Abstract
Alcohol is a carcinogen suspected of increasing lung cancer risk. Therefore, we prospectively evaluated the relationship between alcohol consumption and lung carcinoma in 492,902 persons from the National Institutes of Health-AARP Diet and Health Study. We used Cox models to calculate hazard ratios and 95% confidence intervals, adjusting for tobacco smoking and other potential confounders. Between 1995/1996 and December 31, 2006, there were 10,227 incident cases of lung carcinoma, classified as adenocarcinoma (n = 4,036), squamous cell carcinoma (n = 1,998), small cell carcinoma (n = 1,524), undifferentiated carcinoma (n = 559), and other (n = 2,110). Compared with nondrinking, alcohol consumption was associated with a modest nonlinear reduction in total lung carcinoma risk at lower levels of consumption (for 0.5–<1 drink/day, HR = 0.89, 95% confidence interval: 0.82, 0.96) but a modest increase in risk in the highest category (for ≥7 drinks/day, HR = 1.11, 95% confidence interval: 1.00, 1.24). Regarding histological type, alcohol was associated with a nonlinear reduction in squamous cell carcinoma that became attenuated as consumption increased and a modest increase in adenocarcinoma among heavier drinkers. Cubic spline models confirmed these findings. Our data suggest that the relationship between alcohol consumption and lung carcinoma differs by histological subtype.
Keywords: alcohol consumption, cohort studies, lung cancer
Lung cancer is the leading cause of cancer mortality both in the United States and worldwide (1, 2). While tobacco smoking is the leading cause of lung cancer (3), yearly an estimated 16,000 lung cancer deaths occur in Americans who have never smoked cigarettes (4). Alcohol has been ranked alongside smoking as a group 1 carcinogen by the International Agency for Research on Cancer (5). Alcohol is a cause of esophageal squamous cell carcinoma (SCC) and cancers of the head and neck, colon, rectum, and breast (5–9), and studies have suggested a possible inverse association with renal cell carcinoma (6, 10–12). However, tobacco smoking is potentially a strong confounder of alcohol exposure, and whether alcohol is an independent risk factor for lung cancer remains unclear.
Several studies have evaluated the relationship between alcoholic beverage consumption and lung cancer (13–27), although most have been imited by small sample sizes (27), restricting the ability to detect modest effect sizes. Analyses stratified according to the 4 primary histological subtypes of lung carcinoma—adenocarcinoma, SCC, small cell carcinoma, and undifferentiated carcinoma—have had variable results (16–25, 27), with some studies finding no difference by histological subtype (22, 23, 25) and others reporting differences in the association (13, 16, 19, 21, 25). Investigating histological subtype may be important because of changing patterns of lung cancer incidence by histology over time and because incidence rates of certain subtypes differ between men and women (28–30). For example, SCC, the subtype most highly associated with tobacco smoking (31), has been surpassed by adenocarcinoma as the most common histological subtype in the United States (28–30)—a change which has occurred along with declines in smoking rates and changes in cigarette composition, such as the addition of filters and reconstituted tobacco (29, 32). In addition, it may also be important to assess associations for each type of alcoholic beverage separately. In studies which differentiate between beverage types (13–15, 17, 18, 20, 22–25, 27), risk often appears higher for beer and liquor, while possibly being lower for wine (13, 15, 17, 18, 27).
We investigated associations between alcoholic beverage consumption and risk of lung carcinoma in a US prospective cohort study with detailed information on alcohol consumption and other covariates. The large size of our cohort allowed sufficient statistical power to examine potential differences by lung carcinoma histological subtype, smoking history, and type of alcoholic beverage consumed. We hypothesized that, after adjusting for tobacco smoking, greater total alcohol consumption would be associated with higher risk of lung carcinoma.
METHODS
Study design and population
The National Institutes of Health-AARP (NIH-AARP) Diet and Health Study is a prospective cohort study of 566,398 members of AARP (formerly the American Association of Retired Persons) who completed a comprehensive questionnaire between 1995 and 1996. Participants were aged 50–71 years at baseline and resided in California, Florida, Louisiana, New Jersey, North Carolina, Pennsylvania and the metropolitan areas of Atlanta, Georgia and Detroit, Michigan (33). The Special Studies Institutional Review Board of the National Cancer Institute approved the study. All participants provided written informed consent by completing and returning the self-administered baseline questionnaire.
We excluded 15,760 participants whose questionnaires were completed by a proxy, 51,234 persons with a history of cancer at baseline, 4,417 reporting extreme caloric intake, 46 who died before their questionnaires had been scanned, and 2,039 with missing alcohol data. Our final analytical sample consisted of 492,902 participants.
Outcome assessment
Incident cancers were identified through probabilistic linkage to state cancer registries (33). Case ascertainment has been estimated to be nearly 90% complete (34). Lung cancer cases were defined as those involving first primary lung carcinoma and were categorized by histological subtype: adenocarcinoma, SCC, small cell carcinoma, and undifferentiated carcinoma. Histological types were defined using International Classification of Diseases for Oncology (ICDO) codes. Lung carcinomas included adenocarcinoma (ICDO codes 8140, 8200, 8250, 8251, 8252, 8253, 8254, 8255, 8260, 8310, 8323, 8430, 8480, 8481, 8490, 8550, and 8574), SCC (ICDO codes 8050, 8052, 8070, 8071, 8072, 8073, 8074, 8075, 8083, and 8084), small cell carcinoma (ICDO codes 8002, 8041, 8042, 8043, 8044, and 8045), undifferentiated carcinoma (ICDO codes 8012, 8020, 8021, 8022, 8031, and 8032), lung carcinoma not otherwise specified (ICDO code 8046), and other carcinomas (ICDO codes 8010, 8011, 8030, 8033, 8123, 8560, 8562, and 8575) (35).
Exposure assessment
Typical alcohol consumption over the last 12 months was reported on the baseline questionnaire. Participants provided information on the frequency and quantity of consumption of beer during the summer, beer during the rest of the year, wine and wine coolers, and liquor and mixed drinks. Frequency categories included never, ≤1 time/month, 2–3 times/month, 1–2 times/week, 3–4 times/week, 5–6 times/week, 1 time/day, 2–3 times/day, 4–5 times/day, and ≥6 times/day. Quantity was categorized by beverage type (beer: <1 12-ounce (355-mL) can, 1–2 12-ounce cans, or >2 12-ounce cans; wine: <4 ounces (<118 mL), 4–8 ounces (118–237 mL), or >8 ounces (>237 mL); liquor and mixed drinks: <1 shot (<1.5 ounces or <42.6 mL), 1–2 shots, or >2 shots). These measures were used to calculate the number of standard drinks consumed per day based on the US Department of Agriculture MyPyramid Serving Equivalents Database (36). One standard drink contains 14 g of pure ethanol and is equivalent to 12 fluid ounces (355 mL) of beer, 5 fluid ounces (148 mL) of wine, and 1.5 fluid ounces (42.6 mL) of liquor (37). Binge drinking was not assessed in the questionnaire. We restricted our analysis to alcohol consumed as a beverage and did not include alcohol used in cooking.
Cigarette smoking was also reported on the baseline questionnaire. Participants answered the following questions: Have you smoked 100 or more cigarettes during your entire life (yes, no)? Do you currently smoke cigarettes or have you stopped (currently smoke, stopped within last year, stopped 1–4 years ago, stopped 5–9 years ago, stopped ≥10 years ago)? How many cigarettes per day do or did you usually smoke (1–10, 11–20, 21–30, 31–40, 41–60, ≥61)? For our cigarette smoking adjustment, we converted cigarette smoking history to a categorical variable by cross-classifying cigarette smoking status (current, former, or never smoker), average number of cigarettes smoked per day (1–10, 11–20, 21–30, 31–40, 41–60, ≥61), and, for former cigarette smokers, years since cessation (<1 (within last year), 1–4, 5–9, or ≥10). Because of frequent relapses among recent quitters, we classified participants reporting cessation in the last year as current smokers. Pipe and cigar smoking was assessed by asking the following question: Did you ever smoke pipes or cigars regularly for a year or longer (no; yes, pipes and cigars; yes, pipes only; yes, cigars only)? This yes-no variable was also included in our models.
