Cancer Causes & Control

, Volume 21, Issue 2, pp 269–274

Increasing trend of the incidence of esophageal squamous cell carcinoma, but not adenocarcinoma, in Taiwan

Authors

    • Division of GastroenterologyTaipei Veterans General Hospital
    • Institute of Brain ScienceNational Yang-Ming University
    • School of MedicineNational Yang-Ming University
  • Hui-Chu Lang
    • Institute of Hospital and Health Care AdministrationNational Yang-Ming University
  • Jiing-Chyuan Luo
    • Division of GastroenterologyTaipei Veterans General Hospital
    • School of MedicineNational Yang-Ming University
  • Chun-Chu Liu
    • Department of MedicineKoo Foundation Sun Yat-Sen Cancer Center
    • School of MedicineNational Yang-Ming University
  • Han-Chieh Lin
    • Division of GastroenterologyTaipei Veterans General Hospital
    • School of MedicineNational Yang-Ming University
  • Full-Young Chang
    • Division of GastroenterologyTaipei Veterans General Hospital
    • School of MedicineNational Yang-Ming University
  • Shou-Dong Lee
    • Division of GastroenterologyTaipei Veterans General Hospital
    • School of MedicineNational Yang-Ming University
Original paper

DOI: 10.1007/s10552-009-9458-0

Cite this article as:
Lu, C., Lang, H., Luo, J. et al. Cancer Causes Control (2010) 21: 269. doi:10.1007/s10552-009-9458-0

Abstract

Epidemiologic data on esophageal cancer in Asia are extremely limited. We examined temporal trends in the incidence of esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma (EA) in Taiwan. Patients with esophageal cancer were identified from the Taiwan Cancer Registry between 1979 and 2003. Age-standardized incidences of ESCC and EA were calculated based on the national census and world standard population. Trends in incidence rates were estimated by calculating the annual percentage change (APC). The age-standardized incidence of ESCC increased progressively (from 1979–1983 to 1999–2003: 2.63 to 4.37 per 100,000-year), with an APC higher in male (3.27%, P < 0.0001) than that in female (1.23%, P = 0.03). Though the crude incidence of EA progressively increased in both gender (male: 0.28–0.49 per 100,000-year; female 0.07–0.11 per 100,000-year), the age-standardized incidence of EA is similar along the study period with an APC of 0.72% (P = 0.20) in male and 1.59% (P = 0.30) in female. In Taiwan, the incidence of ESCC significantly increased from 1979 to 2003, whereas the incidence of EA remained unchanged. Although EA incidence has not recently increased, it could in the future.

Keywords

Esophagus cancerSquamous cell carcinomaAdenocarcinomaTaiwan

Introduction

The incidence rates for esophageal adenocarcinoma (EA) have risen rapidly in the USA and Western Europe, especially among white males [1, 2]. A recent paper even indicated that the incidence of EA rose from 1975 through 2004 among not only white men (463% increment) but also white women (335% increment) in all stage and age groups [3]. The underlying mechanism responsible for this increment is unclear, but it is believed that the increased prevalence of gastroesophageal reflux disease (GERD) and obesity may play roles in the observed phenomenon [4, 5]. In contrast, the incidence of esophageal squamous cell carcinoma (ESCC) is declining in the USA, paralleling the downward trend in smoking prevalence from the 1960s to the 1990s [1]. It is reported that the incidence of EA outnumbered ESCC in USA after 1990 [1], and further subgroup analysis revealed that this trend can only be identified in the white or non-Hispanic men, but not all other races of both gender [6].

Population-based epidemiologic data on esophageal cancer in Asia are extremely limited. The ESCC has been reported to be decreasing in Singapore and Hong Kong [7, 8]. The EA incidence is increasing in Singapore [7], but decreasing in Hong Kong [8]. In Japan, the incidence of ESCC and EA seemed to be constant from 1993 to 2001 [9]. In Taiwan, we have demonstrated that the ratio of ESCC and EA remained stable over 15 years (1980–1995) in a single center [10]. However, that study was limited by the fact that the actual incidence cannot be calculated and that the result cannot be applied to the general population in Taiwan. In current study, by acquiring the data from a population-based Taiwan Cancer Registry, we aimed to examine the time trend change in the incidence of ESCC and EA in Taiwan, a country of mainly Chinese descent.

