Single-cell transcriptomics from 5,079 adults of European, British-Pakistani and British-Bangladeshi ancestry resolved 26 circulating immune cell types and 62 transcriptional states. Age, sex, adiposity and smoking each reshaped immune composition, with modifiable exposures associated with advanced immune age. Common diseases converged on shared inflammatory myeloid, adaptive-memory and interferon programmes, cardiovascular therapies were linked to selective shifts in disease-associated states, and immune profiles split cardiometabolic patients into endotypes with distinct metabolic profiles.

The scale and ancestral diversity are real strengths, since most immune aging atlases draw on small, largely European samples. The adiposity and smoking findings fit a large literature on inflammaging, where excess fat tissue and tobacco push the immune system toward myeloid-skewed, memory-heavy states that resemble older age. The practical message is that immune age appears partly modifiable, though this analysis cannot show that losing weight or quitting smoking reverses it. The design is cross-sectional, so exposure and immune state are associated, not causally linked, and the drug findings are vulnerable to confounding by indication, because treated patients differ from untreated ones in disease severity. Hospital-linked cohorts may also over-represent unwell participants. Endotypes are interesting but have no demonstrated clinical use yet. This is a preprint that has not been peer reviewed, so details may shift. Overall it is a confirmatory, resource-style contribution: valuable for its diversity and depth, but not yet something that changes individual advice.