Lung cancer is the leading cause of cancer deaths for both men and women in the United States. Although cigarette smoking is of key importance, factors such as diet also play a role in the development of lung cancer. A diet rich in fruits and vegetables reduces the incidence of lung cancer by approximately 25%, whereas supplementation with vitamins A, C, and E and with beta-carotene offers no protection against the development of lung cancer.1 The role of micronutrients such as iron, zinc, and calcium in the development of lung cancer remains relatively unknown.
Iron, zinc, and calcium are important micronutrients in diet sources as well as in supplements. Iron is critical for many aspects of cellular function, but it may also generate reactive oxygen species and other free radicals, decrease the amounts of copper–zinc and manganese superoxide dismutases, and damage biologic macromolecules.2 Iron catalyzes the generation of hydroxyl radicals from superoxide anions and increases oxidative stress which, in turn, increases free iron concentration. This self-amplifying process may cause damage to lipid membranes, proteins, and DNA; stimulate oxidative damage to pulmonary cell nuclei3; increase oxidative activation of precarcinogens; and support tumor cell growth. Higher diet and supplemental iron intakes have been associated with higher risks of colorectal cancer,4–7 although a case–control study8 suggested that iron is protective for prostate cancer.
In contrast to iron, zinc is 1 of the major antioxidants found in food. Zinc ions may induce the synthesis of metallothionein, protect against free radicals, and retard oxidative processes.9 Zinc is necessary for the activity of superoxide dismutases, which catalyzes the dismutation of superoxide radical to H2O2. In addition to their antioxidant function, zinc finger proteins have been found to be frequently downregulated in lung cancer cells,10 suggesting that these proteins have important tumor-suppressor functions. Higher dietary intake of zinc has been associated with lower risk of breast,11 colon,7 and prostate cancer,12 although supplemental zinc intake at levels far exceeding recommended intakes has been associated with a higher risk of advanced prostate cancer in 1 study.13
Calcium is an important mediator by which oxidative stress may modulate signal transduction pathways. Dose- and time-dependent increases in intracellular reactive oxygen species and calcium levels have been observed concomitantly in human colorectal cancer cells.14 Calcium may also downregulate 1,25(OH)2D, the active form of vitamin D, which in turn may inhibit lung cancer growth (as found in a cell line).15 Very high calcium intake from dietary or supplemental sources has been associated with a higher risk of prostate cancer.16–18 However, dietary calcium intake has also been found to be protective for colorectal cancer and breast cancer.19–22
Based on this epidemiologic and biologic evidence, we hypothesized that dietary iron, zinc, and calcium intake may modify the risk of lung cancer development. Furthermore, the association between micronutrients and lung cancer risks may be modified by smoking status. We tested these hypotheses in a large hospital-based case–control study.
Participants were enrolled between December 1992 and September 2000 as part of an ongoing hospital-based case–control study. Details of this case–control population have been described previously.23–25 In brief, all eligible cases (patients with histologically confirmed lung cancers who were age 18 years or older) at Massachusetts General Hospital were recruited. Controls were recruited among healthy friends and nonblood-related family members (usually spouses) of patients with cancer or friends and family members of patients with a cardiothoracic condition who were undergoing surgery. Over 85% of eligible cases and over 90% of controls participated in this study. The study was approved by the Human Subjects Committees of Massachusetts General Hospital and the Harvard School of Public Health, both in Boston, Massachusetts. Informed consent was collected from each study participant.
At the time of recruitment, a trained interviewer administered a detailed modified American Thoracic Society health questionnaire and semiquantitative food-frequency questionnaire (FFQ) to most cases and controls.26 Approximately 27% of participants preferred to complete the FFQ at home and return it by mail in a prestamped envelope. The 126-food item FFQ, developed by the Nutrition Department at the Harvard School of Public Health, has been validated in a group of white female nurses27 and male health professionals living in Boston.26 A commonly used unit or portion size was specified for each food item, and both cases and controls were asked about their average consumption over the past year before enrollment. Average consumption was based on 9 possible responses ranging from “never, or less than once per month” to “more than 6 servings per day.” Estimated average intakes for each specific food were obtained and nutrient intakes were computed using the Harvard database, which is a modification of the U.S. Department of Agriculture Nutrition Composition Laboratory's food composition database.
