- Email: [email protected]
Cigarette smoking and changing trends of lung cancer incidence by histological subtype among Chinese male population
Lung Cancer 66 (2009) 22–27
Contents lists available at ScienceDirect
Lung Cancer journal homepage: www.elsevier.com/locate/lungcan
Cigarette smoking and changing trends of lung cancer incidence by histological subtype among Chinese male population Lap Ah Tse a,b , Oscar Wai-Kong Mang c , Ignatius Tak-Sun Yu b,∗ , Fan Wu d , Joseph Siu-Kie Au e , Stephen Chun-Key Law c,e a
School of Public Health, The Chinese University of Hong Kong, Hong Kong SAR, China Department of Community & Family Medicine, The Chinese University of Hong Kong, Hong Kong SAR, China Hong Kong Cancer Registry, Hospital Authority, Hong Kong SAR, China d Shanghai Municipal Center for Disease Control & Prevention, China e Department of Clinical Oncology, Queen Elizabeth Hospital, Kowloon, Hong Kong SAR, China b c
a r t i c l e
i n f o
Article history: Received 1 October 2008 Received in revised form 10 December 2008 Accepted 21 December 2008 Keywords: Lung neoplasm Incidence Histology Smoking Time trends Age-period-cohort models
a b s t r a c t We analyzed the time trends of lung cancer by histological subtype in Hong Kong during 1991–2005, and examined how the time trends were influenced by the effects of birth cohort and calendar period of diagnosis. Cancer incidence data were obtained from Hong Kong Cancer Registry and population data from Census and Statistics Department. Age-standardized incidence rates were computed by the direct method using WHO 1966 standard population as reference. Period and cohort effects were assessed by using two separate Poisson regression models adjusting for age. From 1991 to 2005, the incidence rates in Hong Kong Chinese males decreased steadily. The decline in overall lung cancer incidence rates was limited primarily to the decrease in squamous cell carcinoma, which could be explained by the decreasing trend of cigarette smoking. Adenocarcinoma had been the most predominant histological subtype all along. The relatively horizontal trend of adenocarcinoma and the lack of cohort effect implied the important roles of gene-environment interaction and/or the use of low-tar and filter tip cigarettes. Our study suggests that different histological subtypes may represent different disease entities with perhaps some distinct risk factors. The hypotheses generated from this ecological study will need confirmation by subsequent analytic studies. © 2008 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Lung cancer, a rare malignancy at the beginning of the 20th century, has become a leading cause of death throughout the world [1]. In Hong Kong, lung cancer incidence rates among males started to decrease since 1993 but lung cancer remains as the leading cancer [2]. This trend mirrors that of many Western countries a decade ago [3–5]. Numerous epidemiologic studies have been done in response to the global ‘epidemic’ of lung cancer and convincingly estab-
Abbreviations: HKCR, Hong Kong Cancer Registry; WHO, World Health Organization; ADC, adenocarcinoma; SQCC, squamous cell carcinoma; SCC, small cell carcinoma; LCC, large cell carcinoma; OSC, other specified and/or unspecified histological subtypes; UTT, unknown tumour type; APC, annual percentage change; RR, relative risk; CI, confidence interval. ∗ Corresponding author at: Department of Community & Family Medicine, The Chinese University of Hong Kong, 4/F School of Public Health, Prince of Wales Hospital, Shatin, N.T., Hong Kong SAR, China. Tel.: +852 2252 8773; fax: +852 2606 3500. E-mail address: [email protected] (I.T.-S. Yu). 0169-5002/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.lungcan.2008.12.023
lished cigarette smoking as the main cause of lung cancer [6]. Ethnic disparities seem to exist for lung cancer etiology. In Western populations, more than 80% lung cancer risks among the males could be attributable to cigarette smoking [7], whereas the figure was much lower (58%) for Chinese males [8,9]. Other risk factors, including occupational carcinogens (e.g., crystalline silica dust, asbestos), residential radon, passive smoking, ambient air pollution, as well as other unknown risk factors, might have contributed to the other 42% of lung cancer risk among Hong Kong male population. Notable changes of histological subtypes of lung cancer have also occurred over time [4,10]. Squamous cell carcinoma was the most frequent histological subtype of lung cancer in the initial decades of the smoking-induced epidemic of lung cancer but it has gradually been replaced by adenocarcinoma since mid-1970s [4]. Changes in histological subtypes of lung cancer might be indicative of changes in etiology. The aim of this study was to extend a previous analysis describing the epidemiology of lung cancer incidence in Hong Kong [11] by examining how the time trends of male lung cancer incidence rates of different histological subtypes were influenced by the effects of calendar period of diagnosis and cohort year of birth that were not studied previously. Effects of cigarette smoking
Deviance/DF Deviance Deviance/DF
7065.8 220.8 1841.2 83.7 19.1 (p < 0.001) 0.95 184.7 (p < 0.001) 18.5
Deviance/DF Deviance
5481.9 171.3 1070.9 48.7 53.5 (p < 0.001) 2.7 264.1 (p < 0.001) 26.4
Deviance/DF Deviance
5785.6 180.8 24.1 1.1 19.5 (p < 0.001) 1.0 10.4 (p < 0.001) 1.0
Deviance/DF Deviance
15748.9 492.2 29.8 1.4 29.2 (p < 0.001) 1.5 15.5 (p < 0.001) 1.6
Deviance/DF Deviance
16055.3 501.7 231.5 10.5 54.1 (p < 0.001) 2.7 12.3 (p < 0.001) 1.2
p values in parenthesis are for model improvements with the addition of the parameters compared with the age alone models. † DF, degree of freedom.
We used the direct method to calculate age-standardized annual incidence rate for each histological subtype of lung cancer using the WHO standard population (1966) as reference. Annual percentage change (APC) in incidence rate was calculated by using nonlinear regression under the assumption of same rate of change throughout the study period.
Deviance
2.3. Statistical analyses
Adenocarcinoma
Prevalence of daily smoking in different age groups among Hong Kong male population during 1982–2005 was obtained from the General Household Surveys of Hong Kong Census and Statistics Department [13], while the 1974 data were from an ad hoc survey carried out by a research project [14]. Daily smokers were defined as those persons aged 15 and over whom at the time of enumeration had a daily cigarette smoking habit (although they might not smoke on certain days because of illness or other reasons) [13].
Squamous cell carcinoma
2.2. Source of prevalence data on daily smoking
Variables in DF† the model
Lung cancer incidence data during the period 1991–2005 were obtained from Hong Kong Cancer Registry (HKCR) who has continuously collected cancer information by checking the computerised databases and/or medical records with all public and most private hospitals. With the maximum coverage, the incidence data taken from HKCR is considered of good representation of the territory [11]. Lung cancer was classified into six histological subtypes according to the criteria of the World Health Organization (WHO) histological typing of lung tumours [12]; adenocarcinoma (ADC), squamous cell carcinoma (SQCC), small cell carcinoma (SCC), large cell carcinoma (LCC), and other specified and/or unspecified histological subtypes (OSC). Those without pathological verification were regarded as unknown tumour type (UTT). Only the registered male lung cancer cases were included in the present study. Population data during the corresponding period were obtained from Hong Kong Census and Statistics Department. Mid-year population data were employed in the calculation of the incidence rate.
Table 1 Results of Poisson regression models for lung cancer incidence by histological type in Hong Kong, 1991–2005.
