SEARCH

SEARCH BY CITATION

Keywords:

  • time trends;
  • mortality;
  • incidence;
  • esophagus cancer;
  • gastric cardia cancer

Abstract

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

China was one of the countries with the highest esophagus cancer risk in the world during the 1970s. This report provides data on time trends of esophagus cancer incidence and mortality during the 1970s–90s in selected geographic areas of China. Information on newly diagnosed cancer cases and cancer deaths is based on data collected by local population-based registries and Disease Surveillance Points (DSP). For the whole country, esophagus cancer mortality decreased slightly, 17.4 per 105 populations during 1990–92 in contrast to 18.8 per 105 populations in 1973–75. In the Linxian area, trends in the incidence and mortality rates for esophagus+gastric cardia cancer reversed over time; incidence rates increased significantly during 1959–72 but were decreased significantly on average −2.26% (95% Confidence Interval [CI]: −1.74, −2.77) and −1.10% (95% CI = −0.58, −1.62) per year for males and females, respectively, during 1972–97. In urban Shanghai, incidence trend for esophagus cancer decreased monotonically and significantly on average by −4.99% (95% CI = −4.28, −5.70) and −5.18% (95% CI = −4.99, −5.70) per year for males and females, respectively. In Nanao islet, esophagus+gastric cardia cancer mortality rates increased during 1970–82 but decreased slowly from 1982–99 (−0.96% per year; 95% CI = −0.14, −1.78). Our study indicates that incidence and mortality rates for esophagus or esophagus+gastric cardia cancer are now decreasing in China. The declines may be due to an unplanned success of prevention, such as changes in population dietary patterns and food preservation methods. © 2002 Wiley-Liss, Inc.

In the 1970s, the highest mortality rates of esophagus cancer in the world were reported from China,1 a large country with wide geographical variation in the rates. In the 1990s, the mortality for the disease in China estimated by Pisani and her colleagues2 was second worldwide, with South Africa having the highest rates. A sustained decline in China was first reported in urban Shanghai.3, 4, 5 The time trends for esophagus cancer, however, among different areas of China, especially among high risk areas, are not yet fully understood.

The World Health Organization (WHO) has encouraged member states to introduce systems of death registration involving medical certification of cause of death. In consequence, the Disease Surveillance Points (DSP) program was established in China in 1978 (see incidence and mortality rates below).6, 7 The rates from rural China and Shanghai, Tianjin and Qidong registries are available from the WHO databank.6, 7 Thus, registers covering limited populations and sample surveys are becoming more common and there has been a slow increase in the availability of such data, especially in areas with high risk for various cancers.

We present analyses of mortality and incidence trends from esophagus cancer over time by areas of China, to provide an update on the continuing decline in urban Shanghai and to present information regarding time trends for the cancer not reported previously from some high risk areas in China. For these analyses, we have utilized data from the DSP and from cancer registers and also any other existing data on the cancer that could provide information on probable mortality and incidence trends.

MATERIAL AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Selected geographic areas

By level of the mortality rate for esophagus cancer reported in the 1970s, 6 regions have been identified as at highest risk for esophagus cancer in China, 5 of which are located in the mid-north of China, including Linxian of Henan province, Yang City of Shanxi, Huaian of Jiangsu, Yanting of Sichuan and Shenxian of Hebei. Only 1 is located in South China, the Nanao area of Guangdong province. Two of these areas (Linxian and Nanao) and urban Shanghai, with a large change in mortality or incidence for esophagus cancer, will be considered in our study. Data sets from other areas are of very questionable completeness, accuracy or representativeness and were therefore rejected (coverage <70% or medically certified deaths <70%).

The Linxian area, located in North China, is known for its very high incidence of and mortality from esophagus cancer. The local population-based cancer registries were established in 1959. Subsequently, it joined in the DSP program in 1977.8, 9 The Linxian population was approaching 900,000 in 1997.

Nanao is also one of the areas with the highest mortality for esophagus cancer in China, with a population size of more than 72,000 in 1997. It is an island and islet county located in the Chaoshan area of Guangdong province in South China. Nanao joined the DSP program in 1970 and the local population-based cancer registry was established in 1990.