Materials and methods

Patient data on ESCC and EA from 1979 to 2003 were obtained from the Taiwan National Cancer Registry, which was founded in 1979 by the Department of Health of Taiwan. This cancer registry includes data from hospitals with ≥50 beds throughout the country. In Taiwan, the diagnosis of cancer is almost always established in hospitals. Solo practice is the most common form of practice pattern for the primary care physician (PCP) in Taiwan. Therefore, once cancer cases are suspected by the PCPs, they almost always refer their patients to hospital for tissue proof without delay. Thus, this Taiwan Cancer Registry is widely accepted as population-based data bank. The cancer registry data are evaluated for completeness and accuracy on a yearly basis [11]. The quality indicator of this data bank showed that the percentage of death certificate only cases is below 5% and the percentage of morphologically verified cases is above 85% in from year 2000 to 2006 (Taiwan Cancer Registry. Taiwan Cancer Registry Annual Report. http://crs.cph.ntu.edu.tw/). This registry has been utilized by numerous published studies in the research of cancer incidence in Taiwan, including hepatocellular carcinoma, nasopharyngeal carcinoma, cervical cancer, and multiple myeloma. [1115]. This study protocol was approved by the Institutional Review Board of Taipei Veterans General Hospital.

The recording of esophageal cancer was based on the International Classification of Disease for Oncology (ICD-O). All primary malignant neoplasms of the esophagus (ICD-O codes 150.0 –150.9) were included in this analysis. Esophageal cancers were divided into the following histological categories: squamous cell carcinoma (ICD-O codes 8050– 8082), adenocarcinoma (ICD-O codes 8140-8573), others including unspecified morphology (ICD-O codes 8000-8043), and those with missing histology [10]. Population data derived from the Department of Statistics of the Ministry of the Interior of Taiwan (http://www.moi.gov.tw/stat/english/index.asp) were used in the calculation of the crude incidence of esophageal cancer. The age-standardized incidence per 100,000 was calculated according to the year 2000 world standard population from the WHO world health statistics [16]. Cases were grouped into 5 periods, namely 1979–1983, 1984–1988, 1989–1993, 1994–1998, and 1999–2003. Averaged crude and age-adjusted incidence rates were calculated by averaging the incidence of esophageal cancer of the 5-year time periods. Trends in incidence were assessed by calculating the annual percentage changes (APCs) from 1979 to 2003. The APC calculates average percentage increase or decrease in cancer incidence rates per year over selected times. A regression line was fitted under the assumption that the natural logarithm of cancer rates changes at a constant rate when using calendar year as a predictive variable, that is, y = mx +b, where y = ln (rate) and x = calendar year. Then the APC = 100 × (em1). The null hypothesis being tested was that the APC = 0, that is, m = 0, with a 2-sided alpha of 0.05. A sensitivity analysis was conducted assuming cases with missing histology as squamous cell carcinoma.

All analyses were performed using SAS software, version 9.1 (SAS institute Inc, Cary, North Carolina). A value of P < 0.05 was considered statistically significant.

Results

Overall incidence of esophageal cancer in Taiwan from 1979 to 2003

During the period from 1979 to 2003, 17,899 new cases of esophageal cancer were diagnosed. ESCC was the predominant esophageal cancer in Taiwan, while only around 6–7% of all the esophageal cancers belonged to the adenocarcinoma group (Table 1). Most patients were male (n = 16,239). The total numbers of esophageal cancer increased progressively from 1979 to 2003 in both genders. The age-standardized incidence of esophageal cancer can only be seen in male (6.10, 5.64, 6.03, 7.28, and 9.27 per 100,000-year in each cohort) with an APC of 2.26% (95% confidence interval [CI], 2.13– 2.39, P < 0.0001), but not in female (0.89, 0.69, 0.81, 0.82, and 0.85 per 100,000-year in each cohort; APC = 0.20%, 95% CI, 0.02–0.37 P = 0.66). The age-adjusted incidence of other or missing esophageal cancer is decreasing progressively in both genders from 1979 to 2003 (Fig. 1).
Table 1

Numbers and proportion of different histology types of esophageal cancer in 5 cohorts from 1979 to 2003 in Taiwan

 

