- Email: [email protected]
Risk factors associated with lung cancer in Hong Kong
Lung Cancer 40 (2003) 131 /140 www.elsevier.com/locate/lungcan
Risk factors associated with lung cancer in Hong Kong Moira Chan-Yeung a,*, L.C. Koo b,c, J.C.-M. Ho a, K.W.-T. Tsang a, W.-S. Chau d, S.-W. Chiu d, M.S.-M. Ip a, W.-K. Lam a a
Division of Respiratory and Critical Care Medicine, Department of Medicine, The University of Hong Kong, 4/F, Professorial Block, Queen Mary Hospital, Hong Kong SAR, China b Department of Epidemiology and Disease Modeling, AstraZeneca, Wilmington, DE, USA c Clinical Epidemiology Unit, CCEB, University of Pennsylvania, Philadelphia, USA d Cardiothoracic Surgical Unit, The Grantham Hospital, Hong Kong SAR, China Received 1 October 2002; received in revised form 16 December 2002; accepted 19 December 2002
Abstract The purpose of this study was to investigate the risk factors associated with lung cancer in Hong Kong. Three hundred and thirtyone histologically or cytologically proven consecutive cases of lung cancer and the same number of in- and out-patients without cancer matched for age and sex were recruited for this study using a detailed questionnaire completed by a trained interviewer. Smoking was the most important risk factor associated with lung cancer but the attributable risk (AR) was estimated to be 45.8% in men and 6.2% in women, considerably lower compared with those estimated in early 1980s. In addition, among women, exposure to environmental tobacco smoke (ETS) at work9/at home and lack of education, were independent risk factors for lung cancer with adjusted odds ratio (OR) 3.60, (95% confidence interval (CI) 1.52 /8.51) and OR 2.41 (95% CI 1.27 /4.55), respectively. Among men, exposure to insecticide/pesticide/herbicide, ETS exposure at work or at home, and a family history of lung cancer and were independent risk factors with adjusted OR 3.29 (95% CI 1.22 /8.9, OR 2.43, 95% CI 1.24 /4.76 and OR 2.37, 95% CI 1.43 /3.94, respectively). Exposure to incense burning and frying pan fumes were not significant risk factors in both sexes. A moderate or high consumption of fat in the diet was associated with increased risk in men but decreased risk in women. The results of this study suggested that as the prevalence of smoking declined, the influence of smoking as a risk factor for lung cancer decreased even further. Moreover, the contribution of other environmental, occupational and socioeconomic factors may be more apparent as etiological factors for lung cancer in a population with relatively high lung cancer incidence but low AR from active smoking. # 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Lung cancer; Risk factors; Smoking; Environmental tobacco smoke; Insecticide; Family history; Socioeconomic factor; Diet
1. Introduction Lung cancer is the most common cause of cancer mortality in Hong Kong and worldwide. In 1999, the incidence of lung cancer in Hong Kong was 58.3/100 000 in men and 23.2/100 000 in women [1]. While cigarette smoking is the most important cause for lung cancer in both men and women in the Western world, the attributable risk (AR) of cigarette smoking was found to be only 35.8% in women in Hong Kong in 1985 [2]. The etiologic or risk factors associated with lung cancer
* Corresponding author. Tel.: /852-2855-4385; fax: /852-28551143. E-mail address: [email protected] (M. Chan-Yeung).
in nonsmoking women are not entirely clear. Exposure to passive smoke, occupational exposures, air pollution, and dietary factors have been implicated. However, the increased risk for lung cancer due to exposure to environmental tobacco smoke (ETS) was found to be small in Hong Kong; with the odds ratios (OR) in two different studies found to be approximately 1.6 [2,3]. There have been many studies investigating the relation between nutrition and lung cancer [4 /11]. After adjusting for smoking, most of the observational studies suggest that increased vegetable and fruit intake is associated with a lower risk for lung cancer in both men and women. In Hong Kong, Koo et al. from a study conducted in the early 1980s [5] reported that higher consumption of leafy green vegetables, carrots,
0169-5002/03/$ - see front matter # 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0169-5002(03)00036-9
132
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
bean curd, fresh fruits and fresh fish was associated with a lower risk for adenocarcinoma in nonsmoking women. There has been a slow decline in the incidence of lung cancer in Hong Kong during the past two decades [1]. The purpose of this study was to investigate the risk factors associated with lung cancer in Hong Kong and whether risk factors have changed during the past two decades since lifestyle, nutrition, and exposure to environmental carcinogens have changed in Hong Kong.
2. Materials and methods 2.1. Subjects Three hundred and thirty-one consecutive histologically or cytologically proven lung cancer patients and the same number of control patients without lung or other cancers, matched for age (9/5 years) and sex, were recruited from the Queen Mary Hospital, a teaching and regional hospital, from May 1999 to December 2001 for this study. The controls were identified from the same hospital ward or from outpatient clinics for respiratory patients. They were all ethnically Chinese. Among the controls, there were 150 (45.6%) patients with asthma, 119 (36%) with obstructive sleep apnea syndrome, 17 (5%) with chronic obstructive lung disease, 33 (10%) with chest infections, and the remaining 12 patients with endocrine disorders, liver problems and skin rash. 2.2. Methods All patients and controls were interviewed using a questionnaire by one trained interviewer. The questions included the following: place of birth and residence, marital history, history of tuberculosis, family history of cancer and lung cancer, and occupational and environmental exposure to various chemicals such as insecticides, pesticides, herbicides and asbestos. In addition, information related to smoking and amount smoked, and exposure to passive smoke, incense and cooking fumes was collected together with a detailed dietary history. Information on past exposure to passive smoke at home from spouses or other family members who smoked, and at work from co-workers who smoked was only obtained from nonsmokers. Nonsmoking status was checked by interviewing family members living with the patient. The dietary section of the interview inquired about the frequency of consumption of a list of food items or food groups as described previously [5]. Briefly, cases were asked to report on their usual food habits 1 year before their diagnosis of cancer or the appearance of symptoms. Controls were asked to recall their usual habits of eating. Each control was usually interviewed
within 6/8 weeks of the matched case’s interview to reduce the effects of seasonality on food frequency pattern. The average number of portions was calculated for each month for each of the following food item: Green vegetables : all leafy green vegetables, cabbage, cauliflower, green pepper, broccoli, and legumes. Root vegetables : carrot, squash, and yam. Fruit : any fruit. Fatty food : ham, sausage, ribs, pigs feet, pig skin, meatball, fried food, butter, ice-cream, and cheese. Preserved food : Chinese preserved meats (siuyuk ), Chinese sausage, pickled vegetables, salted egg, preserved beancurd, salted fish, dried squid, dried octopus, dried oyster, dried shrimp. 2.3. Definitions Ever-smoker : one who had smoked at least one cigarette a day, pipe, water pipes, cigars, and/or hand rolled cigarettes, for 1 year or more. Current smoker */one who was still smoking (any of the above) 1 month prior to the interview. Ex-smoker : one who had ever smoked at least one cigarette a day and any of the above forms of tobacco, for 1 year or more and had given up smoking 1 month prior to the interview. Nonsmoker : a life-long nonsmoker. ETS exposure at home : life-long nonsmoker exposed to anyone who smoked at home regularly for at least 2 years. ETS exposure at work9/at home : life-long nonsmoker exposed to smokers at work in enclosed space9/at home regularly for at least 2 years. Regular exposure to the following was defined as exposure at least once a week for 2 or more years: Paint exposure: regular exposure to paint or lacquer or varnish or paint-thinner at work. Gas exposure: regular exposure to gasoline or kerosene at work. Insecticide exposure: regular exposure to insecticide, pesticide or herbicide at work. Incense exposure: regular exposure to incense burning in the home. Frying pan fumes exposure: regular exposure to cooking fumes from frying in the home. 2.4. Analysis For men, the amount of smoking was studied by combining ex- and current smokers and categorized into light (B/20 pack-years), medium (20 /39 pack-years) and heavy (/40 pack-years). In women, because of small numbers of ex- and current smokers (ten and three, respectively, among cases and four and two
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
among controls), they were combined in the analysis and the dose relationship between smoking and risk of lung cancer was not determined. For exposure to ETS, the effects of exposure at home and at work (with and without exposure at home) were analyzed separately. In men, the number of nonsmokers with exposure to ETS at home was small and ETS exposure at work or at home was used in the analysis. Regular occupational exposure to paint, gasoline, pesticide, and asbestos was categorized into ‘yes’ and ‘no’. Incense smoke exposure was categorized into no exposure, festival only, daily outdoors, daily indoors9/ outdoors. Exposure to cooking fumes from frying was categorized into no exposure, B/3.5, 3.5 /7 and /7 times a week. For dietary data, cases and controls were grouped into tertile levels of consumption (low, medium and high) whenever possible, and into two levels when the consumption was low. The data were analyzed separately for men and women. The unadjusted OR and 95% confidence intervals (CI) for lung cancer was calculated for each of the following demographic characteristics: birthplace; marital, employment and educational status; previous history of tuberculosis; and family history of cancer. ORs and 95% CI for lung cancer were calculated for each of the following risk factors: smoking, exposure to ETS; occupational exposure to paint, gasoline, asbestos, pesticides, insecticides, and herbicides; environmental exposure to incense burning; cooking fumes from frying pans; and intake of various food items. Adjusted ORs and 95% CI were also calculated for each of the risk factors using conditional logistic regression analyses with adjustment made for significant demographic confounders (place of birth, educational status, and family history of lung cancer in men and educational status in women (see Table 1) and smoking. The adjustment of smoking took into consideration the smoking status (non-, ex-, and current smoking), duration and amount of cigarettes smoked (B/20, 20 /39, / 40 pack-years). Conditional logistic regression analysis was carried out including all exposure risk factors to determine the independence of the risk factors adjusted for demographic confounders. The AR of exposure for lung cancer was estimated as follows: AR /P (OR/1)/{P (OR/1)/1}, where OR is the odds ratio and P is the prevalence of exposure [12]. Statistical analyses were carried out using the SPSS for WINDOWS version 10, statistical package (SPSS, Chicago, IL).
