Rheumatology Advance Access published September 5, 2014
RHEUMATOLOGY
294
Original article
doi:10.1093/rheumatology/keu352
Utility of the American-European Consensus Group and American College of Rheumatology Classification Criteria for Sjo¨gren’s syndrome in patients with systemic autoimmune diseases in the clinical setting
Objective. The aim of this study was to evaluate the feasibility and performance of the AmericanEuropean Consensus Group (AECG) and ACR Classification Criteria for SS in patients with systemic autoimmune diseases. Methods. Three hundred and fifty patients with primary SS, SLE, RA or scleroderma were randomly selected from our patient registry. Each patient was clinically diagnosed as probable/definitive SS or non-SS following a standardized evaluation including clinical symptoms and manifestations, confirmatory tests, fluorescein staining test, autoantibodies, lip biopsy and medical chart review. Using the clinical diagnosis as the gold standard, the degree of agreement with each criteria set and between the criteria sets was estimated. Results. One hundred fifty-four (44%) patients were diagnosed with SS. The AECG criteria were incomplete in 36 patients (10.3%) and the ACR criteria in 96 (27.4%; P < 0.001). Nevertheless, their ability to classify patients was almost identical, with a sensitivity of 61.6 vs 62.3 and a specificity of 94.3 vs 91.3, respectively. Either set of criteria was met by 123 patients (80%); 95 (61.7%) met the AECG criteria and 96 (62.3%) met the ACR criteria, but only 68 (44.2%) patients met both sets. The concordance rate between clinical diagnosis and AECG or ACR criteria was moderate (k statistic 0.58 and 0.55, respectively). Among 99 patients with definitive SS sensitivity was 83.3 vs 77.7 and specificity was 90.8 vs 85.6, respectively. A discrepancy between clinical diagnosis and criteria was seen in 59 patients (17%). Conclusion. The feasibility of the SS AECG criteria is superior to that of the ACR criteria, however, their performance was similar among patients with systemic autoimmune diseases. A subset of SS patients is still missed by both criteria sets. Key words: Sjo¨gren’s syndrome, classification criteria, connective tissue diseases, epidemiology.
1 Department of Immunology and Rheumatology, Instituto Nacional de Ciencias Me´dicas y Nutricio´n Salvador Zubira´n, Mexico City, Mexico, 2 Department of Pathology, University Health Network, Toronto General Hospital, Toronto, ON, Canada, 3Ophthalmology Service, 4Dental Service, Instituto Nacional de Ciencias Me´dicas y Nutricio´n Salvador Zubira´n, Mexico City, Mexico, and 5Mount Sinai Hospital and University Health Network, University of Toronto, Toronto, ON, Canada.
Submitted 25 March 2014; revised version accepted 3 July 2014. Correspondence to: Jorge Sa´nchez-Guerrero, Mount Sinai Hospital and University Health Network, University of Toronto, 60 Murray Street, 2-005, Toronto, ON, Canada M5T 3L9. E-mail: [email protected] y
Deceased.
! The Author 2014. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: [email protected]
1
CLINICAL SCIENCE
Abstract
Downloaded from http://rheumatology.oxfordjournals.org/ at Ohio State University on December 19, 2014
Gabriela Herna´ndez-Molina1, Carmen Avila-Casado2, Carlos Nun˜ez-Alvarez1, Francisco Ca´rdenas-Vela´zquez3,y , Carlos Herna´ndez-Herna´ndez4, Marı´a Luisa Calderillo4, Vero´nica Marroquı´n4, Claudia Recillas-Gispert3, Juanita Romero-Dı´az1 and Jorge Sa´nchez-Guerrero5
Gabriela Herna´ndez-Molina et al.