2.1. Incidence and population data
Small cell carcinoma
Large cell carcinoma
2. Materials and methods
32 22 20 10
and the etiological implications on different histological subtypes of lung cancer were discussed.
Intercept Age Age + period Age + cohort
Others specified or unspecified
Fig. 1. Incidence of male lung cancer by histology among Hong Kong male population during 1991–2005.
22052.4 689.1 1621.0 73.7 306.3 (p < 0.001) 15.3 81.2 (p < 0.001) 8.1
23
Unknown tumour type
L.A. Tse et al. / Lung Cancer 66 (2009) 22–27
24
L.A. Tse et al. / Lung Cancer 66 (2009) 22–27
Fig. 2. Prevalence of male daily cigarette smokers by age in Hong Kong, 1974–2005.
The multiple Poisson regression analyses were carried out to examine the period and birth cohort effects on histology-specific incidence rates of lung cancer using SAS procedure GENMOD after the adjustment of age [15]. Three 5-year periods (1991–1995, 1996–2000, and 2001–2005) were covered. Because of very few lung cancer cases diagnosed below 30, we only included 11 age groups by 5-year interval (ranging from 30–34 to 75–79, and 80 or above) in the multivariate analyses. Based on the period and age groups, we calculated 13 partially overlapping 5-year birth cohorts starting in “before 1915” and ending in 1971–1975. The reference categories were the 50–54 years age group, 1991–1995 calendar period, and 1941–1945 birth cohort. A model including all three time variables was not computed because the use of a full age-period-cohort model may involve serious methodological difficulties [16,17]. The goodness of fit of the models was assessed by the deviance. The closer the deviance to the degrees of freedom, the better the model fit [16,17]. 3. Results 3.1. Incidence rates by histological subtype A total of 38,449 new cases of male lung cancer were diagnosed over the study period, with 80.8% confirmed by histology in 2005 compared to 61.3% in 1991. Sixteen cases with unknown age were excluded from further analyses. The age-standardized incidence rates for all male lung cancers decreased steadily from 73.37 in 1991 to 51.33 per 100,000 in 2005 with an APC of −2.56 (95% confidence interval [CI]: −2.11, −3.02) (Fig. 1). The overall declined rate was mainly due to the significant decline of SQCC (APC = −3.59; 95%CI: −4.51, −2.66), LCC (APC = −8.58; 95%CI: −4.26, −12.70), and UTT (APC = −9.25; 95%CI: −7.34, −11.13). We observed a notable increase for OSC incidence (APC = 16.01; 95%CI: 8.50, 24.04), while the trend for ADC
(APC = 0.31; 95%CI: −0.61, 1.23) and SCC (APC = −0.58; 95%CI: −1.56, 0.40) remained relatively steady. ADC had been the most predominant histological subtype throughout the study period (36–42.8%). 3.2. Trend of daily smoking Prevalence of daily smoking among males decreased substantially during the period 1974–2005. Such decreasing trend was reflected in all age groups with the largest change for males aged above 60 years old (Fig. 2), while the trend was relatively flat for the younger groups. 3.3. Birth cohort and period effects adjusted for age Deviance statistics from Table 1 suggest that the age-cohort model fitted better for SQCC but the age-period model fitted better for LCC and OSC, as shown by the deviance of the better model fits closer to the degrees of freedom. The goodness of fits for age-period models differed slightly from those of age-cohort models for ADC and SCC. Fig. 3a and b illustrated the detailed analyses of the age-period and age-cohort models on the incidence rate of each histological subtype of lung cancer. With the period 1991–1995 as reference, age-adjusted relative risk (RR) of SQCC decreased dramatically from period 1991–1995 to 2001–2005 by 31%. The drop was more rapid for LCC (79%) and UTT (58%), whereas the period effects increased tremendously for OSC. On the other hand, no period effect was observed for ADC and SCC throughout the entire period. With the birth cohort 1941–1945 as reference, RRs of consecutive birth cohorts for SQCC reduced noticeably till the birth cohort 1961–1965, leveled off for birth cohort 1966–1970, but continued decreasing in the most recent generation. We did not observe a clear birth cohort effect for ADC over the study period except for the last generation with a slight decrease. Effects of birth cohort for
L.A. Tse et al. / Lung Cancer 66 (2009) 22–27
Fig. 3. (a) Cohort and period effects obtained from age-period and age-cohort analyses for squamous cell carcinoma, adenocarcinoma, and small cell carcinoma of the lung. (b) Cohort and period effects obtained from age-period and age-cohort analyses for large cell carcinoma, other specified or unspecified carcinoma, and lung cancer without histological verification.
SCC fluctuated but generally decreased over the study period. We observed an enormous decreased RR for LCC and UTT, whereas there appeared a vastly increased birth cohort effect for OSC (Fig. 3b). 4. Discussion A downward trend in age-standardized incidence rate of lung cancer was observed for Hong Kong males during the period 1991–2005. The decline in overall lung cancer incidence was limited primarily to the decrease in squamous and large cell carcinoma. We observed a relatively horizontal time trend for ADC and SCC incidence, and ADC has already supplanted SQCC as the predominant type of lung cancer among Hong Kong male population. These changes parallel those seen in many industrialized countries [3,18–21].
25
Advances in modern imaging technologies and flexible endoscopes over time would be reflected by decreasing numbers of lung cancer cases without histological confirmation (UTT), and this would account for the significantly declining period effect on the time trend of UTT. The proportion of UTT has indeed dropped substantially by an annual reduction rate of 9.25% over the entire study period. We simply assumed non-differential classification of some of these UTT cases to ADC or other subtypes under study since no documentation shows that the pathologists have special preferences in making classification. The remarkably increasing period effect in the more recent periods for OSC was probably the result of the potential misclassification with LCC, as a significantly decreasing period effect was observed for LCC. On the other hand, given the relatively horizontal time trend for ADC, the decreasing trend seen in SQCC was probably genuine, as the actual decrease should be more marked if some UTT cases in the distant past were classified as SQCC. Being a full member of the International Association of Cancer Registries under the WHO, there should be little doubt on the quality and comparability of cancer data between Hong Kong Cancer Registry and other members [22]. Strong birth cohort effects for SQCC but minimum for ADC and SCC among Hong Kong males provide materials for etiology hypotheses. The age-cohort model fitted well for SQCC, suggesting that changing environmental exposures in early life in successive cohorts might be the most likely determinants for this subtype. The progressively protective cohort effects in more recent generations could be explained by the gradual reduction of exposures to environmental risk factors. Tobacco smoking is associated with all major types of lung cancer but the link is stronger with SQCC and SCC [4,23,24]. A rapid and steady decrease in daily smoking prevalence together with switching to low-tar and filter cigarettes decades ago is believed to contribute largely to the declined rates of SQCC in 1990s [5,25,26]. Evidence has shown that filter tip can effectively reduce deposition of larger particles in the central airway resulting in a reduced risk of SQCC but increase in receiving smallsize particles