Of the original 2706 individuals (1325 cases and 1381 controls) enrolled, 2612 (97%) subjects provided complete information on age, sex, smoking, and education levels. We then excluded 507 individuals (19%) who did not have complete information on dietary intake. The demographic distributions were similar for individuals with and without dietary data. Among the remaining 2105 subjects, we restricted our analyses to the 2048 white participants, including 923 patients with lung cancer and 1125 controls, to reduce confounding by race. In this study population, all individuals fell within the cutoff points for reasonable caloric intake (ranging from 800 to 4200 kcal for men and 600 to 3500 kcal for women); therefore, no individuals were excluded as outliers for the specific nutrients.
Micronutrient quintiles were created based on distributions of dietary intake in our control population and evaluated both as categorical and ordinal variables. Ordinal diet variables were based on median values of each quintile for controls. We created energy-adjusted micronutrient residual quintiles by regressing each micronutrient on total calories and obtaining the residual from this model. The residual value for each observation was then added to the mean micronutrient value for our population. Because total calories in men were greater than women, we ran separate regressions for men and women to obtain more appropriate energy-adjusted micronutrient residuals.
Logistic regression models were performed to calculate the odds ratio (OR) and 95% confidence interval (CI) for associations between dietary micronutrients and the risk of lung cancer, adjusting for age, sex, smoking status, pack-years of smoking, years since smoking cessation, education levels, and total caloric intake. Dietary intake of iron, zinc, and calcium were analyzed separately (multivariable model 1) as well as in the same model (multivariable model 2). We tested for trends in association by dietary intake using the Wald test, based on the ordinal diet variables. Total micronutrient intake that includes supplements was also analyzed in a similar manner using logistic regression models. Furthermore, we assessed partitioned iron intake from different food sources (heme iron and nonheme iron).
The distributions of demographic characteristics for cases and controls are summarized in Table 1. Compared with controls, cases were older, had a higher proportion of men, more likely to be current smokers or heavy smokers, had a shorter time since smoking cessation (if a former smoker) and more pack-years of smoking, and had lower education levels. The distribution of smoking variables in our controls was similar to the general Massachusetts population over age 45 years.24,25
Adenocarcinoma, squamous cell carcinoma, large cell carcinoma, and small cell carcinoma represented 51%, 25%, 8%, and 9% of cases, respectively. The remaining 7% were of mixed histologic subtype or had more than 1 primary tumor. More than half (58%) of cases were early-stage patients (stage I or II).
Dietary Micronutrient Intake
Cases reported greater mean intakes of total energy than controls (among women, 1922 kcal for cases and 1819 kcal for controls; among men, 2347 kcal for cases and 2144 kcal for controls). Cases had lower energy-residual-adjusted dietary zinc intake when compared with controls (Table 1). Dietary iron and zinc intake were highly correlated. After energy residual adjustment, the correlation coefficients were 0.61 between iron and zinc, 0.14 between iron and calcium, and 0.26 between zinc and calcium.
Dietary Micronutrient and Lung Cancer Risk
When the 3 dietary micronutrients were analyzed separately (multivariable model 1), iron and calcium were associated with a higher risk of lung cancer and zinc was associated with a lower risk of lung cancer (Table 2). The adjusted odds ratios for dietary iron, zinc, and calcium were 1.45 (95% CI = 1.03–2.06), 0.71 (0.50–0.99), and 1.64 (1.17–2.29), respectively, when comparing the highest with lowest quintile for each micronutrient. These associations were stronger when all 3 micronutrients were analyzed in the same model (multivariable model 2); the adjusted odds ratios of dietary iron, zinc, and calcium were 1.95 (1.33–2.86), 0.46 (0.31–0.68), and 1.81 (1.27–2.56), respectively. Similar associations were found when animal protein and animal fat intake were adjusted in the model, when tertiles and quartiles were used in the analyses instead of quintiles, and among subjects who reported no supplement intake (data not shown). Similar associations were also found when we limited the cases to incident early-stage patients and when we compared subjects who were recruited at different time periods.
Because iron and zinc were highly correlated, we also investigated the joint association of dietary iron and zinc with lung cancer risk (Table 3). When compared with subjects in the lowest tertile of iron intake and the highest tertile of zinc intake (considered as the lowest joint-risk group), subjects who have the highest dietary intake of iron and lowest dietary intake of zinc have a higher risk of lung cancer (2.01, 0.96–4.2; P test for trend <0.0001). This finding confirms that the effect of iron and zinc we observed from multivariable model 1 (in which each micronutrient was analyzed in separate models) is not entirely driven by 1 micronutrient. Consistent trends in lung cancer risk were also seen using joint distributions for calcium and zinc, and iron and calcium (data not shown).