Introduction
Patients and methods Patients The study was conducted in a tertiary care centre where the rheumatology clinic provides regular care to 5942 patients, of whom 4813 (81%) have systemic autoimmune diseases. Between February 2006 and July 2007, 100 of 2527 patients with RA, 100 of 1860 patients with SLE and 100 of 136 patients with scleroderma according to the classification criteria and 50 of 81 patients with a clinical diagnosis of primary SS were selected using random numbers from our patient registry and assessed for SS using a structured approach [1921]. We pursued the completeness of the assessment according to the regular clinical approach, considering the risk and benefits of each test and the preference of patients. Patients were excluded if they were taking any medication that would reduce salivary flow (i.e. antihistamines, sedatives, beta-blockers, diuretics, etc.) within 48 h
2
Assessment of SS Participating patients were subjected to a standardized evaluation designed in three phases: screening, confirmatory phase and lip biopsy. In the evaluation phase, patients were assessed by a structured team composed of rheumatologists, ophthalmologists and dentists with experience in systemic autoimmune diseases including SS. Participants were asked to refrain from eating, drinking, smoking, chewing or oral hygiene procedures for at least 1 h before the evaluation and were seen in a closed room with no air conditioning or heating during the morning. The presence of extraglandular manifestations was assessed from each patient’s medical records and by direct questions during the interview.
Screening phase All patients had a face-to-face interview with a single rheumatologist using a standardized form that included questions about demographic data and the use of medications. In addition, a validated six-item screening questionnaire for oral and ocular sicca symptoms was applied and a Schirmer I test and wafer test were carried out [14, 16, 22]. Also a blood sample was drawn for autoantibody testing. The Schirmer I test was done using two standardized sterile tear measurement strips (Tear Flo; Rose Stone Enterprises, Alta Loma, CA, USA) and the wafer test was done as described [22]. Testing for autoantibodies included anti-Ro/La (IgG isotype) by ELISA (Binding Site, Birmingham, UK). For the current analysis, in November 2013, RF (RF, IgG isotype) and ANA were assessed in all patients using stored serum samples obtained during the recruitment period. RF was measured by a commercially available ELISA method (Orgentec, Mainz, Germany) and ANA by indirect immunofluorescence using HEp-2 cells IgG isotype (Inova Diagnostics, San Diego, CA, USA). Patients with at least one affirmative response to the screening questionnaire, Schirmer I test 45 mm in 5 min or wafer test >4 min were considered to have a positive screening.
Confirmatory phase Patients with positive screening underwent this phase, consisting of a fluorescein staining test and non-stimulated whole salivary flow rate (NSWSF). The fluorescein staining test was performed by two ophthalmologists blinded to the patient’s diagnosis. The test was considered positive with a score 54 according to the Bijsterveld scale in at least one eye [16]. For purposes of the current analysis, the ACR criteria ocular staining score (OSS) 53 was substituted by a van Bijsterveld score 53 [23].
www.rheumatology.oxfordjournals.org
Downloaded from http://rheumatology.oxfordjournals.org/ at Ohio State University on December 19, 2014
SS is a systemic autoimmune disease characterized by immune-mediated damage of exocrine glands, predominantly salivary and lacrimal, with subsequent development of keratoconjunctivitis and xerostomia. Symptoms usually develop insidiously and most patients have mild disease, therefore the treatment is oriented to minimize discomfort by using substitutes for the dysfunctional exocrine secretions [1, 2]. However, in specialized centres, 5090% of patients may develop extraglandular manifestations and require intensive treatment [3, 4]. Despite being the second most prevalent systemic autoimmune disease and overlapping in 1530% of patients with other autoimmune diseases, the diagnosis of SS may be difficult to establish because there is no gold standard test [5]. In the clinical setting, the diagnosis of SS relies on clinical procedures, including integration of patient history, test results and the clinical opinion of a multidisciplinary team. For research purposes, several sets of classification criteria have been proposed [616]. During the past decade the 2002 revised AmericanEuropean Consensus Group (AECG) classification criteria have been the paradigm in clinical and research studies of SS [16]. Recently a new set of criteria endorsed by the ACR has been proposed that are based on objective tests and aim to reduce misclassification of patients in situations where high specificity is required [17]. Two reports from well-characterized multicentre sicca cohorts describe concordant results between the two sets [17, 18]. Whether these results are generalized to patients with systemic autoimmune diseases followed in clinical settings still needs to be determined, since these diseases were not included in the original study [17]. We aimed to evaluate the feasibility and performance of the AECG and ACR classification criteria for SS in a random sample of patients with systemic autoimmune diseases.
before the study or had a history of hepatitis C or HIV infection, sarcoidosis, IgG4-related disease, lymphoma, graft vs host disease or neck/head radiotherapy. The study was approved by the Institutional Biomedical Research Board of the Instituto Nacional de Ciencias Me´dicas y Nutricio´n SZ and all patients gave signed informed consent to participate according to the Declaration of Helsinki.
