Azithromycin intake is associated with immune cell subset changes. (A) Frozen peripheral mononuclear blood cells (PBMCs) from patients included in placebo (PLA; n = 106) or azithromycin (AZM; n = 98) arm and from healthy donors (n = 20) were thawed and analyzed by means of mass cytometry with a panel targeting 43 antigens used to cluster cells according to their phenotype and T-cell functional state (naïve, activated, or exhausted). A preprocessing pipeline was used to normalize data across batches and then identify singlet CD45⁺ living cells. Fifty-five cell-phenotypic subsets were identified with the use of 31 phenotype antigens and FlowSOM algorithm. (B) Circular dendrogram showing the 55 cell-subset hierarchy colored according to the corresponding subsets and sized by frequency among CD45⁺ cells. (C) Uniform manifold approximation and projection, depicting cell clustering colored according to their subsets. (D) Boxplots representing percentage of main PBMC subsets among living CD45⁺ cells according to sample groups. (E) Heatmap representing scaled expression of phenotype antigen across the cell subsets manually ordered and annotated for visualization purposes. Targeted antigens are ordered by hierarchic clustering, and unidentified cell subsets are shown in supplemental Figure 4. Fold changes of immune subsets in AZM group compared with PLA are summarized with a bar plot (∗P < .05; ∗∗P < .01; P values are shown in F). Bold names of subsets indicate significant difference. (F) Boxplots of statistically different subsets between AZM and PLA cohorts. For visualization purposes, square root transformation was applied on the y-axis. All P values were calculated by means of 2-sided Wilcoxon signed rank test. DN, double negative; EM, effector memory; EMRA, effector memory CD45RA⁺; Non-conv, nonconventional; MAIT, mucosal-associated invariant T cells; NK, natural killer.