Dietary zinc, copper and selenium, and risk of lung cancer

Abstract

Zinc, copper and selenium are important cofactors for several enzymes that play a role in maintaining DNA integrity. However, limited epidemiologic research on these dietary trace metals and lung cancer risk is available. In an ongoing study of 1,676 incident lung cancer cases and 1,676 matched healthy controls, we studied the associations between dietary zinc, copper and selenium and lung cancer risk. Using multiple logistic regression analysis, the odds ratios (OR) and 95% confidence intervals (CI) of lung cancer for all subjects by increasing quartiles of dietary zinc intake were 1.0, 0.80 (0.65–0.99), 0.64 (0.51–0.81), 0.57 (0.42–0.75), respectively (p trend = 0.0004); similar results were found for men. For dietary copper, the ORs and 95% CI for all subjects were 1.0, 0.59 (0.49–0.73), 0.51 (0.41–0.64), 0.34 (0.26–0.45), respectively (p trend < 0.0001); similar reductions in risk and trend were observed by gender. Dietary selenium intake was not associated with risk, except for a significant inverse trend (p = 0.04) in men. Protective trends (p < 0.05) against lung cancer with increased dietary zinc intake were also found for all ages, BMI > 25, current smokers, pack-years ≤30, light drinkers and participants without emphysema. Increased dietary copper intake was associated with protective trends (p < 0.05) across all ages, BMI, smoking and vitamin/mineral supplement categories, pack-years ≤30 and 30.1–51.75 and participants without emphysema. Our results suggest that dietary zinc and copper intakes are associated with reduced risk of lung cancer. Given the known limitations of case–control studies, these findings must be interpreted with caution and warrant further investigation. © 2006 Wiley-Liss, Inc.

Lung cancer is the most common cause of cancer deaths in the United States.1 Smoking is the overwhelming exposure associated with this disease, but dietary factors may also modulate lung cancer risk. Although considerable emphasis has been placed on evaluating the role of certain micronutrients, macronutrients and phytochemicals, little emphasis has been placed on examining the role of dietary trace metals in lung cancer risk.

Zinc,2 copper3 and selenium,4 essential dietary trace metals, are components of critical enzyme systems that play important roles in maintaining DNA integrity by preventing oxidative DNA damage. Zinc and copper are key cofactors in several enzymes, including copper/zinc superoxide dismutases and several DNA repair proteins.2 Selenium is also a key component of enzyme systems including glutathione peroxidase (GPx) that removes hydrogen peroxides generated in vivo by free radicals.4

Zinc is found in a wide variety of foods, but meat and shellfish are the best sources while plant foods tend to be low in zinc.5 Plant foods are the major dietary sources of selenium, but selenium is also found in some meats and seafood. The selenium content in food varies with its content in the soil, where the plants are grown or the animals raised.6 The major sources of dietary copper are plant foods, shellfish and organ meats.7

The Nutritional Prevention of Cancer (NPC) trial, which was designed to test the effect of 200 μg/day selenium for an average of 4.5 years on risk of nonmelanoma skin cancer, found in a secondary analysis that the selenium supplement group exhibited significant reductions in total cancer mortality (50%), total cancer incidence (47%) and lung cancer incidence (46%).8 A recent meta-analysis of the epidemiological evidence from 16 studies (of serum, toenail and dietary selenium) found a significantly reduced risk of lung cancer with higher selenium exposure (RR = 0.74; 95% CI; 0.57–0.97).9 Fewer studies are available on the association between dietary zinc and copper intakes and lung cancer risk. Of the 3 published reports on dietary zinc intake and lung cancer risk, 2 reported an inverse association10, 11 and the other a positive association with risk.12 There are no published studies of dietary copper intake and lung cancer risk. Published studies of either serum or hair copper levels and lung cancer include small numbers of cases and yield inconsistent results.13, 14, 15

Because of the similar functions of these dietary trace metals and their potential importance in host defense against the initiation and progression of cancer, we hypothesized that increased dietary intake of these essential nutrients would be protective against lung cancer, and investigated the associations between dietary and supplemental intakes of zinc, selenium and copper, and risk for lung cancer in an ongoing case–control study.

