July 8, 2008
by Jerome Arnett, Jr., M.D.
Exposure to secondhand smoke (SHS) is an unpleasant experience for many nonsmokers, and for decades was considered a nuisance. But the idea that it might actually cause disease in nonsmokers has been around only since the 1970s.
Recent surveys show more than 80 percent of Americans now believe secondhand smoke is harmful to nonsmokers.
Federal Government Reports
A 1972 U.S. surgeon general’s report first addressed passive smoking as a possible threat to nonsmokers and called for an anti-smoking movement. The issue was addressed again in surgeon generals’ reports in 1979, 1982, and 1984.
A 1986 surgeon general’s report concluded involuntary smoking caused lung cancer, but it offered only weak epidemiological evidence to support the claim. In 1989 the Environmental Protection Agency (EPA) was charged with further evaluating the evidence for health effects of SHS.
In 1992 EPA published its report, “Respiratory Health Effects of Passive Smoking,” claiming SHS is a serious public health problem, that it kills approximately 3,000 nonsmoking Americans each year from lung cancer, and that it is a Group A carcinogen (like benzene, asbestos, and radon).
The report has been used by the tobacco-control movement and government agencies, including public health departments, to justify the imposition of thousands of indoor smoking bans in public places.
EPA’s 1992 conclusions are not supported by reliable scientific evidence. The report has been largely discredited and, in 1998, was legally vacated by a federal judge.
Even so, the EPA report was cited in the surgeon general’s 2006 report on SHS, where then-Surgeon General Richard Carmona made the absurd claim that there is no risk-free level of exposure to SHS.
For its 1992 report, EPA arbitrarily chose to equate SHS with mainstream (or firsthand) smoke. One of the agency’s stated assumptions was that because there is an association between active smoking and lung cancer, there also must be a similar association between SHS and lung cancer.
But the problem posed by SHS is entirely different from that found with mainstream smoke. A well-recognized toxicological principle states, “The dose makes the poison.”
Accordingly, we physicians record direct exposure to cigarette smoke by smokers in the medical record as “pack-years smoked” (packs smoked per day times the number of years smoked). A smoking history of around 10 pack-years alerts the physician to search for cigarette-caused illness. But even those nonsmokers with the greatest exposure to SHS probably inhale the equivalent of only a small fraction (around 0.03) of one cigarette per day, which is equivalent to smoking around 10 cigarettes per year.
Low Statistical Association
Another major problem is that the epidemiological studies on which the EPA report is based are statistical studies that can show only correlation and cannot prove causation.
One statistical method used to compare the rates of a disease in two populations is relative risk (RR). It is the rate of disease found in the exposed population divided by the rate found in the unexposed population. An RR of 1.0 represents zero increased risk. Because confounding and other factors can obscure a weak association, in order even to suggest causation a very strong association must be found, on the order of at least 300 percent to 400 percent, which is an RR of 3.0 to 4.0.
For example, the studies linking direct cigarette smoking with lung cancer found an incidence in smokers of 20 to around 40 times that in nonsmokers, an association of 2000 percent to 4000 percent, or an RR of 20.0 to 40.0.
Scientific Principles Ignored
An even greater problem is the agency’s lowering of the confidence interval (CI) used in its report. Epidemiologists calculate confidence intervals to express the likelihood a result could happen just by chance. A CI of 95 percent allows a 5 percent possibility that the results occurred only by chance.
Before its 1992 report, EPA had always used epidemiology’s gold standard CI of 95 percent to measure statistical significance. But because the U.S. studies chosen for the report were not statistically significant within a 95 percent CI, for the first time in its history EPA changed the rules and used a 90 percent CI, which doubled the chance of being wrong.
This allowed it to report a statistically significant 19 percent increase of lung cancer cases in the nonsmoking spouses of smokers over those cases found in nonsmoking spouses of nonsmokers. Even though the RR was only 1.19–an amount far short of what is normally required to demonstrate correlation or causality–the agency concluded this was proof SHS increased the risk of U.S. nonsmokers developing lung cancer by 19 percent.
EPA Study Soundly Rejected
In November 1995 after a 20-month study, the Congressional Research Service released a detailed analysis of the EPA report that was highly critical of EPA’s methods and conclusions. In 1998, in a devastating 92-page opinion, Federal Judge William Osteen vacated the EPA study, declaring it null and void. He found a culture of arrogance, deception, and cover-up at the agency.
Osteen noted, “First, there is evidence in the record supporting the accusation that EPA ‘cherry picked’ its data. … In order to confirm its hypothesis, EPA maintained its standard significance level but lowered the confidence interval to 90 percent. This allowed EPA to confirm its hypothesis by finding a relative risk of 1.19, albeit a very weak association. … EPA cannot show a statistically significant association between [SHS] and lung cancer.”
The judge added, “EPA publicly committed to a conclusion before the research had begun; adjusted established procedure and scientific norms to validate its conclusion; and aggressively utilized its authority to disseminate findings to establish a de facto regulatory scheme to influence public opinion.”
In 2003 a definitive paper on SHS and lung cancer mortality was published in the British Medical Journal. It is the largest and most detailed study ever reported. The authors studied more than 35,000 California never-smokers over a 39-year period and found no statistically significant association between exposure to SHS and lung cancer mortality.
Propaganda Trumps Science
The 1992 EPA report is an example of the use of epidemiology to promote belief in an epidemic instead of to investigate one. It has damaged the credibility of EPA and has tainted the fields of epidemiology and public health.
In addition, influential anti-tobacco activists, including prominent academics, have unethically attacked the research of eminent scientists in order to further their ideological and political agendas.
The abuse of scientific integrity and the generation of faulty “scientific” outcomes (through the use of pseudoscience) have led to the deception of the American public on a grand scale and to draconian government overregulation and the squandering of public money.
Millions of dollars have been spent promoting belief in SHS as a killer, and more millions of dollars have been spent by businesses in order to comply with thousands of highly restrictive bans, while personal choice and freedom have been denied to millions of smokers. Finally, and perhaps most tragically, all this has diverted resources away from discovering the true cause(s) of lung cancer in nonsmokers.
Dr. Jerome Arnett Jr. (firstname.lastname@example.org) is a pulmonologist who lives in Helvetia, West Virginia.
James E. Enstrom and Geoffrey C. Kabat, “Environmental tobacco smoke and tobacco related mortality in a prospective study of Californians, 1960-98,” British Medical Journal, May 2003: .