Monday, April 13, 2009

Study on Cotinine Levels Among NYC Adults Being Used to Implicate Importance of Brief Tobacco Smoke Exposure on Sidewalks and at Building Entrances

A study by the New York City Department of Health & Mental Hygiene that was published online ahead of print last week in the journal Nicotine & Tobacco Research is being used to implicate sidewalk and building entrance exposure to secondhand smoke as an important source of tobacco smoke exposure that may explain the finding of increased exposure among New York City residents (see: Ellis JA et al. Secondhand smoke exposure among nonsmokers nationally and in New York City. Nicotine & Tobacco Research 2009; doi: 10.1093/ntr/ntp021).

The study compared serum cotinine levels among a representative sample of New York City adults in the city's 2004 National Health and Nutrition Examination Survey (NYC-NHANES) with levels among a similar sample of U.S. adults in the national NHANES survey conducted in 2003-2004. Using a cutoff value of 0.05 ng/mL as defining secondhand smoke exposure, the paper reports that exposure among New York City adults was significantly higher than adults in the nation as a whole: 56.7% of New York City adults were exposed to secondhand smoke, compared to 44.9% nationally.

In a press release accompanying the paper, the study authors express shock at the finding that the prevalence of secondhand smoke exposure is much higher in New York City, despite its lower active smoking prevalence, and suggest that one reason for this increased exposure may exposure to high levels of secondhand smoke on sidewalks, at building entrances, and at bus stops and subway entrances.

According to the press release: "On sidewalks, including near building and subways entrances, non-smokers are still exposed to tobacco smoke unavoidably. The federal government has adopted regulations prohibiting smoking within 25 feet of federal building entrances to address complaints and health concerns; other cities and states have similar regulations for all buildings."

According to a New York Times article, health commissioner and study co-author Dr. Thomas Frieden was quoted as stating: "This is not what we expected. It is a shocking number."

Also according to the Times article: "Dr. Frieden suggested that New Yorkers are being exposed primarily through sidewalk contact with smokers, passing through crowds smoking outside doorways or waiting with smokers at bus stops...".

The Rest of the Story

There are two important aspects to the rest of the story.

First, the assertion that the primary explanation for the observed cotinine levels in nonsmokers is exposure through "sidewalk contact with smokers, passing through crowds smoking outside doorways or waiting with smokers at bus stops" is unsupported, either by empirical data or by the principles of toxicology.

The study did not collect data on exposure in the home versus other places and no questions were asked about exposure on sidewalks or other public places, so there is really no way that the study can draw conclusions about the magnitude and importance of exposure on sidewalks. Making an assertion like this has the appearance of trying to make an advocacy point, rather than sticking to the science and drawing solid conclusions.

Moreover, it would not make much sense for sidewalk exposure to be the primary source of the observed exposure because such exposure is so brief. While it may be true that levels of tobacco smoke in such places (sidewalks, bus stops, building entrances, etc.) can be high at times, exposure is transient. Since dose is a function not only of the concentration of the toxicant, but also of the duration of exposure, the overall contribution of these exposures to an individual's cotinine level is likely to be quite small.

The more likely explanation for the observed exposure is smoking that takes place indoors: in particular, in the home. The reason why the proportion of New York City residents with detectable cotinine levels was higher than the national figures despite a lower smoking prevalence in the city is most likely that when smoking does occur, it is more likely to happen inside the home, and more people are likely to be exposed. Housing density is much higher in New York City, there are many high-rise apartment buildings, and therefore it is much more difficult for smokers to smoke outside the home. Thus, for the number of smokers, there is probably disproportionate tobacco smoke exposure due to increased smoking inside the home.

What the study suggests to me is not that efforts must be taken to protect nonsmokers from tobacco smoke exposure on sidewalks, at building entrances, and at bus stops and subway entrances. Instead, the study suggests that there is considerable exposure to secondhand smoke inside the home.

One of the findings emphasized in the paper is that the highest proportion of exposure occurred among Asian Americans. It is not plausible to suggest Asian Americans spend more time on sidewalks and at building entrances and bus stops where smoking is occuring. Instead, the data suggest that the rates of smoking among Asian Americans in New York City are disproportionately high and that the higher exposure is due to more exposure to smoke by the people who Asian American nonsmokers are living with or spending time with. The smoking prevalence among Asian American men was 35.9%, higher than any racial/ethnic/gender subgroup and much higher than the overall smoking prevalence of 23.3%.

The second aspect to the rest of the story is that the study's overall conclusion is suspect because the comparison of very low cotinine levels in the two studies is sketchy. It turns out (if you read the fine print of the paper) that two different mass spectrometers were used with very different levels of detection for cotinine. The equipment used in the NYC NHANES had a level of detection of 0.05 ng/mL cotinine. The equipment used in NHANES was newer and had a level of detection of 0.015 ng/mL cotinine.

The definition of exposure to secondhand smoke in the New York City sample was any detectable cotinine in the serum. In the national sample, the definition of exposure was a cotinine level of greater than 0.05 ng/mL.

These are two very different ways of estimating the prevalence of exposure. In the New York City method, the cutoff is set at the absolute level of detection of the method. Thus, if any cotinine is detected, the individual is defined as exposed. However, in the national method, the level of detection is much lower (0.015 ng/mL), so there are many individuals with detectable cotinine levels who are not classified as exposed to secondhand smoke.

In addition, the blanks for the New York City sample averaged 0.018 ng/mL and were as high as 0.03 ng/mL, which seems a little on the high side (although I don't have comparable data from the CDC national lab).

My point is not that the overall cotinine analysis was in any way invalid; it is simply that a conclusion drawn based on an absolute cutoff at the extreme low end of the cotinine scale is suspect because of the very different sensitivities of the two methods (machines) used. To base the entire conclusion on a comparison of whether or not a sample contains nicotine that is above or below an absolute level of 0.05 ng/mL seems questionable when the limit of detection of one of the methods is 0.05 ng/mL.

I think it is very possible that more individuals in New York City were identified as being exposed to secondhand smoke because there was not the sensitivity to clearly define low cotinine levels and throw out levels that were below an absolute level of 0.05 ng/mL. You can't have confidence in the linearity at the low range of an assay when the cutoff you are using is the limit of detection for that assay!

Interestingly, the study authors were concerned enough about the differences in the two methods that they refused to compare geometric mean levels of cotinine between the two samples. I actually think that would have been a more meaningful assessment of any differences in exposure between the two populations because the central research question is not how many people have detectable levels of cotinine, but instead, what is the difference in the average exposure level?

Regardless of this limitation, I feel the study is still important because it highlights the fact that even with a comprehensive smoking ban, secondhand smoke exposure is still substantial. To me, this points to the importance of smoke exposure inside the home, especially for an urban population with a high housing density and high occupancy density within the home.

Ironically, I'm not sure that the most important (and solid) conclusion from the study is the one that is actually highlighted in the article (that many more people in New York City are exposed to secondhand smoke, likely due to more exposure on sidewalks and at building entrances, subway entrances, and bus stops).

No comments: