Rheumatology Advance Access published April 4, 2014
RHEUMATOLOGY
Concise report
273, 294
doi:10.1093/rheumatology/keu037
Salivary gland ultrasonography improves the diagnostic performance of the 2012 American College of Rheumatology classification criteria for Sjo¨gren’s syndrome
Objective. The aim of this study was to evaluate whether salivary gland ultrasonography (SGUS) improves the diagnostic performance of the 2012 ACR classification criteria for SS. Methods. We studied a cohort of 101 patients with suspected SS seen at a single centre in Brittany, France. An SGUS echostructure score 52 was considered abnormal. The reference standard was a clinical diagnosis of SS made by a group of experts blinded to SGUS findings. Results. SS was diagnosed in 45 patients. Similar proportions of patients with and without SS had an ocular staining score 53. Adding RF positivity and ANA titre 51:320 as an alternative to anti-SSA/SSB positivity increased the sensitivity of the serological item without modifying specificity compared with using anti-SSA/SSB alone. SGUS was 60.0% sensitive and 87.5% specific for SS. Adding the SGUS score to the ACR criteria increased sensitivity from 64.4% to 84.4% and only slightly decreased specificity, from 91.1% to 89.3%. Conclusion. The diagnostic performance of the ACR classification criteria for SS is notably improved by adding the SGUS score. SGUS should be included in future classification criteria for SS. Key words: Sjo¨gren’s syndrome, ultrasonography, classification criteria, diagnosis, sensitivity, specificity.
Introduction In 2012 the ACR issued new classification criteria for SS developed in >1500 patients of the Sjo¨gren’s International Collaborative Clinical Alliance (SICCA) cohort [1]. The new criteria were designed as a diagnostic tool for patients referred to specialists because of signs or symptoms suggesting SS. They were selected based on expert opinion
1
Department of Rheumatology, 2EA 2216 Immunology and Pathology, SFR ScinBios, Labex Immunotherapy, Graft, Oncology, Brest University, 3Department of Odontology and 4Laboratory of Immunology and Immunotherapy, Brest Teaching Hospital, Brest, France. Submitted 23 September 2013; revised version accepted 28 January 2014. Correspondence to: Vale´rie Devauchelle-Pensec, Service de Rhumatologie, Hoˆpital de la Cavale Blanche, BP 824, F 29609 Brest cedex, France. E-mail: [email protected]
in order to cover the oral, ocular and systemic features of SS. Experts in rheumatology, ophthalmology and oral medicine chose what they felt was the most important criterion for diagnosing SS in their specialty. The new set is composed of only objective criteria. A diagnosis of SS requires at least two of the following three criteria: (i) either anti-SSA/SSB antibodies or a combination of ANA titre 51:320 and RF positivity; (ii) a combined fluorescein and lissamine green ocular staining score (OSS) 53, indicating keratoconjunctivitis sicca [2]; and (iii) a minor salivary gland biopsy showing lymphocytic sialadenitis with a focus score 51 focus/4 mm2. Exclusion criteria are a history of head-and-neck radiotherapy, hepatitis C infection, AIDS, sarcoidosis, amyloidosis, graft-vs-host disease and IgG4-related disease. These new criteria were validated within the SICCA cohort using a statistical method known as latent class analysis. However, the diagnostic performance of
! 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 Syracuse University Library on April 6, 2014
Divi Cornec1,2, Sandrine Jousse-Joulin1,2, Thierry Marhadour1, Jacques-Olivier Pers2,3, Sylvie Boisrame´-Gastrin3, Yves Renaudineau2,4, Alain Saraux1,2 and Vale´rie Devauchelle-Pensec1,2
Divi Cornec et al.
served as the reference standard. The diagnoses of other systemic diseases were based on published classification criteria.
