Inside Track: Bacterial Vaginosis: Common, and Silent

For a vaginal infection that is often symptomless, testing options are on the rise

Interview by Steve Halasey

Bacterial vaginosis (BV) is the most common vaginal infection among women aged 15 to 44. Although more than 50% of women with BV have no symptoms, prevalence in the United States has been reported to be around 21 million. Education and diagnostic testing approaches have not been heavily stressed.

Donna Mayne, BS, MT(ASCP), (CLS)NCA, Sacred Heart Hospital Pensacola Laboratory.

Last September, CLP hosted a well-attended webcast on “Bacterial Vaginosis: Complications and Diagnostic Options,” featuring a presentation by Donna Mayne, BS, MT(ASCP), (CLS)NCA, microbiology and molecular laboratory manager at Sacred Heart Hospital Pensacola Laboratory, Pensacola, Fla. A veteran of more than two decades at Sacred Heart, Mayne manages the daily operations of the unit’s microbiology, serology, and molecular laboratories, including their personnel and test menus. It is her responsibility to plan, develop, and implement new tests, equipment, and procedures to align with the current and future testing needs of the medical staff.

In her presentation, Mayne provided an overview of current knowledge about BV, including complications and risks associated with the disease, and currently available diagnostic testing options. The webcast concluded with a live Q&A session that elicited dozens of questions from attendees, most of which could not be addressed during the live event. Abridged and edited results of this Q&A session are reproduced below.

BV CHARACTERISTICS

CLP: Does bacterial vaginosis (BV) increase the risk of human papilloma virus (HPV) infection as it does for gonorrhea, herpes simplex virus (HSV), human immunodeficiency virus (HIV), and syphilis?

Donna Mayne, BS, MT(ASCP), (CLS)NCA: I don’t remember seeing HPV specifically listed in any of the literature I reviewed, but if I had to guess I would say ‘yes.’ BV is typically seen as an inflammation, which all by itself allows the tissues to be more accepting of infectious organisms. So in theory, yes, BV may also increase the risk of HPV infection. But I can’t state that on the basis of any research that I’ve encountered. Similarly, I can’t say whether BV might in any way increase the risk of an HPV infection leading to cervical cancer or other HPV-related conditions.

CLP: You mentioned that BV appears as an inflammation. But isn’t it called bacterial ‘vaginosis’ instead of ‘vaginitis’ because it doesn’t always cause an inflammatory response such as yeast does? In wet prep or Gram stain examination of specimens from patients with BV, not many white blood cells indicative of inflammation are usually seen.

Mayne: That is true. I agree that we don’t see a lot of white cells on our wet preps when BV is present, and that’s something of an inconsistency. However, we do often see inflammatory markers such as interleukin 1 and 8, and other enzymes that cause inflammation. So BV does cause inflammation at least occasionally, or during some periods of the illness.

CLP: Is there any evidence supporting probiotics as a way to prevent or treat BV?

Mayne: A number of studies in the literature support the use of probiotics. Several report success when using probiotics to prevent BV, and others suggest using these in combination with antibiotics to treat BV. However, treating BV by using probiotics alone, without antibiotics, is not recommended.

CLP: The antibiotics used to treat BV are directed against anaerobes. With persistent or recurrent BV, might a resistant organism be remaining and would a culture be helpful?

Mayne: Cultures are not currently recommended for this purpose. Keep in mind that when antibiotics are given to treat BV, we aren’t targeting a single organism as we would in treating strep throat or a urinary tract infection. Similarly, we aren’t trying to eradicate specific bacteria.

BV is basically an imbalance of the normal vaginal flora. It is likely that in most cases of BV, more than one organism at a time is “out of balance.” Consequently, the goal of using antibiotics is simply to reduce the number of such organisms (mostly anaerobes) in order to restore the normal balance. And only significant resistance across multiple species could result in treatment failure.

In order for culture to help in a case of such significant resistance, one would have to culture all of the organisms associated with BV, quantitate them, compare the result to the quantities of Lactobacillus, and then perform sensitivities. This process would be far too labor-intensive for most labs outside of research settings.

