Action on Smoking and Health

November 2002

There has been extensive coverage of the British Lung Foundation's report "A smoking gun?" released to the press on 11 November in which comparisons are made between the hazardousness of tobacco and cannabis use. Some of the media coverage goes beyond what is stated in the report, but in other areas the report itself is quite misleading. The report is an in-house publication and is available at the BLF web site. The report is a literature review, and as such contains no new data – though it has been reported in some places as if it is new research. Here are some comments from ASH on the main headlines...

1. Murky origin of the claim that three cannabis joints equates to 20 cigarettes
The reported claim that three cannabis joints per day equate in risk to 20 cigarettes has a very dubious basis. It is based on data published in two papers in 1987 (see below), but the papers in question did not make that claim and were MUCH more cautious. This relationship was asserted in a later paper that cited the original papers as demonstrating the relationship without adding any new analysis, even though the originals did not contain that interpretation. Note that although the original data papers were actually cited in the BLF report in relation to other points, the BLF did not cite these papers in support of the 3:20 claim – one might assume because these papers didn't make the claim. The evolution of the 3:20 assertion is set out below, starting from the news coverage on launch day, and working back through press release, the report, and its references to secondary sources to find its original source.

2. Limited scope of comparison between tobacco and cannabis
The two 1987 studies on which the claim is based examine only a limited range of respiratory illness symptoms, and did not estimate the risks of lung cancer and chronic obstructive pulmonary disease (COPD - eg. emphysema), which are the main fatal lung diseases caused by smoking tobacco. In the UK, lung cancer and COPD are responsible for almost half of tobacco related deaths (heart disease taking most of the rest). The BLF report acknowledges "conflicting findings" on the link between lung cancer and cannabis, and calls for more research "to establish what link (if any) there is between COPD and cannabis smoking". Given that the data used don't actually cover the main risks and the link between these major risks is acknowledged to be uncertain, it is premature to draw overall risk comparisons between cannabis and tobacco - and certainly not with precision like 3:20. BLF did limit the scope of the claim in its report - but it was widely interpreted as a measure of overall relative risk.

However, one can see how this happens... for example: on BBC Radio's flagship "Today" programme (11 November 2002, 08.28):

Presenter: "...you come to the conclusion that three cannabis joints a day are doing the same damage as 20 cigarettes?"
BLF spokesperson: "Absolutely".

3. Central importance of usage pattern and lifetime exposure ignored
The major mortal risks to the lungs, heart and circulation depend on lifetime exposure as well as the toxicity of the smoke. The importance of lifetime exposure was underlined in a major study of tobacco smokers in 2000 which found that: "People who stop smoking, even well into middle age, avoid most of their subsequent risk of lung cancer, and stopping before middle age avoids more than 90% of the risk attributable to tobacco." [*Peto R et al. Smoking, smoking cessation, and lung cancer in the UK since 1950: combination of national statistics with two case-control studies. BMJ 2000; 321: 323-329]. The very high risks due to tobacco use really arise from its addictiveness, which causes many tobacco smokers to continue to smoke well after they would want to stop. It is common for tobacco smokers to consume 20 cigarettes per day every day for several decades. However, this is not a common pattern of use for cannabis, which appears to be much less addicting than nicotine. The UK government’s Advisory Council on Misuse of Drug noted that:

In general cannabis users smoke fewer cigarettes per day than tobacco smokers and most give up in their 30s, so limiting the long-term exposure that we now know is the critical factor in cigarette-induced lung cancer. (4.4.1)

Advisory Council on the Misuse of Drugs. The classification of cannabis under the Misuse of Drugs Act 1971, (UK Government) Home Office, March 2002.

The House of Lords took evidence on addictiveness of cannabis and concluded that dependence was less serious:

Giving up cannabis is widely believed to be relatively easy: according to the Department of Health, "studies report that of those who had ever been daily users only 15 per cent persisted with daily use in their late twenties" (4.31)

House of Lords Committee on Science and Technology, Ninth Report Session 1997-98, Cannabis, the scientific and medical evidence. November 1998.