1979–1983

1984–1988

1989–1993

1993–1997

1998–2003

Male

 Total No. of esophageal cancer

2,050

2,216

2,753

3,753

5,467

 Squamous cell carcinoma, n (%)

1,497(73%)

1,675(76%)

2,141(78%)

3,095(82%)

4,738(87%)

 Adenocarcinoma, n (%)

134(7%)

132(6%)

167(6.%)

214(6%)

281(5%)

 Other or missing, n (%)

419(20%)

409(18%)

445(16%)

444(12%)

448(8%)

Female

 Total No. of esophageal cancer

243

229

317

387

484

 Squamous cell carcinoma, n (%)

150(62%)

154(67%)

211(67%)

267(69%)

363(75%)

 Adenocarcinoma, n (%)

30(12%)

23(10%)

29(9%)

41(11%)

59(12%)

 Other or missing, n (%)

63(26%)

52(23%)

77(24%)

79(20%)

62(13%)

Overall

 Total No. of esophageal cancer

2,293

2,445

3,070

4,140

5,951

 Squamous cell carcinoma, n (%)

1,647(72%)

1,829(75%)

2,352(77%)

3,362(81%)

5,101(86%)

 Adenocarcinoma, n (%)

164(7%)

155(6%)

196(6%)

255(6%)

340(6%)

 Other or missing, n (%)

482(21%)

461(19%)

522(17%)

523(13%)

510(9%)

https://static-content.springer.com/image/art%3A10.1007%2Fs10552-009-9458-0/MediaObjects/10552_2009_9458_Fig1_HTML.gif
Fig. 1

Age-standardized incidence (with 95% confidence interval at year 2000 world population) of esophageal cancer with other or missing histology males and females from 1969–1973 to 1999–2003 in Taiwan

Trends in the incidence of ESCC in Taiwan from 1979 to 2003

Both gender showed a statistically significant increasing trend in the age-standardized incidence of ESCC (Fig. 2), with a higher APC in male (APC = 3.27%, 95% CI, 3.12–3.42, P < 0.001) than female (APC = 1.23%, 95% CI, 1.02–1.45, P = 0.03).
https://static-content.springer.com/image/art%3A10.1007%2Fs10552-009-9458-0/MediaObjects/10552_2009_9458_Fig2_HTML.gif
Fig. 2

Age-standardized incidence (with 95% confidence interval at year 2000 world population) of esophageal squamous cell carcinoma in males and females from 1969–1973 to 1999–2003 in Taiwan

Trends in the incidence of EA in Taiwan from 1979 to 2003

Both gender showed an increasing crude incidence of EA from 1979 to 2003 (male: 0.28, 0.26. 0.31, 0.39, and 0.49 per 100,000-year in each cohort; female: 0.07, 0.05, 0.06. 0.08, and 0.11 per 100,000-year in each cohort). However, after age standardization, there existed no increasing trend in the EA incidence during the investigated period (male: APC = 0.72%, 95% CI, 0.50–0.93, P = 0.20; female: APC = 1.59%, 95% CI, 1.00–2.18, P = 0.30; Fig 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs10552-009-9458-0/MediaObjects/10552_2009_9458_Fig3_HTML.gif
Fig. 3

Age-standardized incidence (with 95% confidence interval at year 2000 world population) of esophageal adenocarcinoma in males and females from 1969–1973 to 1999–2003 in Taiwan

Recalculation of age-standardized incidence of ESCC by assumption that all missing cancers were ESCC

In Taiwan, a significantly increasing trend of the age-standardized incidence of ESCC was identified between 1979 and 2003. But, a high numbers of unclassified cancers in the early period of our study (around 20% of the total) were also found (Table 1), which might bias the result of the increasing trend of ESCC. Because ESCC accounts for the substantial majority of the cancers in Taiwan, it is then reasonable to assume that better recording of the cell type would reclassify most of these tumors as ESCC. To overcome the potential bias, recalculation of the age-standardized incidence was performed by assuming that all unclassified cancers were in reality ESCC. The increasing trend of ESCC can still be observed after the adjustment (2.00 of 100,000-year in 1979 to 4.78 of 100,000-year in 2003, P < 0.0001), in both males (3.28 of 100,000-year in 1979 to 8.86 of 100,000-year in 2003, P < 0.0001) and female (0.44 of 100,000-year in 1979 to 0.71 of 100,000-year in 2003, P = 0.04).

