Up in Vapor: The real story behind the formaldehyde-cancer link
As a pathologist with 28 years’ formaldehyde exposure via workplace inhalation, I have both a professional and a personal interest in whether formaldehyde causes cancer. It turns out that the formaldehyde-cancer link is weak to nonexistent, and it has been grossly exaggerated by organizations like the International Agency for Research on Cancer (IARC) and the U.S. National Toxicology Program (NTP). In many ways, formaldehyde has been treated much like smokeless tobacco, which the same agencies list as a carcinogen, in spite of weak to nonexistent epidemiologic evidence.
The fact is, IARC and NTP primarily base their formaldehyde carcinogen classification on a single National Cancer Institute (NCI) study claiming that the agent causes nasopharyngeal cancer, or NPC. That study, published in 2004 by NCI’s Michael Hauptmann and colleagues, has been shown to be unreliable.
The Hauptmann study was based on 25,000 workers, most of whom were exposed to formaldehyde at 10 industrial plants over a 30-year period. Compared with unexposed workers, Hauptmann et al. reported that formaldehyde-exposed workers had about a twofold higher risk for NPC, based on eight cases. No other increased cancer risks were observed.
The Hauptmann data was reassessed in 2005 by Gary Marsh, a University of Pittsburgh statistician, who was supported by the Formaldehyde Council, an industry group.
Marsh reported that “reanalysis provided little evidence to support NCI’s suggestion of a causal association between formaldehyde exposure and mortality from NPC.” He pointed out that Hauptmann had based his conclusion on an excess of NPC in only one plant, located in Wallingford, Conn. In the other nine plants, formaldehyde-exposed workers had lower-than-expected NPC rates.
Marsh continued to investigate the NPC cancers at Wallingford. In 2007, he published another study showing the NPC excess at the Wallingford factory was probably not due to formaldehyde, but to workers’ exposure “to several suspected risk factors for upper respiratory system cancer (e.g., sulfuric acid mists, mineral acid, metal dusts and heat).” The NCI team also had claimed that the highest exposures to formaldehyde resulted in the highest NPC risks. Dr. Marsh showed that claim to be wrong in a third study published in 2007.
In 2010, Annette Bachand and colleagues conducted a comprehensive meta-analysis, which included “all relevant cohort and case-control studies published through May 2009.” They found no overall increase in NPC risk after excluding Hauptmann’s Wallingford plant data. Six case-control studies produced a marginally elevated risk for NPC among formaldehyde-exposed workers (OR = 1.22, 95 percent Confidence Interval = 1.00 – 1.50), which may have been due to a failure to adjust for smoking, a known risk factor for NPC. Seven cohort studies resulted in a lower NPC risk among exposed workers (OR = 0.72, CI = 0.40 – 1.29). Bachand’s results were in agreement with an earlier meta-analysis by Bosetti et al.
Inflated claims by NCI epidemiologists are not unprecedented; witness the 1981 assertion by Dr. Deborah Winn that all American smokeless tobacco users were at risk for mouth cancer (completely refuted here and here).
The only bright side of this story is that the NCI formaldehyde-cancer claim could be analyzed because the agency’s data was available to outside investigators. In contrast, the Karolinska Institute and the American Cancer Society have issued warnings about smokeless tobacco, but persist in withholding the underlying data (here and here) – a violation of data-sharing and research-practice standards.