This study was undertaken to determine the risk of lung cancer associated with medium, low, and very low tar filter cigarettes. Using a cohort of patients from the US cancer prevention study II (CPS-II), 364 239 men (100 868 current smokers) and 576 535 women (124 270 current smokers) were studied. The primary outcome measure was death from tracheal or lung cancer over 6 years of follow up. Tar rating per cigarette brand was designated high (⩾22 mg), medium (15–21 mg), low (8–14 mg), or very low (⩽7 mg). Assessment was made using Cox proportional hazards analysis and adjustment made for multiple covariates.

The risk of lung cancer was increased in current smokers, with high tar cigarette smokers experiencing the highest risk. No difference was evident between the medium, low, and very low tar brands. These results withstood sensitivity analysis, including exclusion of people with smoking related diseases. The study results challenge the assumption of a linear association between tar rating and lung cancer risk. “Compensatory smoking” in those smoking lower tar brands (deeper inhalation with increased puff volume, longer duration of smoke retention, and a greater total number of cigarettes smoked) and the poor correlation between tar yield and carcinogen delivery offer possible explanations. Cessation should be advised in all smokers and patients informed that low tar cigarettes are not a safer alternative.