Anda di halaman 1dari 6

Rheumatology 2006;45:iv26iv31 doi:10.

1093/rheumatology/kel309

Cardiac imaging in rheumatic diseases


R. Maksimovic, P. M. Seferovic1, A. D. Ristic1, B. Vujisic-Tesic1, D. S. Simeunovic1,
G. Radovanovic1, M. Matucci-Cerinic2 and B. Maisch3

The majority of the imaging techniques in cardiology could be applied in rheumatic diseases (RDs), such as
echocardiography, single-photon emission computed tomography (SPECT), radionuclide ventriculography, angiography,
cardiovascular MRI and CT. Inflammatory pericardial involvement is the most common cardiac manifestation in various
forms of RD. Echocardiography is the gold standard for diagnosis of pericardial abnormalities, demonstrating location
and amount of pericardial effusion. Cardiac MRI and CT can be used to assess the features of pericardial effusions
and peracardial structures. In patients with valvular heart disease in RD, transoesophageal echocardiography is a superior
method and offers reliable information about valve morphology, the severity of the disease and left ventricular (LV) function.
In addition, cardiac MRI is a valuable tool for the evaluation of valvular stenosis and regurgitation severity. Myocardial

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on November 26, 2015


involvement in RD is demonstrated by abnormalities in LV size and function, indicating myocardial inflammation. In
these patients Doppler echocardiography and myocardial tissue imaging can provide essential diagnostic information. Both
LV angiography and cardiac MRI can provide reliable information on LV size, function and mass. In patients with coronary
disease associated with RD, LV ejection fraction and ventricular wall motion can be assessed by echocardiography,
radionuclide ventriculography, gated SPECT and MRI. Three-dimensional (3D) echocardiography is considered superior
to 2D echocardiographic techniques. Stress echocardiography is the most used method for detection of myocardial
ischaemia. The only accurate visualization of the coronary arteries is by selective coronary arteriography, which remains
the gold standard. Although new non-invasive techniques have been developed, including CT and MRI angiography, some
limitations apply.

Introduction Involvement of pericardium


In patients with rheumatic diseases (RDs), cardiovascular Inflammatory pericardial involvement is the most common
involvement is common, may have serious consequences, and cardiac manifestation in various forms of RD [2, 3]. In RA,
can contribute to worsening of patients outcome (Table 1) [1]. pericardial effusion is seen in 210% of the patients as chronic,
Incidence and prevalence of cardiovascular diseases (CVD) in RD asymptomatic pericardial effusion, while tamponade and
have recently been largely increased due to both the improved constrictive pericarditis are rare (0.5%) [2]. In patients
biochemical and cardiac imaging diagnostic techniques. CVD with SLE, symptomatic pericardial effusion has been reported
can be diagnosed in more than 50% of the patients with systemic in 650%, clinically significant pericarditis and pericardial
lupus erythematosus (SLE) [2], and is the cause of nearly 40% thickening are described in <30%, while post-mortem pericardial
of deaths in rheumatoid arthritis (RA) patients [3]. Patients lesions range from 6080%. In SSc, pericardial involvement
with RA are more likely to have myocardial infarction and is seen in 70% of the patients, mostly as small pericardial
sudden death, and are prone to serious systemic complications. effusion [4].
Systemic sclerosis (SSc) is frequently associated with pulmonary Echocardiography is the most widely used imaging technique in
hypertension and/or systemic hypertension, which can lead to the evaluation of suspected pericardial disease. Transthoracic
subsequent complications. echocardiography represents the gold standard for diagnosis of
The majority of the imaging techniques in cardiology could pericardial abnormalities, demonstrating the location and
be also applied in RD, with variable diagnostic accuracy. amount of even minimal pericardial effusion. The echocardio-
These include echocardiography, single-photon emission com- gaphic features of pericardial effusion are the pericardial
puted tomography (SPECT), radionuclide ventriculography, layer separation with an echo-free space and the decrease in the
angiography, cardiovascular MRI and CT. Imaging techniques parietal pericardium motion. If tamponade develops, which is
are essential in the early detection of CVD in RD, having a rare complication of RD, a several echocardiographic signs
significant therapeutic implications and improving patients can be appreciated such as diastolic compression of right
long-term outcome. heart chambers, lack of inspiratory collapse of the dilated inferior

Center for Magnetic Resonance Imaging and 1Institute of Cardiovascular Diseases, University Clinical Center of Serbia, Belgrade, Serbia, 2Division of
Rheumatology, Department of Internal Medicine, University of Florence, Florence, Italy and 3Department of Internal Medicine-Cardiology, Philipps-
University, Marburg, Germany.

