Background <p>Sepsis is a life-threatening condition caused by a dysregulated host response to infection, with significant biological heterogeneity. Transcriptomic profiling offers insights into immune-related gene expression patterns that may inform patient stratification. This study aimed to identify transcriptomic signatures of sepsis or septic shock.</p> Method <p>Eleven adult patients who presented to the emergency department with sepsis or septic shock were enrolled at a tertiary hospital in Korea. Serial whole-blood samples were collected on days 1, 3, and 7. RNA sequencing was performed using Illumina paired-end protocols. Principal component analysis (PCA) was performed on normalized transcriptomic data to derive a gene signature. Enrichment scores were calculated using single-sample gene set enrichment analysis and validated in external bulk and single-cell RNA-seq datasets.</p> Results <p>PCA identified first principal component (PC1) as the major axis of variation, separating samples by sepsis severity and timepoint. Clustering of the top 50 PC1 contributing genes identified three transcriptomic subgroups with distinct expression patterns. Enrichment scores for this gene set increased with severity and declined over time. In multiple external sepsis cohorts, enrichment scores showed consistent trends and were significantly associated with 28-day mortality. Functional analysis revealed predominant neutrophil-associated pathways, as further supported by external single-cell RNA-seq validation.</p> Conclusions <p>We identified a neutrophil-driven transcriptomic signature associated with sepsis severity and temporal immune dynamics. These findings suggest that longitudinal transcriptomic profiling may provide additional insight into the evolving host immune response during sepsis.</p>