RESPONSE: Re: Mortality From Lymphohematopoietic Malignancies Among Workers in Formaldehyde Industries

Extract

We agree with Casanova et al. that our study ( 1 ) does not provide conclusive evidence of a causal association between formaldehyde exposure and leukemia. However, it is difficult to conceive how our findings of an increasing risk of lymphohematopoietic malignancies, especially leukemia, with increasing average intensity and peak levels of exposure to formaldehyde could be explained solely by bias due to imprecision of exposure metrics or uncontrolled confounding, in the absence of a causal association.

The availability of the peak exposure metric is a unique feature of our study, and peak exposure is the metric that best characterizes exposure patterns similar to those experienced by pathologists and embalmers ( 2 ), for whom increased leukemia mortality has been observed in several studies ( 3 – 5 ) . Estimates of peak exposure were based on the judgment of experts using information on job titles and tasks in combination with measurements of formaldehyde concentrations at selected workplaces. Uncertainties in estimating levels of peak exposure are unlikely to have induced the observed exposure–response gradient because the assessment was done before determining vital status and cause of death and was therefore unlikely to be differential with respect to disease outcome. We also observed increasing risks for average exposure intensity and duration of exposure, although not for cumulative exposure.