TESTS AND METHODS

CLP: Is patient self-collection for BV testing recommended?

Mayne: I haven’t seen any research to support the use of self-collected vaginal swabs for BV, but there is a great deal of work being done on that practice for such other conditions as chlamydia, gonorrhea, and trichomonas. In theory, self-collected swabs should probably be suitable to test for BV, so long as enough material is collected. But suitability may depend on what test format is being used to perform the test.

CLP: Because the test kits currently available for BV require 10 to 15 minutes to produce a result—plus the time needed to conduct a pelvic exam—some physicians treat according to qualitative clue cell results from a microscopic urine examination in conjunction with clinical symptoms. Is this acceptable? Is it best practice?

Mayne: In my opinion, it would not be best practice. Nothing I’ve encountered in the research literature suggests that it is a good practice to diagnose or treat on the basis of clue cells from a urine specimen. That’s partly because it’s not possible to determine whether you’re not seeing clue cells because they truly aren’t present, or because the patient collected a really good clean-catch urine specimen—even though that’s usually not the case. There isn’t very good evidence to support that practice, so I don’t think it’s a good idea.

CLP: When a lab’s technologists perform both a Gram stain and wet prep, what interpretive criteria would you recommend providing as guidance for the clinician? What else would you recommend including in the report?

Mayne: If you’re going to perform a Gram stain, you should definitely include the interpretive criteria, which is frequently the Nugent score. You should discuss with your medical director whether to use the straight-up Nugent Score or its modified version. But reporting both the score and the interpretive criteria would be highly recommended. For instance, a Nugent score of 7 is interpreted as positive for bacterial vaginosis.

Alternatively, if your laboratory information system permits the inclusion of tables, you could opt to report the Nugent score, and below that include a look-up table of the interpretive criteria for various scoring levels.

Either way, if you’re going to perform a Gram stain, you really should be reporting either a Nugent score with interpretation, or some similar criteria that differentiates the Lactobacillus from other organisms and shows the relationships among them.

CLP: When you do a wet prep examination, do you also report whiff test results?

Mayne: No, we don’t. When that test is done in our facility, it is performed by the physician at the patient bedside. Physicians may be performing the whiff test in their own offices, but I don’t think the emergency department is doing it, and we are not doing it in the lab.

CLP: What are the current published sensitivity and specificity of the whiff test?

Mayne: A 2003 study by Myzuik compared the OSOM BVBlue test by Sekisui Diagnostics to the Amsel criteria and each of its individual components.¹ The whiff test for vaginal fluid amines posted respectable specificity of 97.8%, but a very low sensitivity of just 50%—the lowest ranking of all the criteria.

CLP: How are test results for BV affected by racial differences?

Mayne: I haven’t seen any discussion about the testing differences, so I can’t really comment on that. I don’t know that there really was anything noted.

CLP: We are still performing genital cultures and report Gardnerella from the culture. Is this not recommended?

Mayne: Not usually. Although Gardnerella can cause vaginal abscesses, it is typically considered part of the normal vaginal flora. In most cases it should not be identified and reported specifically.

CLP: Since culture can’t be used to diagnose BV, should Gram stain be performed when vaginal culture is ordered?

Mayne: In my opinion, yes—along with reporting the Nugent score.

CLP: If the presence of Gardnerella is not sufficient for diagnosis of BV, what is the advantage of using BD Affirm VPIII for diagnosis?

Mayne: The BD Affirm assay tests for candidiasis, trichomoniasis, and the presence of high bacterial loads of Gardnerella (?2 x 10⁵). A high level of Gardnerella is suggestive of BV, but not diagnostic. In the presence of other clinical findings, however, such a high level can be considered diagnostic.

The advantage of this test is that it also tests for candidiasis and trichomoniasis, which are significant vaginal conditions that produce similar symptoms. Testing for all these syndromes at once can streamline several laboratory tests by combining them into one panel.

CLP: What role does sperm play in BV diagnosis?