Salivary gland ultrasonography All ultrasound scans were performed by the same examiner (S.J.-J.), who was blinded to the clinical data and used an iU22 scanner (Philips Medical Systems, Andover, MA, USA). The parotid and submandibular glands were scanned on both sides. SGUS was performed at the same time as the other tests. The echostructure of each gland was graded on a scale of 04, as previously described [3]. In each patient, the highest SGUS score of the four glands was considered for the analysis and scores 52 were considered abnormal. This definition of abnormal SGUS was chosen after receiver operating characteristic (ROC) curve analysis in our previous report [3].
Statistical analysis
Patients and methods Study population We studied the 101 consecutive patients included between November 2006 and March 2013 in the singlecentre Brittany cohort of patients with suspected SS who had all tests available [3]. Inclusion criteria were subjective ocular or oral dryness, recurrent or bilateral parotidomegaly and/or extraglandular symptoms or laboratory abnormalities suggesting SS. Written informed consent was obtained from all participants before inclusion according to the Declaration of Helsinki and the study was approved by the ethics committee of Brest University Hospital.
Clinical evaluation and laboratory tests done prospectively in the cohort Data were collected using a standardized form. All patients were examined by an ophthalmologist who had experience with dry-eye diseases and who performed both fluorescein and lissamine green stains to assess keratoconjunctivitis sicca. Therefore we were able to determinate retrospectively the OSS. ANA, anti-SSA and anti-SSB antibodies and RF were assessed using ELISA. Minor labial salivary gland biopsy was performed in all patients and graded according to the semi-quantitative Chisholm and Mason score [5]; grades 3 and 4 indicating focus scores 51 focus/4 mm2 met the 2012 ACR criteria requirement [6].
Diagnostic criteria For each patient, a single rheumatologist (V.D.P., 15 years experience of SS) determined whether the symptoms were due to SS, based on the medical history, physical examination, laboratory tests and salivary gland biopsy results; the rheumatologist was not aware of the SGUS findings. All cases were reviewed by a panel of three experts (A.S., S.J.-J., V.D.-P.) blinded to the SGUS findings. The diagnosis by the rheumatologist and the panel review
2
Statistical tests were performed using the Statistical Package for the Social Sciences version 18.0 (SPSS, Chicago, IL, USA). Quantitative variables were described as mean (S.D.) and qualitative variables as number (%). To compare patients with and without SS, we used the MannWhitney test, Fisher’s exact test or chi-square test, as appropriate. ROC curve analysis was performed to determine the optimal cut-off producing the best combination of sensitivity and specificity for each criteria set. On a ROC curve, this optimal cut-off value is the point nearest the upper left end of the curve.
Results Table 1 lists the main patient characteristics. Among the 101 patients, 45 received a diagnosis of SS by the rheumatologist. Diagnoses in the remaining 56 patients were idiopathic sicca syndrome (n = 37), other CTDs (n = 11) and drug-induced sicca syndrome (n = 8). No significant differences were found between the groups with and without SS for age, symptom duration, sex ratio or frequency of sicca symptoms (ocular and oral dryness). The proportion of patients whose OSS was 53 was not significantly different between the groups with and without SS. The sensitivity of this criterion for diagnosing SS was only 55.6% (95% CI 41.1, 70.1) and specificity was only 58.9% (95% CI 46.0, 71.8). Adding RF positivity and ANA titre 51:320 as a possible alternative to anti-SSA/SSB positivity increased the sensitivity of the serological criterion from 53.3% (95% CI 38.7, 67.9) to 60.0% (95% CI 45.7, 74.3) without modifying specificity [96.4% (95% CI 91.5, 100.0); for both serological criteria]. The focus score displayed a sensitivity of 82.2% (95% CI 71.0, 93.4) and a specificity of 82.1% (95% CI 72.1, 92.1). The SGUS score was 52 in 34 of 101 (33.7%) patients. The sensitivity of an SGUS score 52 for the diagnosis of SS was 60.0% (95% CI 45.7, 74.3) and specificity was 87.5% (95% CI 78.8, 96.2). When we added the SGUS score to the three criteria used in the 2012 ACR criteria,
www.rheumatology.oxfordjournals.org
Downloaded from http://rheumatology.oxfordjournals.org/ at Syracuse University Library on April 6, 2014
classification or diagnostic criteria must be evaluated in independent cohorts to assess suitability for use in clinical practice. The Brittany cohort of patients with suspected SS is appropriate for such an evaluation [3]. None of the three ACR criteria reflects salivary gland function and morphology, which are altered in SS. In the 2002 American-European Consensus Group (AECG) classification criteria [4], salivary gland involvement is assessed using salivary flow measurement, salivary gland scintigraphy or parotid sialography. However, all three tests have numerous limitations. In a recent study, we showed that major salivary gland ultrasonography (SGUS) was useful for diagnosing SS and improved the diagnostic performance of the AECG criteria [3]. The objectives of this work were to evaluate the diagnostic performance of the 2012 ACR criteria in an independent cohort of patients with suspected SS and to determine whether this performance was improved by adding SGUS.
