You are here

Article Text

Article menu
Clinical and epidemiological research
Concise report
The SSB-positive/SSA-negative antibody profile is not associated with key phenotypic features of Sjögren's syndrome
  1. Alan N Baer1,
  2. Mara McAdams DeMarco1,
  3. Stephen C Shiboski2,
  4. Mi Y Lam2,
  5. Stephen Challacombe3,
  6. Troy E Daniels2,
  7. Yi Dong4,
  8. John S Greenspan2,
  9. Bruce W Kirkham3,
  10. Hector E Lanfranchi5,
  11. Morten Schiødt6,
  12. Muthiah Srinivasan7,
  13. Hisanori Umehara8,
  14. Frederick B Vivino9,
  15. Cristina F Vollenweider5,
  16. Yan Zhao4,
  17. Lindsey A Criswell2,
  18. Caroline H Shiboski2
  19. for the Sjögren's International Collaborative Clinical Alliance (SICCA) Research Groups
    1. 1Johns Hopkins University School of Medicine, Baltimore, Maryland, USA
    2. 2University of California, San Francisco, California, USA
    3. 3King's College, London, UK
    4. 4Peking Union Medical College, Beijing, China
    5. 5University of Buenos Aires and German Hospital, Buenos Aires, Argentina
    6. 6Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark
    7. 7Aravind Eye Care System, Madurai, India
    8. 8Kanazawa Medical University, Ishikawa, Japan
    9. 9Penn Presbyterian Medical Center and University of Pennsylvania, Philadelphia, Pennsylvania, USA
    1. Correspondence to Dr Alan N Baer, Division of Rheumatology, Johns Hopkins University School of Medicine, Baltimore, MD 21224, USA; alanbaer{at}jhmi.edu

    Abstract

    Objective To determine whether the Sjögren's syndrome B (SSB)-positive/Sjögren's syndrome A (SSA)-negative antibody profile is associated with key phenotypic features of SS.

    Methods Among registrants in the Sjögren's International Collaborative Clinical Alliance (SICCA) with possible or established SS, we compared anti-SSA/anti-SSB reactivity profiles against concurrent phenotypic features. We fitted logistic regression models to explore the association between anti-SSA/anti-SSB reactivity profile and each key SS phenotypic feature, controlling for potential confounders.

    Results Among 3297 participants, 2061 (63%) had negative anti-SSA/anti-SSB, 1162 (35%) had anti-SSA with or without anti-SSB, and 74 (2%) anti-SSB alone. Key SS phenotypic features were more prevalent and had measures indicative of greater disease activity in those participants with anti-SSA, either alone or with anti-SSB, than in those with anti-SSB alone or negative SSA/SSB serology. These between-group differences were highly significant and not explained by confounding by age, race/ethnicity or gender. Participants with anti-SSB alone were comparable to those with negative SSA/SSB serology in their association with these key phenotypic features. Among SICCA participants classified with SS on the basis of the American-European Consensus Group or American College of Rheumatology criteria, only 2% required the anti-SSB-alone test result to meet these criteria.

    Conclusions The presence of anti-SSB, without anti-SSA antibodies, had no significant association with SS phenotypic features, relative to seronegative participants. The solitary presence of anti-SSB antibodies does not provide any more support than negative serology for the diagnosis of SS. This serological profile should thus be interpreted cautiously in clinical practice and potentially eliminated from future classification criteria.

    • Sjögren's Syndrome
    • Autoantibodies
    • Systemic Lupus Erythematosus

    https://doi.org/10.1136/annrheumdis-2014-206683

    Statistics from Altmetric.com

      Request Permissions

      If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

      Anti-Sjögren's syndrome A (SSA) (Ro) and anti-Sjögren's syndrome B (SSB) (La) antibodies are present in up to 75% of patients with primary SS.1 ,2 Two profiles of anti-SSA/anti-SSB reactivity are common, both anti-SSA and anti-SSB and anti-SSA alone, with the former being more common than the latter.3 ,4 Anti-SSB alone is an uncommon serological profile in established SS3 ,5–8 but often prompts an SS evaluation in patients with sicca symptoms. It may reflect the reported 1–15% prevalence of anti-SSB alone in healthy individuals7 ,9 ,10 and the increased sensitivity of current solid-phase immunoassays.11