In Linxian and Nanao, the excess cancers occur not only as squamous cell carcinomas of the esophagus, but also as adenocarcinomas of the gastric cardia. Historically, both of these tumors have been called esophagus cancer, because of their proximity to one another and their similarity.10

Shanghai is the most populous city in China, with more than 10,000,000 people in 1997. The Shanghai Cancer Register is a population-based registry covering the entire city of Shanghai. Shanghai was defined as a medium risk area for esophagus cancer, with reported incidence rates of 28.8 per 105 and 11.3 per 105 populations in males and females, respectively, in 1972–74. Since then, however, the incidence rate for esophagus cancer in urban Shanghai has been changing markedly.

Incidence and death rates

Information on esophagus cancer deaths in China is based on causes of death reported on death certificates collected by the DSP program. Briefly, the DSP program collects cancer mortality data from 29 provinces and municipalities and 145 DSPs, representing an estimated 1% of the Chinese population. A typical surveillance point covers a defined population of about 100,000, in which a team of workers, including a doctor, investigates each death and assess the underlying cause from medical records and interviews with family members. The DSP program has yielded reliable statistics on death and disease for the entire country and has had extensive use.6, 7, 11, 12, 13 Information on newly diagnosed esophagus cancer cases is obtained from the local population-based registries. The data used in our study were derived from the DSP and from the local population-based cancer registries.

For the analysis, we used available mortality and incidence data from 1959–99. For the period 1959–78, the 8th revision of the International Classification of Diseases (ICD-8) was used and for 1979–99 the 9th revision (ICD-9). In the ICD, cancers of the esophagus and gastric cardia were coded as 150 and 151.0, respectively.

We used resident population estimates for each year to compute age-adjusted cancer incidence and death rates. Rates are expressed as per 100,000 populations and age-adjustment was carried out by the direct method using the world standard population for Shanghai and the 1964 China standard population for Linxian and Nanao.

Average annual percent change

The average annual percent change (APC) is estimated by fitting a regression line to the natural logarithm of the rates using calendar year as a regressor variable, i.e., y = mx + b in which y = ln(rate) and x = calendar year. Then the estimated APC = 100 × (em − 1). Testing the hypothesis that the APC is equal to 0 is equivalent to testing the hypothesis that the slope of the line in the above equation is equal to 0 (i.e., that the rate is not increasing or decreasing). The hypothesis test statistic uses the t distribution of m/SEm in which SE is the SE of m and the number of df is equal to the number of calendar years −2. A 100(1 − α)% CI estimate of m is calculated using t distribution with (the number of calendar years −2) df by the formula: m ± t × SE.14 The calculation assumes that rates increase or decrease at a constant rate over time although the validity of this assumption has not been assessed. Differences between incidence or mortality trends for the 2 time periods in the same area have been tested for statistical significance using a t statistic and the number of df is equal to the sum of the number of calendar years of each time period −2 defined as the difference in the regression coefficients divided by the SE of the difference.14 Statistical significance for all hypothesis tests was assessed using p = 0.05 and 95% CI calculated with the above methods.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

The countrywide mortality from esophagus cancer was estimated based on the DSP program for 1973–75 and 1990–92. The countrywide esophagus cancer mortality decreased slightly between 1973–75 (18.8 per 105 populations) and 1990–92 (17.4 per 105 populations). Mortality for each year was not, however, available countrywide.

Incidence and mortality for the Linxian area reported in Figure 1 and Figure 2 were for esophagus cancer and gastric cardia cancer combined. There were about 31,515 esophagus and gastric cardia cancer deaths in Linxian residents and 35,277 newly diagnosed cases were registered during 1959–97. The pattern of mortality somewhat varied by gender (Fig. 1). In males, mortality increased significantly during 1959–75 (2.98% per year) and decreased during 1975–97 (−0.99 per year). The largest annual decreases in mortality occurred during 1986–97 (−2.33% per year; 95% CI = −0.66, −3.96; p < 0.05), suggesting a quickening in the annual decrease in mortality for this disease. In females, a pattern of increasing death rates during 1959–71 was also seen (2.82% per year), followed by slowly decreasing death rates during 1971–97 (−0.61% per year).

thumbnail image

Figure 1. Age-standardized mortality trends from both combination of esophageal cancer and cardiac cancer by gender in Linxian, North China, 1959–97. □, male; ○, female.

