Does Cigarette Smoking Cause 87% of Lung Cancers?
[This politically-incorrect article was submitted to the journal Science and later to the New England Journal of Medicine in 2009, and was rejected in either case, as expected, with no reason given. I have since added more evidences, and revised some paragraphs. BASIC FACTS - Lung cancer kills more people than colorectal, breast, pancreatic, and prostate cancers combined. Lung cancer kills almost twice as many women as breast cancer, and almost three times as many men as prostate cancer, but government funding of lung cancer research used be less than 1/10 a few years ago and is still only half of breast cancer research. Currently in the US, 60+% of all new lung cancer diagnoses are in non-smokers or never-smokers. Lung cancer accounts for 14% of all new cancer diagnoses but 27% of all cancer deaths. TAKE HOME MESSAGE: smoking cessation promotion is not the cure for society's lung cancer problem, but actually the reason why so many today are suffering and dying from this extremely deadly disease. - Jaydad]
Does Cigarette Smoking Cause 87% of Lung Cancer? / 肺癌: 無煙可戒怎麽辦?
Abstract
As more and more non-smokers and never smokers are diagnosed with lung cancer, the landscape in lung cancer etiology has been slowly but surely changing. However, public perception and stigmatization of lung cancer has not. Fueling the prejudice against lung cancer are powerful forces behind aggressive anti-smoking campaigns. While these nameless forces may possess genuine merits otherwise, the injustice they brought upon non-smoker lung cancer patients, the free-passes they seemingly delivered to non-tobacco INDUSTRIAL causes of lung cancer by accident, and the subsequent misappropriation of small-to-start-with lung cancer research fund, are of grand scale and appalling. The author set out to analyze US government data on cigarette smoking and lung cancer incidences over the past half century and tried to answer subjectively if cigarette smoking is major contributory factor in lung cancers, broken down to pathology subcategories and in male and female sexes. Results agree that smoking causes the majority of small cell lung cancer, in both male and female (28,000 cases, Year 2007/CDC data), and that smoking is also significantly responsible for MALE non-small cell lung cancer (93,000 cases), of non-adeno-carcinomas pathology, i.e. squamous cell and large cell carcinomas (57,000 cases), and that smoking could NOT possibly be a MAJORITY (>50%) cause for women non-small cell lung cancer (76,000 cases, 38.6% total lung cancer cases). There is no correlation between smoking rate and time-delayed male lung adenocarcinoma (36,000 cases, 18.3% of total lung cancer cases) which corroborates very well with the fact that overwhelming majority of never smoker lung cancers fall into this NSCLC subcategory. In conclusion, systematic exaggeration of smoking contribution to lung cancer, strongly driven by political lobbying groups and hidden-corporate-sponsored organizations, is utterly self-evident, and perhaps is anything but accidental.
In the US, cigarette smoking reached highest prevalence during the 1960s, with per capita cigarette production & consumption peaked at 4,345 in 1963, and population smoking rate plateaued at 42.4% around 1965 (1965 was the first year national smoking survey was on record at CDC) and backed down ever since. More than half of US men (52%) were current smokers in 1965. As of 2007, current smoker rates are 22.7% men and 17.4% women.1
Small cell lung cancer incidence
The carcinogenic effect of cigarette smoking began to materialize some twenty plus years later, well in agreement with multi-stage carcinogenesis theory.2 In 1986 (21 years after smoking rate peak), small cell lung cancer incidence peaked in US male population at 15.5 per 100,000 people (age adjusted incidence, by SEER-NCI 3). There is a near perfect correlation between US male population smoking rate and male SCLC incidence 21 years afterwards, with an Rsq = 0.96.
