Diet and oral premalignancy in female south Indian tobacco and betel chewers: a case-control study.

The diets of 158 tobacco/betel quid-chewing women diagnosed with oral premalignant lesions and 155 quid-chewing but lesion-free controls, frequency matched for age, tobacco/betel habits, and socioeconomic status, were assessed using a food frequency survey. Index scores generated from the food frequency survey indicated that the mean levels of consumption for foods of animal origin (p < 0.001), total vegetables and fruit (p = 0.001), vegetables alone (p = 0.006), fruits alone (p = 0.006), and green leafy vegetables (p = 0.015) were significantly lower in cases than in controls. The mean index score for cobalamin (vitamin B12) was lower in cases with a borderline significance (p = 0.05), whereas the indexes for folate and carotene were not significantly different. The analysis of index scores estimating the number of 100-g servings per week of foods of animal origin [meat, eggs, milk, curd (yogurt), fish] consumed revealed that women who ate fewer servings were more likely to have premalignant lesions than those who ate more animal foods [odds ratio (OR) 3.38, 95% confidence interval (CI) 2.07-5.54, p = 0.001]. The risk for low consumption of vegetables was not as significant as that for foods of animal origin. However, those eating low levels of vegetables and low levels of foods of animal origin were at the greatest risk for lesions (OR 5.38, 95% CI 1.72-22.17, p < 0.05). In South Indian female tobacco/betel chewers, a diet deficient in foods of animal origin appears to be a more significant risk factor for oral premalignancy than is a diet deficient in fruits and vegetables.