In summary, most public exposure to secondhand smoke had been eliminated by 1993, in hospitals, the white collar workplace, schools, day care centers, airline flights under two hours, malls, etc, but our public health had not improved. Is it reasonable to expect elimination of secondhand smoke in the few remaining public places where smoking is allowed will suddenly make a big difference?
But what about all the studies???
Since about 1990, there has been a constant barrage of media announcements that yet another secondhand smoke study has concluded smoke is a health hazard. What is less well known is many other studies, often ignored, have reached opposite conclusions. I have identified some common features shared by studies which find smoke is a hazard, and by those which do not.
First, to understand what I am about to say, it must be recognized there are many causes of all medical conditions blamed on smoke exposure. Any responsible researcher must address the possibility the health effects being measured might come from another risk factor than smoke exposure.
Smokers’ characteristics do not fit the general population pattern in many ways.
1. They are one-third more likely to reside in an urban area. ( Behavioral Risk Factor Surveillance System)
2. they are half again as likely to earn under $30,000/ year ( Statistical Abstracts 1993, table 213)
3. They are twice as likely to be blue collar workers. (American J of Public Health, 2004)
Not only do smokers fit these patterns, but nonsmoking spouses and children of smokers will fit the same patterns of urban residency and low income. As is the case for smokers, coworkers of smokers also likely to be blue collar workers twice as often. The health risks associated with urban residency, and low income among nonsmokers and children of nonsmokers have to be subtracted from the health risks found among nonsmoking spouses and children of smokers before accurate estimations of health risk due to passive smoking can be measured. The same is true of coworker studies; that risks of two-fold excess blue collar employment must be considered before accurate measurements of the health impact of working with a smoker can be made.
As an example, research has clearly concluded urban residency carries with it an increased lung cancer risk, ACS Monograph #173 reported a study which found nonsmokers living in a city had lung cancer rates 1/4th as much as smokers living in rural areas. It is most notable, nonsmokers living in rural areas had lung cancer rates of zero. Another study found urban/rural elevations of lung cancer were 1.89, males, and 1.64 for females.
This urban residency issue is much more statistically compelling than the excess lung cancer risk the EPA found in it’s 1992 report, Respiratory Health Effects of Passive Smoking, of which I have a copy. The EPA found an excess risk of lung cancer due to passive smoking of 19%, while the excess urban risk of lung cancer in nonsmokers was infinite. Yet, after careful review, of the EPA report, I have found urban residency was not considered by the EPA when concluding secondhand smoke exposure is associated with a 19% increased risk of lung cancer.
If the data the EPA used is from studies representative of our overall population, then nonsmoking spouses of smokers were urban residents 1/3 more often, and taking the conclusions of C. Arden Pope, that urban residency increases lung cancer risk 30 – 50% into account, 40% X 1/3 = 13% of the 19% risk of living with a smoker found by the EPA could have actually been due to urban residency.
As another example, Pope found an 8-18% increased risk of heart disease mortality from each 10-ug/cu M of urban particles in urban air. Since this is a typical particulate amount, an average of 13% increased risk among urban residents with 1/3 more likelihood of living with a smoker would translate into 4% excess heart disease death risk deduced from secondhand smoke studies. For comparison, the current estimate by The American Heart Association is that secondhand smoke accounts for 5% of all heart disease deaths.
The EPA did not consider that the 1/3 excess urban residency of smokers puts their children at more risk from urban pollution too. Low income status is associated with an increased possibility of poor maintainence of the home, which can lead to radon and mold problems and these older homes are more likely to contain building materials composed of asbestos, and contain lead paint. In addition, the lower income status of smokers puts their children at more risk from the lack of health insurance common in this group. Without health insurance, parents are less likely to provide preventative medical care to their children. In summary, the circumstance that secondhand smoke exposure is now related to increased incidence of practically all medical conditions in nonsmoking spouses and children of smokers, indicates that these other more broad-based risks are likely the real causes.
The EPA did consider the possibility personal history of lung disease, family history of lung disease, heat sources, cooking with oil, occupation, and diet might influence the outcome of their spousal lung cancer analysis, but concluded none of these were sufficiently compelling to include in their analysis of the studies. This is surprising, considering other EPA researchers were finding serious lung cancer risks from urban pollution, asbestos, and radon which the EPA did not consider as alternate causes of lung cancer in assessing the passive smoking studies.
The Congressional Research Service of the Library of Congress has also severely criticized the conclusions the EPA drew in this report, and I have attached a copy of a March, 23, 1994 CRS report that finds such a high degree of uncertainity in the estimates made by the EPA, that their results are meaningless.