Statistical analysis
Associations between alcoholic beverage consumption and lung carcinoma could be nonlinear. Therefore, we evaluated linearity in the relationship between alcohol and lung carcinoma risk in our data. We first analyzed total alcohol consumption in the following categories: none (0 drinks/day), >0–<0.5 drink/day, 0.5–<1.0 drink/day, 1.0–<3.0 drinks/day, 3.0–<5.0 drinks/day, 5.0–<7.0 drinks/day, and ≥7 drinks/day. These categories were chosen a priori based on multiples of a single alcoholic drink used in previous analyses of the NIH-AARP cohort. Given our large cohort and the number of lung cancer cases, we were able to further subdivide the lowest drinking category (0–<1.0 drink/day) to better characterize modest alcohol consumption. Nondrinkers served as the referent group. Since heavy alcohol consumption might carry the largest risk for lung carcinoma, we examined alcohol consumption of 3–<5, 5–<7, and ≥7 drinks/day separately, although numbers of participants in the high intake categories were low. Second, we assessed potential nonlinear associations using restricted cubic spline models with continuous drinks per day (38).
We estimated hazard ratios and 95% confidence intervals using multivariable Cox proportional hazards regression models with person-years of follow-up time as the underlying time metric. All tests were 2-sided. Follow-up time extended from the day the returned questionnaire was scanned until first primary lung cancer diagnosis, death, movement out of the registry ascertainment area, or the end of follow-up (December 31, 2006). All models adjusted for sex, age at baseline, cigarette smoking (categorical variable), pipe and cigar smoking (yes, no), education (less than high school; completion of high school; post–high school training other than college; some college; college and postgraduate), physical activity at work (sitting during the day/little walking; sitting much of the day/walking a fair amount; standing/walking a lot—no lifting; lifting/carrying light loads/climbing stairs/climbing hills; doing heavy work/carrying loads), leisure-time physical activity (never, rarely, 1–3 times/month, 1–2 times/week, 3–4 times/week, ≥5 times/week); energy intake (excluding alcohol; kcal/day); body mass index (weight (kg)/height (m)2; <18.5, 18.5–<25, 25–<30, 30–<35, ≥35); race/ethnicity (non-Hispanic white; non-Hispanic black; Hispanic; Asian, Pacific Islander, or American Indian/Alaska Native), and the subject's Healthy Eating Index-2010 score, which was modified to exclude alcohol (39).
The proportional hazards assumption was tested for total lung carcinoma, and we found no violation (P = 0.136). In addition to total lung carcinoma, we also evaluated the 4 major histological subtypes: adenocarcinoma, SCC, small cell carcinoma, and undifferentiated carcinoma. Given that smoking is strongly associated with both alcohol and lung cancer, we included results stratified by smoking status (current, former, or never smoker), irrespective of statistical significance. For analyses of the individual types of alcoholic beverages, we controlled for intake of the other kinds of beverages (each beverage was modeled using the following categories: 0 drinks/day, >0–<0.5 drink/day, 0.5–<1.0 drink/day, 1.0–<3.0 drinks/day, 3.0–<5.0 drinks/day, 5.0–<7.0 drinks/day, or ≥7 drinks/day), using nondrinkers of any alcoholic beverage as the referent category. All analyses were conducted using SAS, version 9.2 TS2M3 (SAS Institute, Inc., Cary, North Carolina).
RESULTS
After exclusions, our analytical cohort was 59.7% male (n =294,246) and 40.3% female (n = 198,656). The mean age of cohort members at baseline was 62.0 years, and the mean age at diagnosis for total lung carcinoma was 69.5 years. The most common racial/ethnic group was non-Hispanic white (92.5%). Most participants had a history of smoking (14.3% current smokers, 49.3% former smokers, 36.5% never smokers). Alcohol was commonly consumed, with 75.9% consuming some alcohol in the 12 months before baseline, 52.9% consuming >0–<1 drink/day, 15.4% consuming 1–<3 drinks/day, and 7.6% consuming ≥3 drinks/day. Consumption of 3 or more alcoholic beverages per day was more common in males (10.9%) than in females (2.7%). Beer, wine, and liquor were consumed by 54.5%, 60.7%, and 53.3% of our population, respectively. Compared with nondrinkers (n = 119,025), higher proportions of alcohol drinkers were ever smokers (66.3% vs. 54.9%) and had post–high school education (95.2% vs. 89.9%) (Table 1).
Table 1.
Selected Characteristics of 492,902 Participants in the National Institutes of Health-AARP Diet and Health Study, by Category of Alcohol Consumption, 1995/1996–2006
| Characteristic | Total Alcohol Consumption, drinks/day |
|||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| None (0) |
0.01–0.49 |
0.50–0.99 |
1.00–2.99 |
3.00–4.99 |
5.00–6.99 |
≥7.00 |
||||||||
| No.a | % | No. | % | No. | % | No. | % | No. | % | No. | % | No. | % | |
| Total cohort | 119,025 | 24.2 | 211,230 | 42.9 | 49,601 | 10.1 | 75,746 | 15.4 | 19,176 | 3.9 | 8,530 | 1.7 | 9,594 | 2.0 |
| Age, yearsb | 62.3 (5.3) | 61.8 (5.4) | 61.9 (5.4) | 62.3 (5.3) | 62.0 (5.3) | 62.4 (5.3) | 61.9 (5.2) | |||||||
| Sex | ||||||||||||||
| Male | 60,781 | 20.7 | 112,921 | 38.4 | 33,224 | 11.3 | 55,355 | 18.8 | 15,822 | 5.4 | 7,522 | 2.6 | 8,621 | 2.9 |
| Female | 58,244 | 29.3 | 98,309 | 49.5 | 16,377 | 8.2 | 20,391 | 10.3 | 3,354 | 1.7 | 1,008 | 0.5 | 973 | 0.5 |
| Cigarette smoking history | ||||||||||||||
| Never smoker | 51,570 | 29.8 | 80,599 | 46.6 | 15,811 | 9.1 | 19,119 | 11.1 | 3,311 | 1.9 | 1,314 | 0.8 | 1,287 | 0.7 |
| Former smoker | ||||||||||||||
| ≤20 cigarettes/day | 25,973 | 19.8 | 57,247 | 43.6 | 15,529 | 11.8 | 23,602 | 18.0 | 4,724 | 3.6 | 2,105 | 1.6 | 2,080 | 1.6 |
| >20 cigarettes/day | 22,180 | 21.6 | 37,970 | 37.0 | 10,633 | 10.4 | 19,476 | 19.0 | 6,071 | 5.9 | 2,949 | 2.9 | 3,292 | 3.2 |
| Current smoker | ||||||||||||||
| ≤20 cigarettes/day | 9,335 | 21.3 | 19,113 | 43.7 | 4,054 | 9.3 | 6,926 | 15.8 | 2,279 | 5.2 | 914 | 2.1 | 1,130 | 2.6 |
