March 1998

Introduction

It is well established that the testing system currently used in Canada to determine tar, nicotine and carbon monoxide levels is outdated and flawed. The resulting information communicated to consumers is inaccurate and misleads them about the nature and level of toxic substances in cigarettes.

There is, however, less consensus on how the system might be changed, and how best to inform consumers about emissions so they can make decisions about their smoking behaviour that will reduce risk.

This paper discusses the specific problems that need to be addressed, and comments on what solutions may or may not be desirable.

Policy Objective

The over-riding policy objective should be to ensure that smokers understand the risk level of smoking and how their behaviour affects that risk. In particular, smokers should understand:

• the ingredients and quantities of ingredients in cigarettes and cigarette smoke;
• the health effects of those ingredients;
• how their smoking behaviour affects risk; and
• how they can change smoking behaviour to reduce risk of harm (including knowing the benefits of quitting relative to changing the way that they smoke).

Principles to Guide Change

1. The system must be looked at comprehensively.

Changes to the testing and reporting system to emissions should ideally occur as a comprehensive review that looks at the test method, cigarette design, how emissions are reported, which emissions are tested and reported, and related labelling issues such as descriptive brand names.

For example, there is limited value in revamping the test system if manufacturers are allowed to continue to undermine or negate emissions information by labelling their products “light” and “ultra light.”

Policy measures should not be limited to addressing the current “toxic trio” of tar, nicotine and carbon monoxide. More substances can be measured and reported. “Tar” contains separate toxins which can and should be reported separately. More detail can be provided on the nature and impact of substances. Eg. “Nicotine - the addictive agent in tobacco.” “Benzene - causes cancer in humans.”

2. The system should emphasize that smoking behaviour, not the cigarette brand, is the primary determinant of amount of emissions inhaled.

A concept that desperately needs to be communicated is that there are not “light” cigarettes, but there may be “light” smokers. In other words, switching down will only have a potential benefit if there is no compensatory smoking behaviour.

The challenge here is that (a) it may be difficult for smokers to know if and how much they are compensating and, (b) even if they are sure they are not compensating, it is not clear how significant the reduced health risks, if any, will be.

In particular, if smokers use either “switching down” or perceived changes in the way they smoke as an alternative to quitting or significantly reducing consumption, the opportunity cost of not quitting may outweigh the much smaller benefit of other behaviour change.

3. The primary message communicated should be that there is no change in smoking behaviour that offers comparable health benefits to quitting smoking.

4. Reducing the number of brands on the market is a legitimate health goal.

Real consumer choice is desirable, and should be pursued for smokers by offering them safer methods of getting nicotine either as aids to quitting or as alternatives to smoking.

However, the choices currently offered to smokers through 180+ cigarette brands are largely illusory. Extended brand families, which offer everything from an unfiltered nicotine hit down to a seemingly more benign “ultra light” cigarette, provide ample rationalization for smokers to continue smoking and calm their health fears.

This false choice misleads smokers in a way that translates into fewer quitters and more deaths. Moving toward a system that eliminates non-beneficial product choice is a legitimate goal of health policy.

A plethora of brand choice also offers up dozens of appealing images to new smokers and increases shelf space and brand recognition. Fewer brand choices could be expected to lead to fewer starters. Think of automobiles. Some people own cars because they need them. But many others own cars because, like cigarette brands, they offer an image. If drivers could only have Ladas -- not Jeeps, Lincolns, Miatas or Accords -- there would, in all likelihood, be fewer cars on the road.

The proposals described below may result in fewer brands for a number of reasons. Improved emissions testing and the removal of descriptive terms will diminish the ability of companies to distinguish between brands. In addition, the cost associated with expanded emissions testing may be prohibitive for brands with small brand share and place pressure on their manufacturers to remove them from the market.

Current Policies

The logical place to begin in considering policy options is to look at systems for testing and reporting currently in place. Three are discussed here: the current ISO method in place in Canada; the proposed changes to the FTC system in the U.S.; and the method recently implemented in Massachusetts.

1. The ISO (International Organization for Standardization) Method (Canada)

Testing

The ISO testing system in place in Canada is used in many countries around the world and is similar to the FTC system used in the U.S.

The system suffers from two major flaws. First, it measures tar, nicotine and carbon monoxide yields under one standardized parameter which does not accurately mimic the actual behaviour of many or even most smokers. Therefore, the absolute yields do not represent those actually inhaled by smokers.

Second, the system does not in any way account for the compensatory behaviour commonly found in smokers of so-called low-tar cigarettes. Lower yields in machine tests are achieved largely through tiny ventilation holes in the cigarette filter. The smoking machines leave the holes open. Smokers, however, often cover the ventilation holes wholly or partially with their fingers or lips, resulting in much higher yields. In many cases, covering the holes may result in a low-tar cigarette delivering as much or more tar as a regular cigarette.