Subject and methods

Study population

Our study population is consisted of 1,676 patients with lung cancer (cases) and 1,676 healthy controls selected from the control database to be frequency matched by age (±5 years), in an ongoing and previously described case–control study of lung cancer.16 Cases with histologically confirmed lung cancer were recruited prior to initiation of radiotherapy or chemotherapy from The University of Texas M. D. Anderson Cancer Center, Houston. There were no age, gender, ethnic or stage restrictions. Healthy controls without a previous diagnosis of cancer (except nonmelanoma skin cancer) were recruited from the Kelsey-Seybold clinics, Houston's largest private multispecialty physician group of 23 clinics and more than 250 physicians. All participants were US residents. To date, the overall response rate among both the case patients and the control subjects has been ∼75%. This research was approved by both MD Anderson Cancer Center and Kelsey-Seybold Institutional Review Boards.

Epidemiologic and diet data

All participants completed a personal interview to obtain information on demographic factors and smoking history. Individuals who had smoked at least 100 cigarettes in their lifetimes were defined as ever smokers; of those, former smokers were defined as ever smokers who had quit smoking at least 1 year before diagnosis (cases) or before interview (controls). Pack-years were calculated from the product of the number of cigarettes smoked per day and the number of years smoked divided by 20. Race/ethnicity information (white, Hispanic, African American or other) was self-reported. Light drinkers were defined as subjects who reported drinking alcohol in amounts ranging from 0.1–15 g/d. Moderate–heavy drinkers were defined as subjects who reported drinking alcohol in the amounts >15 g/d. History of emphysema was based on the subject reporting a physician's diagnosis of emphysema. Family history of cancer was defined as a first-degree relative ever being diagnosed with cancer.

Dietary data were collected from a modified version of the 135-item National Cancer Institute's health habits and history questionnaire (HHHQ) and food frequency questionnaire. The HHHQ includes a semiquantitative food frequency list, an open-ended food section and other dietary-behavior questions pertaining to dining at restaurants and food preparation methods. The questionnaire has been shown to be a valid and reliable tool across various populations.17, 18 Study participants were asked about their diet during the year prior to diagnosis (cases) and the year prior to enrollment into the study (controls). Nutrient intake was calculated using the DIETSYS + Plus version 5.9 dietary analysis program (Block Dietary Data Systems, Berkeley, California). The DIETSYS + Plus database has been expanded to include values for dietary copper. The primary source for the copper values was the SR16-1, a database maintained by the US Department of Agriculture, Agricultural Research Service. For multiingredient food items not available in SR16-1, nutrient values were derived from appropriate recipes from the Continuing Survey of Food Intakes by Individuals, 1994–96, 1998. Recipe adjustments were made, where required, for moisture changes and nutrient loss due to cooking.

Statistical analysis

Pearson's x² test was used to test the differences between the case patients and control subjects in terms of gender, ethnicity and smoking status. Student's t test was used to test differences in mean age, cigarette pack-years, BMI (weight kilogram per height [m²]), total energy intake and intakes of dietary zinc, copper and selenium between cases and controls.

Quartiles of zinc, copper and selenium were created based on distributions of dietary intake in control subjects. Multiple logistic regression analysis was performed to calculate odds ratios (ORs) and 95% confidence intervals (CIs) for associations between dietary zinc, copper and selenium and the risk of lung cancer. The first quartile (lowest intake) was considered as the reference category. Variables were included in the logistic regression models based on a priori knowledge of risk factors for lung cancer; and hence as potential confounders of the association between the dietary trace metals and lung cancer. We tested for trends in association by dietary intake using the Wald test for the categorical diet variables. We also conducted subgroup analyses defined by age (≤60 and >60 years), BMI (kg/m²) (≤25 and >25), smoking status (current, former and never smokers), tertiles of pack-years smoked and alcohol intake categories (nondrinkers, light drinkers and moderate–heavy drinkers), vitamin/mineral supplement use and whether the subjects had emphysema. We also computed Pearson and Spearman rank correlation coefficients in the control population, revealing significant correlations between daily intake of dietary zinc, copper and selenium (p < 0.0001) (data not shown). Indeed, these dietary trace metals may affect the absorption of each other, therefore, in addition to estimate ORs from the models for zinc, copper and selenium, we constructed a second model for zinc that included copper and selenium as covariates; a second model for copper that included zinc and selenium as covariates; and a second model for selenium that included zinc and copper as covariates. The top 10 food sources for zinc, copper and selenium were calculated by the DIETSYS + Plus dietary analysis program. Statistical analyses were performed with SAS (version 8.0; SAS Institute, Cary, NC). All statistical tests were two-sided, and a p value < 0.05 for any test or model was considered statistically significant.