Correspondence to: R. Maksimovic, MD, PhD, Associate Professor of Radiology, Center for Magnetic Resonance Imaging, University Clinical Center
of Serbia, Koste Todorovica 8, 11 000 Belgrade, Serbia. E-mail: rmaksimovic@yahoo.com
iv26
The Author 2006. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org
Cardiac imaging in rheumatic diseases iv27

TABLE 1. Heart diseases in major rheumatic diseases

Pericardium Valves Myocardium Coronary arteries


RA /
SLE /
APLA /
SS /

RA, rheumatoid arthritis; SLE, systemic lupus erythematosus; APLA,


antiphospholipid antibody syndrome; , rare; , low incidence; ,
medium incidence; , high incidence.
PE

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on November 26, 2015


FIG. 2. Transoesophageal echocardiography, four-chamber view.
Pericardial effusion (arrowheads) and two endocardial vegetations
(arrows) in SLE. The vegetations are localized on atrial side of the
base of the mitral leaflets. LA, left atrium; LV, left ventricle; RA,
right atrium; RV, right ventricle; PE, pericardial effusion.

RV
FIG. 1. Echocardiography, parasternal short axis view.
Tamponade in SLE demonstrating diastolic right atrial free wall RA
collapse (arrow). Ao, aorta; ra, right atrium; la, left atrium. LV

LA
vena cava and the swinging of the entire heart [5] (Fig. 1).
Doppler echocardiographic patterns of tamponade include
marked respiratory variations of right and left ventricular (LV)
inflow and pulmonary and hepatic venous flow velocities.
Echocardiography is also very instrumental if combined pericar-
dial and valvular pathology is observed (Fig. 2).
Pericardial thickening resulting from fibrosis or calcification
can be presented as a small echo-free space or a single dense band
of echoes. Since the pericardium is the most echogenic structure in
the image, sensitivity and specificity of transthoracic echocardio-
graphy for pericardial thickening is low. Doppler assessment of FIG. 3. Cardiac MRI, four-chamber view. Cine TrueFISP
diastolic flow patterns and the respiratory changes in these sequence shows high signal intensity of a pericardial effusion in
patients can provide compelling evidence for pericardial constric- a patient with SSc (arrows). LV, left ventricle; LA, left atrium;
tion and rule out the diagnosis of restrictive cardiomyopathy and RV, right ventricle; RA, right atrium.
cardiac tamponade [6].
Both cardiac MRI and CT in RD can also be used to assess the
size and location of pericardial effusions, although they tend to
overestimate the amount of fluid in comparison with echocardio- ascites, hepatomegaly and pleural effusion. MRI has high
graphy [7] (Fig. 3). Pericardial effusion is hypodense in diagnostic accuracy for the definitive non-invasive diagnosis of
comparison with the myocardium, with radiodensity ranging constrictive pericarditis and its separation from restrictive
from 1040 Hounsfield units, depending on the protein content cardiomyopathy. Both the increased pericardial thickness and
(fibrin) and admixture of blood. If the amount of blood in the alterations in the heart structures can contribute to the diagnosis
pericardium is small, differentiation from fibrosis (chronic with sensitivity of 88%, specificity 100% and accuracy of 93% [9].
constrictive pericarditis), which displays about the same density, Furthermore, MRI could quantify functional abnormalities,
may be impossible [8]. namely LV systolic and diastolic dysfunction, which may be
MRI is useful in patients with loculated or complex configura- associated with pericardial diseases [10].
tions of pericardial effusions, which can be often found in RD.
MRI has an advantage over CT in differentiating small pericardial
effusions from pericardial thickening. However, MRI cannot
Involvement of heart valves
distinguish between chronic pericardial thickening and calcifica-
tions. CT and MRI diagnosis of constrictive pericarditis is Echocardiography is superior and the most frequently used
established only if the findings of systemic venous hypertension technique for the evaluation of the patients with valvular heart
are confirmed, such as dilated superior and inferior vena cava, disease in RD. This method offers the non-invasive and reliable
iv28 R. Maksimovic et al.

