Cannabis smoking and lung cancer risk

From our recent review of the impact of cannabis use on respiratory health (Gates et al 2014), we found that the most consistent finding was an association between frequent smoking and symptoms of chronic bronchitis. These symptoms are likely caused by injury and inflammation to our central airways and by a reduction in the amount of ciliated epithelium. Interestingly associations between cannabis use and its impact on actual lung functioning were not as clear. This includes impact on airflow obstruction, ability to forcefully expire air and the development of lung cancer.

This lack of clarity regarding a link between cannabis use and lung cancer is particularly interesting given that tobacco smoking is recognised to be the main risk factor for lung cancer. Moreover, like tobacco smoke, cannabis smoke contains known carcinogens but is generally smoked with longer inhalation times, resulting in greater retention in the lungs. Notably, previous epidemiological research on the associations between cannabis smoking and lung cancer risk have been criticised for suffering from serious limitations such as confounding tobacco use, small sample sizes, and a failure to account for the frequency of cannabis use between follow-up assessments.

Recognising the mixed findings and limitations of this research, the International Lung Cancer Consortium (ILCCO) - an international group of lung cancer researchers with the key goal of exploring potential lung cancer risk factors – has pooled together the findings from six relevant studies. Each of these studies considered primary, incident and histologically confirmed lung cancer cases (chiefly including squamous cell carcinoma, adenocarcinoma and small cell lung cancer), with a pooled analysis sample consisting of 2,159 cases and 2,985 controls (including a pooled sample of 370 cases and 1,358 controls who had never smoked tobacco).

Each of these cases and controls were compared to assess the odds that a cannabis smoker would develop lung cancer compared to a non-smoker (referred to as an odds ratio; OR) while matching on the individual’s age, sex, race, highest level of education, and use of tobacco (ever smoked compared to never smoked and a continuous measure of smoking frequency). The estimated OR of any cannabis use compared to never or rarely smoking ranged from 0.57 to 2.17 (meaning individuals were about half as likely in one study to twice as likely to have any lung cancer in another study). The ILCCO overall combined OR was estimated to be 0.96 (0.66-1.38), although this increased to be 1.03 (0.54-1.98) when considering those who smoked for 20 years or more. Interestingly, no overall association was found between cannabis smoking and any lung cancer when considering those who never smoked tobacco although this was a comparatively small sample.

Of all the lung cancer types investigated the strongest association with cannabis use was regarding adenocarcinoma. For this type of lung cancer, compared to never or rarely smoking, daily smoking was estimated to increase the health risk by 1.73 (0.75-4.00).

Some limitations of the included studies could not be avoided and may have contributed to the range in the estimates presented by the ILCCO. First, the representativeness of the included studies was limited in that half the studies were from the US (and one from Canada) and participants were over 75% White/Caucasian. Second, the studies relied on self-report of cannabis use and this was not verified with bioanalysis (although the self-report of drug use has been supported to be reliable in a number of studies and, in this case, the frequency of cannabis use reported did not have a considerable impact on the findings). Third, the studies did not assess the way in which the cannabis was smoked (inhalation techniques) or any aspects of the cannabis used (species/type, or how much of a joint was smoked).

Unfortunately, as is too often the case when attempting to combine studies which have been criticised for certain methodological limitations, the pooled result does not always increase the clarity of results. Importantly, particularly for comparisons between heavy, frequent and long-term use (chronic use) and never or rarely smoking cannabis; it was clear that the risk of developing any lung cancer, and particularly adenocarcinoma, was increased. As such, it appears that it is not a question of whether cannabis use can increase the risk of lung cancer, but rather, how much does the risk increase? Naturally, future study, with more accurate and standardised measures of cannabis use and controls for confounding variables, will provide greater clarity in terms of how much risk is associated with different frequencies of cannabis use.

References:

Gates P, Jaffe A, & Copeland J. (2014). Cannabis smoking and respiratory health: Consideration of the literature. Respirology 19, 655-662.

Zhang, L.R., Morgenstern, H., Greenland, S., et al. (2015). Cannabis smoking and lung cancer risk: Pooled analysis in the International Lung Cancer Consortium. International Journal of Cancer 136, 894-903.