Results

Population characteristics

Data from 1,676 patients with lung cancer and 1,676 age-matched controls were available for this analysis (Table I). Overall, cases and controls did not differ by ethnicity, but men were overrepresented in cases (53.8%) versus controls (49.5%). Cases compared to controls had fewer never smokers and former smokers, but more current smokers. Cases also reported higher pack-years of smoking than controls (42.1 vs. 32.5, p < 0.0001). Cases also generally had lower BMI than controls (26.2 vs. 28.2, p < 0.0001). Total caloric intake was similar in cases and controls. Overall and by gender, cases compared to controls had significantly lower dietary zinc (men: 11.08 vs. 11.67 mg/d; women: 8.73 vs. 9.09 mg/d) and copper (men: 1.32 vs. 1.43 mg/d; women: 1.10 vs. 1.18 mg/d) intakes. Differences in selenium intake were not statistically significant.

Table I. Distribution of Select Demographic and Dietary Variables by Case–Control Status

Variable	Cases (N = 1676)	Controls (N = 1676)	p value1
Values are mean (SD) unless otherwise indicated. 1 χ^2 test for categorical variables and t test for continuous variables.
Age, years
Overall	61.13 (10.49)	60.96 (10.35)	0.64
Men	61.48 (10.03)	61.97 (9.80)	0.31
Women	60.72 (10.99)	59.97 (10.77)	0.17
Gender, N (%)
Men	902 (53.82)	829 (49.46)
Women	774 (46.18)	847 (50.54)	0.02
Ethnicity, N (%)
Caucasian	1314 (78.59)	1279 (76.45)
Hispanic	108 (6.46)	129 (7.71)
African American	250 (14.95)	265 (15.84)	0.25
Smoking status, N (%)
Never	256 (15.27)	313 (18.68)
Former	693 (41.35)	779 (46.48)
Current	747 (43.38)	584 (34.84)	<0.0001
BMI (kg/m^2)
Overall	26.2 (5.2)	28.2 (5.5)	<0.0001
Men	26.4 (4.6)	28.2 (4.7)	<0.0001
Women	26.0 (5.8)	28.1 (6.1)	<0.0001
Pack-year smoked
Overall	42.09 (34.27)	32.52 (30.71)	<0.0001
Former smoker	45.00 (31.12)	39.21 (30.38)	0.0003
Current smoker	54.13 (31.74	41.03 (27.90)	<0.0001
Total calories (kcal/d)
Overall	2033.3 (684.5)	2019.7 (650.8)	0.56
Men	2274.0 (691.2)	2246.9 (660.7)	0.41
Women	1752.8 (558.6)	1797.3 (557.9)	0.11
Zinc (mg/d)
Overall	10.00 (3.99)	10.36 (4.17)	0.009
Men	11.08 (4.17)	11.67 (4.39)	0.004
Women	8.73 (3.36)	9.09 (3.49)	0.04
Copper (mg/d)
Overall	1.22 (0.51)	1.30 (0.48)	<0.0001
Men	1.32 (0.51)	1.43 (0.49)	<0.0001
Women	1.10 (0.48)	1.18 (0.42)	0.0004
Selenium (μg/d)
Overall	90.40 (32.01)	91.74 (33.01)	0.23
Men	99.80 (32.41)	102.34 (34.24)	0.11
Women	79.44 (27.80)	81.35 (28.13)	0.17

Main effects analysis

When the dietary intakes of zinc, copper and selenium were analyzed separately (Table II), increasing zinc and copper intakes were associated with lower risk of lung cancer, whereas selenium showed nonsignificant risk reductions only in the highest quartile of intake. Overall, increased dietary intake of zinc was associated with a monotonically decreasing risk of lung cancer, with a 20, 36 and 43% significant reduction in risk with increasing quartile of intake. These associations were similar for both men and women, but significant only in men. When dietary zinc intake was reciprocally adjusted for dietary copper and selenium intakes, both the risk reduction and the trend remained significant for all subjects (data not shown).

Table II. Adjusted Odds Ratios for Associations of Dietary Intakes of Zinc, Copper and Selenium, and Lung Cancer Risk