| >20 cigarettes/day | 5,376 | 22.5 | 8,655 | 36.2 | 1,704 | 7.1 | 3,856 | 16.1 | 2,022 | 8.5 | 921 | 3.9 | 1,409 | 5.9 |
| Regular pipe/cigar smoker | 14,616 | 17.2 | 31,325 | 36.8 | 10,184 | 12.0 | 18,320 | 21.5 | 5,207 | 6.1 | 2,591 | 3.0 | 2,972 | 3.5 |
| Education | ||||||||||||||
| Less than high school | 11,526 | 39.5 | 11,118 | 38.1 | 1,822 | 6.2 | 2,817 | 9.7 | 885 | 3.0 | 308 | 1.1 | 708 | 2.4 |
| Completion of high school | 30,059 | 31.2 | 42,225 | 43.9 | 7,192 | 7.5 | 10,714 | 11.1 | 3,073 | 3.2 | 1,110 | 1.2 | 1,868 | 1.9 |
| Some post–high school training | 38,235 | 23.5 | 72,670 | 44.6 | 15,890 | 9.7 | 23,702 | 14.5 | 6,524 | 4.0 | 2,707 | 1.7 | 3,346 | 2.1 |
| Some college | 17,316 | 18.6 | 38,365 | 41.3 | 11,030 | 11.9 | 17,720 | 19.1 | 4,413 | 4.8 | 2,236 | 2.4 | 1,835 | 2.0 |
| College/postgraduate school | 17,499 | 18.0 | 40,881 | 41.9 | 12,509 | 12.8 | 19,196 | 19.7 | 3,833 | 3.9 | 2,007 | 2.1 | 1,567 | 1.6 |
| Usual daily physical activity | ||||||||||||||
| Sitting during the day/little walking | 10,159 | 26.1 | 17,459 | 44.8 | 3,423 | 8.8 | 5,133 | 13.2 | 1,373 | 3.5 | 670 | 1.7 | 720 | 1.9 |
| Sitting during the day/walking a fair amount | 36,175 | 22.8 | 70,483 | 44.4 | 16,449 | 10.4 | 24,126 | 15.2 | 5,896 | 3.7 | 2,711 | 1.7 | 2,773 | 1.8 |
| Standing/walking a lot—no lifting | 44,785 | 24.1 | 78,354 | 42.2 | 19,157 | 10.3 | 29,192 | 15.7 | 7,336 | 4.0 | 3,245 | 1.8 | 3,633 | 2.0 |
| Lifting/carrying light loads, climbing stairs, climbing hills | 20,074 | 23.7 | 34,897 | 41.3 | 8,586 | 10.2 | 14,085 | 16.7 | 3,583 | 4.2 | 1,506 | 1.8 | 1,814 | 2.2 |
| Doing heavy work/carrying loads | 4,130 | 29.1 | 5,496 | 38.8 | 1,125 | 7.9 | 1,988 | 14.0 | 691 | 4.9 | 273 | 1.9 | 478 | 3.4 |
| Frequency of (leisure-time) physical activity | ||||||||||||||
| Never | 8,775 | 40.2 | 8,224 | 37.6 | 1,203 | 5.5 | 2,096 | 9.6 | 735 | 3.4 | 310 | 1.4 | 513 | 2.4 |
| Rarely | 19,305 | 28.9 | 29,554 | 44.3 | 4,904 | 7.4 | 7,676 | 11.5 | 2,533 | 3.8 | 1,176 | 1.8 | 1,613 | 2.4 |
| 1–3 times/month | 15,295 | 22.9 | 30,549 | 45.7 | 6,463 | 9.7 | 9,314 | 13.9 | 2,680 | 4.0 | 1,187 | 1.8 | 1,367 | 2.0 |
| 1–2 times/week | 23,055 | 21.8 | 47,029 | 44.4 | 11,255 | 10.6 | 16,707 | 15.8 | 4,191 | 4.0 | 1,842 | 1.7 | 1,940 | 1.8 |
| 3–4 times/week | 28,310 | 21.5 | 56,458 | 42.8 | 15,183 | 11.5 | 22,417 | 17.0 | 5,098 | 3.9 | 2,211 | 1.7 | 2,220 | 1.7 |
| ≥5 times/week | 22,606 | 23.9 | 37,415 | 39.5 | 10,208 | 10.8 | 17,013 | 18.0 | 3,803 | 4.0 | 1,746 | 1.8 | 1,845 | 2.0 |
| Body mass indexc | ||||||||||||||
| ≤18.49 | 5,427 | 32.2 | 6,755 | 40.1 | 1,382 | 8.2 | 2,169 | 12.9 | 577 | 3.4 | 225 | 1.3 | 306 | 1.8 |
| 18.50–24.99 | 36,478 | 22.1 | 69,427 | 42.0 | 18,863 | 11.4 | 28,979 | 17.5 | 6,190 | 3.7 | 2,605 | 1.6 | 2,894 | 1.8 |
| 25.00–29.99 | 45,950 | 22.4 | 86,576 | 42.3 | 21,522 | 10.5 | 33,335 | 16.3 | 8,975 | 4.4 | 4,064 | 2.0 | 4,478 | 2.2 |
| 30.00–34.99 | 20,480 | 27.1 | 34,255 | 45.3 | 6,288 | 8.3 | 9,128 | 12.1 | 2,723 | 3.6 | 1,320 | 1.7 | 1,514 | 2.0 |
| ≥35.00 | 10,690 | 35.6 | 14,217 | 47.4 | 1,546 | 5.2 | 2,135 | 7.1 | 711 | 2.4 | 316 | 1.1 | 402 | 1.3 |
| Race/ethnicity | ||||||||||||||
| Non-Hispanic white | 104,357 | 23.2 | 192,475 | 42.8 | 46,412 | 10.3 | 71,675 | 15.9 | 18,115 | 4.0 | 8,150 | 1.8 | 9,013 | 2.0 |
| Non-Hispanic black | 7,477 | 39.3 | 7,835 | 41.2 | 1,246 | 6.6 | 1,533 | 8.1 | 502 | 2.6 | 155 | 0.8 | 274 | 1.4 |
| Hispanic | 2,201 | 23.4 | 4,568 | 48.5 | 960 | 10.2 | 1,197 | 12.7 | 259 | 2.8 | 101 | 1.1 | 131 | 1.4 |
| Asian, Pacific Islander, or American Indian/Alaska Native | 3,036 | 37.4 | 3,682 | 45.4 | 499 | 6.2 | 643 | 7.9 | 128 | 1.6 | 59 | 0.7 | 70 | 0.9 |
| Healthy Eating Indexd | ||||||||||||||
| Quartile 1 | 35,073 | 28.5 | 55,556 | 45.1 | 9,756 | 7.9 | 14,132 | 11.5 | 4,090 | 3.3 | 1,435 | 1.2 | 3,183 | 2.6 |
| Quartile 2 | 29,115 | 23.6 | 53,349 | 43.3 | 12,029 | 9.8 | 18,033 | 14.6 | 5,107 | 4.1 | 2,345 | 1.9 | 3,248 | 2.6 |
| Quartile 3 | 27,341 | 22.2 | 51,740 | 42.0 | 13,478 | 10.9 | 20,759 | 16.9 | 5,206 | 4.2 | 2,563 | 2.1 | 2,139 | 1.7 |
| Quartile 4 | 27,496 | 22.3 | 50,585 | 41.1 | 14,338 | 11.6 | 22,822 | 18.5 | 4,773 | 3.9 | 2,187 | 1.8 | 1,024 | 0.8 |
| Energy intake, kcal/dayd,e | 1,617.0 (948.8) | 1,575.7 (872.6) | 1,606.8 (841.5) | 1,644.8 (852.7) | 1,743.0 (926.0) | 1,754.1 (912.7) | 2,071.8 (1,075.5) | |||||||
a Numbers may not sum to totals because of missing values.
b Values are presented as mean (standard deviation).
c Weight (kg)/height (m)2.
d Excluding alcohol.
e Values are presented as median (interquartile range).
During follow-up, 10,227 participants were diagnosed with lung carcinoma. The malignancies were classified as adenocarcinoma (n = 4,036), SCC (n = 1,998), small cell carcinoma (n = 1,524), undifferentiated carcinoma (n = 559), and other or not otherwise specified (n = 2,110).
Overall, we observed alcohol consumption (versus none) to be associated with a modest reduction in risk for lung carcinoma at lower levels (>0–<0.5 drink/day: hazard ratio (HR) = 0.92 (95% confidence interval (CI): 0.87, 0.97); 0.5–<1 drink/day: HR = 0.89 (95% CI: 0.82, 0.96); 1–<3 drinks/day: HR = 0.94 (95% CI: 0.88, 1.00); global P < 0.001) (Table 2) and suggestive of a modest increase in risk in the highest category (≥7 drinks/day: HR = 1.11, 95% CI: 1.00, 1.24). In histology-specific models, we observed an inverse association for SCC (>0–<0.5 drink/day: HR = 0.81 (95% CI: 0.73, 0.91); 0.5–<1 drink/day: HR = 0.70 (95% CI: 0.58, 0.84); 1–<3 drinks/day: HR = 0.81 (95% CI: 0.70, 0.93); 3–<5 drinks/day: HR = 0.90 (95% CI: 0.74, 1.09); 5–<7 drinks/day: HR = 0.79 (95% CI: 0.60, 1.05); ≥7 drinks/day: HR = 0.99 (95% CI: 0.79, 1.24); global P < 0.001) but an elevation in risk for adenocarcinoma in the highest category of consumption (≥7 drinks/day: HR = 1.30, 95% CI: 1.09, 1.55; global P = 0.116). No association was found for small cell carcinoma. Undifferentiated carcinoma followed a pattern similar to that of SCC, but our study was underpowered to detect all but a strong association.