Small cell lung cancer incidences, age-adjusted
SC O M F
1975 6.6 10.4 3.8
1976 7.1 11.4 3.8
1977 7.7 12.4 4.2
1978 8.6 13.4 5
1979 8.7 13.1 5.4
1980 9.1 14.2 5.2
1981 9.7 14.5 6.1
1982 10.4 14.9 7.1
1983 10.4 14.9 7.1
1984 10.7 15.2 7.4
1985 10.4 15 7.1
1986 11.2 15.5 8
1987 10.6 14.1 8.2
1988 11.4 15.3 8.5
1989 11.2 14.8 8.7
1990 10.7 13.8 8.4
1991 11.3 14 9.3
1992 10.5 13.4 8.4
1993 10 12.3 8.3
1994 9.8 11.7 8.4
1995 9.8 12.2 8.1
1996 9.7 12.1 8
1997 9.5 10.9 8.6
1998 9.3 10.8 8.1
1999 8.8 10 8
2000 8.8 10.4 7.6
2001 8.3 9.8 7.3
2002 8.2 9.3 7.5
2003 8.1 9 7.5
2004 8 8.8 7.4
2005 7.4 8.4 6.7
2006e 7.2 8.1 6.5
e - estimated; peak incidences shown in bold
Male smoking and SCLC
smoking y rate cancer y incidence
1965 51.9 1986 15.5
1970 44.1 1991 14
1974 43.1 1995 12.2
1978 38.1 1999 10
1980 37.6 2001 9.8
1983 35.1 2004 8.8
1985 32.6 2006 8.2
Rsq 0.96*
In 1988, 23 years after overall population smoking rate peak in 1965, US SCLC reached peak at 11.4 per 100,000 people (both sexes). Again, the incidence and smoking rate correlation is nearly perfect, with Rsq = 0.97, using a 23-year response lapse. Small cell lung cancer, of which patients are 99% long time heavy smokers, is a truly smoking caused disease. Percent of SCLC as of all lung cancers has dropped precipitously over the past 2 decades (from 16% down to 12%), reflecting a 10% decrease in smoking rate in the population some 20+ years earlier.
Population smoking and SCLC (both sexes)
smoking y rate cancer y incidence
1965 42.4 1988 11.4
1970 37.4 1993 10
1974 37.1 1997 9.5
1978 34.1 2001 8.3
1980 33.2 2003 8.1
1983 32.1 2006 7.2
Rsq 0.97*
In the female population, smoking peak took longer to translate into SCLC spike. With a 26-year lapse, we have a moderately strong correlation, with Rsq = 0.92. There seems to be a sex difference in the carcinogenic effect of smoking - an interesting topic but outside current discussion.4
Female smoking and SCLC
smoking y rate cancer y incidence
1965 33.9 1991 9.3
1970 31.5 1996 8
1974 32.1 2000 7.6
1978 30.7 2004 7.4
1980 29.3 2006 6.5
Rsq 0.92*
Non-small cell lung cancer incidence
Similar to SCLC, NSCLC in US male population reached peak around same time. Using a 22-year response lapse, we are able to calculate an Rsq = 0.9, indicating there was statistically significant correlation between male smoking and male NSCLC incidence.
Non-small cell lung cancer incidences, age-adjusted
NSC O M F
1975 45.6 79.2 20.7
1976 48.4 82.5 23.5
1977 49 83.1 24.1
1978 49.2 83.3 24.6
1979 49.9 82.7 26.2
1980 51.6 85.8 27
1981 52.3 84.9 29
1982 52.9 85.8 29.6
1983 53.1 84.6 30.8
1984 54.8 86.9 32
1985 54.2 83.7 33.1
1986 54.6 83.5 34.3
1987 57.2 87.4 36
1988 56.6 83.4 37.8
1989 56.3 83.1 37.4
1990 57.4 83.1 39.3
1991 57.9 83.3 40.2
1992 58.9 83.8 41.4
1993 57.7 81.6 40.8
1994 57.3 79.2 42.1
1995 56.9 77.5 42.3
1996 56.7 75.9 43.1
1997 57 75.4 43.9
1998 58.2 77.1 44.7
1999 57 74.5 44.5
2000 55.2 71.5 43.6
2001 55.5 71.1 44.2
2002 55.4 70.5 44.7
2003 56 71.1 45.1
2004 53.2 66 43.9
2005 53.6 64.8 45.5
2006e 54 46
e - estimated; peak incidences shown in bold
Male smoking and NSCLC
smoking y rate cancer y incidence
1965 51.9 1987 87.4
1970 44.1 1992 83.8
1974 43.1 1996 75.9
1978 38.1 2000 71.5
1980 37.6 2002 70.5
1983 35.1 2005 64.8
Rsq 0.90*
However, among US female population, there is (a negative correlation) no positive correlation between US female smoking rate and female NSCLC incidence. Female NSCLC has been on steady increase since 1975, with age-adjusted incidence more than doubled within the last 30 years, while female smoking rate has been on steady decline since 1965. If smoking (at 1978 rate of 30.7%) had caused majority (>=50%) of female NSCLC in 2004, it would have caused at least 0.5*(33.9/30.7)*40.2=22.2 cases in 1991. Even at the same 50% contribution (with higher smoking rate at 1965), the 1991 female NSCLC incidence should be at least 44.4 in 1991, not 40.2. The math will never add up, for anything bigger than 50%. Therefore, cigarette smoking is not a majority contributing factor in US female NSCLC etiology.