SS classification in systemic autoimmune diseases
NSWSF collection was performed as reported [24]. Saliva was collected during 5 min and the volume was expressed as millilitres per 5 min. The confirmatory phase was also performed in a random sample of 18 patients with negative screening. The assessment protocol did not include scintigraphy or sialography, but several patients had these studies done as part of the regular clinical evaluation.
Lip biopsy
Diagnosis of SS Diagnosis of SS was made on a clinical basis by two rheumatologists. For this purpose, each of them independently evaluated every patient considering the results of the six-item screening questionnaire, history of parotiditis, Schirmer I test, wafer test, NSWSF rate, fluorescein staining test, autoantibodies, lip biopsy, medical notes from ophthalmology and oral medicine and a medical chart review. Each patient was diagnosed as SS (definitive or probable) or non-SS. The agreement from the independent evaluation by the two rheumatologists was 92.2% for definitive SS and 79.1% for probable SS. Patients in whom discrepancy existed underwent further review and discussion between the two rheumatologists and a final diagnosis was reached.
Results Three hundred and fifty patients participated in the study; all of them had the screening questionnaire for oral/ocular symptoms, Schirmer I test and wafer test and were tested for ANA, RF and anti-SSA/B antibodies. The flow of patients through the three study phases is shown in Fig. 1. Three hundred patients (85.7%) were positive for the screening phase and advanced to the confirmatory phase; the NSWSF was measured in all the patients and the fluorescein staining test was done in 292 (97.3%). Two hundred and fifty-one patients met the criteria for lip biopsy; it was performed in 107 patients (42.6%), 128 (50.9%) declined it and 16 (6.3%) had a contraindication (anticoagulation therapy or severe thrombocytopenia). None of the patients with negative screening in whom the confirmatory phase was done met the criteria for lip biopsy.
Diagnosis of SS Clinical diagnosis of SS was made by the study investigators in 154 patients (44%) (definitive SS in 99, probable SS in 55). Primary SS was seen in 47 patients and associated with another systemic autoimmune disease in 107. The demographic, clinical, serological and histopathological data of the patients with and without SS are presented in Table 1.
Classification of patients according to the AECG and ACR criteria In order to evaluate the feasibility of each set of criteria, we estimated the proportion of patients classified by each of them. Patients were classified as SS if they fulfilled the criteria and as non-SS when the criteria were not met independently of having the criteria set fully appraised.
FIG. 1 Flowchart of patients through the study phases
Statistical analysis The clinical diagnosis of SS was the gold standard. The AECG and ACR criteria were applied to each study participant, including those with incomplete features for any set. The ACR criteria were defined using the results of anti-SSA, anti-SSB, ANA 51:320 plus RF, lip biopsy and substituting the fluorescein staining test for the sicca OSS because this method is not implemented in our institution [23]. We estimated the sensitivity, specificity, positive predictive value and negative predictive value with 95% CIs, as well as the accuracy for each criteria set. The kappa statistic was used to quantify the degree of agreement between the clinical diagnosis and each criteria set and between the two sets of criteria. Categorical variables were compared by chi-square or Fisher’s exact test when appropriate; continuous variables were compared using Student’s t test or the MannWhitney U test. A two-tailed P-value 4 min, presence of keratitis by the fluorescein staining test, NSWSF 50 lymphocytes/4 mm2) [16].
were performed using SPSS for Windows 20.0 (IBM, Armonk, NY, USA).
Gabriela Herna´ndez-Molina et al.
