In Finnish data from 1995-2000, 1006 fatal poisonings due to alcohol (ethanol), a single drug or both were statistically analysed in retrospect to evaluate the interaction between alcohol and drugs. In 53% of these cases, low concentrations of some common benzodiazepines were present. The median postmortem blood alcohol concentration (BAC) was 3.3 percent per thousand (w/w) in the 615 alcohol poisonings, but significantly lower, ranging from 1.3 to 1.7 percent per thousand, when promazine, doxepin, amitriptyline or propoxyphene were found together with alcohol. When levomepromazine, temazepam or zopiclone were present, the median BAC was also significantly lower, 2.5-2.7 percent per thousand. Citalopram and diltiazem did not exhibit a significant effect. The median BAC was significantly lower in cases with high concentrations than in those with low concentrations of a drug (excluding citalopram), suggesting a positive concentration-effect relationship. Fatal toxicity indices (FTIs) were calculated by relating the number of deaths caused by a drug to the corresponding sales figures. Promazine had an extremely high FTI, followed by levomepromazine, propoxyphene, doxepin and amitriptyline. The other drugs had relatively low FTIs. The results reflect not only the acute toxicity of a given drug-alcohol combination but also the manners of use and abuse of these drugs.