Female smoking and NSCLC
smoking y rate cancer y incidence
1965 33.9 1991 40.2
1970 31.5 1996 43.1
1974 32.1 2000 43.6
1978 30.7 2004 43.9
1980 29.3 2006 46
Rsq 0.93**
** negative correlation
Using the same 23 year response lapse as with SCLC, we do not find a correlation between US population smoking rate and NSCLC incidence in both sexes, Rsq=0.43. This again reflects the non-correlation of female data between smoking rate and NSCLC incidence.
Population smoking and NSCLC (both sexes)
smoking y rate cancer y incidence
1965 42.4 1988 56.6
1970 37.4 1993 57.7
1974 37.1 1997 57
1978 34.1 2001 55.5
1980 33.2 2003 56
1983 32.1 2006 54
Rsq 0.43
Male and female adenocarcinomas
Almost all never smokers lung cancer fall into this subcategory5 of NSCLC, therefore it is not surprising to find that there is no correlation between smoking rate and adenocarcinoma incidences in either male (male adenocarcinoma accounts for about 18.5% of all lung cancer cases) or female.
Female smoking and Adenocarcinomas
smoking y rate cancer y incidence
1965 33.9 1985 14.7
1970 31.5 1990 17.8
1974 32.1 1994 18.8
1978 30.7 1998 20.3
1980 29.3 2000 18.7
1983 29.5 2003 18.8
1985 27.9 2005 18.6
RSQ 0.4**
** negative correlation
Male smoking and Adenocarcinomas
smoking y rate cancer y incidence
1965 51.9 1985 24.9
1970 44.1 1990 26
1974 43.1 1994 27.5
1978 38.1 1998 27.6
1980 37.6 2000 24.7
1983 35.1 2003 23.8
1985 32.6 2005 21.5
RSQ 0.2**
** negative correlation
Population smoking and combined male and female Adenocarcinomas
smoking y rate cancer y incidence
1965 42.4 1985 19
1970 37.4 1990 21.2
1974 37.1 1994 22.4
1978 34.1 1998 23.3
1980 33.2 2000 21.2
1983 32.1 2003 20.8
1985 30.1 2005 19.8
RSQ 0.04
** negative correlation
Overall lung cancer incidence
As can be expected, US male overall lung cancer incidence (both SC and NSC combined) correlates well with male smoking rate, at Rsq=0.95, with a 19-year response lapse. However, there is no positive correlation between female smoking rate and female total LC incidence, which has remained flat for the past decade after reaching plateau.