We also investigated the associations between dietary micronutrients and the risk of lung cancer in subgroups of age, sex, smoking status, histologic cell type, and clinical stage patients (Table 4). The associations between micronutrients and lung cancer were similar among women and men, and across education levels, histologic cell types, and clinical stage of patients (comparing each group of patients with all controls). Slightly stronger micronutrients–lung cancer associations were observed in younger subjects (less than 55 years compared with 55+) and in current smokers compared with former smokers, specifically for dietary iron and calcium intake.
Total Micronutrient Intake and Lung Cancer Risk
In addition to dietary micronutrient intake, we also analyzed the effects of total iron, zinc, and calcium intake including both food and supplements (Table 5). A total of 198 cases (21%) and 236 controls (21%) were users of iron supplements, 193 (21%) cases and 247 (22%) controls took zinc supplements, and 244 cases (26%) and 338 controls (30%) took calcium supplements. Similar to the findings from dietary micronutrients, total iron and calcium were associated with higher risk of lung cancer (for iron, 2.06, 1.31–3.24; for calcium, 1.45, 1.01–2.08), whereas total zinc appeared to be associated with a lower risk of lung cancer (0.67, 0.43–1.06).
Heme and Nonheme Iron Intake and the Risk of Lung Cancer
We further studied the association for iron from different food sources, ie, heme iron (from red meat, poultry, and seafood) and nonheme iron. Heme iron accounted for 7% of dietary iron for cases and 8% of dietary iron for controls. We found that the increased risk of lung cancer conferred by dietary iron was the result of nonheme iron intake. Heme iron, on the other hand, was associated with decreased risk of lung cancer. The adjusted odds ratios for nonheme iron was 1.49 (1.04–2.14, P test for trend = 0.007) and for heme iron was 0.39 (0.28–0.56, P test for trend <0.0001) for highest quintile versus lowest quintile. Consistent results were found when dietary zinc and calcium were adjusted in the model and in different subgroups of age, sex, smoking status, histologic cell type, and clinical stage patients (data not shown).
Iron, zinc, and calcium are all involved in the metabolizing of reactive oxygen species. Use of a modest zinc supplement may induce cellular iron deficiency and further reduce iron status,28 and iron intake may also impair zinc absorption,29 suggesting a close interrelationship between the 2 nutrients. Despite the correlation between dietary iron and zinc, a higher intake of iron is associated with higher risk of lung cancer, whereas a higher intake of zinc is associated with a lower risk. A similar pattern of contrasting associations between iron and zinc has been reported for the risk of colon cancer.7 In addition to meat, fish, and poultry, other important food sources of zinc are beans, nuts, whole grains, fortified breakfast cereals, and dairy products. It is possible that the relatively high intake of dietary zinc was accounted for by higher intake of the nonmeat food mentioned here and may help to explain the inverse associations between dietary zinc intake and the risk of lung cancer.
We observed that high calcium intakes from dietary or supplemental sources are associated with higher risk of lung cancer. We do not know whether the underlying mechanism also involves the potential role of 1,25(OH)2D, as suggested in the development of prostate cancer, because calcium intake can modify 1,25(OH)2D serum levels. Previous in vitro and animal studies have suggested that 1,25(OH)2D may inhibit lung cancer cell proliferation15 and increase intratumoral T-cell immune reactivity and limit metastasis and locoregional tumor recurrence.30
In our population, the association between dietary micronutrients and lung cancer risk was stronger in current smokers than in former smokers (specifically for iron and calcium; Table 4). Tobacco smoke is a major source of exogenous and endogenous reactive oxygen species in the lungs. Therefore, current exposure to reactive oxygen species or other carcinogens from cigarette smoke may enhance the effect of reactive oxygen species from dietary sources, which may account for the stronger associations found among current smokers when compared with former smokers. We also found slightly stronger associations between iron, zinc, and calcium intakes and lung cancer risk in younger individuals than in older individuals (Table 4) both in women and in men (data not shown). One possible explanation is that younger individuals are more likely to be current smokers (37%) than former smokers (32%), whereas for older individuals, the pattern is reversed (26% current vs 56% former).
