in this disease,
although their role in disease pathogenesis is unclear. Interest
in the frequency of the 3 interleukin-10 (IL-10) promoter
alleles (defined by base substitutions at positions ⫺1082, ⫺819,
and ⫺592) in SS patients increased following a report that the
GCC allele was associated with a small subset of systemic lupus
erythematosus (SLE) patients with anti-Ro autoantibodies (1).
Recently, 2 conflicting studies on IL-10 promoter
polymorphisms in SS patients have been published. In an
article in Arthritis & Rheumatism, Hulkkonen et al reported
that the frequency of the IL-10 promoter GCC allele was
increased in a cohort of Finnish SS patients compared with
controls (2). In contrast, our earlier study demonstrated no
differences in the frequency of the IL-10 promoter allele
between Australian SS patients and controls (3). Furthermore,
in SS patients, no associations between these alleles and the
presence of anti-Ro and/or anti-La autoantibodies were observed. Therefore, the role of IL-10 promoter polymorphisms
in SS remains unclear.
Both of the above-mentioned studies (2,3) had similar
power to detect an increase in frequency of the GCC allele;
Table 1. Results of individual and combined study log-likelihood ratio (G2) tests for analysis of interleukin-10 (IL-10) promoter polymorphisms
in Sjögren’s syndrome (SS) patients compared with controls
Degrees of
Allele effect
Genotype effect
Total IL-10 effect
Study ⫻ allele interaction
Study ⫻ genotype interaction
Total study ⫻ IL-10 interaction
Hulkkonen et al (2)
Limaye et al (3)
therefore, any difference in outcome might be attributable to
genetic and/or environmental differences between the Australian and Finnish populations or to differences in the diagnostic
criteria used. The IL-10 genotype and allele frequencies in the
2 studies are remarkably similar, however, suggesting that
chance fluctuations and perhaps the type of statistical analysis
used may have played a role in the differing conclusions.
Analysis of this type of data can be problematic,
because differences in both allele frequencies and genotype
frequencies between the patient and control groups are of
interest, and analysis complexity increases with increasing
numbers of alleles. It is common practice to construct multiple
2 ⫻ 2 tables to compare each allele and each genotype
frequency in the patient and control groups by chi-square
analysis. However, this approach may lead to incorrect inferences for several reasons. First, each comparison will have
different power to detect a significant effect, depending on the
frequency of the allele or genotype being tested. Second,
multiple comparisons result in an increased risk of detecting a
significant result that may be attributable to chance alone.
Although the Bonferroni correction for multiple comparisons
is sometimes used, this approach still ignores the correlation
within and between allele and genotype frequencies. That
correlation arises because allele and genotype frequencies
must sum to 1, and allele frequencies are calculated by direct
counting from genotype frequencies. Lack of consideration of
such a correlation results in confounded and redundant comparisons that may overestimate the significance of any apparent effects. Also, it is impossible to distinguish between “allele
effects,” in which the presence of a particular allele in any
genotype may confer susceptibility to disease, and “genotype
effects,” in which only a subset of genotypes carrying this allele
may be associated with disease.
IL-10 promoter genotype frequencies for SS patients
and controls form a 6 ⫻ 2 contingency table with 5 degrees of
freedom (df). The best overall test of the null hypothesis of
equivalence in the IL-10 promoter polymorphism distribution
between SS and control groups is the chi-square test of this
contingency table. The deviance statistic G2 (log-likelihood
ratio) from an “intercept only” logistic regression model, in
which the dichotomous outcome is either SS or controls, is
equivalent. The G2 and chi-square statistics are approximately
equal in value and are interpreted in the same manner. The
population differentiation test used by our group (3) is an exact
version of this test and is preferable when sample numbers are
The G2 statistic with 5 df is a composite statistic with
additive components comprising allele-specific (with 2 df) and
genotype-specific (with 3 df) effects. Partitioning can be
achieved by constructing log-likelihood ratio tests from a series
of appropriately formulated nested logistic regression models
(4). Analysis can be readily extended to estimate a “common
effect” or combined analysis from multiple studies by including
additional study stratifying variables.