Lung cancer incidences, age-adjusted
LC O M F
1973 49 85.9 20.9
1974 50.6 87.3 22.8
1975 52.3 89.6 24.5
1976 55.4 93.8 27.3
1977 56.7 95.5 28.3
1978 57.8 96.7 29.7
1979 58.6 95.8 31.6
1980 60.7 100 32.2
1981 62 99.4 35.1
1982 63.3 100.7 36.7
1983 63.4 99.5 37.9
1984 65.5 102.1 39.5
1985 64.6 98.7 40.2
1986 65.8 99.1 42.3
1987 67.9 101.5 44.1
1988 68 98.7 46.3
1989 67.5 97.9 46.1
1990 68 96.9 47.7
1991 69.2 97.3 49.6
1992 69.4 97.2 49.8
1993 67.8 93.9 49.2
1994 67.1 90.9 50.5
1995 66.8 89.7 50.4
1996 66.4 87.9 51.1
1997 66.6 86.2 52.5
1998 67.5 87.9 52.9
1999 65.8 84.5 52.4
2000 64 81.9 51.1
2001 63.9 80.9 51.5
2002 63.6 79.8 52.1
2003 64.1 80.1 52.6
2004 61.3 74.9 51.3
2005 61 73.2 52.2
2006e 61 71 52
e - estimated; peak incidences shown in bold
Male smoking and overall LC incidence (both SC and NSC)
smoking y rate cancer y incidence
1965 51.9 1984 102.1
1970 44.1 1989 97.9
1974 43.1 1993 93.9
1978 38.1 1997 86.2
1980 37.6 1999 84.5
1983 35.1 2002 79.8
1985 32.6 2004 74.9
1987 31.2 2006 71
Rsq 0.95*
Also, when applying the 23-year response lapse, correlation between overall population smoking rate and lung cancer incidence turned out to be rather weak, Rsq=0.77. This again showed the dilution of correlation effect with non-compliant female data.
Population smoking and overall LC incidence (both SC and NSC)
smoking y rate cancer y incidence
1965 42.4 1988 68
1970 37.4 1993 67.8
1974 37.1 1997 66.6
1978 34.1 2001 63.9
1980 33.2 2003 64.1
1983 32.1 2006 61
Rsq 0.77
Summary
In conclusion, the author showed that cigarette smoking, while strongly responsible for small cell lung cancer in both male and female populations (11% of all lung cancers today), can NOT be majorly responsible for female non-small lung cancer (which accounts for 38% of all lung cancers today). There is no correlation between smoking rate and delayed male lung adenocarcinoma (36,000 cases, 18.3% of all lung cancers today) which corroborates very well with the fact that overwhelming majority of male never smoker lung cancer falls into this category. Therefore, even if you can assume that 100% all smoker's lung cancers are caused by their tobacco smoker exposure or smokers are immune to any other carcinogens besides tobacco, you still can not derive a total of 87% cause. The claim by both governmental agencies such as NCI 6 and non-governmental organizations such as ACS 7 that cigarette smoking causes 87% of all lung cancer cases and lung cancer deaths, is scientifically unfounded, and ethically wrong.
Additional evidence
In the UK, female lung cancer incidence (combined SC and NSC, by CancerResearchUK 8) goes against the trend in smoking rate decrease resulting in a strong negative correlation.
UK smoking rates and overall LC incidence (SC and NSC combined)
Male rate cancer y incidence
1970 55 1993 86.9
1974 51 1997 77
1978 45 2001 68.5
1982 38 2005 61.3
1986 35 2009e 54.8
Rsq 0.98*
Female rate cancer y incidence
1970 44 1993 34.8
1974 41 1997 35.4
1978 37 2001 35.6
1982 33 2005 36.8
1986 31 2009e 37.5
Rsq 0.94**
** negative correlation
References
1. Surveillance for Selected Tobacco-Use Behaviors -- United States, 1900-1994.http://www.cdc.gov/mmwr/preview/mmwrhtml/00033881.htm
2. Wogan GN, Hecht SS, et al. Environmental and chemical carcinogenesis. Semin Cancer Biol. 2004; 14:473-86.
3. Fast Stats - National Cancer Institute Surveilance Epidemiology and End Results http://seer.cancer.gov/faststats/selections.php
4. Melikian AA, Djordjevic MV, et al. Gender differences relative to smoking behavior and emissions of toxins from mainstream cigarette smoke. Nicotine Tob Res. 2007; 9:377-87.
5. Subramanian J, Govindan R. Lung cancer in never smokers: a review. J Clin Oncol. 2007; 25:561-70.
6. National Cancer Institute - Cigarette Smoking and Cancer: Questions and Answers.http://www.cancer.gov/cancertopics/factsheet/tobacco/cancer
7. American Cancer Society - Tobacco-Related Cancers Fact Sheet.http://www.cancer.org/docroot/PED/content/PED_10_2x_Tobacco-Related_Cancers_Fact_Sheet.asp
8. Cancer Research UK - UK Lung Cancer incidence statistics.http://info.cancerresearchuk.org/cancerstats/types/lung/incidence/
