A polygenic risk score (PRS) for systemic lupus erythematosus (SLE)—trained with PRS-CS on a genome-wide association studies of 6,748 cases and 11,516 controls of European ancestry—was applied to Electronic Medical Records and Genomics (eMERGE) Network biobank participants and evaluated against electronic health record phenotypes and Systemic Lupus International Collaborating Clinics (SLICC) criteria. In Northwestern University’s cohort (n=1,061), the PRS predicted SLE with an area under the curve of 0.724 and showed strong associations with lupus phecodes (P=1.6×10⁻⁹; 3.0×10⁻⁹). Nominal associations extended to cardiac complications (P=2.8×10⁻³) and impaired renal function (P=4.6×10⁻³), with enrichment across circulatory (P=9.7×10⁻⁵) and hematopoietic (P=3.2×10⁻²) traits—mirroring common SLE organ systems.
Across eMERGE (n=32,329), higher genetic risk correlated with 5/6 immunologic and 3/11 clinical SLICC criteria, notably acute cutaneous lupus, thrombocytopenia, and leukopenia. The strongest signal was low complement (P=7.4×10⁻³¹): 55% of patients in the highest PRS decile had low complement vs 30% in the lowest. These findings suggest that genetic risk partially stratifies SLE endophenotypes and could augment prediction models for earlier prevention, triage, and management. Future work, researchers note, should expand to multi-ancestry cohorts and integrate longitudinal labs/encounters to improve generalizability and clinical utility.
Reference: Forrest N, Tuteja S, Pacheco J, et al. Polygenic risk of lupus is differentially associated with individual EHR-derived classification criteria [abstract]. Arthritis Rheumatol. 2025; 77 (suppl 9).