Adjunction of SGUS to ACR 2012 criteria for SS
TABLE 1 Comparison of patients with and without SS Overall (n = 101) Age, mean (S.D.), years Symptom duration, mean (S.D.), years Female, n (%) Xerophthalmia, n (%)a Xerostomia, n (%)a OSS 53, n (%) Focus score 51, n (%) Anti-SSA or -SSB positivity, n (%) Anti-SSA/SSB or RF and ANA 51:320, n (%) SGUS score 52, n (%) ACR criteria 52, n (%)
57.4 6.7 95 94 96 48 47 26 29 34 34
(13.5) (6.1) (94.1) (93.1) (95.1) (47.5) (46.5) (25.7) (28.7) (33.7) (33.7)
SS (n = 45) 59.4 7.4 42 43 44 25 37 24 27 27 29
(11.2) (6.7) (93.3) (95.6) (97.8) (55.6) (82.2) (53.3) (60.0) (60.0) (64.4)
No SS (n = 56) 55.7 6.2 53 51 52 23 10 2 2 7 5
(15.1) (5.6) (94.6) (91.1) (92.9) (41.1) (17.9) (3.6) (3.6) (12.5) (8.9)
P-value 0.18 0.31 0.78 0.38 0.26 0.15
RHEUMATOLOGY
Concise report
273, 294
doi:10.1093/rheumatology/keu037
Salivary gland ultrasonography improves the diagnostic performance of the 2012 American College of Rheumatology classification criteria for Sjo¨gren’s syndrome
Objective. The aim of this study was to evaluate whether salivary gland ultrasonography (SGUS) improves the diagnostic performance of the 2012 ACR classification criteria for SS. Methods. We studied a cohort of 101 patients with suspected SS seen at a single centre in Brittany, France. An SGUS echostructure score 52 was considered abnormal. The reference standard was a clinical diagnosis of SS made by a group of experts blinded to SGUS findings. Results. SS was diagnosed in 45 patients. Similar proportions of patients with and without SS had an ocular staining score 53. Adding RF positivity and ANA titre 51:320 as an alternative to anti-SSA/SSB positivity increased the sensitivity of the serological item without modifying specificity compared with using anti-SSA/SSB alone. SGUS was 60.0% sensitive and 87.5% specific for SS. Adding the SGUS score to the ACR criteria increased sensitivity from 64.4% to 84.4% and only slightly decreased specificity, from 91.1% to 89.3%. Conclusion. The diagnostic performance of the ACR classification criteria for SS is notably improved by adding the SGUS score. SGUS should be included in future classification criteria for SS. Key words: Sjo¨gren’s syndrome, ultrasonography, classification criteria, diagnosis, sensitivity, specificity.
