Secondhand Smoke-and-Mirrors

(How the Pros Tamper with Jury Pools)

Ichiro Kawachi, Graham A. Colditz, Frank E. Speizer, JoAnn E. Manson,
Meir J. Stampfer, Walter C. Willet, Charles H. Hennekens
Circulation 1997;95:2374-239

On June 2, 1997, jury selection starts in the Florida trial involving flight attendants suing tobacco companies for various health effects allegedly caused by secondhand smoke.

But Harvard University's Ichiro Kawachi et al. (funded by the U.S. Government which is at "war" with the tobacco industry), the (fiercely anti-tobacco) American Heart Association and Miami trial lawyer Stanley Rosenblatt (who stands to make a bundle representing tens of thousands nonsmoking flight attendants) are making sure prospective jurors make up their minds before hearing any evidence.

A scant two weeks before the trial begins, the American Heart Association published in its journal Circulation Kawachi et al.'s study of 32,000 nurses. The researchers claim that secondhand smoke increased the risk of heart attack by 91 percent. How convenient. And in the understatement of the year, attorney Rosenblatt said "'s very helpful to us."

But is it the substance of the study that's "very helpful"? Or just its timing and publicity?

First, the study relies on statistics, not science. Kawachi et al. claim to identify a 91 percent increase in risk of heart attack among those "regularly" exposed to secondhand smoke. But even the National Cancer Institute views such statistics with suspicion:

In epidemiologic research, [increases in risk of less than 100 percent] are considered small and are usually difficult to interpret. Such increases may be due to chance, statistical bias, or the effects of confounding factors that are sometimes not evident.[Source: National Cancer Institute, Press Release, October 26, 1994.]

Second, Kawachi et al. have NO data — i.e., NONE, NOTHING, ZERO, NADA, NIL — on how much secondhand smoke even one study subject was exposed to. Instead the study subjects were asked ONCE whether they were "never," "occasionally," or "regularly" exposed to secondhand smoke. Their responses were never verified. The researchers simply went on their word. Also, the responses were never updated. So the researchers assumed that someone "regularly" exposed to secondhand smoke when the question was first asked was ALWAYS "regularly" exposed throughout the study.

And what do you suppose the exposure categories mean? If I'm exposed to secondhand smoke every Monday is that "regular" (i.e., it happens every Monday) or "occasional" (i.e., it happens only once a week)? And who is "never" exposed to secondhand smoke? What bubble do they live in?

Also, the study subjects were supposed to be nonsmokers. But were they? Studies report that up to 15 percent of nonsmokers are actually —are you ready for this — SMOKERS! It's a phenomenon called "smoker misclassification." Many people are "closet smokers" who won't admit they smoke for insurance or social reasons. This could be particularly true with health professionals (like these nurse study subjects) who might fear admission of smoking would stigmatize them in the workplace. So because smoking status was not verified, Kawachi et al. may have studied a mixed bag of smokers and nonsmokers, not just nonsmokers as claimed.

Self-reported data is SO bad that we can't even have confidence that the reported heart attacks actually occurred. What!? According to a 1994 study published in the journal Epidemiology, self-reported history of heart attack may be horrendously unreliable. In that study, 40 percent of self-reports of heart attack were WRONG! [Source" Epidemiology 1995;6:67-69]. Kawachi et al. acknowledged that at least 16 percent of the claimed heart attacks in their study aren't definite.

Third, Kawachi et al. claimed to have considered a number of confounding risk factors for coronary heart disease, including histories of high blood pressure, diabetes, and hypercholesterolemia; menopausal status; current use of post-menopausal hormones, past use of oral contraceptives, vigorous exercise (defined as working up a sweat more than once per week); intake levels for alcohol, fat, vitamin E, and aspirin; parental history of heart attack; and father's occupation.

Sounds exhaustive, right? But once again, how much confidence can you have in unverified, self- reported data? For example, do you know what your intake of saturated fat is? What about vitamin E? Could you accurately remember your alcohol intake — even if you were willing to make that admission? What about high blood pressure being touted as the "silent killer." A study subject could have had high blood pressure, not have realized it, and reported having normal blood pressure.

Then what about defining "vigorous exercise" as working up a sweat more than once a week? Beneficial exercise is generally considered to involve raising the heart rate to certain target levels for at least 20 minutes, at least 3 times per week. I've never heard of "working up a sweat more than once a week" as being the "gold standard" of beneficial exercise.

Fourth, the study subjects were asked whether and how long they lived with smokers. But contrary to the authors' conclusion, no link was identified between coronary heart disease and how long study subjects lived with smokers — even though some lived with smokers for more than 30 years!

Fifth, there is no known mechanism by which secondhand smoke could cause heart attack or coronary heart disease. What's even odder is the alleged risk of heart disease from secondhand smoke is equivalent to the reported risk of heart disease from smoking. Huh? Somebody's smokin' somethin' and there ain't nothin' secondhand about it.

And last, but not least, why didn't the peer review process catch (or care about) these flaws? What do you expect given the study was peer reviewed under the auspices of the American Heart Association? The association is hardly a disinterested observer when it comes to tobacco issues.

But don't take my word that this is junk science. Take the word of Charles Hennekens, one of the study's authors:

Epidemiology is a crude and inexact science. Eighty percent of cases are almost all hypotheses. We tend to overstate findings either because we want attention or more grant money. [Source: New York Times, October 11, 1995]

So Chuck... is this study in the 80 percent part, or the 20 percent part? Are the odds 4:1 it's junk science?

But its the timing of the study that's giving me a coronary. This study was submitted to Circulation on October 1, 1996. It was put through the peer review process, revised, re-submitted and finally accepted for publication by December 14, 1996 — a very short timeframe, particularly given the Thanksgiving holiday. But then Circulation (I mean the American Heart Association) witheld the study from publication until May 19, 1997 — just two weeks before the flight attendant trial.

Why rush to accept the study, and then delay publication for five months? Just remember what attorney Rosenblatt stated "'s very helpful to us."

Material presented on this home page constitutes opinion of the author.

Copyright © 1997 Steven J. Milloy. All rights reserved. Site developed and hosted by WestLake Solutions, Inc.