2.5. Ethics The study was granted approval by the Ethics Committee, Faculty of Medicine, of the University of Hong Kong.
133
3. Results Three hundred and thirty-one consecutive patients with histologically proven lung cancer and the same number of age (9/5 years) and sex-matched controls were recruited for this study. Table 1 shows the characteristics of the study subjects by sex. There were no significant differences between patients and controls in marital status, place of birth and employment status. However, having no education or only primary school education was associated with a higher risk for lung cancer in both sexes, but after adjustment for smoking, was only found to be significant among women without formal education (OR 3.78, 95% CI 1.62 /8.84). In men, those with a family history of cancer or lung cancer had a significantly higher risk for lung cancer but this was not found in women. Table 2 show the association between smoking and exposure to ETS and lung cancer, adjusted for confounding demographic variables (place of birth, educational status and a family history of lung cancer in men; educational status in women). In men, ex-smokers and current smokers had higher lung cancer risks which were statistically significant for current smokers (adjusted OR 5.53, 95% CI 2.9 /10.6). Among ever smokers, the risk of lung cancer increased significantly with increased amount of smoking in a dose-dependent manner. Among nonsmoking men, exposure to ETS (at home or at work) showed an increase in risk for lung cancer (adjusted OR 3.38, 95% CI 0.86 /13.3) but not statistically significant. In women, ever smokers (ex- and current) showed a significantly higher risk (adjusted OR 3.78, 95% CI 1.11 /12.9). Among nonsmoking women, exposure to ETS at work9/at home significantly increased the risk for lung cancer (adjusted OR 3.23, 95% CI 1.25 /8.36) but not ETS at home only. The estimated ARs of active smoking for lung cancer were 45.8% in men and 6.2% in women. The AR of ETS exposure at work (9/at home) was estimated to be 17.6% in nonsmoking women. The risks of lung cancer from various occupational and environmental exposures are shown in Table 3, adjusted for smoking and confounding demographic variables. In men, exposure to insecticide/pesticide/ herbicides regularly at work significantly increased the risk for lung cancer (adjusted OR 5.57, 95% CI 1.66 / 18.7). The number of women with occupational exposure to these chemicals at work was too small for analysis to be meaningful. Exposure to incense burning during festivals showed an increased risk in women (adjusted OR 2.95, 95% CI 1.10 /7.87) but not daily exposure. Exposure to frying pan fumes from cooking did not increase the risk of lung cancer in either men or women. The association of consumption of various food items and lung cancer is shown in Table 4 adjusted for
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
134 Table 1 Characteristics of the study subjects by sex Men Patients
N Age (years)
Controls
212 57.59/11.1
Marital status Married 176 (83.0) Others 37 (17.0)
OR (95%CI)
Adjusted OR (95% CI)a
Patients
212 57.49/11.2
Controls
119 56.09/11.2
OR (95% CI)
Adjusted OR (95% CI)a
119 55.99/11.3
172 (81.1) 40 (18.9)
1 1 0.88 (0.54 /1.45) 0.70 (0.39 /1.26)
86 (72.3) 33 (27.7)
82 (68.9) 37 (31.1)
1 0.85 (0.9 /1.49)
105 (49.5) 107 (50.5)
1 1 1.55 (1.06 /2.29) 1.39 (0.86 /2.25)
55 (46.2) 64 (53.8)
55 (46.6) 63 (53.4)
1 1 1.02 (0.61 /1.69) 0.96 (0.57 /1.63)
88 (41.5) 34 (16.0)
78 (36.8) 21 (9.9)
1 1 1.44 (0.77 /2.68) 0.75 (0.33 /1.68)
29 (24.4) 10 (8.4)
32 (26.9) 8 (6.7)
1 1 1.38 (0.48 /3.97) 1.67 (0.51 /5.59)
90 (42.5)
113 (53.3)
0.71 (0.47 /1.07) 0.60 (0.30 /1.21)
80 (67.2)
79 (66.4)
1.12 (0.62 /2.02) 1.22 (0.62 /2.38)
127 (59.9)
1
44 (37.0)
58 (49.1)
1
32 (26.9) 43 (36.1)
38 (32.2) 22 (18.6)
1.11 (0.60 /2.05) 1.39 (0.71 /2.72) 2.58 (1.35 /4.92) 3.78 (1.62 /8.84)
111 (93.3) 8 (6.7)
1 1 1.70 (0.68 /4.27) 1.62 (0.65 /4.03)
Place of birth Hong Kong 82 (38.7) China/ 130 (61.3) others Employment Retired Part time/ unemployed Full time/ house wife
Women
Education level Secondary/ 97 (45.7) post secondary Primary 81 (38.3) None 34 (16.0)
63 (29.7) 22 (10.4)
1
1.68 (1.10 /2.57) 1.25 (0.75 /2.09) 2.02 (1.11 /3.68) 1.68 (0.75 /3.76)
1 0.79 (0.44 /1.44)
1
Previous history of TB No 179 (84.8) Yes 32 (15.2)
193 (91.0) 19 (9.0)
1 1 106 (89.1) 1.82 (0.99 /3.32) 1.97 (0.98 /3.94) 13 (10.9)
Family history of cancer No 141 (66.5) Yes 71 (33.5)
162 (76.4) 50 (23.6)
1 1 1.63 (1.07 /2.50) 1.95 (1.16 /3.31)
75 (63.0) 44 (37.0)
83 (69.7) 37 (30.3)
1 1 1.35 (0.79 /2.32) 1.52 (0.83 /2.86)
Family history of Ca lung No 184 (86.8) Yes 28 (13.2)
196 (92.5) 16 (7.5)
1 1 104 (87.4) 1.86 (0.98 /3.56) 3.03 (1.38 /6.68) 15 (12.6)
109 (91.6) 10 (8.4)
1 1 1.57 (0.68 /3.66) 1.82 (0.74 /4.48)
a
Adjusted for smoking amount and duration using conditional logistic regression analysis.
smoking and confounding demographic variables. Higher consumption of carrots and of green vegetables was each associated with a higher risk for lung cancer in both sexes but significant only for the highest level of consumption of green vegetables in men after adjustment. When consumption of all vegetables including green and root vegetables and fruits was combined, no significant increased or reduced risk for lung cancer was found in either men or women. Higher consumption of preserved food and fatty food was each associated with a significantly higher lung cancer risk in men only, but a dose-dependent relationship was not found. Similar analyses were conducted on nonsmoking women only but the results were not different the data are not shown. To study the effect of multiple exposure factors, conditional logistic regression analyses were carried
out with models that included all demographic and exposure variables that affected lung cancer risk with P values /0.15, separately for men and women, as shown in Table 5. In men, smoking, exposure to insecticide/ pesticide/herbicide, medium or high fat content diet, and a family history of lung cancer remained as significant risk factors after adjustment for the other exposure variables; ETS exposure became significant while high consumption of preserved food was no longer significant. In women, ever smoking, ETS exposure at work9/ at home and lack of education remained as risk factors after adjustment for other exposure variables but exposure to incense smoke during festivals was no longer significant. A diet with high fat content in women was associated with a lower risk for lung cancer. High consumption of vegetables and fruits in diet had no influence on the risk of lung cancer in both sexes.
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
135
Table 2 Smoking and exposure to ETS and risk for lung cancer in men/women, adjusted and unadjusted OR (95% CI) N
Patients
Controls
OR (95% CI)
Adjusted OR (95%CI)
(a ) Men Smoking habit Nonsmoker Ex-smoker Current smoker
52 (24.5) 68 (32.1) 92 (43.4)
96 (45.3) 87 (41.0) 29 (13.7)
1 1.44 (0.91 /2.29) 5.86 (3.42 /10.0)
1b 1.26 (0.70 /2.26)b 5.53 (2.90 /10.6)a,b
Ever smokers Nonsmoker Light ( B/20 pack-years) Medium (20 /39 pack-years) Heavy (/40 pack-years)
52 25 50 85
96 41 32 43
(45.3) (19.3) (15.1) (20.3)
1 1.13 (0.62 /2.05) 2.88 (1.65 /5.04) 3.65 (2.22 /6.01)
1b 0.91 (0.44 /1.89)b 3.40 (1.77 /6.50)b 4.76 (2.41 /9.40)a,b
Exposure to ETS in nonsmokers No ETS or B/2 years At work or at home ]/2 years
25 (48.1) 27 (51.9)
67 (69.8) 29 (30.2)
1 2.50 (1.24 /5.01)
1b 3.38 (0.86 /13.28)b
106 (89.1) 13 (10.9)
113 (95.0) 6 (5.0)
1 2.31 (0.85 /6.30)
1e 3.78 (1.11 /12.92)e
53 (50.0) 31 (29.2) 22 (20.8)
69 (61.1) 35 (31.0) 9 (8.0)
1 1.15 (0.63 /2.10) 3.18 (1.35 /7.48)
1e 1.01 (0.47 /2.18)e 3.23 (1.25 /8.36)d,e
(24.5) (11.8) (23.6) (40.1)
(b ) Women Smoking habit Nonsmoker Ex- and current smokersc Exposure to ETS in nonsmokers No ETS or B/2 years ETS at home ]/2 years At work9/at home ]/2 years
a A significant (P B/0.05) increasing trend in the OR was found between nonsmokers, ex- and current smokers; and increasing amount of smoking among the ever smokers. b Adjusted for place of birth, educational status, and a family history of lung cancer by conditional logistic regression analysis. c Ex- and current smokers combined because of small numbers. d A significant (P B/0.05) increasing trend in OR was found between no ETS exposure, exposure to ETS at home only, and ETS exposure at work9/at home. e Adjusted for educational status by conditional logistic regression analysis.