Discussion

Our study showed that the age-standardized incidence of ESCC was increasing from 1979 to 2003, while that incidence for EA in both genders remained unchanged during the same period in Taiwan.

The incidence of ESCC is significantly increased in Taiwan, especially in males. This result is in contrast to the reports from developed countries, where a decreased trend in the ESCC incidence of is almost always observed [1]. A decline in ESCC incidence has been reported in USA [16], Western Europe [17], Singapore [7], and Hong Kong [7]. Smoking has been identified as a risk factor leading to ESCC [18], and the decline in the incidence of ESCC in the USA and Singapore has been attributed to a drop in the prevalence of cigarette smoking. On the contrary, in Taiwan, the per capita cigarette consumption rose from 1,071 cigarettes per capita in the year 1969 and 1,604 in the year 1986 to 2,082 cigarettes in the year 2001 [19]. Secondly, alcohol intake is significantly associated with ESCC risk (hazard ratio = 4.93, 95% confidence interval: 2.69, 9.03), but it is not a risk factor for EA [20]. Alcohol consumption has skyrocketed in Taiwan in the past decades, in parallel with local economic development, industrialization, and modernization [21]. The prevalence of alcoholism in Taiwan increased from 0.01% in 19461948 to 4.8% in metropolitan Taipei (the largest city in Taiwan) and 9.8% in two small towns in 19811986 [21]. Betel nut chewing may be another potential factor contributing to the increasing trend of ESCC in Taiwan. In Taiwan, about 1 in every 10 people chews betel nut. The prevalence of betel nut chewing among the Chinese in Taiwan over 15 years of age is around 8.8–16.1% [22, 23]. The rapidly increasing incidence of oropharyngeal cancer in Taiwan is probably heavily influenced by this rising use of betel nuts [24]. Several studies from Taiwan have also demonstrated that betel nut use is an independent risk factor for ESCC [25, 26]. In addition, there is a synergistic effect of smoking/alcohol drinking and betel nut use in ESCC, and almost all betel nut chewers (92.6%) are smokers in Taiwan [25, 27]. Furthermore, gender difference is extremely prominent in the prevalence of smoking and betel nut chewing in Taiwan, with male 17–18 folds higher than female (betel nut chewing: 15.85% vs. 0.96%; smoking: 41.13% vs. 2.25%) [22]. Study also demonstrated that males were more vulnerable in developing alcohol abuse or alcohol dependence, and had higher alcoholism prevalence than females in Taiwan [28]. Based on the earlier observations, it is not surprising to find a higher increasing trend of ESCC in the males than females of Taiwan.

The prevalence of GERD, leading to Barrett’s esophagus, is lower in Asian countries than in Western countries [29]. The GERD prevalence, however, is reported to be increasing in Asia, including Taiwan. For example, in a questionnaire survey in 2002, we have shown that the prevalence of weekly reflux symptoms in the healthy adult population is 6.6% [30], which increased to 14.2% in 2005 [31]. For reflux esophagitis, its prevalence is reported to be as low as 1.8% in 1979 and up to 18.4% in 2005 [31, 32]. This increment of GERD prevalence might be the results of increasing physicians’ awareness of GERD, westernization of lifestyle and increased prevalence of obesity, and overweight in Taiwan. A probable causal relationship has also been reported between GERD and EA [5]; thus, it will be interesting to know whether there is a concurrent rise in the incidence of EA in Taiwan. Though we did find the crude incidence of EA is increasing in Taiwan, its age-standardized incidence is unchanged during the period from 1979 to 2003. The discrepancy increasing GERD and unchanging trend of EA may be explained by the long clinical interval between GERD and occurrence of EA. It has been estimated that the prevalence of GERD rises dramatically after the age of 40 [33]. Furthermore, EA occurs most commonly around 65–84 years of age in Taiwan (data not shown). Therefore, it may take 20–30 years for GERD to develop into EA, and the observation period in our study may thus be too short to find an increasing trend for EA. An increasing tendency may possibly be observed in the near future in Taiwan.

