Lung Cancer Epidemiology in Mainland China

Incidence, prevalence and trends

Lung cancer incidence has increased rapidly in China over the last 20 years. In 2000, Lin et al.^[1] examined the world wide lung cancer incidence using data from the global network of cancer registries. The incidence of lung cancer has increased over recent decades, especially in females and lung cancer was the most common female cancer in Tianjin China. The world age-adjusted incidence rates (ASR) in male and female were 56.1/100000 and 18.2/100000 in Shanghai, 55.9/ 100000 and 37.0/100000 in Tianjin, 35.0/100000 and 11.0/100000 in Qidong (a small city near Shanghai). In most areas of the world, the incidence of lung cancer has increased or stayed relatively stable in both sexes. See Tables 1 and 2.^[2]

Chinese specificity compared to other countries

Among the 183 registered worldwide populations, lung cancer incidence in males was ranked as the 73rd, 74th, 127th and 23rd respec-tively for Shanghai, Tianjin, Qidong and Hong Kong, and in females the 52nd, 13th, 102nd and 23rd. The sex ratio (M/F) ranged from 1.5 to 3.5 for most areas. See Tables 1 and 2.

High rates of lung adenocarcinoma among women

The main pathological patterns were squamous carcinoma, adenocarcinoma and small cell carcinoma. Due to a lack of population-based data of cell types from Shanghai and Qidong, only data from Tianjin were compared to that from Hong Kong and the United States. In China, because fewer lung cancer patients were diagnosed/confirmed by pathology, unknown cell types make up a large proportion in Tianjin and may cause potential bias. The ratio of squamous/adenocarcinoma was 2.01 in males and 0.67 in females in Tianjin, 0.97 and 0.28 in Hong Kong, 1.00 and 0.61 in the US white population and 1.18 and 0.49 in US blacks. A very high proportion of adenocarcinoma was observed in Hong Kong females. In comparison with adenocarcinoma, a two fold squamous cell lung cancer incidence was observed in Tianjin males. See Table 3.^[1]

Regional differences in the Northern provinces and in industrial cities

According to current available data, higher incidences of lung cancer were observed at similar rates in large cities like Shanghai, Tianjin and Beijing. In the large cities, higher smoking rates plus more severe indoor and outdoor air pollution could be the main reasons related to the higher lung cancer incidence/mortality rate. In Northern China, air pollution caused by burning coal indoors and/or outdoors during the long heating season may play an important role in lung cancer etiology. Indoor cooking with little ventilation in the Northern China during the winter results in indoor air pollution worse than Southern China. Coal consumption in Northern China is much higher compared to Southern China. See Table 4.^[3]

A case-control study in four metropolitan areas

Yu et al. ^[4] published a multi-center case-control study of four metropolitan areas, Beijing, Tianjin, Shanghai and Chongqing, the four largest cities in China. This study aimed to investigate the risk factors for lung cancer and to determine the essential variables that could be used for setting up a national risk surveillance system. The paper used the data collected from inperson interviews based on the National Research on Malignant Tumor Morbidity, Mortality and Surveillance Methods in 1996. The collaborating group carried out a 1:1 matched case-control study based on a population involving 1,312 cases and 1,312 neighborhood controls aged 35-74 from Beijing, Tianjin, Shanghai and Chongqing urban areas. The associations between risk factors and lung cancer were analyzed by conditional univariate and multivariate logistic regression. Cigarette smoking, a lower body-mass index (BMI), a longtime depressed mood, lower capability of emotion control, previous respiratory diseases and family history of cancer in first-degree relatives significantly increased the risk of lung cancer after adjusting for age, education and family average income. The odds ratios of these risk factors were 3.46 (95%CI 2.704.45), 1.18(95% Cl 1.02~1.40), 2.64(95% Cl 1.953.57), 2.71(95% Cl 2.02-3.65), 2.28 (95% Cl 1.812.88), 1.79 (95% Cl 1.38-2.32), respectively for the four cities. The odds ratio associated with the highest quartile of fresh vegetable consumption compared with the lowest was 0.75 (95% Cl 0.65-0.87), showing an inverse relationship. Physical exercise more often had a significant inverse association between cases and controls, with an odds ratio of 0.81 (95% Cl 0.74-0.89). The authors concluded that cigarette smoking, psychological factors, previous respiratory diseases and a family history of cancer were the main risk factors for lung cancer in the four cities. They made up the important surveillance index for lung cancer. The relation between lung cancer and environmental tobacco smoke, dietary factors and the factors related with female lung cancer should be further studied with a special design to carefully and exactly define the degree of exposure.

