Ferris, Daron G. MD; Nyirjesy, Paul MD; Sobel, Jack D. MD; Soper, David MD; Pavletic, Adriana MD; Litaker, Mark S. PhD
An accurate diagnosis of vaginitis is contingent upon patient information, clinical findings, and interpretation of vaginal specimen laboratory analyses. However, patient symptoms and clinical findings alone are often unreliable for determining the etiology of a woman's complaint.1–6 In particular, vulvovaginal candidiasis is more commonly misdiagnosed than are vaginal trichomoniasis and bacterial vaginosis.1,6 Inaccurate diagnosis of vulvovaginal symptoms may delay properly diagnosing more serious lower genital tract diseases or coexisting infections.
Until the past decade, treatment of vulvovaginal candidiasis remained entirely within the domain of qualified health care providers. However, the reclassification of former prescription‐only vulvovaginal candidiasis drugs to over‐the‐counter products provided women easy access to self‐treatment. The availability appeals to women in general, the pharmaceutical industry, health insurance providers, and many clinicians. This popular approach is believed to reduce health care expenditures, empower women in addressing their own wellness, and improve pharmaceutical product convenience for the public.7–9
Appropriate selection and use of over‐the‐counter products for vulvovaginal candidiasis may be problematic, particularly for some women who have no medical training.10 Universal drug availability does not invariably insure proper selection, use, or satisfactory outcomes.11 The Food and Drug Administration (FDA) assumes that vulvovaginal candidiasis medical education provided to patients during a prior visit to a medical provider, or health information written on the over‐the‐counter package label enable women to treat themselves appropriately. Yet, clinicians frequently encounter patients who have used these drugs to treat vaginal symptoms not caused by vulvovaginal candidiasis.
The magnitude of errant use of over‐the‐counter anti‐fungal products remains subject to speculation because of a paucity of population‐based data.12,13 The purpose of this study was to estimate what proportion of a time‐location sample of symptomatic women purchasing over‐the‐counter antifungal products for immediate treatment of presumed vulvovaginal candidiasis had vulvovaginal candidiasis or other genitourinary conditions. In addition, we determined whether the history of a previous diagnosis of vulvovaginal candidiasis by a clinician or reading the over‐the‐counter package label health education instructions contributed to a more accurate self‐diagnosis of vulvovaginal candidiasis.
MATERIALS AND METHODS
Women 18 years of age and older who purchased an over‐the‐counter product for treatment of self‐diagnosed vulvovaginal candidiasis at pharmacies and grocery stores in the metropolitan areas of Augusta, Georgia, Philadelphia, Pennsylvania, Detroit Michigan, Charleston, South Carolina, and Omaha, Nebraska, were referred for enrollment in the study between September 1997 and December 1999. Subjects were included if they: 1) had symptoms which prompted them to purchase an over‐the‐counter antifungal product, 2) intended to use the product immediately, 3) and presented to the clinics with an untampered over‐the‐counter product purchased recently. Subjects who had been diagnosed as having vulvovaginal candidiasis by a health care provider or who had broken the tamper proof seals on the over‐the‐counter antifungal products were excluded from participation. The study was approved by the respective Institutional Review Boards (Medical College of Georgia, Temple University, Wayne State University, Medical University of South Carolina, and University of Nebraska).
A time‐location sample of women with symptoms of presumed vulvovaginal candidiasis who purchased an over‐the‐counter antifungal product for their personal and immediate use were informed about the study and asked to volunteer for the study by the pharmacists, cashiers, or clerks at the pharmacy/grocery stores in the respective metropolitan areas. Each woman received a pamphlet that described the trial, her involvement, and potential benefits. If the woman agreed to participate, two tamper proof seals were placed over the ends of the over‐the‐counter package, and the woman was asked to take this over‐the‐counter product to the clinic site for enrollment. All women were evaluated within 24 hours of presentation or product purchase.
