Background: Herpes zoster (HZ) arises in older people due to age-related decline in immunity. We assessed whether HZ, as a marker of immune suppression, is associated with increased cancer risk. Methods: We conducted a case-control study in U.S. adults with ages ≥ 65 years using the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database. Cases (n = 1,108,986) were people with first cancers identified in cancer registries (1992-2005). Controls (n = 100,000) were cancer-free individuals frequency matched to cases on age, sex, and year of selection. We identified HZ diagnosis using Medicare claims. Logistic regression models were constructed to determine adjusted associations between cancer and HZ. Results: HZ prevalence was modestly higher in cases than controls (1.4% vs. 1.2%). We identified significant associations between HZ and oral cavity/pharyngeal [adjusted OR (aOR) = 1.21], colon (aOR = 1.10), lung (aOR = 1.11), and nonmelanoma skin (aOR = 1.46) cancers; myeloma (aOR = 1.38); diffuse large B-cell lymphoma (aOR = 1.30); lymphoplasmacytic lymphoma (aOR = 1.99); and chronic lymphocytic leukemia/small lymphocytic lymphoma (aOR = 1.55). Among solid cancers, HZ was mostly associated with regional and/or distant stage tumors. Associations were strongest when HZ was diagnosed 13 to 35 months before cancer diagnosis/selection; they were significant for some cancers in the 36 to 59 months period, and 60+ months for lymphoplasmacytic lymphoma (OR = 1.99). Conclusion: HZ is associated with modestly increased risk of a few cancers, particularly hematologic malignancies. Associations were strongest at short latency intervals for many cancers, and for regional/distant stages among solid cancers, perhaps reflecting reverse causality. Impact: Age-related immune decline does not play a major role in cancer development in older people, but it may be important for some lymphomas.