Article Text
PDF
Abstract
Objectives To report methodology and overall clinical, laboratory and radiographic characteristics for Henoch–Schönlein purpura (HSP), childhood polyarteritis nodosa (c-PAN), c-Wegener granulomatosis (c-WG) and c-Takayasu arteritis (c-TA) classification criteria.
Methods The preliminary Vienna 2005 consensus conference, which proposed preliminary criteria for paediatric vasculitides, was followed by a EULAR/PRINTO/PRES - supported validation project divided into three main steps. Step 1: retrospective/prospective web-data collection for HSP, c-PAN, c-WG and c-TA, with age at diagnosis ≤18 years. Step 2: blinded classification by consensus panel of a subgroup of 280 cases (128 difficult cases, 152 randomly selected) enabling expert diagnostic verification. Step 3: Ankara 2008 Consensus Conference and statistical evaluation (sensitivity, specificity, area under the curve, κ-agreement) using as ‘gold standard’ the final consensus classification or original treating physician diagnosis.
Results A total of 1183/1398 (85%) samples collected were available for analysis: 827 HSP, 150 c-PAN, 60 c-WG, 87 c-TA and 59 c-other. Prevalence, signs/symptoms, laboratory, biopsy and imaging reports were consistent with the clinical picture of the four c-vasculitides. A representative subgroup of 280 patients was blinded to the treating physician diagnosis and classified by a consensus panel, with a κ-agreement of 0.96 for HSP (95% CI 0.84 to 1), 0.88 for c-WG (95% CI 0.76 to 0.99), 0.84 for c-TA (95% CI 0.73 to 0.96) and 0.73 for c-PAN (95% CI 0.62 to 0.84), with an overall κ of 0.79 (95% CI 0.73 to 0.84).
Conclusion EULAR/PRINTO/PRES propose validated classification criteria for HSP, c-PAN, c-WG and c-TA, with substantial/almost perfect agreement with the final consensus classification or original treating physician diagnosis.
Statistics from Altmetric.com
Introduction
The patient groups upon which the American College of Rheumatology (ACR) based the vasculitis classification criteria did not include children.1 2 However, since then paediatricians have relied on these and other adult-based criteria3,–,7 for their patients. Although children/adolescents and adults with vasculitis share many signs and symptoms of disease, they differ in the relative frequency of some clinical manifestations and concomitant diseases.8 9 Therefore, it cannot be assumed a priori that the classification criteria developed for adults are suitable for children and adolescents.
In 2005 the vasculitis working group of the Paediatric Rheumatology European Society (PRES) proposed preliminary classification criteria, for some of the most common childhood vasculitides10—namely, Henoch–Schönlein purpura (HSP), childhood polyarteritis nodosa (c-PAN), c-Wegener granulomatosis (c-WG), c-Takayasu arteritis (c-TA) and Kawasaki disease. Subsequently, with support from the European League Against Rheumatism (EULAR), the Paediatric Rheumatology International Trials Organisation (PRINTO)11 and PRES, PRINTO/PRES established a formal statistical validation process with a large-scale data collection that culminated in the final 2008 Ankara Consensus Conference.
