than in controls [f(GG) = 1.2% vs. f(GG) = 2.6%; OR 2.3, 95% CI = 0.99–5.32]. After adjustment for age and sex, that association was no longer significant (P>0.1). As can be seen in Table 2, genotypes of the two other markers did not differ between cases and controls, nor did the allele frequencies of the three markers. Measures of LD and haplotype blocks across BCHE did not produce any evidence for an association with disease (data not shown). Results of the general tests of association for the preferred route of cocaine administration can be seen in Table 3. Of the three markers, only rs1803274 was associated with distinct genotypes among the subgroups of cocaine users. Among the a priori genetic models, the recessive model (genotype AA) best accounted for the significant difference between crack users and powder cocaine users (P = 0.027; OR = 4.36; 95% CI = 1.18–16.04), as well as for that observed between crack users and dual users (P = 0.001; OR = 5.83; 95% CI = 2.10–16.16). Those associations remained significant after adjustment for age and sex (data not shown).