Download figure to PowerPoint

thumbnail image

Figure 2. Age-standardized incidence trends from both combination of esophageal cancer and cardiac cancer by gender in Linxian, North China, 1959–97. □, male; ○, female.

Download figure to PowerPoint

The incidence rates in Linxian significantly increased since the 1950s, at 3.18% and 2.35% per year for males and females, respectively and appear to have peaked in 1972 (Fig. 2). The incidence rates decreased significantly, on average by −2.26% and −1.10% per year for males and females, respectively, during 1972–97.

For the Nanao area, the number of esophagus and gastric cardia cancer deaths has been estimated to have been 1686 during 1970–99. Mortality from the disease increased for both genders combined during 1970–82 (2.51% per year) and decreased (−0.96% per year) during 1982–99 (Fig. 3).

thumbnail image

Figure 3. Age-standardized mortality for both genders from both combinations of esophageal and cardiac cancer in Nanao Island, South China, 1970–99.

Download figure to PowerPoint

Incidence data based on the cancer register were only available for 1990 and later. During 1995–99, 244 esophagus and 120 gastric cardia cancer cases were registered among residents of Nanao. In the 5-year period, the incidence rates for esophagus and gastric cardia cancers combined for both genders ranged between 92–102 per 105, with a non-significant reduction of −1.54% per year.

A total of 23,809 esophagus cancer cases were diagnosed among residents of urban Shanghai during 1972–98. In the 27-year period, the incidence reported by the population-based cancer registry decreased monotonically and significantly on average by −4.99% per year for males and −5.18% per year for females, respectively (Fig. 4). Further, the populations studied, the years covered, % trends and 95% CIs were totaled in Table I.

thumbnail image

Figure 4. Age-standardized incidence trends from esophageal cancer by gender in urban Shanghai, China, 1972–98. □, male; ○, female.

Download figure to PowerPoint

Table I. Summary Table for Trends of Incidence and Mortality of Esophageal Cancer (Shanghai) or Esophagus + Gastric Cardia Cancer (Linxian and Nanao)
Population studied and areasYears coveredAverage annual change (%)195% CI
  • 1

    The minus sign implies a decreasing trend or else an increasing trend

  • 2

    p < 0.05.

Incidence rate   
 Linxian   
  Male1959–19723.1821.53, 4.85
  Male1972–1997−2.262−1.74, −2.77
  Female1959–19722.3520.14, 4.61
  Female1972–1997−1.102−0.58, −1.62
 Nanao island   
  Both genders1995–1999−1.54−5.56, 2.65
 Urban Shanghai   
  Male1972–1998−4.992−4.28, −5.70
  Female1972–1998−5.182−4.48, −5.88
Mortality rate   
 Linxian   
  Male1959–19752.9821.55, 4.42
  Male1975–1997−0.992−0.04, −1.94
  Female1959–19712.82−0.44, 6.19
  Female1971–1997−0.612−0.03, −1.19
 Nanao island   
  Both genders1970–19822.51−0.77, 5.90
  Both genders1982–1999−0.962−0.14, −1.78

Information on the proportions by site and corresponding changes over time in urban Shanghai is available (data not shown). The proportions for cancers of esophagus (ICD code: 150) and stomach (ICD code: 151) of all subsites in both males and females declined in parallel with time. There was also a small decline for the unspecified subsites (ICD code: 159) and a little for the unspecified site (ICD code: 199).