      Anti-SSA/anti-SSB serology is a criterion for the classification of SS in both the American-European Consensus Group (AECG) and American College of Rheumatology (ACR) sets.12 ,13 However, it is not known whether the solitary presence of anti-SSB has validity equivalent to anti-SSA, present either alone or with anti-SSB, in supporting SS classification. Accordingly, we sought to determine the association of three different anti-SSA/anti-SSB serological profiles with SS phenotypic features among participants in the Sjögren's International Collaborative Clinical Alliance (SICCA) registry.14

      Methods

      SICCA cohort and study population

      SICCA is a registry of individuals with symptoms or signs indicative of possible, early to well-established SS, each of whom underwent a systematic and extensive evaluation for SS using uniform protocol-driven data collection methods.14 Methodological details of this registry are provided in the online supplementary materials. There were 3514 SICCA participants enrolled as of 6 September 2013. We excluded 217 for whom data were lacking on at least one of the three objective criteria for SS, as defined by the ACR criteria,4 leaving 3297 participants for the current cross-sectional analyses.

      All SICCA serological testing was performed centrally by Quest Diagnostics (Madison, New Jersey, USA). For the initial 876 registrants, anti-SSA and anti-SSB were tested using the Bio-Rad Autoimmune EIA (Bio-Rad, Hercules, California, USA), a semiquantitative enzyme immunoassay (EIA) that used purified native antigens and reported results in either enzyme units or index values. The subsequent 2421 registrants had anti-SSA/anti-SSB testing performed with the newly introduced Bio-Rad Bioplex 2200 multiplex flow immunoassay (MFIA). For this assay, the Ro52 antigen was recombinant, while Ro60 and SSB were native in origin.15 Positive results were expressed in ‘antibody index’ (AI) units and provided as continuous variable measures up to 8 AI.

      Statistical analysis

      Proportions for categorical variables and median (range) for continuous variables were used to describe the sociodemographic features in the cohort. We used Fisher's exact tests to evaluate associations between categories of anti-SSA/anti-SSB serological reactivity (anti-SSA with or without anti-SSB, only anti-SSB, and lacking both anti-SSA and anti-SSB at baseline) and key SS phenotypic features. Rank-sum tests were used to compare the continuous SS phenotypic measures between the three groups (Kruskal–Wallis) and for pairwise comparisons between selected groups of interest (Wilcoxon) defined by serological reactivity.

      We used logistic regression models to further investigate the association between anti-SSA/anti-SSB serological profile (included as the primary predictor) on each key SS phenotypic feature (treated as binary outcomes), controlling for potential confounders, and allowing for the possibility of interaction between the presence of anti-SSA and of anti-SSB. Confounding variables included age (in years), sex and race (with the Caucasian subgroup as reference). All statistical analyses were performed using SAS (V.9.3, SAS Institute, Cary, North Carolina, USA) and Stata (V.13.1, StataCorp, College Station, Texas, USA).

      Results

      The sociodemographic characteristics of the 3297 participants are shown in table 1. A total of 1490 (45%) met the ACR criteria (129 with secondary SS) and 1457 (44%) met the AECG criteria for SS (119 with secondary SS).

      View this table:
      Table 1

      Sociodemographic characteristics of 3297 SICCA participants*

      The majority of participants had negative SSA/SSB serology, 35% had anti-SSA with or without anti-SSB and 2% had only anti-SSB (table 2). The anti-SSB alone serological profile was more prevalent among the participants who had their testing performed with MFIA (3% vs 0.3%; p<0.0001). However, these two groups were not equivalent since the EIA group had a greater proportion of individuals with SS defined by ACR criteria (49% vs 44%; p=0.0113). Anti-SSB was more commonly present in low titre (levels of 1 to ≤6 index units by EIA, and 1 to ≤8 AI units by MFIA) when occurring alone than when associated with anti-SSA of any titre (91% vs 49%; p<0.0001).

      View this table:
      Table 2

      Phenotypic characteristics (categorical variables) of 3297 SICCA participants by anti-SSA and anti-SSB reactivity patterns

      The prevalence of key SS phenotypic features differed significantly among the three anti-SSA/anti-SSB serological groups (table 2). These differences stemmed primarily from comparisons of the group with anti-SSA with or without anti-SSB and the groups of patients with anti-SSB alone or with negative SSA/SSB. In contrast, the frequency of these disease markers was statistically comparable between the groups with only anti-SSB and those with negative SSA/SSB. The only exceptions were low unstimulated whole saliva flow (UWSF) rate, dry mouth symptoms and antinuclear antibody positivity that were actually more prevalent among the participants with negative SSA/SSB than in those with anti-SSB alone.

