LETTERS
Bricfly, we would emphasize that wc did test a number in a genctically susceptible host. Future studies directed at
of methods for isolation of D N A from synovial fluids and
tissues. Based on extensive background testing, we selected the
methods reported as the most sensitive, by actual measure-
ment of sensitivity. Furthcrmorc, we would point out that the
method selected for isolation of DNA from synovial fluid is an
approach that has becn validated in the literature for extrac-
tion of microbial DNA from synovial fluid (2). Regarding the
primers used in the PCR, a broad set of primers was designed
and validated as described in our report. To determine sensi-
tivity and specificity of the primers selected for use, most
known pathogcnic human mycoplasmas were tested. As de-
scribed, extensive studies were done to optimizc the PCR
conditions and to document the specificity and sensitivity of
the assay. This included the use of nonhomologous internal
standard (molecular mimic) and the USK of intact mycoplasmas
and purified Mycoplasrnu DNA of known concentrations in
mixing expcriments. In our study, extensive testing was also
done to overcome the potential influence of synovial fluid and
synovial tissue inhibitors,
Regarding the use of molecular mimics for diagnostic
PCR analysis, we would recmphasize the importancc of such
constructs both as a control for DNA amplification and as a
critical measure of minimum effective sensitivity. The argu-
ment by Schaeverbeke et al that mimic added to sample prior
to DNA extraction may not be effectively "separated froin
bactcrial cell components and correctly extracted" is puzzling,
not only because it is difficult to imagine, but more impor-
tantly, because the result would be a lower, rather than higher,
apparent sensitivity if only a proportion of original input were
being measured. We continue to support the inclusion of
mimics prior to sample extraction, and argue that they serve as
valuable measures of thc actual (versus theoretical) minimal
sensitivity being measured. This would be a useful number to
determine in any study, including that of Schacverbcke et al.
Since any method has ii limit of detection, both theorctical and
practical, we behevc this actual level of sensitivity measured is
a critical parameter that should be reported.
We went to great lengths to avoid contamination of our
samples with kfycoplusrnii or PCR amplicons. While this is a
notorious problem, wc assumed that it was not the basis for the
rcsults of Schaeverbeke et al (I), and thcrefore we suggested
that passive carriage of an organism which we and others (3)
found to bc a common human commensal or pathogen was a
possible explanation for their findings. Our mention of possi-
ble differences in the 2 study populations was meant to refer to
normal parameters known to affect such studies, such as age,
bias in major histocompatibility complex backgrounds, history
of infectious diseases, ctc. It was not meant to reflect differ-
ences in nationality, as implied by Schaeverbekc and cowork-
ers. Furthermore, the finding by Schaeverheke et al (1) of the
prcsencc of M/ermetztutzs in the joints of paticnts with a variety
of rheumatic diseases of presumed divcrsc pathogenesis is
itself evidencc against chronic local Mferzcrifuns infection as
the etiology of RA. Our serologic finding that a n immune
reaponsc against M honzitzis and Mferrnenlun.~i s conimon, and
the results reportcd by Schaevcrbeke et al, are potentially
consistent with the alternative hypothesis that KA is caused by
an immune rcsponse against a previous Mycoplu.snza infection
757
testing these hypotheses are ncedcd.
Robert W. Hoffman, DO, FACP
University of Missouri and Huwy S Truman
Menlorial Veterans Hospital
Kim S. Wise, PhD
University o j Missouri
Columbia, MO
1. Schaevcrbeke T, Gilroy CB, BCbCar C, Dehais J, Taylor-Robinson
D. Mycoplasma fermcntans in joints of patients with rheumatoid
arthritis and other joint diseases [letter]. Lancet 1996;347:1418.
2. Nocton JJ, Dressler F, Rutledge BJ, Rys PN, Persing DH, Steere
AC. Detection of Borrellin hctl-gdofmi DNA by polymerase chain
reaction in synovial fluid from patients with Lyme arthritis. N Engl
J Med 1994;330:229-34.
3. Chingbingyong MI, Hughes CV. Detection of Mycoplasma fcr-
mentans in human saliva with a polymerase chain reaction-based
assay. Arch Oral Biol 1996;41:311-4.
