Microbiology of vaginal discharge with emphasis on gardnerella vaginalis.

Abstract

The microbiological aetiology of vaginal discharge was studied in 208 women attending various outpatient clinics at King Edward VIII Hospital. Specimens from the lower genital tract were collected for microscopy and culture. Vaginal wet smear examination, amine liberation test and vaginal pH estimation were performed and assessed for their reliability for the rapid diagnosis of vaginal infections. Vaginal and endo-cervical infections were present in 163 (78,4%) women. G. vaginalis (65,4%), T. vaginalis (37,9%), genital yeasts (37,0%), M. hominis (59,6%), g. urealyticum (48,1%), anaerobic bacteria (32,6%), N. gonorrhoeae (11,1%), f. trachomatis (22,1%) and Herpes simplex virus (0,9%) were detected. Of the 104 women in • whom vaginal infections were detected, bacterial vaginosis was present as the sole infection in 32 (22,2%), I. vaginalis in 35 (24,3%) and C. albicans in 23 (15,9%). Bacterial vaginosis occurred concurrently with T. vaginalis and f. albicans in 24 (16,5%) and 11 (7,5%) women respectively; whilstT. vaginalis and f. albicans occurred concurrently in 14 (9,7%) women. In 6 (4,1%) women all three infections were present. No vaginal or endo-cervical pathogens were detected in 45 (21,6%) women. Women with bacterial vaginosis were found to be significantly colonised with G. vaginalis, M, hominis, anaerobic bacteria and curved Gram-negative bacilli (p < 0,05). Vaginal wet smear microscopy detected T.. vaginalis in 29% and "clue" cells in 41,3% of smears. The presence of "clue" cells (91,8%) and a positive amine test (76,7%) was significantly associated with bacterial vaginosis. Although a raised vaginal pH was also significantly associated with bacterial vaginosis, this test was less specific (65,2%) than "clue" cells (85,9%) and the amine test (95,5%). The vaginal Gram stain, as performed in this study, was found to be unreliable for the detection of "clue" cells. G. vaginalis biotypes 1 and 5 were significantly associated with bacterial vaginosis, however the serotyping scheme did not distinguish between strains isolated from women with and without bacterial vaginosis. The antimicrobial susceptibility pattern of 93 strains of G. vaginalis was not typical of either Gram-positive or Gram-negative bacteria. Serological tests revealed reactive syphilis serology in 47 (22,6%) and the presence of hepatitis B surface antigen in 16 (7,7%) women. Antibody to human immunodeficiency virus was detected in 4 (1,9%) women attending the colposcopy clinic. This study clearly demonstrates the high prevalence of vaginal and/or endo-cervical infections in women locally, the majority of whom were asymptomatic. The high frequency of concurrent infections is of concern and there is a need for the recognition, and appropriate management of such infections.