	Quartiles of intake
	1 (<7.59)	2 (7.59–9.71)	3 (9.72–12.31)	4 (>12.31)	p trend
The values inside parentheses are 95% CI. Adjusted for age, gender, race and family history of cancer in first-degree relatives, pack-years smoked, BMI, total calories and fruits + vegetable intake.
Zinc from foods (mg/d)
Overall
Cases	494	436	374	372
Controls	420	419	420	417
OR	1.0	0.80 (0.65–0.99)	0.64 (0.51–0.81)	0.57 (0.42–0.75)	0.0004
Men
Cases	166	220	237	279
Controls	124	164	233	308
OR	1.0	0.96 (0.69–1.34)	0.68 (0.48–0.96)	0.56 (0.37–0.84)	0.006
Women
Cases	328	216	137	93
Controls	296	255	187	109
OR	1.0	0.76 (0.58–0.99)	0.68 (0.48–0.95)	0.73 (0.47–1.13)	0.11
	1 (<0.99)	2 (0.99–1.22)	3 (1.23–1.56)	4 (>1.56)	p trend
Copper from food (mg/d)
Overall
Cases	594	388	376	318
Controls	430	422	408	416
OR	1.0	0.59 (0.49–0.73)	0.51 (0.41–0.64)	0.34 (0.26–0.45)	<0.0001
Men
Cases	243	191	233	235
Controls	140	177	225	287
OR	1.0	0.58 (0.42–0.79)	0.47 (0.33–0.66)	0.32 (0.21–0.48)	<0.0001
Women
Cases	351	197	143	83
Controls	290	245	183	129
OR	1.0	0.64 (0.49–0.84)	0.60 (0.43–0.82)	0.41 (0.27–0.64)	0.0002
	1 (<67.78)	2 (67.79–87.56)	3 (87.57–111.61)	4 (>111.61)	p trend
Selenium from food (μg/d)
Overall
Cases	417	443	446	370
Controls	421	417	419	419
OR	1.0	1.09 (0.89–1.34)	1.08 (0.86–36)	0.86 (0.64–.15)	0.14
Men
Cases	136	212	265	289
Controls	140	158	234	297
OR	1.0	1.41 (1.01–1.97)	1.13 (0.79–1.59)	0.93 (0.61–1.40)	0.04
Women
Cases	281	231	181	81
Controls	281	259	185	122
OR	1.0	0.99 (0.76–1.29)	1.18 (0.85–1.65)	0.87 (0.54–1.38)	0.35

With increased dietary copper intakes, we found 41, 49 and 66% reduced risk, respectively, for all subjects (p trend < 0.0001). Similar risk reductions were observed in men and women. The direction and magnitude of these associations were similar when dietary copper intake was reciprocally adjusted for zinc and selenium (data not shown).

For dietary selenium intake, we did not find any meaningful associations with lung cancer, although the highest quartile of intake was consistently associated with the lowest OR.

We also evaluated the top 10 food sources for zinc, copper and selenium (data not shown) in our population. Foods such as beef, other meats, whole wheat bread and cereals were the major sources of zinc. Dark bread, potatoes, nuts and seeds were major sources of copper. The major sources of selenium were dark bread, eggs, seafood, meat and cereals.

Subgroup analysis

In this section, we present subgroup analyses for dietary zinc and copper intake and lung cancer risk. We do not report subgroup results for dietary selenium and lung cancer risk because these analyses revealed a similar lack of effect as the main effects analysis for selenium.

We conducted subgroup analyses defined by age (≤60 and >60 years), BMI (≤25 and >25), smoking status (current, former and never smokers) and alcohol intake categories (nondrinkers, light drinkers and moderate–heavy drinkers). There were significant (p < 0.05) trends for decreased risk with increased dietary zinc intakes in both age strata (Table III). Among the younger subjects (≤60 years), those in the highest quartile of dietary zinc intake had a 46% reduced risk when compared with a 40% reduced risk in older subjects (>60 years). We observed a significant inverse trend (p = 0.007) among subjects with BMI >25 (Table III), corresponding to a 21, 39 and 42% reduction in risk for the second, third and fourth quartiles of dietary zinc intake, respectively. There was a decreased risk among the lighter subjects (BMI ≤ 25), but these were not statistically significant (data not shown).

Table III. Risk Estimates for Dietary Zinc Intake (mg/d) and Lung Cancer Defined by Selected Variables