In this first paper we describe the general methodology and overall clinical, laboratory and radiographic characteristics, while the specific details of the final classification criteria for each of the four vasculitides analysed (HSP, c-TA, c-PAN and c-WG) are reported in the accompanying paper.12
Patients and methods
Consensus formation techniques
Two well-recognised consensus formation methodologies, specifically designed to combine judgments from a group of experts in a particular field, were used to attain the goals: the Delphi technique and nominal group technique (NGT).13 14 The Delphi technique uses a series of well-defined mail questionnaires while NGT is a structured face-to-face meeting designed to facilitate reaching consensus, through round robin discussion.7 15,–,22
The 2005 Vienna Consensus conference
The general objective was to reach a consensus on classification criteria for childhood vasculitis, as previously described.10 In brief, consensus (≥80%) was reached to base the general working classification for childhood vasculitides on vessel size: predominantly large-, medium- or small-vessel vasculitis (granulomatous and non-granulomatous), and other vasculitides (box 1). Preliminary classification criteria based on consensus discussion were proposed for HSP, c-PAN, c-WG, or c-TA and Kawasaki disease as reported by Ozen et al.10
Box 1 Classification of childhood vasculitis
Predominantly large-vessel vasculitis
Takayasu arteritis
Predominantly medium-sized vessel vasculitis
Childhood polyarteritis nodosa (c-PAN)
Cutaneous polyarteritis
Kawasaki disease
Predominantly small-vessel vasculitis
Granulomatous
Wegener granulomatosis
Churg–Strauss syndrome
Non-granulomatous
Henoch–Schönlein purpura
Microscopic polyangiitis
Microscopic polyangiitis
Hypocomplementaemic urticarial vasculitis
Other vasculitides
Behçet disease
Vasculitis secondary to infection (including Hep B associated PAN), malignancies and drugs, including hypersensitivity vasculitis
Vasculitis associated with connective tissue diseases
Isolated vasculitis of the central nervous system
Cogan syndrome
Unclassified
Reproduced from Ozen et al.10
Validation of the preliminary criteria
After obtaining consent from parent(s)/child and ethics committee approval (as required by the national law in each participating country), 97 PRINTO/PRES institutions in 36 countries (see complete list at the end of the paper) enrolled children with selected vasculitides, into a three-step study.
Step 1: Web-based data collection
Inclusion criteria were children with age at diagnosis ≤18 years, diagnosed by their treating physician, as HSP, c-PAN, c-WG, c-TA or other c-primary systemic vasculitis (c-other). The other rarest forms of vasculitis (cryoglobulinaemic vasculitis, Churg–Strauss syndrome) were excluded because of their low prevalence while Kawasaki syndrome was excluded because it is under evaluation by another paediatric group (personal communication).
The five-page, web-based case report form included demographic data, clinical diagnosis made by the attending physician, presence of signs/symptoms in 12 broad organs/systems (general, mucosa and skin, eye, ears/nose/throat, respiratory, cardiovascular, neurological, gastrointestinal, renal, genital, musculoskeletal, other features of active vasculitis) before or at the date of diagnosis and at least 3 months later (in this series data after diagnosis are not presented), laboratory values (antineutrophilic cytoplasmic antibody (ANCA) testing by immunofluorescence/MPO/PR3, renal features, inflammation indices, platelets), physician global assessment of disease activity on a 10 cm visual analogue scale, biopsy findings (renal, skin, upper airway, other) and imaging reports (chest x-ray, CT scan, MRI, conventional, CT or MRI angiography findings). A three-page glossary, adapted from the Birmingham Vasculitis Activity Score glossary,23,–,25 provided definition of all items included in the case report forms. Data were abstracted by review of the clinical charts. Data collection was both retrospective for diagnosis before May 2007 and prospective for the remaining patients (figure 1).
The final enrolment figures. #More than one reason possible for exclusion. *Includes 45 cutaneous PAN, 14 microscopic polyangiitis. ∧222/280 (79%) Cases analysed by the consensus panel on which consensus was achieved (see step 2 in the “Patients and methods” and “Results” sections for further details). c-other, childhood-unknown vasculitis; c-PAN, c-polyarteritis nodosa; c-TA, c-Takayasu arteritis; c-WG, c-Wegener granulomatosis; HSP, Henoch–Schönlein purpura.
Data entry was carried out online through the member area of the PRINTO website (http://www.printo.it, accessed 11 March 2010) on a secured https platform. The electronic forms contained some predefined rules to avoid errors and missing data, but were also reviewed for consistency by a dedicated PRINTO research assistant.
The aim was to collect data for approximately 150 patients with HSP, 50 with c-TA, 80 with c-PAN and 40 with c-WG using each group of vasculitis as comparator for the others.2 These numbers were realistic projections of patient numbers required to provide reliable statistical analyses based on the relative frequencies of each disease.
