Risk Factors in the Development of Esophageal Adenocarcinoma

Heiko Pohl MD; Katharina Wrobel MD; Christian Bojarski MD; Winfried Voderholzer MD; Amnon Sonnenberg MD; Thomas Rösch MD; Daniel C Baumgart MD


Am J Gastroenterol. 2013;108(2):200-207. 

In This Article


In the present study, we examined risk factors associated with different disease stages in the development of esophageal adenocarcinoma. We found that different combinations of risk factors were associated with separate disease stages. Hiatal hernia was the only risk to be strongly associated with the development of GERD. Hiatal hernia, male gender, old age, increased BMI, and duration of reflux symptoms were all associated with the development of Barrett's esophagus in patients with GERD. Finally, male gender, smoking, decreased intake of fruit and vegetables, and increasing length of Barrett's esophagus were all associated with the progression to esophageal cancer in patients with Barrett's esophagus.

It has long been known that esophageal adenocarcinoma is more common in the elderly and in white men.[3] In concordance with prior observations, our results suggest that rising age increases the risk for developing Barrett's esophagus.[18] However, we did not find that age alone was associated with further progression from Barrett's esophagus to cancer.[19] The observation that esophageal adenocarcinoma tends to occur in the elderly may be therefore primarily related to the development of the underlying Barrett's esophagus. Our results also confirm prior observations of a male predominance in both, Barrett's esophagus and esophageal adenocarcinoma. We found that male gender more than doubled the risk for patients with GERD to develop Barrett's esophagus, and further doubled the risk for Barrett patients to develop cancer of HGD.

Although smoking has been shown to be a risk factor for esophageal adenocarcinoma, doubling of its overall risk,[4] it is unknown at what stage smoking actually exerts most of its influence.[20,21] Our results indicate that smoking has no effect on the development of GERD or the transition from GERD to Barrett's esophagus. However, smoking appears to increase the risk for the progression from Barrett's esophagus to cancer. Similar results were reported in several European studies,[22,23,24,25] but are in contrast with one prior US study.[7] A recently published cohort study by Coleman et al.[25] found a doubling hazard ratio for former or current smokers as compared with never smokers. Similar to our study their study also failed to find any association with the length of smoking history.

Our study confirms prior case–control studies that showed an association between an increased BMI and the risk for esophageal adenocarcinoma.[5] While there is a general consensus that overweight and obesity increase the risk for the development of reflux disease,[5] we failed to confirm this association in our own study population. This may be related to an underlying lesser variation of the BMI in a German population with fewer overweight patients. The effect of a high BMI on the progression to Barrett's esophagus and cancer is still being debated.[24,26,27] We found a strong association between an increased BMI and the progression from reflux disease to Barrett's esophagus, but not from Barrett's esophagus to cancer. These results suggest that obesity mediates its risk for esophageal adenocarcinoma primarily through the development of Barrett's esophagus. It should be noted, however, that we did not assess central obesity or intra-abdominal fat as separate risk factors. Although central obesity has been suggested to be a more important risk measure than BMI, there have been no data to show that central obesity increases the risk of transition from Barrett's esophagus to cancer.[27,28,29]

Case–control studies have shown a protective dose-dependent influence of fruit and vegetable intake on the development of esophageal adenocarcinoma,[6,30,31,32] which may be mediated through the effect of antioxidants.[33] Our results confirm prior studies in showing its lack of influence on the development of GERD.[34] In contrast with a prior study by Kubo et al.[35] we did not find that a high intake of fruit and vegetables protected against the development of Barrett's esophagus. The lack of effect in our study may be related to fewer study participants and a different assessment of dietary habits. The study by Kubo et al. asked for dietary habits within the year before Barrett diagnosis, while our study assessed dietary habits at age 40 introducing the higher potential of recall bias. With respect to the progression to cancer, our study suggests that a high fruit and vegetable intake may have a protective effect against the development of cancer in patients with Barrett's esophagus.

H. pylori infection has been reported to decrease the risk of Barrett's esophagus[36] and its progression to cancer, possibly as a result of reduced acid secretion in H. pylori-associated corpus predominant gastritis.[37] Although our results did not reveal a statistically significant association, we observed an overall trend suggesting some protective influence on both the progression to Barrett's esophagus and to cancer.

Our study has several limitations. A large number of the initially selected participants could not be contacted and included into the final study population (Figure 1). Although we could not gather sufficiently detailed information to assess whether there were any substantial differences between participants and non-participants, the basic demographic characteristics (age, gender, BMI, and reflux symptoms) of our study population were similar to those reported in previous cohort or population-based studies,[24,25,38] except for a slightly larger proportion of men in our GERD group than typically seen in this age group.[38] Furthermore, we asked about weight and dietary habits at age 40, introducing the possibility of a recall bias. The direction of its effect can go both ways, depending on how much a patient (particularly a cancer patient) believes the risk factor is associated with the diagnosis. In our analysis, recall bias may have increased the variation around the estimate.

It also needs to be mentioned that our no-GERD group does not represent an asymptomatic population. These patients had an endoscopy for other reasons. While they did not have esophagitis or typical reflux symptoms some patients in this group may have had atypical symptoms that would have been revealed by an abnormal pH study. This limitation likely leads to an underestimation of the observed effects.

Our results should be viewed as hypothesis generating. Our study examined the influences of various risk factors by comparing each two consecutive disease stages. Ideally, a large cohort of asymptomatic (no-GERD) patients should be followed over long time to understand transitions and risks along the assumed disease stages to cancer. Because such study would be impractical and extremely difficult to carry out, we have to rely on results from cross-sectional studies. It is plausible that seemingly intermediary stages (GERD and Barrett's metaplasia) may present themselves at the onset as the most severe and final stage of reflux disease.[40] ORs that were determined for different disease stages, therefore, indicate an epidemiologic risk association rather than a time-dependent progression among different disease stages. Although a time-dependent progression may be truly mediated by different sets of risk factors, it could also be that combinations of different risk factors or varying strengths of individual risk factors govern the outcome from the beginning onwards. According to the second possibility, such patterns of risk factors would function more like as a set of switches that lead toward its pre-determined end point from the onset and within a short time period. Our cross-sectional study design does not allow us to differentiate between these varying possibilities for the natural progression to cancer.

In general, the strength of association calculated for individual risk factors depends on the type of comparison groups and their individual sizes. Some risk factors may appear weak or insignificant when contrasted with the findings of previous investigators. However, when we compared esophageal adenocarcinoma patients with controls without GERD (rather than Barrett's esophagus), the pattern changed with most ORs increasing in size or changing their level of significance. Because of the high number of exclusions our control group without GERD turned out to be smaller than anticipated. A larger control population might have rendered a larger number of risk factors statistically significant.

In conclusion, our study shows that different sets of risk factors are associated with different disease stages in the development to esophageal adenocarcinoma. While some risk factors act predominantly on the initial development of reflux disease, others are associated with more advanced disease stages, such as Barrett's esophagus or esophageal adenocarcinoma. Our results suggest that different combinations or varying strengths of individual risk factors may govern the final outcome from the onset. Alternatively, it is also possible that a gradual progression through different stages of disease severity may be mediated by the consecutive and time-dependent action of varying risk factors.