Send to

Choose Destination
Obstet Gynecol. 2017 Jul;130(1):181-189. doi: 10.1097/AOG.0000000000002090.

Clinical Validation of a Test for the Diagnosis of Vaginitis.

Author information

Johns Hopkins University, Baltimore, Maryland; Pathology Inc, Torrance, California; the University of Alabama at Birmingham, Birmingham, Alabama; Planned Parenthood Southeastern Pennsylvania, Philadelphia, Pennsylvania; Planned Parenthood Gulf Coast, Houston, Texas; Sidney and Lois Eskenazi Hospital and Indiana University School of Medicine, Indianapolis, Indiana; Drexel University College of Medicine, Philadelphia, Pennsylvania; Planned Parenthood of Southern New England, New Haven, Connecticut; Planned Parenthood Mar Monte, San Jose, California; and BD Diagnostics, Qu├ębec, Canada, and Sparks, Maryland.



Vaginitis may be diagnosed as bacterial vaginosis, vulvovaginal candidiasis, trichomoniasis, or coinfection. A new molecular test assays the vaginal microbiome and organisms that cause three common infections. The objective of the trial was to evaluate the clinical accuracy of the investigational test for vaginal swabs collected by patients (self) or clinicians. The primary and secondary outcomes were to compare the investigational test with reference methods for the three most common causes of vaginitis and compare clinician-collected with self-collected swabs.


We conducted a cross-sectional study in which women with symptoms of vaginitis were recruited at ten clinical centers and consented to the investigation between May and September 2015. The woman collected a vaginal swab, sheathed, and then handed it to the clinician. These swabs were to evaluate how self-collected swabs compared with clinician-collected swabs. The clinician collected an investigational test swab and reference test swabs. From 1,740 symptomatic patients, clinician-collected and self-collected vaginal swabs were evaluated by the molecular test and six tests. The reference methods for bacterial vaginosis were Nugent's score and Amsel's criteria for intermediate Nugent results. The reference methods for Candida infection were isolation of any potential Candida microorganisms from inoculation of two culture media: chromogenic and Sabouraud agar and sequencing. The reference methods for trichomoniasis were wet mount and culture.


For clinician-collected swabs, by reference methods, bacterial vaginosis was diagnosed in 56.5%, vaginal candidiasis in 32.8%, trichomoniasis in 8%, and none of the three infections in 24% with a coinfection rate of 20%. The investigational test sensitivity was 90.5% (95% confidence interval [CI] 88.3-92.2%) and specificity was 85.8% (95% CI 83.0-88.3%) for bacterial vaginosis. The investigational test sensitivity was 90.9% (95% CI 88.1-93.1%) and specificity was 94.1% (95% CI 92.6-95.4%) for the Candida group. Sensitivity for Candida glabrata was 75.9% (95% CI 57.9-87.8%) and specificity was 99.7% (95% CI 99.3-99.9%). Investigational test sensitivity was 93.1% (95% CI 87.4-96.3%) and specificity was 99.3% (95% CI 98.7-99.6%) for trichomoniasis. Results from self-collected swabs were similar to clinician-collected swabs.


A molecular-based test using vaginal swabs collected by clinicians or patients can accurately diagnose most common bacterial, fungal, and protozoan causes of vaginitis. Women and their clinicians seeking accurate diagnosis and appropriate selection of efficacious treatment for symptoms of vaginitis might benefit from this molecular test.

Comment in

[Indexed for MEDLINE]
Free PMC Article

Supplemental Content

Full text links

Icon for Wolters Kluwer Icon for PubMed Central
Loading ...
Support Center