Smokers also compensate in other ways -- inhaling more deeply, taking more puffs, smoking more of each cigarette, smoking more cigarettes -- none of which are accounted for by machine testing.

Reporting

Smokers in Canada learn about the yields of cigarettes from the package. Until 1997, regulations authorized under the Tobacco Products Control Act required tar, nicotine and carbon monoxide yields to be printed on the side of the package in bold lettering under the heading “Toxic constituents (Average).” A single value in milligrams was listed for each of the three substances.

While these legal requirements are currently in limbo(2), most manufacturers continue to adhere to this general format.

Because the yields are reported as absolute measures, without yield ranges and without an explanation of how smoking behaviour determines yield, consumers largely take the yields at face value.

Smokers are not told and generally do not know that the way they smoke determines the yield they get. They also are not told that the way they smoke can completely wipe out any reduced yield they think they are getting from a lower tar cigarette.

2. The Proposed FTC (Federal Trade Commission) Method (United States)

The FTC recently (Sept. 1997) proposed changes, currently under review, to its testing and reporting system for tar and nicotine emissions.(3)

The current FTC system suffers from the same flaws as the ISO system. The proposed system makes marginal improvements, but retains significant problems.

Testing

The FTC tests levels of tar, nicotine and carbon monoxide, but only tar and nicotine levels are reported to the consumer. The FTC has proposed to test yields under two standardized parameters. One would reflect what the FTC calls “less intensive smoking behaviour,” the other “more intensive smoking behaviour.” The variations under the two parameters would be accomplished by varying the volume and the frequency of puffs.

While this change is a marginal improvement in that it implicitly suggests that yields are not absolute but are dependent on smoking behaviour, it fails to account for some of the more common compensatory behaviour, such as hole blocking. In failing to do so, the proposed system will still result in yield levels (albeit yield ranges) that suggest a higher range for regular cigarettes, and a lower range for low-tar cigarettes. And smokers will still be left with the impression that lower tar brands deliver relatively less tar and nicotine than regular brands.

This hypothetical representation illustrates the problem:

Despite the range, Brand Light will still be viewed as the lower-yield, and thus in some sense safer, cigarette.

Reporting

Tobacco companies have never been required by law to list tar and nicotine yields on tobacco packages. When they do, it is generally because there is a marketing advantage. (Brands that have gained market share as low-tar cigarettes often print the yields; regular brands rarely do.)(4)

Tobacco companies are required to print tar and nicotine yields (not carbon monoxide) on tobacco ads. The yields generally appear in very small print at the bottom of ads. As in Canada, a single yield value is reported as follows: “17 mg ‘tar,’ 1.2 mg nicotine av. by FTC method.”

The proposed changes would not require reporting on the package. The FTC proposes a consumer education campaign to inform smokers of the changes to the system and about how smoking behaviour affects yield. It does not, however, detail the nature of the campaign and how it proposes to reach all smokers.

The FTC proposes to change the required text on cigarette ads to carry one of the following two warnings:

2 mg. - 6 mg. tar, 0.2 mg. - 0.6 mg. Nicotine per cigarette by FTC method

How much tar and nicotine you get from a cigarette
depends on how intensely you smoke it.

These messages are of questionable benefit to public health. They raise the question of compensatory smoking, but implicitly resolve it for the smoker by providing an upper range value that is quite low and that, in the case of “lighter,” more heavily ventilated brands, is likely much lower than the upper range value for a regular brand.

The messages also imply that “smoking intensely” will yield the upper range listed. However, because the proposed FTC test does not take into account much compensatory smoking behaviour, the actual upper range for many smokers will be significantly higher than the one listed.

The FTC proposals do not address the use of descriptive labelling. This sets up an anomalous situation where cigarette manufacturers are allowed to misinform smokers on each and every cigarette package through misleading terminology, while regulators limit themselves to correcting the misinformation in a medium guaranteed to either miss many smokers or reach them far less often.

3. Massachusetts Method

In 1997 Massachusetts implemented changes to the way it tests and rates cigarette emissions levels.(5) The results of the revised system for 1997 were released in January of 1998.(6) The Massachusetts system is now the best in the world in terms of accuracy. While not perfect, it accounts for many of the ways smokers compensate, and results in yields that are closer to what smokers actually inhale.

Testing

Massachusetts has focussed on nicotine levels in its testing and reporting. The test method increases the puff volume and decreases the puff interval from the FTC method, and requires that 50% of any ventilation holes be blocked during testing. While this doesn’t eliminate all yield differences between “light” and regular brands, it reduces the size of the discrepancy between many brands. It also provides a much more realistic estimate of what smokers inhale.