4. Discussion In this study, we found that smoking was the most important risk factor for lung cancer in both sexes, as expected. However, among men, additional significant risk factors included ETS exposure at work or at home, occupational exposure to pesticide/insecticide/herbicide, diets with medium to high fat content, and a family history of cancer or lung cancer. In women, exposure to ETS at work with or without exposure at home, and having no education were other significant risk factors found in this study. Smoking is a well-recognized risk factor in both men and women for lung cancer. In 1985, the AR of smoking for lung cancer was found to be 95.4% in men and 35.8% in women [2]. In this study, the AR of smoking has declined to 45.8 and 6.2%, respectively. These figures, however, are similar to the lower AR reported in China in the 1989/1990 (52.3 and 19.4%, respectively) [13]. In Hong Kong, the prevalence of smoking has declined from 1982 to 2000 to be 39.7 and 5.6% to 25.2 and 4.4%, respectively, in men and women over the age of 15 years [14]. The incidence of lung cancer in Hong Kong has also declined from 84.6/100 000 in 1983 to 58.6/100 000
in 1999 in men, and from 31.5 to 23.2/100 000, respectively, in women [1]. As there is a latent period between the onset of smoking and the development of lung cancer, a firm conclusion cannot be made that the decline in lung cancer incidence is due to the reduction in the prevalence of smoking in Hong Kong as smoking data from the general population was not available before 1983. It is possible that some patients in this study failed to report their smoking habit to avoid blame as it is more widely known now compared with early 1980s that smoking can cause lung cancer. This is especially true for women as it is associated with more social stigma for a Chinese woman than a man to be a smoker. We attempted to verify the nonsmoking status by interviewing with family members living with the patient. Because of limited resources, we did not examine urinary or hair cotinine level to validate the patients’ current smoking status. In this study, our controls were patients from the same hospital who mostly suffered from chronic lung disease such as asthma and chronic obstructive lung disease and did not have lung cancer or other types of cancer. The prevalence of current smokers among our male control subjects in this study is higher than those of
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
136
Table 3 Environmental exposures and risk for lung cancer, unadjusted and adjusted OR (95% CI) Men Patients
Women Controls
OR (95%CI)
Adjusted OR (95%CI)a
Patients
172 (81.1)
173 (81.6)
1
1
113 (95.0)
116 (97.5)
40 (16.9)
39 (18.4)
1.03 (0.63 /1.68)
0.93 (0.53 /1.62)
6 (5.0)
3 (2.5)
161 (75.9)
166 (78.3)
1
1
111 (93.3)
115 (96.6)
51 (24.1)
46 (21.7)
1.14 (0.73 /1.80)
1.02 (0.59 /1.78)
8 (6.7)
4 (3.4)
192 (90.6)
206 (97.2)
1
1
116 (97.5)
118 (99.2)
20 (9.4)
6 (2.8)
3.58 (1.41 /9.09)
5.57 (1.66 /18.7)
3 (2.5)
1 (0.8)
206 (97.2)
209 (99.1)
1
1
6 (2.8)
2 (0.9)
3.04 (0.61 /15.25)
4.91 (0.73 /32.9)
Controls
OR (95% CI)
Adjusted OR (95% CI)a
1
1
2.05 (0.50 /8.41)
1.53 (0.34 /6.82)
1
1
2.07 (0.61 /7.07)
1.36 (0.38 /4.91)
1
1
3.05 (0.31 /29.8)
7.70 (0.52 /114.8)
Occupational exposure Paints No or B/2 years Yes Gasoline No or B/2 years Yes Pesticide No or B/2 years Yes Asbestos No or B/2 years Yes
117 (98.3)
118 (100)
1
1
2 (1.7)
/
/
/
Environmental exposure Incense No or B/2 years Festival only Daily Frying food No or B/2 years B/3.5 3.5 /7 /7
87 (41.0)
86 (43.2)
1
1
30 (25.2)
42 (37.2)
1
1
25 (11.8) 100 (47.2)
20 (10.1) 93 (46.7)
1.24 (0.64 /2.39) 1.06 (0.71 /1.60)
1.03 (0.47 /2.26) 0.94 (0.56 /1.56)
18 (15.1) 71 (59.7)
11 (9.7) 60 (53.1)
2.29 (0.95 /5.55) 1.66 (0.93 /2.96)
2.95 (1.10 /7.87) 1.58 (0.77 /3.26)
146 (70.2)
131 (67.2)
1
1
34 (28.6)
34 (30.1)
1
1
27 (13.0) 22 (10.6) 13 (6.3)
33 (16.9) 24 (12.3) 7 (3.6)
0.73 (0.42 /1.29) 0.82 (0.44 /1.54) 1.67 (0.65 /4.30)
0.69 (0.32 /1.49) 0.83 (0.38 /1.80) 1.22 (0.38 /3.99)
37 (31.1) 27 (22.7) 21 (17.6)
36 (31.9) 26 (23.0) 17 (15.0)
1.03 (0.53 /1.99) 1.04 (0.51 /2.13) 1.24 (0.56 /2.74)
1.08 (0.50 /2.32) 1.05 (0.46 /2.42) 1.54 (0.57 /4.13)
a Adjusted for place of birth, educational status, a family history of lung cancer and smoking in men; educational status and smoking status in women using conditional logistic regression analysis.
the general population, 43.4 and 25.2%, respectively; the corresponding figures for women were 1.7 and 4.4%, respectively. Thus the use of male patients with a higher prevalence of smoking may have underestimated the effect of active smoking in men and overestimated its effects in women. The female controls were mostly nonsmoking asthmatics. Exposure to ETS as a risk factor for lung cancer in nonsmokers has been studied extensively during the past few decades and many comprehensive reviews of the health consequences of ETS have been published [15 / 20]. It is generally agreed that similar carcinogens are present in mainstream tobacco smoke as in ETS, and that nonsmokers are exposed to carcinogens in the ETS. The meta-analysis carried out by the US Environmental Protection Agency in 1992 concluded that ETS exposure is associated with an increased risk for lung cancer in nonsmoking adults accounting for approximately 3000 deaths annually in the US [17]. The majority of the
studies on ETS exposure were from spousal smoke rather than exposure at work. In this study, we found ETS exposure at work with or without exposure at home increased the risk of lung cancer in nonsmoking men and women. It should be noted that the number of nonsmoking women with ETS exposure in this study is small. Compared with two decades ago, ETS exposure at work has become more important as a risk factor in Hong Kong where there is no regulation to prohibit smoking in the workplace. Although there might have been recall bias and the actual number of subjects involved was small for patient and control groups, our data showed that regular (at least once weekly for 2 or more years) workplace exposure to insecticides, pesticides and herbicides was associated with a significantly higher lung cancer risk for men. Exposure to DDT has been found to increase the risk of lung cancer among orchard and construction workers [21,22]. A recent study in Gaza also showed
Table 4 Consumption of food (number of portions per week) and risk of lung cancer, unadjusted and adjusted OR and 95% CI Portion/month
Men
Tomato Carrots
Green vegetables
Fruits
Green vegetables and tubers and fruits
Preserved food
Fatty food
a
B/2 ]/2 B/2 2 /4 /4 B/45 45 /65 65/ B/15 15 /30 30/ B/70 70 /100 100/ B/1 1 /2 /2 B/1 1 /2 2/
96 114 55 80 77 72 43 97 92 54 66 68 88 56 70 89 53 51 72 89
Controls
OR (95% CI)
(45.7) 94 (44.3) 1 (54.5) 117 (55.7) 0.95 (0.65 /1.4) (25.9) 76 (35.8) 1 (37.7) 89 (42.0) 1.24 (0.78 /1.97) (36.4) 47 (22.2) 2.26 (1.37 /3.74) (34.0) 94 (44.3) 1 (20.2) 52 (24.5) 1.08 (0.65 /1.79) (45.8) 66 (31.2) 1.92 (1.24 /2.97) (43.4) 77 (36.3) 1 (25.5) 56 (26.4) 0.81 (0.50 /1.31) (31.1) 79 (37.3) 0.70 (0.45 /1.09) (32.1) 79 (37.3) 1 (41.5) 82 (38.7) 1.25 (0.80 /1.94) (26.4) 51 (24.1) 1.28 (0.77 /2.10) (33.0) 81 (38.6) 1 (42.0) 60 (28.6) 1.72 (1.09 /2.71) (25.0) 69 (32.8) 0.89 (0.55 /1.44) (24.1) 80 (38.1) 1 (34.0) 53 (25.2) 2.13 (1.29 /3.51) (41.9) 77 (36.7) 1.81 (1.14 /2.89)
Adjusted OR (95% CI)a
Patients
Controls
OR (95%CI)
Adjusted OR (95%CI)a
1 0.97 1 1.46 2.22 1 1.15 2.23 1 0.89 0.65 1 1.43 1.42 1 1.83 0.86
60 59 36 42 41 20 36 63 36 22 61 21 43 55 61 35 23 50 47 22
46 73 39 49 30 34 27 58 35 37 47 32 36 51 63 34 19 43 39 34
1 0.62 1 0.93 1.48 1 2.27 1.85 1 0.58 1.26 1 1.82 1.64 1 1.06 1.25 1 1.04 0.56
1 0.74 1 1.12 2.05 1 2.11 1.74 1 0.58 1.25 1 1.60 1.86 1 1.46 1.19 1 0.94 0.45
(0.62 /1.52) (0.82 /2.60) (1.22 /4.06) (0.60 /2.23) (1.28 /3.89) (0.49 /1.59) (0.36 /1.16) (0.84 /2.43) (0.78 /2.58) (1.06 /3.16) (0.47 /1.55)
2.37 (1.26 /4.46) 1.61 (0.89 /2.94)
(50.4) (49.6) (30.3) (35.3) (34.3) (16.8) (30.3) (52.9) (30.3) (18.5) (51.3) (17.6) (36.1) (46.2) (51.3) (29.4) (19.3) (42.0) (39.5) (18.5)
(38.7) (61.3) (33.1) (41.5) (25.4) (28.6) (22.7) (48.7) (29.4) (31.1) (39.5) (26.9) (30.3) (42.9) (54. 3) (29.3) (16.4) (37.1) (33.6) (29.3)
(0.37 /1.04) (0.50 /1.71) (0.77 /2.85) (1.08 /4.77) (0.96 /3.56) (0.29 /1.17) (0.69 /2.30) (0.90 /3.69) (0.84 /3.21) (0.59 /1.92) (0.62 /2.52) (0.58 /1.87) (0.28 /1.09)
(0.43 /1.28) (0.56 /2.23) (0.95 /4.42) (0.91 /4.86) (0.86 /3.52) (0.27 /1.22) (0.65 /2.44) (0.74 /3.47) (0.89 /3.90) (0.71 /3.01) (0.56 /2.52)
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
Patients
Women
(0.45 /1.96) (0.20 /1.02)
Adjusted for place of birth, educational status, a family history of lung cancer and smoking in men; educational status and smoking in women using conditional logistic regression analysis.
137
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
138
Table 5 Conditional logistic regression analysis: model included all exposure variables adjusted for demographic confounders Risk factors
Men OR (95% CI)
Women OR (95% CI)
History of tuberculosis Yes vs. no
1.89 (0.95 /3.76)
Family history of cancer Yes vs. no
2.37 (1.43 /3.94)
Education None vs. secondary and above
1.78 (0.86 /3.76) 2.41 (1.27 /4.55)
Smoking Medium vs. nonsmoking Heavy vs. nonsmoking Any smoking vs. nonsmoking ETS in nonsmokers Any exposure vs. none
5.13 (2.73 /9.65) 6.39 (3.56 /11.5) 5.08 (1.32 /19.5) 2.43 (1.24 /4.76)
ETS Exposure at work9/home vs. none
3.60 (1.52 /8.51)
Diet Medium fat vs. low fat High fat vs. low fat
2.71 (1.51 /4.87) 2.53 (1.46 /4.39) 0.43 (0.22 /0.85)
Exposure to insecticide/pesticide Yes vs. no
3.29 (1.22 /8.90)
Variables with P -value /0.15 were removed from the final model.