Table 2.
Adjusted Hazard Ratios for the Relationship Between Total Alcohol Consumption and Lung Carcinoma, by Histological Subtype, Among 492,902 Participants in the National Institutes of Health-AARP Diet and Health Study, 1995/1996–2006
| Total Alcohol Consumption, drinks/day | Histological Type |
||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Total Lung Cancer |
Adenocarcinoma |
Squamous Cell Carcinoma |
Small Cell Carcinoma |
Undifferentiated Carcinoma |
|||||||||||
| No. of Cases | HRa | 95% CI | No. of Cases | HRa | 95% CI | No. of Cases | HRa | 95% CI | No. of Cases | HRa | 95% CI | No. of Cases | HRa | 95% CI | |
| None (0) | 2,442 | 1.00 | Referent | 891 | 1.00 | Referent | 525 | 1.00 | Referent | 372 | 1.00 | Referent | 147 | 1.00 | Referent |
| 0.01–0.49 | 3,908 | 0.92 | 0.87, 0.97 | 1,615 | 1.02 | 0.94, 1.11 | 730 | 0.81 | 0.73, 0.91 | 583 | 0.91 | 0.80, 1.03 | 197 | 0.76 | 0.62, 0.95 |
| 0.50–0.99 | 882 | 0.89 | 0.82, 0.96 | 361 | 0.96 | 0.85, 1.09 | 146 | 0.70 | 0.58, 0.84 | 132 | 0.95 | 0.78, 1.16 | 54 | 0.88 | 0.64, 1.20 |
| 1.00–2.99 | 1,700 | 0.94 | 0.88, 1.00 | 699 | 1.06 | 0.95, 1.17 | 320 | 0.81 | 0.70, 0.93 | 231 | 0.90 | 0.76, 1.06 | 91 | 0.80 | 0.61, 1.05 |
| 3.00–4.99 | 624 | 1.01 | 0.92, 1.10 | 226 | 1.07 | 0.93, 1.25 | 132 | 0.90 | 0.74, 1.09 | 101 | 1.05 | 0.84, 1.32 | 35 | 0.88 | 0.60, 1.28 |
| 5.00–6.99 | 272 | 0.96 | 0.85, 1.09 | 96 | 1.02 | 0.82, 1.26 | 55 | 0.79 | 0.60, 1.05 | 40 | 0.93 | 0.67, 1.30 | 15 | 0.81 | 0.47, 1.39 |
| ≥7.00 | 399 | 1.11 | 1.00, 1.24 | 148 | 1.30 | 1.09, 1.55 | 90 | 0.99 | 0.79, 1.24 | 65 | 1.14 | 0.87, 1.49 | 20 | 0.87 | 0.54, 1.39 |
| P for trendb | 0.002 | 0.007 | 0.437 | 0.151 | 0.993 | ||||||||||
| Global P (6 df) | <0.001 | 0.116 | <0.001 | 0.434 | 0.386 | ||||||||||
Abbreviations: CI, confidence interval; df, degrees of freedom; HR, hazard ratio.
a Adjusted for sex, age at baseline, cigarette smoking (categorical variable), pipe and cigar smoking (yes, no), education (less than high school; completion of high school; post–high school training other than college; some college; college and postgraduate), physical activity at work (sitting during the day/little walking; sitting much of the day/walking a fair amount; standing/walking a lot—no lifting; lifting/carrying light loads/climbing stairs/climbing hills; doing heavy work/carrying loads), leisure-time physical activity (never, rarely, 1–3 times/month, 1–2 times/week, 3–4 times/week, ≥5 times/week); energy intake (excluding alcohol; kcal/day); body mass index (weight (kg)/height (m)2; ≤18.49, 18.5–24.99, 25.00–29.99, 30.00–34.99, ≥35.00); race/ethnicity (non-Hispanic white; non-Hispanic black; Hispanic; Asian, Pacific Islander, or American Indian/Alaska Native), and Healthy Eating Index-2010 score, which was modified to exclude alcohol (39). We adjusted for cigarette smoking using a detailed categorical variable which included cigarette smoking status, average number of cigarettes smoked per day, and, for former cigarette smokers, years since cessation.
b We tested for trend using an ordinal variable containing the median value of each alcohol category (0 drinks/day, 0.10 drinks/day, 0.69 drinks/day, 1.47 drinks/day, 3.66 drinks/day, 6.10 drinks/day, and 11.50 drinks/day, respectively).
To further explore nonlinear relationships, we fitted restricted cubic spline models (Figure 1). These models also suggested a nonlinear inverse association for total lung carcinoma at lower levels of alcohol consumption, with a modest increase in risk for adenocarcinoma and an inverse association with SCC. The Akaike Information Criterion values for these spline models were lower than those for the models that modeled alcohol as a categorical variable (Table 2). Examination of the spline models showed no evidence of nonlinear associations for small cell carcinoma or undifferentiated carcinoma.
Figure 1.
Hazard ratios for the association between total alcohol consumption (drinks/day) and lung carcinoma in the National Institutes of Health-AARP Diet and Health Study, 1995/1996–2006. A) Total lung cancer; B) adenocarcinoma; C) squamous cell carcinoma. Estimates were obtained from restricted cubic spline models, and nondrinkers comprised the reference group in all analyses. Knot locations were derived from exposure percentiles (5th, 27.5th, 50th, 72.5th, and 95th percentile levels) and occurred at 0 drinks/day, 0.02 drinks/day, 0.13 drinks/day, 0.72 drinks/day, and 4.03 drinks/day, respectively. All models adjusted for sex, age at baseline, education, physical activity at work, leisure-time physical activity, energy intake (excluding alcohol; kcal/day), body mass index, race/ethnicity, Healthy Eating Index-2010 score (excluding alcohol), and tobacco smoking. The curves for all spline models were significant overall (P < 0.0001). Nonlinear associations were observed for total lung carcinoma (P for nonlinearity = 0.003) and squamous cell carcinoma (P for nonlinearity <0.001) but not for adenocarcinoma (P for nonlinearity = 0.694). Dashed lines, 95% confidence intervals.
There was no statistical evidence for interaction between number of alcoholic drinks per day (continuous) and tobacco smoking (categorical: current, former, or never) (P = 0.164). When results were stratified by smoking status, the inverse associations with total lung cancer and SCC were similar in current and former smokers, but they appeared stronger among former smokers who had smoked >20 cigarettes/day and in lower drinking categories (Table 3). Nearly all SCC cases occurred in current or former smokers (94.1%), so we had minimal statistical power to test for an association in never smokers. The elevation in adenocarcinoma risk appeared stronger in former smokers who had smoked ≤20 cigarettes/day, but it was not present in never smokers or current smokers.
Table 3.
