Recent reports on a potential association between the K-variant of the gene for butyrylcholinesterase (BCHE-K) and Alzheimer's disease (AD) are discordant. An initial finding of association through a synergistic enhancement of risk of APOE epsilon4 with late-onset AD has not been confirmed by others. We have conducted a case-control study of histopathologically confirmed AD (n=135) and non-AD (n=70) cases (age of death > or =60 years), in which we have genotyped for APOE epsilon4, BCHE-K, and BCHE-A1914G, a silent polymorphism 299 bp downstream of the BCHE-K mutation. The allelic frequency of BCHE-K was 0.13 in the controls and 0.23 in the AD cases, giving a carrier odds ratio (OR(c)) of 2.1 (95% C.I. 1.1-4.1) for BCHE-K in confirmed AD. The allelic frequency for the BCHE-1914G variant was 0.19 and 0.33 in controls and AD cases, respectively (OR(c)=2.4; 95% C.I. 1.3-4.5). In an older sub-sample of 27/70 controls and 89/135 AD patients with ages of death > or =75 years, the OR(c) was increased to 4.5 (95% C.I. 1.4-15) for BCHE-K and 2.7 (95% C.I. 1.0-7.2) for BCHE-1914G carriers. The BCHE-K association with AD became even stronger in carriers of at least one APOE epsilon4 allele. Only three out of 19 controls compared with 39/81 AD cases carried BCHE-K in addition to APOE epsilon4, giving an odds ratio of confirmed AD of 5.0 (95% C.I. 1.3-19) for BCHE-K carriers within APOE epsilon4 carriers. Five out of 19 controls and 52/81 AD cases carried BCHE-1914G, giving the same odds ratio of confirmed AD of 5.0 (95% C.I. 1.6-16) for BCHE-1914G carriers within APOE epsilon4 carriers. In addition, our results suggest strong linkage disequilibrium between BCHE-K and BCHE-1914G but no major association of the sole BCHE-1914G chromosome with AD. We conclude that BCHE through its K-variant, rather than a nearby marker, is a susceptibility factor for AD and enhances the AD risk defined by APOE epsilon4 alone in an age-dependent manner.