Introduction In 2012 the ACR issued new classification criteria for SS developed in >1500 patients of the Sjo¨gren’s International Collaborative Clinical Alliance (SICCA) cohort [1]. The new criteria were designed as a diagnostic tool for patients referred to specialists because of signs or symptoms suggesting SS. They were selected based on expert opinion
1
Department of Rheumatology, 2EA 2216 Immunology and Pathology, SFR ScinBios, Labex Immunotherapy, Graft, Oncology, Brest University, 3Department of Odontology and 4Laboratory of Immunology and Immunotherapy, Brest Teaching Hospital, Brest, France. Submitted 23 September 2013; revised version accepted 28 January 2014. Correspondence to: Vale´rie Devauchelle-Pensec, Service de Rhumatologie, Hoˆpital de la Cavale Blanche, BP 824, F 29609 Brest cedex, France. E-mail: [email protected]
in order to cover the oral, ocular and systemic features of SS. Experts in rheumatology, ophthalmology and oral medicine chose what they felt was the most important criterion for diagnosing SS in their specialty. The new set is composed of only objective criteria. A diagnosis of SS requires at least two of the following three criteria: (i) either anti-SSA/SSB antibodies or a combination of ANA titre 51:320 and RF positivity; (ii) a combined fluorescein and lissamine green ocular staining score (OSS) 53, indicating keratoconjunctivitis sicca [2]; and (iii) a minor salivary gland biopsy showing lymphocytic sialadenitis with a focus score 51 focus/4 mm2. Exclusion criteria are a history of head-and-neck radiotherapy, hepatitis C infection, AIDS, sarcoidosis, amyloidosis, graft-vs-host disease and IgG4-related disease. These new criteria were validated within the SICCA cohort using a statistical method known as latent class analysis. However, the diagnostic performance of
! 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 Syracuse University Library on April 6, 2014
Divi Cornec1,2, Sandrine Jousse-Joulin1,2, Thierry Marhadour1, Jacques-Olivier Pers2,3, Sylvie Boisrame´-Gastrin3, Yves Renaudineau2,4, Alain Saraux1,2 and Vale´rie Devauchelle-Pensec1,2
Divi Cornec et al.
served as the reference standard. The diagnoses of other systemic diseases were based on published classification criteria.
Salivary gland ultrasonography All ultrasound scans were performed by the same examiner (S.J.-J.), who was blinded to the clinical data and used an iU22 scanner (Philips Medical Systems, Andover, MA, USA). The parotid and submandibular glands were scanned on both sides. SGUS was performed at the same time as the other tests. The echostructure of each gland was graded on a scale of 04, as previously described [3]. In each patient, the highest SGUS score of the four glands was considered for the analysis and scores 52 were considered abnormal. This definition of abnormal SGUS was chosen after receiver operating characteristic (ROC) curve analysis in our previous report [3].
Statistical analysis
Patients and methods Study population We studied the 101 consecutive patients included between November 2006 and March 2013 in the singlecentre Brittany cohort of patients with suspected SS who had all tests available [3]. Inclusion criteria were subjective ocular or oral dryness, recurrent or bilateral parotidomegaly and/or extraglandular symptoms or laboratory abnormalities suggesting SS. Written informed consent was obtained from all participants before inclusion according to the Declaration of Helsinki and the study was approved by the ethics committee of Brest University Hospital.
Clinical evaluation and laboratory tests done prospectively in the cohort Data were collected using a standardized form. All patients were examined by an ophthalmologist who had experience with dry-eye diseases and who performed both fluorescein and lissamine green stains to assess keratoconjunctivitis sicca. Therefore we were able to determinate retrospectively the OSS. ANA, anti-SSA and anti-SSB antibodies and RF were assessed using ELISA. Minor labial salivary gland biopsy was performed in all patients and graded according to the semi-quantitative Chisholm and Mason score [5]; grades 3 and 4 indicating focus scores 51 focus/4 mm2 met the 2012 ACR criteria requirement [6].
Diagnostic criteria For each patient, a single rheumatologist (V.D.P., 15 years experience of SS) determined whether the symptoms were due to SS, based on the medical history, physical examination, laboratory tests and salivary gland biopsy results; the rheumatologist was not aware of the SGUS findings. All cases were reviewed by a panel of three experts (A.S., S.J.-J., V.D.-P.) blinded to the SGUS findings. The diagnosis by the rheumatologist and the panel review
2
Statistical tests were performed using the Statistical Package for the Social Sciences version 18.0 (SPSS, Chicago, IL, USA). Quantitative variables were described as mean (S.D.) and qualitative variables as number (%). To compare patients with and without SS, we used the MannWhitney test, Fisher’s exact test or chi-square test, as appropriate. ROC curve analysis was performed to determine the optimal cut-off producing the best combination of sensitivity and specificity for each criteria set. On a ROC curve, this optimal cut-off value is the point nearest the upper left end of the curve.