in the deeper parts of the lung where adenocarcinoma preferentially occurs [25,27]. Moreover, the use of low-tar cigarettes may lead smokers to taking larger puffs and/or inhaling more deeply to satisfy their craving for nicotine thus further increases the occurrence of ADC [28,29]. Results from American Cancer Society’s Cancer Prevention Study provided supportive evidence on changes in smoking behavior and cigarette design contributing to the rise in the occurrence of ADC over 20 years of follow-up [30]. However, with the notable decreasing trend of smoking prevalence and the strong association between smoking and SCC, it would be difficult to explain the relatively stable SCC incidence trend observed. Other factors related to the increased risk of male lung cancer include radon, environmental tobacco smoke, occupational agents (e.g., asbestos and silica dusts), cooking and oil vapors or smoky coal, virus, and dietary factors [31–37]. Few epidemiological studies had been conducted to link specific occupational and environmental agents with histological subtypes of lung cancer among male population, largely due to the difficulties in collecting nonsmoking lung cancer cases. Studies reported that workers exposed to chloromethyl ethers and underground miners exposed to radon progeny had excess risk of SCC [4], while crocidolite asbestos exposure was linked with increased risks of ADC and SQCC [38–40]. With more stringent legislations and standards in public and workplace and massive health promotions at the population level, we believe that these risk factors possibly playing important roles in lung cancer etiology in the past are expected to contribute less to Hong Kong male population over time. The lack of birth cohort effects for ADC in Hong Kong males suggests that this histological type may not be related to environmental exposures or that environmental exposures associated with it have remained stable through generations. Studies showed that
26
L.A. Tse et al. / Lung Cancer 66 (2009) 22–27
long-term exposure to some components of polluted air, especially nitrogen oxides (NOx ), might play the major role in the increasing trend of ADC incidence in US [28]. NOx can be generated from a combustion process of burning fuels at high temperature and the major sources of local emission are from motor vehicles and electric utilities. Report from Hong Kong Environmental Protection Department showed that the annual average of NOx in urban areas has remained quite constant over the last decade [41]. The ecological relationship between ADC incidence and NOx level among Hong Kong males was consistent with that was reported in US [28]. On the other hand, genetic risk factors shifting across birth cohorts might influence the trends of lung cancer incidence. However, there have been no substantial changes in the place of origin (ethnic origin) among Hong Kong residents in recent decades [42], implying a stable genetic susceptibility in Hong Kong population. Genetic susceptibility may play an important role in modifying the effect of cigarette smoking (or other environmental risk factors) on the risk of ADC, as can be seen from the markedly lower incidence rates of Asian immigrants to Hawaii than those of other ethic origins [43,44], and gene and environment interactions were thus hypothesized as the likely explanation in the causation of ADC. However, it is difficult to see why changing environmental factors could not lead to a change in ADC incidence trend among Hong Kong males. It is possible that traditional environmental risk factors for ADC have decreased over time, which should lead to a decreasing trend in the incidence, but this could have been counteracted by more smokers using filter tips—expected to increase risk of ADC and reduce risk of SQCC. A hospital-based case-control study among German revealed a significant effect of the XPA polymorphism in heavy smokers and occupationally exposed individuals, suggesting an important gene-environment interaction in the causation of lung ADC [45]. Although the underlying mechanisms between specific genes and the environmental carcinogens for the development of lung ADC remain unclear, the horizontal time trend of ADC incidence and the lack of birth cohort effects in the Hong Kong male population suggest gene-environment interactions might have played an important role in the etiology of lung ADC. Our previous study showed an overall 2.5–2.8 times the risk of age-standardized incidence rate of lung cancer for Hong Kong males when compared with the female population [11]. A significant annual increase in the incidence rate of lung cancer (4.15%) during 1976–1990 was observed for Hong Kong females but the incidence significantly decreased by 2.30% annually during 1991–2000. The decrease in the 1990s occurred in all age groups, with the largest annual decrease of 8.26% in the 35–39 years old group [46]. Similar to those of the Hong Kong males, female ADC has been the most frequent histological subtype of lung cancer since 1990, and an overall decline in the incidence rate of SQCC, SCC, and LCC also occurred in females. ADC incidence rate among Hong Kong females increased significantly by 5.9% annually during 1983–1988, but followed by an insignificant annual decline of 0.9%. The later was somewhat inconsistent with that of the Hong Kong males (with a slight annual increase of 0.31%) [11]. The difference in the time trend of ADC incidence in the two genders could have resulted from differential prevalence of gender-specific polymorphisms; however, little has been known about the role of sex chromosomes in the etiology of lung cancer. Our recent research showed that cumulative exposure to household cooking fumes generated from any form of frying was strongly associated with an increased risk of ADC of the lung in females [47]. In Hong Kong, many families have employed domestic helpers from Southeast Asian countries since early 1980s and more local Chinese women (especially for the young and middle aged working population) have hence been removed from household cooking fumes exposure. This might provide an explanation for the small observed decrease in the trend of ADC incidence among local female population in the recent years.
In conclusion, the overall decline in male lung cancer incidence rate was limited primarily to a decline in the rate of SQCC, which could mainly be explained by the decreasing trend of cigarette smoking in the population. ADC had been the most frequent histological subtype all along since 1990. The relatively horizontal trend of incidence rate of ADC and the lack of birth cohort effect implied the important roles of gene-environment interaction and/or the use of low-tar and filter tip cigarettes. Our study suggests that different histological subtypes may represent different disease entities with perhaps some shared and/or some distinct risk factors. The hypotheses generated from this ecological study will need confirmation by subsequent analytic studies. Conflict of interest statement The author(s) indicated no potential conflicts of interest. References [1] Doll SR. Introduction and overview. Epidemiology of lung cancer. New York/Basel/Hong Kong: Marcel Dekker, Inc.; 1994. p. 1–14. [2] Hong Kong Cancer Registry HA. Hong Kong cancer stat. Hong Kong Cancer Registry; 2004 http://www3.ha.org.hk/cancereg/eng/canstat2004.pdf [Accessed on May 31, 2008]. [3] Harkness EF, Brewster DH, Kerr KM, Fergusson RJ, MacFarlane GJ. Changing trends in incidence of lung cancer by histologic type in Scotland. Int J Cancer 2002;102:179–83. [4] Alberg AJ, Brock MV, Samet JM. Epidemiology of lung cancer: looking to the future. J Clin