Ab0288 comorbidity clusters in patients with rheumatoid arthritis are associated with disease activity and predict survival prognosis

Background Comorbidities are common in patients with rheumatoid arthritis (RA), but there is little information regarding distinct comorbidity patterns in patients with RA and how these patterns impact disease activity and mortality. Objectives To examine clusters of patients with RA based on comorbidities, the association between these clusters and RA disease activity, and their impact on mortality. Methods In this retrospective, population-based study, residents of a geographically well-defined area with prevalent RA on 1-1-2015 were identified. Patients were followed for vital status until death, last contact or 12-31-2021. Clinical Disease Activity Index (CDAI) was obtained from the medical visit closest to the prevalence date (±1 year). Diagnostic codes were retrieved for 5 years prior to the prevalence date. Using 2 codes ≥30 days apart, 55 comorbidities were defined. Latent class analysis (LCA) was used to cluster patients based on comorbidity patterns with the optimal number of clusters determined by Bayesian Information Criteria. Standardized mortality ratios (SMR) were used to assess differences in mortality across the clusters. Results A total of 1643 patients with prevalent RA (72% female; 94% white; median age 64 years, median RA duration 7 years) were studied. Comorbidities were present in 1548 (94%) of the subjects with 5+ comorbidities present in 980 (59%) subjects. LCA identified 4 clusters of patients. Although only comorbidities were used in the LCA, the clusters differed by age, sex, number of comorbidities, smoking status, body mass index (BMI) and RF/ anti-CCP positivity (Table 1). Cluster 1 (n=686) included younger patients who were less likely to smoke and had fewer comorbidities compared to other clusters. In contrast, cluster 4 (n=134) included older patients with 6+ comorbidities who were more likely to have smoked. Cluster 2 (n=200) included younger patients with 6+ comorbidities and high prevalences of depression and obesity, while cluster 3 (n=623) included the remaining patients with 3+ comorbidities consisting mainly of hypertension, hyperlipidemia and back pain). Cluster 1 had no increased mortality based on US lifetables (SMR 0.8; 95% confidence interval [CI: 0.55-1.1). Cluster 4 had the worst survival (SMR 3.5; 95% CI 2.8-4.4). Cluster 2 also had poor survival (SMR 1.6; 95% CI 1.03-2.2), while the poor survival in cluster 3 was less pronounced (SMR 1.2; 95% CI 0.99-1.4). Data on CDAI were available on 725 patients who were mostly like those without CDAI except they were slightly younger (mean age 62 vs 64 years), fewer ever smoked (51% vs 58%) and more were seropositive (71% vs 63%) compared to those without CDAI. CDAI differed across the clusters (p<0.001) with the highest proportion of remission in Cluster 1 and the lowest proportion of remission and the largest proportion of high disease activity in Cluster 2. Conclusion Clustering of patients with RA based on comorbidity profiles identified distinct groups exhibiting different prognosis. This demonstrates comorbidities congregate in some patients, leading to poor prognosis, while others experience fewer comorbidities and better survival than people without RA. The cluster with both better survival and more remission despite a high proportion of seropositivity is intriguing. Table 1. Characteristics of four clusters of patients with RA Characteristic C1 (N=686) C2 (N=200) C3 (N=623) C4 (N=134) p value Age, years 55 (46, 63) 60 (51, 68) 72 (63, 78) 77 (71, 82) < 0.001 Female sex 75% 86% 64% 66% < 0.001 RF/anti-CCP positive 70% 62% 64% 65% 0.073 RA duration, years 6 (3, 12) 5 (3, 11) 8 (3, 17) 9 (4, 21) < 0.001 BMI, kg/m 2 27 (24, 33) 31 (27, 35) 29 (25, 34) 28 (25, 34) < 0.001 Ever smoker 50% 60% 57% 62% 0.005 Number of comorbidities < 0.001 0-2 43% 0% 0% 0% 3-5 43% 0% 29% 0% 6+ 4% 100% 71% 100% Disease activity N=327 N=94 N=260 N=44 CDAI 4 (1, 13) 11 (6, 20) 8 (3,13) 7 (4, 13) <0.001 Remission (≤2.8) 40% 11% 26% 25% Low (>2.8 to 10.0) 29% 36% 38% 39% Moderate (>10.0 to 22.0) 19% 31% 23% 20% High (>22.0) 12% 22% 13% 16% Values in the table are median (25 th , 75 th %ile) or % REFERENCES: NIL. Acknowledgements: NIL. Disclosure of Interests Cynthia S. Crowson: None declared, Elizabeth Atkinson: None declared, Elena Myasoedova: None declared, Vanessa Kronzer: None declared, Bradly Kimbrough: None declared, Chanakya Kodishala: None declared, Edward Lovering: None declared, Rakesh Kumar: None declared, John M Davis III Grant/research support from: Pfizer.