	Quartiles of dietary zinc intake				p trend
	1 (<7.59)	2 (7.59–9.71)	3 (9.72–12.31)	4 (>12.31)
The values inside parentheses are 95% CI. OR adjusted for age, gender, race, family history of cancer in first-degree relatives, pack years smoked, BMI, total calories, and fruits + vegetable intake. 1 For drinkers, total calories in the multivariate model exclude energy from alcohol sources. 2 Reported use of any vitamin/mineral supplement. 3 Self-reported physician-diagnosed emphysema. Note: Because of the space limitations, data are not shown for BMI ≤25, former smokers and never smokers, pack-years (30.1–51.75 and >51.75), nondrinkers and moderate–heavy drinkers (>15 g alcohol/d) and emphysema (yes) because these associations were not significant.
Age (years)
≤60
Cases	204	192	163	187
Controls	169	192	186	199
OR	1.0	0.73 (0.53–1.00)	0.61 (0.43–0.88)	0.54 (0.34–0.84)	0.03
>60
Cases	290	244	211	185
Controls	251	227	234	218
OR	1.0	0.88 (0.67–1.16)	0.69 (0.51–0.95)	0.60 (0.41–0.88)	0.04
BMI
>25
Cases	257	240	220	226
Controls	271	292	321	319
OR	1.0	0.79 (0.61–1.03)	0.61 (0.45–0.82)	0.58 (0.39–0.84)	0.007
Smoking status
Current smokers
Cases	212	173	179	163
Controls	128	152	150	154
OR	1.00	0.57 (0.40–0.81)	0.51 (0.34–0.75)	0.36 (0.22–0.57)	0.0004
Pack-years
≤30
Cases	219	177	132	132
Controls	246	222	228	199
OR	1.00	0.59 (0.40–0.88)	0.48 (0.31–0.76)	0.45 (0.26–0.79)	0.009
Alcohol
Light drinkers1 (0.1–15 g/d)
Cases	194	148	117	138
Controls	164	179	186	164
OR	1.00	0.61 (0.43–0.85)	0.42 (0.28–0.62)	0.46 (0.28–0.75)	0.0003
Supplement use2
Nonusers
Cases	164	179	147	138
Controls	133	135	157	143
OR	1.00	0.71 (0.47–1.08)	0.40 (0.25–0.65)	0.31 (0.17–0.57)	0.0004
Users
Cases	330	257	227	234
Controls	287	284	263	274
OR	1.00	0.71 (0.54–0.95)	0.73 (0.53–1.01)	0.66 (0.45–0.97)	0.09
Emphysema3
No
Cases	407	356	301	306
Controls	396	398	391	383
OR	1.00	0.67 (0.52–0.86)	0.57 (0.43–0.76)	0.49 (0.34–0.69)	0.0003

The protective effect of zinc was largely restricted to the current smokers (Table III), for whom the highest quartile of zinc intake was associated with a 64% reduced risk. When stratified by pack-years smoked, those participants with higher pack-years smoked exhibited less protection against lung cancer with increasing dietary intake of zinc. Higher dietary zinc intake was associated with a significant trend (p = 0.009) for decreased risk only in the lowest tertile of pack-years (≤30) (Table III).

For drinking categories, a significant (p = 0.0005) protective dose effect for dietary zinc intake was observed only among light drinkers (Table III). Dietary zinc intake was associated with a significant trend for reduced risk among those who did not use vitamin/mineral supplements, and borderline significant trend (p = 0.09) for reduced risk among users of vitamin/mineral supplements. The highest quartile of zinc intake among nonusers was associated with a significant 69% reduced risk versus a 44% reduction in risk in users (Table III). We also stratified by history of emphysema and found that higher levels of dietary zinc were associated with significant trends (p < 0.05) for reduced risk only among subjects who reported no prior diagnosis of emphysema.

There were significant (p < 0.05) trends for decreased risk with increased dietary copper intakes for both age strata (Table IV). Among younger subjects, those in the highest quartile of dietary copper intake had a 72% reduced risk, whereas among the older subjects, those in the highest quartile of dietary copper had a 58% reduced risk. There were significant trends for reduced risk with increasing intake of dietary copper among subjects in both BMI strata, with a suggestion of a slightly greater protection for dietary copper among heavier subjects. Across all smoking categories, there were significant trends for decreased risk with increasing copper intake (Table IV). Greater protection was afforded to those with the least pack-years smoked and least protection was in the heaviest smokers. Among the nondrinkers and moderate–heavy drinkers, there were significant reduced risks with increasing quartile of dietary copper intake (Table IV). Significant trends were evident for both nonusers and users of vitamin/mineral supplements, and among subjects without emphysema, but less evident in those with emphysema.

Table IV. Risk Estimates for Dietary Copper Intake (mg/d) and Lung Cancer Defined by Selected Variables

