Background

The incidence of esophageal carcinoma has been rising rapidly during the past three decades, and it is the seventh leading cause of cancer-related death in males in the USA.1 This is largely due to increasing numbers of adenocarcinoma cases, which are now 600 % higher than in the 1970s.2 In fact, the absolute number of these cancers is even higher if adenocarcinoma of the cardia, often erroneously excluded from these analyses, is included. A similar trend has also been reported in other western countries.3 – 5

Previous studies have suggested that this trend in esophagus and the gastric cardia (EAC) incidence may be slowing and may even have plateaued over the past decade.6 Reasons behind these changes are not completely understood, although several mechanisms, including dietary and environmental causes, have been suggested.7

Given the poor survival rates of advanced EAC, detection and treatment in an early, potentially precancerous stage is paramount. On the other hand, since widely applicable, inexpensive screening techniques are not available, the population that would benefit the most from a screening program has to be defined.

The aim of this brief study was to update the latest trends and analyze the racial and gender differences in esophageal adenocarcinoma incidence in the USA. Data presented in this study could form the basis for the better allocation of resources in future esophageal cancer screening programs.

Methods

Surveillance Epidemiology and End Results (SEER) 9 database of the National Cancer Institute was queried for all cases of histologically verified primary adenocarcinomas of the EAC. SEER 9 has collected information on all newly diagnosed malignancies since 1973 among residents of the nine original SEER geographic areas (areas within the states of Connecticut, Hawaii, Iowa, New Mexico, and Utah and the cities of Atlanta, Detroit, San Francisco, and Seattle) and represents approximately 10 % of the population from the USA.8 Our analysis included data from all available years (1973–2009). EAC was defined by using tumor site codes C15.0–15.5, 15.8, 15.9, and 16.0 and ICD-O-3 histology codes 8140–8573. From this group, we excluded all patients that were diagnosed at autopsy or from death certificate data. The final study cohort comprised 26,334 patients with EAC. Information about age-adjusted incidence, sex, race, tumor stage, and population was obtained from SEER. Racial information was dichotomized into “white” and “nonwhite” categories. Stage-specific distribution was analyzed using the SEER historic stage A variable, which defines in situ stage as noninvasive cancer, localized stage as cancer limited to the primary organ, regional stage as disease extending beyond the primary organ into adjacent tissue or regional nodes, and distant stage as metastatic cancer detected. To evaluate time trends, we used joinpoint analysis9 developed by the National Cancer Institute for the analysis of SEER data. Joinpoint analysis identifies inflection points (so-called jointpoints) where a statistically significant change over time in linear slope of the trend occurred. The level of statistical significance was set at p < 0.01. All analyses were conducted using R 2.15.1 (R Development Core Team,10 version 2.15.1).

Results

The overall incidence of adenocarcinoma of the esophagus and the gastric cardia increased from 13.4 per million in 1973 to 51.4 per million in 2009, a nearly 400 % increase (Fig. 1a). Figure 1b shows a significant change of slope in the incidence curve of adenocarcinoma of the gastric cardia. After a steep annual increase of 5.2 % since 1973, the incidence seems to have plateaued after 1985. During the same period, the incidence of esophageal adenocarcinoma changed from 3.9 per million in 1973 to 28.9 per million in 2009 surpassing the incidence of adenocarcinoma of the gastric cardia in 1996.

Fig. 1
figure 1

a Incidence of adenocarcinoma of the esophagus and the gastric cardia. Using joinpoint analysis, a statistically significant change can be detected in the incidence trend between the two time periods (period I vs. period II). b Incidence of adenocarcinoma of the esophagus vs. incidence of adenocarcinoma of the gastric cardia

Full size image

Stage-Specific Trends

The incidence of noninvasive cancer (SEER stage A) decreased after 2003 with a yearly decrease of 0.22. The percentage of patients diagnosed with in situ carcinoma declined after 2000 and remained under 2.5 % through the study period. Yearly increase of incidence of localized cancer has slowed after 1999 and was 0.13 per million between 1999 and 2009. Similarly, from 1973 to 1985, the incidence of regional disease increased from 4.7 to 12.1 per million, an annual increase of 5.2 %. From 1985 to 2009, the incidence increased only with 0.16 per million/year (1.2 %), reflecting a significant decline in the slope of the incidence curve. During the study period, the incidence of metastatic disease has continued to rise and was slightly more than 20 per million in 2009. Furthermore, 38 % of all patients were diagnosed in metastatic stage (Fig. 2).

Fig. 2
figure 2

Trends in stages of esophageal adenocarcinoma at diagnosis in 1973–2009. Using joinpoint analysis, a statistically significant change can be detected in the incidence trend between the two time periods (period I vs. period II)

Full size image

Incidence Trends by Sex and Race

Further analysis revealed substantial differences in incidence trends according to race and gender (Fig. 3a–d). Between 1973 and 2009, EAC incidence in white males increased sharply with a yearly increase in incidence of 2.3 per million reaching 107.3 per million at the end of the study period. A markedly smaller and decreasing yearly increase in incidence is observed in white women (0.7 per million before 1988 to 0.4 per million after 1988). On the other hand, the incidence of adenocarcinoma of the esophagus and the gastric cardia in nonwhite males has been decreasing with −0.3 per million since 1992 reaching an incidence of 30 per million in 2009, less than a third of that of white males. EAC incidence in nonwhite females shows no trend and remains relatively constant between 7 and 17 per million.

