We conducted a preliminary case-control investigation of the association of pancreatic polypeptide (PP) with mild cognitive impairment (MCI) in 202 MCI cases (mean age, 81.6 years) and 202 age- and sex-matched cognitively normal controls in the Mayo Clinic Study of Aging. Plasma PP was measured and examined as the natural logarithm or dichotomized at the median. The OR (95% CI) of MCI increased with increasing PP (1.46 [1.04-2.05]). There was a negative interaction of PP with APOE ε4 allele; compared to the reference group (no APOE ε4 allele and low PP), the OR (95% CI) for combinations of ε4 and PP were: 2.64 (1.39-5.04) for APOE ε4 plus low PP; 2.09 (1.27-3.45) for no APOE ε4 plus high PP; and 1.91 (1.04-3.53) for no APOE ε4 plus high PP (P for interaction = .017). There was also a trend toward a negative interaction with type 2 diabetes (P for interaction=.058). Compared to no diabetes and low PP, the OR (95% CI) was 3.02 (1.22-7.46) for low PP plus diabetes but 1.80 (1.01-3.22) for high PP plus diabetes. Participants with high PP had a greater mean weight loss (kilograms per decade) than persons with low PP (-2.26 [4.08] vs. -1.72 (5.28); P=.034). MCI cases had a non-significantly greater weight loss per decade compared to controls. These findings suggest that high PP alone or jointly with APOE e4 allele or type 2 diabetes is associated with MCI, and that high PP may mitigate some effects of APOE e4 allele and type 2 diabetes on cognition. Potential mechanisms may involve PP-related weight loss and centrally mediated effects of PP on cognition. These findings remain to be validated in other studies.