Dietary habits and vaginal environment: can a beneficial impact be expected?

Abstract

Introduction: In reproductive-aged women, a vaginal microbiota dominated by several Lactobacillus species is crucial for maintaining vaginal health. Among the various factors affecting the composition of the vaginal ecosystem, the impact of dietary habits has rarely been explored. Thus, in this cross-sectional study, we assessed the role of macronutrient intake on the vaginal microbiota in a cohort of 113 young women, independently from potential confounders.

Methods: For each subject, we characterized (i) the vaginal bacterial community-state type (CST) by 16S rRNA gene profiling, (ii) the vagina lmetabolic profile by 1H-NMR spectroscopy, and (iii) the energy, nutrient and alcohol intake through a validated food frequency questionnaire.

Results: We found that the increase in animal protein intake, mainly derived from red and processed meat, was positively associated with the dysbiotic condition of CST IV and, similarly, alcohol consumption was significantly associated with the levels of Gardnerella spp. and Ureaplasma spp. On the other hand, we noticed a beneficial effect of a-linolenic acid, with its increase inversely associated with CST III, dominated by the 'less-protective' species Lactobacillus iners. Moreover, linolenic acid was related to the abundance of Lactobacillus crispatus, in turn related tovarious vaginal metabolites such as 4-hydroxyphenyllactate and several amino acids. Total carbohydrates, vegetable proteins, total fiber, and starch were negatively correlated with Gardnerella spp.

Discussion: We highlighted that specific dietary habits (i.e., reduced consumption of alcohol and animal proteins, higher intake of linolenic acid) can have a beneficial impact on the vaginal environment, through the maintenance of a microbiota mainly dominated by 'protective' Lactobacillus species.

Keywords: diet; macronutrients; metabolome; vaginal microbiota; women’s health.

Figure 1

(left) Histograms of the average relative abundance (%) of bacteria over CST. (right) Stacked bars of the average relative abundance (%), grouping together samples classified in CSTs I, II and V. Only the main taxa (average abundance >0.4% over all the samples) are represented here, with Lactobacillus genus further subclassified to species level. Less abundant taxa are grouped in the “Other” category.

Figure 2

Multinomial logistic regression coefficient (b-coefficient) and standard errors (SE) of CST in relation to the macronutrient balances, total energy, fiber and alcohol intake (n=113). The model also included terms for age, BMI, marital status, and hormonal contraception use. Asterisks (*) indicate statistical significance of the model: * p<0.1; ** p<0.05, *** p<0.01.

Figure 3

Network representing the Spearman correlation values among bacterial, metabolite and macronutrient abundances. Bacterial taxa are depicted as red circles, metabolites as cyan squares and macronutrients as green diamonds. Edges thickness is proportional to the correlation coefficient. Only significant (p<0.05) correlations are represented. Bacteria, metabolites and macronutrients are clustered according to the indicator species analysis over the three CST groups considered.