Such an analysis (see Table 1) was performed by fitting
3 nested logistic regression models (model 1, no effects
[intercept only]; model 2, allele effects only; model 3, both
allele and genotype effects). The deviance statistic G2 from
model 1 is a test of no overall association between IL-10
promoter polymorphisms for SS patients versus controls. The
successive reduction in the G2 statistic between each of these 3
models is the log-likelihood ratio test for the significance of the
allele and genotype components, respectively.
The formulation of the models was done according to
the protocol described by Weir (4). Briefly, 2 regression
coefficients for allele effects were estimated using dummy
variables coded 0, 1 (heterozygotes), or 2 (homozygotes) for 2
of the alleles. Three regression coefficients for genotype
effects were estimated using dummy variables coded 0 or 1 for
each heterozygote genotype. The remaining coefficients were
set to 0 for identifiability. In this format, the allele effects test
is equivalent to the analysis of the 3 ⫻ 2 contingency table of
allele frequencies. The genotype effects test estimates any
difference in departure from Hardy-Weinberg equilibrium
between the 2 groups.
The analysis of a “common effect” over both studies
was performed with the inclusion of a study stratifying variable.
This analysis had an additional 5 df, corresponding to interaction terms between the study variable and IL-10 promoter
polymorphisms, and provided a test of “between study” heterogeneity.
The data from the study by Hulkkonen et al are
suggestive of an association between IL-10 promoter polymorphisms and SS (G2 ⫽ 10.85, 3 df, P ⫽ 0.055), but the
association is not statistically significant. The separate allelespecific and genotype-specific effects also fail to reach statistical significance, although, again, the results of the allele
effects test are suggestive of a difference in allele frequencies
between SS patients and controls (G2 ⫽ 5.13, 2 df, P ⫽ 0.08).
No evidence of a significant association between IL-10 promoter polymorphisms and SS was observed in either our study
or in the combined, stratified analysis of both studies (G2 ⫽
7.90, 5 df, P ⫽ 0.16). Importantly, there is no evidence of study
heterogeneity, as assessed by the study interaction terms (G2 ⫽
6.71, 5 df, P ⫽ 0.24).
The standardized odds (i.e., the ratio of SS-to-control
group frequencies), obtained by alternative parameterization
of the regression coefficients to zero-sum constraints, are a
useful way to interpret the relative contribution of each allele
or genotype to any differences between the 2 groups. The
standardized odds (95% confidence interval) obtained from
the “allele only” regression model for the combined data were
1.11 (0.94–1.32), 0.84 (0.70–1.02), and 1.07 (0.87–1.31) for the
GCC, ACC, and ATA alleles, respectively. There was a
nonsignificant trend toward a decrease in the ACC allele
frequency in SS patients compared with controls, with a
concomitant and equally proportionate increase in both the
GCC and ATA allele frequencies. There was no evidence of a
selective increase in the GCC allele frequency in SS patients.
We conclude that there is currently no evidence for an
association between IL-10 promoter polymorphisms and SS.
However, this does not preclude the possibility of associations
between IL-10 promoter polymorphisms and clinical subgroups within the population of SS patients. We were unable to
show any association with Ro/La autoantibody status in our
study, although the power of this analysis was limited. Furthermore, the relationship between IL-10 promoter polymorphism
and autoantibodies is potentially more complex. Estrogen
enhances IL-10 production in in vitro cultures of peripheral
blood mononuclear cells from patients with SLE, resulting in
an increase in total IgG and anti–double-stranded DNA
antibodies (5). These effects are less pronounced in patients
with inactive disease or healthy controls. Hulkkonen et al
demonstrated that the IL-10 promoter GCC allele was associated with elevated IL-10 levels in the sera of SS patients but
not in healthy controls; however, the sex of these patients was
not reported.
It is plausible that IL-10 levels in patients with autoimmune disease may be influenced by both estrogen status and
IL-10 promoter polymorphisms, which in turn may influence
autoantibody production. In support of this notion, we observed an apparent association between the ACC allele (associated with low serum IL-10 levels in the study by Hulkkonen
et al) and a limited anti-Ro/La response (defined as an anti-La
antibody detected only by enzyme-linked immunosorbent assay and undetectable by immunodiffusion) in female SS patients only (3). We believe that judicious interpretation and
reporting of candidate genetic association studies underpins
their utility in the study of complex diseases.