Yet the EPA report remains the one most often cited as justification for smoking bans by the CDC, and anti-smoking groups, despite the fact other lung cancer causes could easily explain the risk the EPA found.
Another study, conducted by the World Health Organization, and often cited as proof secondhand smoke causes lung cancer found an upside risk of 16% more lung cancer from spousal smoking and 17% more risk from occupational exposure. Yet they found a decreased risk of 22% from ETS exposure during childhood. It cannot be overemphasized the decreased risk found from childhood exposure characterizes the errors inheret in these kinds of studies. It is luderocus to assume childhood exposure to smoke has a protective effect. Otherwise one could argue smoking should be allowed in public places to protect children. It is much more likely the WHO study had measurement errors in excess of 22% and that since the risks found from spousal and occupational exposure were less than 22%, that the findings were within experimental error of zero. The WHO study also found no increased risk from secondhand smoke from other sources such as bar and restaurant exposure.
Another study often referred to as the Helena Heart Attack Study found heart attacks dropped 40% after a local smoking ban took effect and that heart attacks went back to former prevalence after the ban was lifted. This science is not predictable and reproducible. On Jan 1, 1995, the State of California banned all restaurant smoking. According to data from the California Dept. of Health, and posted on the state’s website, the states’ heart attack death total for 1994 was 68,312 ; and for 1995, 67,990. California was also conducting a hospital performance study during that time. Most hospitals were involved, and those hospitals reported 41,927 patients were admitted to their ER’s in 1994, and 42,183 were admitted in 1995. California’s population is 500 times larger than Helena’s. It is therefore 500 times more likely we will experience no fewer heart attacks if this ban passes.
A recent study, known as the UCLA study, conducted by Enstrom and Kabat, surveyed 118,094 adults enrolled in an American Cancer Society study begun in 1959, and focused on 35,561 never smokers who had a spouse with known smoking status. This was a study with a particularly large number of study subjects. Never-smoking females married to nonsmokers had the same incidence of lung cancer and heart disease as those married to smokers. Never-smoking males married to smokers had the same incidence of heart disease, but 25% less lung cancer. Again, as in the case of the WHO study, studies can produce negative results indicating measurement errors are in the realm of 25%. Most noteworthy, the study was initiated prior to the development of the social gradient between urban residency and economic status of smokers vs. nonsmokers. In addition, this study controlled for urban residency, underscoring the importance of controlling for other risk factors of heart and lung disease.
The authors made one other important discovery. Study funding is likely to be denied when results do not fit with preconceived notions of outcome. The authors stated in the report that the University of California withdrew funding after it became known results would be non-significant.
Dr. Givens, analyzed this phenomenon, and found by a bayesian statistical model that probably enough secondhand smoke studies were never published to explain an overestimation of the health effects of passive smoking by about 30%.
Other studies have attempted to estimate secondhand smoke exposure based on urinary metabolite markers and then estimate disease rates in nonsmokers by extrapolating disease rates from active smokers. Most research has concluded nicotine and it’s metabolites are the best markers and this is in part because other components in secondhand smoke are commonly present from other sources, such as food and urban pollution. The EPA report concluded nicotine was the best marker because many other potential markers are present from other sources, or were present in smoking environments at concentrations too low. However, Domino reported in the New England J. of Medicine nicotine was present in potatoes and tomatoes in sufficient amounts to falsely elevate nicotine in nonsmokers equivalent to three hours of heavy passive smoke exposure. Eating normal amounts of these foods, therefore completely confounds urinary marker tests. Apart from calling into question all conclusions concerning smoke exposure in nonsmokers, these circumstances indicate nonsmokers will still be exposed to all the components in smoke even if smoking is eliminated from indoor environments.
The best studies of smoke exposure in nonsmokers were carried out by OakRidge National Laboratory. Non-smoking workers at bars and restaurants wore air sampling pumps and it was found these workers were exposed to only one-fifth as much smoke as the eight-hour limit set by OSHA allows.
. In summary, most studies including those not addressed here, find excess heart disease and lung cancer risk of 20-30% in nonsmokers, but considering that other studies often generate negative results in the 20-30% range, the only reasonable conclusion is that measurement errors are as large as the study outcomes are. Other studies with negative results are briefly summarized below:
Brownson, 1992 reported many subsets with 70% risk. EPA 1992 reported: 81% risk Buffler 1984; 75% risk Chan1982; 86% risk Janerich 1990; 79% risk Kabat 1984; 74% risk Liu 1991; and 79% risk Wu-Williams 1990.