Risk among non-smoking females: meta-analysis

Zhang et al.^[5] evaluated the risk factors for lung cancer in non-smoking Chinese women based on the da-ta from 7 case-control studies from 1990-1999 (Xuanwei 1990, 2 studies in Shanghai 1995, Harbin 1996, Shenyang 1996, Beijing 1997 and Guangzhou 1997). The cumulative cases and controls were 1,115 and 1,520 respectively. The risk factors and those pooled ORs were family history of lung cancer (3 studies, case/control: 724/969, OR 2.79, 95% Cl 2.22-3.48), personal history of benign lung diseases (2 studies, case/control: 220/368, OR 2.87, 95% Cl 2.08-3.95), cooking oil fume indoor air pollution (2 studies, case/control: 670/767, OR 2.52, 95% Cl 2.12-2.91), coal soot indoor air pollution (4 studies, case/control: 846/1091, OR 1.42, 95% Cl 1.30-1.55) and exposure to environmental tobacco smoke (4 studies, case/control: 489/797, OR 1.64, 95% Cl 0.84-3.22).

Difference of risk factors between smokers and non-smokers: Tianjin study

Wang et al.^ra investigated the risk factors for lung cancer in females in the Tianjin urban area by using a case-control design. A total of 264 female incident cases aged 30-70 and living in urban Tianjin for more than 3 years were included. According to the age frequency of cases at various ages, a random sample of 250 residents for a control was drawn from the same area. An odds ratio calculation and unconditional logistic regression was used to evaluate the cancer risks. As an approach, smoking was used as a grouping variable in the data analysis to explore the difference between smokers and non-smokers. Factors including smaller BMI, smoking, cooking times, heavy cooking filmes and less vitamin C and E intake were independent risk factors for female lung cancer after adjusting for age. In the smoking group, they had smaller BMIs, heavy cooking fumes, less vitamin C and E; and in the non-smoking group, they had smaller BMIs, long cooking times, heavy cooking firmes and less vitamin E intake. The author concluded that the occurrence of female lung cancer was influenced by many risk factors. The risk factors of smokers and non-smokers were different. Vitamin C, as a protective factor may play a more important role in smokers compared to non-smoker. See Table 5.

Genetic predisposition/polymorphism and susceptibility to lung cancer

Li et al.^m investigated the genetic polymorphism of cytochrome P4502E1 (CYP2E1) as a susceptibility factor for lung cancer in Chinese population using the PCR-RFLP method. A total of 92 lung cancer cases and 137 frequency-matched normal controls were included in this study. The CYP2E1 cl/cl genotype was found in 72.8% of the lung cancer cases, which was significantly higher (P<0.01) than the controls (54.7%). By multivariate analysis, this genotype was found to be associated with a 2.5-fold increased risk of lung cancer (adjusted OR for sex, age and smoking: 2.5, 95% Cl 1.8-3.8), squamous cell carcinoma (adjusted OR: 3.5, 95%CI 2.3~5.8), adenocarcinoma (adjusted OR: 2.2, 95% CI 0.8-4.7) and other carcinoma cell types (adjusted OR: 0.9, 95% Cl 0.4-2.5). Stratified analysis suggested an interaction between the CYP2E1 cl/cl genotype and cigarette smoking. The odds ratio (OR) for the CYP2E1 cl/cl genotype, cigarette smoking and both factors combined were 3.9 (95% Cl 1.1-15.2), 4.1 (95% Cl 1.1-15.9) and 7.9 (95% Cl 2.4-29.4), respectively. The interaction between CYP2E1 cl/cl genotype and pack-years of cigarette smoking followed the same pattem, with a remarkably increased risk in the heaviest smokers. The authors concluded that the CYP2E1 cl/cl genotype is a susceptibility factor for development of lung cancer in Chinese, and there is an apparent gene and environmental interaction between the susceptible genotype and cigarette smoking.

Current molecular epidemiology of lung cancer

The research on lung cancer in the field of molecular epidemiology has focused on the following topics in Mainland China.^[8]

Viral infedion/HPV infection and lung cancer

In 1994, Xing et al.^[9] looked for human papilloma virus (HPV) in 49 squamous lung cancers using PCR and in situ hybridization methods. The positive rate was 7/49 (four HPV6+HPV11, one HPV6+HPV16 and two HPV6+HPV11+HPV16).