highly significant correlations between cancer risk and the use of insecticides, fungicides, herbicides, nematicides and all types of pesticides [23]. Another study from Costa Rica showed that heavy pesticide usage was associated with an overall increase of cancer incidence including that of lung cancer [24]. Potent insecticides, such as pyrethroids, which account for more than 30% of global use, are known to be carcinogenic [25]. Further study is needed to explore the types of insecticide/ pesticide/herbicide used locally at present and in the past and their association with lung cancer. Exposure to incense smoke and exposure to frying pan fumes were found to be associated with a higher risk for lung cancer in nonsmoking Chinese women due to the presence of combustion-related carcinogens [26 /30]. We have not been able to find such a relationship with either exposure in this study. Even though about 70% of women claimed that they were exposed to incense burning, many were exposed during festivals and outdoors only. Indoor incense burning less commonly practice in Hong Kong with progressive westernization. The use of extraction fans and the practice of opening kitchen windows during cooking reduced exposure to cooking fumes in Hong Kong. The association between lung cancer and frying pan fumes has been found in Shanghai with the use of rapeseed oil [29], which is not commonly used in Hong Kong. Increased lung cancer
risk was also found to be associated with exposure to fumes from kerosene stoves, which are no longer used in Hong Kong [3]. Thus some of the sources of exposure that carry an increased risk for lung cancer nonsmoking women locally are no longer common. A reported family history of cancer or lung cancer increased the risk of lung cancer significantly in men and a similar trend in women. Most of the published literature have shown a smoking-adjusted increased risk of lung cancer by about 2 /4-fold associated with a family history of lung cancer [31 /34]. As only a small proportion of smokers develop lung cancer, genetic susceptibility is likely to be important, although recall bias could also contribute to this finding. The issue of genetic predisposition of lung cancer is being actively investigated by many researchers at present [35 /37]. We found that the lack of any education, a surrogate for socioeconomic status, is an independent risk factor for lung cancer in women, increasing the risk by twofold after adjusting for smoking and exposure to ETS. This association has also been found in a previous study on lung cancer risk in Hong Kong [38]. Socioeconomic status has been reported as a risk factor for lung cancer [39,40]. A careful study of the exposure of uneducated and poor women with lung cancer may give us some clues as to the etiology of lung cancer in women. There have been many epidemiologic studies investigating the relationship between lung cancer and nutrition. In general, the majority of the case /control studies demonstrated that a higher consumption of vegetables and fruits is associated with a lower lung cancer risk in both women and men after adjusting for smoking [5 / 11]. In Hong Kong, Koo and colleagues [5] studied 88 nonsmoking women with lung cancer and 137 matched controls and found that consumption of leafy green vegetables, carrots, bean curd, fresh fruits and fresh fish was associated with a protective effect on adenocarcinoma. By using the same methodology of data collection, we found a contrary effect, that f higher consumption of green vegetables, fruits and carrots were associated with higher risk. On the other hand, we found high fat consumption was associated with an increase in lung cancer risk in men but a lower risk in women, after adjustment for smoking and other demographic and exposure variables. A positive association between high fat consumption and lung cancer risk has been reported [41,42]. The discrepancy in findings between the previous study and the present one despite using the same method of data collection could be due to several reasons. In the previous study, the controls were recruited from a healthy population while in the present study the controls were recruited from patients with mostly lung disease but without any cancer in the hospital or outpatient clinics. Dietary intake of leafy green vegetables and fruits in patients with chronic lung disease and obstructive sleep apnea may be quite similar
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
to those with lung cancer and different from healthy controls. In this study we did not stratify patients according to histological types because of sample size limitations. It is also possible that there has been a change in the dietary pattern in Hong Kong with a higher consumption of meat and fatty food compared with 20 years ago. The differential effect of high fat consumption between men and women on lung cancer risk is hard to explain but the number of women among cases and controls who had a high fat consumption was small. In summary, our study has shown that while cigarette smoking is still a major cause of lung cancer in men, its influence is decreasing. Our results also suggest that other environmental, occupational and socioeconomic issues may be important in the development of lung cancer. As lung cancer is the commonest lethal malignant disease worldwide, it is necessary to continue to search for other relevant environmental exposures, combined with genetic predispositions. One obvious one is long term exposure to combustion-related fine particles from air pollution, which has been shown to be associated with increase in lung cancer mortality recently [43].
Acknowledgements The authors wish to thank the Anti-Cancer Society of Hong Kong for its generous support for this study and Anne DyBuncio for data analysis.
References [1] Hong Kong Cancer Registry. Hospital Authority, Hong Kong, SAR 2000. [2] Koo LC, Ho JH-C, Lee N. An analysis of some risk factors for lung cancer in Hong Kong. Int J Cancer 1985;35:149 /55. [3] Lam TH, Kung ITM, Wong CM, et al. Smoking, passive smoking and histological types in lung cancer in Hong Kong Chinese women. Br J Cancer 1987;56:673 /8. [4] Koo LC, Ho JH-C. Diet as a confounder of the association between air pollution and female lung cancer: Hong Kong studies on exposures to environmental tobacco smoke, incense, and cooking fumes as examples. Lung Cancer 1996;14(Suppl. 1):S47 /61. [5] Koo LC. Dietary habits and lung cancer risk among Chinese females in Hong Kong who never smoked. Nutr Cancer 1988;11:155 /72. [6] Wu-Williams AH, Dai XD, Blot W, et al. Lung cancer among women in Northeast China. Br J Cancer 1990;62:982 /7. [7] Darby S, Whitley E, Doll R, Key T, Silcocks P. Diet, smoking and lung cancer: a case /control study of 1000 cases and 1500 controls in south-west England. Br J Cancer 2001;84:728 /35. [8] Feskanich D, Ziegler RG, Michaud DS, et al. Prospective study of fruit and vegetable consumption and risk of lung cancer among men and women. J Natl Cancer Inst 2000;92:1812 /23.
139
[9] Jansen MCJF, Bueno-de-Mesquita HB, Rasanen L, et al. Cohort analysis of fruit and vegetable consumption and lung cancer mortality in European men. Int J Cancer 2001;92:913 /8. [10] Feskanich D, Ziegler RG, Michaud DS, et al. Prospective study of fruit and vegetable consumption and risk of lung cancer among men and women. J Natl Cancer Inst 2000;92:1812 /23. [11] Voorrips LE, Goldbohm RA, Verhoeven DTH, et al. Vegetables and fruit consumption and lung cancer risk in the Netherland cohort study on diet and cancer. Cancer Causes Control 2000;11:101 /15. [12] Breslow NE, Day N. Statistical methods in cancer research. In: Davis W, editor. The analysis of case /control studies, vol. 1. Lyons: International Agency for Research on Cancer, 1980:70. [13] Liu BQ, Peto R, Chen ZM, et al. Emerging tobacco hazards in China: 1. Retrospective proportional mortality study of 1 million deaths. Br Med J 1998;317:1411 /22. [14] Census and Statistics Department, Hong Kong Government. Hong Kong: General Household Surveys, 2000. [15] Koo LC, Ho JH. Worldwide epidemiological patterns of lung cancer in nonsmokers. Int J Epidemiol 1990;19:S14 /23. [16] US Department of Health Education and Welfare. The health consequences of smoking. A report of the Surgeon General. Washington, DC: US Government Printing Office, 1971. DHEW publication No HSM 73-8704. [17] US Environmental Protection Agency. Respiratory health effects of passive smoking: Lung cancer and other disorders. Smoking and Tobacco Control Monograph No 4, Bethesda, MD: US Department of Health, Public Health Service, National Institute of Health, 1993. NIH Publication No 93-3605. [18] Samet JM, Yang G. Passive smoking, women and children, In: Samet JM, Yoon SY, editors. Women and the tobacco epidemic. Challenges for the 21st century. The World Health Organization in collaboration with the Institute for Global Tobacco Control and the Johns Hopkins School of Public Health, 2001. [19] Hackshaw AK, Law MR, Wald NJ. The accumulated evidence of lung cancer and environmental tobacco smoke. Br Med J 1997;315:980 /8. [20] Boffetta P, Agudo A, Ahrens W, et al. Multicentre case /control study of exposure to environmental case /control study of exposure to environmental tobacco smoke and lung cancer. J Natl Cancer Inst 1998;90:1440 /50. [21] Wicklund KG, Daling JR, Allard J, Weiss NS. Respiratory cancer among orchardists in Washington State, 1968 /1980. J Occup Med 1988;30:561 /4. [22] De Stafani E, Kogevinas M, Boffetta P, Ronco A, Mendilaharsu M. Occupation and the risk of lung cancer in Uruguay. Scand J Work Environ Health 1996;22:346 /52. [23] Safi JM. Association between chronic exposure to pesticides and recorded cases of human malignancy in Gaza Governorates (1990 /1999). Sci Total Environ 2002;284:75 /84. [24] Wesseling C, Antich D, Hogstedt C, Rodrigues AC, Ahlbom A. Geographical differences of cancer incidence in Costa Roca in relation to environmental and occupational pesticide exposure. In J Epidemiol 1999;28:365 /74. [25] Shukla Y, Yadav A, Arora A. Carcinogenic and cocarcinogenic potential of cypermethrin on mouse skin. Cancer Lett 2002;182:33 /41. [26] Gao YT, Blot WJ, Zheng W, et al. Lung cancer among Chinese women. Int J Cancer 1987;40:604 /9. [27] Koo LC, Matsushita H, Ho JHC, et al. Carcinogens in the indoor air of Hong Kong homes: levels, sources, and ventilation effects on 7 polynuclear aromatic hydrocarbons. Environ Technol 1994;15:401 /18. [28] Seow A, Poh WT, The M, et al. Fumes from meat cooking and lung cancer risk in Chinese women. Cancer Epidemiol Biomark Prev 2000;9:1215 /21.
140
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
[29] Chiang TA, Wu PF, Ko YC. Identification of carcinogens in cooking oil fumes. Environ Res 1999;81:18 /22. [30] Schoental R, Gibbard S. Carcinogens in Chinese incense smoke. Nature 1967;216:612. [31] Tokuhata GK, Lilienfeld AM. Familial aggregation of lung cancer in humans. J Natl Cancer Inst 1963;30:289 /312. [32] Wu AH, Fontham ETH, Reynolds P, et al. Family history of cancer and risk of lung cancer among lifetime nonsmoking women in the United States. Am J Epidemiol 1996;143:535 /42. [33] Brownson RC, Alavanja MCR, Caporaso N, et al. Family history of cancer and risk of lung cancer in lifetime nonsmokers and long term ex-smokers. Int J Epidemiol 1997;26:256 /63. [34] Liu ZY, He XZ, Chapman RS. Smoking and other risk factors for lung cancer in Xuanwei. Int J Epidemiol 1991;20:26 /31. [35] Bartsch H, Nair U, Risch A, Rojas M, Wikman H, Alexandrov K. Genetic polymorphism of CYP genes, alone or in combination, as a risk modifier of tobacco-related cancers. Cancer Epidemiol Biomark Prev 2000;9:3 /28. [36] Bouchardy C, Benhamou S, Jourenkova N, Dayer P, Hirvoren A. Metabolic genetic polymorphism and susceptibility to lung cancer. Lung Cancer 2001;32:109 /12. [37] Lan Q, He X, Costa DJ, et al. Indoor coal combustion emissions, GSTM1 and GSTT1 genotypes, and lung cancer risk: a case /
[38]
[39]
[40]
[41]
[42]
[43]
control study in Xuan Wei, China. Cancer Epidemiol Biomark Prev 2000;9:605 /8. Koo LC, Ho JH, Saw D, Ho CY. Measurements of passive smoking and estimates of lung cancer risk among non-smoking Chinese females. Int J Cancer 1987;39:162 /9. Mao Y, Hu J, Ugnat AM, Semenciw R, Fincham S. The Canadian Cancer Registries Epidemiology Research Group. Int J Epidemiol 2001;30:809 /17. Hart CL, Hole DJ, Gillis CR, Smith GD, Watt GC, Hawthorne VM. Social class differences in lung cancer mortality: risk factor explanations using two Scottish cohort studies. Int J Epidemiol 2001;30:268 /74. Alvanja MCR, Brown CC, Swanson C, Brownson RC. Saturated fat intake and lung cancer risk among nonsmoking women in Missouri. J Natl Cancer Inst 1993;85:1905 /16. De Stefani E, Brennan P, Boffetta P, et al. Diet and adenocarcinoma of the lung: a case /control study in Uruguay. Lung Cancer 2002;35:43 /51. Pope CA, Burnett RT, Thun MJ, et al. Lung cancer, cardiopulmonary mortality and long term exposure to fine particle air pollution. J Am Med Assoc 2002;287:1132 /41.
