The study, published in the journal Cancer Causes & Control, drew on the Multiethnic Cohort Study, a prospective population-based investigation that enrolled more than 215,000 residents of Hawaii and California between 1993 and 1996. Participants aged 45 to 75 completed detailed baseline questionnaires covering diet, medications, occupation, physical activity, family history of cancer, and immigration history. The research team, led by Haejin In of the Rutgers Cancer Institute together with Meira Epplein of Duke University, restricted their analysis to individuals who identified as Japanese or non-Hispanic White, ultimately following 94,232 participants for a median of just over 20 years. During that time, 716 primary gastric cancer cases were identified through statewide SEER cancer registries in Hawaii and California.

The central innovation of the study was its use of generational immigration status, a classification built from the birthplaces of the participants and their parents. Japanese descendants were divided into four groups: those born in Japan themselves, US-born individuals with both parents born in Japan, US-born individuals with one parent born in Japan, and US-born individuals whose parents were both born in the United States. Non-Hispanic White participants served as the reference group. The researchers then used Cox proportional hazards models, with age as the time scale, to estimate how gastric cancer risk varied across these groups while adjusting for a wide range of demographic, dietary, and lifestyle variables, including education, body mass index, smoking, alcohol consumption, fruit and sodium intake, diabetes history, and family history of the disease.

The results were unambiguous. Japan-born participants had a 4.62-fold higher hazard of developing gastric cancer compared with non-Hispanic Whites. Among US-born participants, the risk declined in a clear gradient: those with both parents born in Japan had a 3.66-fold higher hazard, those with one Japan-born parent had a 3.39-fold higher hazard, and those whose parents were both born in the United States had a 2.57-fold higher hazard. The trend across generations was highly statistically significant, with a p-value for trend below 0.0001. When the two second-generation groups were combined, their pooled risk was 3.59 times that of non-Hispanic Whites, confirming a progressive attenuation of risk across generations rather than a sharp drop confined to the transition from the first to the second generation.

Perhaps the most striking finding emerged when the researchers broke the results down by anatomical subsite. The generational association was driven almost entirely by non-cardia gastric cancer, the type that arises in the lower portions of the stomach and is most strongly linked to chronic infection with the bacterium Helicobacter pylori. First-generation Japanese descendants had a nearly six-fold increased risk of non-cardia cancer compared with non-Hispanic Whites, second-generation descendants a 4.5-fold increased risk, and third-generation descendants roughly a three-fold increased risk. When the analysis was restricted to precisely defined non-cardia subsites, the associations grew even stronger, with the first-generation hazard ratio reaching 7.13. In contrast, the associations for cardia cancer, the type arising near the junction with the esophagus, were weaker and showed no statistically significant linear trend across generations.

What could explain such a persistent, generational gradient? The leading hypothesis centers on Helicobacter pylori, the bacterium considered one of the major drivers of gastric cancer development. People who migrate from countries where H. pylori is highly prevalent are more likely to have acquired the infection during childhood, a critical window when chronic gastric inflammation can become established and shape long-term susceptibility to carcinogenesis. Persistent childhood infection is also associated with more virulent bacterial strains, such as those carrying the cagA and vacAs1m1 genes, which may disrupt early-life microbiota and immune regulation. Although H. pylori prevalence is known to decline across generations within immigrant families, the fact that third-generation Japanese descendants still face nearly triple the risk of non-Hispanic Whites suggests the bacterium alone cannot account for the full picture.

The authors point to a constellation of additional mechanisms that may be transmitted across generations or interact with the post-migration environment. Dietary acculturation is one candidate: first-generation immigrants in the cohort consumed the most sodium, a known risk factor for non-cardia gastric cancer, with intake declining in each subsequent generation, while body mass index rose across generations. Prior research has shown that a Westernized diet and lifestyle alter the gut microbiome, causing US-associated microbial communities to replace native Asian ones, with effects that intensify the longer a person lives in the United States. Studies of Japanese descendants have also documented a preference for Japanese food in the second generation giving way to Western foods in the third, alongside microbiome alterations linked to insulin resistance. Epigenetic programming may play a role as well; a cross-sectional study of Italian migrants found differences in DNA methylation compared with non-migrants even after adjusting for diet and lifestyle. Genetic variation adds yet another layer, with certain polymorphisms in inflammatory pathway genes, such as interleukin-1, reported among Japanese individuals to modify gastric cancer risk independently of H. pylori infection or smoking.

Notably, the researchers found no evidence that established risk factors, including family history of gastric cancer, education, body mass index, smoking status, and sodium intake, modified the association between generational status and cancer risk. This suggests that the generational gradient reflects distinct environmental, biological, or early-life influences rather than simple differences in these conventional exposures. The findings echo earlier work, including a landmark study by Kamineni and colleagues showing that first-generation Japanese immigrants to the United States had higher gastric cancer incidence than US Whites but lower rates than people in Japan, and a Hawaiian study by Kolonel and colleagues finding higher incidence among Japan-born than US-born Japanese residents. The new study extends this evidence with individual-level exposure data and more than two decades of prospective follow-up, advantages that registry-based studies could not provide.