information about valve morphology, the severity of valve


stenosis or regurgitation, and the impact of valvular lesion on
LV size and function. Transoesophageal echocardiography is
considered to be more sensitive than transthoracic echocardio-
graphy in revealing valve abnormalities in RD [11]. The
haemodynamic data are derived from Doppler echocardiography,
and valvular regurgitation can be assessed by colour flow imaging
of regurgitant jet.
Valvular disease can be frequently diagnosed in SLE, as
valvular thickening, vegetations and valvular insufficiency.
As the most typical lesion, the LibmanSacks endocarditis
(atypical verrucous endocarditis) is described, echocardiographi-
cally characterized by vegetations usually <10 mm in diameter,
which vary in shape, have irregular borders, and have various
echodensities (Fig. 4). Lesions are easily echocardiographically
detected and are usually attached to the basilar portion of the
aortic and mitral valve leaflets. The verrucae are near the edge and
in only 34% of the patients haemodynamically significant.
Echocardiographic anatomical abnormalities were reported in 40 FIG. 4. Transoesophageal echocardiography in a patient with

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on November 26, 2015


50% of these cases, with much higher sensitivity for transoeso- SLE. Endocardial vegetations (arrow) on the ventricular side of
phageal echocardiography in 5060% patients [2]. the mitral leaflet (LibmanSacks endocarditis). lv, left ventricle;
Valvular involvement in RA patients can be also confirmed la, left atrium.
by echocardiography, which usually reveal aortic valve lesions in
514% and mitral valve abnormalities (thickening and prolapse)
in 535% of the patients, depending on the imaging technique congestive heart failure due to SLE and acute myocarditis are not
applied [6]. common, the latter could be initial clinical presentation of the
In SSc, nodular thickening of the mitral valve could be seen in disease. Acute myocarditis has been reported in 825% of these
38% of the patients, while in only a few cases mitral and tricuspid patients, but due to introduction of steroid therapy decreased to
vegetations are reported [12]. In patients with antiphospholipid 5% [18]. In SS, myocardial involvement can manifest as incipient
syndrome, heart valve abnormalities are the most common cardiomyopathy that could progress to ventricular dysfunction
cardiac manifestations and could be presented as vegetations, and heart failure [4].
thickening and valvular dysfunction [13]. Mitral valve thickening Echocardiography is considered an accurate tool for the
is found on echocardiogaphy in 63% of the patients, followed by assessment of LV size and function, which are indicative of
the aortic (32%) and tricuspid valve thickening in 8% of the myocardial inflammation and/or dysfunction. 2D echocardiogra-
patients [14]. Ankylosing spondylitis is associated with aortic root phy could be of limited value, but Doppler echocardiography and
and valve disease [15]. Aortic valve thickening is predominantly myocardial tissue imaging can provide essential diagnostic
echocardiographically manifested as nodularities, usually single, information. The analysis of impaired myocardial relaxation,
homogeneously echoreflectant, with well-defined borders, located decreased compliance and increased filling pressure, are important
on any portion of the three cusps. Mitral valve thickening could be clues in revealing LV diastolic dysfunction [2].
also common and is predominantly localized to the basal anterior LV angiography demonstrates LV volumes and function and is
mitral leaflet. routinely used in the assessment of these patients. LV angiography
Cardiac MRI is a valuable tool for evaluation of valvular using technetium 99 m-labelled red cells can be used to evaluate
regurgitation severity and quantification of the effects of valvular LV function, but has largely been replaced by echocardiography
lesion on LV volumes, function and mass [10]. The accepted MRI and recently MRI [10]. CT is not considered the tool of choice for
method for qualitative assessment of severity of valvular the assessment of LV function in RD due to radiation and
regurgitation is based on the dark signal extending from the application of contrast media.
diseased valve into the receiving chamber during a variable time Cardiac MRI is superior to other techniques in assessment
period of the cardiac cycle. For the valve stenosis evaluation, of LV size, function and mass because it is non-invasive, has high
signal loss distal to an abnormal valve is appreciated, but could spatial resolution, high reproducibility and low interobserver
not be used as an accurate parameter because it is dependent on and intraobserver variability. The results are comparable to
MRI settings. Non-invasive quantitative assessment of blood flow SPECT and the method is more accurate than 2D echocardio-
can be performed by velocity-encoded cine MRI, not yet available graphy and radionuclide ventriculography [10]. Myocardial
for clinical use. Furthermore, in the absence of a cardiac shunt, fibrosis, which is considered as a non-specific finding, is
any discrepancy between stroke volumes in a patient with demonstrated as a low signal intensity area on T2-weighted
regurgitation will identify the regurgitant LV volume. images and as a delayed enhanced region after contrast admin-
Disadvantage of this method is that only patients with a single istration. Inflammatory oedema consistent with myocarditis,
regurgitant valve can be assessed although combination with could be seen as multiple focal areas of increased myocardial
velocity-encoded cine MRI allows assessment and quantification signal intensity on T2-weighted images, and early and late
of multivalvular disease [10]. gadolinium enhancement images [19] (Fig. 5). Similar to this,
increased gallium-67 citrate uptake may be indicative of a
myocardial interstitial inflammation [20].
Involvement of myocardium
The myocardium is frequently affected in RD by means of
different immunological mechanisms. In RA, myocardium was
Involvement of coronary arteries
involved in 19% of the cases on post-mortem study by Lebowitz In SLE, risk of developing coronary artery disease (CAD) is
[16]. Although myocarditis is not usually associated with RA, 48 times higher than in controls, while in the middle-aged
secondary amyloidosis could cause cardiomyopathy and women with long-standing disease and a long period of
atrioventricular conduction abnormalities [17]. Acute or chronic corticosteroid intake the risk is increased 50-fold than in the
Cardiac imaging in rheumatic diseases iv29

