Socioeconomic status and risk of adenocarcinoma of the oesophagus and cancer of the gastric cardia in Scotland

The incidence of oesophageal adenocarcinoma and gastric cardia cancer increased strikingly in Scotland between 1977 and 1996. In contrast to other cancers of the oesophagus and stomach, which showed a clear relationship with socioeconomic status, no trend was evident for oesophageal adenocarcinoma or gastric cardia cancer during 1987–1996. © 2000 Cancer Research Campaign

Mid-year population estimates were derived from the Annual Reports of the Registrar General for Scotland (Registrar General for Scotland, 1978. Annual and quinquennial sex-specific agestandardized incidence rates were calculated for the period 1977-1996 by direct standardization to the European Standard Population (Waterhouse et al, 1976). In addition, for the total 10-year period 1987-1996, 1991 census-derived Carstairs deprivation category- (Carstairs and Morris, 1991) and sex-specific age-standardized incidence rates were calculated. The percentage changes in incidence between 1977 and 1996, were estimated by linear regression on annual log age-standardized rates. P-values reflect the statistical significance of the gradients of each of the fitted regression lines.
Sex-specific temporal trends in incidence rates of the seven diagnostic entities chosen for analysis are shown in Table 1. Rates of all seven diagnostic entities were, as expected, higher in males than females. Among males, between 1977 and 1996, the estimated percentage changes in incidence rates of oesophageal adenocarcinoma and gastric cardia cancer were +139.5% (P < 0.0001) and +94.9% (P < 0.0001) respectively. Corresponding estimates for females were +124.6% (P < 0.0001) and +65.3% (P = 0.01).
Age-standardized incidence rates by deprivation category and sex are shown for each of four diagnostic categories in Figures 1A-D. Although there is a clear association between deprivation and the risk of oesophageal cancers other than adenocarcinoma, and stomach cancers other than those of the gastric cardia, no trend is seen for oesophageal adenocarcinoma or gastric cardia cancer. The incidence of oesophageal adenocarcinoma has continued to increase strikingly in Scotland in recent years. Clearly, there is scope for at least some of the increase to be an artefact of diagnostic and/or coding practice. However, a previous analysis of Scottish data suggests that the observed trend is unlikely to be due entirely to artefact (McKinney et al, 1995). Certainly, in the time period covered by the present analysis, the estimated absolute changes in incidence of oesophageal cancers other than adenocarcinoma, and of gastric cardia cancers, are insufficient to account for the estimated absolute increase in incidence of oesophageal adenocarcinoma. It is also of interest to note that rates of oesophageal adenocarcinoma and gastric cardia cancer combined have, since the early 1980s, begun to exceed rates of other malignant neoplasms of the oesophagus in males, but not in females. This sex difference, which has also been seen in the Merseyside and Cheshire area (Dolan et al, 1999), is not completely consistent with a diagnostic or coding artefact, which might be expected to have a similar effect in both males and females. However, it must be acknowledged that subsite misclassification could, at least in part, explain the observed increase in incidence of gastric cardia cancer (Ekström et al, 1999;Levi et al, 1999).
While many studies in diverse populations have found an association between socioeconomic deprivation (using a variety of measures) and risk of all cancers of the oesophagus and all cancers of the stomach (Faggiano et al, 1997), the relationship with tumour subsites and subtypes has been much less studied and results are inconsistent. Analysis of population-based cancer registration data for the West Midlands Region of the UK showed that a higher proportion of male cases of adenocarcinoma of the oesophagus and gastric cardia were classified as higher social class (based on occupation) compared both to other sites of gastric cancer, and to all sites of cancer combined (Powell and McConkey, 1990). Conversely, in the same comparison, a lower than expected proportion of cases of adenocarcinoma of the oesophagus and gastric cardia were classified as lower social class. However, in a recent population-based case-control study, Ye et al (1999) did not find any striking differences in the distribution of socioeconomic status between 90 cases of gastric cardia cancer and 1164 controls. The results of several other case-control studies have suggested an inverse association between risk of cancers arising in the vicinity of the oesophago-gastric junction and socioeconomic status in terms of education (Palli et al, 1992;Vaughan et al, 1995;Gammon et al, 1997), occupation (Palli et al, 1992;Brown et al, 1994) and income (Brown et al, 1994;Gammon et al, 1997). However, Brown et al (1994) found a non-significant increase in risk of adenocarcinoma of the oesophagus and oesophago-gastric junction with a higher level of education. Wu-Williams et al (1990) also found a non-significant increase in risk of cardia cancer with a higher level of education, although their use of neighbourhood controls may not have been ideal for studying socioeconomic differences in risk. In a hospital-based casecontrol study, compared to controls, Kabat et al (1993) found higher proportions of cases of adenocarcinoma of the oesophagus or cardia with higher levels of education and (in males) occupation. A prospective cohort study in The Netherlands found a lower risk both of cardia cancer and of other subsites of stomach cancer in men with the highest level of education, although there was no significant trend across the levels of educational attainment for either diagnostic entity, and the lower risk of cardia cancer was attenuated by adjustment for a variety of confounding factors in a multivariate analysis (van Loon et al, 1998). It is important to remember that all of the above studies are based on individual measures of socioeconomic status whereas the Carstairs score used in our analysis is an area (postcode sector) based indicator of deprivation based on levels of overcrowding, male unemployment, low social class and household car ownership derived from the decennial census (Carstairs and Morris, 1991). Although areabased measures are prone to misclassification and the 'ecological fallacy', they have been shown to be independent predictors of morbidity (Davey Smith et al, 1998) (Spencer et al, 1999), may be more discriminating than individual measures of socioeconomic status. Recent analytical studies have implicated cigarette smoking, obesity, gastro-oesophageal reflux disease and Barrett's oesophagus as the main risk factors or predisposing conditions for adenocarcinoma of the oesophagus and gastric cardia (Devesa and Fraumeni, 1999). A recent population-based case-control study in Scotland and other regions of the UK identified obesity in early adulthood and low consumption of fruit as risk factors for oesophageal adenocarcinoma in women (Cheng et al, 2000).
Given the strong inverse relationship between cigarette smoking and social class (Dong and Erens, 1997), the absence of a clear association between deprivation and risk of oesophageal adenocarcinoma and gastric cardia cancer in Scotland seems surprising. However, it is important to note that the increased risk of these cancers has been shown to persist for at least 30 years after smoking cessation (Gammon et al, 1997) so that current deprivationcategory specific incidence rates may reflect smoking patterns at a time when the prevalence of smoking was distributed more evenly across the social classes in the UK (Wald and Nicolaides-Bouman, 1991). Similarly, current evidence suggests that there is not a clear relationship between obesity and social class in Scotland (Dong and Erens, 1997). Evidence is also emerging of an inverse relationship between infection with cag+ Helicobacter pylori strains and risk of adenocarcinoma of the oesophagus and gastric cardia (Blaser, 1999). Indeed, the declining prevalence of H. pylori infection has been offered as a possible explanation for the increasing incidence of adenocarcinoma of the oesophagus and gastric cardia, and the decreasing incidence of cancers of the distal stomach. If this explanation is correct, it would help to explain the tendency towards lower incidence of gastric cardia cancer among people living in more deprived communities in Scotland, since there is a well established relationship between lower social class and risk of infection with H. pylori in the UK .
In summary, in contrast to the findings for other cancers of the oesophagus and stomach, the recent striking increase in incidence of oesophageal adenocarcinoma and possibly also gastric cardia cancer, coupled with the absence of a clear association with deprivation support the view that, in Scotland as in other countries, the aetiology of these tumours is distinct.