Risk factors associated with lung cancer in Hong Kong Moira Chan-Yeung a,*, L.C. Koo b,c, J.C.-M. Ho a, K.W.-T. Tsang a, W.-S. Chau d, S.-W. Chiu d, M.S.-M. Ip a, W.-K. Lam a a
Division of Respiratory and Critical Care Medicine, Department of Medicine, The University of Hong Kong, 4/F, Professorial Block, Queen Mary Hospital, Hong Kong SAR, China b Department of Epidemiology and Disease Modeling, AstraZeneca, Wilmington, DE, USA c Clinical Epidemiology Unit, CCEB, University of Pennsylvania, Philadelphia, USA d Cardiothoracic Surgical Unit, The Grantham Hospital, Hong Kong SAR, China Received 1 October 2002; received in revised form 16 December 2002; accepted 19 December 2002
Abstract The purpose of this study was to investigate the risk factors associated with lung cancer in Hong Kong. Three hundred and thirtyone histologically or cytologically proven consecutive cases of lung cancer and the same number of in- and out-patients without cancer matched for age and sex were recruited for this study using a detailed questionnaire completed by a trained interviewer. Smoking was the most important risk factor associated with lung cancer but the attributable risk (AR) was estimated to be 45.8% in men and 6.2% in women, considerably lower compared with those estimated in early 1980s. In addition, among women, exposure to environmental tobacco smoke (ETS) at work9/at home and lack of education, were independent risk factors for lung cancer with adjusted odds ratio (OR) 3.60, (95% confidence interval (CI) 1.52 /8.51) and OR 2.41 (95% CI 1.27 /4.55), respectively. Among men, exposure to insecticide/pesticide/herbicide, ETS exposure at work or at home, and a family history of lung cancer and were independent risk factors with adjusted OR 3.29 (95% CI 1.22 /8.9, OR 2.43, 95% CI 1.24 /4.76 and OR 2.37, 95% CI 1.43 /3.94, respectively). Exposure to incense burning and frying pan fumes were not significant risk factors in both sexes. A moderate or high consumption of fat in the diet was associated with increased risk in men but decreased risk in women. The results of this study suggested that as the prevalence of smoking declined, the influence of smoking as a risk factor for lung cancer decreased even further. Moreover, the contribution of other environmental, occupational and socioeconomic factors may be more apparent as etiological factors for lung cancer in a population with relatively high lung cancer incidence but low AR from active smoking. # 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Lung cancer; Risk factors; Smoking; Environmental tobacco smoke; Insecticide; Family history; Socioeconomic factor; Diet
1. Introduction Lung cancer is the most common cause of cancer mortality in Hong Kong and worldwide. In 1999, the incidence of lung cancer in Hong Kong was 58.3/100 000 in men and 23.2/100 000 in women [1]. While cigarette smoking is the most important cause for lung cancer in both men and women in the Western world, the attributable risk (AR) of cigarette smoking was found to be only 35.8% in women in Hong Kong in 1985 [2]. The etiologic or risk factors associated with lung cancer
* Corresponding author. Tel.: /852-2855-4385; fax: /852-28551143. E-mail address: [email protected] (M. Chan-Yeung).
in nonsmoking women are not entirely clear. Exposure to passive smoke, occupational exposures, air pollution, and dietary factors have been implicated. However, the increased risk for lung cancer due to exposure to environmental tobacco smoke (ETS) was found to be small in Hong Kong; with the odds ratios (OR) in two different studies found to be approximately 1.6 [2,3]. There have been many studies investigating the relation between nutrition and lung cancer [4 /11]. After adjusting for smoking, most of the observational studies suggest that increased vegetable and fruit intake is associated with a lower risk for lung cancer in both men and women. In Hong Kong, Koo et al. from a study conducted in the early 1980s [5] reported that higher consumption of leafy green vegetables, carrots,
0169-5002/03/$ - see front matter # 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0169-5002(03)00036-9
132
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
bean curd, fresh fruits and fresh fish was associated with a lower risk for adenocarcinoma in nonsmoking women. There has been a slow decline in the incidence of lung cancer in Hong Kong during the past two decades [1]. The purpose of this study was to investigate the risk factors associated with lung cancer in Hong Kong and whether risk factors have changed during the past two decades since lifestyle, nutrition, and exposure to environmental carcinogens have changed in Hong Kong.
2. Materials and methods 2.1. Subjects Three hundred and thirty-one consecutive histologically or cytologically proven lung cancer patients and the same number of control patients without lung or other cancers, matched for age (9/5 years) and sex, were recruited from the Queen Mary Hospital, a teaching and regional hospital, from May 1999 to December 2001 for this study. The controls were identified from the same hospital ward or from outpatient clinics for respiratory patients. They were all ethnically Chinese. Among the controls, there were 150 (45.6%) patients with asthma, 119 (36%) with obstructive sleep apnea syndrome, 17 (5%) with chronic obstructive lung disease, 33 (10%) with chest infections, and the remaining 12 patients with endocrine disorders, liver problems and skin rash. 2.2. Methods All patients and controls were interviewed using a questionnaire by one trained interviewer. The questions included the following: place of birth and residence, marital history, history of tuberculosis, family history of cancer and lung cancer, and occupational and environmental exposure to various chemicals such as insecticides, pesticides, herbicides and asbestos. In addition, information related to smoking and amount smoked, and exposure to passive smoke, incense and cooking fumes was collected together with a detailed dietary history. Information on past exposure to passive smoke at home from spouses or other family members who smoked, and at work from co-workers who smoked was only obtained from nonsmokers. Nonsmoking status was checked by interviewing family members living with the patient. The dietary section of the interview inquired about the frequency of consumption of a list of food items or food groups as described previously [5]. Briefly, cases were asked to report on their usual food habits 1 year before their diagnosis of cancer or the appearance of symptoms. Controls were asked to recall their usual habits of eating. Each control was usually interviewed
within 6/8 weeks of the matched case’s interview to reduce the effects of seasonality on food frequency pattern. The average number of portions was calculated for each month for each of the following food item: Green vegetables : all leafy green vegetables, cabbage, cauliflower, green pepper, broccoli, and legumes. Root vegetables : carrot, squash, and yam. Fruit : any fruit. Fatty food : ham, sausage, ribs, pigs feet, pig skin, meatball, fried food, butter, ice-cream, and cheese. Preserved food : Chinese preserved meats (siuyuk ), Chinese sausage, pickled vegetables, salted egg, preserved beancurd, salted fish, dried squid, dried octopus, dried oyster, dried shrimp. 2.3. Definitions Ever-smoker : one who had smoked at least one cigarette a day, pipe, water pipes, cigars, and/or hand rolled cigarettes, for 1 year or more. Current smoker */one who was still smoking (any of the above) 1 month prior to the interview. Ex-smoker : one who had ever smoked at least one cigarette a day and any of the above forms of tobacco, for 1 year or more and had given up smoking 1 month prior to the interview. Nonsmoker : a life-long nonsmoker. ETS exposure at home : life-long nonsmoker exposed to anyone who smoked at home regularly for at least 2 years. ETS exposure at work9/at home : life-long nonsmoker exposed to smokers at work in enclosed space9/at home regularly for at least 2 years. Regular exposure to the following was defined as exposure at least once a week for 2 or more years: Paint exposure: regular exposure to paint or lacquer or varnish or paint-thinner at work. Gas exposure: regular exposure to gasoline or kerosene at work. Insecticide exposure: regular exposure to insecticide, pesticide or herbicide at work. Incense exposure: regular exposure to incense burning in the home. Frying pan fumes exposure: regular exposure to cooking fumes from frying in the home. 2.4. Analysis For men, the amount of smoking was studied by combining ex- and current smokers and categorized into light (B/20 pack-years), medium (20 /39 pack-years) and heavy (/40 pack-years). In women, because of small numbers of ex- and current smokers (ten and three, respectively, among cases and four and two
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
among controls), they were combined in the analysis and the dose relationship between smoking and risk of lung cancer was not determined. For exposure to ETS, the effects of exposure at home and at work (with and without exposure at home) were analyzed separately. In men, the number of nonsmokers with exposure to ETS at home was small and ETS exposure at work or at home was used in the analysis. Regular occupational exposure to paint, gasoline, pesticide, and asbestos was categorized into ‘yes’ and ‘no’. Incense smoke exposure was categorized into no exposure, festival only, daily outdoors, daily indoors9/ outdoors. Exposure to cooking fumes from frying was categorized into no exposure, B/3.5, 3.5 /7 and /7 times a week. For dietary data, cases and controls were grouped into tertile levels of consumption (low, medium and high) whenever possible, and into two levels when the consumption was low. The data were analyzed separately for men and women. The unadjusted OR and 95% confidence intervals (CI) for lung cancer was calculated for each of the following demographic characteristics: birthplace; marital, employment and educational status; previous history of tuberculosis; and family history of cancer. ORs and 95% CI for lung cancer were calculated for each of the following risk factors: smoking, exposure to ETS; occupational exposure to paint, gasoline, asbestos, pesticides, insecticides, and herbicides; environmental exposure to incense burning; cooking fumes from frying pans; and intake of various food items. Adjusted ORs and 95% CI were also calculated for each of the risk factors using conditional logistic regression analyses with adjustment made for significant demographic confounders (place of birth, educational status, and family history of lung cancer in men and educational status in women (see Table 1) and smoking. The adjustment of smoking took into consideration the smoking status (non-, ex-, and current smoking), duration and amount of cigarettes smoked (B/20, 20 /39, / 40 pack-years). Conditional logistic regression analysis was carried out including all exposure risk factors to determine the independence of the risk factors adjusted for demographic confounders. The AR of exposure for lung cancer was estimated as follows: AR /P (OR/1)/{P (OR/1)/1}, where OR is the odds ratio and P is the prevalence of exposure [12]. Statistical analyses were carried out using the SPSS for WINDOWS version 10, statistical package (SPSS, Chicago, IL).