In looking at iron from different food sources, we found that the increased lung cancer risk was the result of nonheme iron. Unexpectedly, heme iron was associated with decreased risk of lung cancer. Note that although heme iron is more absorbable in the body, more than 90% of iron is nonheme. The exact mechanism for this observation is unclear. One possible explanation is that heme iron also comes from food sources other than red meat such as poultry, fish, and seafood. Intake from these sources have been found to be protective of lung cancer risk in previous studies.31–33
There are some limitations in this study. First, this is a hospital-based case–control study in which a subset of the controls included healthy spouses and friends of lung cancer cases. Such controls have a tendency to be more similar to cases than population controls because they may share similar health behaviors (eg, smoking and diet); thus, potential overmatching may be present. However, this potential overmatching should reduce our ability to detect a true effect of these lifestyle factors with a bias toward the null. We have compared the dietary intake of our population with several North American cohort studies, and found that the total energy intake and specific micronutrient intake are very similar.34 Second, recall bias may exist in our population. We observed a higher mean intake of total calories in cases than controls. This observation may result from cases recalling more food items in general and could lead to an upward bias of odds ratios for the main effect of dietary intake. We did attempt to control for this effect by using energy-residual adjustment for all of the micronutrients, as well as adjustment for total caloric intake, in logistic regression models. Third, the FFQ were administered at the time of recruitment, and it is possible that the patients with cancer may have changed their dietary habits after cancer symptoms appeared, which could lead to an ascertainment bias. If this bias truly exists, we would expect to see different associations between early- and late-stage patients, because late-stage patients are more likely to change their dietary styles. However, we observed similar associations between early- and late-stage patients (Table 4), among subjects who were recruited at different time periods, and for incident early-stage patients only (data not shown). This consistency across subgroups argues somewhat against ascertainment bias. Last, we lacked some information that could have been helpful. For example, although associations were similar for dietary and total (dietary plus supplemental) intakes, we do not have complete information on the length of time these supplements were used or on dietary history. Similarly, although we found a stronger association between iron and lung cancer risk in younger women than in older women (data not shown), we did not collect the information on menopausal status or postmenopausal hormone use.
In conclusion, our study demonstrates consistent associations of dietary iron, zinc, and calcium with the risk of lung cancer. Our study findings suggest that dietary iron and calcium intake may be associated with higher risk of lung cancer, whereas dietary zinc intake is associated with lower risk. These results need to be confirmed in large prospective studies in which the potential for bias is minimized. Possible biologic mechanisms should be studied further.
We thank the following staff members of the Lung Cancer Susceptibility Group: Barbara Bean, Jessica Shinn, Andrea Solomon, Thomas Van Geel, Lucy Ann Principe, Salvatore Mucci, Richard Rivera-Massa, Kofi Asomaning, and Li Su. We also appreciate generous support of Panos Fidias and other physicians and surgeons of the Massachusetts General Hospital Cancer Center.
1. Fabricius P, Lange P. Diet and lung cancer. Monaldi Arch Chest Dis
2. Reddy MB, Clark L. Iron, oxidative stress, and disease risk. Nutr Rev
3. Yano T, Obata Y, Yano Y, et al. Stimulating effect of excess iron feeding on spontaneous lung tumor promotion in mice. Int J Vitam Nutr Res
4. Kato I, Dnistrian AM, Schwartz M, et al. Iron intake, body iron stores and colorectal cancer risk in women: a nested case–control study. Int J Cancer
5. Nelson RL. Iron and colorectal cancer risk: human studies. Nutr Rev
