Lung cancer mortality among industrial workers exposed to formaldehyde: a Poisson regression analysis of the National...

The Formaldehyde Institute (FI) sponsored additional Poisson regression analysis of lung cancer mortality data from the joint National Cancer Institute (NCI)/FI cohort study of workers exposed to formaldehyde to investigate the previously reported effects of plant and latency period and to assess the impact of short-term workers (under 1 yr employment) on the results. There were 242 lung cancer deaths in this cohort of 20,067 white male workers. With OCMAP software, lung cancer death rates for the white males in this cohort were computed by plant, age, calendar time, and job type for several time-dependent formaldehyde exposures, including formaldehyde exposure in the presence of 12 selected co-exposures: ammonia (AM), antioxidants (AN), asbestos (AS), carbon black (CB), dyes/inks/pigments (DY), hexamethylenetetramine (HX), melamine (ME), particulates (PT), phenol (PH), plasticizers (PL), urea/urea compounds (UR), wood dust (WD), and a composite co-exposure (X5) involving AN, HX, ME, PH, and UR.A 1.6-fold increase in lung cancer risk was found, beginning approximately 16-20 yr after first employment in the study plants with no evidence of a differential effect of latency between hourly and salaried workers or among the various categories of formaldehyde exposure as measured by cumulative average intensity or length of exposure. The statistically significant heterogeneity in lung cancer risk among the 10 plants could not be explained by interplant differences in cumulative or average intensity of exposure to formaldehyde, either without regard to co-exposures or in the presence of any of the 12 co-exposures considered individually. Plant was not a statistically significant predictor of lung cancer risk when cumulative exposure to the composite X5 was included in the model, suggesting that some component of X5, or a correlate, could at least partly account for the overall heterogeneity. No significant associations were found for cumulative, average, or length of exposure to formaldehyde without regard to co-exposure, but positive associations were found for cumulative exposure to formaldehyde in the presence of several co-exposures (AN, HX, ME, PH, and UR). For workers who were never exposed to any of 10 co-exposures associated with an increased lung cancer risk, there was a decreasing pattern of estimated lung cancer risk ratios relative to cumulative formaldehyde exposure. Similar patterns were seen when the analysis was restricted to the long-term workers. Analysis of the internal cohort rates corroborates previous analyses of NCI/FI cohort data in that significant positive associations were found between the risk of lung cancer and cumulative exposure to formaldehyde in the presence of several of the same co-exposures. No such associations were found in the absence of these co-exposures.