At the clinical site, women were informed about the study, and then read and signed the Institutional Review Board approved consent form. Demographic information was obtained and a self‐administered survey completed. The survey included questions pertaining to the subject's symptoms, history of vaginitis, sexual history, and experience with over‐the‐counter products. After visualization of the cervix and vagina by use of a vaginal speculum, cervical and vaginal specimens were obtained for Neisseria gonorrhoeae (N. gonorrhoeae) and Chlamydia trachomatis (C. trachomatis) DNA testing, 10% potassium hydroxide and saline wet preparation, sniff test, pH determination, Gram stain, and fungal culture. An abdominal, external genitalia, and pelvic examination completed the assessment.
A vaginal specimen sample for a saline wet preparation was examined microscopically for the presence of leukocytes, clue cells, Lactobacillus sp., pseudohyphae and spores, and motile Trichomonas vaginalis. Another vaginal specimen sample was examined similarly for pseudohyphae and spores after the addition of potassium hydroxide. A sniff test of the vaginal discharge was performed immediately after 10% potassium hydroxide placement to detect an amine odor. A small amount of the vaginal specimen was applied to a strip of pH paper, and the resulting colorimetric reaction was interpreted by comparison with the corresponding pH reference scale. A small sample of vaginal secretions/discharge was collected, then inoculated on media, for fungal culture. A Gram stain was prepared from a small sample of vaginal fluid smeared on a glass slide. The slide was heat fixed and processed by the standard Gram stain technique. Gram stains were interpreted at a central site, and results included Trichomonas vaginalis, fungal elements, clue cells, inflammatory cells, red blood cells, sperm, gram‐negative diplococci, Lactobacillus sp., Gardnerella vaginalis, and Mobiluncus sp.
A cervical swab specimen was obtained and tested for the presence of C. trachomatis and N. gonorrhoeae by a polymerase chain reaction technique. Urine specimens were analyzed by use of a dipstick test and microscopic examination, when appropriate. Other than Gram stains, which were evaluated at a central site (JDS), all specimens were processed and interpreted at the local clinical center laboratory.
A positive diagnosis of vulvovaginal candidiasis in this symptomatic population was defined as gram‐positive for vulvovaginal candidiasis or a positive fungal culture with or without a positive 10% potassium hydroxide preparation (budding yeast, pseudohyphae, or hyphal forms). A positive diagnosis of Trichomonas vaginalis was defined as the presence of characteristic motile trichomonads on a saline wet preparation or nonmotile trichomonads on the Gram stain. The positive clinical diagnosis of bacterial vaginosis was defined as the presence of three of four Amsel's criteria (adherent off‐white vaginal discharge, pH greater than 4.5, clue cells on saline wet preparation, a positive sniff test), or a positive Nugent's Gram stain score of greater than 7.14,15 A positive diagnosis of C. trachomatis and N. gonorrhoeae infection was defined as a positive polymerase chain reaction test for C. trachomatis and N. gonorrhoeae, respectively.
The percentage of women in the sample who were diagnosed with vulvovaginal candidiasis and other gynecologic conditions was calculated, including conditions coexisting with vulvovaginal candidiasis. The proportion of women with vulvovaginal candidiasis was compared between the groups with and without a prior diagnosis of vulvovaginal candidiasis, using the χ2 statistic. The odds ratio (OR) was calculated as a measure of association between prior diagnosis and correct self‐diagnosis. A 95% confidence interval (CI) for the true OR was calculated by the logit method. The sample percentage of women reporting having read the package label was calculated, along with a 95% CI. The proportion of women who were diagnosed with vulvovaginal candidiasis among those who reported reading the package label were compared with the corresponding proportion among those who reported not reading the label using Fisher exact test. The OR and a 95% confidence limit were calculated as measures of association. The Wilcoxon rank sum test was used to compare continuous and Likert scale variables between groups. Mantel‐Haenszel χ2 statistic was used to compare ordinal categoric responses between groups.