Mayne: I haven’t seen any studies relating specifically to sperm’s role, if any. However, a 1997 study by Hay followed 18 women over a number of months, and asked the women to prepare daily vaginal smears and record symptoms, sexual activity, menstruation, and other activities.² The study found that a good number of BV cases resolved spontaneously following unprotected sexual activity.

GOING MOLECULAR

CLP: What organisms would you include if you wanted to design a molecular-based multiplex test for vaginitis or vaginosis?

Mayne: Based on what I’ve read, Atopobium, Gardnerella, a couple different species of Lactobacillus, and Mycoplasma would probably be an excellent mix. Those are the organisms that have already been studied and appear in the literature. With or without a couple additional organisms, studies have shown that mixtures of that group of organisms would create an excellent test.

CLP: Can using a molecular method to detect a particular BV-causing organism help in guiding the selection of a drug for use in treatment?

Mayne: Maybe, or maybe not. I don’t know that we will ever get to the point at which we can say, ‘this is the specific organism that is causing vaginosis in this patient.’ There are so many species present, and I don’t really think there’s just one causative agent.

Instead, what we’re really looking for is a shift in the vaginal flora. But pinpointing the cause to a single species and then trying to personalize the therapy in that direction—I don’t really see that happening.

CLP: Would self-collected swabs be more or less suitable for BV if the test format were a nucleic acid amplification test, such as those being used to test for chlamydia and gonorrhea?

Mayne: Again, I haven’t seen any research in the literature to support the practice. But at least theoretically, a self-collected swab should work if the lab is using an amplified molecular assay. We are seeing a lot of development in that area—especially for chlamydia, gonorrhea, and trichomonas—and I can’t imagine why it would be a bad idea.

Perhaps this approach will encourage more women to be tested. Women are generally resistant to going through a full pelvic exam. But if they’re able to self-collect vaginal swabs, perhaps we’ll have better cooperation with testing.

CLP: What are the acceptable sample types for molecular-based tests for BV?

Mayne: It would be awesome if a self-collected vaginal swab could be developed and proven effective, as previously suggested. A self-collected vaginal swab using a generic collection and transport device, such as the ESwab by BD Diagnostics, would be a good fit.

It would be equally nice if it were possible to perform the testing from a urine specimen. If the goal is to develop the perfect test using the perfect collection device, using urine specimens or self-collected swabs comes close, since those samples are easy to come by. You can also consider running multiple tests from the same sample.

CLP: Are there any new recommendations regarding the requirement to begin testing for trichomonas via a polymerase chain reaction (PCR) test?

Mayne: Several manufacturers have either recently released molecular assays for trichomonas, or have them under development or pending FDA clearance. Pretty soon, they will become standard for trichomonas, just as they are for chlamydia and gonorrhea. But I don’t know when a practice guideline might be published to recommend the use of molecular diagnostics for trichomonas. I haven’t seen anything about that.

CLP: If utilizing a molecular test with cut-off value for positive versus negative, should a standard volume of specimen be required? Is this the case with the current molecular tests you mentioned in your presentation?

Mayne: Most molecular tests that I’m familiar with do require standard volumes of the specimen, such as 100 µl of urine, or 50 µl of blood from a UTM tube. When collecting a specimen with a swab, however, one doesn’t have much control over the amount of sample being collected, and in a test looking for elevated levels of an organism, that could be an issue.

I imagine that limitation is the reason that the BD Affirm test is not considered “diagnostic” but only “suggestive” of BV when positive for Gardnerella. As more intuitive molecular tests are developed for BV diagnosis, I would think a comparison between cycle times (Ct) for the targeted organisms (such as Gardnerella) and expected normal flora (such as Lactobacillus) would be one way to reduce the effect of specimen collection variations.

Steve Halasey is chief editor of CLP. He can be reached via [email protected].

REFERENCES

1. Myzuik L, Romanowski B, Johnson SC. BVBlue test for diagnosis of bacterial vaginosis. J Clin Microbiol. 2003;41(5):1925?1928; doi: 10.1128/jcm.41.5.1925-1928.2003.

2. Hay PE, Ugwumadu A, Chowns J. Sex, thrush, and bacterial vaginosis. Int J STD AIDS. 1997;8(10):603–608.