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Before the 1970s, in the high risk areas of China, gastric cardia cancer used to be classified with esophagus cancer mainly to avoid classification biases, so the incidence or mortality reported for esophagus cancer in Linxian and Nanao is for esophagus and gastric cardia cancer combined. The esophagus cancer to gastric cardia cancer ratio was about 2:1 in the data presented in our study. The urban Shanghai, however, data were for esophagus cancer only in Figure 4.

Trends in both incidence and mortality changed to be in the opposite direction with time in China; they first increased and then decreased. The downturn in esophagus cancer risk after long term increases varied by gender and region. Decreases in mortality occurred later than in incidence; whereas in the high risk areas with a more rural population, the decrease occurred later than in urban Shanghai and in females later than in males.

The earliest decreases in incidence occurred for males in urban Shanghai in the early 1970s,3, 4 and for females during almost the same time period. In Linxian, decreases in incidence in males and females began in the mid-1970s, which confirmed results obtained in a previous investigation.15 In Nanao, although incidence rates and mortality are not yet significantly decreasing in both genders, they were no longer significantly increasing as was mortality before the early 1980s.

The reasons for the decreasing incidence and mortality from cancers of the esophagus in Shanghai and esophagus and gastric cardia in Linxian are not well understood. The very low utilization of screening tests (e.g., endoscopy16, 17 or balloon-mesh cytology18), suggests that factors other than screening are probably involved.

The trends could reflect changes in diagnostic accuracy, e.g., because of improved diagnostic abilities, cancers previously attributed to the esophagus, or even esophagus plus gastric cardia, may now be called non-cardia stomach cancers. If there had been confusion in the diagnosis of both, one would expect that stomach cancer would have shown an increase, or a lesser rate of decrease. There have, however, been the same time trends in rates from esophagus and gastric cancer in the corresponding areas.3, 4, 5, 9, 19 Thus, the stomach cancer to esophagus cancer case ratios were relatively constant over time. Also, if diagnostic accuracy has improved because of more accurate tumour localization methods (lower third of the esophagus or gastric cardia) the rates for cancer of esophagus or gastric cardia would have increased. This was not observed, however, so improving diagnostic accuracy is probably not responsible for the trends observed.

Another explanation for the time trends might be changes in medical seeking behavior/access and population mobility. Clearly, the medical seeking behavior/access for people have improved over time, but such improvement should not cause a decreasing time trend. Changes in immigration patterns in Shanghai may have influenced the downward trend of esophagus cancer, however. Rates for this cancer were particularly elevated in certain urban districts, with a concentration of immigrants form high-risk areas north of the Yangtze River.20

Changes in specific dietary and other environmental factors are likely to be involved, however. The potential for changing dietary habits of populations may represent an unplanned success of prevention as for gastric cancer.21 This would presumably represent the results of replacing traditional methods of preservation and storage with refrigeration, together with an increased consumption of vitamin-rich food. From 1956–82, the per person consumption of fresh fruits and vegetables in Shanghai increased more than 60%.22 High consumption of fruits and vegetables has been associated with a reduced risk of both esophagus and gastric cancers in many epidemiology investigations. A nutrition intervention trial of multiple vitamins and minerals was carried out in some villages of Linxian during 1985–91 and the findings from our study indicated that a combination of beta-carotene, vitamin E and selenium may effect a reduction in esophagus+gastric cardia cancer risk in this population.23, 24

Other studies of Shanghai and Linxian data have considered the effects of period and birth cohort on the time trends.3, 8 The results showed that although the downward trends were observed for all age groups and birth cohorts in the Shanghai population, there were greater rates of decline in the younger age groups and the more recent birth cohorts among both men and women, whereas in Linxian, a decline was observed only in younger age groups and the most recent birth cohorts. This indicates that the time trends can be related to a period effect, a decrease in exposure to risk factors or an increase in exposure to protective factors in these areas.

Considering the above factors, it is easy to explain why an urban region first experiences the greatest decreases in esophagus cancer incidence. By this deduction the incidence trends for esophagus or gastric cardia cancers in Nanao and other rural areas in China will also decrease quickly in the coming years.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

We thank H.L. Yang for assistance in study and P. Yu for her excellent secretarial assistance.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES