Bacterial Vaginosis Workup

Updated: Oct 25, 2018
  • Author: Philippe H Girerd, MD; Chief Editor: Michel E Rivlin, MD  more...
  • Print
Workup

Laboratory Studies

Clinical diagnosis of bacterial vaginosis (BV) relies on history, vaginal examination, and microscopic examination (see the Table under "Diagnostic tools" for a summary of the differential diagnoses). Emerging data support sending vaginal cultures in recalcitrant cases. [14]

Obtain historical information regarding the patient's symptoms and nature of the discharge. Patients may report the following:

  • Most patients experience a vaginal malodor and/or an increase in vaginal discharge.

  • Vulvar irritation may be present or absent.

  • Dysuria, dyspareunia, and abdominal pain are lacking.

Vaginal examination

Findings on vaginal examination include the following:

  • Typical BV discharge characteristics

  • Lack of significant vulvovaginal inflammation

Microscopic examination of the discharge

Demonstrating three of the following four Amsel's criteria is considered necessary to diagnose BV most accurately [2] :

  • Demonstration of clue cells on a saline smear is the most specific criterion for diagnosing BV. Clue cells are vaginal epithelial cells that have bacteria adherent to their surfaces. The edges of the squamous epithelial cells, which normally have a sharply defined cell border, become studded with bacteria. The epithelial cells appear to be peppered with coccobacilli.

  • A pH greater than 4.5 indicates infection, and pH may be elevated in up to 90% of patients with BV.

  • Characteristic discharge appearance is thin, gray, and homogeneous.

  • The whiff test may be positive in up to 70% of BV patients. This test is performed by placing a drop of 10% KOH on the speculum after the vaginal examination or mixing vaginal fluid with a drop of KOH on a microscope slide. The KOH, by virtue of its alkaline properties, causes the release of volatile amines from the vaginal fluid. The amines are products of anaerobic bacterial metabolism.

The vaginal discharge of patients with BV is notable for its lack of polymorphonuclear leukocytes (PMNs), typically 1 or less than 1 PMN per vaginal epithelial cell.

Diagnosing BV accurately is more difficult when a coinfection is present. However, finding an increase in the number of PMNs per epithelial cell may lead the clinician to consider BV as a possibility.

Microscopic evaluation of the bacteria flora

The bacterial flora may be examined microscopically for evidence of changes in the overall bacterial predominance. The healthy vagina has a predominance of lactobacilli (large gram-positive rods). The flora of a patient with BV changes to become dominated by coccobacilli, reflecting an increase in the growth of Gardnerella vaginalis and other anaerobes.

Many use Nugent's criteria to quantify or grade bacteria via Gram stain of vaginal samples. In brief, Nugent's criteria evaluate 3 types of bacteria via Gram stain: Lactobacillus, Bacteroides/Gardnerella, and Mobiluncus. They are each graded on a scale of 1-4 (1+ is < 1 cell per field, 2+ is 1-5 cells per field, 3+ is 6-30 cells per field, and 4+ is >30 cells per field). In this system, Lactobacillus and Bacteroides/Gardnerella are given scores between 0-4 but Mobiluncus is only graded from 0-2. Total scores are then calculated and used as follows: 0-3 (Normal), 4-6 (intermediate bacterial count), and 7-10 (bacterial vaginosis). [15]

Vaginal cultures

Obtaining routine vaginal cultures in patients with BV has no utility, because this is a polymicrobial infection and some women may have asymptomatic carriage of G vaginalis organisms. Although G vaginalis has been demonstrated to grow in up to 100% of vaginal cultures of women with BV, it has also been cultured in up to 70% of asymptomatic women. However, obtaining cultures to exclude other infectious etiologies (eg, Trichomonas species, C trachomatis, N gonorrhoeae) is appropriate. In recurrent cases that have not resolved with standard regimens, cultures may be appropriate.

Diagnostic tools

FemExam [16] has been shown to have variable sensitivity (38-90%) and specificity (12.5-97%) depending on the study and study population (pre- vs postmenopausal). [17] The test provides a result in 2 minutes.

Other diagnostic tools use DNA extracted from samples of vaginal fluid using Instagene Matrix (Bio-Rad Laboratories, Hercules, CA). Results of these tests often show a difference in the bacterial components in women with BV who are postmenopausal compared with those who are premenopausal. This information is certainly helpful, particularly in treatment failures to better tailor antibiotic treatment. [17]

Table. Differential Diagnosis of the Vaginitides (Open Table in a new window)

Clinical Elements

Bacterial Vaginosis

Trichomoniasis

Vaginal Candidiasis

Symptoms

Vaginal odor

+

+/-

-

Vaginal discharge

Thin, gray, homogenous

Green-yellow

White, curdlike

Vulvar irritation

+/-

+

+

Dyspareunia

-

+

-

Signs

Vulvar erythema

-

+/-

+/-

Bubbles in vaginal fluid

+

+/-

-

Strawberry cervix

-

+/-

-

Microscopy

Saline wet mount

Clue cells

+

-

-

Motile protozoa

-

+

-

KOH test

Pseudohyphae

-

-

+

Whiff test

+

+/-

-

pH

>4.5

>4.5

< 4.5

Next:

Other Tests

Recent research indicates that genital cytokine profiles may be used as a biomarker of sexually transmitted infections (STIs) and BV to identify women with asymptomatic, treatable infections, which may allow improvement in treatment of these conditions and possibly reduce the risk of HIV infection in high-risk women. [18]  Masson et al found that that evaluating clinical signs/symptoms in conjunction with evaluation of IL-1beta and IP-10 as biomarkers of genital inflammation is more sensitive and specific than clinical signs/symptoms alone (increased IL-1beta and decreased IP-10 concentrations). [18]

Although metronidazole is a first-line agent for treatment of BV, it is not unusual for treatment failure and recurrent disease to occur. [3] Of the four clades in the population structure of G vaginalis clades 1 and 3 are associated with BV. Schyuler et al found that metronidazole resistance was highly associated with clade 3 (< 0.0001) compared to clade 1, which may have treatment implications. [3]

Abramovici et al suggest that quantitative polymerase chain reaction (qPCR) bacterial load measurement is useful in the evaluation of BV treatment response and the risk of preterm birth in pregnant women. [19] The investigators found that qPCR correlated with Nugent score and demonstrated decreased bacterial load after antibiotic treatment.

Previous