Background:
Protection from second-hand smoke (SHS) is one of the fundamental principles of the World Health Organization Framework Convention for Tobacco Control. Scarce data are available on SHS exposure in private areas, such as vehicles. This study aimed to estimate prevalence of smoking inside vehicles and assess the consequent levels of exposure.

Methods:
Levels of fine respirable particles with a diameter of 2.5 micrometres or less (PM_2.5) were measured in five experimental models in smokers' and non-smokers' vehicles. Prevalence of smoking in vehicles was estimated in two steps: direct observation of vehicles in different socio economic status areas, and further correction for detectability.

Results:
Median PM_2.5 concentration was 181 µg/m3 in “smoking vehicles” and 0 µg/m3 in “non-smoking vehicles” (p˂0.001). The highest concentration reached 2900 µg/m3 in a parked car with driver's window partially open. We observed 10,011 vehicles. In 219 (2.2%; 95% confidence interval, 1.91-2.49) of them, smoking was observed, and in 29.2% of these, another person was exposed to SHS. According to the multiplying factor we constructed, direct observation detected one of six to nine vehicles in which smoking occurred. The observed prevalence of smoking in vehicles (2.2%) could reflect a real prevalence between 12% and 19%.

Conclusions:
Smoking was found to occur in 12% to 19% of vehicles, with involuntary exposure in one of three. Concentration of particular matter in vehicles reached very high levels, similar to those at certain sites in countries with weak tobacco control policies. These facts underscore a need for new public policies to eliminate SHS in vehicles to protect public health.