A collaborative case-control study was conducted in France in order to determine the prevalence of alcohol, cannabinoids, opiates, cocaine metabolites, amphetamines and therapeutic psychoactive drugs in blood samples from drivers injured in road accidents and to compare these values with those of a control population. Recruitment was performed in emergency departments of six university or general hospitals and comprised 900 drivers involved in a non-fatal accident and 900 patients (controls) who attended the same emergency units for a non-traumatic reason. Drivers and controls were matched by sex and age. Alcohol was determined by flame ionization-gas chromatography, drugs of abuse (DOA) by gas chromatography-mass spectrometry with the same analytical procedures in the six laboratories, and medicines by high performance liquid chromatography with diode array detection. Blood alcohol concentration exceeding 0.5 g/l (i.e. the legal French threshold) was found in 26% of drivers and 9% of controls. In the 18-27 years age range, alcohol was the only toxic found in blood samples of 17% drivers and 5% controls, leading to an odds-ratio (OR) of 3.8. A significant relationship was found between alcohol blood concentrations and OR values. All age groups confounded, the main active substance of cannabis, Delta(9) tetrahydrocannabinol (THC), was found in 10% of drivers and 5% of controls. In the less than 27 years old, THC (>1 ng/ml) was detected alone in the blood of 15.3% drivers and of 6.7% controls, giving OR=2.5, whereas there was no link between THC blood concentrations and OR value. THC was found alone in 60% of cases and associated with alcohol in 32%, with OR=4.6 between drivers and controls for this association. The difference in morphine prevalence between drivers (2.7%) and controls (0.03%) was highly significant (P<0.001), with OR=8.2. The number of positive cases for amphetamines and cocaine metabolites was too low for reaching any interpretation. The most frequently observed psychoactive therapeutic drugs were by far benzodiazepines, that were found alone in 9.4% of drivers and 5.8% of controls, which led to OR=1.7 (P<0.01). This study demonstrates a higher prevalence of opiates, alcohol, cannabinoids and the combination of these last two compounds in blood samples from drivers involved in road accidents than in those from controls, which suggests a causal role for these compounds in road crashes.