The restoration or conservation of predators could reduce seroprevalences of certain diseases in prey if predation selectively removes animals exhibiting clinical signs. We assessed disease seroprevalences and blood parameters of 115 adult female elk (Cervus elaphus) wintering on the northern range of Yellowstone National Park [YNP] during 2000–2005 and compared them to data collected prior to wolf (Canis lupus) restoration (WR) in 1995 and to two other herds in Montana to assess this prediction. Blood parameters were generally within two standard deviations of the means observed in other Montana herds (Gravelly-Snowcrest [GS] and Garnet Mountain [GM]), but Yellowstone elk had higher seroprevalences of parainfluenza-3 virus (95% CI YNP  =  61.1–78.6, GS  =  30.3–46.5) and bovine-virus-diarrhea virus type 1 (95% CI YNP  =  15.9–31.9, GM  =  0). In comparisons between pre-wolf restoration [pre-WR] (i.e., prior to 1995) seroprevalences with those post-wolf restoration [post-WR] in Yellowstone, we found lower seroprevalences for some disease-causing agents post-wolf restoration (e.g., bovine-virus-diarrhea virus type-1 [95% CI pre-WR  =  73.1–86.3, post-WR  =  15.9–31.9] and bovine-respiratory syncytial virus [95% CI pre-WR  =  70.0–83.8, post-WR  =  0]), but similar (e.g., Brucella abortus [95% CI pre-WR  =  0–4.45, post-WR  =  0–4.74] and epizootic hemorrhagic disease virus [95% CI pre-WR  =  0, post-WR  =  0]) or higher for others (e.g., Anaplasma marginale [95% CI pre-WR  =  0, post-WR  =  18.5–38.7] and Leptospira spp. [95% CI pre-WR  =  0.5–6.5, post-WR  =  9.5–23.5]). Though we did not detect an overall strong predation effect through reduced disease seroprevalence using retrospective comparisons with sparse data, our reference values will facilitate future assessments of this issue.