had no objective parameters of cocaine use in cases and controls, such as toxicological measurements of cocaine in hair or urine. The subjective ratings reported by cocaine users are known to be unreliable markers of the true amount and frequency of illegal substance use, and that holds true for self-reported drug use by non-dependent subjects, such as our controls [37]. Although we consider it unlikely, it is possible that our controls underreported cocaine use and dependence, which would have decreased the strength of the association found in the present study. Finally, it is possible that the BCHE SNP that is associated with vulnerability to crack use is not the actual causative SNP; instead, other nearby SNPs in LD could be the alternative causation.
