Background: Evidence is accumulating that immune cells play a prominent role in pancreatic cancer etiology but prospective investigations are missing. Methods: We conducted a nested case–control study within the European Prospective Investigation into Cancer and Nutrition (EPIC) study with 502 pairs of incident pancreatic cancer cases and matched controls. Relative counts of circulating immune cells (neutrophils and lymphocyte sublineages: total CD3+, CD8+, CD4+, and FOXP3+ regulatory T cells (Tregs) relative to nucleated cells, (white blood cells) were measured by qRT-PCR. ORs with 95% confidence intervals were estimated using logistic regressions, modeling relative counts of immune cells on a continuous scale. Results: Neither relative counts of immune cell types taken individually, nor mutually adjusted for each other were associated with pancreatic cancer risks. However, in subgroup analyses by strata of lag-time, higher relative counts of Tregs and lower relative counts of CD8+ were significantly associated with an increased pancreatic cancer risks in participants diagnosed within the first 5 years of follow-up. Conclusions: These results might reflect reverse causation, due to higher relative counts of Tregs and lower counts of CD8+ cells among individuals with more advanced stages of latent pancreatic cancer, who are closer to the point of developing clinical manifest disease. Impact: We have shown, for the first time, that increased relative counts of regulatory T cells and lower relative counts of CD8+, cytotoxic T cells may be associated with pancreatic cancer risk or relatively late-stage tumor development. See related commentary by Michaud and Kelsey, p. 2176