      Key phenotypic features of SS expressed as continuous variables also differed by SSA/SSB serological profile (table 2). The median ocular staining score (OSS) and focus score (FS) were higher and Schirmer test and UWSF values were lower among participants with anti-SSA with or without anti-SSB than among participants with only anti-SSB or with negative anti-SSA/anti-SSB. The differences between the two groups with respect to these SS phenotypic features were all statistically significant (p<0.0001). Levels of OSS and FS did not differ significantly between the group with only anti-SSB and that with negative anti-SSA/anti-SSB. In contrast, the group with only anti-SSB had significantly higher median Schirmer test and UWSF results than the group lacking both antibodies, opposite to what was expected.

      With a logistic regression analysis of the joint effects of anti-SSA and anti-SSB on selected SS phenotypic features, we quantified the between-group differences observed in table 2 and showed that these are not explained by confounding by age, race/ethnicity or gender. Results are summarised in table 3 as adjusted ORs comparing the odds of occurrence for each feature between specified groups defined using anti-SSA and anti-SSB status.

      View this table:
      Table 3

      Effect of anti-SSA and/or anti-SSB status on each phenotypic feature of SS*

      Positive SSA/SSB serology was present in 1138 (76%) of those participants who met the ACR criteria and in 1067 (73%) of those who met the AECG criteria for SS. In each of these two groups, 52 participants had anti-SSB alone (ACR: 3%; AECG: 4%).

      Classification of subjects with SS was dependent on the presence of anti-SSB alone (due to the absence of the positive lip biopsy criterion) in 33 (2.2%) of the subjects using the ACR criteria and in 34 (2.3%) using the AECG criteria.

      Discussion

      In this large registry of individuals with suspected or established SS, anti-SSB in the absence of anti-SSA was found in 2% and had no association with SS phenotypic features compared with those who lacked both antibodies. This is an important and novel observation since current SS classification schema include these antibodies as criteria but do not differentiate among the various possible anti-SSA and anti-SSB serological profiles, giving equal weight to all.12 ,13

      The rare detection of the ‘anti-SSB alone’ pattern of reactivity among the SICCA registrants may be a function of the assay methodology or a true entity. The Bio-Rad Bioplex 2200 MFIA has greater sensitivity for anti-SSB relative to double immunodiffusion and HEp-2 immunofluorescence assays.11 ,16 In our cohort, anti-SSB alone was 10 times more commonly detected with this MFIA than with EIA. A false-positive reaction to SSB may result from low-titre, weak-affinity antibodies with no pathological relevance15 or cross-reacting antibodies, such as those to histones, cardiac myosin and spectrin.17–19 Recently, Danda et al20 analysed 29 anti-Ro60 (SSA)-negative/anti-La (SSB)-positive sera from their cohort of 468 patients with SS; in 25, the sera were unequivocally negative for anti-Ro60 with four different assay methods, including an immunofluorescent assay using HEp-2000 cells transfected with human Ro60.

      As part of a separate study, 5 of the 74 SSB-positive/SSA-negative sera from the SICCA cohort were assayed in the Johns Hopkins Rheumatic Diseases Research Core Center for antibodies to Ro52 and SSB using ELISA (QUANTA Lite, Inova Diagnostics) and for Ro60 by immunoprecipitation using in vitro transcription/translated protein.21 ,22 Each of the sera was negative for anti-Ro52 and anti-Ro60 antibodies. Anti-SSB antibodies were confirmed by ELISA in three of the sera, with weak positive (20–39 units) results in two and moderate positive (40–80 units) in one. These findings support those of Danda et al and suggest that the SSB-positive/SSA-negative serological profile can be a true positive, although the anti-SSB reactivity is typically low titre.

      The presence of anti-SSA and/or anti-SSB is a mandatory criterion in the absence of a ‘positive’ labial salivary gland (LSG) biopsy in the 2002 AECG and 2012 ACR criteria sets.12 ,13 The findings in the current study have implications for the serological diagnosis of SS as stipulated in these criteria. Thus, our data argue that the SSB-positive/SSA-negative antibody profile should be interpreted cautiously in a patient with suspected SS. In this circumstance, additional support for the diagnosis is advisable in the form of a LSG biopsy showing focal lymphocytic sialadenitis with an FS ≥1. Elimination of anti-SSB alone as a serological criterion for SS should be considered by the international community as it prepares new consensus criteria for SS; the impact would be limited to only 2% of subjects currently classified with SS by the AECG and ACR criteria sets.

