Stomach cancer incidence in a cohort of fishermen in Singapore.

PHOON WOoStomach cancer incidence in a cohort of fishermen in Singapore. Scand J Work Environ Health 13 (1987) 524-526. A retrospective cohort study of 279 Chinese fisherm en in Singapore was undertaken to examine the possibility of an elevated incidence of stomach cancer. The fishermen as a group had a greater than twofold excess in stomach cancer relative to Chinese men of similar age and over the same calendar years in Singapore, but the difference was not statistically significant. On further examination it was revealed that the four stomach cancer cases were observed among the subgroup of divers. The standardized incidence ratio for the divers was 4.3, which was significantly greater than unity. There was no evidence of an increased risk for cancer of any other site (lung, nasopharynx and liver) in this cohort. Dietary factors are suggested as a possible explan ation for the observation of an increased risk for stomach cancer among fishermen, but larger studies are required 10 test this hypothesis.

In Singap ore stomach cancer rank s second on the list of canc ers for men and fourth for women (20). In add ition there is a high incidence of stomach can cer in Japan and other parts of Asia a nd South America, but in North America and Europe the incidence is low and is decreasing (22,23). Stud ies indicate a reasonably co nsistent association between sto mac h cancer and various aspect s of environment and life-style, particularly dietary int ake . For instance , it has been do cumented that a decl ine in the inciden ce of sto mach cancer ha s occurred over several generations among Japanese migrants to the United States, the indication being that environmental or life-style relat ed factors are invol ved (12). Dietar y factors such as sa lt intake in food (10), salted or pre served foods , and smoked foods (3,7,9,II,24) and oth er personal habit s (8) have been shown to be correlated with stomac h cancer.
A relationship between occupation and sto mach cancer is not well establi shed. A slight excess risk for selected occupational groups such as min ers (5,14), asbestos workers (6,19), rubber workers (15), and metal product workers (13,17) has been noted, however. Th ese studies ha ve sugges ted a po ssible link between stoma ch cancer and var ious factors in th e work environment. Non e of these studies incriminat ed a possible diet ar y factor as bein g respon sible for the observed ass ociatio n. The pre sent study was initiated to examine the incidence of stomach canc er among diver fishe rm en in Sing apore becau se of a clinical

Materials and methods
The study cohort com prised 320 Chinese registered as members of an association of fishermen in Singapore. The se men fish (17) in the South China Sea and the Indian Ocea n about 300 to I 500 km from Singapore. The subjects were registe red as fishermen with the associat ion during the period between 1947 and 1980, and the cohort was monitored until Au gust 1985. The fishermen so identifi ed were tr aced, and information was obt ained on the ir age, date of commencement of wo rk , durat ion o f em ployme nt if no lon ger working as a fisherman , nature of work (diver or nondi ver) , and illnesses. Incid ence dat a were obtained fr om the asso ciat ion registe r, medical records, and interviews of fam ily member s during the survey. The diagnoses for cancer were verified from the database of the Singapore Cancer Registry. In formation was available on 279 (180 divers a nd 99 nondi vers) of the 320 fishe rmen , the response rate bein g 87 0/ 0. The total per sonyears o f follow-up was 3 548 for the divers (average o f 19.7 years per diver ) and 2 050 for the nondivers (average o f 20.7 yea rs per nondi ver) .
A mod ified life-table method (16) was employed in which person-years of observation were cumulated for each exposure dur ation in five-year age gro ups and five-year calenda r inte rva ls. The se dat a were then used to genera te expected numbers o f site-specific ca ncer incidence cases based on the age and calendar-spec ific incid ence rates of C hinese men in Singa pore. Th e standardized incidence rati o (SIR) was co mput ed as the ratio of the ob served number of cases to the expected number of ca ses of site-specific cancer. Statistical significan ce for the SIR was determined under the assum ption of a Poisson distribution for the ob served number of cancer cases, as described by Bailar & Ede rer (I) . All the statistical computations were performed with the man-years FORTRAN program (4). Table I shows the observed and expected numbers of sto mach cancer among the fishermen and the ir corresponding SIR values. The SIR for all the subjects was 2.3 . Thus the fishermen as a group had a greater than twofold increase in risk for stomach cancer relative to Chinese men of similar age and over the same calendar years in Singapore, but the SIR was not close to statistical significance (95 lifo confidence interval 0.6-5.8). However, it is noteworthy that all four stomach cancer cases occu rred among the diver s, for an SIR of 4.3, which was significantly greater than un ity (95 % con fidence interval 1.2-11.0).

Results
All four stomach cancer cases occurred in men between the ages of 35 and 59 years. One case was identified with in five years of the man becoming a fisherman , one between five and ten years , and two after more than ten years. As of August 1985, the mean age of the divers was 48.1 (SO 10.6) years, and that of the nondivers was 51.3 (SO 12.8) years.
In addition to the four stomach cancer cases , there were two cases of lung can cer, one of nasopharyngeal cancer , and on e of liver cancer among the fishermen. The SIR of each cancer site for all the subjects was lower than unity, although the difference was not statistically significant. It was again noted that these cancers occurred only in the diver s, but none of their SIR values was significa ntly in excess of unit y.

Discussion
This study showed that Chinese fishermen in Singapore have a greater than twofold likelihood of developing stomach cancer in comparison to the co rresponding general population. When th is group was divided into diver s and nondivers , it was observed that all o f the stomach cancer cases occurred only among the diver s, with a greater than fourfold risk relative to the general population (table I). It is no teworthy that there was no evidence of an increased risk for cancer of other sites amo ng this group of fishermen in comparison to the general population.
A possible explanation for the ob ser vation o f a higher incidence of stomach cancer in thi s group could be associated with the known practice of these fishermen having an increased intake of salted fish or "rotted" fish or both in their diet. The process of salting fish increases the content of nitrates in the diet. Furthermore, th is particular type o f "rotted" fish could result in bacterial contamination o f the stomach. The nitrates are subsequently reduced to nitrites and the bacterial flora could lead to the conversion of the nitrites to carcinogenic N-nitroso compounds. A similar mechanism has been postulated for stomach cancer associated with iron deficiency and pernicious anemi a (2,18). In addition to this possible carcinogenic mechanism, the situation is further aggravated by the fact that the fishermen's diet lacks fresh vegetables, a source of vitamin A (21, 25). It has been recognized that the presence of vitamin A in the diet may have a protective effect on the gastric mucosa and thereby may prevent the occurrence of stomach cancer. The present study points toward an elevated incidence of stomach cancer for Chinese fishermen in Singapore, particularly among divers. The possibility of an associated dietary factor is considered a plausible explanation for the observation. We plan to investigate this possibility in the future by undertaking a comparative study of the dietary habits and life-s tyle of fishermen and non fishermen.