Sue Lester, BSc (Hons)
Vidya Limaye, FRACP
Sarah Downie-Doyle, PhD
Kevin D. Pile, MD, FRACP
The Queen Elizabeth Hospital
Woodville, Australia
Peter Bardy, FRACP, FRCPA
Red Cross Blood Transfusion Service
Tom P. Gordon, MD, FRACP, FRCPA
Flinders Medical Centre
Bedford Park, Australia
Maureen Rischmueller, FRACP
The Queen Elizabeth Hospital
Woodville, Australia
1. Lazarus M, Hajeer AH, Turner D, Sinnott P, Worthington J, Ollier
WE, et al. Genetic variation in the interleukin 10 gene promoter
and systemic lupus erythematosus. J Rheumatol 1997;24:2314–7.
2. Hulkkonen J, Pertovaara M, Antonen J, Lahdenpohja N, Pasternack A, Hurme M. Genetic association between interleukin-10
promoter region polymorphisms and primary Sjögren’s syndrome.
Arthritis Rheum 2001;44:176–9.
3. Rischmueller M, Limaye V, Lester S, Downie-Doyle S, Pile K,
Bardy P, et al. Polymorphisms of the interleukin 10 gene promoter
are not associated with anti-Ro autoantibodies in primary Sjögren’s
syndrome. J Rheumatol 2000;27:2945–6. Correction in: J Rheumatol 2001;28:684.
4. Weir B. Genetic data analysis II. Sunderland MA: Sinauer Associates; 1996.
5. Kanda N, Tsuchida T, Tamaki K. Estrogen enhancement of
anti–double-stranded DNA antibody and immunoglobulin G production in peripheral blood mononuclear cells from patients with
systemic lupus erythematosus. Arthritis Rheum 1999;42:328–37.
DOI 10.1002/art.10205
To the Editor:
Dr. Lester and colleagues have retrospectively analyzed our data on IL-10 gene polymorphism in primary SS and
also performed combined logistic regression analysis using our
data and their own (1). In conclusion, they suggest that IL-10
genotypes are not associated with susceptibility to primary SS,
which is contradictory to our findings.
We appreciate the interest of Lester et al in our results.
However, their analytical approach raises some major points of
concern. Some differences between the 2 studies must be taken
into account when evaluating the results and, particularly,
before performing any kind of combined analyses.
The first and most important difference between these
2 studies is patient selection. Although the diagnostic criteria
for primary SS in the study by Rischmueller et al (1) are not
presented in the text, it can be concluded from the references
that they used the European criteria. In our study, however,
patients were originally selected using the modified Californian criteria; additionally, all patients fulfilled the European
criteria. Because only a minority (10–15%) of primary SS
patients diagnosed by the European criteria fulfill the California criteria (for review, see ref. 2), the patient groups in these
studies are not comparable. The patient group selected by the
European criteria is more heterogeneous and genetically
closer to the normal population than is a cohort selected
according to the more strict Californian criteria. This differTable 1. Observed frequencies of individual interleukin-10 (IL-10)
genotypes in healthy controls and patients with primary Sjögren’s
(n ⫽ 62)
(n ⫽ 400)
15 (⫺0.9)
16 (1.0)
18 (2.2)
5 (⫺0.7)
7 (⫺0.8)
1 (⫺1.3)
126 (0.4)
77 (⫺0.4)
63 (⫺0.9)
46 (0.3)
63 (0.3)
25 (0.3)
* Values are the number (standardized residual).
ence may be the fundamental reason for the different result
regarding IL-10 allele distribution between our study and that
of Rischmueller et al.