      Strengths of our study include the large size and geographic heterogeneity of the SICCA cohort. Inclusion criteria for the registry were intentionally broad to ensure representation of the spectrum of individuals evaluated for SS. Each participant underwent a uniform evaluation with all LSG biopsies being read centrally and SSA/SSB serological testing being performed in the same clinical laboratory, thereby assuring results representative of those encountered in clinical practice.

      A limitation of the study is the small number of individuals with anti-SSB alone, resulting in lower power for statistical analyses related to this group. Nevertheless, this is the largest number studied to date. Our analysis was also cross sectional.

      In conclusion, the presence of anti-SSB alone is rare and has no association with key SS phenotypic features. These findings are particularly relevant with the advent of MFIAs for anti-SSA/anti-SSB testing since they may reveal low levels of anti-SSB with no clinical significance in the evaluation of a patient with suspected SS.

      Acknowledgments

      Dr Baer is supported by NIH RO1 DE12354-15A1. The Johns Hopkins Rheumatic Diseases Research Core Center, which is supported by NIH grant P30-AR-053503, performed the validation assays.

      References

      1. ↵
        1. Bournia VK,
        2. Vlachoyiannopoulos PG
        . Subgroups of Sjögren syndrome patients according to serological profiles. J Autoimmun 2012;39:15–26. doi:10.1016/j.jaut.2012.03.001
        OpenUrlCrossRefPubMedWeb of Science
      2. ↵
        1. Malladi AS,
        2. Sack KE,
        3. Shiboski SC, et al
        . Primary Sjögren's syndrome as a systemic disease: a study of participants enrolled in an international Sjögren's syndrome registry. Arthritis Care Res 2012;64:911–18. doi:10.1002/acr.21610
        OpenUrlCrossRefWeb of Science
      3. ↵
        1. Ramos-Casals M,
        2. Solans R,
        3. Rosas J, et al
        . Primary Sjögren syndrome in Spain: clinical and immunologic expression in 1010 patients. Medicine 2008;87:210–19. doi:10.1097/MD.0b013e318181e6af
        OpenUrlCrossRefPubMedWeb of Science
      4. ↵
        1. Fauchais AL,
        2. Martel C,
        3. Gondran G, et al
        . Immunological profile in primary Sjögren syndrome: clinical significance, prognosis and long-term evolution to other auto-immune disease. Autoimmun Rev 2010;9:595–9. doi:10.1016/j.autrev.2010.05.004
        OpenUrlCrossRefPubMed
      5. ↵
        1. Gerli R,
        2. Muscat C,
        3. Giansanti M, et al
        . Quantitative assessment of salivary gland inflammatory infiltration in primary Sjögren's syndrome: its relationship to different demographic, clinical and serological features of the disorder. Br J Rheumatol 1997;36:969–75. doi:10.1093/rheumatology/36.9.969
        OpenUrlAbstract/FREE Full Text
      6. ↵
        1. Wahren M,
        2. Solomin L,
        3. Pettersson I, et al
        . Autoantibody repertoire to Ro/SSA and La/SSB antigens in patients with primary and secondary Sjögren's syndrome. J Autoimmun 1996;9:537–44. doi:10.1006/jaut.1996.0072
        OpenUrlCrossRefPubMedWeb of Science
      7. ↵
        1. Garberg H,
        2. Jonsson R,
        3. Brokstad KA
        . The serological pattern of autoantibodies to the Ro52, Ro60, and La48 autoantigens in primary Sjögren's syndrome patients and healthy controls. Scand J Rheumatol 2005;34:49–55. doi:10.1080/03009740510017940
        OpenUrlCrossRefPubMedWeb of Science
      8. ↵
        1. Peene I,
        2. Meheus L,
        3. Veys EM, et al
        . Diagnostic associations in a large and consecutively identified population positive for anti-SSA and/or anti-SSB: the range of associated diseases differs according to the detailed serotype. Ann Rheum Dis 2002;61:1090–4. doi:10.1136/ard.61.12.1090
        OpenUrlAbstract/FREE Full Text
      9. ↵
        1. Gaither KK,
        2. Fox OF,
        3. Yamagata H, et al
        . Implications of anti-Ro/Sjogren's syndrome A antigen autoantibody in normal sera for autoimmunity. J Clin Invest 1987;79:841–6. doi:10.1172/JCI112892
