Diet and Human Leukemia: An Analysis of International Data

Abstract

Background. There has been little direct exploration of the relationships between dietary factors and human leukemia; however, a number of literature reports from animal studies and correlation analyses between countries suggest that diet can influence leukemia risk. Methods. Statistical analyses of international food supply data and leukemia incidence data were performed to define better the diet–human leukemia relationship. Incidence data for lymphoid, myeloid, and total leukemia (all leukemia subtypes combined) from 24 countries with reliable cancer registry data were regressed with estimates of per capita disappearance of macronutrients and alcohol, as well as gross national product and average height. Results. Simple correlation analyses showed that the dietary variables were associated with leukemia in a gender- and site-specific fashion. Males consistently had higher correlations than females. The strongest correlations were found between total calorie intake and both lymphoid and total leukemia incidence, especially among males. Myeloid leukemia in either gender was most strongly associated with gross national product. To control for potential confounding, multiple regression analyses were performed, with all regression models adjusted for gross national product, height, and dietary covariates. Total calorie supply was the only significant explanatory variable for the international variation in lymphoid and total leukemia in these analyses. The calorie–leukemia association was stronger among males than among females. No significant association was observed between myeloid leukemia and any of the dietary variables studied, after adjusting for height and gross national product. Conclusion. The findings from this rigorous analysis of international data strengthen and expand the hypothesis based on previous simple correlation analyses and animal experiments that an underlying biological relationship exists between diet, particularly energy intake, and international variations in the incidence of certain types of human leukemia. Possible mechanisms for the calorie–leukemia associations are discussed.