Background: Many individuals in the community are prescribed psychoactive drugs with sedative effects. These drugs may affect their daily functions, of which automobile driving is a major component. Objective: To examine the association of three classes of commonly used psychoactive drugs (viz. benzodiazepines and newer non-benzodiazepine hypnotics, antidepressants and opioids) with (i) the risk of traffic accidents (as indexed by epidemiological indicators of risk); and (ii) driving performance (as indexed by experimental measures of driving performance). Methods: A literature search for material published in the English language between January 1966 and January 2010 in PubMed and EMBASE databases was combined with a search for other relevant material referenced in the retrieved articles. Retrieved articles were systematically reviewed, carrying out meta-analyses where possible. Twenty-one epidemiological studies (13 case-control and 8 cohort studies) fulfilled the inclusion criteria by estimating the accident risk associated with drug exposure (ascertained by blood/urine analysis or prescription records). Sixty-nine experimental studies fulfilled the inclusion criteria by testing actual or simulated driving performance after administering a single dose or multiple doses. Results: Two meta-analyses showed that benzodiazepines are associated with a 60% (for case-control studies: pooled odds ratio [OR] 1.59; 95% CI 1.10, 2.31) to 80% (for cohort studies: pooled incidence rate ratio 1.81; 95% CI 1.35, 2.43) increase in the risk of traffic accidents and a 40% (pooled OR 1.41; 95% CI 1.03, 1.94) increase in ‘accident responsibility’. Co-ingestion of benzodiazepines and alcohol was associated with a 7.7-fold increase in the accident risk (pooled OR 7.69; 95% CI 4.33, 13.65). Subgroup analysis of case-control studies showed a lower benzodiazepine-associated accident risk in elderly (>65 years of age) drivers (pooled OR 1.13; 95% CI 0.97, 1.31) than in drivers <65 years of age (pooled OR 2.21; 95% CI 1.31, 3.73), a result consistent with age-stratified risk differences reported in cohort studies. Anxiolytics, taken in single or multiple doses during the daytime, impaired driving performance independent of their half-lives. With hypnotics, converging evidence from experimental and epidemiological studies indicates that diazepam, flurazepam, flunitrazepam, nitrazepam and the short half-life non-benzodiazepine hypnotic zopiclone significantly impair driving, at least during the first 2–4 weeks of treatment. The accident risk was higher in the elderly (>65 years of age) who use tricyclic antidepressants (TCAs); however, the evidence for an association of antidepressants with accident risk in younger drivers was equivocal. Sedative but not non-sedative antidepressants were found to cause short-term impairment of several measures of driving performance. Limited epidemiological research reported that opioids may be associated with increased accident risk in the first few weeks of treatment. Conclusions: Benzodiazepine use was associated with a significant increase in the risk of traffic accidents and responsibility of drivers for accidents. The association was more pronounced in the younger drivers. The accident risk was markedly increased by co-ingestion of alcohol. Driving impairment was generally related to plasma half-lives of hypnotics, but with notable exceptions. Anxiolytics, with daytime dosing, impaired driving independent of their half-lives. TCAs appeared to be associated with increased accident risk, at least in the elderly, and caused short-term impairment in driving performance. Opioid users may be at a higher risk of traffic accidents; however, experimental evidence is limited on their effects on driving.