The clinical implications are significant. Because the United States does not operate population-wide gastric cancer screening programs, unlike Japan and South Korea, identifying which groups stand to benefit most is essential. The marked differences in risk across immigrant generations, even within a single self-identified ethnic group, suggest that generational status could serve as a practical marker for risk-stratified screening. Prior studies have demonstrated the cost-effectiveness of gastric cancer screening among high-risk US groups, including Asian Americans undergoing upper endoscopy, opportunistic endoscopic examination during routine colonoscopy, and non-invasive biomarker approaches such as serum pepsinogen testing. Concentrating surveillance resources on the most at-risk generations could improve both the effectiveness and the accessibility of gastric cancer prevention for minority populations.

The study has limitations worth noting. Self-reported race and ethnicity may not perfectly reflect genetic ancestry, information on H. pylori infection itself was not available, and the cohort was drawn from Hawaii and California, regions with long-established Japanese American communities whose migration and acculturation patterns may differ from those of Japanese descendants elsewhere in the country. The authors call for replication in geographically diverse cohorts with detailed data on age at migration and duration of US residence. Still, the core message stands: the legacy of where a family came from, and when, is written into cancer risk for at least three generations. Understanding these transgenerational determinants, the authors argue, could reshape how the United States screens, detects, and prevents one of the world’s deadliest cancers.

Subject of Research: Generational immigration status and gastric cancer incidence among Japanese descendants in the Multiethnic Cohort Study

Article Title: Generational immigration status and gastric cancer incidence among Japanese descendants: an analysis of the Multiethnic Cohort (MEC) Study

Article References: In, H., De la Torre-Cisneros, K., Adams, A., Chen, C., Rana, B., Wilkens, L., & Epplein, M. (2026). Generational immigration status and gastric cancer incidence among Japanese descendants: an analysis of the Multiethnic Cohort (MEC) Study. Cancer Causes & Control, 37(10), Article 168. https://doi.org/10.1007/s10552-026-02250-0

Image Credits: AI Generated

DOI: 10.1007/s10552-026-02250-0

Keywords: gastric cancer, Japanese descendants, immigration generation, Multiethnic Cohort Study, Helicobacter pylori, non-cardia gastric cancer, health disparities, cancer screening, epidemiology, dietary acculturation, Cox proportional hazards, cancer risk

The Multiethnic Cohort, established in 1993–1996 in Hawaii and Los Angeles, was designed precisely to overcome a weakness that has plagued nutritional epidemiology: study populations dominated by White, well-educated participants whose habits and biology may not generalize. Colorectal cancer is the third most common cancer worldwide and the second leading cause of cancer death in the United States, and the burden is unevenly distributed. Black men and women in the US experience roughly 40 percent and 24 percent higher colorectal cancer mortality than their White counterparts, while disaggregated data show elevated rates among Japanese and Native Hawaiian populations. At the same time, national dietary surveys show that coffee consumption itself varies sharply by group, with non-Hispanic White individuals averaging about 1.5 cups per day, Hispanic individuals about one cup, and Black individuals about half a cup. Any study that hopes to draw honest conclusions about coffee and cancer survival must therefore contend with the fact that who drinks coffee, how much, and in what form differs profoundly across communities.

The new analysis capitalized on the cohort’s structure. Participants completed a validated food frequency questionnaire at enrollment and again a decade later, between 2003 and 2008. The researchers identified 1,098 individuals who were free of colorectal cancer at baseline, were diagnosed with the disease before the follow-up survey, and survived long enough to complete that second questionnaire, which captured post-diagnosis intake of regular coffee, decaffeinated coffee, and cappuccino-style drinks across nine frequency categories ranging from never to four or more times daily. Mortality was tracked through linkage with SEER cancer registries, state vital records, and the National Death Index through December 31, 2019, yielding 628 deaths, of which 107 were attributable to colorectal cancer. The median time from diagnosis to the post-diagnosis survey was nearly five years, and the average follow-up after that survey exceeded ten years, giving the analysis unusual depth.

Methodologically, the team used multivariable Cox proportional hazards regression, the workhorse of survival analysis, to estimate hazard ratios adjusted for a carefully assembled battery of confounders. The fully adjusted models accounted for age, sex, race and ethnicity, body mass index, time from diagnosis to the dietary assessment, tumor site and SEER stage, educational attainment, marital status, smoking status, physical activity, a comorbidity index spanning conditions from diabetes to Parkinson’s disease, history of another cancer, alcohol intake, and omega-3 fatty acid consumption. A third model mutually adjusted caffeinated and decaffeinated coffee against each other to disentangle their independent effects. Sensitivity analyses excluding participants who died within one or two years of the survey, those with distant-stage disease, those who did not undergo surgical resection, and records with missing data all left the core results intact, addressing the ever-present specter of reverse causation, in which sicker patients simply drink less coffee.