Of 104 women enrolled in the study, complete data were available for 95 who presented with a self‐diagnosis of vulvovaginal candidiasis and intended to use the product immediately for themselves. The mean age of subjects was 36.6 years, and the range was 18–67 years. The majority (53.7%) of subjects was black, 40.0% were white, and 6.3% were Hispanic or Asian. With respect to marital status, 39.0% of women were married, 36.8% single, 17.9% divorced, and 6.3% separated or widowed. The majority of subjects (72.3%) were employed, 88.5% were high school graduates, and 67% had a household income of greater than or equal to $20,000. Approximately half (50.5%) of the women used some form of contraception. Of this group, 38.3% used condoms, 27.7% oral birth control pills, 17.0% were sterilized, 4.3% used DepoProvera, and 12.8% used other forms of contraception.
The final diagnoses for women who intended to treat their self‐diagnosed vulvovaginal candidiasis with an over‐the‐counter antifungal product are listed in Table 1. Only 32 of 95 (33.7%) women made the correct diagnosis of vulvovaginal candidiasis. However, an additional 19 (20.0%) women had vulvovaginal candidiasis plus another type of vaginitis, primarily coexisting bacterial vaginosis (18.9%). Fifty‐one percent of women had a vaginal pH greater than or equal to 4.7, 30% had a positive sniff test, 36% had evidence of vulvovaginal candidiasis by the 10% potassium hydroxide test, 30% had evidence of vulvovaginal candidiasis by Gram stain, and 45% had a positive fungal culture. One woman had a positive test for C. trachomatis.
Women who made the correct diagnosis of vulvovaginal candidiasis (pure vulvovaginal candidiasis or mixed infection) were no more confident of their self‐diagnosis (median Likert score = 7, range 1 = not certain and 10 = very certain) than were women who did not have vulvovaginal candidiasis (median Likert score = 8, P = .16, Wilcoxon rank sum test). This was also true if only women with pure vulvovaginal candidiasis infections were compared with women who did not have only vulvovaginal candidiasis, medians 7 and 8, respectively (P = .43, Wilcoxon rank sum test).
The effect of a prior diagnosis of vulvovaginal candidiasis on a current correct diagnosis of vulvovaginal candidiasis was analyzed (Table 2). Women with pure vulvovaginal candidiasis were no more likely to have a previous clinically based diagnosis of vulvovaginal candidiasis than were women without a current diagnosis of a pure vulvovaginal candidiasis infection (χ2 = 0.27, P = .6). The odds of having a prior diagnosis of vulvovaginal candidiasis was equivalent for women with and without a current pure vulvovaginal candidiasis infection (OR 0.76, 95% CI 0.26, 2.18). These data differed if the current diagnosis was defined as any vulvovaginal candidiasis instead.
Of the 60 women who consulted with a clinician for their last diagnosis of vulvovaginal candidiasis, 72.3% received the diagnosis after a clinical examination, 17.3% were diagnosed by telephone, and 11.4% had both a telephone and then clinically based diagnosis. A prior telephone diagnosis of vulvovaginal candidiasis did not improve women's self‐diagnosis ability (χ2 = 0.88, P = .3).
Of women who diagnosed themselves correctly, more read the over‐the‐counter package label (63.3%) than women who did not read the label (36.7%) (Table 3). However, women who read the label were no more likely to have vulvovaginal candidiasis than were women who did not read the label (P = .39, Fisher exact test). The odds of reading the label was equivalent for women with and without vulvovaginal candidiasis (OR 1.1, 95% CI 0.4, 2.7, P = .8).
