(A) scRNA-Seq comparing atrial nonmyocytes from Sh (n = 3) versus TAF (n = 3) mice revealed MIF as a top communication pathway. (B) Reclustering of AFB and macrophages with (C) quantification of Mif expression. Data shown as mean ± IQR. (D and E) MIF inhibition with 4-IPP (50 mg/kg/d) prevented poAF in mice (χ² test). Number of mice per group is denoted in E. Atrial macrophage accumulation by (F) F4/80 protein (N = 6 per group) and (G) flow cytometry (N = 4 per group) were attenuated by 4-IPP without changes in CXCR2 (N = 6 per group). P values were derived from 1-way ANOVA followed by Tukey’s. (H) MIF was elevated in PF from poAF (N = 11) versus SR (N = 6) patients 24 hours after cardiac surgery (2-sample 2-tailed t test). (I) Treatment of THP-1 monocytes (N = 4 per group) with PF (1:10 dilution) for 6 hours increased STAT3-Tyr705 phosphorylation, which was attenuated by 30 minutes of 100 μM 4-IPP pretreatment (1-way ANOVA followed by Tukey’s). AFB, atrial fibroblast; EC, endothelial cell; FB, fibroblast; Mac, macrophage; MIF-i, macrophage migration inhibitory factor inhibitor; norm exp, normalized expression; PF, pericardial fluid; poAF, postoperative atrial fibrillation; scRNA-Seq, single-cell RNA sequencing; Sh, sham; SR, sinus rhythm; TAF, thoracotomy atrial fibrillation; UMAP, uniform manifold approximation and projection.