Step 2: Classification by consensus panel
This step was included in order to further clean the original dataset collected, thus allowing a robust final patient dataset on which to base statistical validation of the classification criteria. Three physicians from the PRINTO international coordinating centre (NR, AP, GF) screened out all submitted cases. A total of 128 patients who did not fulfil ACR or EULAR classification criteria or were misclassified3 10 (so called ‘difficult cases’) but were nevertheless diagnosed by the treating physicians as HSP, c-PAN, c-WG, c-TA or c-other were retrieved; in addition, 152 randomly selected patients were added (38 HSP, 23 c-PAN, 41 c-WG, 50 c-TA). Using three Delphi web-rounds, the representative subgroup of 280 were classified by a panel of 11 paediatric rheumatologists/nephrologists into HSP, c-PAN, c-WG, c-TA or c-other. The panel was presented with the complete time-of-diagnosis case report in web-based form, but was blinded to the original diagnosis assigned by the treating physician. A consensus ≥80% within the 11 panellists was required to establish the final classification of each difficult case. Patients for whom consensus did not reach ≥80% in the first round, were re-evaluated in the second and third rounds, where panellists were made aware of each other's evaluations and were allowed to engage in a web-based collegial discussion of the specific cases. All patients for whom consensus was not achieved (<80%) were excluded from further analysis.
We then calculated the κ level of agreement,26 27 with 95% CI between the consensus panel classification and the attending physician diagnosis; we expected to find a κ>0.8 (almost perfect agreement)27 between the consensus panel and the treating physician. The expected almost perfect agreement was crucial in order to allow combination, in the third step, of both the difficult cases for whom the panellists reached classification consensus and the remaining patients diagnosed by the referring physician (see figure 1). The final diagnosis attribution for each child in the third step was given therefore by the consensus classification or by the original treating physician diagnosis.
Step 3: Statistical and consensus evaluations
An NGT consensus conference was convened in Ankara in October 2008, to discuss the statistical performance (frequency, sensitivity/specificity, area under the curve (AUC) and κ level of agreement) of clinical/laboratory findings (criteria) and of different classification criteria (about 50 for each disease) for HSP, c-TA, c-PAN and c-WG. All classification criteria tested were developed by the PRINTO international coordinating centre (NR, AP, GF) based on combinations of criteria derived from the literature (eg, ACR classification criteria, Vienna childhood criteria, etc) or created ad hoc (eg, for HSP 2/5 criteria, 3/5 criteria, etc). For the purposes of the statistical analysis, cases (eg, HSP classified by the panel or HSP diagnosed by the treating physician) were compared with the control group represented by the remaining patients with vasculitis with other diagnoses (eg, c-PAN, c-WG, c-TA, c-others). This use of other vasculitides as controls was employed in the original ACR adult vasculitis classification exercise.1 2
Statistics
Descriptive statistics, using the final classification criteria, were reported as means ± SD, whereas categorical variables were reported as absolute frequencies and percentages. Laboratory values were standardised based on the normal values provided by each local laboratory, as previously described.16 Comparison of frequencies was made by the χ2 or the Fisher exact text (post hoc with Bonferroni's correction), while comparisons of means were made by the analysis of variance (parametric or non-parametric) with post hoc comparisons (Scheffé or Dunn's test) as appropriate.
For each individual criterion/classification definition, we calculated the sensitivity (ability to recognise patients with the particular disease calculated as number of patients positive for the criterion/definition divided by the total number of patients with the disease) and specificity (ability to exclude patients with other diseases calculated as number of patients negative for the criterion/definition divided by the total number of patients without the disease). Sensitivity and specificity for either the individual criterion (eg, purpura for HSP), reported in the accompanying paper,12 and the classification definitions (eg, Vienna preliminary HSP criteria) were calculated using as ‘gold standard’ the final disease consensus classification or original treating physician diagnosis. Moreover, we evaluated in the second step, the agreement between the consensus panel classification and the diagnosis made by the treating physician, and in the third step, the agreement between the definitions tested and the final disease consensus panel or original diagnosis by the treating physician by means of the κ-coefficient,26 with the following cut-off levels proposed by Landis and Koch27: 0.01–0.2 = slight; 0.21–0.4 = fair; 0.41–0.6 = moderate; 0.61–0.8 = substantial; 0.81–1 = almost perfect agreement.
