# Limited Predictive Utility of Baseline Peripheral Blood Bulk Transcriptomics for Influenza Vaccine Responsiveness in Older Adults

## TL;DR

Baseline blood RNA profiles don't reliably predict which older adults will respond to flu vaccines.

## Contribution

Identified transcriptomic differences between vaccine responders and non-responders in older adults.

## Key findings

- 1152 genes differed between responders and non-responders at baseline.
- Responder genes were linked to B cell activation and protein synthesis.
- Non-responders showed genes related to innate immunity and platelet activation.

## Abstract

Background: Older adults face increased risks of influenza infection and related complications due to declining immunity and reduced vaccine responsiveness. Despite widespread vaccination, only 30–40% mount immune response due to immunosenescence. However, no biomarkers exist to identify potential non-responders, limiting the ability to target vaccine strategies, like high-dose or adjuvanted formulations, to those unlikely to benefit from standard options. Methods: We analysed publicly available baseline bulk RNA sequencing data from peripheral blood mononuclear cells of individuals aged ≥65 years to determine baseline transcriptomic signatures predictive of influenza vaccine response. Using two independent cohorts (discovery and validation), we classified individuals as triple responders (TRs) or triple non-responders (TNRs) based on hemagglutination inhibition assay titers at Day 0 and Day 28 for three components: A/H1N1, A/H3N2, and B/Yamagata. Results: We identified 1152 differentially expressed genes between TRs and TNRs at baseline. TRs exhibited enrichment of genes involved in B cell activation and protein synthesis, while TNRs showed enrichment of genes associated with innate immune responses and platelet activation. A response score derived from gene expression achieved high predictive accuracy in the discovery cohort (area under the curve [AUC] = 0.98). However, performance declined in the validation cohort (AUC = 0.69), and did not outperform clinical predictors, such as baseline titers, sex and vaccine dose. Conclusions: While baseline transcriptomic profiles may reveal mechanistic insights into vaccine responsiveness in the elderly, they offer limited predictive utility. Future work should prioritise higher-resolution or combined cell-specific approaches, such as single-cell RNA-sequencing or flow cytometry.