OBJECTIVES: To to study the risk of cancer, particularly of lung cancer and bladder cancer, among workers in the printing industry according to different occupations. METHODS: This is a population based retrospective cohort study. The cohort comprised 1332 men and 426 women employed in the printing industry in Iceland according to a published union registry. A computerised file of the cohort was record linked to the Cancer Registry by making use of personal identification numbers. Expected numbers of cases of cancer were calculated on the basis of number of person-years and specific incidences of cancer sites for men and women provided by the Cancer Registry. RESULTS: Among the men (36 217.5 person-years at risk) there were 125 observed cancers versus 123.66 expected, standardised incidence ratio (SIR) 1.01, 95% confidence interval (95% CI) was 0.84 to 1.20. The SIR (95% CI) for liver cancer was 1.97 (0.55 to 5.20) and the SIR for non-Hodgkin's lymphoma was 2.26 (0.96 to 4.41). No excess risk for cancer was found among women (8631.0 person-years at risk). The SIR (95% CI) for liver cancer was 4.21 (0.47 to 15.20) and for non-Hodgkin's lymphoma it was 4.99 (1.61 to 11.63) among the typesetters. A survey on smoking habits among active and retired union members showed that they smoked less than a random sample of the general population. CONCLUSION: The cancer site most often reported to show excess risk among printing industry workers has been the lung and the urinary bladder; however, this was not found in the present study. This may be explained by difference in smoking habits among union members compared with the general population. There is a high occurrence of non-Hodgkin's lymphoma, particularly among typesetters, which warrants further studies.