HLA phenotype and systemic sclerosis-rheumatoid
arthritis overlap syndrome: comment on the article by
Horiki et a1
To the Editor:
We read with interest the article by Horiki et al
proposing that sclcroderma-rheumatoid arthritis (SSc-RA)
overlap syndrome may be a distinct clinical cntity with gener-
alized skin sclerosis, severe seropositive polyarthritis, pulmo-
nary fibrosis, anti-topoisomerase I antibodies, and a common
HLA-phenotype (1). The authors note that all 4 of their
HLA-typed patients carried the HLA-DR4,53;DQA1"03;
D Q B l V 4 haplotype (1). We havc seen 3 patients with
SSc-RA overlap syndrome who had very similar clinical man-
ifestations; however, their HLA phenotypes were not similar to
thosc of the patients described by Horiki and colleagues. We
determined our patients' HLA-DR and DQ genotypes by the
polymerase chain reaction-sequence-specific oligonucleotide
method as previously described (2,3). As shown in Table 1,
only patient 3 carried the DRBl"O405 (DK4),DRB4*01
(DKS3);DQA1*0303;DQB1*0401haplotype. Considering that
more than 60% of Japanese R A patients carry DRBl"O4
(DR4),DRB4"01 (DR53);DQA1"0303;DQBl "04 (4), it is
likely that all 4 patients of Horiki ct al had this haplotype by
chance.
HLA-DR4,53;DQAI *03;DQBlY04 in the Japanese
of 2 different haplotypes, I~RB1*0405,
03;DQB1"0401 and DRBl"0410,DRB4*01;
DQAI*O303;DQB1*0402, found in 26.5% and 3.8% of normal
unrelated Japanesc subjects, respectively (3). DRBl"0405 is
different from DRBl*0410 only by a single amino acid substi-
tution at codon 86 (Gly versus Val), and DQBI"0401 is
differcnt from DQBIV402 only at codon 23 (Leu versus Arg)
( 5 ) . However, thesc 2 haplotypes show contrasting association
with the genetic predisposition to insulin-dependent diabetes
mellitus (3). It is also well known that the DKBl"040S;
DQB I','O4OI haplotype is strongly positively associated with
R A in the Japanese population (4), but no significant associ-
ation between the DRB1"0410;DQB1"0402 haplotype and KA
has been reported to datc. I n the report by Horiki et al, 2 of the
758
Table 1. HLA-DRIDQ haplotypes of 3 patients with systemic
sclerosis-rheumatoid arthritis (SSc-RA) overlap syndrome
Patient 1 DRB1*1502,DRB5*0102,DQA1*0103,DQB1"06011
DRBl*0901,DRB4*0l,DQA1*0302,DQBl1 03032
Patient 2 DRB1"1502,DRB5*0102,DQAl0103,DQBl *06O11
DRBl*1201,DRB3"0202,DQAlA05O13,DQB1+0301
Patient 3 DRBl"0405,DRB4"01,DQAl "O303,DQB1'0401
DRB1*1401,DRB3*0202,DQAl+0104,DQB1h0502 netics of susceptibility to rheumatoid arthritis. Arthritis
4 patients with SSc-RA overlap syndrome carried
DRB1"0405;DQB1*0401 and 2 carried D R B l ^ 0 4 1 0 ;
DQBlV402. It does not seem reasonable to suggest a definite
association between a particular HLA phenotype and suscep-
tibility to SSc-RA overlap syndrome from a report with such a
small sample size.
Shin'ichiro Yasunaga, MD, PhD
Masanori Higuchi, MD, PhD
Hiroaki Nishizaka, MD, PhD
Shigeru Yoshizawa, MD, PhD
Takahiko Horiuchi, MD, PhD
Kyushu University
Fukuoka, Japan
1. Horiki T, Moriuchi .I,Takaya M, Uchiyama M, Hoshina Y, Inada K,
et al. The coexistence of systemic sclerosis and rheumatoid arthritis
in five patients: clinical and immunogenetic features suggest a
distinct entity. Arthritis Rheum 1996;39:152-6.
2. Kimura A, Sasazuki T. Eleventh international histocompatibility
workshop reference protocol for the HLA DNA-typing technique.
In: Tsuji K, Aizawa M, Sasazuki T, editors. HLA 1991: proceedings
of the eleventh international workshop and conference. Vol. 1.
Oxford: Oxford University Press; 1992. p. 397-419.
3. Yasunaga S, Kimura K, Hamaguchi K, Ronningen KS, Sasazuki T.
Different contribution of HLA-DR and -DQ genes in susceptibility
and resistance to insulin-dependent diabetes mellitus (IDDM).