Fig. 3
figure 3

Incidence of EAC according to race and sex between 1973 and 2009. a White males. b Nonwhite males. c White females. d Nonwhite females

Full size image

Discussion

In this study, we provide an update on recent incidence changes through 2009 using the population-based SEER 9 registry. We have found that the incidence of esophageal adenocarcinoma continues to rise in the twenty-first century in the USA. Our results are in contrast with the findings of a previous report by Pohl et al.6 and indicate that this increase involves all stages of invasive EAC. On the other hand, like other reports,6 we have found that the slope of the incidence curve declines, which represents a significant change in trend. Furthermore, there seems to be a recent decrease in the incidence of noninvasive (in situ) disease, but this reduction has only a marginal effect on EAC incidence, since the proportion of patients diagnosed with in situ cancer remains only around 2 % of the total population.

Perhaps one of the more surprising findings of this study is the fact that, while adenocarcinoma rates are rising at a similar pace among both white men and women, the incidence of EAC remained stable in nonwhite women and actually declined in nonwhite men. We have also found that the previously reported significant epidemiological differences between adenocarcinoma of the esophageal body and that of the gastric cardia (GCA) persist in the first decade of the twenty-first century.11 While the incidence of adenocarcinoma of the cardia remained stable between 1987 and 2009, the incidence of EAC continues to rise and surpassed the incidence of GCA already in the late 1990s.

Similar trends have been observed in other countries.3 – 5 , 12 In contrast, Lagergren et al. reported that after decades of rising incidence in Sweden, the rate of EAC was stabilizing after 2001 and was showing a decreasing trend after 2005.13

Reasons for the initial dramatic rise in the incidence of EAC are largely unknown; rising prevalence of smoking, gastroesophageal reflux, and obesity have all been discussed as possible explanations. On the other hand, it is also possible that these risk factors might have already exerted their maximal effect, thus resulting in the contemporary decline in the slope of the incidence curve. Additionally, the decline in the incidence of in situ disease might be a result of recent progress in Barrett screening and possibly of an increased access to ablation therapy.14

Based on the natural history of esophageal cancer, a change in local or systemic exposures must have predated this rise in EAC incidence by several years. However, according to results from the National Health and Nutrition Examination Survey (NHANES), a significant increase in the prevalence of obesity in the USA was not observed until 1976–1980.14 Although SEER data is available since 1973 only, analysis of the Connecticut Tumor Registry Database (the oldest population-based tumor registry in the USA) indicates that the incidence of EAC has already begun to rise in the late 1960s; therefore, obesity could not have played a deciding role in the initial phase.15 Furthermore, despite the higher documented prevalence of obesity among African–Americans,16 , 17 the incidence of EAC remained stable in nonwhite women and declined in nonwhite men in the last 20 years in the USA.

The roles of gastroesophageal reflux disease (GERD) and Barrett’s esophagus (BE) in the development of EAC are well established, but their contribution to the rising incidence of EAC is uncertain since there is no incident data available on these risk factors. In a review of 45 population-based studies, El Serag found a significant increase in the prevalence of GERD between 1982 and 2005, and it seems likely that an increase in prevalence is a reflection of increase over time in the incidence of GERD.18 Additionally, a study of Olmsted County, Minnesota, residents found that the incidence of BE increased 28-fold during a period of three decades, and a study of endoscopy and histology reports between 1977 and 1996 found that frequency of BE increased from 0.2 to 1.6 % of all endoscopies in the UK.19 , 20

Tobacco smoking is an established risk factor for EAC, and the prevalence of smoking rose markedly during the first half of the twentieth century.21 Two large population-based studies found moderately increased risk of adenocarcinoma of the esophagus among cigarette smokers, but no significant association was identified in a Swedish study.22 – 24 Furthermore, it is unlikely that smoking was a major contributor in the dramatic rise in incidence of EAC since similar temporal trends were not observed in the incidence of other malignancies associated with tobacco use such as noncardia gastric cancer and esophageal squamous cell cancer.25

Our study shares the limitations of all investigations using a population-based dataset. First, despite being the most representative cancer database in the USA, the SEER 9 regions represent approximately only 10 % of the US population.8 Second, although SEER conducts regular audits to evaluate both the quality and completeness of data, a possible misclassification of patient information particularly regarding tumor site must always be considered when using a population-based database. Studies evaluating the role of misclassification in incidence trends of esophageal and cardia cancer have shown that a possible substantial misclassification of anatomic site exists, yet this does not explain the observed incidence trends.2 , 26

Despite these limitations, we can conclude that our results strongly indicate that the incidence of esophageal adenocarcinoma continues to rise in the twenty-first century in the USA, most dramatically among white males, while its incidence in nonwhite males is decreasing. The significant epidemiological differences between adenocarcinoma of the esophageal body and that of the gastric cardia persist.