TABLE 1 Clinical and serological characteristics of patients with and without a clinical diagnosis of SS SS (n = 154)
No SS (n = 196)
P-value
Age, mean (S.D.), years Females, n (%) Oral symptoms, n (%) Ocular symptoms, n (%) Parotid enlargement, n (%) Schirmer I test, n (%) Wafer test, n (%) Keratoconjuctivitis sicca, n/N (%) NSWSF
RHEUMATOLOGY
294
Original article
doi:10.1093/rheumatology/keu352
Utility of the American-European Consensus Group and American College of Rheumatology Classification Criteria for Sjo¨gren’s syndrome in patients with systemic autoimmune diseases in the clinical setting
Objective. The aim of this study was to evaluate the feasibility and performance of the AmericanEuropean Consensus Group (AECG) and ACR Classification Criteria for SS in patients with systemic autoimmune diseases. Methods. Three hundred and fifty patients with primary SS, SLE, RA or scleroderma were randomly selected from our patient registry. Each patient was clinically diagnosed as probable/definitive SS or non-SS following a standardized evaluation including clinical symptoms and manifestations, confirmatory tests, fluorescein staining test, autoantibodies, lip biopsy and medical chart review. Using the clinical diagnosis as the gold standard, the degree of agreement with each criteria set and between the criteria sets was estimated. Results. One hundred fifty-four (44%) patients were diagnosed with SS. The AECG criteria were incomplete in 36 patients (10.3%) and the ACR criteria in 96 (27.4%; P < 0.001). Nevertheless, their ability to classify patients was almost identical, with a sensitivity of 61.6 vs 62.3 and a specificity of 94.3 vs 91.3, respectively. Either set of criteria was met by 123 patients (80%); 95 (61.7%) met the AECG criteria and 96 (62.3%) met the ACR criteria, but only 68 (44.2%) patients met both sets. The concordance rate between clinical diagnosis and AECG or ACR criteria was moderate (k statistic 0.58 and 0.55, respectively). Among 99 patients with definitive SS sensitivity was 83.3 vs 77.7 and specificity was 90.8 vs 85.6, respectively. A discrepancy between clinical diagnosis and criteria was seen in 59 patients (17%). Conclusion. The feasibility of the SS AECG criteria is superior to that of the ACR criteria, however, their performance was similar among patients with systemic autoimmune diseases. A subset of SS patients is still missed by both criteria sets. Key words: Sjo¨gren’s syndrome, classification criteria, connective tissue diseases, epidemiology.
1 Department of Immunology and Rheumatology, Instituto Nacional de Ciencias Me´dicas y Nutricio´n Salvador Zubira´n, Mexico City, Mexico, 2 Department of Pathology, University Health Network, Toronto General Hospital, Toronto, ON, Canada, 3Ophthalmology Service, 4Dental Service, Instituto Nacional de Ciencias Me´dicas y Nutricio´n Salvador Zubira´n, Mexico City, Mexico, and 5Mount Sinai Hospital and University Health Network, University of Toronto, Toronto, ON, Canada.
Submitted 25 March 2014; revised version accepted 3 July 2014. Correspondence to: Jorge Sa´nchez-Guerrero, Mount Sinai Hospital and University Health Network, University of Toronto, 60 Murray Street, 2-005, Toronto, ON, Canada M5T 3L9. E-mail: [email protected] y
Deceased.
! The Author 2014. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: [email protected]
1
CLINICAL SCIENCE
Abstract
Downloaded from http://rheumatology.oxfordjournals.org/ at Ohio State University on December 19, 2014
Gabriela Herna´ndez-Molina1, Carmen Avila-Casado2, Carlos Nun˜ez-Alvarez1, Francisco Ca´rdenas-Vela´zquez3,y , Carlos Herna´ndez-Herna´ndez4, Marı´a Luisa Calderillo4, Vero´nica Marroquı´n4, Claudia Recillas-Gispert3, Juanita Romero-Dı´az1 and Jorge Sa´nchez-Guerrero5
Gabriela Herna´ndez-Molina et al.