Two other Chinese population studies regarding the incidence of esophageal cancer are available from Singapore and Hong Kong [7, 8]. The time trend changes for ESCC and EA differ among these three Asia Chinese societies. The ESCC is decreasing in Singapore and Hong Kong [7, 8], but increasing in Taiwan. The EA incidence is increasing in Singapore [7], decreasing in Hong Kong [8], and is unchanged in Taiwan. We speculated that lifestyle factors may lead to such a huge difference in the time trend change of esophageal cancers among the Chinese in different regions, since the Chinese in three regions have their own local culture. For example, betel nut chewing is rarely seen in Singapore or Hong Kong, but is common in Taiwan. This phenomenon may partly explain the rapid increase of ESCC only in Taiwan. On the other hand, although the prevalence of GERD is reported to be increasing in all these three Chinese-predominant and Westernized societies, the EA incidences are not as expected to be increasing in all three countries/regions. The factor leading to this extreme discrepancy is unclear and needs further investigation. Nevertheless, our findings would suggest that lifestyle factors may be important in the pathogenesis of esophageal cancer among Chinese.

Esophageal squamous cell carcinoma (ESCC) comprises around 75–85% of all the esophageal cancer among the Chinese in Taiwan, which findings are comparable to those from Hong Kong, Singapore, and Japan [79]. In USA, the incidence of EA has been reported to be outnumbered ESCC since 1990 in the USA, especially in the non-Hispanic white man [1, 6]. But other races of both gender in USA, including American Indian/Alaska Native; Asian or Pacific Islander, still have a higher ESCC incidence [6]. The reason for this discrepancy between West and East is unclear, although certain factors may be responsible for the differences. First, obesity is reported to be a major risk factor for EA, but is inversely associated with ESCC [4, 34], and Chinese people have a smaller body mass index than Caucasian people [35]. Second, the different prevalence of Helicobacter pylori (H pylori) and its genotype in Taiwan and Western countries might contribute to the different distribution of the esophageal neoplasm. A recent metanalysis showed an inverse relationship of prevalence of the H pylori and H pylori cagApositive strains with both Barrett’s esophagus and EA, suggesting that H pylori infection might play a protective role in EA occurrence [36]. In Taiwan, H pylori infection has occurred in 72.3% of the general population of over 40 years old [37], and more than 98% of the organisms have been found to be cagA strains [38]. In contrast, the prevalence of H pylori infection is relatively low in developed countries (about 35.4% in the general population over 40 years old in the United States) [39], and only around 60% of the H pylori is cagA strain. Third, Barrett’s esophagus is the only known premalignant condition for EA [40], but it rarely occurs in Asians [29], which may also be a result of the protective effect of H pylori.

People may argue that the high numbers of unclassified cancers in the early period of our study (around 20% of the total) might bias the result of the increasing trend of ESCC. Actually, similar situation is also existed in two previous studies from Singapore (44.5% during 1968–1972), Hong Kong (22.5% during 1984–1988), and Japan (24.9% during 1993–1995) [79]. By recalculation of age-standardized incidence assuming that all unclassified cancers were in reality ESCC, we still noted an increasing trend of ESCC after this adjustment. Therefore, we do believe that the increasing trend of ESCC in Taiwan is real. Another limitation of this study is that we found a drop in the incidence of ESCC and EA from 1st to 2nd time period. The reason for this drop is unknown. Since there is no protocol change during those periods, the possible explanation may be the errors in recording or cancer miss-classification during the early stages of cancer registration. Nevertheless, the drop will actually not affect the increasing trend in the incidence of ESCC in Taiwan. Miss-classification of EA into cardiac cancer is possible in early days, but this will actually increase the EA incidence and will hardly affect the unchanging trend of the EA incidence in Taiwan. Despite of the fact, we should still be cautious that there seems to have a steady increasing trend of EA after 1984 till 2003.

In conclusion, our data indicate that the incidence of ESCC in Taiwan significantly increased from 1979 to 2003. Increased consumption of cigarettes, alcohol, and betel nuts may be associated with the increasing trend. The development of policies and programs targeting a reduction of exposure to pivotal risk factors may be essential for the prevention of ESCC in Taiwan. Though the EA incidence remained unchanged during the examined period, a caution should be sounded that an increasing tendency could be observed in Taiwan in the near future. Lifestyle factors may be important in the pathogenesis of esophageal cancer among Chinese.

Acknowledgment

We thank the Taiwan National Cancer Registry for the help in acquiring data used in this study.

Copyright information

© Springer Science+Business Media B.V. 2009