Results Table 1 lists the main patient characteristics. Among the 101 patients, 45 received a diagnosis of SS by the rheumatologist. Diagnoses in the remaining 56 patients were idiopathic sicca syndrome (n = 37), other CTDs (n = 11) and drug-induced sicca syndrome (n = 8). No significant differences were found between the groups with and without SS for age, symptom duration, sex ratio or frequency of sicca symptoms (ocular and oral dryness). The proportion of patients whose OSS was 53 was not significantly different between the groups with and without SS. The sensitivity of this criterion for diagnosing SS was only 55.6% (95% CI 41.1, 70.1) and specificity was only 58.9% (95% CI 46.0, 71.8). Adding RF positivity and ANA titre 51:320 as a possible alternative to anti-SSA/SSB positivity increased the sensitivity of the serological criterion from 53.3% (95% CI 38.7, 67.9) to 60.0% (95% CI 45.7, 74.3) without modifying specificity [96.4% (95% CI 91.5, 100.0); for both serological criteria]. The focus score displayed a sensitivity of 82.2% (95% CI 71.0, 93.4) and a specificity of 82.1% (95% CI 72.1, 92.1). The SGUS score was 52 in 34 of 101 (33.7%) patients. The sensitivity of an SGUS score 52 for the diagnosis of SS was 60.0% (95% CI 45.7, 74.3) and specificity was 87.5% (95% CI 78.8, 96.2). When we added the SGUS score to the three criteria used in the 2012 ACR criteria,
www.rheumatology.oxfordjournals.org
Downloaded from http://rheumatology.oxfordjournals.org/ at Syracuse University Library on April 6, 2014
classification or diagnostic criteria must be evaluated in independent cohorts to assess suitability for use in clinical practice. The Brittany cohort of patients with suspected SS is appropriate for such an evaluation [3]. None of the three ACR criteria reflects salivary gland function and morphology, which are altered in SS. In the 2002 American-European Consensus Group (AECG) classification criteria [4], salivary gland involvement is assessed using salivary flow measurement, salivary gland scintigraphy or parotid sialography. However, all three tests have numerous limitations. In a recent study, we showed that major salivary gland ultrasonography (SGUS) was useful for diagnosing SS and improved the diagnostic performance of the AECG criteria [3]. The objectives of this work were to evaluate the diagnostic performance of the 2012 ACR criteria in an independent cohort of patients with suspected SS and to determine whether this performance was improved by adding SGUS.
Adjunction of SGUS to ACR 2012 criteria for SS
TABLE 1 Comparison of patients with and without SS Overall (n = 101) Age, mean (S.D.), years Symptom duration, mean (S.D.), years Female, n (%) Xerophthalmia, n (%)a Xerostomia, n (%)a OSS 53, n (%) Focus score 51, n (%) Anti-SSA or -SSB positivity, n (%) Anti-SSA/SSB or RF and ANA 51:320, n (%) SGUS score 52, n (%) ACR criteria 52, n (%)
57.4 6.7 95 94 96 48 47 26 29 34 34
(13.5) (6.1) (94.1) (93.1) (95.1) (47.5) (46.5) (25.7) (28.7) (33.7) (33.7)
SS (n = 45) 59.4 7.4 42 43 44 25 37 24 27 27 29
(11.2) (6.7) (93.3) (95.6) (97.8) (55.6) (82.2) (53.3) (60.0) (60.0) (64.4)
No SS (n = 56) 55.7 6.2 53 51 52 23 10 2 2 7 5
(15.1) (5.6) (94.6) (91.1) (92.9) (41.1) (17.9) (3.6) (3.6) (12.5) (8.9)
P-value 0.18 0.31 0.78 0.38 0.26 0.15
