OP0081 Phenotypic subgroups in igg4-related disease – a cluster analysis

• Published in: Annals of the rheumatic diseases Vol. 77; no. Suppl 2; p. 91
• Main Authors: Wallace, Z.S., Zhang, Y., Perugino, C.A., Naden, R., Choi, H., Stone, J.H.
• Format: Journal Article
• Language: English
• Published: Kidlington Elsevier Limited 01.06.2018
• Subjects: Age; Aorta; Bayesian analysis; Biopsy; Cluster analysis; Diagnosis; Ethnicity; Etiology; Fibrosis; Hypocomplementemia; Immunoglobulin G; Lung diseases; Mathematical models; Minority & ethnic groups; Morbidity; Patients; Phenotypes; Risk factors; Salivary gland
• ISSN: 0003-4967; 1468-2060
• DOI: 10.1136/annrheumdis-2018-eular.4749

BackgroundIgG4-related disease (IgG4-RD) is a multi-organ immune-mediated condition of uncertain etiology characterised by substantial organ-specific morbidity if not diagnosed and treated promptly. Identifying IgG4-RD subgroups based on the distribution of organ involvement may influence the understanding of pathogenesis and guide clinical management.ObjectivesTo identify phenotypic clusters of IgG4-RD that may differentiate clinically meaningful subgroups using an unbiased method.MethodsThe study cohort consisted of 493 IgG4-RD subjects diagnosed by 76 IgG4-RD specialists from North America, South America, Europe, and Asia. For each case, investigators included details regarding age at disease onset and diagnosis, race/ethnicity, organ involvement, biopsy findings, and lab results. We performed latent class analysis (LCA) using SAS procedure PROC LCA to identify subgroups representing distinct patterns of organ involvement by IgG4-RD (figure 1). We fitted LCA models with 2–5 subgroups and chose the best model based on Akaike information criteria and adjusted Bayesian information criterion. The posterior probability of subgroup (cluster) membership for all cases was determined and cases were assigned to the cluster in which they had the highest probability of membership. We compared the distribution of organ involvement and other baseline features between clusters using Chi square tests and analysis of variance, where appropriate.ResultsOf the 493 IgG4-RD subjects, 65% were male, 40% were Caucasian, 45% were Asian, and 12% were Hispanic. The mean age at diagnosis was 59.5 (±14.0) years. Using LCA, we identified four clusters of IgG4-RD (table 1), each of which accounted for between 19% and 32% of the cohort. Cluster 1 (‘Hepatobiliary’) included 158 (32%) patients characterised by hepatobiliary involvement. Cluster 2 (‘Orbital’) included 88 (19%) patients characterised by orbital and/or sinus disease. Cluster 3 (‘Mikulicz’) included 109 (22%) patients who had features of classic Mikulicz (dacryoadenitis plus major salivary gland involvement), often accompanied by renal and lung disease. Cluster 4 (‘Retroperitoneal Fibrosis (RPF)’) included 138 (28%) patients with RPF and/or aortic involvement. The clusters differed significantly with regard to age at symptom onset (p<0.001), gender and race distribution (p<0.001), serum IgG4 concentration (p=0.02), and presence of hypocomplementemia (p<0.001). In contrast to the other clusters, cluster 2 (‘Orbital’) included a majority of female patients who tended to be younger. Cluster 3 (‘Mikulicz’) was characterised by the highest serum IgG4 concentrations and cluster 4 (‘RPF’) by the lowest. Hypocomplementemia, which occurred in only a minority of patients overall (9%), tended to segregate in cluster 3 (‘Mikulicz’), a group in which renal disease was common.Abstract OP0081 – Table 1ConclusionsUsing an unbiased method, we identified four phenotypic clusters of IgG4-RD patients. In addition to the differences in organ involvement, clusters were distinguished by age at diagnosis as well as race/ethnicity and gender distribution, serum IgG4 concentrations, and frequency of hypocomplementemia. These clusters may identify patients with IgG4-RD resulting from different risk factors or exposures and those likely to respond differently to treatment.Disclosure of InterestNone declared