Thursday 15 December 2011

Anna Gilmore returns

Anna Gilmore's new study doesn't seemed to have generated much media attention, perhaps because any editor looking at the headline of the press release is likely to think "well, duh":

Smoke free legislation linked to drop in second-hand smoke exposure among adults

Levels of second-hand smoke exposure among non-smoking adults fell by almost 30 per cent after smoke free legislation was introduced in England in 2007, researchers in the Department for Health have found.

...Professor Anna Gilmore, who directed the study, said: “The importance of this study is that it examines the impacts of smoke free policies on adults’ exposure using a specific biological-marker of smoke exposure (rather than self-reported exposure) while simultaneously controlling for underlying declines in exposure.

“To our knowledge it is the first study to do this. The fact it shows marked declines in adult exposure provides further evidence of the important public health benefits of smoke-free policies.”

What, if any, are these "public health benefits"? The study looked at cotinine readings in nonsmokers before and after the English smoking ban and found that they fell by 27%. Cotinine itself is perfectly harmless, but it is a bio-marker for nicotine which is, in turn, is a proxy for "secondhand smoke exposure". Nothing wrong with that, nor is there anything surprising about cotinine readings falling as a result of a total smoking ban in 'public' places.*

The graph below (which comes from the study) shows saliva cotinine levels in nonsmokers before and after the ban (click, as ever, to enlarge).

'SFL' indicates the start of the 'SmokeFree Legislation'. What is most striking about this graph is how much cotinine and, it must be assumed, secondhand smoke exposure declined before the smoking ban. After the ban, cotinine levels did not change for people living in smoking households and Gilmore found that people in social classes IV and V experienced no reduction in secondhand smoke exposure at all (inevitably, this leads to her calling for "further efforts to reduce SHS exposure to benefit those who remain most exposed.")

Only nonsmokers from social classes I to III who live in nonsmoking households saw a decline in their saliva cotinine levels. Gilmore claims that this decline was greater than would be expected from the long-term trend, although none of her graphs appear to support this.

Gilmore's track record gives us no particular reason to trust her assertion that the smoking ban accelerated the existing trend towards less secondhand smoke exposure. However, it is obviously very plausible that a smoking ban would have this effect so, for the sake of argument, let us agree that there was a 27% drop in saliva cotinine readings as a result of the ban.

Her data show that before the ban (1998 to 2007), average cotinine levels in nonsmokers' saliva fell from 0.36 ng/ml to 0.14 ng/ml. After the ban, this declined continued and, by the end of 2008, cotinine levels were at 0.071 ng/ml.

The question is: so what? Is this decline—which sounds impressive when described as a 30% fall in secondhand smoke exposure—of any practical significance? How do these levels compare with actually smoking?

A systematic review of cotinine readings found that the average smoker has a saliva cotinine reading of 318 ng/ml. This is more than 2,250 times greater than levels found in nonsmokers before the ban.

This difference is so vast that it is difficult to show it visually. The graph below shows cotinine levels of nonsmokers in 1998, 2007 (pre-ban) and 2008 (post-ban) compared with a typical cotinine reading from a smoker. If you click to enlarge, you may just be able to see the nonsmokers' data.

While smokers have average cotinine readings of 318 ng/ml, the smoking ban reduced the average nonsmokers' levels by 0.0019 ng/ml. This is beyond negligible. Whether before or after the ban, we are talking about truly homeopathic levels of exposure. It takes a leap of faith to believe that reducing 'exposure' levels from 0.03% of a smoker's level to 0.02% of a smoker's level really constitutes "the most significant and beneficial public health intervention for a generation". Far from showing us how effective the smokefree legislation has been in tackling the passive smoking peril, this study reminds us how overhyped the peril was in the first place.

* However, it is wrong of the press release to describe a 27% fall in nonsmokers' cotinine levels as meaning that "second-hand smoke exposure among non-smoking adults fell by almost 30 per cent". This suggests that if all secondhand smoke was eliminated, cotinine levels would fall to zero. This would never happen because nicotine, and therefore cotinine, exists at low levels in various nightshade vegetables.


JJ said...

Smokers well-being

It would be nice for once if a study was carried out to examine how smokers feel about being completely ostracised from social society.

PJH said...

"Cotinine itself is perfectly harmless, but it is a bio-marker for nicotine which is, in turn, is a proxy for "secondhand smoke exposure"."