4. Unwarranted and unsubstantiated scare about increased strength of modern cannabis
The report claims that: "The cannabis smoked today is much more potent than that smoked in the 1960s" and states that the average cannabis cigarette in 1960s contained 10mg THC (the active substance) compared to 150mg today. However, this increase in the strength of cannabis is not obviously a cause for alarm. Even if there is, as claimed, 15 times as much THC in modern cannabis it is unlikely that today's users are 15 times as stoned as their predecessors in the 1960s. It is plausible that cannabis users control the dose they receive by varying their smoking pattern - as it has been shown that tobacco smokers do for nicotine. Stronger cannabis may therefore mean that LESS smoke is inhaled for a given dose of the active ingredient. There is a large literature on 'compensation' and the tendency of smokers to titrate nicotine, though the subject is much less well understood for cannabis. A better working assumption that dose is controlled by the user rather than by the cigarette, and anecdotally, cannabis smokers say that they smoke rather than eat the drug because it is easier to control the dose. However, the report jumps without evidence to the opposite conclusion, leading to some extremely misleading reporting – for example on CNN, one of the world’s largest news networks made the following report, which is thoroughly flawed. See Cannabis smoke health warning, 11 November 2002.

LONDON, England --Health risks from smoking cannabis have risen dramatically since the 1960s because of changes to the way the drug is produced, a health charity says. […] The BLF report, published on Monday, said the health risks were worse now than in the 1960s because there is more THC (tetrahydrocanabinol), the ingredient which accounts for the psychoactive properties of cannabis, in the substance consumed today.

5. Claimed higher toxicity of cannabis tar ignores dramatic variations in tobacco toxin concentrations
The claim that there are 50% more carcinogens in cannabis tar in tobacco smoke also demands caution given the wide variation in carcinogens within even the same cigarette brand. The BLF report identified two carcinogens - "tar from cannabis cigarettes contains up to 50% higher concentrations of the carcinogens benzanthracenes and benzpyrenes". But carcinogens do vary wildly within cigarette brands... Gray et al (2000) measured two nitrosamine yields, NNK and NNN, in Camel, Marlboro and Lucky Strike cigarettes, and found that a:

"three to nine fold variation in carcinogen dose can be given to the smoker... ".

Gray N, Zaridze D, Robertson C, et al. Variation within global cigarette brands in tar, nicotine, and certain nitrosamines: analytic study. Tob Control 2000;9: 351.

Measurements made in British Columbia also show marked carcinogen variations between brands of similar magnitude to that reported for cannabis - see: British Columbia Tobacco Testing and Disclosure: What's in Cigarettes?

6. Need for education…
The BLF report states that “The British Lung Foundation recommends a public health education campaign aimed at young people…”. However, misleading and simplistic comparisons of risk, and wholly flawed interpretation of data related to ‘strength’ of cannabis do little to educate or inform anyone – young or old. Most of the criticisms above were drawn to the attention of BLF, when a preview of the “A smoking gun?” was reported in the Mail on Sunday on 21 July 2002. ASH provided comments to BLF at the time making most of the points above on the basis of the news article and urged caution about drawing these conclusions. These comments were ignored. In July 2002 The British Lung Foundation was even more explicitly claiming that cannabis was more harmful than tobacco. See BBC report of 10 July 2002: Cannabis “worse than tobacco”

BLF chief executive Dame Helena Shovelton said: "Many young people are simply not aware that smoking cannabis may put them at increased risk of respiratory cancers and infections. The government spends millions of pounds a year on smoking cessation and public education about the dangers of smoking, yet smoking cannabis is at least as harmful as smoking tobacco and, indeed, may carry a higher risk of some respiratory cancers."