        OpenUrlCrossRefPubMedWeb of Science
      10. ↵
        1. Metskula K,
        2. Salur L,
        3. Mandel M, et al
        . Demonstration of high prevalence of SS-A antibodies in a general population: association with HLA-DR and enterovirus antibodies. Immunol Lett 2006;106:14–18. doi:10.1016/j.imlet.2006.03.005
        OpenUrlCrossRefPubMedWeb of Science
      11. ↵
        1. Hanly JG,
        2. Thompson K,
        3. McCurdy G, et al
        . Measurement of autoantibodies using multiplex methodology in patients with systemic lupus erythematosus. J Immunol Methods 2010;352:147–52. doi:10.1016/j.jim.2009.10.003
        OpenUrlCrossRefPubMedWeb of Science
      12. ↵
        1. Vitali C,
        2. Bombardieri S,
        3. Jonsson R, et al
        . Classification criteria for Sjögren's syndrome: a revised version of the European criteria proposed by the American-European Consensus Group. Ann Rheum Dis 2002;61:554–8. doi:10.1136/ard.61.6.554
        OpenUrlAbstract/FREE Full Text
      13. ↵
        1. Shiboski SC,
        2. Shiboski CH,
        3. Criswell L, et al
        . American College of Rheumatology classification criteria for Sjögren's syndrome: a data-driven, expert consensus approach in the Sjogren's International Collaborative Clinical Alliance cohort. Arthritis Care Res 2012;64:475–87. doi:10.1002/acr.21591
        OpenUrlCrossRefWeb of Science
      14. ↵
        1. Daniels TE,
        2. Criswell LA,
        3. Shiboski C, et al
        . An early view of the international Sjögren's syndrome registry. Arthritis Rheum 2009;61:711–14. doi:10.1002/art.24397
        OpenUrlCrossRefPubMedWeb of Science
      15. ↵
        1. Infantino M,
        2. Bentow C,
        3. Seaman A, et al
        . Highlights on novel technologies for the detection of antibodies to Ro60, Ro52, and SS-B. Clin Dev Immunol 2013;2013:978202. doi:10.1155/2013/978202
        OpenUrlPubMed
      16. ↵
        1. Desplat-Jego S,
        2. Bardin N,
        3. Larida B, et al
        . Evaluation of the BioPlex 2200 ANA screen for the detection of antinuclear antibodies and comparison with conventional methods. Ann NY Acad Sci 2007;1109:245–55. doi:10.1196/annals.1398.030
        OpenUrlCrossRefPubMedWeb of Science
      17. ↵
        1. Touloupi E,
        2. Routsias JG,
        3. Tzioufas AG
        . Cross-recognition between histones and La/SSB may account for anti-DNA reactivity in SLE patients. Clin Exp Immunol 2005;142:172–9. doi:10.1111/j.1365-2249.2005.02892.x
        OpenUrlCrossRefPubMed
      18. ↵
        1. Horsfall AC,
        2. Rose LM
        . Cross-reactive maternal autoantibodies and congenital heart block. J Autoimmun 1992;5:479–93. doi:10.1016/0896-8411(92)90007-D
        OpenUrlCrossRefPubMedWeb of Science
      19. ↵
        1. Kurien BT,
        2. Dorri Y,
        3. Bachmann M, et al
        . Induction of anti-Ro60/anti-La by immunization with spectrin and induction of anti-spectrin by immunization with Ro60 and 4-hydroxy-2-nonenal-modified Ro60 immunisation. Clin Exp Rheumatol 2012;30:886–93.
        OpenUrlPubMed
      20. ↵
        1. Danda D,
        2. Truong D,
        3. Shaw M, et al
        . La positive, Ro60 negative subset of primary Sjögren's syndrome is a reality. Arthritis Rheum 2014;66:S1111. (abstract).
      21. ↵
        1. Nagaraj K,
        2. Cox A,
        3. Casciola-Rosen L, et al
        . Novel fragments of the Sjogren's syndrome autoantigens alpha-fodrin and type 3 muscarinic acetylcholine receptor generated during cytotoxic lymphocyte granule-induced death. Arthritis Rheum 2001;44:2376–86. doi:10.1002/1529-0131(200110)44:10<2376::AID-ART402>3.0.CO;2-E
        OpenUrlCrossRefPubMedWeb of Science
      22. ↵
        1. Schachna L,
        2. Wigley FM,
        3. Morris S, et al
        . Recognition of granzyme B-generated autoantigen fragments in scleroderma patients with ischemic digital loss. Arthritis Rheum 2002;46:1873–84. doi:10.1002/art.10407
        OpenUrlCrossRefPubMedWeb of Science

