GLP-1RA Exposure and Incident Persistent Cytopenias

Glucagon-like peptide-1 receptor agonists (GLP-1RAs) have been associated with a lower incidence of several hematologic malignancies. Whether this drug class influences earlier stages of transformation to myeloid neoplasms (MN) is unknown. Persistent cytopenias (PC) represent an important intermediate phenotype preceding MN development and serve as a marker of malignant evolution. Here, we evaluated the association between GLP-1RA exposure and incident PC, lineage-specific cytopenias, and MN risk in Vanderbilt University Medical Center's BioVU.

We performed a retrospective active-comparator incident-user cohort study using de-identified electronic health records and whole-genome sequencing data stored in BioVU from 340,702 individuals. Adults ≥ 40 years with type 2 diabetes mellitus (T2DM) and/or obesity initiating a GLP-1RA, SGLT2 inhibitor (SGLT2i), DPP-4 inhibitor (DPP4i), sulfonylurea (SU), or metformin therapy were eligible. An exposure episode required ≥ 90 consecutive therapy days with ≤ 30-day refill gaps. Patients with cytopenia or MN prior to the 90 day post-index episode date were excluded. Primary outcomes included persistent anemia (PA), leukopenia (PL), thrombocytopenia (PT), or any PC, which required two values below the lower limit of normal in the same lineage ≥ 120 days apart as previously defined, and the secondary outcome was incident MN. Overlap propensity weighting was used to balance baseline demographics, drug indication, smoking history, baseline blood counts, and calendar year of index episode. Associations were estimated using overlap-weighted cause-specific Cox proportional hazards models.

Four active-comparator cohorts were constructed comparing GLP-1RA users with DPP-4i (GLP1RA: n=8,466; DPP-4i: n=2,116), SGLT2i (GLP-1RA: n=8,935; SGLT2i: n=6,678), SU (GLP-1RA: n=8,157; SU: n=5,281), and metformin (GLP-1RA: n=5,342; metformin: n=13,413) with a median follow-up of 2.48 (IQR:1.22-3.96) years across all participants. Overlap propensity weighting achieved strong covariate balance across all pairwise comparisons, with post-weighting standardized mean differences approaching 0 for all measured baseline characteristics. Compared with DPP-4i, GLP-1RA exposure was associated with a lower risk of PC (HR 0.79, 95% CI 0.70-0.90, p<0.001), PA (HR 0.83, 95% CI 0.73-0.95, p=0.007), PL (HR 0.33, 95% CI 0.16-0.66, p=0.002), and PT (HR 0.68, 95% CI 0.49-0.96, p=0.026). Compared with SU and metformin, GLP-1RA exposure was similarly associated with reduced risks of PC (SU: HR 0.83, 95% CI 0.74-0.92, p<0.001; metformin: HR 0.89, 95% CI 0.80-0.98, p=0.017) and PA (SU: HR 0.83, 95% CI 0.74-0.93, p=0.001; metformin: HR 0.89, 95% CI 0.80-0.99, p=0.028). In contrast, compared with SGLT2i, GLP1RA users had a higher risk of PC (HR 1.28, 95% CI 1.15-1.43, p<0.001), with PA (HR 1.33, 95% CI 1.18-1.49, p<0.001) accounting for most events. Incidence of clonal hematopoiesis (CH) across each active comparator cohort ranged from 5.0-5.4%, however there was no significant reduction in incident MN risk due to the short follow up and drug exposure durations as well as low number of incident MN events.

In this cohort, GLP-1RA initiation was associated with lower rates of PC and PA compared to DPP-4i, SU, and metformin, but not with SGLT2i, whose established erythropoietin-stimulating effects may account for the comparatively lower anemia risk. GLP1-RAs may relieve T2DM and obesity-related inflammation by decreasing NF-kB activation and adipose macrophage infiltration as well as suppressing TNF-alpha, IL-6, and IL-1 beta. These pathways may restrain erythropoiesis directly and through hepcidin-mediated functional iron restriction, thereby biasing hematopoietic output towards inflammation-adapted myelopoiesis. Additional analyses investigating whether GLP-1RA's hematologic benefit is concentrated in TET2 and JAK2 relative to DNMT3A clones are indicated to discriminate between modification of clone evolution and an effect on counts alone. These findings provide one of the largest real-world evaluations of hematologic outcomes associated with GLP1RA therapy. Future prospective studies with longer follow-up are needed to determine whether GLP1RAs alter clonal evolution and ultimately prevent myeloid neoplasm development.