	Quartiles of dietary copper intake				p trend
	1 (<0.98)	2 (0.98–1.22)	3 (1.23–1.56)	4 (>1.56)
The values inside parentheses are 95% CI. OR adjusted for age, gender, race, family-history of cancer in first-degree relatives, pack-years smoked, BMI, total calories and fruits + vegetable intake. 1 For drinkers, total calories in the multivariate model exclude energy from alcohol sources. 2 Reported use of any vitamin/mineral supplement. 3 Self-reported physician diagnosed emphysema. Note: Because of space limitations, data are not shown for pack-years (>51.75), light drinkers (0.1–15 g alcohol/d) and emphysema (yes) because these associations were not significant.
Age (years)
≤60
Cases	244	179	179	144
Controls	170	185	187	204
OR	1.00	0.58 (0.43–0.79)	0.49 (0.35–0.69)	0.28 (0.18–0.43)	<0.0001
>60
Cases	350	209	197	174
Controls	260	237	221	212
OR	1.00	0.61 (0.46–0.80)	0.54 (0.39–0.75)	0.42 (0.28–0.62)	<0.0001
BMI
≤25
Cases	286	151	158	138
Controls	154	120	107	92
Model 1 OR	1.00	0.60 (0.43–0.84)	0.64 (0.44–0.93)	0.49 (0.30–0.78)	0.007
>25
Cases	308	237	218	180
Controls	276	302	301	324
OR	1.00	0.61 (0.47–0.79)	0.49 (0.37–0.67)	0.31 (0.21–0.45)	<0.0001
Smoking status
Current smoker
Cases	258	168	163	138
Controls	157	148	143	136
OR	1.00	0.63 (0.45–0.88)	0.54 (0.37–0.78)	0.38 (0.24–0.60)	0.0004
Former smokers
Cases	239	151	160	143
Controls	182	187	197	213
OR	1.00	0.58 (0.42–0.80)	0.55 (0.39–0.79)	0.39 (0.25–0.61)	0.0003
Never smokers
Cases	97	69	53	37
Controls	91	87	68	67
OR	1.00	0.55 (0.33–0.90)	0.46 (0.25–0.83)	0.25 (0.11–0.59)	0.012
Pack-years
≤30
Cases	255	160	139	106
Controls	236	238	209	212
OR	1.00	0.47 (0.32–0.69)	0.43 (0.27–0.68)	0.24 (0.13–0.44)	<0.0001
30.1–51.75
Cases	156	97	102	97
Controls	111	102	101	112
OR	1.00	0.63 (0.42–0.96)	0.57 (0.36–0.91)	0.39 (0.22–0.69)	0.01
Alcohol
Nondrinkers
Cases	292	159	152	132
Controls	179	160	148	138
OR	1.00	0.54 (0.39–0.75)	0.49 (0.34–0.71)	0.36 (0.23–0.58)	<0.0001
Moderate–heavy drinkers1 (>15 g/d)
Cases	97	87	91	69
Controls	69	83	85	121
OR	1.00	0.65 (0.40–1.06)	0.49 (0.28–0.85)	0.20 (0.09–0.41)	0.0001
Supplement use2
Nonusers
Cases	224	144	139	121
Controls	155	137	138	138
OR	1.00	0.58 (0.39–0.86)	0.48 (0.31–0.76)	0.35 (0.20–0.61)	0.002
Users
Cases	370	244	237	197
Controls	275	285	270	278
OR	1.00	0.62 (0.47–0.83)	0.57 (0.41–0.79)	0.40 (0.26–0.61)	0.0002
Emphysema3
No
Cases	487	313	309	261
Controls	405	397	380	386
OR	1.00	0.58 (0.45–0.74)	0.53 (0.39–0.70)	0.36 (0.25–0.51)	<0.0001

Discussion

Our study, as with all case–control studies, raises concern about recall bias and residual confounding. Athough the FFQ is practical for large epidemiology studies, it is well known that use of the FFQ is associated with inherent measurement errors.19, 20 In an effort to improve the accuracy of dietary reporting, interviewers were trained in the FFQ administration and FFQ responses were reviewed and requeried by staff nutritionists; and portion sizes (for meat only) were assessed with visual aids. Another source of error could arise from the source of nutrient values for dietary trace metals in the food composition database. Although the DIETSYS + plus database that we have used constitute a wide cross section of foods including ethnic foods, we do not have data to compare zinc, copper and selenium composition of foods in the Houston metropolitian area where most of our participants reside. Values for dietary trace metals may vary by region, soil content and other factors, enhancing the biologic variability of the nutrient composition in foods, requiring larger samples of foods across regions to accurately estimate values for the database. Despite these limitations, the FFQ has been shown to reliably assess dietary intake of populations and for classifying individuals by quartiles of intake.21

In a large case–control study of lung cancer, we observed that increasing dietary intakes of zinc and copper were associated with significantly reduced lung cancer risk, ranging from 20 to 43% reductions for zinc and 41 to 66% reductions for copper. We did not find a consistent association between dietary selenium intake and lung cancer risk.