A B

LV
LV

RV
RV

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on November 26, 2015


FIG. 5. Cardiac MRI in myocarditis in patient with SLE. (A) Hypocontractility is demonstrated in the lateral wall of the left ventricle in
cine TrueFISP sequence as an evidence of ongoing inflammation (arrows). (B) Late enhancement image shows intramural high signal
intensity areas (arrows) due to focal inflammatory necrosis. LV, left ventricle; RV, right ventricle.

matching age group [2]. In patients with RA, risk of CAD is The diagnostic accuracy of adenosine stress cardiac MRI varies
23-fold increased in comparison with the general population. significantly according to the MRI parameter evaluated.
In addition, silent ischaemia and sudden cardiac death are more Ischaemic territories can be identified with high sensitivity and
likely to be experienced by RA patients [1]. specificity by using intravenous gadolinium for a first-pass
For assessment of LV ejection fraction and ventricular wall perfusion, when perfusion abnormalities were confined to the
motion in CAD, radionuclide ventriculography, echocardio- subendocardium. Dark areas of low perfusion could be identified
graphy, gated SPECT and MRI could be used. According to visually or using computer analysis (Fig. 6). Usually, both stress
American Heart Association/American College of Cardiology and rest myocardial perfusion scans are used and are comparable
(AHA/ACC) Task Force on Practice Guidelines in CAD, radio- with myocardial perfusion reserve using flow wire probe, positron
nuclide ventriculography should be used for serial measurements emission tomography (PET) and SPECT [22]. More simple
of LV ejection fraction [21]. Gated SPECT myocardial perfusion approach is to assess stress myocardial perfusion only during
nuclear imaging is preferred for simultaneous assessment of vasodilatation, and using late gadolinium enhancement to define
myocardial perfusion and function [22]. areas of non-viability.
2D echocardiography is suggested for routine assessment of LV Severity of stenotic coronary lesion could be measured by
ejection fraction in patients with CAD in RD. 3D echocardio- coronary flow reserve and direct visualization of coronary artery
graphy is superior to both quantitative and qualitative 2D tree. Coronary flow reserve could be estimated with the
echocardiographic techniques and correlates well with measure- combination of coronary flow assessment by Doppler and
ments made using cardiac MRI. Tissue Doppler imaging study vasodilatator stress, which increases accuracy of this technique
[23] revealed that early diastolic velocities and mitral E/A ratio are [14]. Determination of coronary flow by cardiac MRI in humans
lower in patients with CAD in comparison with the controls. has been only reported to identify stenosis of the left anterior
Recently, a new global index of left ventricular function (TEI descending artery [10].
The most accurate visualization of the coronary arteries is by
index), was introduced by Turiel et al. [14] aimed to assess systolic
selective coronary arteriography, which remains a gold standard.
and diastolic LV function, and was confirmed to be higher in
Although new non-invasive techniques have been developed,
patients with RA and SS compared with control subjects.
including CT and MRI angiography, some limitations apply.
Stress echocardiography is a useful method for detection of
Coronary angiography using 64-slice CT scanners, does not
myocardial ischaemia and for risk stratification of patients with
identify all coronary segments (88%), and has sensitivity of
suspected or known CAD. Stressors include exercise, pharmaco- 88100%, and specificity of 8597% per patient. Angiography
logical agents and pacing. In patients with LV dysfunction and with MRI has even lower diagnostic accuracy than multislice CT
documented CAD, stress echocardiography can differentiate angiography [25]. However, neither of the techniques is an
viable from scarred myocardium. In patients who are unable to alterantive to selective coronary angiography and neither is
exercise, pharmacological stressors can be used: sympathomimetic recommended for screening asymptomatic patients, including
agents (dobutamine) or vasodilatators (dipyridamole or adeno- those at high risk [26].
sine). This technique is important to differentiate myocardial
stunning from necrosis, and it is considered safe, relatively The authors have declared no conflicts of interest.
inexpensive and easy to perform.
Cardiac MRI pharmacological stress test is well-established to
identify ischaemia-induced wall motion abnormalities. Acknowledgements
Dobutamine MRI has greater sensitivity than dobutamine This work is part of the project supported by the Ministry of
echocardiography (86 vs 74%) and specificity (86 vs 70%) [24]. Science and Environmental Protection, Republic of Serbia,
iv30 R. Maksimovic et al.

A B

LV
RV

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on November 26, 2015


FIG. 6. Cardiac MRI in myocardial ischaemia in a patients with SLE. (A) TrueFISP sequence with saturationrecovery preparation
shows perfusion defect in the lateral wall of the left ventricle (arrows). (B) This finding corresponds to angiographically proven
high-grade stenosis of left circumflex coronary artery (arrow). LV, left ventricle; RV, right ventricle.