Oncol 2005;23:3175–85. [5] Zheng TZ, Holford TR, Boyle P, Chen YT, Ward BA, Flannery J, et al. Time trend and the age-period-cohort effect on the incidence of histologic types of lung cancer in Connecticut, 1960–1989. Cancer 1994;74:1556–67. [6] International Agency of Research on Cancer. IARC Monographs on the evaluation of carcinogenic risks to humans, vol. 83, tobacco smoke and involuntary smoking Lyon, France: IARC Press; 2004. [7] Malarcher AM, Schulman J, Epstein LA, Thun MJ, Mowery P, Pierce B, et al. Methodological issues in estimating smoking-attributable mortality in the United States. Am J Epidemiol 2000;152:573–84. [8] Liu ZY. Smoking and lung cancer in China: combined analysis of eight casecontrol studies. Int J Epidemiol 1992;21:197–201. [9] Yang GH, Fan L, Tan L, Qi GM, Zhang Y, Samet JM, et al. Smoking in China, findings of the 1996 National Prevalence Survey. JAMA 1999;282:1247–53. [10] Charloux A, Quoix E, Wolkove N, Small D, Pauli G, Kreisman H. The increasing incidence of lung adenocarcinoma: reality or artefact? A review of the epidemiology of lung adenocarcinoma. Int J Epidemiol 1997;26:14–23. [11] Au JS, Mang OW, Foo W, Law SC. Time trends of lung cancer incidence by histologic types and smoking prevalence in Hong Kong 1983–2000. Lung Cancer 2004;45:143–52. [12] World Health Organization. The World Health Organization histological typing of lung tumours. Second edtion. Am J Clin Pathol 1982;77:123–36. [13] Tobacco Control Office Department of Health Hong Kong. Prevalence of male daily cigarette smokers by age. http://www.tco.gov.hk/english/ infostation/infostation sta 01.html#a4 [Accessed on August 5, 2008]. [14] Millar S. The biosocial survey in Hong Kong. UNESCO/UNEP 1979:359–76. [15] SAS Institute Inc. SAS/STAT user’s guide. Release 6. 03. Edition. Carry, NC: SAS Institute Inc.; 1988. [16] Clayton D, Schifflers E. Models for temporal variation in cancer rates. II: ageperiod-cohort models. Stat Med 1987;6:469–81. [17] Clayton D, Schifflers E. Models for temporal variation in cancer rates. I: ageperiod and age-cohort models. Stat Med 1987;6:449–67. [18] Borras J, Borras JM, Galceran J, Sanchez V, Moreno V, Gonzalez JR. Trends in smoking-related cancer incidence in Tarragona, Spain, 1980–96. Cancer Causes Control 2001;12:903–8. [19] Li X, Mutanen P, Hemminki K. Gender-specific incidence trends in lung cancer by histological type in Sweden, 1958–1996. Eur J Cancer Prev 2001;10: 227–35. [20] Myrdal G, Lambe M, Bergstrom R, Ekbom A, Wagenius G, Stahle E. Trends in lung cancer incidence in Sweden with special reference to period and birth cohorts. Cancer Causes Control 2001;12:539–49. [21] Nguyen AM, Luke CG, Roder D. Time trends in lung cancer incidence by histology in South Australia: likely causes and public health implications. Aust NZ J Public Health 2003;27:596–601. [22] Lee AWM, Foo W, Mang O, Sze WM, Chappell R, Lau WH, et al. Changing epidemiology of nasopharyngeal carcinoma in Hong Kong over a 20-year period (1980–99): an encouraging reduction in both incidence and mortality. Int J Cancer 2003;103:680–5. [23] Khuder SA, Mutgi AB. Effect of smoking cessation on major histologic types of lung cancer. Chest 2001;120:1577–83. [24] Lubin JH, Caporaso NE. Cigarette smoking and lung cancer: modeling total exposure and intensity. Cancer Epidemiol Biomarkers Prev 2006;15:517–23.
L.A. Tse et al. / Lung Cancer 66 (2009) 22–27 [25] Blizzard L, Dwyer T. Lung cancer incidence in Australia: impact of filter-tip cigarettes with unchanged tar yields. Int J Cancer 2002;97:679–84. [26] Joshua AM, Boyer MJ, Subramanian R, Clarke SJ. Smoking reduction does work: resulting alterations in the incidence and histological subtypes of lung cancer in New South Wales in the last 20 years. Respirology 2005;10:233–8. [27] Stellman SD, Muscat JE, Thompson S, Hoffmann D, Wynder EL. Risk of squamous cell carcinoma and adenocarcinoma of the lung in relation to lifetime filter cigarette smoking. Cancer 1997;80:382–8. [28] Chen F, Cole P, Bina WF. Time trend and geographic patterns of lung adenocarcinoma in the United States, 1973–2002. Cancer Epidemiol Biomarkers Prev 2007;16:2724–9. [29] Wynder EL, Kabat GC. The effect of low-yield cigarette smoking on lung cancer risk. Cancer 1988;62:1223–30. [30] Thun MJ, Lally CA, Flannery JT, Calle EE, Flanders WD, Heath CW. Cigarette smoking and changes in the histopathology of lung cancer. J Natl Cancer Inst 1997;89:1580–6. [31] Lissowska J, Bardin-Mikolajczak A, Fletcher T, Zaridze D, Szeszenia-Dabrowska N, Rudnai P, et al. Lung cancer and indoor pollution from heating and cooking with solid fuels: the IARC international multicentre case-control study in Eastern/Central Europe and the United Kingdom. Am J Epidemiol 2005;162:326–33. [32] Boffetta P, Ahrens W, Nyberg F, Mukeria A, Bruske-Hohlfeld I, Fortes C. Exposure to environmental tobacco smoke and risk of adenocarcinoma of the lung. Int J Cancer 1999;83:635–9. [33] Boffetta P, Agudo A, Ahrens W, Benhamou E, Benhamou S, Darby SC, et al. Multicenter case-control study of exposure to environmental tobacco smoke and lung cancer in Europe. J Natl Cancer Inst 1998;90:1440–50. [34] Steenland K, Stayner L. Silica, asbestos, man-made mineral fibers, and cancer. Cancer Causes Control 1997;8:491–503. [35] Berry G, Newhouse ML, Turok M. Combined effect of asbestos exposure and smoking on mortality from lung cancer in factory workers. Lancet 1972;2:476–9. [36] Steenland K, Mannetje A, Boffetta P, Checkoway H, DeKlerk N, Koskela R. Pooled exposure-response analyses and risk assessment for lung cancer in 10 cohorts
[37]
[38]
[39] [40]
[41]
[42]
[43]
[44]
[45]
[46] [47]
27
of silica-exposed workers: an IARC multicentre study. Cancer Causes Control 2001;12:773–84. Chen W, Hnizdo E, Chen JQ, Chen W, Hnizdo E, Lu J, et al. Risk of silicosis in cohorts of Chinese tin and tungsten miners, and pottery workers (I): an epidemiological study. Am J Ind Med 2005;48:1–9. Lee BW, Wain JC, Kelsey KT, Wiencke JK, Christiani DC. Association of cigarette smoking and asbestos exposure with location and histology of lung cancer. Am J Respir Crit Care Med 1998;157(3 Pt 1):748–55. Raffn E, Villadsen E, Engholm G, Lynge E. Lung cancer in asbestos cement workers in Denmark. Occup Environ Med 1996;53:399–402. de Klerk NH, Musk AW, Eccles JL, Hansen J, Hobbs MS. Exposure to crocidolite and the incidence of different histological types of lung cancer. Occup Environ Med 1996;53:157–9. Environmental Protection Department. Air Quality in Hong Kong; 2003 http://www.epd.gov.hk/epd/english/environmentinhk/air/air quality/files/ aqr03e.pdf [Accessed on August 7, 2008]. Census and Statistics Department. Population Aged 5 and Over by Usual Language, 1991, 1996 and 2001 http://www.info.gov.hk/censtatd/eng/hkstat/ fas/01c/cd0062001 index.html [Accessed January 5, 2005]. Le Marchand L, Wilkens LR, Kolonel LN. Ethnic differences in the lung cancer risk associated with smoking. Cancer Epidemiol Biomarkers Prev 1992;1: 103–7. Haiman CA, Stram DO, Wilkens LR, Pike MC, Kolonel LN, Henderson BE, et al. Ethnic and racial differences in the smoking-related risk of lung cancer. N Engl J Med 2006;354:333–42. Butkiewicz D, Popanda O, Risch A, Edler L, Dienemann H, Schulz V, et al. Association between the risk for lung adenocarcinoma and a (-4) G-to-A polymorphism in the XPA gene. Cancer Epidemiol Biomarkers Prev 2004;13:2242–6. Chiu YL, Yu IT, Wong TW. Time trends of female lung cancer in Hong Kong: age, period and birth cohort analysis. Int J Cancer 2004;111:424–30. Yu IT, Chiu YL, Au JS, Wong TW, Tang JL. Dose-response relationship between cooking fumes exposures and lung cancer among Chinese nonsmoking women. Cancer Res 2006;66:4961–7.