2.5. Ethics The study was granted approval by the Ethics Committee, Faculty of Medicine, of the University of Hong Kong.
133
3. Results Three hundred and thirty-one consecutive patients with histologically proven lung cancer and the same number of age (9/5 years) and sex-matched controls were recruited for this study. Table 1 shows the characteristics of the study subjects by sex. There were no significant differences between patients and controls in marital status, place of birth and employment status. However, having no education or only primary school education was associated with a higher risk for lung cancer in both sexes, but after adjustment for smoking, was only found to be significant among women without formal education (OR 3.78, 95% CI 1.62 /8.84). In men, those with a family history of cancer or lung cancer had a significantly higher risk for lung cancer but this was not found in women. Table 2 show the association between smoking and exposure to ETS and lung cancer, adjusted for confounding demographic variables (place of birth, educational status and a family history of lung cancer in men; educational status in women). In men, ex-smokers and current smokers had higher lung cancer risks which were statistically significant for current smokers (adjusted OR 5.53, 95% CI 2.9 /10.6). Among ever smokers, the risk of lung cancer increased significantly with increased amount of smoking in a dose-dependent manner. Among nonsmoking men, exposure to ETS (at home or at work) showed an increase in risk for lung cancer (adjusted OR 3.38, 95% CI 0.86 /13.3) but not statistically significant. In women, ever smokers (ex- and current) showed a significantly higher risk (adjusted OR 3.78, 95% CI 1.11 /12.9). Among nonsmoking women, exposure to ETS at work9/at home significantly increased the risk for lung cancer (adjusted OR 3.23, 95% CI 1.25 /8.36) but not ETS at home only. The estimated ARs of active smoking for lung cancer were 45.8% in men and 6.2% in women. The AR of ETS exposure at work (9/at home) was estimated to be 17.6% in nonsmoking women. The risks of lung cancer from various occupational and environmental exposures are shown in Table 3, adjusted for smoking and confounding demographic variables. In men, exposure to insecticide/pesticide/ herbicides regularly at work significantly increased the risk for lung cancer (adjusted OR 5.57, 95% CI 1.66 / 18.7). The number of women with occupational exposure to these chemicals at work was too small for analysis to be meaningful. Exposure to incense burning during festivals showed an increased risk in women (adjusted OR 2.95, 95% CI 1.10 /7.87) but not daily exposure. Exposure to frying pan fumes from cooking did not increase the risk of lung cancer in either men or women. The association of consumption of various food items and lung cancer is shown in Table 4 adjusted for
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
134 Table 1 Characteristics of the study subjects by sex Men Patients
N Age (years)
Controls
212 57.59/11.1
Marital status Married 176 (83.0) Others 37 (17.0)
OR (95%CI)
Adjusted OR (95% CI)a
Patients
212 57.49/11.2
Controls
119 56.09/11.2
OR (95% CI)
Adjusted OR (95% CI)a
119 55.99/11.3
172 (81.1) 40 (18.9)
1 1 0.88 (0.54 /1.45) 0.70 (0.39 /1.26)
86 (72.3) 33 (27.7)
82 (68.9) 37 (31.1)
1 0.85 (0.9 /1.49)
105 (49.5) 107 (50.5)
1 1 1.55 (1.06 /2.29) 1.39 (0.86 /2.25)
55 (46.2) 64 (53.8)
55 (46.6) 63 (53.4)
1 1 1.02 (0.61 /1.69) 0.96 (0.57 /1.63)
88 (41.5) 34 (16.0)
78 (36.8) 21 (9.9)
1 1 1.44 (0.77 /2.68) 0.75 (0.33 /1.68)
29 (24.4) 10 (8.4)
32 (26.9) 8 (6.7)
1 1 1.38 (0.48 /3.97) 1.67 (0.51 /5.59)
90 (42.5)
113 (53.3)
0.71 (0.47 /1.07) 0.60 (0.30 /1.21)
80 (67.2)
79 (66.4)
1.12 (0.62 /2.02) 1.22 (0.62 /2.38)
127 (59.9)
1
44 (37.0)
58 (49.1)
1
32 (26.9) 43 (36.1)
38 (32.2) 22 (18.6)
1.11 (0.60 /2.05) 1.39 (0.71 /2.72) 2.58 (1.35 /4.92) 3.78 (1.62 /8.84)
111 (93.3) 8 (6.7)
1 1 1.70 (0.68 /4.27) 1.62 (0.65 /4.03)
Place of birth Hong Kong 82 (38.7) China/ 130 (61.3) others Employment Retired Part time/ unemployed Full time/ house wife
Women
Education level Secondary/ 97 (45.7) post secondary Primary 81 (38.3) None 34 (16.0)
63 (29.7) 22 (10.4)
1
1.68 (1.10 /2.57) 1.25 (0.75 /2.09) 2.02 (1.11 /3.68) 1.68 (0.75 /3.76)
1 0.79 (0.44 /1.44)
1
Previous history of TB No 179 (84.8) Yes 32 (15.2)
193 (91.0) 19 (9.0)
1 1 106 (89.1) 1.82 (0.99 /3.32) 1.97 (0.98 /3.94) 13 (10.9)
Family history of cancer No 141 (66.5) Yes 71 (33.5)
162 (76.4) 50 (23.6)
1 1 1.63 (1.07 /2.50) 1.95 (1.16 /3.31)
75 (63.0) 44 (37.0)
83 (69.7) 37 (30.3)
1 1 1.35 (0.79 /2.32) 1.52 (0.83 /2.86)
Family history of Ca lung No 184 (86.8) Yes 28 (13.2)
196 (92.5) 16 (7.5)
1 1 104 (87.4) 1.86 (0.98 /3.56) 3.03 (1.38 /6.68) 15 (12.6)
109 (91.6) 10 (8.4)
1 1 1.57 (0.68 /3.66) 1.82 (0.74 /4.48)
a
Adjusted for smoking amount and duration using conditional logistic regression analysis.
smoking and confounding demographic variables. Higher consumption of carrots and of green vegetables was each associated with a higher risk for lung cancer in both sexes but significant only for the highest level of consumption of green vegetables in men after adjustment. When consumption of all vegetables including green and root vegetables and fruits was combined, no significant increased or reduced risk for lung cancer was found in either men or women. Higher consumption of preserved food and fatty food was each associated with a significantly higher lung cancer risk in men only, but a dose-dependent relationship was not found. Similar analyses were conducted on nonsmoking women only but the results were not different the data are not shown. To study the effect of multiple exposure factors, conditional logistic regression analyses were carried
out with models that included all demographic and exposure variables that affected lung cancer risk with P values /0.15, separately for men and women, as shown in Table 5. In men, smoking, exposure to insecticide/ pesticide/herbicide, medium or high fat content diet, and a family history of lung cancer remained as significant risk factors after adjustment for the other exposure variables; ETS exposure became significant while high consumption of preserved food was no longer significant. In women, ever smoking, ETS exposure at work9/ at home and lack of education remained as risk factors after adjustment for other exposure variables but exposure to incense smoke during festivals was no longer significant. A diet with high fat content in women was associated with a lower risk for lung cancer. High consumption of vegetables and fruits in diet had no influence on the risk of lung cancer in both sexes.
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
135
Table 2 Smoking and exposure to ETS and risk for lung cancer in men/women, adjusted and unadjusted OR (95% CI) N
Patients
Controls
OR (95% CI)
Adjusted OR (95%CI)
(a ) Men Smoking habit Nonsmoker Ex-smoker Current smoker
52 (24.5) 68 (32.1) 92 (43.4)
96 (45.3) 87 (41.0) 29 (13.7)
1 1.44 (0.91 /2.29) 5.86 (3.42 /10.0)
1b 1.26 (0.70 /2.26)b 5.53 (2.90 /10.6)a,b
Ever smokers Nonsmoker Light ( B/20 pack-years) Medium (20 /39 pack-years) Heavy (/40 pack-years)
52 25 50 85
96 41 32 43
(45.3) (19.3) (15.1) (20.3)
1 1.13 (0.62 /2.05) 2.88 (1.65 /5.04) 3.65 (2.22 /6.01)
1b 0.91 (0.44 /1.89)b 3.40 (1.77 /6.50)b 4.76 (2.41 /9.40)a,b
Exposure to ETS in nonsmokers No ETS or B/2 years At work or at home ]/2 years
25 (48.1) 27 (51.9)
67 (69.8) 29 (30.2)
1 2.50 (1.24 /5.01)
1b 3.38 (0.86 /13.28)b
106 (89.1) 13 (10.9)
113 (95.0) 6 (5.0)
1 2.31 (0.85 /6.30)
1e 3.78 (1.11 /12.92)e
53 (50.0) 31 (29.2) 22 (20.8)
69 (61.1) 35 (31.0) 9 (8.0)
1 1.15 (0.63 /2.10) 3.18 (1.35 /7.48)
1e 1.01 (0.47 /2.18)e 3.23 (1.25 /8.36)d,e
(24.5) (11.8) (23.6) (40.1)
(b ) Women Smoking habit Nonsmoker Ex- and current smokersc Exposure to ETS in nonsmokers No ETS or B/2 years ETS at home ]/2 years At work9/at home ]/2 years
a A significant (P B/0.05) increasing trend in the OR was found between nonsmokers, ex- and current smokers; and increasing amount of smoking among the ever smokers. b Adjusted for place of birth, educational status, and a family history of lung cancer by conditional logistic regression analysis. c Ex- and current smokers combined because of small numbers. d A significant (P B/0.05) increasing trend in OR was found between no ETS exposure, exposure to ETS at home only, and ETS exposure at work9/at home. e Adjusted for educational status by conditional logistic regression analysis.
4. Discussion In this study, we found that smoking was the most important risk factor for lung cancer in both sexes, as expected. However, among men, additional significant risk factors included ETS exposure at work or at home, occupational exposure to pesticide/insecticide/herbicide, diets with medium to high fat content, and a family history of cancer or lung cancer. In women, exposure to ETS at work with or without exposure at home, and having no education were other significant risk factors found in this study. Smoking is a well-recognized risk factor in both men and women for lung cancer. In 1985, the AR of smoking for lung cancer was found to be 95.4% in men and 35.8% in women [2]. In this study, the AR of smoking has declined to 45.8 and 6.2%, respectively. These figures, however, are similar to the lower AR reported in China in the 1989/1990 (52.3 and 19.4%, respectively) [13]. In Hong Kong, the prevalence of smoking has declined from 1982 to 2000 to be 39.7 and 5.6% to 25.2 and 4.4%, respectively, in men and women over the age of 15 years [14]. The incidence of lung cancer in Hong Kong has also declined from 84.6/100 000 in 1983 to 58.6/100 000
in 1999 in men, and from 31.5 to 23.2/100 000, respectively, in women [1]. As there is a latent period between the onset of smoking and the development of lung cancer, a firm conclusion cannot be made that the decline in lung cancer incidence is due to the reduction in the prevalence of smoking in Hong Kong as smoking data from the general population was not available before 1983. It is possible that some patients in this study failed to report their smoking habit to avoid blame as it is more widely known now compared with early 1980s that smoking can cause lung cancer. This is especially true for women as it is associated with more social stigma for a Chinese woman than a man to be a smoker. We attempted to verify the nonsmoking status by interviewing with family members living with the patient. Because of limited resources, we did not examine urinary or hair cotinine level to validate the patients’ current smoking status. In this study, our controls were patients from the same hospital who mostly suffered from chronic lung disease such as asthma and chronic obstructive lung disease and did not have lung cancer or other types of cancer. The prevalence of current smokers among our male control subjects in this study is higher than those of
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
136
Table 3 Environmental exposures and risk for lung cancer, unadjusted and adjusted OR (95% CI) Men Patients
Women Controls
OR (95%CI)
Adjusted OR (95%CI)a
Patients
172 (81.1)
173 (81.6)
1
1
113 (95.0)
116 (97.5)
40 (16.9)
39 (18.4)
1.03 (0.63 /1.68)
0.93 (0.53 /1.62)
6 (5.0)
3 (2.5)
161 (75.9)
166 (78.3)
1
1
111 (93.3)
115 (96.6)
51 (24.1)
46 (21.7)
1.14 (0.73 /1.80)
1.02 (0.59 /1.78)
8 (6.7)
4 (3.4)
192 (90.6)
206 (97.2)
1
1
116 (97.5)
118 (99.2)
20 (9.4)
6 (2.8)
3.58 (1.41 /9.09)
5.57 (1.66 /18.7)
3 (2.5)
1 (0.8)
206 (97.2)
209 (99.1)
1
1
6 (2.8)
2 (0.9)
3.04 (0.61 /15.25)
4.91 (0.73 /32.9)
Controls
OR (95% CI)
Adjusted OR (95% CI)a
1
1
2.05 (0.50 /8.41)
1.53 (0.34 /6.82)
1
1
2.07 (0.61 /7.07)
1.36 (0.38 /4.91)
1
1
3.05 (0.31 /29.8)
7.70 (0.52 /114.8)
Occupational exposure Paints No or B/2 years Yes Gasoline No or B/2 years Yes Pesticide No or B/2 years Yes Asbestos No or B/2 years Yes
117 (98.3)
118 (100)
1
1
2 (1.7)
/
/
/
Environmental exposure Incense No or B/2 years Festival only Daily Frying food No or B/2 years B/3.5 3.5 /7 /7
87 (41.0)
86 (43.2)
1
1
30 (25.2)
42 (37.2)
1
1
25 (11.8) 100 (47.2)
20 (10.1) 93 (46.7)
1.24 (0.64 /2.39) 1.06 (0.71 /1.60)
1.03 (0.47 /2.26) 0.94 (0.56 /1.56)
18 (15.1) 71 (59.7)
11 (9.7) 60 (53.1)
2.29 (0.95 /5.55) 1.66 (0.93 /2.96)
2.95 (1.10 /7.87) 1.58 (0.77 /3.26)
146 (70.2)
131 (67.2)
1
1
34 (28.6)
34 (30.1)
1
1
27 (13.0) 22 (10.6) 13 (6.3)
33 (16.9) 24 (12.3) 7 (3.6)
0.73 (0.42 /1.29) 0.82 (0.44 /1.54) 1.67 (0.65 /4.30)
0.69 (0.32 /1.49) 0.83 (0.38 /1.80) 1.22 (0.38 /3.99)
37 (31.1) 27 (22.7) 21 (17.6)
36 (31.9) 26 (23.0) 17 (15.0)
1.03 (0.53 /1.99) 1.04 (0.51 /2.13) 1.24 (0.56 /2.74)
1.08 (0.50 /2.32) 1.05 (0.46 /2.42) 1.54 (0.57 /4.13)
a Adjusted for place of birth, educational status, a family history of lung cancer and smoking in men; educational status and smoking status in women using conditional logistic regression analysis.