Despite this message, delivered earlier in the year in response to the government’s plans to decriminalize cannabis, the report as published does not substantiate that claim. It is difficult therefore to see much educative value in these pronouncements, especially as they contrast with more carefully and credibly formulated information and analysis in recent independent assessments, some of which we describe below.

The usually cautious Advisory Council on Misuse of Drugs, which concluded in March 2002 after a thorough review of the evidence...

"The high use of cannabis is not associated with major health problems for the individual or society." (5.1)

Advisory Council on the Misuse of Drugs. The classification of cannabis under the Misuse of Drugs Act 1971, (UK Government) Home Office, March 2002.

Further, in 1998 The Lancet commissioned a thorough review of the health impacts of cannabis use and concluded in an editorial:

…on the medical evidence available, moderate indulgence in cannabis has little ill-effect on health, and that decisions to ban or to legalise cannabis should be based on other considerations.

Anon (editorial). Dangerous habits. Lancet 1998;352:1565.

Hall W. and Solowij N. Adverse effects of cannabis, Lancet 1998; 352:1611-­16

The Police Foundation’s extensive inquiry (‘The Runciman Commission’) concludes:

When cannabis is systematically compared with other drugs against the main criteria of harm (mortality, morbidity, toxicity, addictiveness and relationship with crime), it is less harmful to the individual and society than any of the other major illicit drugs or than alcohol and tobacco.

Drugs and the Law, Report of the Independent Inquiry into the Misuse of Dugs Act (1971): Chairman: Viscountess Runciman DBE, 1999.

For a discussion of wider issues in relation to cannabis and tobacco, analysis of policy implications and extensive links to reviews of scientific evidence see ASH’s: Legalisation of cannabis – a discussion document.

Warning! – and further information
None of this is to argue that cannabis is harmless! Far from it. Prolonged and heavy cannabis use should be expected to cause respiratory diseases and other ill-effects. However, when making a comparison with tobacco or educating tobacco and cannabis users about the risks it is important to have the harm in perspective. For a better understanding of the risks associated with cannabis we advise consulting the sources listed above and a special reviews in:

Journal of Clinical Pharmacology 42 (11 supplement), November 2002 (abstracts)
Advisory Council on the Misuse of Drugs. The classification of cannabis under the Misuse of Drugs Act 1971, (UK Government) Home Office, March 2002.
Hall W. and Solowij N. Adverse effects of cannabis, Lancet 1998; 352:1611-­16

For an understanding of tobacco, we advise consulting the following:

UK Scientific Committee on Tobacco and Health
US Surgeon General reports
Royal College of Physicians: Nicotine Addiction in Britain

This is the evolution of the 3:20 claim, tracing it back to the data papers published in 1987.

1. The reporting...

"The evidence indicates that three cannabis joints does the same damage to the lining of the lungs as 20 cigarettes" The Independent
"Smoking three joints per day might do as much damage to the lungs as do 20 cigarettes" The Guardian
"Three or four cigarettes are as damaging as 20 cigarettes" The Times
"Three or four cannabis cigarettes are equivalent to smoking 20 tobacco cigarettes a day in terms of the risk of lung damage" The Telegraph
"A study by the British Lung Foundation found that just three cannabis joints a day cause the same damage as 20 cigarettes" BBC.

2. BLF Press release

"Three cannabis joints a day cause the same damage to the lining of the airways as 20 cigarettes."

3. BLF Report Summary

"3-4 Cannabis cigarettes a day are associated with the same evidence of acute and chronic bronchitis and the same degree of damage to the bronchial mucosa as 20 or more tobacco cigarettes a day.

4. BLF report body

"It has been calculated that smoking 3-4 cannabis cigarettes a day is associated with the same evidence of acute and chronic bronchitis and the same degree of damage to the bronchial mucosa as 20 or more tobacco cigarettes a day [14][ 15].