Contents lists available at ScienceDirect
Lung Cancer journal homepage: www.elsevier.com/locate/lungcan
Cigarette smoking and changing trends of lung cancer incidence by histological subtype among Chinese male population Lap Ah Tse a,b , Oscar Wai-Kong Mang c , Ignatius Tak-Sun Yu b,∗ , Fan Wu d , Joseph Siu-Kie Au e , Stephen Chun-Key Law c,e a
School of Public Health, The Chinese University of Hong Kong, Hong Kong SAR, China Department of Community & Family Medicine, The Chinese University of Hong Kong, Hong Kong SAR, China Hong Kong Cancer Registry, Hospital Authority, Hong Kong SAR, China d Shanghai Municipal Center for Disease Control & Prevention, China e Department of Clinical Oncology, Queen Elizabeth Hospital, Kowloon, Hong Kong SAR, China b c
a r t i c l e
i n f o
Article history: Received 1 October 2008 Received in revised form 10 December 2008 Accepted 21 December 2008 Keywords: Lung neoplasm Incidence Histology Smoking Time trends Age-period-cohort models
a b s t r a c t We analyzed the time trends of lung cancer by histological subtype in Hong Kong during 1991–2005, and examined how the time trends were influenced by the effects of birth cohort and calendar period of diagnosis. Cancer incidence data were obtained from Hong Kong Cancer Registry and population data from Census and Statistics Department. Age-standardized incidence rates were computed by the direct method using WHO 1966 standard population as reference. Period and cohort effects were assessed by using two separate Poisson regression models adjusting for age. From 1991 to 2005, the incidence rates in Hong Kong Chinese males decreased steadily. The decline in overall lung cancer incidence rates was limited primarily to the decrease in squamous cell carcinoma, which could be explained by the decreasing trend of cigarette smoking. Adenocarcinoma had been the most predominant histological subtype all along. The relatively horizontal trend of adenocarcinoma and the lack of cohort effect implied the important roles of gene-environment interaction and/or the use of low-tar and filter tip cigarettes. Our study suggests that different histological subtypes may represent different disease entities with perhaps some distinct risk factors. The hypotheses generated from this ecological study will need confirmation by subsequent analytic studies. © 2008 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Lung cancer, a rare malignancy at the beginning of the 20th century, has become a leading cause of death throughout the world [1]. In Hong Kong, lung cancer incidence rates among males started to decrease since 1993 but lung cancer remains as the leading cancer [2]. This trend mirrors that of many Western countries a decade ago [3–5]. Numerous epidemiologic studies have been done in response to the global ‘epidemic’ of lung cancer and convincingly estab-
Abbreviations: HKCR, Hong Kong Cancer Registry; WHO, World Health Organization; ADC, adenocarcinoma; SQCC, squamous cell carcinoma; SCC, small cell carcinoma; LCC, large cell carcinoma; OSC, other specified and/or unspecified histological subtypes; UTT, unknown tumour type; APC, annual percentage change; RR, relative risk; CI, confidence interval. ∗ Corresponding author at: Department of Community & Family Medicine, The Chinese University of Hong Kong, 4/F School of Public Health, Prince of Wales Hospital, Shatin, N.T., Hong Kong SAR, China. Tel.: +852 2252 8773; fax: +852 2606 3500. E-mail address: [email protected] (I.T.-S. Yu). 0169-5002/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.lungcan.2008.12.023
lished cigarette smoking as the main cause of lung cancer [6]. Ethnic disparities seem to exist for lung cancer etiology. In Western populations, more than 80% lung cancer risks among the males could be attributable to cigarette smoking [7], whereas the figure was much lower (58%) for Chinese males [8,9]. Other risk factors, including occupational carcinogens (e.g., crystalline silica dust, asbestos), residential radon, passive smoking, ambient air pollution, as well as other unknown risk factors, might have contributed to the other 42% of lung cancer risk among Hong Kong male population. Notable changes of histological subtypes of lung cancer have also occurred over time [4,10]. Squamous cell carcinoma was the most frequent histological subtype of lung cancer in the initial decades of the smoking-induced epidemic of lung cancer but it has gradually been replaced by adenocarcinoma since mid-1970s [4]. Changes in histological subtypes of lung cancer might be indicative of changes in etiology. The aim of this study was to extend a previous analysis describing the epidemiology of lung cancer incidence in Hong Kong [11] by examining how the time trends of male lung cancer incidence rates of different histological subtypes were influenced by the effects of calendar period of diagnosis and cohort year of birth that were not studied previously. Effects of cigarette smoking
Deviance/DF Deviance Deviance/DF
7065.8 220.8 1841.2 83.7 19.1 (p < 0.001) 0.95 184.7 (p < 0.001) 18.5
Deviance/DF Deviance
5481.9 171.3 1070.9 48.7 53.5 (p < 0.001) 2.7 264.1 (p < 0.001) 26.4
Deviance/DF Deviance
5785.6 180.8 24.1 1.1 19.5 (p < 0.001) 1.0 10.4 (p < 0.001) 1.0
Deviance/DF Deviance
15748.9 492.2 29.8 1.4 29.2 (p < 0.001) 1.5 15.5 (p < 0.001) 1.6
Deviance/DF Deviance
16055.3 501.7 231.5 10.5 54.1 (p < 0.001) 2.7 12.3 (p < 0.001) 1.2
p values in parenthesis are for model improvements with the addition of the parameters compared with the age alone models. † DF, degree of freedom.
We used the direct method to calculate age-standardized annual incidence rate for each histological subtype of lung cancer using the WHO standard population (1966) as reference. Annual percentage change (APC) in incidence rate was calculated by using nonlinear regression under the assumption of same rate of change throughout the study period.
Deviance
2.3. Statistical analyses
Adenocarcinoma
Prevalence of daily smoking in different age groups among Hong Kong male population during 1982–2005 was obtained from the General Household Surveys of Hong Kong Census and Statistics Department [13], while the 1974 data were from an ad hoc survey carried out by a research project [14]. Daily smokers were defined as those persons aged 15 and over whom at the time of enumeration had a daily cigarette smoking habit (although they might not smoke on certain days because of illness or other reasons) [13].
Squamous cell carcinoma
2.2. Source of prevalence data on daily smoking
Variables in DF† the model
Lung cancer incidence data during the period 1991–2005 were obtained from Hong Kong Cancer Registry (HKCR) who has continuously collected cancer information by checking the computerised databases and/or medical records with all public and most private hospitals. With the maximum coverage, the incidence data taken from HKCR is considered of good representation of the territory [11]. Lung cancer was classified into six histological subtypes according to the criteria of the World Health Organization (WHO) histological typing of lung tumours [12]; adenocarcinoma (ADC), squamous cell carcinoma (SQCC), small cell carcinoma (SCC), large cell carcinoma (LCC), and other specified and/or unspecified histological subtypes (OSC). Those without pathological verification were regarded as unknown tumour type (UTT). Only the registered male lung cancer cases were included in the present study. Population data during the corresponding period were obtained from Hong Kong Census and Statistics Department. Mid-year population data were employed in the calculation of the incidence rate.
Table 1 Results of Poisson regression models for lung cancer incidence by histological type in Hong Kong, 1991–2005.
2.1. Incidence and population data
Small cell carcinoma
Large cell carcinoma
2. Materials and methods
32 22 20 10
and the etiological implications on different histological subtypes of lung cancer were discussed.