Selection bias maybe an important problem in case–control studies. Cases may recall their diet differently from healthy controls and may change their dietary habits after symptoms appear. In an attempt to reduce differential misclassification of dietary intake between cases and controls, cases were recruited at diagnosis and they were asked about their diet during the year prior to diagnosis, and controls were asked about their diet the year prior to enrollment into the study. Further, when we accounted for SES factors (education and income in the models), the risk estimates generally remained stable.

The top 10 food contributors to zinc, copper and selenium suggest that no single food is a major contributor to zinc, copper and selenium content of the diet, but various foods contribute small amounts of these trace metals, and there were significant intercorrelations among the 3 trace metals. In our study, the healthy control population consumed comparable daily amounts of dietary zinc, copper and selenium to values reported by the National Health and Nutrition Examination Survey (NHANES), 1999–2000, for the US population.22 For example, the controls, on average, consumed 10.4 mg/d dietary zinc when compared with 11.4 mg/d in NHANES (1999–2000). Our controls, on average, consumed 1.3 mg/d dietary copper when compared with 1.2 mg/d in NHANES (1999–2000). Comparable data for dietary selenium were 91.7 μg/d in our control population and 103.1 μg/d in NHANES data.

Zinc, copper and selenium are important constituents of fruits and vegetables and it is possible that these trace metals could be proxies for other constituents in the vegetable–fruit group. Several studies have reported with relative consistency that consumption of fruits and vegetables is protective against lung cancer.23, 24, 25, 26, 27, 28, 29, 30 To date, 2 epidemiological studies of dietary zinc intake and lung cancer risk have been reported.10, 12 Our results for dietary zinc intake are in agreement with the findings from a large US case–control study (923 cases and 1,125) by Zhou et al.10 who found that higher intakes of dietary zinc were associated with a significant protective trend (p = 0.03) for all subjects as well as by gender. Specifically, the highest quintile of dietary zinc intake was associated with a 29% (OR = 0.71, 95% CI, 0.50–0.99) reduced risk compared to a 50% reduced risk in our study. The other study,12 a case–control study (227 cases and 227 controls) in China did not report a significant association. In a US-prospective study (700 cases of lung cancer during 16 years of follow-up) of zinc-vitamin C supplement interactions among postmenopausal women, Lee and Jacobs11 reported that high dietary zinc intake together with vitamin C supplements were associated with nonsignificant deceased risks of lung cancer. Four case–control studies13, 14, 31, 32 and one nested case–control study of serum zinc33 and lung cancer risk yielded mixed results.

To our knowledge, there are no published reports of dietary copper intake and lung cancer risk. A few small studies have reported inconsistent results for serum copper levels and lung cancer risk.14, 15, 32, 34 Among the subjects with lung cancer (n = 30), nonmalignant lung disease (n = 30) and healthy controls (n = 30),14 serum copper was higher in benign lung disease followed by bronchial carcinoma. In 2 small case–control studies15, 32 serum copper levels were significantly higher in lung cancer cases than in controls; however, another case–control study13 reported that lung cancer cases had significantly lower hair copper levels than controls.

We found no main effect of dietary selenium intake on lung cancer risk. In secondary analysis of the nutritional prevention of cancer (NPC) trial,8, 35 selenium supplementation resulted in a significant reduction in lung cancer incidence (46%), while a meta-analysis by Zhou et al.9 of 7 studies of serum/plasma selenium, 3 studies of toenail selenium and 3 studies of dietary selenium measured by FFQ, the summary relative risk for subjects with higher versus lower selenium status was 0.74 (95% CI, 0.57–0.97). Of the 3 studies of dietary selenium, participants in the highest category of dietary selenium intake in one Chinese case–control study had 30% increased risk (OR = 1.30; 95% CI, 0.70–2.20; p trend = 0.48),12 while participants in another Chinese case–control had a 24% decreased risk (OR = 0.76; 95% CI, 0.47–1.15; p trend = 0.11).36 In a cohort study in the Netherlands, those in the highest category of selenium intake had a 2% decreased risk (RR = 0.98; 95% CI, 0.41–2.36; p trend < 0.01).37 Thus, results of the dietary selenium intake-lung cancer association are inconsistent.