Sample size calculation for the diagnostic accuracy study (study in which the main objective is to evaluate the diagnostic accuracy of clinical, laboratory and other variables for the chosen outcome—in this case classification criteria; parameters for diagnostic accuracy are sensitivity, specificity, AUC) was based on the assumption that the required minimum AUC of the receiving operating curves was 85% with standard error of 3.5% giving a sample of 60 pairs of patients (60 cases vs 60 controls) for each specific vasculitis.28
Data were entered in an Access XP database and analysed with Excel XP (Microsoft), XLSTAT 6.1.9 Addinsoft, Statistica 6.0 (StatSoft), and Stata 7.0 (Stata Corporation).
Results
Step 1: Web-based data collection
Figure 1 shows the final enrolment figures. Of the 1398 children enrolled, 860 (62%) were diagnosed by the treating physician as HSP, 172 (12%) as c-PAN, 67 (5%) as c-WG, 99 (7%) as c-TA and 200 (14%) as c-other. Of the patients with HSP, c-PAN, c-WG and c-TA entered in the PRINTO web system, 74 (5%) were excluded for the following multiple reasons: 19 were beyond age limit, eight had associated co-morbid conditions, nine had incomplete data and 45 were excluded by the consensus panel classification (see step 2). Of the 200 remaining patients (c-other), 45 patients had cutaneous PAN; 14 microscopic polyangiitis; 13 were excluded by the consensus classification exercise because consensus could not be reached (k<80%); and the remaining 128 were excluded because a specific diagnosis was not given by the treating physician (unknown form of vasculitis).
Step 2: Classification by consensus panel
A total of 280/1398 (20%) children enrolled in the web-based system were classified by the consensus panel (figure 1 and table 1), blinded to the centre diagnosis. Three round robin web-based classification cycles allowed consensus ≥80% to be reached for the classification of 56 HSP, 38 c-PAN, 53 c-WG, 48 c-TA and 27 patients with c-other vasculitis. For the remaining 58 (20%) consensus was <80% and therefore these patients were discarded from further consideration (figure 1). Table 1 shows the grid of agreement for the four vasculitides, as well as the κ-coefficient between the consensus panel classification and original patient diagnosis made by the treating physician. The overall κ-agreement was 0.79 (substantial, 95% CI 0.73 to 0.84) for the 280 patients classified considered altogether. The κ-agreement was >0.8 (almost perfect agreement) for HSP, c-WG and c-TA and substantial (0.73) for c-PAN. These high levels of agreement justified the combination of either the subgroup of patients classified by the consensus panel and the remaining patients diagnosed by the treating physician into the next step.
κ Level of agreement (95% CI) between the consensus panel classification (columns) and the diagnosis of the treating physician (rows).
Step 3: Statistical and consensus evaluations
For the face-to-face Ankara 2008 Consensus Conference there were therefore a total of 1183/1398 (85%) of the original patients available for the analysis (figure 1). Final disease classification for these 1183 children, was represented by the original diagnosis made by the treating physician for 961 (81%) patients (771 HSP, 112 c-PAN, 7 c-WG, 39 c-TA and 32 c-other) and the panel consensus classification for the remaining 222 (19%) (56 HSP, 38 c-PAN, 53 c-WG, 48 c-TA and 27 c-other). The 59 c-other comprised 45 patients with cutaneous PAN and 14 children with microscopic polyangiitis.
Sensitivity, specificity, AUC and κ level of agreement (measured between each definition and the final disease classification) of the final EULAR/PRINTO/PRES classification definitions and the related individual criteria for the four childhood vasculitides are reported in the accompanying paper12 and in an online supplementary table.
Demographic and clinical characterisation according to the final EULAR/PRINTO/PRES classification criteria
As shown in table 2, children with HSP had younger age at onset, younger age at first visit and the shortest disease duration, followed sequentially by c-PAN, c-TA and c-WG. The time from onset of signs or symptoms to diagnosis was shortest in HSP (0.1±0.3 years), followed by c-PAN and c-WG while the longest period was observed for c-TA (1.3±1.6 years). Patients were predominantly Caucasian (82%) (p values <0.0001 for all comparisons).