Next, we examined various factors, which might affect a correct self‐diagnosis of vulvovaginal candidiasis (Table 4). Women who did not correctly self‐diagnose pure or mixed vulvovaginal candidiasis had a significantly greater mean and median number of lifetime vulvovaginal candidiasis infections (17.8 and 7.5, respectively) compared with women who made a correct self‐diagnosis of vulvovaginal candidiasis (6.5 and 4.0, respectively, P = .01, Wilcoxon rank sum test). Women who incorrectly self‐diagnosed vulvovaginal candidiasis also had a significantly greater mean and median number of prior clinically diagnosed vulvovaginal candidiasis infections (9.9 and 4.0, respectively) compared with women who diagnosed vulvovaginal candidiasis correctly (4.0 and 3.0, respectively, P = .02, Wilcoxon rank sum test). However, there were no significant differences noted between median group scores for the other variables. Because a large number of women (38.9%) also had bacterial vaginosis, we assessed the role of a prior clinical diagnosis of bacterial vaginosis on establishing a correct current diagnosis of vulvovaginal candidiasis. A prior diagnosis of bacterial vaginosis did not improve women's ability to diagnose a pure and mixed vulvovaginal candidiasis infection correctly (χ2 = 0.009, P = .9), nor a pure vulvovaginal candidiasis infection (χ2 = 3.1, P =.8).
Finally, there were 44 (46.3%) patients whose correct diagnoses were considered delayed by initial intent to use the over‐the‐counter antifungal product. The consequence of this delay was determined by the investigators to be minor for 61%, moderate for 19%, severe for 4%, and of no consequence for the remaining 16% of patients. Acute salpingitis requiring hospitalization is an example of a severe delay.
We determined that approximately one‐third of women who self‐diagnosed a vulvovaginal candidiasis infection and purchased an over‐the‐counter product to treat their condition actually had vulvovaginal candidiasis confirmed by a nearly immediate clinical and laboratory examination. An additional 20% of women had vulvovaginal candidiasis, but also with another type of vaginitis. Self‐treatment with an over‐the‐counter antifungal product will generally cure vulvovaginal candidiasis in women with mixed infections, but these women may likely continue to have other vaginal symptoms because of a second coexisting infection that requires clinical and laboratory evaluation. These women, and an additional 32% (or 51% total), needed a clinician‐prescribed pharmaceutical product for adequate therapy. Therefore, half of the women who use over‐the‐counter products for self‐diagnosed vulvovaginal candidiasis may eventually need to visit a clinician to receive proper therapy. The remaining 14% of women also require a clinical examination and pertinent laboratory tests to reassure them that they are normal and no treatment is necessary. The implications of these findings are unnecessarily wasted financial expenditures, unfulfilled expectations, potential precipitation of lower genital tract allergic and contact dermatitis, and a delay in diagnosis and cure for the majority of women.11,16
The FDA suggests that only women who have had a prior diagnosis of vulvovaginal candidiasis or have read the over‐the‐counter package label should use these products. However, our data clearly indicate that a prior clinician‐based diagnosis of vulvovaginal candidiasis and reading the label do not improve women's ability to properly diagnose vulvovaginal candidiasis. Furthermore, the type of prior diagnostic assistance, whether by telephone or even by clinical examination, did not increase women's self‐diagnosis capability. Although telephone diagnosis is becoming increasingly popular in capitated health care settings, the approach is fraught with inaccurate diagnoses. The number of prior clinically diagnosed vulvovaginal candidiasis infections, frequency of over‐the‐counter vulvovaginal candidiasis product use, and prior clinician recommendation to use over‐the‐counter products did not influence accurate decision making by these women. The latter finding is particularly troubling because many clinicians tell their patients to use an over‐the‐counter product in the future if the same present symptoms return. The educational benefit appears to be overestimated for many women and may be explained by either a failure of clinicians to successfully impart the characteristics of vulvovaginal candidiasis to their patients or some patients' general inability to learn how to self‐diagnose future vulvovaginal candidiasis based on the current clinician‐ and laboratory‐derived diagnosis. Amazingly, the majority of women reported reading the label, even though it did not increase the probability of appropriate product selection and proper use. Did women miscomprehend the instructions or are the instructions of no value for helping women discriminate whether they have vulvovaginal candidiasis? Eye‐tracking studies of people reading product labels have shown that individuals rarely read them.17 Among those who do read them, abbreviated duration of observation and haphazard eye pattern suggest poor comprehension.
