Lung Cancer Risk Among Non-Smoking Firefighters
To the Editor:
Firefighters are exposed to numerous IARC Group 1 and 2A carcinogens1 and so investigators have searched for evidence of a group elevation in risk for lung cancer.2 The recent article “Lung cancer among firefighters: smoking adjusted risk estimates in a pooled analysis of case-control studies”3 confirms that in case-referent studies the group risk is not discernible against the risk conferred by smoking for all firefighters but does not answer the question of lung cancer risk among non-smoking firefighters.
Although an enormous study in the total number of subjects, the non-smoking subcohort consists only of 190 men firefighters in total, only two of which are cases. It also relies heavily on duration of employment as a firefighter as the sole indicator of exposure, although duration, being temporal, is heavily confounded by age, era of firefighting technology and protective equipment, composition of smoke, and prevalence of smoking habits both in the community and among firefighters. No analysis could be made of job assignment, which clarifies duration.
It is not surprising that among smoking firefighters the risk of lung cancer is driven by smoking, as the more frequent and repeated exposure, rather than exposure to fire smoke or diesel exhaust. Any excess risk of lung cancer from these exposures is predictably small compared with the much greater risk conferred by cigarette smoking. However, the studies on which the pooled estimates are based lack detail on firefighting assignment and era and judging from the high reported smoking rates appear skewed in distribution to subjects who entered the fire service before the modern era. There is abundant evidence for a biologically significant effect of firefighting in the cohort literature, including trends consistent with exposure–response relationships for job assignment (which requires knowledge of firefighting practices) and era (associated with use of protective equipment, composition of fire smoke, and smoking prevalence).2
The contribution of firefighting-related exposures to risk among non-smoking or minimally-smoking firefighters is probably not negligible, as implied. With such small numbers and consequent low power (demonstrated by confidence interval in Table 3 of over an order of magnitude), one can only tentatively extrapolate risk to non-smoking firefighters. On the other hand, using the method introduced by Fletcher for attributing lung cancer risk to smoking in occupational cohorts4 derives an estimate of relative risk for non-smokers of about three, approximate but clearly in excess of a doubling.5
This is not an insignificant additional risk for individuals who are at no or low risk from smoking. Non-smoking firefighters are much more common in the fire service today than in the past, in part due to secular trends and in part due to the introduction of wellness programs.
As a practical matter, the risk for this non-smoking group is also important in assessing causation for all firefighters in compensation programs in which the contribution of risk from occupation is compared with personal risk. Fair compensation is based on the civil legal standard of the weight of evidence (or “balance” of probabilities), not scientific certainty, which is difficult to prove and which under workers compensation and other legislation represents too high a bar. Indeed, some legal systems, such as the state of Pennsylvania, have adopted a standard other than strict causation, recognizing occupational exposures that “substantially contribute” to cancer risk. By any standard, more than doubling of a reference or community risk is a substantial contribution.