|| Checking for direct PDF access through Ovid
To examine the occurrence of bacterial vaginosis (BV) over the menstrual cycle and in relation to menstrual protection materials and sexual intercourse in a rural African setting.Married, regularly menstruating female volunteers were asked to collect self administered swabs on alternate days through four menstrual cycles. BV was assessed using Nugent scores. Menstruation and reported sexual intercourse data were recorded contemporaneously. A crossover design comparing traditional and modern menstrual protection methods was incorporated. Multivariate logistic regression was used to examine associations with BV.30 women completed four menstrual cycles in the study. Completeness and validity of data from the self administered swabs was high. Greater frequencies of BV were found for all women in the second week of the menstrual cycle relative to days 14+, and markedly higher frequencies of BV were found in the first week in women with infrequent BV. BV was (non-significantly) more frequent when modern pads were used compared with traditional cloths. No association was found between BV and intercourse reported in the previous 4 days; or between the frequency of reported intercourse in one menstrual cycle and BV in either the same menstrual cycle or the next.Similar transient fluctuations over the menstrual cycle were found to those in industrialised countries. We found no evidence that sexual intercourse was associated with increased frequency of BV. Our data do not support hypotheses that menstrual hygiene materials might explain the high prevalences of BV found in sub-Saharan Africa compared to industrialised countries.