The effect of tobacco smoking on the frequency of micronuclei (MN) in human lymphocytes has been the object of many population studies. In most reports, the results were unexpectedly negative, and in many instances smokers had lower frequencies of MN than non-smokers. A pooled re-analysis of 24 databases from the HUMN international collaborative project has been performed with the aim of understanding the impact of smoking habits on MN frequency. The complete database included 5710 subjects, with 3501 non-smokers, 1409 current smokers, and 800 former smokers, among subjects in occupational and environmental surveys. The overall result of the re-analysis confirmed the small decrease of MN frequencies in current smokers (frequency ratio (FR) = 0.97, 95% confidence interval (CI) = 0.93-1.01) and in former smokers (FR = 0.96, 95% CI = 0.91-1.01), when compared to non-smokers. MN frequency was not influenced by the number of cigarettes smoked per day among subjects occupationally exposed to genotoxic agents, whereas a typical U-shaped curve is observed for non-exposed smokers, showing a significant increase of MN frequency in individuals smoking 30 cigarettes or more per day (FR = 1.59, 95% CI = 1.35-1.88). This analysis confirmed that smokers do not experience an overall increase in MN frequency, although when the interaction with occupational exposure is taken into account, heavy smokers were the only group showing a significant increase in genotoxic damage as measured by the micronucleus assay in lymphocytes. From these results some general recommendations for the design of biomonitoring studies involving smokers can be formulated. Quantitative data about smoking habit should always be collected because, in the absence of such data, the simple comparison of smokers versus non-smokers could be misleading. The sub-group of heavy smokers (> or =30 cigarettes per day) should be specifically evaluated whenever it is large enough to satisfy statistical requirements. The presence of an interaction between smoking habit and occupational exposure to genotoxic agents should be always tested.