Adjusted Hazard Ratiosa for the Relationship Between Total Alcohol Consumption and Lung Carcinoma, by Histological Subtype and Smoking Status, Among 492,902 Participants in the National Institutes of Health-AARP Diet and Health Study, 1995/1996–2006
| Histological Type | Total Alcohol Consumption, drinks/day |
P for Trendb | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| None (0) |
0.01–0.49 |
0.50–0.99 |
1.00–2.99 |
≥3.00 |
||||||||||||
| No. of Cases | HR | 95% CI | No. of Cases | HR | 95% CI | No. of Cases | HR | 95% CI | No. of Cases | HR | 95% CI | No. of Cases | HR | 95% CI | ||
| Never Smoker | ||||||||||||||||
| (n = 51,570) | (n = 80,599) | (n = 15,811) | (n = 19,119) | (n = 1,287) | ||||||||||||
| Total lung cancer | 148 | 1.00 | Referent | 253 | 1.16 | 0.95, 1.43 | 39 | 0.92 | 0.64, 1.32 | 72 | 1.37 | 1.02, 1.84 | 13 | 0.81 | 0.46, 1.44 | 0.927 |
| Adenocarcinoma | 90 | 1.00 | Referent | 152 | 1.13 | 0.87, 1.48 | 22 | 0.86 | 0.54, 1.39 | 42 | 1.36 | 0.92, 1.99 | 8 | 0.91 | 0.44, 1.91 | 0.440 |
| SCC | 12 | 1.00 | Referent | 15 | 1.21 | 0.45, 3.24 | 5 | 1.18 | 0.40, 3.46 | 7 | 1.21 | 0.45, 3.24 | 0 | 0.413 | ||
| Small cell carcinoma | 11 | 1.00 | Referent | 18 | 1.20 | 0.55, 2.58 | 1 | 0.32 | 0.04, 2.55 | 4 | 0.98 | 0.29, 3.26 | 2 | 1.44 | 0.30, 6.85 | 0.966 |
| Undifferentiated carcinoma | 10 | 1.00 | Referent | 10 | 0.64 | 0.26, 1.57 | 3 | 0.96 | 0.25, 3.66 | 2 | 0.53 | 0.11, 2.57 | 1 | 0.72 | 0.09, 5.95 | 0.604 |
| Former Smoker of ≤20 Cigarettes/Day | ||||||||||||||||
| (n = 25,973) | (n = 57,247) | (n = 15,529) | (n = 23,602) | (n = 8,909) | ||||||||||||
| Total lung cancer | 359 | 1.00 | Referent | 711 | 0.99 | 0.87, 1.12 | 159 | 0.84 | 0.70, 1.02 | 293 | 0.97 | 0.83, 1.14 | 142 | 1.14 | 0.94, 1.40 | 0.067 |
| Adenocarcinoma | 147 | 1.00 | Referent | 336 | 1.11 | 0.91, 1.35 | 82 | 1.04 | 0.79, 1.36 | 140 | 1.13 | 0.89, 1.43 | 73 | 1.51 | 1.13, 2.01 | 0.031 |
| SCC | 67 | 1.00 | Referent | 115 | 0.90 | 0.67, 1.23 | 21 | 0.63 | 0.39, 1.04 | 53 | 1.00 | 0.69, 1.45 | 24 | 0.95 | 0.59, 1.53 | 0.496 |
| Small cell carcinoma | 48 | 1.00 | Referent | 77 | 0.81 | 0.56, 1.16 | 12 | 0.49 | 0.26, 0.94 | 23 | 0.59 | 0.35, 0.98 | 12 | 0.69 | 0.36, 1.32 | 0.763 |
| Undifferentiated carcinoma | 27 | 1.00 | Referent | 31 | 0.56 | 0.33, 0.94 | 7 | 0.48 | 0.21, 1.12 | 12 | 0.50 | 0.25, 1.02 | 7 | 0.69 | 0.29, 1.61 | 0.832 |
| Former Smoker of >20 Cigarettes/Day | ||||||||||||||||
| (n = 22,180) | (n = 37,970) | (n = 10,633) | (n = 19,476) | (n = 12,312) | ||||||||||||
| Total lung cancer | 786 | 1.00 | Referent | 1,109 | 0.87 | 0.80, 0.96 | 296 | 0.87 | 0.76, 1.00 | 583 | 0.91 | 0.81, 1.01 | 446 | 1.03 | 0.91, 1.15 | 0.035 |
| Adenocarcinoma | 313 | 1.00 | Referent | 489 | 0.95 | 0.82, 1.09 | 129 | 0.92 | 0.75, 1.13 | 252 | 0.95 | 0.80, 1.12 | 192 | 1.11 | 0.93, 1.34 | 0.060 |
| SCC | 190 | 1.00 | Referent | 210 | 0.69 | 0.57, 0.85 | 48 | 0.59 | 0.43, 0.81 | 119 | 0.76 | 0.60, 0.96 | 85 | 0.77 | 0.59, 1.00 | 0.676 |
| Small cell carcinoma | 80 | 1.00 | Referent | 136 | 1.07 | 0.81, 1.42 | 39 | 1.28 | 0.86, 1.88 | 64 | 1.12 | 0.80, 1.57 | 50 | 1.26 | 0.88, 1.81 | 0.105 |
| Undifferentiated carcinoma | 52 | 1.00 | Referent | 54 | 0.64 | 0.43, 0.93 | 19 | 0.80 | 0.47, 1.37 | 24 | 0.54 | 0.33, 0.88 | 30 | 0.99 | 0.62, 1.57 | 0.302 |
| Current Smoker of ≤20 Cigarettes/Day | ||||||||||||||||
| (n = 9,335) | (n = 19,113) | (n = 4,054) | (n = 6,926) | (n = 4,323) | ||||||||||||
| Total lung cancer | 575 | 1.00 | Referent | 993 | 0.88 | 0.79, 0.97 | 214 | 0.91 | 0.78, 1.07 | 378 | 0.90 | 0.78, 1.02 | 251 | 0.92 | 0.79, 1.07 | 0.313 |
| Adenocarcinoma | 182 | 1.00 | Referent | 364 | 1.00 | 0.84, 1.20 | 74 | 0.99 | 0.76, 1.31 | 139 | 1.06 | 0.85, 1.33 | 81 | 1.00 | 0.76, 1.30 | 0.278 |
| SCC | 127 | 1.00 | Referent | 191 | 0.78 | 0.62, 0.98 | 35 | 0.69 | 0.47, 1.00 | 70 | 0.75 | 0.55, 1.01 | 64 | 1.00 | 0.73, 1.36 | 0.439 |
| Small cell carcinoma | 107 | 1.00 | Referent | 189 | 0.89 | 0.70, 1.13 | 45 | 1.04 | 0.73, 1.48 | 60 | 0.78 | 0.56, 1.07 | 44 | 0.88 | 0.62, 1.27 | 0.758 |
| Undifferentiated carcinoma | 32 | 1.00 | Referent | 48 | 0.77 | 0.49, 1.21 | 14 | 1.08 | 0.57, 2.04 | 29 | 1.23 | 0.73, 2.07 | 12 | 0.80 | 0.41, 1.59 | 0.753 |
| Current Smoker of >20 Cigarettes/Day | ||||||||||||||||
| (n = 5,376) | (n = 8,655) | (n = 1,704) | (n = 3,856) | (n = 4,352) | ||||||||||||
| Total lung cancer | 467 | 1.00 | Referent | 702 | 0.95 | 0.84, 1.07 | 144 | 1.00 | 0.83, 1.20 | 319 | 0.95 | 0.83, 1.10 | 400 | 1.02 | 0.89, 1.17 | 0.447 |
| Adenocarcinoma | 129 | 1.00 | Referent | 221 | 1.07 | 0.86, 1.33 | 41 | 1.00 | 0.70, 1.42 | 106 | 1.11 | 0.86, 1.45 | 100 | 0.93 | 0.71, 1.22 | 0.667 |
| SCC | 97 | 1.00 | Referent | 176 | 1.16 | 0.91, 1.49 | 30 | 1.01 | 0.67, 1.52 | 60 | 0.87 | 0.63, 1.21 | 98 | 1.13 | 0.85, 1.51 | 0.949 |
| Small cell carcinoma | 109 | 1.00 | Referent | 141 | 0.82 | 0.64, 1.05 | 32 | 0.98 | 0.66, 1.46 | 72 | 0.96 | 0.71, 1.30 | 89 | 1.06 | 0.79, 1.42 | 0.317 |
| Undifferentiated carcinoma | 21 | 1.00 | Referent | 44 | 1.33 | 0.79, 2.25 | 10 | 1.43 | 0.67, 3.05 | 18 | 1.11 | 0.59, 2.11 | 18 | 0.88 | 0.46, 1.67 | 0.139 |
Abbreviations: CI, confidence interval; HR, hazard ratio; SCC, squamous cell carcinoma.