Uncovering etiopathogenesis of heart muscle disease, pericardial 9. Masui T, Finck S, Higgins CB. Constrictive pericarditis and restrictive
disease, and congestive heart failure as a comprehensive back- cardiomyopathy: evaluation with MR imaging. Radiology
ground for diagnosis and treatment: application of neuronal 1992;182:36973.
network and artificial intelligence, (No. 145063) and Joint 10. Pennel D, Sechtem PU, Higgins BC et al. Clinical indications for
European Project, Tempus (No. 19046-2004), Restructuring of cardiovascular magnetic resonance (CMR): consensus panel report.
postgraduate studies and training in medicine in Serbia. Eur Heart J 2004;25:194065.
11. Gonzales-Gay AM, Gonzales-Juanatey C, Martin J. Rheumatoid
arthritis: a disease associated with accelerated atherogenesis. Semin
Arthritis Rheum 2005;35:817.
References 12. DAngelo WA, Fries JF, Masi AT, Shulman LE. Pathologic
1. Mandell FB, Hoffman SG. Rheumatic diseases and the cardiovas- observations in systemic sclerosis (scleroderma). Am J Med
cular system. In: Zippes PD, Libby P, Bonow OR, Braunwald E, eds. 1969;46:42840.
Heart disease: a textbook of cardiovascular medicine, International, 13. Tenedios F, Erkan D, Lockshin MD. Cardiac involvement in the
7th edn. Elsevier Saunders, 2005;210116. antiphospholipid syndrome. Lupus 2005;14:6916.
2. Doria A, Iaccarino L, Sarzi-Puttini P, Atzeni F, Turriel M, Petri M. 14. Turiel M, Muzzupappa S, Gottardi B et al. Evaluation of cardiac
Cardiac involvement in systemic lupus erythematosus. Lupus abnormalities and embolic sources in primary antiphospholipid
2005;14:6836. syndrome by transeophageal echocardiography. Lupus 2000;9:
3. Gerli R, Goodson NJ. Cardiovascular involvement in rheumatoid 40612.
arteritis. Lupus 2005;14:67982. 15. Roldan C, Chavez F, Wiest PH, Qualls C, Crawford M. Aortic root
4. Ferri C, Giuggioli D, Sebastiani M, Colaci M, Emdin M. Heart disease and valve disease associated with ankylosing spondylitis. J Am
involvement and systemic sclerosis. Lupus 2005;14:7027. Coll Cardiol 1998;32:1397404.
5. Arslan S, Bozkurt E, Sari RA, Erol MK. Diastolic function 16. Lebowitz WB. The heart in rheumatoid arthritis (rheumatoid disease):
abnormalities in active rheumatoid arthritis evaluation by conven- a clinical and pathological study of 62 cases. Ann Intern Med
tional Doppler and tissue Doppler: relation with duration of disease. 1963;58:10223.
Clin Rheumatol 2006;25:2949. 17. Riboldi P, Gerosa M, Luzzana C, Catelli L. Cardiac involvement
6. Wislowska M, Sypula S, Kowalic I. Echocardiographic findings, in systemic autoimmune diseases. Clin Rev Allergy Immunol