Zhang et al.^[10] reported that the positive rate of HPV DNA (HPV 16, HPV 18 and HPV1 l/6b) in 34 lung carcinoma cases was 4/34 (4/12 squamous cell, 0/10 ade-nocarcinoma, 0/3 small cell, 0/1 large cell, 0/5 alveolar cell and 0/3 adenosquamous.

Chen et al.^[11] studied 40 cases of primary lung cancer using in situ hybridization and PCR to detect HPV-DNA in different types of lung cancers. The positive rates were 22/40 in lung cancers (all types), 9/9 in small cell lung cancer, 8/16 in squamous cell carcinoma and 5/12 in adenocarcinoma. Amplification of the p53 exon 7 was seen in 5/22 HPV-DNA positive cases. Small cell and squamous carcinomas had a higher HPV infection rates than other types of lung cancers. It was concluded that an association existed between HPV infection and p53 mutations.

In 2000, Niyaz et al.^[2l] using PCR and immunohistochemistry studied that HPV 16-DNA/HPV18-DNA and p53 in 110 specimens of lung cancers. Telomerase activity in 22 specimens of fresh lung cancer tissues was examined using PCR-TRAP. The positive rate of HPV16-DNA and HPV 18-DNA in the lung cancers was higher than the normal tissues and inflammatory lesions (all P<0.01). The positive rates of HPV 16 and HPV 18 were 0/20 in normal tissue specimens, 1/20 in inflammatory lesions, 10/24 in small cell lung cancers and 34/86 in non-small cell lung cancers. The positive rate of p53 protein overexpression was 57% in the lung cancers and negative in the normal tissues (P<0.01). The differences of HPV16 and HPV18 infection and p53 overexpression were not statistically significant among different ethnic groups (all P>0.05). The telomerase activity was higher in the cancer tissues (86% ) compared with the adjacent normal tissues (22%, P<0.01). The author concluded that HPV 16 and HPV 18 infection might have an étiologie significance for lung cancer by enhancing p53 gene overexpression resulting in carcinogenesis, but each entity might have its own influence on the carcinogenic process. The high telomerase activity could be a marker of biomolecular change in lung cancers.

High background Radiation

Tao et al.^[13] from the Laboratory of Industrial Hygiene, Ministry of Public Health in 1999 reported on the association between cancer risk and exposure to the average annual effective dose of 6.4 mSv in the high background radiation area (HBRA) of Yangjiang, China. The cancer mortality data from 1979-1986 were collected from a dynamic cohort by a prospective survey. The data from 1987-1995 were obtained from a fixed cohort by a retrospective and/or prospective survey. The mortality investigation at the site consisted of two steps, i.e. the follow-up of the members in the cohort and the ascertainment of the death causes. The estimate of the cumulative individual radiation dose of the cohort members included that of the exposure from natural external and internal sources. Both direct (TLD measurement) and indirect (environmental measurement and occupancy pattem) approaches were used for the individual external radiation dose estimate. On the basis of the area-specific average annual external dose, the cohort members were classified into four groups for internal comparison: high, medial and low dose groups from the HBRA and a control group from the control area (CA). A relative risk (RR) and an excess relative risk coefficient and its 95% confidence interval^) were estimated using the AMFIT program in Epicure. They found that during 1979-1995, there were 10,415 total deaths and 1,003 cancer deaths among 1,698,350 person-years at risk in the cohort of 125,079 subjects. The adjusted RR (95% Cl) with sex and age group for all cancers of the whole HBRA was 0.99 (0.87-1.14). As for the site-specific cancer of the whole HBRA, the RRs of cancers of the stomach, colon, liver, lungs, bone, female breast, and thyroid were less than one, while the RRs of cancers of the nasopharynx, esophagus, rectum, pancreas, skin, cervix uterus, brain and central nervous system and lymphoid tissue were larger than one. However, all tissue sites except for esophageal cancer were not different statistically from one (P>0.05). The homogeneity tests of RRs for all cancers and for site-specific cancer among the three-dose groups in the HBRA revealed that the RRs in these dose groups were for all not different statistically (P>0.05). The ERR (95% CI)/Sv of all cancers for both sexes and all ages was -0.10 (-0.67, 0.69). The authors concluded that an increased cancer risk as well as lung cancer risk associated with the high levels of natural radiation in the HBRA was not found. On the contrary, the mortality from all cancers in the HBRA was generally lower than that in the CA, but not enough for a statistically significant difference.

Risk factors-family history

A family history is associated with higher risk of lung cancer“3,^[14] and it was reported ^[15] that the OR was 7.65 for female lung adenocarcinoma. The association between family history and lung cancer was summarized in a meta-analysis study (see related part in this review under the “Non-smoking female: meta-analysis").