That's nice. It's also a bio-marker for potatoes, tomatoes, eggplants, and red peppers, since they're the same family of plants. And also nicotine replacement products (as mentioned in the wiki article.) Granted they claim to be able to tell the difference between SHS and tomatoes, but I can't seem to find any hard numbers for the level they deem to be SHS.

I note too that she comes up with the usual canard of:
"There is now a large body of evidence documenting the adverse effects of second-hand smoke exposure. In adults it is now known to be linked with coronary heart disease, lung and various other cancers, stroke, chronic respiratory symptoms and adverse pregnancy outcomes."

but when pushed, there doesn't seem to be anyone who's actually died of SHS. Indeed, (from the previous link) SHS was explicitly refused as the reason for one person's lung cancer due to, get this, lack of evidence.

Anonymous said...

I seem to recall a study on food born nicotine levels and they were in the range od .07 to 1.5 nanograms. Cant remember if that was blood or saliva. My point is the levels their citing are so low theyd be eaten up by food born nicotine levels in the test subjects!


PJH said...

Found one:

"In fact, another study showed that eating a normal (4.9 oz) portion of mashed potatoes produced the same amount of cotinine as 4 hours spent having drinks in a smoky bar. And similarly, a rather long lunch with a smoking friend can be easily equaled by a third of an ounce of eggplant.
-"Relevance Of Nicotine Content Of Common Vegetables To The Identification Of Passive Tobacco Smokers," Domino Et Al, Med. Sci. Res, 1993; 21, 571-2; Similarly, New England Jnl Med, Aug, 1993

Anonymous said...

According to your figures, the average non smoker cotinine level is equivalent to smoking around 3 cigarettes a year. As this includes both heavy "passive smokers" and those who have no contact with smoker, it seems consistent with the estimates of about 10 cigarettes a year being equivalent to spending time in a smoky bar each day. JB.

Anonymous said...

Frok the same journal - on thirdhand smoke

Anonymous said...

Gilmore is claiming that smoke exposure is directly proportional to cotine level. She has inadvertently admitted that the average non smoker was exposed to the equivalent of 3 cigarettes a year before the ban. This includes those who made no effort to avoid smoke. Words fail me.

Anonymous said...

Regulatory Toxicology and Pharmacology, Vol. 14, No. 1. (August 1991), pp. 88-105.

Environmental tobacco smoke (ETS) is derived from cigarette smoldering and active smoker exhalation. Its composition displays broad quantitative differences and redistributions between gas and respirable suspended particulate (RSP) phases when compared with the mainstream smoke (MSS) that smokers puff. This is because of different generation conditions and because ETS is diluted and ages vastly more than MSS. Such differences prevent a direct comparison of MSS and ETS and their biologic activities. However, even assuming similarities on an equal mass basis, ETS-RSP inhaled doses are estimated to be between 10,000- and 100,000-fold less than estimated average MSS-RSP doses for active smokers. Differences in effective gas phase doses are expected to be of similar magnitude. Thus the average person exposed to ETS would retain an annual dose analogous to the active MSS smoking of considerably less than one cigarette dispersed over a 1-year period. By contrast, consistent epidemiologic data indicate that active smoking of some 4–5 cigarettes per day may not be associated with a significantly increased risk of lung cancer.

Anonymous said...

In other words it wasnt going to haRM anyone to begin with, No wonder they had to say 'NO SAFE LEVEL'' LMAO!

Anonymous said...

Meant to add, just how close to nothing can you get to! nanogram is the smallest they can read anyway.

Hope I'm In The Firing Squad said...

A staggering achievement. And they only had to destroy society and wreck the economy to do it. Woo-hoo.

Anonymous said...

