Does high gastric cancer risk associated with low serum ferritin level reflect achlorhydria? An examination via cross-sectional study.

Abstract

With respect to the inverse association of serum ferritin level (SFL) with the risk of gastric cancer (GC) observed in some recent epidemiologic studies, possible mediation by achlorhydria as well as atrophic gastritis (AG), both of which are strongly associated with GC risk at not only the individual but also the population level, was examined in a cross‐sectional study of 634 men aged 40 to 49 years randomly selected from 5 populations in Okinawa, Iwate, Nagano, Akita and Tokyo. AG and achlorhydria were serologically diagnosed based on the criteria of pepsinogen (Pep) I level <70 ng/ml and Pep I/Pep II ratio <3.0, as described previously, and a serum gastrin level of over 140 pg/ml, respectively. In the results, while the mean SFL for all the subjects differed significantly by area, similar areal differences in SFL were also found even when only the non‐AG cases were considered. However, both of the above differences were eliminated with the exception of those between Okinawa and each of the other 4 areas, when adjustments were made for medical histories of diabetes mellitus, ulcers and liver disease, body mass index and γ‐glutamyltranspeptidase level. Therefore, no correlation among the 5 areas was observed between the adjusted areal mean SFLs and GC mortality in either case. However, in 17 (45%) achlorhydric cases identified among the 38 subjects with severe AG, which were further discriminated with stricter criteria (Pep I <30 ng/ml and Pep I/Pep II ratio <2.0), their averaged SFL was significantly lower than the values for the non‐AG cases (3.78 vs. 4.64 In(ng/ml), P<0.001) and mild/moderate AG cases (3.78 vs. 4.47 ln(ng/ml), P<0.01), while the differences were not significant with respect to the other 21 (55%) severe AG cases without achlorhydria. Thus, it is suggested that the negative association of SFL, for which no mechanism has been shown in cancer etiology, with GC risk is indirect, since both of them are associated primarily with achlorhydria found among the severe AG cases. Moreover, areal mean SFLs, which are not determined only on the basis of the prevalence of achlorhydric AG cases, could not be related to GC mortality rates, suggesting that no direct association exists at the population level either.