Introduction
Patients and methods Patients The study was conducted in a tertiary care centre where the rheumatology clinic provides regular care to 5942 patients, of whom 4813 (81%) have systemic autoimmune diseases. Between February 2006 and July 2007, 100 of 2527 patients with RA, 100 of 1860 patients with SLE and 100 of 136 patients with scleroderma according to the classification criteria and 50 of 81 patients with a clinical diagnosis of primary SS were selected using random numbers from our patient registry and assessed for SS using a structured approach [1921]. We pursued the completeness of the assessment according to the regular clinical approach, considering the risk and benefits of each test and the preference of patients. Patients were excluded if they were taking any medication that would reduce salivary flow (i.e. antihistamines, sedatives, beta-blockers, diuretics, etc.) within 48 h
2
Assessment of SS Participating patients were subjected to a standardized evaluation designed in three phases: screening, confirmatory phase and lip biopsy. In the evaluation phase, patients were assessed by a structured team composed of rheumatologists, ophthalmologists and dentists with experience in systemic autoimmune diseases including SS. Participants were asked to refrain from eating, drinking, smoking, chewing or oral hygiene procedures for at least 1 h before the evaluation and were seen in a closed room with no air conditioning or heating during the morning. The presence of extraglandular manifestations was assessed from each patient’s medical records and by direct questions during the interview.
Screening phase All patients had a face-to-face interview with a single rheumatologist using a standardized form that included questions about demographic data and the use of medications. In addition, a validated six-item screening questionnaire for oral and ocular sicca symptoms was applied and a Schirmer I test and wafer test were carried out [14, 16, 22]. Also a blood sample was drawn for autoantibody testing. The Schirmer I test was done using two standardized sterile tear measurement strips (Tear Flo; Rose Stone Enterprises, Alta Loma, CA, USA) and the wafer test was done as described [22]. Testing for autoantibodies included anti-Ro/La (IgG isotype) by ELISA (Binding Site, Birmingham, UK). For the current analysis, in November 2013, RF (RF, IgG isotype) and ANA were assessed in all patients using stored serum samples obtained during the recruitment period. RF was measured by a commercially available ELISA method (Orgentec, Mainz, Germany) and ANA by indirect immunofluorescence using HEp-2 cells IgG isotype (Inova Diagnostics, San Diego, CA, USA). Patients with at least one affirmative response to the screening questionnaire, Schirmer I test 45 mm in 5 min or wafer test >4 min were considered to have a positive screening.
Confirmatory phase Patients with positive screening underwent this phase, consisting of a fluorescein staining test and non-stimulated whole salivary flow rate (NSWSF). The fluorescein staining test was performed by two ophthalmologists blinded to the patient’s diagnosis. The test was considered positive with a score 54 according to the Bijsterveld scale in at least one eye [16]. For purposes of the current analysis, the ACR criteria ocular staining score (OSS) 53 was substituted by a van Bijsterveld score 53 [23].
www.rheumatology.oxfordjournals.org
Downloaded from http://rheumatology.oxfordjournals.org/ at Ohio State University on December 19, 2014
SS is a systemic autoimmune disease characterized by immune-mediated damage of exocrine glands, predominantly salivary and lacrimal, with subsequent development of keratoconjunctivitis and xerostomia. Symptoms usually develop insidiously and most patients have mild disease, therefore the treatment is oriented to minimize discomfort by using substitutes for the dysfunctional exocrine secretions [1, 2]. However, in specialized centres, 5090% of patients may develop extraglandular manifestations and require intensive treatment [3, 4]. Despite being the second most prevalent systemic autoimmune disease and overlapping in 1530% of patients with other autoimmune diseases, the diagnosis of SS may be difficult to establish because there is no gold standard test [5]. In the clinical setting, the diagnosis of SS relies on clinical procedures, including integration of patient history, test results and the clinical opinion of a multidisciplinary team. For research purposes, several sets of classification criteria have been proposed [616]. During the past decade the 2002 revised AmericanEuropean Consensus Group (AECG) classification criteria have been the paradigm in clinical and research studies of SS [16]. Recently a new set of criteria endorsed by the ACR has been proposed that are based on objective tests and aim to reduce misclassification of patients in situations where high specificity is required [17]. Two reports from well-characterized multicentre sicca cohorts describe concordant results between the two sets [17, 18]. Whether these results are generalized to patients with systemic autoimmune diseases followed in clinical settings still needs to be determined, since these diseases were not included in the original study [17]. We aimed to evaluate the feasibility and performance of the AECG and ACR classification criteria for SS in a random sample of patients with systemic autoimmune diseases.
before the study or had a history of hepatitis C or HIV infection, sarcoidosis, IgG4-related disease, lymphoma, graft vs host disease or neck/head radiotherapy. The study was approved by the Institutional Biomedical Research Board of the Instituto Nacional de Ciencias Me´dicas y Nutricio´n SZ and all patients gave signed informed consent to participate according to the Declaration of Helsinki.
