Smoking and Lung Cancer Associations

Lung Cancer

Smoking responsible for 90% of cancer deaths

Incidence of lung cancer is rising. Lung cancer is responsible for more cancer deaths than colorectal cancer, breast cancer, and prostate cancer combined.

Lung cancer primarily strikes people over age 45. By the time that an individual develops symptoms, spread has usually occurred.

 

SmokingLung cancer is directly related to smoking. Over 40 carcinogens have been identified in cigarette smoke. The risk of developing lung cancer is directly related to the number of cigarettes smoked. The change in consumption from unfiltered high tar cigarettes to filtered low tar cigarettes parallels the change in incidence from squamous cell carcinoma to adenocarcinoma. There is a long interval between quitting smoking and elimination of lung cancer risk. Up to 40% of newly diagnosed lung cancer occurs in former smokers. (median abstinence duration 9 years).

Most common diseases result from interplay between genes and environment, although the relationship between these factors is not always clear. Three recent papers^[ tackle the relationship between genetics, smoking, and smoking-related diseases like lung cancer, yet they come to different conclusions: One suggests that a locus influences smoking behavior, which in turn affects lung cancer; while two others suggest that the same locus influences lung cancer independently of smoking.

Although smoking rates are on the decline, 21% of all adults (45 million people) in the United States smoke cigarettes. In addition to being the primary cause of lung cancer, smoking also influences heart disease, with smokers increasing their risk for coronary heart disease by 2-4 times compared with nonsmokers. Despite these strong associations, however, not all individuals who smoke will go on to develop these diseases. Genetics may play a role here, making some individuals more susceptible by increasing their risk for disease or through influencing smoking behavior itself.

Thorgeirsson and colleagues^. approached smoking behavior directly and performed a genome-wide association study (GWAS) on the number of cigarettes smoked in 10,995 Icelandic smokers. They found that a locus on 15q24 containing 3 nicotinic acetylcholine receptors was most strongly associated with number of cigarettes smoked (P = 5 x 10^-16), although there was no relationship for initiating smoking. For each additional risk allele carried, individuals smoked on average 1 more cigarette per day. The nicotine receptors at this locus are good candidates for variation in smoking behavior and previously have been associated with nicotine dependence.The authors go on to show that the most strongly associated single nucleotide polymorphism (SNP) is also associated with lung cancer and peripheral arterial disease, both diseases strongly influenced by smoking. The authors do not specifically quantitate how much risk for disease they expect to be caused by smoking 1 additional cigarette per day, however, and so we do not know whether the locus' actions are primarily mediated through smoking behavior or through another mechanism.

Amos and colleagues and Hung and colleagues approached this problem with lung cancer, performing a GWAS on lung cancer cases and controls. Both studies found the same region on 15q24 to carry the strongest association with lung cancer with remarkably similar estimates of risk (combined allelic ORs = 1.32 and 1.30). Both studies attempted to explain their estimates of risk by addressing smoking behavior. Amos and colleagues had carefully frequency-matched their 1154 cases and 1137 controls according to smoking behavior and years of cessation, making it less likely that they would find loci related to smoking behavior. Adjusting for pack-years of smoking within their primary dataset did not greatly influence the observed odds ratio, suggesting that the effect of the locus is not strongly dependent on smoking patterns. Amos and colleagues did observe, however, that within the small number of individuals who had never smoked in a replication sample, there was no increase in lung cancer risk. Additional studies of lung cancer in nonsmokers will need to be performed to confirm this. Hung and colleagues performed a larger study with 1989 lung cancer cases and 2625 controls. Stratification by smoking in this case suggested that current smokers, past smokers, and nonsmokers all show similar risk for lung cancer (ORs range from 1.25 to 1.32), although due to small sample size, the P value for nonsmokers was relatively large (P = .013). Thus, although this locus shows association with smoking behavior, there is little evidence to currently suggest that the risk for disease is mediated through this behavior.

That genetics would influence both smoking and lung cancer independently was once an argument against the relationship between smoking and lung cancer. R.A. Fisher, the preeminent statistician, proposed this idea in 1957, albeit while a scientific consultant to the Tobacco Manufacturers' Standing Committee.The theory was ultimately dismissed, and there is now a plethora of evidence that smoking does cause lung cancer, but there may be a grain of truth to the old argument. The locus described in these papers acts like a locus that is associated both with smoking quantity and with lung cancer, yet the risk for lung cancer does not appear to be strongly dependent on smoking. Thus, those carrying the risk allele may be more likely to smoke and to get lung cancer, increasing the association between smoking and lung cancer. That is not to say that smoking itself is not the main causative agent for lung cancer; rather, the relationship between genotype and phenotype has the potential to be quite complex. These studies are prototypic of the challenges in understanding complex gene-environment interactions. Furthermore, discovery of the common pathway linked to nicotine addiction and lung cancer may serve as a foundation for better drug intervention in the future.