The use of noncancer endpoints as a basis for establishing a reference concentration for formaldehyde

Published studies involving formaldehyde were selected for quality and relevance for determining whether noncancer endpoints could be used to derive a reference concentration for formaldehyde. Chamber studies provided the highest quality data for determining the presence of eye, nose, or throat irritation at a known level of formaldehyde. Some individuals begin to sense irritation at about 0.5 ppm, 5-20% report eye irritation at 0.5 to 1 ppm, and greater certainty for sensory irritation appears at 1 ppm or greater. These levels of formaldehyde do not appear to impact asthmatics even though these individuals are thought to be more sensitive to irritants. Mild, reversible changes in pulmonary function (forced expiratory volume at 1 s and midexpiratory flow) can occur in sensitized individuals at levels approaching 2 ppm. Studies in the manufacturing setting, while confounded by multiple exposures, provide useful information for setting boundaries for sensory irritation or changes in pulmonary function. Community surveys do not provide the specificity nor sensitivity needed to establish a reference concentration. Histological studies of the nasal mucosa suffer significant methodological and technological shortcomings in addition to issues commonly associated with the design of residential and workplace studies. Based on the review of chamber, community, and workplace studies of human exposures to formaldehyde, it is not possible to identify a specific no observed adverse effect level or lowest observed adverse effect level for formaldehyde. Ranges of exposures associated with acute sensory irritation can be derived and do include sensitive subpopulations. However, given the quality and variability of the data, human studies alone, especially those involving sensory irritation, are not adequate to serve as a reference concentration for estimating risk, or lack thereof, for a lifetime of exposure to formaldehyde. Alternative approaches, such as modeling cellular changes observed in animal studies, may be more useful for quantitative risk assessment of noncancer endpoints and should be used as an adjunct to interpreting human sensory studies.