a Adjusted for sex, age at baseline, cigarette smoking (categorical variable), pipe and cigar smoking (yes, no); education (less than high school; completion of high school; post–high school training other than college; some college; college and postgraduate), physical activity at work (sitting during the day/little walking; sitting much of the day/walking a fair amount; standing/walking a lot—no lifting; lifting/carrying light loads/climbing stairs/climbing hills; doing heavy work/carrying loads), leisure-time physical activity (never, rarely, 1–3 times/month, 1–2 times/week, 3–4 times/week, ≥5 times/week); energy intake (excluding alcohol; kcal/day); body mass index (weight (kg)/height (m)2; ≤18.49, 18.5–24.99, 25.00–29.99, 30.00–34.99, ≥35.00); race/ethnicity (non-Hispanic white; non-Hispanic black; Hispanic; Asian, Pacific Islander, or American Indian/Alaska Native), and Healthy Eating Index-2010 score, which was modified to exclude alcohol (39). We adjusted for cigarette smoking using a detailed categorical variable which included cigarette smoking status, average number of cigarettes smoked per day, and, for former cigarette smokers, years since cessation.
b We tested for trend using an ordinal variable containing the median value of each alcohol category (0 drinks/day, 0.10 drinks/day, 0.69 drinks/day, 1.47 drinks/day, 3.66 drinks/day, 6.10 drinks/day, and 11.50 drinks/day, respectively).
When results were stratified by beverage type, with each estimate adjusted for all other beverage types, we found no pronounced differences between wine and liquor, but there was a suggestion of increased risk in the highest category of beer drinking (Table 4). There was also no clear evidence of effect modification by sex (P = 0.074), body mass index (P = 0.945), or education (P = 0.104).
Table 4.
Adjusted Hazard Ratiosa for the Relationship Between Alcohol Consumption and Lung Carcinoma by Type of Beverage, After Adjustment for Other Beverage Types, Among 492,902 Participants in the National Institutes of Health-AARP Diet and Health Study, 1995/1996–2006
| Alcohol Consumption, drinks/day |
|||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| None (0) |
0.01–0.49 |
0.50–0.99 |
1.00–2.99 |
≥3.00 |
|||||||||||
| No. of Cases | HR | 95% CI | No. of Cases | HR | 95% CI | No. of Cases | HR | 95% CI | No. of Cases | HR | 95% CI | No. of Cases | HR | 95% CI | |
| Beer | |||||||||||||||
| (n = 119,025) | (n = 216,451) | (n = 17,462) | (n = 21,127) | (n = 13,399) | |||||||||||
| Total lung cancer | 2,442 | 1.00 | Referent | 4,241 | 0.98 | 0.93, 1.03 | 343 | 0.93 | 0.83, 1.04 | 574 | 1.07 | 0.98, 1.17 | 534 | 1.20 | 1.09, 1.31 |
| Adenocarcinoma | 891 | 1.00 | Referent | 1,707 | 0.97 | 0.90, 1.04 | 128 | 0.87 | 0.72, 1.05 | 219 | 1.08 | 0.93, 1.25 | 189 | 1.21 | 1.03, 1.41 |
| SCC | 525 | 1.00 | Referent | 793 | 0.92 | 0.83, 1.03 | 69 | 0.91 | 0.70, 1.17 | 125 | 1.06 | 0.88, 1.29 | 128 | 1.23 | 1.02, 1.50 |
| Small cell carcinoma | 372 | 1.00 | Referent | 608 | 1.00 | 0.88, 1.13 | 50 | 0.97 | 0.72, 1.31 | 88 | 1.12 | 0.89, 1.41 | 85 | 1.23 | 0.97, 1.56 |
| Undifferentiated carcinoma | 147 | 1.00 | Referent | 237 | 1.10 | 0.89, 1.36 | 22 | 1.18 | 0.75, 1.86 | 25 | 0.90 | 0.59, 1.38 | 29 | 1.20 | 0.80, 1.79 |
| Wine | |||||||||||||||
| (n = 119,025) | (n = 230,920) | (n = 35,039) | (n = 31,579) | (n = 1,528) | |||||||||||
| Total lung cancer | 2,442 | 1.00 | Referent | 4,250 | 0.93 | 0.89, 0.98 | 527 | 0.87 | 0.79, 0.96 | 591 | 0.92 | 0.84, 1.01 | 42 | 1.05 | 0.77, 1.42 |
| Adenocarcinoma | 891 | 1.00 | Referent | 1,775 | 1.02 | 0.94, 1.10 | 246 | 1.02 | 0.88, 1.17 | 268 | 1.06 | 0.93, 1.22 | 21 | 1.42 | 0.92, 2.18 |
| SCC | 525 | 1.00 | Referent | 755 | 0.83 | 0.74, 0.92 | 85 | 0.73 | 0.58, 0.92 | 97 | 0.76 | 0.62, 0.95 | 2 | 0.24 | 0.06, 0.96 |
| Small cell carcinoma | 372 | 1.00 | Referent | 601 | 0.86 | 0.76, 0.97 | 67 | 0.80 | 0.62, 1.04 | 71 | 0.78 | 0.61, 1.00 | 6 | 0.99 | 0.44, 2.21 |
| Undifferentiated carcinoma | 147 | 1.00 | Referent | 219 | 0.85 | 0.70, 1.05 | 26 | 0.75 | 0.49, 1.14 | 31 | 0.83 | 0.56, 1.23 | 6 | 2.57 | 1.14, 5.81 |
| Liquor | |||||||||||||||
| (n = 119,025) | (n = 200,992) | (n = 14,528) | (n = 29,770) | (n = 17,265) | |||||||||||
| Total lung cancer | 2,442 | 1.00 | Referent | 3,720 | 0.93 | 0.88, 0.97 | 295 | 0.87 | 0.77, 0.98 | 820 | 0.97 | 0.90, 1.05 | 639 | 1.01 | 0.93, 1.10 |
| Adenocarcinoma | 891 | 1.00 | Referent | 1,545 | 0.96 | 0.89, 1.04 | 113 | 0.84 | 0.69, 1.02 | 326 | 1.01 | 0.89, 1.14 | 228 | 1.01 | 0.87, 1.16 |
| SCC | 525 | 1.00 | Referent | 681 | 0.91 | 0.81, 1.01 | 60 | 0.93 | 0.71, 1.22 | 158 | 0.94 | 0.79, 1.13 | 133 | 1.00 | 0.82, 1.21 |
| Small cell carcinoma | 372 | 1.00 | Referent | 549 | 0.94 | 0.83, 1.07 | 39 | 0.83 | 0.60, 1.15 | 119 | 0.97 | 0.79, 1.19 | 105 | 1.08 | 0.87, 1.34 |
| Undifferentiated carcinoma | 147 | 1.00 | Referent | 187 | 0.77 | 0.62, 0.95 | 16 | 0.76 | 0.45, 1.27 | 45 | 0.85 | 0.61, 1.18 | 28 | 0.69 | 0.46, 1.03 |
Abbreviations: CI, confidence interval; HR, hazard ratio; SCC, squamous cell carcinoma.
a Adjusted for sex, age at baseline, cigarette smoking (categorical variable), pipe and cigar smoking (yes, no); education (less than high school; completion of high school; post–high school training other than college; some college; college and postgraduate), physical activity at work (sitting during the day/little walking; sitting much of the day/walking a fair amount; standing/walking a lot—no lifting; lifting/carrying light loads/climbing stairs/climbing hills; doing heavy work/carrying loads), leisure-time physical activity (never, rarely, 1–3 times/month, 1–2 times/week, 3–4 times/week, ≥5 times/week); energy intake (excluding alcohol; kcal/day); body mass index (weight (kg)/height (m)2; ≤18.49, 18.5–24.99, 25.00–29.99, 30.00–34.99, ≥35.00); race/ethnicity (non-Hispanic white; non-Hispanic black; Hispanic; Asian, Pacific Islander, or American Indian/Alaska Native), and Healthy Eating Index-2010 score, which was modified to exclude alcohol (39). We adjusted for cigarette smoking using a detailed categorical variable which included cigarette smoking status, average number of cigarettes smoked per day, and, for former cigarette smokers, years since cessation.