6. Deneo-Pellegrini H, De Stefani E, Boffetta P, et al. Dietary iron and cancer of the rectum: a case–control study in Uruguay. Eur J Cancer Prev
7. Lee DH, Anderson KE, Harnack LJ, et al. Heme iron, zinc, alcohol consumption, and colon cancer: Iowa Women's Health Study. J Natl Cancer Inst
8. Vlajinac HD, Marinkovic JM, Ilic MD, et al. Diet and prostate cancer: a case–control study. Eur J Cancer
9. Powell SR. The antioxidant properties of zinc. J Nutr
10. Ito G, Uchiyama M, Kondo M, et al. Kruppel-like factor 6 is frequently down-regulated and induces apoptosis in non-small cell lung cancer cells. Cancer Res
11. Adzersen KH, Jess P, Freivogel KW, et al. Raw and cooked vegetables, fruits, selected micronutrients, and breast cancer risk: a case–control study in Germany. Nutr Cancer
12. Kristal AR, Stanford JL, Cohen JH, et al. Vitamin and mineral supplement use is associated with reduced risk of prostate cancer. Cancer Epidemiol Biomarkers Prev
13. Leitzmann MF, Stampfer MJ, Wu K, et al. Zinc supplement use and risk of prostate cancer. J Natl Cancer Inst
14. Zhang S, Ong CN, Shen HM. Critical roles of intracellular thiols and calcium in parthenolide-induced apoptosis in human colorectal cancer cells. Cancer Lett
15. Higashimoto Y, Ohata M, Nishio K, et al. 1 alpha, 25-dihydroxyvitamin D3 and all-trans-retinoic acid inhibit the growth of a lung cancer cell line. Anticancer Res
16. Chan JM, Stampfer MJ, Ma J, et al. Dairy products, calcium, and prostate cancer risk in the Physicians’ Health Study. Am J Clin Nutr
17. Rodriguez C, McCullough ML, Mondul AM, et al. Calcium, dairy products, and risk of prostate cancer in a prospective cohort of United States men. Cancer Epidemiol Biomarkers Prev
18. Kristal AR, Cohen JH, Qu P, et al. Associations of energy, fat, calcium, and vitamin D with prostate cancer risk. Cancer Epidemiol Biomarkers Prev
19. Wu K, Willett WC, Fuchs CS, et al. Calcium intake and risk of colon cancer in women and men. J Natl Cancer Inst
20. Satia-Abouta J, Galanko JA, Martin CF, et al. Associations of micronutrients with colon cancer risk in African Americans and whites: results from the North Carolina Colon Cancer Study. Cancer Epidemiol Biomarkers Prev
21. Slattery ML, Neuhausen SL, Hoffman M, et al. Dietary calcium, vitamin D, VDR genotypes and colorectal cancer. Int J Cancer
22. Boyapati SM, Shu XO, Jin F, et al. Dietary calcium intake and breast cancer risk among Chinese women in Shanghai. Nutr Cancer
23. Xu LL, Wain JC, Miller DP, et al. The NAD(P)H:quinone oxidoreductase 1 gene polymorphism and lung cancer: differential susceptibility based on smoking behavior. Cancer Epidemiol Biomarkers Prev
24. Miller DP, Liu G, de Vivo I, et al. Combinations of the variant genotypes of GSTP1, GSTM1, and p53 are associated with an increased lung cancer risk. Cancer Res
25. Zhou W, Liu G, Park S, et al. Gene-smoking interaction associations for the ERCC1 polymorphisms in the risk of lung cancer. Cancer Epidemiol Biomarkers Prev
26. Rimm EB, Giovannucci EL, Stampfer MJ, et al. Reproducibility and validity of an expanded self-administered semiquantitative food frequency questionnaire among male health professionals. Am J Epidemiol
27. Willett WC, Sampson L, Stampfer MJ, et al. Reproducibility and validity of a semiquantitative food frequency questionnaire. Am J Epidemiol
28. Donangelo CM, Woodhouse LR, King SM, et al. Supplemental zinc lowers measures of iron status in young women with low iron reserves. J Nutr
29. Troost FJ, Brummer RJ, Dainty JR, et al. Iron supplements inhibit zinc but not copper absorption in vivo in ileostomy subjects. Am J Clin Nutr
30. Wiers KM, Lathers DM, Wright MA, et al. Vitamin D3 treatment to diminish the levels of immune suppressive CD34+ cells increases the effectiveness of adoptive immunotherapy. J Immunother
31. Marchand JL, Luce D, Goldberg P, et al. Dietary factors and the risk of lung cancer in New Caledonia (South Pacific). Nutr Cancer
32. Fortes C, Forastiere F, Farchi S, et al. The protective effect of the Mediterranean diet on lung cancer. Nutr Cancer
33. Takezaki T, Inoue M, Kataoka H, et al. Diet and lung cancer risk from a 14-year population-based prospective study in Japan: with special reference to fish consumption. Nutr Cancer
34. Lee DH, Folsom AR, Jacobs DR Jr. Dietary iron intake and type 2 diabetes incidence in postmenopausal women: the Iowa Women's Health Study. Diabetologia