In addition to the differences in the diagnostic criteria
used, there are other differences between the 2 study populations. The Finnish and the Australian populations are genetically entirely different. Finland is a genetic isolate and thus an
ideal ground for studies concerning multifactorial and polygenic diseases (3). This benefit results from an increased
degree of linkage disequilibrium (LD) between single nucleotide polymorphisms (SNPs) in founder populations compared
with outbred populations (4). Founder populations also provide one valuable advantage when multiple comparisons are
carried out. Finns have been shown to have exceptionally low
genetic diversity (3), which is demonstrable also as a reduced
number of SNPs (4). In a genetically homogeneous founder
population, fewer potential alleles and haplotypes might correlate with the disease phenotype. Thus, the reduced genetic
heterogeneity and the increased LD in founder populations
can significantly increase the genotypic relative risk and increase statistical power for identification of genes (4). Based
on these facts, the sample sizes needed to identify genetic
associations in founder populations are smaller than those
needed in outbred populations.
When the above-mentioned differences in patient selection and population genetics in these 2 studies are taken
into account, it is obvious that the study populations differ.
Therefore, we cannot accept the conclusions drawn from the
combined statistics presented by Lester et al in Table 1 of their
However, there are also some differences between the
statistical analyses of these 2 studies. We agree with Lester et
al that chi-square statistics may be problematic when calculations are carried out with small numbers of patients, but the
logistic regression analysis provides no essential benefit in this
respect. Still, as Lester and colleagues mention, the G2 and
chi-square statistics are approximately equal in value and are
interpreted in the same manner. Based on this fact, we
reanalyzed the data from our original study by carrying out
chi-square and log-likelihood ratio tests for a 6 ⫻ 2 table and
then calculated standardized residuals.
Using standardized residuals, which have an approximate standard normal distribution, we identified the individual
cells that are not homogeneous according to observed distributions. The results from these calculations are summarized in
Table 1, which also shows that in the GCC/ATA cell, the
standardized residual is 2.2, indicating significant difference
according to the hypothesis of homogeneity of patients and
controls. This result suggests that among patients, observations
in this cell are overrepresented; i.e., the GCC/ATA genotype is
significantly more common among patients than among controls. The P value for the chi-square and log-likehood ratio
tests is 0.055, which may be regarded as a borderline result.
Furthermore, the exact 2-sided P value for standardized residual 2.2 is 0.029. In our opinion, these figures are in accordance
with our original observations.
It is a basic principle of statistics in medicine that the
observation of potential new pathogenetic factors should be
confirmed in separate independent studies before establishing
the association. The results reported by Rischmueller et al. (1)
are contradictory to ours, but based on the points presented
above, the analyses described in their letter do not essentially
change our results. More studies in this field are needed to
achieve confirmatory data from other study populations. We
hope that further studies concerning cytokine genes will bring
new perspectives to this discussion. Unfortunately, as demonstrated here, the lack of uniformly accepted criteria for primary SS is still causing problems in research involving this
Janne Hulkkonen, MD
Marja Pertovaara, MD, PhD
Pekka Laippala, PhD
Jaakko Antonen, MD, PhD
Amos Pasternack, MD, PhD
Mikko Hurme, MD, PhD
University of Tampere, Medical School
and Tampere University Hospital
Tampere, Finland
1. Rischmueller M, Limaye V, Lester S, Downie-Doyle S, Pile K,
Bardy P, et al. Polymorphisms of the interleukin 10 gene promoter
are not associated with anti-Ro autoantibodies in primary Sjögren’s
syndrome. J Rheumatol 2000;27:2945–6. Correction in: J Rheumatol 2001;28:684.
2. Fox RI, Tornwall J, Maruyama T, Stern M. Evolving concepts of
diagnosis, pathogenesis, and therapy of Sjögren’s syndrome. Curr
Opin Rheumatol 1998;10:446–56.
3. Peltonen L, Jalanko A, Varilo T. Molecular genetics of the Finnish
disease heritage. Hum Mol Genet 1999;8:1913–23.
4. Shifman S, Darvasi A. The value of isolated populations. Nat Genet
DOI 10.1002/art.10284
A chance meeting of two arthritis patients
To the Editor:
The 25th annual meeting of the American College of
Rheumatology had been unfolding in the expected manner,
during a pleasant week in San Francisco. I was scheduled to
give 2 presentations related to the cutaneous manifestations of
rheumatic diseases and to attend several administrative meetings of 2 patient advocacy organizations. However, this year’s
meeting will be memorable to me for an unexpected reason.
