The cervical microbiota and diet assessed by pyrosequencing and a food-frequency questionnaire, respectively, of 65 women with CIN and 72 control women were used in this study. Principal component analysis and cluster analysis were used to identify dietary patterns and microbiome community types, respectively. The association between dietary pattern and CIN risk was assessed using multivariable logistic regression analysis. The combined effect of dietary pattern and microbiome on CIN risk was determined using relative excess risk due to interaction (RERI) and synergy index (S).
Two dietary patterns and four community types were identified: prudent diet characterized by higher intake of vegetables and fishes; semi-Western diet characterized by higher intake of bread, dairy products, eggs, and soft drinks and relatively higher fat intake ratio; and Lactobacillus crispatus–, L. iners–, Atopobium vaginae–, and Prevotella bivia–dominant types. The high-scoring group of participants with a semi-Western diet had a higher risk of CIN (odds ratio [OR] 3.44, 95% confidence interval [CI] 1.11–10.7, p = 0.03), compared with the low or medium-scoring group of those with a semi-Western diet. L. iners–dominant (OR 6.39, 95% CI 1.52–26.7, p = 0.01) and A. vaginae–dominant (OR 4.99, 95% CI 1.17–21.3, p = 0.03) dominant types had a higher risk of CIN, compared with the L. crispatus–dominant type. The synergistic effect of semi-Western diet and A. vaginae–dominant type on CIN risk was observed (OR 20.8, 95% CI 2.21–195.6, p = 0.01, RERI/S 9.64/1.96).
Our findings suggest that semi-Western diet and its combination with A. vaginae–dominant microflora may represent an important risk factor for cervical neoplasia.