      Supplementary materials

      • Supplementary Data

        This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

        Files in this Data Supplement:

      Footnotes

      • Handling editor Tore K Kvien

      • Collaborators Collaborators of the Sjögren’s International Collaborative Clinical Alliance are as follows: D. Cox, R. Jordan (oral pathology), D. Lee (rheumatology), Y. DeSouza (operations director), D. Drury (clinical coordinator/phlebotomy), A. Do (clinical coordinator), L. Scott (clinical assistant), M. Lam (statistician/programmer), J. Nespeco (data manager), J. Whiteford (finance director), M. Margaret (administrative assistant): University of California, San Francisco; I. Adler, A. C. Smith, A. M. Bisio, M. S. Gandolfo (stomatology), A. M. Chirife, A. Keszler (oral pathology), A. M. Heidenreich (ophthalmology), S. Daverio (specimen processing), V. Kambo (group coordinator): University of Buenos Aires and German Hospital, Buenos Aires, Argentina; Y. Jiang, D. Xu, J. Su (rheumatology), D. Du (stomatology/pathology), H. Wang, Z. Li, J. Xiao (stomatology/labial salivary gland biopsies), Q. Wu (specimens/rheumatology), C. Zhang, W. Meng (phlebotomy), J. Zhang (project assistant): Peking Union Medical College Hospital, Beijing, China; S. Johansen, S. Hamann (ophthalmology), J. Schiødt, H. Holm (oral medicine), P. Ibsen (oral pathology), A. M. Manniche, S. P. Kreutzmann, J. Villadsen (group coordinators/specimen handling), Rigshospitalet, Copenhagen, Denmark; Y. Masaki, T. Sakai (rheumatology), N. Shibata (ophthalmology), M. Honjo (stomatology), N. Kurose, T. Nojima (oral pathology), T. Kawanami (specimen processing), T. Sawaki (haematology/immunology), K. Fujimoto (group coordinator): Kanazawa Medical University, Ishikawa, Japan; E. Odell, P. Morgan (pathology), L. Fernandes-Naglik (specimen processing), B. Varghese-Jacob, S. Ali (oral medicine), M. Adamson (project coordinator): King's College London, London, UK; S. Seghal, R. Mishra (rheumatology), V. Bunya, M. Massaro-Giordano (ophthalmology), S. K. Abboud (otolaryngology), A. Pinto, Y. W. Sia (oral medicine), K. Dow (group coordinator): University of Pennsylvania, Philadelphia; E. Akpek, S. Ingrodi (ophthalmology), W. Henderson (oral medicine), C. Gourin (otolaryngology), A. Keyes, R. Ozl (group coordinators): Johns Hopkins University, Baltimore, Maryland; M. Srinivasan (group director), J. Mascarenhas, M. Das, A. Kumar (co-directors), P. Joshi (ophthalmology), R. Banushree (physician), U. Kim (surgeon), B. Babu (oral medicine), A. Ram, R. Saravanan, K. N. Kannappan (administration), N. Kalyani (group coordinator): Aravind Eye Hospital, Madurai, India.

      • Contributors Each of the authors made substantial contributions to the conception and design of the SICCA study, as well as to the acquisition, analysis and interpretation of the data. Each reviewed this manuscript critically and provided revisions where appropriate. Each will approve the final version for publication. Each is accountable for the accuracy and integrity of this work.

      • Funding This work was supported by a contract from the NIH/NIDCR NO1-DE-32636, the Jerome L. Greene Foundation and the Rosalind Russell/Ephraim P. Engleman Rheumatology Research Center.

      • Competing interest s JSG has received an unrestricted educational gift from GlaxoSmithKline.

      • Ethics approval The Sjögren's International Collaborative Clinical Alliance (SICCA) research study was approved by the institutional review boards at the University of California, San Francisco (study centre) and at each of the research sites.

      • Provenance and peer review Not commissioned; externally peer reviewed.

      • Data sharing statement Data from the SICCA registry are available to all investigators through a dissemination process highlighted on the SICCA website http://sicca.ucsf.edu/.