Intercept Age Age + period Age + cohort
Others specified or unspecified
Fig. 1. Incidence of male lung cancer by histology among Hong Kong male population during 1991–2005.
22052.4 689.1 1621.0 73.7 306.3 (p < 0.001) 15.3 81.2 (p < 0.001) 8.1
23
Unknown tumour type
L.A. Tse et al. / Lung Cancer 66 (2009) 22–27
24
L.A. Tse et al. / Lung Cancer 66 (2009) 22–27
Fig. 2. Prevalence of male daily cigarette smokers by age in Hong Kong, 1974–2005.
The multiple Poisson regression analyses were carried out to examine the period and birth cohort effects on histology-specific incidence rates of lung cancer using SAS procedure GENMOD after the adjustment of age [15]. Three 5-year periods (1991–1995, 1996–2000, and 2001–2005) were covered. Because of very few lung cancer cases diagnosed below 30, we only included 11 age groups by 5-year interval (ranging from 30–34 to 75–79, and 80 or above) in the multivariate analyses. Based on the period and age groups, we calculated 13 partially overlapping 5-year birth cohorts starting in “before 1915” and ending in 1971–1975. The reference categories were the 50–54 years age group, 1991–1995 calendar period, and 1941–1945 birth cohort. A model including all three time variables was not computed because the use of a full age-period-cohort model may involve serious methodological difficulties [16,17]. The goodness of fit of the models was assessed by the deviance. The closer the deviance to the degrees of freedom, the better the model fit [16,17]. 3. Results 3.1. Incidence rates by histological subtype A total of 38,449 new cases of male lung cancer were diagnosed over the study period, with 80.8% confirmed by histology in 2005 compared to 61.3% in 1991. Sixteen cases with unknown age were excluded from further analyses. The age-standardized incidence rates for all male lung cancers decreased steadily from 73.37 in 1991 to 51.33 per 100,000 in 2005 with an APC of −2.56 (95% confidence interval [CI]: −2.11, −3.02) (Fig. 1). The overall declined rate was mainly due to the significant decline of SQCC (APC = −3.59; 95%CI: −4.51, −2.66), LCC (APC = −8.58; 95%CI: −4.26, −12.70), and UTT (APC = −9.25; 95%CI: −7.34, −11.13). We observed a notable increase for OSC incidence (APC = 16.01; 95%CI: 8.50, 24.04), while the trend for ADC
(APC = 0.31; 95%CI: −0.61, 1.23) and SCC (APC = −0.58; 95%CI: −1.56, 0.40) remained relatively steady. ADC had been the most predominant histological subtype throughout the study period (36–42.8%). 3.2. Trend of daily smoking Prevalence of daily smoking among males decreased substantially during the period 1974–2005. Such decreasing trend was reflected in all age groups with the largest change for males aged above 60 years old (Fig. 2), while the trend was relatively flat for the younger groups. 3.3. Birth cohort and period effects adjusted for age Deviance statistics from Table 1 suggest that the age-cohort model fitted better for SQCC but the age-period model fitted better for LCC and OSC, as shown by the deviance of the better model fits closer to the degrees of freedom. The goodness of fits for age-period models differed slightly from those of age-cohort models for ADC and SCC. Fig. 3a and b illustrated the detailed analyses of the age-period and age-cohort models on the incidence rate of each histological subtype of lung cancer. With the period 1991–1995 as reference, age-adjusted relative risk (RR) of SQCC decreased dramatically from period 1991–1995 to 2001–2005 by 31%. The drop was more rapid for LCC (79%) and UTT (58%), whereas the period effects increased tremendously for OSC. On the other hand, no period effect was observed for ADC and SCC throughout the entire period. With the birth cohort 1941–1945 as reference, RRs of consecutive birth cohorts for SQCC reduced noticeably till the birth cohort 1961–1965, leveled off for birth cohort 1966–1970, but continued decreasing in the most recent generation. We did not observe a clear birth cohort effect for ADC over the study period except for the last generation with a slight decrease. Effects of birth cohort for
L.A. Tse et al. / Lung Cancer 66 (2009) 22–27
Fig. 3. (a) Cohort and period effects obtained from age-period and age-cohort analyses for squamous cell carcinoma, adenocarcinoma, and small cell carcinoma of the lung. (b) Cohort and period effects obtained from age-period and age-cohort analyses for large cell carcinoma, other specified or unspecified carcinoma, and lung cancer without histological verification.
SCC fluctuated but generally decreased over the study period. We observed an enormous decreased RR for LCC and UTT, whereas there appeared a vastly increased birth cohort effect for OSC (Fig. 3b). 4. Discussion A downward trend in age-standardized incidence rate of lung cancer was observed for Hong Kong males during the period 1991–2005. The decline in overall lung cancer incidence was limited primarily to the decrease in squamous and large cell carcinoma. We observed a relatively horizontal time trend for ADC and SCC incidence, and ADC has already supplanted SQCC as the predominant type of lung cancer among Hong Kong male population. These changes parallel those seen in many industrialized countries [3,18–21].
25
Advances in modern imaging technologies and flexible endoscopes over time would be reflected by decreasing numbers of lung cancer cases without histological confirmation (UTT), and this would account for the significantly declining period effect on the time trend of UTT. The proportion of UTT has indeed dropped substantially by an annual reduction rate of 9.25% over the entire study period. We simply assumed non-differential classification of some of these UTT cases to ADC or other subtypes under study since no documentation shows that the pathologists have special preferences in making classification. The remarkably increasing period effect in the more recent periods for OSC was probably the result of the potential misclassification with LCC, as a significantly decreasing period effect was observed for LCC. On the other hand, given the relatively horizontal time trend for ADC, the decreasing trend seen in SQCC was probably genuine, as the actual decrease should be more marked if some UTT cases in the distant past were classified as SQCC. Being a full member of the International Association of Cancer Registries under the WHO, there should be little doubt on the quality and comparability of cancer data between Hong Kong Cancer Registry and other members [22]. Strong birth cohort effects for SQCC but minimum for ADC and SCC among Hong Kong males provide materials for etiology hypotheses. The age-cohort model fitted well for SQCC, suggesting that changing environmental exposures in early life in successive cohorts might be the most likely determinants for this subtype. The progressively protective cohort effects in more recent generations could be explained by the gradual reduction of exposures to environmental risk factors. Tobacco smoking is associated with all major types of lung cancer but the link is stronger with SQCC and SCC [4,23,24]. A rapid and steady decrease in daily smoking prevalence together with switching to low-tar and filter cigarettes decades ago is believed to contribute largely to the declined rates of SQCC in 1990s [5,25,26]. Evidence has shown that filter tip can effectively reduce deposition of larger particles in the central airway resulting in a reduced risk of SQCC but increase in receiving smallsize particles in the deeper parts of the lung where adenocarcinoma preferentially occurs [25,27]. Moreover, the use of low-tar cigarettes may lead smokers to taking larger puffs and/or inhaling more deeply to satisfy their craving for nicotine thus further increases the occurrence of ADC [28,29]. Results from American Cancer Society’s Cancer Prevention Study provided supportive evidence on changes in smoking behavior and cigarette design contributing to the rise in the occurrence of ADC over 20 years of follow-up [30]. However, with the notable decreasing trend of smoking prevalence and the strong association between smoking and SCC, it would be difficult to explain the relatively stable SCC incidence trend observed. Other factors related to the increased risk of male lung cancer include radon, environmental tobacco smoke, occupational agents (e.g., asbestos and silica dusts), cooking and oil vapors or smoky coal, virus, and dietary factors [31–37]. Few epidemiological studies had been conducted to link specific occupational and environmental agents with histological subtypes of lung cancer among male population, largely due to the difficulties in collecting nonsmoking lung cancer cases. Studies reported that workers exposed to chloromethyl ethers and underground miners exposed to radon progeny had excess risk of SCC [4], while crocidolite asbestos exposure was linked with increased risks of ADC and SQCC [38–40]. With more stringent legislations and standards in public and workplace and massive health promotions at the population level, we believe that these risk factors possibly playing important roles in lung cancer etiology in the past are expected to contribute less to Hong Kong male population over time. The lack of birth cohort effects for ADC in Hong Kong males suggests that this histological type may not be related to environmental exposures or that environmental exposures associated with it have remained stable through generations. Studies showed that