Tissue Antigens 1996;47:37-48.
4. Tsuchiya K, Kondo M, Kimura A, Nishimura Y, Sasazuki T. The
HLA-DRB1 and/or the DQBl locus controls susceptibility and the
DRBl locus controls resistance to rheumatoid arthritis in the
Japanese. In: Tsuji K, Aizawa M, Sasazuki T, editors. HLA 1991:
proceedings of the eleventh international workshop and confer-
ence. Vol. 1. Oxford: Oxford University Press; 1992. p. 509-12.
5. Marsh SGE, Bodmer JG. HLA class TI region nucleotide scquences.
1905. Tissue Antigens 1995;45:258-80.
Reply
To the Editor:
We thank Drs. Yasunaga et a1 for their interest in our
report on the 5 patients with SSc and R A who showed severe
articular destruction. Interestingly, 10 patients reported to date
as having SSc-RA overlap and our 5 patients showed similar
clinical manifestations, such as pulmonary fibrosis and positive
anti-topoisomerase I (anti-topo I) antibodies. HLA studies of
4 of our 5 patients further revealed that all the patients had
HLA-DR4, DR53, DQA1*0301, and DQBl*04. In contrast,
this haplotype did not show a positive association in 42 other
SSc patients including 10 who were anti-topo 1 positive.
Drs. Yasunaga et al suggest that 2 of our patients
had DRB1*0405,DRB4*01;DQAl*0303;DQB Lw401 and
LETTERS
the remaining 2 had DRBl"'0410,DRB4:k01;DQA1"0303;
DQB l"0402, and that these 2 haplotypes show contrasting
association with insulin-dependent diabetes mellitus. Never-
theless, we would like to emphasize that the 2 haplotypes have
a common R A susccptibility D R P cpitope (amino acids 70-74)
(Gregersen PK, Silver J, Winchester RJ. The shared epitope
hypothcsis: an approach to understanding the molecular ge-
Rheum 1937;30:1205-1.3.) (Kimura A, Sasazuki T. Eleventh
international histocompatibility workshop reference protocol
for the HLA DNA-typing tcchniquc. In: Tsuji K, Aizawa M,
Sasazuki T, editors. HLA 1991: proceedings of the eleventh
international workshop and conference. Vol. 1. Oxford: Ox-
ford University Press; 1992. p. 397-419.).
Since we do not know the clinical details of the 3
patients mentioned by Yasunaga et al, we would like to reserve
further comment. We, howcvcr, indeed agrcc with the com-
ment that our report included too small a sample size to draw
definite conclusions on HLA ociations with SSc-RA over-
lap syndrome. To analyze HLA susceptibility to SSc-RA
overlap syndrome, studies of more patients, including precise
information on their clinical features, will be needed.
Terumi Horiki, MD, PhD
Junko Moriuchi, MD, PhD
Masatoshi Takaya, MD, PhD
Mitsuaki Uchiyama, MD, PhD
Yuichi Hoshina, MD, PhD
Hidetoshi Inoko, PhD
Kimiyoshi Tsuji, MD, PhD
Yukinobu Ichikawa, MD, PhD
Tokai Univer~sitySclzool of Medicine
Kanugawu, Japan
Kenichi Inada, MD, PhD
Aichi Cancer Center
Aiclzi, Japaii
Neck injury and chronic pain syndromes: comment on
the article by Buskila et a1
To the Editor:
Buskila et al demonstrated an increased rate of fibro-
myalgia after a neck injury as compared with the rate after a
lower extremity fracture (1). This is an important study and we
would like to suggest a potential explanation for their findings.
First, demonstrating an association is not equivalent to dem-
onstrating a causal relationship. To do this is often a difficult
task. One can reasonably accept that trauma caused a fracture,
because the 2 events are so closely related in time. On the
other hand, since fibromyalgia is usually diagnosed several
months or years following the trauma, the temporal relation-
ships are not close. There is thus the possibility that other
factors have intervened to lead to the diagnosis of fibromy-
algia.
Another means by which to demonstrate a causal
association, in a scientific sense, would be to reproduce the
exposure experimentally. As we have reviewed elsewhere, in
the more specific scenario of whiplash patients (2),none of the
patients studied has ever developed a chronic pain syndrome
despite more than 4 decades of simulated collisions to induce