The central result was a null that speaks volumes. Total and caffeinated coffee intake showed no statistically significant association with all-cause or colorectal cancer-specific mortality in the overall cohort, although the association for total coffee and all-cause mortality hovered at the edge of significance, with a hazard ratio of 0.92 per additional cup per day and a p-value of 0.057. Decaffeinated coffee, by contrast, was clearly associated with survival: each additional cup per day carried a hazard ratio of 0.82 for all-cause mortality, and participants drinking at least one cup daily had a 29 percent lower risk compared with non-drinkers. The finding persisted even after controlling for caffeinated coffee consumption, suggesting that compounds other than caffeine, or the behaviors of people who choose decaf, may be at work.

The caffeine analysis added a twist that cuts against a simple decaf story. When the researchers aggregated caffeine from every source, including coffee, black tea, green tea, regular soda, and diet soda, participants in the top quartile of intake had a 22 percent lower risk of all-cause mortality and a 52 percent lower risk of colorectal cancer-specific mortality compared with those in the bottom quartile. The authors suggest that measuring caffeine across all dietary sources may have captured a wider range of exposure than coffee alone, where the cohort’s relatively modest consumption levels, topping out well below the four or more cups per day examined in prior studies, may have limited the ability to detect an effect.

The stratified results are arguably the study’s most consequential contribution. Among White participants, who reported the highest coffee and caffeine intake in the cohort, both total and caffeinated coffee were associated with lower colorectal cancer-specific mortality, with hazard ratios of 0.48 and 0.45 per additional daily cup, and decaffeinated coffee was again linked to lower all-cause mortality. Among Native Hawaiian participants, total and caffeinated coffee were associated with lower all-cause mortality, with hazard ratios of 0.62 and 0.55 per cup per day, an observation never before reported for this group. No statistically significant associations emerged among Black, Japanese American, or Latino participants. Formal tests for interaction did not reach significance, meaning the differences between groups could reflect chance, but the pattern is biologically and culturally plausible. Coffee strength, preparation methods, and additives such as sugar and cream vary across cultures, and coffee interacts with diet-dependent gut microbiome compositions and insulin dynamics in ways that could differ across populations.

Mechanistically, the plausibility of a coffee-cancer survival link rests on a substantial experimental foundation. In vitro and animal studies have shown that coffee and its constituents are antiproliferative, pro-apoptotic, antioxidant, and anti-inflammatory, capable of slowing the division of cancer cells and promoting their self-destruction. Coffee consumption has also been shown to reshape the gut microbiome and improve insulin sensitivity, both of which are implicated in colorectal carcinogenesis and progression, since hyperinsulinemia fuels tumor growth and inflammation accelerates it. Caffeine specifically may act by enhancing insulin sensitivity and by exerting anti-inflammatory effects that inhibit carcinogenesis. Prior clinical studies lend support: among 953 patients with stage III colon cancer in the CALGB 89803 trial, two or more cups of coffee daily were associated with longer disease-free survival, and among 1,599 patients with stage I–III colorectal cancer in the Nurses’ Health Study and Health Professionals Follow-up Study, four or more cups daily were linked to lower cancer-specific and overall mortality. A 2024 Global Cancer Update Programme meta-analysis graded the evidence as limited-suggestive, and a Dutch study of 1,719 patients reported a U-shaped relationship with the lowest mortality at four cups per day.

The authors are candid about limitations. Coffee consumption in this cohort was modest, few participants drank at the extremes where earlier studies found effects, and the very high consumption seen in some prior cohorts may itself signal healthier, more resilient patients, a bias this more moderate sample partly avoids. Some racial and ethnic subgroups were small, survival bias is unavoidable in a design requiring participants to live long enough to complete a post-diagnosis survey, residual confounding cannot be excluded in any observational study, and the multiplicity of exposures and outcomes examined raises the possibility of chance findings. Coffee recurrence data were unavailable. Still, the study is the first to interrogate the coffee-survival question in a genuinely multiracial population, and its message to patients and clinicians is refreshingly grounded: coffee after a colorectal cancer diagnosis, whether caffeinated or not, does not appear to be harmful, decaffeinated coffee and caffeine may both be associated with survival advantages, and future research should focus less on coffee as a monolith and more on caffeine itself and on how dietary factors interact with the biology of diverse populations.

Subject of Research: Association of post-diagnosis coffee and caffeine intake with mortality among colorectal cancer patients in a racially diverse cohort

Article Title: Association of coffee intake with risk of all-cause and colorectal cancer-specific mortality among individuals with colorectal cancer in the Multiethnic Cohort Study

Article References: Benson, K. R. K., Tolstykh, I., Chan, J. M., Piawah, S., Wilkens, L. R., Park, S.-Y., Kenfield, S. A., Le Marchand, L., Haiman, C. A., Cheng, I., & Van Blarigan, E. L. (2026). Association of coffee intake with risk of all-cause and colorectal cancer-specific mortality among individuals with colorectal cancer in the Multiethnic Cohort Study. Cancer Causes & Control, 37(10), Article 157. https://doi.org/10.1007/s10552-026-02246-w

Image Credits: AI Generated

DOI: 10.1007/s10552-026-02246-w

Keywords: colorectal cancer, coffee, caffeine, decaffeinated coffee, cancer survival, mortality, Multiethnic Cohort Study, nutritional epidemiology, cancer disparities, insulin sensitivity, gut microbiome, prospective cohort