In addition to testing for nicotine yield under the revised method, brands are classified into one of four categories: high (>1.2 mg), moderate (0.2-1.2), low (0.1-.2), or nicotine free (<.01). This is a novel approach, and diminishes the problem of smokers perceiving that lower ratings translate into safer smoking. Since the vast majority of brands tested (85%) fall into the “high nicotine” category, smokers will be sent the message that most cigarettes are the same.

Reporting

Unfortunately, federal law pre-empts Massachusetts’ ability to print test results and categories on tobacco packages. Although information is being publicized on the Internet and in paid media, the inability to put the information on or in the package means fewer smokers will benefit from the new system.

However, the information that Massachusetts has produced in support of its system is comprehensive, clearly explained, and offers specific information that smokers can act upon to reduce risk.

Improvements could be made to the Massachusetts system. In terms of testing, it is possible that more vent blocking is justified based on actual smoking behaviour. Smokers who block with their fingers may only cover 50% of the holes, but smokers who block with their lips may cover 80% of the holes.

In terms of reporting, providing information about yields, nicotine categories, and the relationship between smoking behaviour and yield on and inside the package would obviously be preferable. Jurisdictions with the flexibility to print information on the package, such as British Columbia, should use it.

Some Solutions

1. Ban deceptive descriptive labelling.

Whatever occurs with the testing and reporting system for yields, there is no justification for continuing to allow deceptive descriptive labelling. Terms like “light,” “ultra light,” “mild,” and “smooth” are meaningless, mislead smokers into perceiving these cigarettes as less risky, and deter quit attempts.

In fact, a strong argument can be made that the use of these terms is in violation of Section 2(1)(b) of the Tobacco Sales Act:

2. (1) A person must not deal in, sell offer for sale, distribute, advertise or promote the use of tobacco

(b) in a manner that allows a consumer or purchaser of tobacco to be deceived or misled concerning its character, toxicity, composition, merit or safety

BC should enforce this provision by requiring the removal of deceptive descriptive brand names from tobacco packaging. An ancillary benefit may be, ultimately, a reduced number of brands on the market.

2. Amend the test method to more accurately mimic actual smoking behaviour.

Machine tests should reflect a range of smoking conditions and compensatory behaviours, including hole blocking. Massachusetts currently offers the best real-world model, but consultations and research should occur to see how this can be further improved. Some alternative protocols have been developed.(7)

One way of arriving at a range of yields would be to study a representative sample of smokers on which a reported yield range could be based. Another way would be to study smokers in order to determine parameters that could be incorporated into the machine tests. Some researchers have found greater actual intake of tar and nicotine by smokers from “low-yield” cigarettes than from cigarettes with higher yields on the FTC scale.(8)

One issue in particular that should be examined is the extent of vent hole blocking, both in terms of the proportion of smokers that block vents, and the degree to which an individual smoker blocks holes. Studies of prevalence of vent blocking have shown a median of 19% for extreme vent blocking and a median of 50% for some vent blocking.(9) The Massachusetts test uses 50% vent blocking, which may accurately represent blocking by fingers. However, some smokers block vents with their mouths. In this instance, a much higher percentage of blocking probably occurs. Any testing model should research these questions and incorporate the findings.

3. Report yields so smokers understand how to interpret them.

Reporting yield ranges is a first step to helping smokers understand that it is they, not the cigarette, that determine the yield they receive. This should be supplemented with detailed information about the variations in smoking behaviour that interpret yield, along with accurate information about risk reduction due to quitting versus brand switching. An analogy that might be useful to the smoker is that of gas mileage. How far you go on a tank of gas depends to some extent on how you drive.

While detailed information can be provided on a package insert, it is more of a challenge to present this information in short form on the package. One approach is to show what the filter looks like depending on how much has been inhaled.(10),(11),(12) Figure 1 provides an example of such a representation.

Figure 1

4. Create tools to assist the smoker in monitoring smoking behaviour and reducing risk.

It has been a failing of health initiatives that, even when smokers are told that how they smoke affects yield, markers to help smokers determine how they smoke are not provided. Unlike drivers, who can look at the speedometer and know how to adjust speed to improve gas mileage, smokers have few objective cues to help them monitor and adjust smoking behaviour to reduce risk.

The “bulls-eye” representation or colour matching technique described above has the advantage of offering smokers a visual with which they can compare their cigarettes. While the comparison will not be exact, it will give smokers an idea of the yields they inhale. Ads for certain low-tar brands have focused on the “bulls-eye” appearance of the filter to convey to smokers the concept of lower tar. If smokers are urged to examine the filter after they have finished a cigarette and consistently see that it is not a bulls-eye, but mostly brown, they will have a better understanding that the low-tar cigarette does not always result in a low-tar smoke. This could motivate quit attempts.