The null findings for dietary selenium and lung cancer risk were unexpected. Several lines of evidence support selenium as a potent antioxidant, notably, selenium is essential to GPx and other selenoproteins. We do not believe that this null finding is due to an inadequate range of selenium intake, because the selenium intake in our control population is comparable to national estimates of selenium intake in the US (NHANES 1999–2000: mean 103 μg; median 90 μg).22

Dietary trace metals differ from occupational or environmental exposures to metals, which could lead to toxicity and increased cancer risk. Dietary trace metals such as zinc, copper and selenium are essential for biochemical functions required for health maintenance throughout the life cycle. Zinc is a component of the copper/zinc superoxide dismutase (CuZnSOD) and several proteins involved in DNA repair.38, 39 Zinc is also necessary for normal immune functions, and even mild zinc deficiency depresses immunity.40

Our study also demonstrates a protective effect of dietary copper intake on lung cancer risk. Copper may be released as copper ions to catalyze the formation of ROS, however, in vitro studies have demonstrated that copper ions and complexes can induce ROS production, and oxidative damage has been linked to chronic copper overload.41 In animal studies, restricting dietary copper increases cellular susceptibility to oxidative damage.41 Under normal levels of intake, the cellular fate of copper is under strict control; transport is in the reduced state to prevent oxidative catalysis and radical production.3, 42 Although chronic copper overload in animal studies induces oxidative stress, in human trials, 2 months of supplementation with up to 6 mg copper/L in drinking water (which represented an exposure of up to 20 mg copper/d)43 or a single daily dose of 10 mg copper for 2 months44 was not associated with detectable liver dysfunction. In animal studies, under conditions of copper restriction, CuZnSOD and other antioxidant defense systems such as selenium-dependent GPx and catalase can be compromised.3, 41 In rats, copper deficiency also induces proinflammatory effects on neutophils and lung endothelial cells.45 However, it is possible that copper may be similar to β-carotene in that too much may result in a prooxidant state, too little may lead to oxidative stress and increased inflammation, and adequate levels maintain biosystem adaptations to stress and inflammation.

In our study population, the associations between dietary zinc and copper and lung cancer risk were generally similar across age groups, with a suggestion that the effects may be slightly stronger in younger subjects. For dietary zinc intake, the association was stronger in current smokers than in former and never smokers; however, a significant protective trend was only confined to the lowest tertile of pack-years smoked (≤30) (Table IV), similar to another study.10 For dietary copper intake, results were similar across smoking status, but significant trends for reduced lung cancer risk were only observed in lower tertiles of pack-years smoked (Table IV). Since cigarette smoke is an important source of ROS and carcinogen in the lungs, it is plausible that in heavy smokers, the concentrations of ROS or carcinogens may overwhelm any protective effects associated with dietary zinc and copper. Also, the protective inverse associations for dietary zinc and copper were stronger in heavier (BMI > 25) than thinner subjects (BMI < 25) as expected from earlier research46, 47, 48, 49; however, the mechanism for this association remains unclear.

Significant protective effects were observed in the second, third and fourth quartiles of zinc intake (also a strong protective trend; p = 0.0003) only in light drinkers, but no such association appeared for dietary copper intake. Alcohol can act as a prooxidant in lung tissue.50 Thus, both dietary zinc and copper may function by counteracting oxidative stress induced by alcohol consumption.

For both dietary zinc and copper intakes, we observed protective trends in both nonusers and users of vitamin/mineral supplements. However, the amount and frequency of use of vitamins/minerals were unavailable to compute total (dietary plus supplemental) zinc, copper and selenium intakes and to analyze the associations of supplemental dose and risk.

Emphysema, which is strongly influenced by smoking,51 is a chronic inflammatory condition52 and a risk factor for lung cancer.53 The dietary trace metals, zinc and copper, may have antiinflammatory properties.45, 54 We observed significant protective trends against lung cancer risk with increased dietary intake of zinc and copper only for the subjects who did not report a diagnosis of emphysema. Chronic inflammation in emphysema results in a cycle of lung injury and repair that may overwhelm the antiinflammatory effects associated with dietary zinc and copper intakes.

Phytates, the phosphorylated forms of inositol and phytic acid, can inhibit zinc absorption.55 The highest concentrations of phytates are found in whole grains and legumes.55 In both the low and high consumers of whole grains, there were significant inverse trends (p < 0.05) between intake and lung cancer risk (data not shown). Similar results were found for low and high legume consumers (data not shown). Thus, in our population, consumption of foods rich in phytates did not modify the zinc-lung cancer association.

Because of the complexity of the human diet, a high correlation among different nutrients often exists. Zinc, copper and selenium were highly correlated and thus we computed reciprocal adjustments. Reciprocal adjustments of this kind will change the risk estimates depending on the extent of the variation it will introduce in the model. In general, these reciprocal adjustments did not typically modify the risk estimates and for these reasons, we present results only for the model without the reciprocal covariates.

In conclusion, we found that the estimated intake of zinc and copper from food sources is associated with a lower risk of lung cancer. Our findings suggest that these dietary trace metals should be considered when lung cancer risks are studied.