SS classification in systemic autoimmune diseases
NSWSF collection was performed as reported [24]. Saliva was collected during 5 min and the volume was expressed as millilitres per 5 min. The confirmatory phase was also performed in a random sample of 18 patients with negative screening. The assessment protocol did not include scintigraphy or sialography, but several patients had these studies done as part of the regular clinical evaluation.
Lip biopsy
Diagnosis of SS Diagnosis of SS was made on a clinical basis by two rheumatologists. For this purpose, each of them independently evaluated every patient considering the results of the six-item screening questionnaire, history of parotiditis, Schirmer I test, wafer test, NSWSF rate, fluorescein staining test, autoantibodies, lip biopsy, medical notes from ophthalmology and oral medicine and a medical chart review. Each patient was diagnosed as SS (definitive or probable) or non-SS. The agreement from the independent evaluation by the two rheumatologists was 92.2% for definitive SS and 79.1% for probable SS. Patients in whom discrepancy existed underwent further review and discussion between the two rheumatologists and a final diagnosis was reached.
Results Three hundred and fifty patients participated in the study; all of them had the screening questionnaire for oral/ocular symptoms, Schirmer I test and wafer test and were tested for ANA, RF and anti-SSA/B antibodies. The flow of patients through the three study phases is shown in Fig. 1. Three hundred patients (85.7%) were positive for the screening phase and advanced to the confirmatory phase; the NSWSF was measured in all the patients and the fluorescein staining test was done in 292 (97.3%). Two hundred and fifty-one patients met the criteria for lip biopsy; it was performed in 107 patients (42.6%), 128 (50.9%) declined it and 16 (6.3%) had a contraindication (anticoagulation therapy or severe thrombocytopenia). None of the patients with negative screening in whom the confirmatory phase was done met the criteria for lip biopsy.
Diagnosis of SS Clinical diagnosis of SS was made by the study investigators in 154 patients (44%) (definitive SS in 99, probable SS in 55). Primary SS was seen in 47 patients and associated with another systemic autoimmune disease in 107. The demographic, clinical, serological and histopathological data of the patients with and without SS are presented in Table 1.
Classification of patients according to the AECG and ACR criteria In order to evaluate the feasibility of each set of criteria, we estimated the proportion of patients classified by each of them. Patients were classified as SS if they fulfilled the criteria and as non-SS when the criteria were not met independently of having the criteria set fully appraised.
FIG. 1 Flowchart of patients through the study phases
Statistical analysis The clinical diagnosis of SS was the gold standard. The AECG and ACR criteria were applied to each study participant, including those with incomplete features for any set. The ACR criteria were defined using the results of anti-SSA, anti-SSB, ANA 51:320 plus RF, lip biopsy and substituting the fluorescein staining test for the sicca OSS because this method is not implemented in our institution [23]. We estimated the sensitivity, specificity, positive predictive value and negative predictive value with 95% CIs, as well as the accuracy for each criteria set. The kappa statistic was used to quantify the degree of agreement between the clinical diagnosis and each criteria set and between the two sets of criteria. Categorical variables were compared by chi-square or Fisher’s exact test when appropriate; continuous variables were compared using Student’s t test or the MannWhitney U test. A two-tailed P-value 4 min, presence of keratitis by the fluorescein staining test, NSWSF 50 lymphocytes/4 mm2) [16].
were performed using SPSS for Windows 20.0 (IBM, Armonk, NY, USA).
Gabriela Herna´ndez-Molina et al.
TABLE 1 Clinical and serological characteristics of patients with and without a clinical diagnosis of SS SS (n = 154)
No SS (n = 196)
P-value
Age, mean (S.D.), years Females, n (%) Oral symptoms, n (%) Ocular symptoms, n (%) Parotid enlargement, n (%) Schirmer I test, n (%) Wafer test, n (%) Keratoconjuctivitis sicca, n/N (%) NSWSF