26
L.A. Tse et al. / Lung Cancer 66 (2009) 22–27
long-term exposure to some components of polluted air, especially nitrogen oxides (NOx ), might play the major role in the increasing trend of ADC incidence in US [28]. NOx can be generated from a combustion process of burning fuels at high temperature and the major sources of local emission are from motor vehicles and electric utilities. Report from Hong Kong Environmental Protection Department showed that the annual average of NOx in urban areas has remained quite constant over the last decade [41]. The ecological relationship between ADC incidence and NOx level among Hong Kong males was consistent with that was reported in US [28]. On the other hand, genetic risk factors shifting across birth cohorts might influence the trends of lung cancer incidence. However, there have been no substantial changes in the place of origin (ethnic origin) among Hong Kong residents in recent decades [42], implying a stable genetic susceptibility in Hong Kong population. Genetic susceptibility may play an important role in modifying the effect of cigarette smoking (or other environmental risk factors) on the risk of ADC, as can be seen from the markedly lower incidence rates of Asian immigrants to Hawaii than those of other ethic origins [43,44], and gene and environment interactions were thus hypothesized as the likely explanation in the causation of ADC. However, it is difficult to see why changing environmental factors could not lead to a change in ADC incidence trend among Hong Kong males. It is possible that traditional environmental risk factors for ADC have decreased over time, which should lead to a decreasing trend in the incidence, but this could have been counteracted by more smokers using filter tips—expected to increase risk of ADC and reduce risk of SQCC. A hospital-based case-control study among German revealed a significant effect of the XPA polymorphism in heavy smokers and occupationally exposed individuals, suggesting an important gene-environment interaction in the causation of lung ADC [45]. Although the underlying mechanisms between specific genes and the environmental carcinogens for the development of lung ADC remain unclear, the horizontal time trend of ADC incidence and the lack of birth cohort effects in the Hong Kong male population suggest gene-environment interactions might have played an important role in the etiology of lung ADC. Our previous study showed an overall 2.5–2.8 times the risk of age-standardized incidence rate of lung cancer for Hong Kong males when compared with the female population [11]. A significant annual increase in the incidence rate of lung cancer (4.15%) during 1976–1990 was observed for Hong Kong females but the incidence significantly decreased by 2.30% annually during 1991–2000. The decrease in the 1990s occurred in all age groups, with the largest annual decrease of 8.26% in the 35–39 years old group [46]. Similar to those of the Hong Kong males, female ADC has been the most frequent histological subtype of lung cancer since 1990, and an overall decline in the incidence rate of SQCC, SCC, and LCC also occurred in females. ADC incidence rate among Hong Kong females increased significantly by 5.9% annually during 1983–1988, but followed by an insignificant annual decline of 0.9%. The later was somewhat inconsistent with that of the Hong Kong males (with a slight annual increase of 0.31%) [11]. The difference in the time trend of ADC incidence in the two genders could have resulted from differential prevalence of gender-specific polymorphisms; however, little has been known about the role of sex chromosomes in the etiology of lung cancer. Our recent research showed that cumulative exposure to household cooking fumes generated from any form of frying was strongly associated with an increased risk of ADC of the lung in females [47]. In Hong Kong, many families have employed domestic helpers from Southeast Asian countries since early 1980s and more local Chinese women (especially for the young and middle aged working population) have hence been removed from household cooking fumes exposure. This might provide an explanation for the small observed decrease in the trend of ADC incidence among local female population in the recent years.
In conclusion, the overall decline in male lung cancer incidence rate was limited primarily to a decline in the rate of SQCC, which could mainly be explained by the decreasing trend of cigarette smoking in the population. ADC had been the most frequent histological subtype all along since 1990. The relatively horizontal trend of incidence rate of ADC and the lack of birth cohort effect implied the important roles of gene-environment interaction and/or the use of low-tar and filter tip cigarettes. Our study suggests that different histological subtypes may represent different disease entities with perhaps some shared and/or some distinct risk factors. The hypotheses generated from this ecological study will need confirmation by subsequent analytic studies. Conflict of interest statement The author(s) indicated no potential conflicts of interest. References [1] Doll SR. Introduction and overview. Epidemiology of lung cancer. New York/Basel/Hong Kong: Marcel Dekker, Inc.; 1994. p. 1–14. [2] Hong Kong Cancer Registry HA. Hong Kong cancer stat. Hong Kong Cancer Registry; 2004 http://www3.ha.org.hk/cancereg/eng/canstat2004.pdf [Accessed on May 31, 2008]. [3] Harkness EF, Brewster DH, Kerr KM, Fergusson RJ, MacFarlane GJ. Changing trends in incidence of lung cancer by histologic type in Scotland. Int J Cancer 2002;102:179–83. [4] Alberg AJ, Brock MV, Samet JM. Epidemiology of lung cancer: looking to the future. J Clin Oncol 2005;23:3175–85. [5] Zheng TZ, Holford TR, Boyle P, Chen YT, Ward BA, Flannery J, et al. Time trend and the age-period-cohort effect on the incidence of histologic types of lung cancer in Connecticut, 1960–1989. Cancer 1994;74:1556–67. [6] International Agency of Research on Cancer. IARC Monographs on the evaluation of carcinogenic risks to humans, vol. 83, tobacco smoke and involuntary smoking Lyon, France: IARC Press; 2004. [7] Malarcher AM, Schulman J, Epstein LA, Thun MJ, Mowery P, Pierce B, et al. Methodological issues in estimating smoking-attributable mortality in the United States. Am J Epidemiol 2000;152:573–84. [8] Liu ZY. Smoking and lung cancer in China: combined analysis of eight casecontrol studies. Int J Epidemiol 1992;21:197–201. [9] Yang GH, Fan L, Tan L, Qi GM, Zhang Y, Samet JM, et al. Smoking in China, findings of the 1996 National Prevalence Survey. JAMA 1999;282:1247–53. [10] Charloux A, Quoix E, Wolkove N, Small D, Pauli G, Kreisman H. The increasing incidence of lung adenocarcinoma: reality or artefact? A review of the epidemiology of lung adenocarcinoma. Int J Epidemiol 1997;26:14–23. [11] Au JS, Mang OW, Foo W, Law SC. Time trends of lung cancer incidence by histologic types and smoking prevalence in Hong Kong 1983–2000. Lung Cancer 2004;45:143–52. [12] World Health Organization. The World Health Organization histological typing of lung tumours. Second edtion. Am J Clin Pathol 1982;77:123–36. [13] Tobacco Control Office Department of Health Hong Kong. Prevalence of male daily cigarette smokers by age. http://www.tco.gov.hk/english/ infostation/infostation sta 01.html#a4 [Accessed on August 5, 2008]. [14] Millar S. The biosocial survey in Hong Kong. UNESCO/UNEP 1979:359–76. [15] SAS Institute Inc. SAS/STAT user’s guide. Release 6. 03. Edition. Carry, NC: SAS Institute Inc.; 1988. [16] Clayton D, Schifflers E. Models for temporal variation in cancer rates. II: ageperiod-cohort models. Stat Med 1987;6:469–81. [17] Clayton D, Schifflers E. Models for temporal variation in cancer rates. I: ageperiod and age-cohort models. Stat Med 1987;6:449–67. [18] Borras J, Borras JM, Galceran J, Sanchez V, Moreno V, Gonzalez JR. Trends in smoking-related cancer incidence in Tarragona, Spain, 1980–96. Cancer Causes Control 2001;12:903–8. [19] Li X, Mutanen P, Hemminki K. Gender-specific incidence trends in lung cancer by histological type in Sweden, 1958–1996. Eur J Cancer Prev 2001;10: 227–35. [20] Myrdal G, Lambe M, Bergstrom R, Ekbom A, Wagenius G, Stahle E. Trends in lung cancer incidence in Sweden with special reference to period and birth cohorts. Cancer Causes Control 2001;12:539–49. [21] Nguyen AM, Luke CG, Roder D. Time trends in lung cancer incidence by histology in South Australia: likely causes and public health implications. Aust NZ J Public Health 2003;27:596–601. [22] Lee AWM, Foo W, Mang O, Sze WM, Chappell R, Lau WH, et al. Changing epidemiology of nasopharyngeal carcinoma in Hong Kong over a 20-year period (1980–99): an encouraging reduction in both incidence and mortality. Int J Cancer 2003;103:680–5. [23] Khuder SA, Mutgi AB. Effect of smoking cessation on major histologic types of lung cancer. Chest 2001;120:1577–83. [24] Lubin JH, Caporaso NE. Cigarette smoking and lung cancer: modeling total exposure and intensity. Cancer Epidemiol Biomarkers Prev 2006;15:517–23.