the general population, 43.4 and 25.2%, respectively; the corresponding figures for women were 1.7 and 4.4%, respectively. Thus the use of male patients with a higher prevalence of smoking may have underestimated the effect of active smoking in men and overestimated its effects in women. The female controls were mostly nonsmoking asthmatics. Exposure to ETS as a risk factor for lung cancer in nonsmokers has been studied extensively during the past few decades and many comprehensive reviews of the health consequences of ETS have been published [15 / 20]. It is generally agreed that similar carcinogens are present in mainstream tobacco smoke as in ETS, and that nonsmokers are exposed to carcinogens in the ETS. The meta-analysis carried out by the US Environmental Protection Agency in 1992 concluded that ETS exposure is associated with an increased risk for lung cancer in nonsmoking adults accounting for approximately 3000 deaths annually in the US [17]. The majority of the
studies on ETS exposure were from spousal smoke rather than exposure at work. In this study, we found ETS exposure at work with or without exposure at home increased the risk of lung cancer in nonsmoking men and women. It should be noted that the number of nonsmoking women with ETS exposure in this study is small. Compared with two decades ago, ETS exposure at work has become more important as a risk factor in Hong Kong where there is no regulation to prohibit smoking in the workplace. Although there might have been recall bias and the actual number of subjects involved was small for patient and control groups, our data showed that regular (at least once weekly for 2 or more years) workplace exposure to insecticides, pesticides and herbicides was associated with a significantly higher lung cancer risk for men. Exposure to DDT has been found to increase the risk of lung cancer among orchard and construction workers [21,22]. A recent study in Gaza also showed
Table 4 Consumption of food (number of portions per week) and risk of lung cancer, unadjusted and adjusted OR and 95% CI Portion/month
Men
Tomato Carrots
Green vegetables
Fruits
Green vegetables and tubers and fruits
Preserved food
Fatty food
a
B/2 ]/2 B/2 2 /4 /4 B/45 45 /65 65/ B/15 15 /30 30/ B/70 70 /100 100/ B/1 1 /2 /2 B/1 1 /2 2/
96 114 55 80 77 72 43 97 92 54 66 68 88 56 70 89 53 51 72 89
Controls
OR (95% CI)
(45.7) 94 (44.3) 1 (54.5) 117 (55.7) 0.95 (0.65 /1.4) (25.9) 76 (35.8) 1 (37.7) 89 (42.0) 1.24 (0.78 /1.97) (36.4) 47 (22.2) 2.26 (1.37 /3.74) (34.0) 94 (44.3) 1 (20.2) 52 (24.5) 1.08 (0.65 /1.79) (45.8) 66 (31.2) 1.92 (1.24 /2.97) (43.4) 77 (36.3) 1 (25.5) 56 (26.4) 0.81 (0.50 /1.31) (31.1) 79 (37.3) 0.70 (0.45 /1.09) (32.1) 79 (37.3) 1 (41.5) 82 (38.7) 1.25 (0.80 /1.94) (26.4) 51 (24.1) 1.28 (0.77 /2.10) (33.0) 81 (38.6) 1 (42.0) 60 (28.6) 1.72 (1.09 /2.71) (25.0) 69 (32.8) 0.89 (0.55 /1.44) (24.1) 80 (38.1) 1 (34.0) 53 (25.2) 2.13 (1.29 /3.51) (41.9) 77 (36.7) 1.81 (1.14 /2.89)
Adjusted OR (95% CI)a
Patients
Controls
OR (95%CI)
Adjusted OR (95%CI)a
1 0.97 1 1.46 2.22 1 1.15 2.23 1 0.89 0.65 1 1.43 1.42 1 1.83 0.86
60 59 36 42 41 20 36 63 36 22 61 21 43 55 61 35 23 50 47 22
46 73 39 49 30 34 27 58 35 37 47 32 36 51 63 34 19 43 39 34
1 0.62 1 0.93 1.48 1 2.27 1.85 1 0.58 1.26 1 1.82 1.64 1 1.06 1.25 1 1.04 0.56
1 0.74 1 1.12 2.05 1 2.11 1.74 1 0.58 1.25 1 1.60 1.86 1 1.46 1.19 1 0.94 0.45
(0.62 /1.52) (0.82 /2.60) (1.22 /4.06) (0.60 /2.23) (1.28 /3.89) (0.49 /1.59) (0.36 /1.16) (0.84 /2.43) (0.78 /2.58) (1.06 /3.16) (0.47 /1.55)
2.37 (1.26 /4.46) 1.61 (0.89 /2.94)
(50.4) (49.6) (30.3) (35.3) (34.3) (16.8) (30.3) (52.9) (30.3) (18.5) (51.3) (17.6) (36.1) (46.2) (51.3) (29.4) (19.3) (42.0) (39.5) (18.5)
(38.7) (61.3) (33.1) (41.5) (25.4) (28.6) (22.7) (48.7) (29.4) (31.1) (39.5) (26.9) (30.3) (42.9) (54. 3) (29.3) (16.4) (37.1) (33.6) (29.3)
(0.37 /1.04) (0.50 /1.71) (0.77 /2.85) (1.08 /4.77) (0.96 /3.56) (0.29 /1.17) (0.69 /2.30) (0.90 /3.69) (0.84 /3.21) (0.59 /1.92) (0.62 /2.52) (0.58 /1.87) (0.28 /1.09)
(0.43 /1.28) (0.56 /2.23) (0.95 /4.42) (0.91 /4.86) (0.86 /3.52) (0.27 /1.22) (0.65 /2.44) (0.74 /3.47) (0.89 /3.90) (0.71 /3.01) (0.56 /2.52)
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
Patients
Women
(0.45 /1.96) (0.20 /1.02)
Adjusted for place of birth, educational status, a family history of lung cancer and smoking in men; educational status and smoking in women using conditional logistic regression analysis.
137
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
138
Table 5 Conditional logistic regression analysis: model included all exposure variables adjusted for demographic confounders Risk factors
Men OR (95% CI)
Women OR (95% CI)
History of tuberculosis Yes vs. no
1.89 (0.95 /3.76)
Family history of cancer Yes vs. no
2.37 (1.43 /3.94)
Education None vs. secondary and above
1.78 (0.86 /3.76) 2.41 (1.27 /4.55)
Smoking Medium vs. nonsmoking Heavy vs. nonsmoking Any smoking vs. nonsmoking ETS in nonsmokers Any exposure vs. none
5.13 (2.73 /9.65) 6.39 (3.56 /11.5) 5.08 (1.32 /19.5) 2.43 (1.24 /4.76)
ETS Exposure at work9/home vs. none
3.60 (1.52 /8.51)
Diet Medium fat vs. low fat High fat vs. low fat
2.71 (1.51 /4.87) 2.53 (1.46 /4.39) 0.43 (0.22 /0.85)
Exposure to insecticide/pesticide Yes vs. no
3.29 (1.22 /8.90)
Variables with P -value /0.15 were removed from the final model.