DISCUSSION
We found evidence that alcohol consumption is associated with a modest reduction in the risk of total lung carcinoma at moderate levels of consumption (>0–<1 drink/day) and associated with a modest increase in risk in the highest category of consumption (≥7 drinks/day). We observed some differences by histological subtype. There was a modest nonlinear inverse association at all levels of alcohol consumption for SCC. In contrast, we observed an association showing an elevated risk of lung adenocarcinoma that was limited to the highest consumption category (≥7 drinks/day). The findings for total lung carcinoma are a combination of apparently distinct associations by histological subtype.
Our finding that modest alcohol consumption may be inversely associated with risk of total lung carcinoma and SCC contrasts with some previous studies which have found either increased risk (15, 20) or no association (13, 14, 22, 25). However, we note that the inverse associations in our study persisted in analyses that were stratified by smoking status (Table 3) and beverage type (Table 4). Findings for increased risk with beer drinking are consistent with prior beverage-specific analyses, which have often found inverse associations for wine but higher risks for beer and liquor (13, 15, 17, 18, 27). We also found an increased risk for adenocarcinoma, which suggests potential differences by histological subtype. Previous studies that also found differences by histological subtype have differed in their results (17), with risk reported to be higher for SCC (16), adenocarcinoma (13, 25), or both (19, 21). Most of these prior studies have been limited by modest sample sizes, which probably contributed to the observed heterogeneity.
Our observed inverse association for total lung carcinoma and SCC remained evident at very low levels of consumption, such as 0.1 drink/day. Such a finding seems implausible and suggests that this result may not be due to a biological effect of the alcoholic beverage but rather due to confounding. However, residual confounding by smoking is unlikely to explain the observed association, because we adjusted our analyses for a highly detailed smoking variable and smokers are more likely to drink alcohol than the converse. Residual confounding by socioeconomic status seems more likely, since alcohol use is positively correlated with socioeconomic status (40), and we could only adjust for socioeconomic status by way of education. It is possible that this single measure did not fully capture socioeconomic status. Because our data were observational, other noncausal explanations are certainly possible.
Our study had several notable strengths. First, we used a prospective design with a large sample size (n = 492,902), giving rise to a large number of lung carcinoma cases (n =10,227) and allowing good power for the evaluation of associations by histological subtype. Similarly, we had sufficient data to evaluate specific types of alcoholic beverages in addition to total alcohol. Second, our range of alcohol consumption was wide, allowing for detailed exposure categorization and the assessment of nonlinear relationships with cubic splines. The baseline questionnaire provided us with detailed information on smoking status, smoking cessation, average number of cigarettes smoked per day, and pipe/cigar use, allowing us to comprehensively adjust for the effects of tobacco smoking. We also used the information on other potential confounders such as age, sex, education, physical activity at work, leisure-time physical activity, body mass index, race/ethnicity, and diet.
Our study also had several limitations. Despite a wide range of alcohol intakes, only a small proportion of the cohort had exposures in the heaviest drinking categories, which limited our statistical power to examine very heavy drinking. We did not assess lifetime drinking history or binge drinking. We lacked information on secondhand smoke exposure as well as detailed information on use of noncigarette tobacco products, such as pipes and cigars, that are associated with lung carcinoma. We also lacked information on duration of smoking for the majority of participants, and we cannot exclude the possibility of residual confounding by smoking. However, in a subcohort of 190,587 participants who provided information on smoking duration in a follow-up questionnaire between October 7, 2004, and December 30, 2004, the median age at initiation of smoking was 17 years in both men (interquartile range, 13–22) and women (interquartile range, 17–22). Because of this limited variability in age at initiation, smoking duration was highly correlated with age among current smokers.
In summary, our analysis of the NIH-AARP Diet and Health Study cohort provides evidence that alcohol consumption is associated with a modest reduction in risk of lung carcinoma at moderate levels of consumption and that any level of alcohol drinking could be inversely associated with SCC, but this finding may have been due to uncontrolled confounding by socioeconomic status or some other unrecognized confounder. There was also evidence that higher levels of alcohol consumption are associated with increased risk of total lung carcinoma and adenocarcinoma. However, it is reassuring that we found no increased risk at levels observed to be protective against cardiovascular disease and total mortality (41, 42), suggesting that adherence to the current public health recommendation—to consume alcohol in moderation for those who choose to drink (43)—is unlikely to adversely impact risk of lung carcinoma. The evidence for increased risks of total lung carcinoma and adenocarcinoma was strongest in beer drinkers, but further study is needed to replicate these findings. Our results suggest potential differences by histological subtype, with the largest inverse relationship being present when examining SCC. Given this heterogeneity, future analyses of the relationship between alcohol drinking and lung carcinoma need to examine this association by histological subtype.
ACKNOWLEDGMENTS
Author affiliations: Department of Chronic Disease Epidemiology, Yale School of Public Health, New Haven, Connecticut (Jose Ramon Troche, Susan T. Mayne, Fatma M. Shebl); Division of Cancer Epidemiology and Genetics, National Cancer Institute, Rockville, Maryland (Jose Ramon Troche, Neal D. Freedman, Christian C. Abnet); and Yale Cancer Center, New Haven, Connecticut (Susan T. Mayne, Fatma M. Shebl).
This work was supported by the Intramural Research Program of the National Cancer Institute and the National Institutes of Health (grant T32 CA105666).
The opinions and conclusions expressed in this article are solely the views of the authors and do not necessarily reflect those of any US government agency.
Conflict of interest: none declared.
REFERENCES
- 1.Ferlay J, Soerjomataram I, Ervik M et al. All cancers (excluding non-melanoma skin cancer): estimated incidence, mortality and prevalence worldwide in 2012. In: GLOBOCAN 2012: Estimated Cancer Incidence, Mortality and Prevalence Worldwide in 2012. Version 1.0. (IARC CancerBase no. 11) Lyon, France: International Agency for Research on Cancer; 2013. http://globocan.iarc.fr/Pages/fact_sheets_cancer.aspx Updated January 6, 2015. Accessed June 1, 2015. [Google Scholar]
- 2.American Cancer Society. Cancer Facts & Figures 2013. Atlanta, GA: American Cancer Society; 2013. [Google Scholar]
- 3.National Cancer Institute. Lung cancer prevention (PDQ®). 2014. http://www.cancer.gov/cancertopics/pdq/prevention/lung/Patient/page1/AllPages Updated May 12, 2015. Accessed June 1, 2015.