Baseline demographic and clinical data for the 1124 patients considered in the final analysis (see also figure 1)
In table 3 are reported the absolute frequencies, percentage frequencies and related p values (all <0.0001 except testicular pain or tenderness equal to 0.019) of the signs/symptoms (individually or in combination) for the 1124 patients included in the final analysis (data for the 59 c-other are not shown in the table). Signs/symptoms reported refer to those most frequently observed in each different form of vasculitis. General signs/symptoms (malaise or fever) were more frequently observed in the three non-HSP systemic vasculitides 245 (82%) (122 c-PAN, 53 c-WG and 70 c-TA) and were less common in HSP 386 (47%). For mucosa and skin involvement, oral/nasal ulceration were typical of c-WG, occurring in 29 (48%) cases; purpura in 827 (100%) HSP, and to a lesser extent in c-PAN 68 (45%) and c-WG 14 (23%); while specified forms of skin involvement (livedo/nodules/superficial or deep skin infarction) were observed mainly in c-PAN 104 (69%). Upper airway/laryngo-tracheo-bronchial involvement was more frequent in c-WG, occurring in 52 (87%), cardiovascular in c-TA 74 (85%) and neurological involvement in c-PAN 39 (26%). Gastrointestinal (mainly abdominal pain) and musculoskeletal complaints were evenly distributed among all four vasculitides.
Baseline signs/symptoms, individually or in combination, for the 1124 patients considered in the final analysis (see also figure 1)
Table 4 shows the absolute values, percentage frequencies and related p values (all <0.0001) of the laboratory, biopsy and imaging data. ANCA positivity was reported mainly in c-WG (any ANCA in 90%). Systemic inflammation, as measured by erythrocyte sedimentation rate or C-reactive protein, was most common in c-PAN, c-WG and c-TA 277 (combined mean 93%) as compared with patients with HSP 430 (52%). Renal involvement (haematuria/proteinuria, serum creatinine elevation, abnormal glomerular filtration rate) was most frequent in c-WG (68%), followed by HSP (33%) and c-PAN (29%). Deposition of IgA was present in 83/89 (93%) biopsies performed in HSP, small-/medium-size necrotising vasculitis 67/79 biopsies (85%) for c-PAN, while granulomatous inflammation lesions were observed in 27/50 (54%) c-WG biopsies. Similarly, abnormal chest x-ray or CT findings (nodules, cavities or fixed infiltrates) were mainly seen in 47/60 (78%) images available in c-WG. Angiography findings demonstrated that aneurysms/stenosis of medium/small arteries were present in 64/79 (81%) images in c-PAN, while in c-TA aneurysms/stenosis of large arteries were present in 87/87 (100%) of the images available.
Baseline laboratory, biopsy and imaging data, individually or in combination, for the 1124 patients considered in the final analysis (see also figure 1)
Discussion
The results of this project allow the classification of children with HSP, c-PAN, c-WG or c-TA, using different combinations of signs/symptoms, laboratory, biopsy and imaging reports. The proposed definitions robustly classify and discriminate patients with the specific type of vasculitis (sensitivity) and exclude a high proportion of patients with other diseases (specificity).
After the ACR publication in 1990 paediatric researchers had to rely on classification criteria developed for adult patients. Because of the peculiarities of childhood vasculitis and its observed differences from adult vasculitis, in 2005 the vasculitis working group of PRES decided to assess the applicability of adult-onset vasculitis classification criteria in children, with a view to determining if new or modified classification criteria should be developed for children. Indeed, it was not the intention of the PRES working group to replace existing criteria for adult patients, but rather to formally evaluate if they apply to children, and to determine if changes in classification were necessary.
The process started with a literature review and a consensus-based preliminary proposal,10 followed by a formal validation process described in this paper. The large-scale data collection has been more successful in recruiting patient numbers than initially expected. Our validation process resembles the methodology followed by the ACR in part, since we collected a large series of paediatric cases to evaluate the sensitivity and specificity of each individual criterion and several different combinations of the most sensitive and specific items. Our process differed from the ACR methodology in that we introduced, as ‘gold standard’ a blinded (to the original diagnosis assigned by the treating physician) consensus classification of the most difficult cases (about 20% of the cases) by a panel of 11 experienced paediatric rheumatologists/nephrologists. The purpose of this blinded classification exercise was to confirm/exclude, with an unbiased revision, the diagnosis made by the patient's physician and to calculate the κ-agreement between the consensus panel and the treating physician. The results of the blinded evaluation showed almost perfect κ-agreement, justifying the combination into one single group of both the patients classified by the consensus panel and the patients diagnosed by the primary treating physician. The almost perfect κ was also an indirect confirmation that the diagnosis made by the treating physicians was consistent with the clinical presentation of each child as abstracted from the patients' charts.