pg 110

n a study of 211 people in London, England, a plasma cutoff of 13.7 ng/mL provided an optimal classification with 94 percent sensitivity and 81 percent specificity based on self-reported exposure levels (Jarvis et al. 1987). The authors attributed the relatively poor specificity to “deception” in the self-reports of some participants with high serum cotinine levels. When the investigators reclassified those believed to be deceptive as smokers, sensitivities were 96 to 97 percent and specificities were 99 to 100 percent using plasma, saliva, or urine cotinine as the biomarker for comparison. The optimal cutoff values in this study were 14.2 ng/mL in saliva and 49.7 ng/mL in urine (Jarvis et al. 1987).
Pirkle and colleagues (1996) used a serum cotinine cutoff level of 15 ng/mL in a large U.S. epidemiologic study. They found a strong agreement with the self-reported nonsmoking status of the participants: those with serum cotinine levels above 15 ng/mL who claimed no tobacco use comprised only about 1.3 percent of the adult participants and 2.6 percent of the adolescents. Caraballo and colleagues (2001) examined the participants in this study aged 17 years
and older in detail and used the same nominal cutoff
of 15 ng/mL. There was a 92.5 percent agreement between serum cotinine concentrations and self-reported active smoking status and a 98.5 percent agreement among self-reported nonsmokers. The researchers regarded the infrequent or low rate of
cigarette use as an explanation for the disagreement with serum cotinine levels among self-reported smokers in most cases. However, there may have been some deception in the 1.5 percent with discrepant results between their serum cotinine levels and self-reported

Anonymous said...

Now we get to the heart of the matter,it was REPACE who went from 14 ng to confirm a non-smoker against food born nicotine to the 1 ng amount and it got put in the surgeon generals report of all things!

Cotinine Measurements as an Index
of Nicotine Exposure
Although the potential for overlap of levels always exists between nonsmokers with an extensive exposure to secondhand smoke and occasional or currently abstinent smokers, the use of cotinine measurements to separate smokers from nonsmokers provides a generally valid approach. Benowitz and colleagues (1983) originally proposed 10 ng/mL as a reasonable cutoff level for cotinine in serum to distinguish between smokers and nonsmokers. Consistent with that proposal, Repace and Lowrey (1993) estimated median serum cotinine levels to be about 1 ng/mL for U.S. adult nonsmokers and about 10 ng/mL for the most heavily exposed nonsmokers.

still page 110

Anonymous said...

Exposure to Second-Hand Smoke Widespread - 1996

"Nearly 9 out of 10 non-smoking Americans are exposed to environmental tobacco smoke (ETS, or second-hand smoke), as measured by the levels of cotinine in their blood, according to a study conducted by the Centers for Disease Control and Prevention (CDC).

"The data, reported by the CDC in this week's edition of the Journal of the American Medical Association, shows measurable levels of cotinine in the blood of 88 percent of all non-tobacco users.

The presence of cotinine, a chemical the body metabolizes from nicotine, is documentation that a person has been exposed to tobacco smoke."

Anonymous said...

When you look at the 1ng comparison as the cutoff point as per repace and the chart,the highest pre-ban reading was about 1.25 ng or 1.4 ng then a drop to about 1 ng in the smoking homes.

The levels only persist for about 2-4 hours then they went into half life comparisons. If the non-smoking homes were already within the food detectable range although repace trys to make it at .020ng from food born after some fancy dancing with figures. Why does he need to toss up 1 ng as the cut off, I gather becuase the food born amounts can always spike way above the 1ng mark from food.

Anonymous said...

But we are dealing with amounts so small,why would anyone stoop this low and set their own bio-markers so frigging low just to show an exposure. This is beyond insanity!

If this is what tobacco control basis their game plan on,how the hell did it ever get off the ground! Then youve got exposure bias in all the questionaire studies they did.... at what point does the media jump up and say ENUF!

Anonymous said...


To begin with, by far the most plentiful-- and measurable-- part of a cigarette (either first or secondhand) is simply: the nicotine. The body then metabolizes nicotine into cotinine. Smoking, however, isn't the only source of cotinine. The body also metabolizes foods into cotinine--tomatoes, potatoes, eggplant, and various teas.

In fact, another study showed that eating a normal (4.9 oz) portion of mashed potatoes produced the same amount of cotinine as 4 hours spent having drinks in a smoky bar. And similarly, a rather long lunch with a smoking friend can be easily equaled by a third of an ounce of eggplant.
-"Relevance Of Nicotine Content Of Common Vegetables To The Identification Of Passive Tobacco Smokers," Domino Et Al, Med. Sci. Res, 1993; 21, 571-2; Similarly, New England Jnl Med, Aug, 1993

Further, too, it's been noted that cotinine is only used as a "marker" (for exposure to ETS) because it's possible to measure. Even the California EPA has admitted, as most studies do, that the harmless "cotinine does not serve well as a marker for the presence of other tobacco smoke compounds."