5. Reference 15 used in BLF report

Ref 15 in BLF report is: Tashkin, DP, Effects of marijuana smoking profile on respiratory deposition of tar and absorption of CO and D-9 tratrahydrocanabinol, In: Pulmonary pathophysiology and immune consequences of smoked substance abuse, FASEB Summer Research Conference, July 18-23, 1999, Copper Mountain, CO

This not a peer reviewed paper and not easy to obtain. However, Tashkin is an author of the papers that offer the original data.

6. Reference 14 used in BLF report

Ref 14 in BLF report: Ashton H, 2001 Pharmacology and effects of cannabis: a brief review Br Journal of Psyschiatry 178, 101-106

Ashton doesn’t make the case herself, but cites an earlier paper, (Benson & Bentley, 1995) to make this claim.

"It has been calculated that smoking 2-4 cannabis cigarettes a day is associated with the same evidence of acute and chronic bronchitis and the same degree of damage to the bronchial mucosa as 20 or more cigarettes a day (Benson and Bentley, 1995)

7. Benson & Bentley (cited as a source in BLF ref 14)

Benson M and Bentley AM, (1995). Lung disease induced by drug addiction. Thorax, 50, 1125-1127 - cited in Ashton (2001)

Benson & Bentley's article does not contain this calculation or make this claim. The nearest Benson and Bentley get is to cite the ubiquitous Wu et al (1988) study...

All Benson and Bentley say is "Smoking a cannabis cigarette results in an approximately five-fold greater increase in carboxyhaemoglobin concentration than with a tobacco cigarette, with increases in inhaled tar content and the amount retained in the respiratory tract. [23]". This is a repeat of the Wu et al conclusion and does not in itself substantiate the 3:20 relationship.

8. Wu et al (1998) cited in Benson and Bentley as ref 23

Wu TC, Tashkin DP, Djahed B, et al. Pulmonary hazards of smoking marijuana as compared with tobacco. N Engl J Med 1988;318: 347-351.

The Wu et al paper is also cited in the BLF report at ref 16 - but this is not used to substantiate the 3:20 estimate. Wu et al. measured tar deposition in the lungs of 15 smokers of both cannabis and tobacco. The Wu et al paper is also not the original source, but refers to an earlier claim and starts with the following...

"We have previously shown that the habitual smoking of 3 or 4 marijuana cigarettes per day is associated with the same frequency of the symptoms of acute and chronic bronchitis [1] and same type and extent of epithelial damage in the central airways [2] as the regular smoking of 20 tobacco cigarettes a day.

9. The original data papers

It seems that references [1] and [2] in Wu et al are the original data sources for this claim...

[1] Tashkin, DP, Coulson, AH, Clark, VA, et al, 1987, Respiratory symptoms and lung function in habitual, heavy smokers of marijuana alone, smokers of marijuana and tobacco, smokers of tobacco alone and nonsmokers, Am Rev Respir Dis 135, 209-216

[2] Gong, H, Fligiel, S, Tashkin, DP, Barbers, RG, 1987, Tracheobronchial changes in habitual heavy smokers of marijuana with and without tobacco, Am Rev Respir Dis 136, 209-216

The abstracts of these papers are reproduced below. Neither paper makes a numerical comparison between cannabis and tobacco.

These papers are referenced at [38] and [33] respectively in the BLF report, but are not used in the BLF report as citations to back the 3:20 claim. One good reason for that may be that these papers do not actually make this claim. It is not until these papers are cited in the introduction to Wu et al, that the 3:20 (or more accurately 3 or 4 to 20) comparison starts to be made, and this is then subsequently repeated in other later papers as if it is established. The author that is common to the two data papers, and to Wu et al is Donald Tashkin - and Tashkin's conference paper is also cited by BLF in support of this comparison. Why the claim was not included in the basic papers but then cited as fact subsequently is unclear to me.