The main advantage of this project is the large dataset sample analysed and the unique worldwide nature of the data collected, thus capturing a broad spectrum of cases and ensuring that the classification criteria can be applied universally. A possible limitation is related to the exclusion from the analysis of about 15% of the children enrolled who were, however, scattered proportionally among all vasculitides. Also, to obtain a more homogeneous sample we excluded the cases of unclassified vasculitides that will be more precisely characterised in future analysis.
In conclusion, EULAR/PRINTO/PRES propose validated definitions for the classification of HSP, c-PAN, c-WG and c-TA, each of which demonstrates high sensitivity/specificity and an almost perfect agreement with the final consensus classification or original treating physician diagnosis.
Acknowledgments
The authors thank the EULAR/PRINTO/PRES for their financial support for the project. The authors thank Professor Ihsan Dogramacı for his support for the study. Drs Silvia M Iusan, MD, Anuela Kondi, MD, Valentin Pop, MD and Iris Vilca, MD were recipients of a EULAR training bursary. The authors thank Dr Laura Carenini and Dr Luca Villa from the PRINTO coordinating centre who managed the data collection.
References
Footnotes
-
Funding EULAR/PRINTO/PRES.
-
Competing interests None.
-
Ethics approval This study was conducted with the approval of all participating centres if required by the national laws of the specific country.
-
Provenance and peer review Not commissioned; externally peer reviewed.
-
Principal investigator Seza Ozen, MD, Ankara, Turkey. Project coordinator: Nicolino Ruperto, MD, MPH, Genoa, Italy. Vienna (Austria) Consensus Conference10 Consensus moderator: Nicolino Ruperto, MD, MPH, Genoa, Italy. Consensus participants: Seza Ozen, MD, Ankara, Turkey, Principal Investigator; Michael J Dillon, MD, London, UK; Arvind Bagga, MD, New Delhi, India; Karyl Barron, MD, Bethesda, MD, USA; Jean Claude Davin, MD, Amsterdam, The Netherlands; Tomisaku Kawasaki, MD, Tokio, Japan; Carol Lindsley, MD, Kansas City, KS, USA; Ross E Petty, MD, Vancouver, BC, Canada; Anne Marie Prieur, MD, Paris, France; Angelo Ravelli, MD, Genoa, Italy; Patricia Woo, MD, London, UK. Ankara Consensus Conference: Consensus moderator: Nicolino Ruperto, MD, MPH, Genoa, Italy. Moderator assistant: Giovanni Filocamo, MD, Genoa, Italy. Consensus bio-statistician: Angela Pistorio, MD, PhD, Genoa, Italy. PRINTO analysis committee: Nicolino Ruperto, MD, MPH, Genoa, Italy; Seza Ozen, MD, Ankara, Turkey, Principal Investigator; Angela Pistorio, MD, PhD, Genoa, Italy; Angelo Ravelli MD, Genoa, Italy; Alberto Martini MD, Genoa, Italy; Silvia M Iusan, MD, Genoa, Italy; Giovanni Filocamo, MD, Genoa, Italy; Claudia Saad Magalhães, Genoa, Italy. Consensus participants: Paul Brogan, MD, London, UK; David A Cabral, MD, Vancouver, BC, Canada; Ruben Cuttica, MD, Buenos Aires, Argentina; Pavla Dolezalova, MD, Prague, Czech Republic; Raju