Vaginal colonization with Escherichia coli in healthy women. Determination of relative risks by quantitative culture and multivariate statistical analysis.

The rate of vaginal colonization with Escherichia coli in 495 healthy women was 12% in a prospective study with use of selective media and semiquantitative culture techniques. Computer-assisted multivariate analysis revealed that vaginal E. coli was significantly associated with the menstrual phase of the cycle, prior use of antibiotics, use of diaphragm or cervical cap for contraception, history of previous urinary tract infection, and coisolation of Staphylococcus aureus that was positive for the toxic shock syndrome toxin-1 (p less than 0.05, multiple stepwise logistic regression analysis). No significant association was observed with tampon use or brand, other contraceptive methods, sexual activity, genital symptoms, recent vaginal infection, or other personal habits. Quantitative cultures obtained sequentially throughout the menstrual cycle in 12 unselected women confirmed higher E. coli counts in menstrual or midcycle samples compared to paired premenstrual specimens (p less than 0.05, Wilcoxon paired rank sign test). These data emphasize the hormonal and other host determinants in vaginal colonization by E. coli and may explain the high rate of vaginal E. coli (64%), in addition to toxicogenic S. aureus, in acute toxic shock syndrome and the higher incidence of urinary tract infection in women with diaphragm or cervical cap for contraception compared to other contraceptive methods.