While taking a midafternoon break during a busy day
of meeting activities, I encountered a 30-something–year-old
black woman who was working as a cashier in the refreshments
area in the lower level of the Moscone Convention Center.
There was no rush, she was smiling, and a brief conversation
ensued. She was surprised that I, a dermatologist interested in
diseases like lupus, was attending this “rheumatology” meeting. Speaking of lupus, she told me that she had been having
problems with a sun-sensitive rash and joint aches, and that her
doctors had been having trouble making a diagnosis. She had
even seen a dermatologist, who had performed a biopsy of the
rash on the front of her upper chest to no avail. My curbside
recommendation was that she should return to her dermatologist for another biopsy the next time the rash appeared.
Often, those lupus butterflies and wolves can be coy with their
new victims. She smiled and thanked me.
Several days later, on a particularly warm and sunny
afternoon, while dodging autos in an intersection on my way to
the convention center, I happened upon a small elderly woman
holding a cane, who had just tripped on the curb and fallen.
While several others had stopped to ask if she was OK, mine
was the first hand to reach hers. After slowly standing up and
testing her weight on her left knee, she said that she thought
she would be able to walk back to her apartment. We very
cautiously began to retrace her steps back toward her apartment, which she said was only 1 block away. I had assumed
that, after regaining her composure, she would release my
hand and return home on her own. However, as we creeped
along together, her very warm and soft left hand did not even
begin to release its firm grip on mine.
As we inched along, she apologized for delaying me,
and I reassured her that I was in no particular hurry. She was
pleasantly surprised by the irony that I was attending a meeting
of arthritis specialists right across the street from her apartment on a day that had started for her with a visit to her own
doctor about the arthritis pain in her knees. Along our slow
journey together, I learned several things about her. She had
been a dentist in the Philippines but had not practiced in the
US since moving here 20 years ago. She was a poet and was
currently working on her second book. As we neared the
intersection where her apartment building stood, literally a
stone’s throw away from the convention center, she asked if I
would like to join her in her apartment for a cup of tea. I
declined, indicating that I needed to get along to the meeting.
We then parted at the doorway to her apartment building—the
very first time that her warm grip showed even the faintest
interest in loosening.
As I think back now, I question my decision favoring
another afternoon of subterranean science over a nice cup of
tea with an interesting Filipino dentist-poet in a high-rise
apartment building on a bright sunny day in downtown San
Richard D. Sontheimer, MD
University of Iowa College of Medicine
Iowa City, IA
DOI 10.1002/art.10278
Clinical Images: Hydroxychloroquine-associated mucocutaneous hyperpigmentation
Newly discovered therapeutic benefits and ongoing research have prompted an increase in the use of antimalarial therapy for
rheumatic disease. The images presented here demonstrate late clinical findings and classic histopathologic features of
mucocutaneous hyperpigmentation associated with long-term hydroxychloroquine administration. A, Lower legs of a patient with
longstanding, seropositive rheumatoid arthritis with multiple joint deformities. Chronic infection associated with shoulder
arthroplasty and osteomyelitis precluded the use of immunosuppressive agents. The patient was treated with hydroxychloroquine for
many years with good clinical results, but developed large areas of hyperpigmentation over all extremities, the torso, and the hairline.
The hyperpigmentation was so pronounced that in one instance, it resulted in a report of elder abuse by health care personnel. B,
Skin biopsy specimen from the same patient, demonstrating sharply defined, red-brown fibers in the deep dermis and the classic
“banana-shaped body” (inset) described to occur in exogenous ochronosis from antimalarial therapy.
The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views
of the United States Air Force or the Department of Defense.
Capt. David G. True, DO
Lt. Col. L. Ray Bryant, MD
Maj. Mark D. Harris, MD
Maj. Richard A. Bernert, MD
Lackland Air Force Base, Texas
Без категории
Размер файла
698 Кб
hulkkonen, promote, articles, syndromecomment, associations, allele, evidence, sjgren, primary, interleukin, gcc
Пожаловаться на содержимое документа