highly significant correlations between cancer risk and the use of insecticides, fungicides, herbicides, nematicides and all types of pesticides [23]. Another study from Costa Rica showed that heavy pesticide usage was associated with an overall increase of cancer incidence including that of lung cancer [24]. Potent insecticides, such as pyrethroids, which account for more than 30% of global use, are known to be carcinogenic [25]. Further study is needed to explore the types of insecticide/ pesticide/herbicide used locally at present and in the past and their association with lung cancer. Exposure to incense smoke and exposure to frying pan fumes were found to be associated with a higher risk for lung cancer in nonsmoking Chinese women due to the presence of combustion-related carcinogens [26 /30]. We have not been able to find such a relationship with either exposure in this study. Even though about 70% of women claimed that they were exposed to incense burning, many were exposed during festivals and outdoors only. Indoor incense burning less commonly practice in Hong Kong with progressive westernization. The use of extraction fans and the practice of opening kitchen windows during cooking reduced exposure to cooking fumes in Hong Kong. The association between lung cancer and frying pan fumes has been found in Shanghai with the use of rapeseed oil [29], which is not commonly used in Hong Kong. Increased lung cancer
risk was also found to be associated with exposure to fumes from kerosene stoves, which are no longer used in Hong Kong [3]. Thus some of the sources of exposure that carry an increased risk for lung cancer nonsmoking women locally are no longer common. A reported family history of cancer or lung cancer increased the risk of lung cancer significantly in men and a similar trend in women. Most of the published literature have shown a smoking-adjusted increased risk of lung cancer by about 2 /4-fold associated with a family history of lung cancer [31 /34]. As only a small proportion of smokers develop lung cancer, genetic susceptibility is likely to be important, although recall bias could also contribute to this finding. The issue of genetic predisposition of lung cancer is being actively investigated by many researchers at present [35 /37]. We found that the lack of any education, a surrogate for socioeconomic status, is an independent risk factor for lung cancer in women, increasing the risk by twofold after adjusting for smoking and exposure to ETS. This association has also been found in a previous study on lung cancer risk in Hong Kong [38]. Socioeconomic status has been reported as a risk factor for lung cancer [39,40]. A careful study of the exposure of uneducated and poor women with lung cancer may give us some clues as to the etiology of lung cancer in women. There have been many epidemiologic studies investigating the relationship between lung cancer and nutrition. In general, the majority of the case /control studies demonstrated that a higher consumption of vegetables and fruits is associated with a lower lung cancer risk in both women and men after adjusting for smoking [5 / 11]. In Hong Kong, Koo and colleagues [5] studied 88 nonsmoking women with lung cancer and 137 matched controls and found that consumption of leafy green vegetables, carrots, bean curd, fresh fruits and fresh fish was associated with a protective effect on adenocarcinoma. By using the same methodology of data collection, we found a contrary effect, that f higher consumption of green vegetables, fruits and carrots were associated with higher risk. On the other hand, we found high fat consumption was associated with an increase in lung cancer risk in men but a lower risk in women, after adjustment for smoking and other demographic and exposure variables. A positive association between high fat consumption and lung cancer risk has been reported [41,42]. The discrepancy in findings between the previous study and the present one despite using the same method of data collection could be due to several reasons. In the previous study, the controls were recruited from a healthy population while in the present study the controls were recruited from patients with mostly lung disease but without any cancer in the hospital or outpatient clinics. Dietary intake of leafy green vegetables and fruits in patients with chronic lung disease and obstructive sleep apnea may be quite similar
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
to those with lung cancer and different from healthy controls. In this study we did not stratify patients according to histological types because of sample size limitations. It is also possible that there has been a change in the dietary pattern in Hong Kong with a higher consumption of meat and fatty food compared with 20 years ago. The differential effect of high fat consumption between men and women on lung cancer risk is hard to explain but the number of women among cases and controls who had a high fat consumption was small. In summary, our study has shown that while cigarette smoking is still a major cause of lung cancer in men, its influence is decreasing. Our results also suggest that other environmental, occupational and socioeconomic issues may be important in the development of lung cancer. As lung cancer is the commonest lethal malignant disease worldwide, it is necessary to continue to search for other relevant environmental exposures, combined with genetic predispositions. One obvious one is long term exposure to combustion-related fine particles from air pollution, which has been shown to be associated with increase in lung cancer mortality recently [43].
Acknowledgements The authors wish to thank the Anti-Cancer Society of Hong Kong for its generous support for this study and Anne DyBuncio for data analysis.
References [1] Hong Kong Cancer Registry. Hospital Authority, Hong Kong, SAR 2000. [2] Koo LC, Ho JH-C, Lee N. An analysis of some risk factors for lung cancer in Hong Kong. Int J Cancer 1985;35:149 /55. [3] Lam TH, Kung ITM, Wong CM, et al. Smoking, passive smoking and histological types in lung cancer in Hong Kong Chinese women. Br J Cancer 1987;56:673 /8. [4] Koo LC, Ho JH-C. Diet as a confounder of the association between air pollution and female lung cancer: Hong Kong studies on exposures to environmental tobacco smoke, incense, and cooking fumes as examples. Lung Cancer 1996;14(Suppl. 1):S47 /61. [5] Koo LC. Dietary habits and lung cancer risk among Chinese females in Hong Kong who never smoked. Nutr Cancer 1988;11:155 /72. [6] Wu-Williams AH, Dai XD, Blot W, et al. Lung cancer among women in Northeast China. Br J Cancer 1990;62:982 /7. [7] Darby S, Whitley E, Doll R, Key T, Silcocks P. Diet, smoking and lung cancer: a case /control study of 1000 cases and 1500 controls in south-west England. Br J Cancer 2001;84:728 /35. [8] Feskanich D, Ziegler RG, Michaud DS, et al. Prospective study of fruit and vegetable consumption and risk of lung cancer among men and women. J Natl Cancer Inst 2000;92:1812 /23.
139
[9] Jansen MCJF, Bueno-de-Mesquita HB, Rasanen L, et al. Cohort analysis of fruit and vegetable consumption and lung cancer mortality in European men. Int J Cancer 2001;92:913 /8. [10] Feskanich D, Ziegler RG, Michaud DS, et al. Prospective study of fruit and vegetable consumption and risk of lung cancer among men and women. J Natl Cancer Inst 2000;92:1812 /23. [11] Voorrips LE, Goldbohm RA, Verhoeven DTH, et al. Vegetables and fruit consumption and lung cancer risk in the Netherland cohort study on diet and cancer. Cancer Causes Control 2000;11:101 /15. [12] Breslow NE, Day N. Statistical methods in cancer research. In: Davis W, editor. The analysis of case /control studies, vol. 1. Lyons: International Agency for Research on Cancer, 1980:70. [13] Liu BQ, Peto R, Chen ZM, et al. Emerging tobacco hazards in China: 1. Retrospective proportional mortality study of 1 million deaths. Br Med J 1998;317:1411 /22. [14] Census and Statistics Department, Hong Kong Government. Hong Kong: General Household Surveys, 2000. [15] Koo LC, Ho JH. Worldwide epidemiological patterns of lung cancer in nonsmokers. Int J Epidemiol 1990;19:S14 /23. [16] US Department of Health Education and Welfare. The health consequences of smoking. A report of the Surgeon General. Washington, DC: US Government Printing Office, 1971. DHEW publication No HSM 73-8704. [17] US Environmental Protection Agency. Respiratory health effects of passive smoking: Lung cancer and other disorders. Smoking and Tobacco Control Monograph No 4, Bethesda, MD: US Department of Health, Public Health Service, National Institute of Health, 1993. NIH Publication No 93-3605. [18] Samet JM, Yang G. Passive smoking, women and children, In: Samet JM, Yoon SY, editors. Women and the tobacco epidemic. Challenges for the 21st century. The World Health Organization in collaboration with the Institute for Global Tobacco Control and the Johns Hopkins School of Public Health, 2001. [19] Hackshaw AK, Law MR, Wald NJ. The accumulated evidence of lung cancer and environmental tobacco smoke. Br Med J 1997;315:980 /8. [20] Boffetta P, Agudo A, Ahrens W, et al. Multicentre case /control study of exposure to environmental case /control study of exposure to environmental tobacco smoke and lung cancer. J Natl Cancer Inst 1998;90:1440 /50. [21] Wicklund KG, Daling JR, Allard J, Weiss NS. Respiratory cancer among orchardists in Washington State, 1968 /1980. J Occup Med 1988;30:561 /4. [22] De Stafani E, Kogevinas M, Boffetta P, Ronco A, Mendilaharsu M. Occupation and the risk of lung cancer in Uruguay. Scand J Work Environ Health 1996;22:346 /52. [23] Safi JM. Association between chronic exposure to pesticides and recorded cases of human malignancy in Gaza Governorates (1990 /1999). Sci Total Environ 2002;284:75 /84. [24] Wesseling C, Antich D, Hogstedt C, Rodrigues AC, Ahlbom A. Geographical differences of cancer incidence in Costa Roca in relation to environmental and occupational pesticide exposure. In J Epidemiol 1999;28:365 /74. [25] Shukla Y, Yadav A, Arora A. Carcinogenic and cocarcinogenic potential of cypermethrin on mouse skin. Cancer Lett 2002;182:33 /41. [26] Gao YT, Blot WJ, Zheng W, et al. Lung cancer among Chinese women. Int J Cancer 1987;40:604 /9. [27] Koo LC, Matsushita H, Ho JHC, et al. Carcinogens in the indoor air of Hong Kong homes: levels, sources, and ventilation effects on 7 polynuclear aromatic hydrocarbons. Environ Technol 1994;15:401 /18. [28] Seow A, Poh WT, The M, et al. Fumes from meat cooking and lung cancer risk in Chinese women. Cancer Epidemiol Biomark Prev 2000;9:1215 /21.
140
M. Chan-Yeung et al. / Lung Cancer 40 (2003) 131 /140
[29] Chiang TA, Wu PF, Ko YC. Identification of carcinogens in cooking oil fumes. Environ Res 1999;81:18 /22. [30] Schoental R, Gibbard S. Carcinogens in Chinese incense smoke. Nature 1967;216:612. [31] Tokuhata GK, Lilienfeld AM. Familial aggregation of lung cancer in humans. J Natl Cancer Inst 1963;30:289 /312. [32] Wu AH, Fontham ETH, Reynolds P, et al. Family history of cancer and risk of lung cancer among lifetime nonsmoking women in the United States. Am J Epidemiol 1996;143:535 /42. [33] Brownson RC, Alavanja MCR, Caporaso N, et al. Family history of cancer and risk of lung cancer in lifetime nonsmokers and long term ex-smokers. Int J Epidemiol 1997;26:256 /63. [34] Liu ZY, He XZ, Chapman RS. Smoking and other risk factors for lung cancer in Xuanwei. Int J Epidemiol 1991;20:26 /31. [35] Bartsch H, Nair U, Risch A, Rojas M, Wikman H, Alexandrov K. Genetic polymorphism of CYP genes, alone or in combination, as a risk modifier of tobacco-related cancers. Cancer Epidemiol Biomark Prev 2000;9:3 /28. [36] Bouchardy C, Benhamou S, Jourenkova N, Dayer P, Hirvoren A. Metabolic genetic polymorphism and susceptibility to lung cancer. Lung Cancer 2001;32:109 /12. [37] Lan Q, He X, Costa DJ, et al. Indoor coal combustion emissions, GSTM1 and GSTT1 genotypes, and lung cancer risk: a case /
[38]
[39]
[40]
[41]
[42]
[43]
control study in Xuan Wei, China. Cancer Epidemiol Biomark Prev 2000;9:605 /8. Koo LC, Ho JH, Saw D, Ho CY. Measurements of passive smoking and estimates of lung cancer risk among non-smoking Chinese females. Int J Cancer 1987;39:162 /9. Mao Y, Hu J, Ugnat AM, Semenciw R, Fincham S. The Canadian Cancer Registries Epidemiology Research Group. Int J Epidemiol 2001;30:809 /17. Hart CL, Hole DJ, Gillis CR, Smith GD, Watt GC, Hawthorne VM. Social class differences in lung cancer mortality: risk factor explanations using two Scottish cohort studies. Int J Epidemiol 2001;30:268 /74. Alvanja MCR, Brown CC, Swanson C, Brownson RC. Saturated fat intake and lung cancer risk among nonsmoking women in Missouri. J Natl Cancer Inst 1993;85:1905 /16. De Stefani E, Brennan P, Boffetta P, et al. Diet and adenocarcinoma of the lung: a case /control study in Uruguay. Lung Cancer 2002;35:43 /51. Pope CA, Burnett RT, Thun MJ, et al. Lung cancer, cardiopulmonary mortality and long term exposure to fine particle air pollution. J Am Med Assoc 2002;287:1132 /41.