Then too there's this:

"There exists between individuals a genetically determined variation (up to 50 fold) in the level of metabolism to cotinine from a given exposure. "
-Nyberg Et Al "Misclassification Of Smoking Status,.." Epidemiology, May 1997

And this:

Men metabolize cotinine faster than women, and whites faster than blacks (by 30%) This would also apply to the cotinine metabolized out of potatoes. And would indicate that either-- and/or-- women and blacks would register higher levels than their white / male counterparts to identical exposures. Further, eating a meal can affect metabolic speed.
-"Nicotine Metabolism Variables," Ahijevych Et Al/Agency For Health Care Policy & Research (HS07373)

Nonetheless, it's used as a marker.

Okay-- let's get a baseline. Again, according to the government (NIOSH Bulletin 54) a study by Cummings et al in 1990, showed the average concentration of urinary cotinine in a pack a day smoker is approximately 1,200 ng/ml.

Another study, conducted by the National Research Council in 1986, showed the average amount in a smoker was 1,825 ng/ml.

Another study, also reported by NIOSH, found that restaurant workers who themselves didn't smoke, had urine concentrations of 56 ng/ml. About 1/20th to 1/37th of the cotinine in a smoker.

And please remember that cotinine is a harmless metabolite and a rather poor indicator of anything else in smoke.

Nonetheless, in a recent study (Wortley et al, a study the Mayor likes) the blood levels of cotinine were carefully measured in 50 nonsmoking restaurant workers, for a mean serum level of 0.62 (ng/ml.) or 6/10th of a nanogram. This was higher than the levels recorded in, say, farmers (0.13)

Yes? And?

Anonymous said...

People who work in restaurants are exposed to more smoke than, say, a farmer who works in the field. Um...okay.

-Wortley Et Al, "Exposure To Secondhand Smoke In The Workplace: Serum Cotinine By Occupation," Jnl OCC & Env Med, 2002; 44 (6) (*) Personal Correspondence w/ Lead Author

Nor is this study a breakthrough in any way. Remember, the serum measured here is blood, not urine. According to a study sponsored by ALA, the blood cotinine levels in less than pack-a-day smokers ranged from 182- 249 ng/ml. That's a quantum difference from the 0.62 reported by Wortley among nonsmoking wait staff.
- Ahijevych Et Al, In "Addiction Behaviors, " Feb, 1999

And yet.

And yet. A study by James Repace measured the urine cotinine in waiters in Hong Kong and projected, solely on the basis of their cotinine, that 150 would die.
-"Passive Smoking And Risk For Heart Disease And Cancer," Hedley, Repace

But cotinine (have we mentioned?) is a harmless metabolite that indicates nothing except exposure to nicotine (or nightshade vegetables) But, hey, let's say it again:

"It is not known how levels of cotinine measured relate to the biologically important components of ETS other than nicotine." -Nyberg, Op Cit

Which leaves us with the general workplace studies--all epidemiological-- of which there aren't many. The largest American study ever done (and sponsored in part by The National Cancer Institute) "Passive Smoking and Lung Cancer in Nonsmoking Women," Brownson et al., 1992, reported:

"... no elevated lung cancer risk associated with passive exposure in the
-American Jnl Of Public Health, Nov. 1992

Alas, complained EPA, this study (published in November '92) arrived "too late" to be included in their (1993) report. Had it been so (the largest study ever done!) and been thrown into the hopper, it would have so tilted the weight that ETS could not, not even with lots of fudging, have been classified as a risk or a Class A carcinogen.

As of 1995, there were 8 extant studies that looked into the workplace/ lung cancer data. According to one set of researchers:

"An EPA-style meta-analysis...demonstrates no statistically significant
association for the data in its entirety--or for any sub-group of American workers. The pooled risk estimate [is] 1.0."
-"ETS Exposure, Lung Cancer And Heart Disease", Sears, Steichen, OSHA

The studies were pooled into that "EPA-style meta-analysis" because, the authors say, their results were so contradictory and mixed. Two showing a slightly increased risk; two showing a slightly decreased risk, and the rest indeterminate.

The same authors point out that "it is difficult to draw any conclusions concerning heart-disease risk in the US workplace because virtually no such information on the US workplace exists."

None of this, however, prevents the zealots from hollering Wolf or stops them from conjuring baseless death tolls, easily shattered by science and sense. Take, for example, this crowning lulu from Mayor Bloomberg of NYC.