Another marker to assist smokers could be numbered lines around the circumference of the cigarette itself. This would help smokers determine how much of the cigarette they actually smoke. Butt length is not the sole determinant of how much was smoked, as number and duration of puffs come into play, but providing this marker to smokers is better than the complete lack of guidelines they currently receive.

Some have suggested textualizing and/or making visible vent holes in the filter.(13) Making vent holes visible, perhaps through colour, helps smokers understand how they work and where they are located so as to avoid blocking them. Textualizing vent holes (for example, raised holes as in the raised dots used in braille) would be even better. Smokers could feel with their lips or fingers whether or not they were blocking the holes, and could even experiment to see the correlation between the ease of draw and the blocking of holes.

Whether the above or other mechanisms are used to assist smokers in risk reduction behaviour, such tools should be a component of product/package redesign.

5. Expand ingredient disclosure.

The list of ingredients that currently appears on the package needs to be expanded to include other smoke components and ingredients in cigarettes themselves. Canada’s Expert Committee on Cigarette Modifications has provided some guidance on the ingredients in smoke to be tested and reported.(14)

It may not be feasible to list on the package all ingredients in smoke. To ensure that smokers receive full information, there should be an insert in the package to provide additional ingredient information. Alternatively, as a first step toward greater disclosure, the outside of the package should at least state that the ingredients listed are only a portion of those contained in smoke.

Notes

(1) William Rickert of Labstat Incorporated provided valuable comments on an earlier version of this paper. Any errors or omissions are the responsibility of the author.

(2) The current format for yield information printed on Canadian packages, mandated by regulations under the Tobacco Products Control Act (TPCA), was legally required from September 1994 to April 1997. On September 21, 1995, the Supreme Court of Canada struck down significant portions of the TPCA. Although the portion of the law requiring testing and reporting of smoke yields stood, the TPCA was repealed with the enactment of a new federal law, the Tobacco Act, on April 25, 1997. The Tobacco Act reinstates the legislative authority to print smoke yields on packages. While regulations reinstating the yield information have not yet been promulgated, they are expected to be issued in the near future. In the meantime, most manufacturers have voluntarily continued to print the previously mandated information on packages.

(3) Federal Trade Commission. Cigarette Testing: Request for Public Comment, FTC File No. P944509, September 9, 1997.

(4) Davis RM Healy P Hawk SA. Information on Tar and Nicotine Yields on Cigarette Packages. American Journal of Public Health, May 1990, 80: 5, 551-553.

(5) Massachusetts General Laws, Chapter 307B, CMR 660.000

(6) The Massachusetts Department of Public Health’s 1997 Cigarette Nicotine Disclosure Report is available from the American Cancer Society’s web site: www.cancer.org.

(7) Rickert WS. Smoking Under Realistic Conditions: Development of Minimum and Maximum Values for Toxic Constituents in Tobacco Smoke. Project Report: DSS File No 35SS.H4078-5-C105, for Health Canada, September 30, 1996.

(8) Djordjevic MV Fan J Ferguson S Hoffmann D. Self-regulation of smoking intensity. Smoke yields of low-nicotine, low-’tar’ cigarettes. Carcinogenesis 1995, 16(9), 2015-2021.

(9) Kozlowski LT Goldberg ME Yost BA Ahern FM Aronson KR Sweeney CT. Smokers are unaware of the filter vents now on most cigarettes: results of a national survey. Tobacco Control 1996, 5, 265-270.

(10) Kozlowski LT Pope MA Rickert WS Robinson JC. A Colour-Matching Technique for Monitoring Tar/Nicotine Yields to Smokers. American Journal of Public Health 1982, 72, 597-599.

(11) Rickert WS Robinson JC Kaiserman MJ. Quantitation of “Tar” Colour with Specific Reference to Estimating Yields, Quantifying ETS and the Production of Colour Scales. 48th Tobacco Chemists’ Research Conference, Greensboro, North Carolina, 1994.

(12) Kozlowski LT Pillitteri JL. Compensation for Nicotine by Smokers of Lower Yield Cigarettes. Chapter 12 in The FTC Cigarette Test Method for Determining Tar, Nicotine, and Carbon Monoxide Yields of U.S. Cigarettes: Report of the NCI Expert Committee. NIH Publication No. 96-4028. National Cancer Institute, National Institutes of Health, US Department of Health and Human Services, August 1996.

(13) Kozlowski LT. In Report of Canada’s Expert Committee on Cigarette Modifications. Conference Proceedings. Rickert WS (Ed.) March 1-3, 1996, Toronto, Ontario, Canada.

(14) Report of Canada’s Expert Committee on Cigarette Modifications. Conference Proceedings. ibid.