Risk factors-tobacco consumption

Smoking as a key risk factor for lung cancer has been reported from studies in Guangzhou^[14],(OR=5.06, 95% Cl 2.77-9.24 in males, Tianjin ^[6](OR=4.02, 95% Cl 2.61-6.18 in females), Beijing (OR=4.47, 95% Cl 2.24-8.89).^[15] The association between passive smoking and lung cancer was summarized a meta-analysis study (see related part in this report under the “Non-smoking femalemeta-analysis").

Risk fadors-mentol health

Depression was shown to be an increased risk factor for lung cancer (OR=4.30, 95% Cl 1.53-12.11 in males).“^[4] Mental trauma due to a poor marriage, death of a relative and job stress associated with an elevated risk of lung cancer.^[15] Is it a confounder or causal factor? More studies should be conducted to clarify this question.

Risk factors-prior lung diseases

Previous benign lung diseases are associated with a higher risk for lung cancer as reported by a number of studies in Mainland China, including Guangzhou (OR=2.129, P=0.0005 in males).^[4-15] The association between prior lung disease and lung cancer was summarized in a meta-analysis study (see related part in this report under “Non-smoking female: meta-analysis").

Risk factors-coal soot indoor air pollution

It was reported that if coal had been used for fuel previously 10 years ago, it was a risk factor (OR=2.71, P=0.0001 in males).^[14] The association between coal soot indoor air pollution and lung cancer was summarized in a meta-analysis study (see related part in this report under the “Non-smoking female: meta-analysis").

Risk fodors-cooking fumes indoor air pollution

Zhou et al.^[16] reported that cooking fumes were associated with a higher risk for female lung adenocarcinoma (OR=4.53). The association between indoor air pollution from cooking fumes and lung cancer was summarized in a meta-analysis study (see related part in this report under “Non-smoking female: meta-analysis" and “Difference of risk factors between smokers and non-smokers: Tianjin Study").

Risk foctors-hormones

No direct and reliable data are available concerning the association between hormones and lung cancer. Yu et al.^[14] found an association between female lung cancer and a longer menstrual period (OR:Beijing= 1.08, Tianjin=0.76, Shanghai=2.00, P<0.05, Chongqing=0.94, pooled=1.15) and oral contraceptive pill usage (OR: Beijing=0.86, Tianjin=0.75, Shanghai= 1.80, Chongqing=1.38, pooled=1.12).

Risk fadors-diet

Vitamin C and E intake was associated with lung cancer as a protective factor and related to the level of smoking exposure (see related part in this report under “Difference of risk factors between smoker and non-smoker: Tianjin Study"). A diet of fresh vegetables showed a protective effect against lung cancer.

^[15,16] A study in miners (a high lung cancer risk mine in Yunan province) showed that fresh vegetables were a protective factor for lung cancer.^[17] The protective effect was of statistical significance after adjusting for age, number of years as a miner and smoking. B-Carotene has been shown to be a preventive factor for female lung adenocarcinoma.^[16]

Risk factors-occupational exposures

Occupations related to elevated risk for lung cancer were workers exposed to asphalt,^[18] miners exposed to radon and arsine,^[19] and crocidolite.1201 Many occupations related to chemicals, metals, rubber/plastic, working in a mine and exposure to other polluted air/environments or having a higher smoking prevalence showed higher risk for lung cancer.^[2l]

Risk factors-outdoor air pollution

Outdoor air pollution was associated with higher risk of lung cancer in Beijing.^[15] The association between lung cancer and outdoor air pollution has been reported in many cities. Although no one doubts that lung cancer is related to air pollution, one must remember that in most studies there are some limitations due to research methods.

Risk factors-sodoeconomic level/educcrtion

Different socioeconomic or education level may determine one’s life style and behavior. No research has shown an association between those factors and lung cancer.

Risk factors-alcohol consumption

Lu et al.^[22] showed there was an association between alcohol consumption and lung cancer in tin miners. A negative association was found in miners who drank small amounts of alcohol but there was a positive association with heavy drinkers. After adjusting for age, number of years employment and smoking, only heavy drinkers showed a statistically positive association.

Risk factors and their interactions (addition/synergy)

According to previous research regarding lung cancer, some interactions have been confirmed as follows.

1. Gene and smoking or carcinogens.

2. Preventive factors like vitamin C is more important for those exposed to smoking or carcinogens.

3. Smoking and occupational exposure.

4. Cancer is produced in multiple stages and by multiple factors, so that the etiology of lung cancer is based on the sum of unbalanced inter and outer environments.