- 4.Thun MJ, Hannan LM, Adams-Campbell LL et al. Lung cancer occurrence in never-smokers: an analysis of 13 cohorts and 22 cancer registry studies. PLoS Med. 2008;59:e185. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.International Agency for Research on Cancer. Consumption of Alcoholic Beverages. (IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, vol. 100 E) Lyon, France: International Agency for Research on Cancer; 2012:472. [Google Scholar]
- 6.World Cancer Research Fund/American Institute for Cancer Research. Food, Nutrition, Physical Activity, and the Prevention of Cancer: A Global Perspective. Washington, DC: World Cancer Research Fund/American Institute for Cancer Research; 2007. [Google Scholar]
- 7.Bagnardi V, Bla M, La Vecchia C et al. A meta-analysis of alcohol drinking and cancer risk. Br J Cancer. 2001;8511:1700–1705. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Corrao G, Bagnardi V, Zambon A et al. A meta-analysis of alcohol consumption and the risk of 15 diseases. Prev Med. 2004;385:613–619. [DOI] [PubMed] [Google Scholar]
- 9.Bagnardi V, Rota M, Botteri E et al. Alcohol consumption and site-specific cancer risk: a comprehensive dose-response meta-analysis. Br J Cancer. 2015;1123:580–593. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Bellocco R, Pasquali E, Rota M et al. Alcohol drinking and risk of renal cell carcinoma: results of a meta-analysis. Ann Oncol. 2012;239:2235–2244. [DOI] [PubMed] [Google Scholar]
- 11.Cheng G, Xie L. Alcohol intake and risk of renal cell carcinoma: a meta-analysis of published case-control studies. Arch Med Sci. 2011;74:648–657. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Song DY, Song S, Song Y et al. Alcohol intake and renal cell cancer risk: a meta-analysis. Br J Cancer. 2012;10611:1881–1890. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.De Stefani E, Correa P, Deneo-Pellegrini H et al. Alcohol intake and risk of adenocarcinoma of the lung. A case-control study in Uruguay. Lung Cancer. 2002;381:9–14. [DOI] [PubMed] [Google Scholar]
- 14.Li Y, Yang H, Cao J. Association between alcohol consumption and cancers in the Chinese population—a systematic review and meta-analysis. PLoS One. 2011;64:e18776. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Prescott E, Grønbaek M, Becker U et al. Alcohol intake and the risk of lung cancer: influence of type of alcoholic beverage. Am J Epidemiol. 1999;1495:463–470. [DOI] [PubMed] [Google Scholar]
- 16.Bandera EV, Freudenheim JL, Marshall JR et al. Diet and alcohol consumption and lung cancer risk in the New York State Cohort (United States). Cancer Causes Control. 1997;86:828–840. [DOI] [PubMed] [Google Scholar]
- 17.Bandera EV, Freudenheim JL, Vena JE. Alcohol consumption and lung cancer: a review of the epidemiologic evidence. Cancer Epidemiol Biomarkers Prev. 2001;108:813–821. [PubMed] [Google Scholar]
- 18.Benedetti A, Parent ME, Siemiatycki J. Consumption of alcoholic beverages and risk of lung cancer: results from two case-control studies in Montreal, Canada. Cancer Causes Control. 2006;174:469–480. [DOI] [PubMed] [Google Scholar]
- 19.Dosemeci M, Gokmen I, Unsal M et al. Tobacco, alcohol use, and risks of laryngeal and lung cancer by subsite and histologic type in Turkey. Cancer Causes Control. 1997;85:729–737. [DOI] [PubMed] [Google Scholar]
- 20.Freudenheim JL, Ritz J, Smith-Warner SA et al. Alcohol consumption and risk of lung cancer: a pooled analysis of cohort studies. Am J Clin Nutr. 2005;823:657–667. [DOI] [PubMed] [Google Scholar]
- 21.Koo LC. Dietary habits and lung cancer risk among Chinese females in Hong Kong who never smoked. Nutr Cancer. 1988;113:155–172. [DOI] [PubMed] [Google Scholar]
- 22.Rohrmann S, Linseisen J, Boshuizen HC et al. Ethanol intake and risk of lung cancer in the European Prospective Investigation into Cancer and Nutrition (EPIC). Am J Epidemiol. 2006;16411:1103–1114. [DOI] [PubMed] [Google Scholar]
- 23.Shimazu T, Inoue M, Sasazuki S et al. Alcohol and risk of lung cancer among Japanese men: data from a large-scale population-based cohort study, the JPHC study. Cancer Causes Control. 2008;1910:1095–1102. [DOI] [PubMed] [Google Scholar]
- 24.Tse LA, Yu IT, Wang XR et al. Synergistic effect between alcohol consumption and familial susceptibility on lung cancer risk among Chinese men. PLoS One. 2012;77:e40647. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Woodson K, Albanes D, Tangrea JA et al. Association between alcohol and lung cancer in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study in Finland. Cancer Causes Control. 1999;103:219–226. [DOI] [PubMed] [Google Scholar]
- 26.Bagnardi V, Rota M, Botteri E et al. Alcohol consumption and lung cancer risk in never smokers: a meta-analysis. Ann Oncol. 2011;2212:2631–2639. [DOI] [PubMed] [Google Scholar]
- 27.Chao C. Associations between beer, wine, and liquor consumption and lung cancer risk: a meta-analysis. Cancer Epidemiol Biomarkers Prev. 2007;1611:2436–2447. [DOI] [PubMed] [Google Scholar]
- 28.Devesa SS, Shaw GL, Blot WJ. Changing patterns of lung cancer incidence by histological type. Cancer Epidemiol Biomarkers Prev. 1991;11:29–34. [PubMed] [Google Scholar]
- 29.Wynder EL, Muscat JE. The changing epidemiology of smoking and lung cancer histology. Environ Health Perspect. 1995;103(suppl 8):143–148. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Houston KA, Henley SJ, Li J et al. Patterns in lung cancer incidence rates and trends by histologic type in the United States, 2004–2009. Lung Cancer. 2014;861:22–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.American Cancer Society. What is non-small cell lung cancer? 2014. http://www.cancer.org/cancer/lungcancer-non-smallcell/detailedguide/non-small-cell-lung-cancer-what-is-non-small-cell-lung-cancer Updated March 4, 2015. Accessed June 1, 2015.
- 32.Centers for Disease Control and Prevention. How Tobacco Smoke Causes Disease: The Biology and Behavioral Basis for Smoking-Attributable Disease. A Report of the Surgeon General. Atlanta, GA: Centers for Disease Control and Prevention; 2010. [PubMed] [Google Scholar]
- 33.Schatzkin A, Subar AF, Thompson FE et al. Design and serendipity in establishing a large cohort with wide dietary intake distributions: the National Institutes of Health-American Association of Retired Persons Diet and Health Study. Am J Epidemiol. 2001;15412:1119–1125. [DOI] [PubMed] [Google Scholar]
- 34.Michaud DS, Midthune D, Hermansen S et al. Comparison of cancer registry case ascertainment with SEER estimates and self-reporting in a subset of the NIH-AARP Diet and Health Study. J Registry Manag. 2005;322:70–75. [Google Scholar]
- 35.Smith L, Brinton LA, Spitz MR et al. Body mass index and risk of lung cancer among never, former, and current smokers. J Natl Cancer Inst. 2012;10410:778–789. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Bowman SA, Friday JE, Moshfegh AJ. MyPyramid Equivalents Database, 2.0 for USDA Survey Foods, 2003–2004: Documentation and User Guide. Beltsville, MD: Food Surveys Research Group, Beltsville Human Nutrition Research Center, Agricultural Research Service, US Department of Agriculture; 2008. http://www.ars.usda.gov/SP2UserFiles/Place/80400530/pdf/mped/mped2_doc.pdf Updated September 2008. Accessed July 3, 2014. [Google Scholar]
- 37.Centers for Disease Control and Prevention. Alcohol and public health. Frequently asked questions. http://www.cdc.gov/alcohol/faqs.htm Updated November 7, 2014. Accessed June 1, 2015.
- 38.Li R, Hertzmark E, Louie M et al. The SAS LGTPHCURV9 Macro. Boston, MA: Harvard School of Public Health; 2011. https://cdn1.sph.harvard.edu/wp-content/uploads/sites/271/2012/09/lgtphcurv9_7-3-2011.pdf Updated October 4, 2010. Accessed November 26, 2014. [Google Scholar]
- 39.Guenther PM, Casavale KO, Reedy J et al. Update of the Healthy Eating Index: HEI-2010. J Acad Nutr Diet. 2013;1134:569–580. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Substance Abuse and Mental Health Services Administration. Results from the 2009 National Survey on Drug Use and Health: Volume I. Summary of National Findings. Rockville, MD: 2010. http://archive.samhsa.gov/data/NSDUH/2k9NSDUH/2k9Results.htm Updated September 2010. Accessed June 1, 2015. [Google Scholar]
- 41.Di Castelnuovo A, Costanzo S, Bagnardi V et al. Alcohol dosing and total mortality in men and women: an updated meta-analysis of 34 prospective studies. Arch Intern Med. 2006;16622:2437–2445. [DOI] [PubMed] [Google Scholar]
- 42.Ronksley PE, Brien SE, Turner BJ et al. Association of alcohol consumption with selected cardiovascular disease outcomes: a systematic review and meta-analysis. BMJ. 2011;342:d671. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.US Department of Agriculture; US Department of Health and Human Services. Dietary Guidelines for Americans, 2010. Washington, DC: US Government Printing Office; 2010. http://www.cnpp.usda.gov/sites/default/files/dietary_guidelines_for_americans/PolicyDoc.pdf Accessed June 1, 2015. [Google Scholar]