L.A. Tse et al. / Lung Cancer 66 (2009) 22–27 [25] Blizzard L, Dwyer T. Lung cancer incidence in Australia: impact of filter-tip cigarettes with unchanged tar yields. Int J Cancer 2002;97:679–84. [26] Joshua AM, Boyer MJ, Subramanian R, Clarke SJ. Smoking reduction does work: resulting alterations in the incidence and histological subtypes of lung cancer in New South Wales in the last 20 years. Respirology 2005;10:233–8. [27] Stellman SD, Muscat JE, Thompson S, Hoffmann D, Wynder EL. Risk of squamous cell carcinoma and adenocarcinoma of the lung in relation to lifetime filter cigarette smoking. Cancer 1997;80:382–8. [28] Chen F, Cole P, Bina WF. Time trend and geographic patterns of lung adenocarcinoma in the United States, 1973–2002. Cancer Epidemiol Biomarkers Prev 2007;16:2724–9. [29] Wynder EL, Kabat GC. The effect of low-yield cigarette smoking on lung cancer risk. Cancer 1988;62:1223–30. [30] Thun MJ, Lally CA, Flannery JT, Calle EE, Flanders WD, Heath CW. Cigarette smoking and changes in the histopathology of lung cancer. J Natl Cancer Inst 1997;89:1580–6. [31] Lissowska J, Bardin-Mikolajczak A, Fletcher T, Zaridze D, Szeszenia-Dabrowska N, Rudnai P, et al. Lung cancer and indoor pollution from heating and cooking with solid fuels: the IARC international multicentre case-control study in Eastern/Central Europe and the United Kingdom. Am J Epidemiol 2005;162:326–33. [32] Boffetta P, Ahrens W, Nyberg F, Mukeria A, Bruske-Hohlfeld I, Fortes C. Exposure to environmental tobacco smoke and risk of adenocarcinoma of the lung. Int J Cancer 1999;83:635–9. [33] Boffetta P, Agudo A, Ahrens W, Benhamou E, Benhamou S, Darby SC, et al. Multicenter case-control study of exposure to environmental tobacco smoke and lung cancer in Europe. J Natl Cancer Inst 1998;90:1440–50. [34] Steenland K, Stayner L. Silica, asbestos, man-made mineral fibers, and cancer. Cancer Causes Control 1997;8:491–503. [35] Berry G, Newhouse ML, Turok M. Combined effect of asbestos exposure and smoking on mortality from lung cancer in factory workers. Lancet 1972;2:476–9. [36] Steenland K, Mannetje A, Boffetta P, Checkoway H, DeKlerk N, Koskela R. Pooled exposure-response analyses and risk assessment for lung cancer in 10 cohorts
[37]
[38]
[39] [40]
[41]
[42]
[43]
[44]
[45]
[46] [47]
27
of silica-exposed workers: an IARC multicentre study. Cancer Causes Control 2001;12:773–84. Chen W, Hnizdo E, Chen JQ, Chen W, Hnizdo E, Lu J, et al. Risk of silicosis in cohorts of Chinese tin and tungsten miners, and pottery workers (I): an epidemiological study. Am J Ind Med 2005;48:1–9. Lee BW, Wain JC, Kelsey KT, Wiencke JK, Christiani DC. Association of cigarette smoking and asbestos exposure with location and histology of lung cancer. Am J Respir Crit Care Med 1998;157(3 Pt 1):748–55. Raffn E, Villadsen E, Engholm G, Lynge E. Lung cancer in asbestos cement workers in Denmark. Occup Environ Med 1996;53:399–402. de Klerk NH, Musk AW, Eccles JL, Hansen J, Hobbs MS. Exposure to crocidolite and the incidence of different histological types of lung cancer. Occup Environ Med 1996;53:157–9. Environmental Protection Department. Air Quality in Hong Kong; 2003 http://www.epd.gov.hk/epd/english/environmentinhk/air/air quality/files/ aqr03e.pdf [Accessed on August 7, 2008]. Census and Statistics Department. Population Aged 5 and Over by Usual Language, 1991, 1996 and 2001 http://www.info.gov.hk/censtatd/eng/hkstat/ fas/01c/cd0062001 index.html [Accessed January 5, 2005]. Le Marchand L, Wilkens LR, Kolonel LN. Ethnic differences in the lung cancer risk associated with smoking. Cancer Epidemiol Biomarkers Prev 1992;1: 103–7. Haiman CA, Stram DO, Wilkens LR, Pike MC, Kolonel LN, Henderson BE, et al. Ethnic and racial differences in the smoking-related risk of lung cancer. N Engl J Med 2006;354:333–42. Butkiewicz D, Popanda O, Risch A, Edler L, Dienemann H, Schulz V, et al. Association between the risk for lung adenocarcinoma and a (-4) G-to-A polymorphism in the XPA gene. Cancer Epidemiol Biomarkers Prev 2004;13:2242–6. Chiu YL, Yu IT, Wong TW. Time trends of female lung cancer in Hong Kong: age, period and birth cohort analysis. Int J Cancer 2004;111:424–30. Yu IT, Chiu YL, Au JS, Wong TW, Tang JL. Dose-response relationship between cooking fumes exposures and lung cancer among Chinese nonsmoking women. Cancer Res 2006;66:4961–7.