24-hour electrocardiographic Holter monitoring in patients with 2002;23:24761.
rheumatoid arthritis according to Steinbrockers criteria, functional 18. Roman MJ, Salmon JE, Sobel R et al. Prevalence and relation to risk
index, value of Waaler-Rose titre and duration of disease. Clin factors of carotid atherosclerosis and left ventricular hypertrophy
Rheumatol 1998;17:36977. in systemic lupus erythematosus and antiphospholipid antibody
7. Maisch B, Seferovic PM, Ristic AD et al. Guidelines on the diagnosis syndrome. Am J Cardiol 2001;87:6636.
and management of pericardial diseases. The Task Force on the 19. Dziadzio M, Giovagnoni A, Pomponio G et al. Acute myocarditis
Diagnosis and Management of Pericardial Diseases of the European associated with adenoviral infection in a patient with scleroderma.
Society of Cardiology. Eur Heart J 2004;25:587610. Clin Rheumatol 2003;22:48790.
8. Maksimovic R, Goldner B. Computed tomography and magnetic 20. Gaal J, Hegedus I, Devenyi K, Czirjak L. Myocardial gallium-67
resonance imaging in pericardial disease. In: Seferovic PM, citrate scintigraphy in patients with systemic sclerosis. Ann Rheum
Spodick DH, Maisch B, Shabetai R, Maksimovic R, Ristic AD, Dis 1995;54:8568.
co-eds. Pericardiology: contemporary answers to continuing chal- 21. Selvanayagam JB, Jerosch-Herold M, Porto I et al. Resting
lenges. Beograd: Nauka, 2000;12736. myocardial blood flow is impaired in hibernating myocardium: a
Cardiac imaging in rheumatic diseases iv31

magnetic resonance study of quantitative perfusion assessment. 24. Nagel E, Lorenz C, Baer F et al. Stress cardiovascular magnetic
Circulation 2005;112:328996. resonance: consensus panel report. J Cardiovasc Magn Reson
22. Schwitter J, Nanz D, Kneifel S et al. Assessment of myocardial 2001;3:26781.
perfusion in coronary artery disease by magnetic resonance: a 25. Kim WY, Danias PG, Stuber M et al. Coronary magnetic resonance
comparison with positron emission tomography and coronary angiography for the detection of coronary stenosis. N Engl J Med
angiography. Circulation 2001;103:22305. 2001;345:18639.
23. Specker C, Perniok A, Brauckmann U, Siebler M, Schneider M. 26. Kefer J, Coche E, Legros G et al. Head-to-head comparison of three-
Detection of cerebral microemboli in APS-introducing a novel dimensional navigator-gated magnetic resonance imaging and 16-slice
investigation method and implications of analogies with carotid computed tomography to detect coronary artery stenosis in patients.
artery disease. Lupus 1998;(Suppl 2):S7580. J Am Coll Cardiol 2005;46:92100.

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on November 26, 2015

Anda mungkin juga menyukai