Abstract
Asbestos exposure and laryngeal cancer
I read with interest your editorial in the October issue, “Asbestos exposure and laryngeal cancer: Is there an association?” 1 I completely agree with the central point, that the simple association of two events does not prove that the first event causes the second. This point is important to understanding epidemiologic studies suggesting disease associations and highlights the timelessness of considering all of Koch's postulates and Hill's criteria of disease causality.
As to the specific question the authors raise regarding whether asbestos is a cause of laryngeal cancer, I share a concern that associations at a population/group level may be interpreted as proof that an individual's cancer has a specific cause. However, a critical monograph on the subject from the National Academies of Sciences was not referenced or reviewed. 2 This monograph represents the review and conclusions of an Institute of Medicine multidisciplinary committee of 12 distinguished members, backed by research staff as well as a separate and independent 12-person review committee. Neither I nor a member of my institution was a member of these committees.
The conclusion of the committee regarding laryngeal cancer was that despite “… the lack of confirmatory evidence from animal studies or documentation of fiber deposition in the larynx…evidence is sufficient to infer a causal relationship between asbestos exposure and laryngeal cancer.” In making this conclusion, the committee considered 35 cohort populations and 18 case-control studies. This work is a comprehensive review of this literature and a summary of the principal epidemiologic criteria supporting asbestos as a cause of laryngeal cancer, namely: (1) consistency of the epidemiologic studies, (2) biologic plausibility, (3) strength of association, (4) dose-response effect, and (5) effect modification.
We would hope the authors and readership will give this monograph careful review when considering the potential carcinogenic effects of asbestos at a population/group level while at the same time recognizing the limitations in drawing inferences about etiology for specific individuals.