Khubchandani, MD, Mumbai, India; Daniel Lovell, MD, Cincinnati, OH, USA; Alberto Martini, MD, Genoa, Italy; Kathleen M O'Neil, MD, Oklahoma City, OK, USA; Seza Ozen, MD, Ankara, Turkey, Principal Investigator; Pierre Quartier, MD, Paris, France; Angelo Ravelli, MD, Genoa, France. We are also indebted to all PRINTO/PRES members (97 centres in 36 countries) and to the families for their participation in the study: Albania Anuela Kondi, MD; Argentina Maria Teresa Apaz, MD, Esteban Rodolfo Castell, MD, Ruben Cuttica, MD, Graciela Espada, MD, Stella Maris Garay, MD, Ricardo Russo, MD, Iris Vilca, MD; Austria Jurgen Brunner, MD; Brazil Maria Odete Hilario, MD, Sheila Knupp Feitosa de Oliveira, MD, Claudia Saad Magalhães, MD, Clovis Artur Silva, MD, Flavio Sztajnbok, MD; Bulgaria Dimitrina Mihaylova, MD; Chile Benito Gonzalez, MD, Ximena Norambuena, MD; Croatia Lana Tambic-Bukovac, MD; Czech Republic Pavla Dolezalova, MD; Denmark Troels Herlin, MD, Susan Nielsen, MD; El Salvador Mauricio Alegria, MD; Estonia Chris Pruunsild, MD; France Alexandre Belot, MD, Rolando Cimaz, MD, Djamal Djeddi, MD, Michel Fischbach, MD, Erick Grouteau, MD, Isabelle Kone Paut, MD, Pierre Quartier, MD; Georgia Maka Ioseliani, MD, Karaman Pagava, MD; Germany Frank Dressler, MD, Ivan Foeldvari, MD, Dirk Foell, MD, Hermann Girschick, MD, Gerd Horneff, MD, Rolf Keitzer, MD, Rolf-Michael Kuester, MD, Hartmut Michels, MD, Ralf Trauzeddel, MD, Claudia Sengler MD; Greece Florence Kanakoudi-Tsakalidou, Polyxeni Pratsidou, MD, Olga Vougiouka, MD; India Arvind Bagga, MD, Raju Khubchandani, MD; Iraq Saad Alhasani, MD; Israel Judith Barash, MD, Yackov Berkun, MD, Riva Brik, MD, Tsvia Tauber, MD, Yosef Uziel, MD; Italy Maria Giannina Alpigiani, MD, Giancarlo Barbano, MD, Patrizia Barone, MD, Claudia Bracaglia, MD, Luciana Breda, MD, Antonella Buoncompagni, MD, Giuseppina Calcagno, MD, Luca Cantarini, MD, Rolando Cimaz, MD, Elisabetta Cortis, MD, Gianfranco D'Angelo, MD, Sara Garozzo, MD, Loredana Lepore, MD, Silvia Magni Manzoni, MD, Alma Nunzia Olivieri, MD, Salvatore Scarpato, MD, Achille Stabile, Donato Rigante, MD, Francesco Zulian, MD; Latvia Ingrida Rumba, MD, Valda Stanevicha, MD; Libya Halima Ben Amer, MD; Mexico Adriana Cespedes Cruz, MD, Maria del Rocio Maldonado, MD; Netherlands Esther Hoppenreijs, MD, Marion A J van Rossum, MD, Nico Wulffraat, MD; Peru Amparo Ibanez Estrella, MD; Poland Henryka Mazur-Zielinska, MD; Romania Nicolae Iagaru, MD, Valentin Pop, MD, Calin Lazar, MD; Russia Vyacheslav Chasnyk, MD; Saudi Arabia Sulaiman M Al-Mayouf, MD; Serbia Srdjan Pasic, MD, Gordana Susic, MD; Slovakia Richard Vesely, MD; Slovenia Tadej Avcin, MD; Spain Jordi Anton, MD, María Luz Gamir, MD, Julia Garcia Consuegra, MD, Laura Nuno, MD; Sweden Anders Fasth, MD; Turkey Nuray Aktay Ayaz MD, Esra Baskin, MD, Ilmay Bilge, MD, Zelal Bircan, MD, Faysal Gok, MD, Balahan Makay MD, Seza Ozen, MD, Erbil Unsal, MD; UK Paul Brogan, MD, Kjell Tullus, MD; USA Marisa Klein-Gitelman, MD
Linked Articles
- Miscellaneous