Possible reason why the 3:20 interpretation was not made in the original paper
A possible reason why the claim wasn't made (or accepted by peer reviewers) in the original data paper is as follows: There are only really two data points - these are the average consumption of the tobacco smokers and average consumption of the marijuana smokers recruited into the study (3.9 and 22 per day respectively in [1]). There was no attempt in either paper to see how the symptoms varied with consumption. So it is unknown whether people smoking three tobacco cigs per day would have experienced the same or less of the symptoms than someone smoking 20 cigs. If the symptoms were not that sensitive or proportional to consumption, then entirely different number may have been found. These are also unusually heavy users of cannabis but quite 'normal' users of tobacco and it is possible can imagine quite a bit of potential confounding and potential bias.

Abstracts for the original papers
It may be of interest to know what was actually said in the data papers: it turns out to be far more cautious.

Abstract for Tashkin et al [1] above...

To evaluate the possible pulmonary effects of habitual marijuana smoking with and without tobacco, we administered a detailed respiratory and drug use questionnaire and/or lung function tests to young, habitual, heavy smokers of marijuana alone (n = 144) or with tobacco (n = 135) and control subjects of similar age who smoked tobacco alone (n = 70) or were nonsmokers (n = 97). Mean amounts of marijuana and/or tobacco smoked were 49 to 57 joint-years marijuana (average daily number of joints times number of years smoked) and 16 to 22 pack-years of tobacco. Among the smokers of marijuana and/or tobacco, prevalence of chronic cough (18 to 24%), sputum production (20 to 26%), wheeze (25 to 37%) and greater than 1 prolonged acute bronchitic episode during the previous 3 yr (10 to 14%) was significantly higher than in the nonsmokers (p less than 0.05, chi square). No difference in prevalence of chronic cough, sputum production, or wheeze was noted between the marijuana and tobacco smokers, nor were there additive effects of marijuana and tobacco on symptom prevalence. We noted significant worsening effects of marijuana but not to tobacco on specific airway conductance and airway resistance (tests of mainly large airways function) in men and of tobacco but not of marijuana on carbon monoxide diffusing capacity and on closing volume, closing capacity, and the slope of Phase III of the single- breath nitrogen washout curve (tests reflecting mainly small airways function) (p less than 0.03, two-way ANCOVA). No adverse interactive effects of marijuana and tobacco on lung function were found

Abstract for Gong et al [2] above...

We performed flexible fiberoptic bronchoscopy in 29 habitual, heavy marijuana smokers 25 to 45 yr of age, with and without concomitant tobacco smoking, to inspect and biopsy their proximal tracheobronchial tree for the evaluation of histopathologic changes. Control tobacco smokers (TS) and nonsmokers (NS) residing in the same metropolitan area were similarly studied and compared with the marijuana smokers (MS) and marijuana-tobacco smokers (MTS). Respiratory and drug histories, physical examination, and pulmonary function tests were obtained prior to bronchoscopy. The prevalence of respiratory symptoms and pulmonary function abnormalities was generally higher in the 3 smoking groups than in the NS group but was not statistically different across all groups. However, bronchoscopic inspection revealed airway hyperemia and other visible abnormalities in 32 (91%) subjects in the 3 smoking groups, unlike the unremarkable findings in the NS group. Light microscopy showed 2 or more histopathologic changes in the bronchial epithelium of all MS, MTS, and TS. Squamous metaplasia was observed in all MTS, a prevalence that was significantly different from that in MS, TS, and NS. Hyperplasia of basal and goblet cells was more prevalent in the MS than in the NS, whereas cellular disorganization was more prevalent in the MS than in the TS. A direct relationship between cumulative marijuana use (joint-years) and bronchoscopic and histopathologic changes was not apparent in this study sample. These results indicate that relatively young, habitual, heavy marijuana smokers have a high prevalence of abnormal airway appearance and histologic findings, irrespective of concomitant tobacco smoking. the results suggest a causal relationship between marijuana smoking and histologic lesions in the airways. The long term clinical importance of these histopathologic findings is unclear. In view of the widespread use of marijuana in this country, however, the finding of histopathologic airway change in young adult marijuana smokers justifies a serious concern about the development of chronic airy disease in these smokers.