The total number of lifetime vulvovaginal candidiasis infections was associated with women's ability to self‐diagnose vulvovaginal candidiasis, but not as might be expected. Women who diagnosed vulvovaginal candidiasis incorrectly reported a significantly greater number of lifetime vulvovaginal candidiasis infections compared with women who made the correct self‐diagnosis. It is possible that many of these presumed vulvovaginal candidiasis infections were in fact not vulvovaginal candidiasis. Repetitive treatment with an over‐the‐counter product for a nonvulvovaginal candidiasis infection would mistakenly convey the impression of recurrent and numerous vulvovaginal candidiasis infections. Perhaps the women who incorrectly self‐diagnosed vulvovaginal candidiasis also tended to overestimate the frequency because of their frustrations in dealing with the matter.
A potential for delay in proper diagnosis was noted for half of these women. Although the consequences of the diagnostic delays were primarily minor, several were believed to be potentially severe. A patient who presented with vaginal discharge determined to be caused by acute salpingitis, requiring hospital admission and treatment, is one example from our study of a potentially serious diagnostic delay. Had this woman instead used the over‐the‐counter product for self‐diagnosed vulvovaginal candidiasis, more serious complications could have occurred. New data indicating women with bacterial vaginosis are at increased risk for acquiring human immunodeficiency virus underscores the importance of what was previously believed to represent a minor patient inconvenience.18 An unhealthy vaginal ecosystem places women at risk for human immunodeficiency virus and other serious medical conditions, including postoperative infections and pregnancy‐related complications.19–24
Our study was limited by a somewhat small sample size. However, recruitment was extremely difficult for our experienced research group because most women who use these over‐the‐counter products do so to avoid being seen by a health care provider. Even though we were available to see women immediately and provided transportation money, the vast majority chose to ignore our offer. Therefore, some may infer that we may have attracted women who were unsure of their self‐diagnosis. Yet, it was noteworthy to find a high confidence level for correct self‐diagnosis of vulvovaginal candidiasis in both groups. This high level of confidence has been noted by others.25 Such a finding would argue against the contention of an unintentional selective recruitment. Black women were also represented at a greater percentage than found in the general US population. However, this percentage reflects the normal population expected at the major enrollment site in the southern United States. Whether this increased percentage also represents an actual barrier to health care is unknown.
Finally, the overall economic impact of these misdiagnoses may be less than reverting back to a prescription‐only era. Much of the costs to the health care industry have been clearly reduced and assumed by consumers. In fact, the health care industry is now pushing for other prescription‐only products to become classified as over the counter. Our intent was not to determine the cost‐effectiveness of the self‐diagnosis, self‐treatment scenario.7,8 However, it is also obvious that the pharmaceutical industry has benefited economically by empowering women. The FDA has allowed the use of over‐the‐counter antifungal drugs for treatment of self‐diagnosed vulvovaginal candidiasis, although our study found the FDA's proper use stipulations ineffective. In the current era of self‐diagnosis and treatment for vulvovaginal candidiasis, American women have been led to believe that they are able to fully handle this condition on their own. Our data strongly dispute this assumption and suggest that further analysis should address whether women truly benefit from the current approach towards treating vulvovaginal candidiasis.
1. Ferris DG, Hendrich J, Payne PM, Getts A, Rassekh R, Mathis D, et al. Office laboratory diagnosis of vaginitis: Clinician performed tests compared with a rapid nucleic acid hybridization test. J Fam Pract 1995;41:575–81.
2. Berg AO, Heidrich FE, Fihn SD, Bergman JJ, Wood RW, Stamm WE, et al. Establishing the cause of genitourinary symptoms in women in a Family Practice: Comparison of clinical examination and comprehensive microbiology. JAMA 1984;251:620–5.
3. Schaaf VM, Perez-Stable EJ, Borchardt K. The limited value of symptoms and signs in the diagnosis of vaginal infections. Arch Intern Med 1990;150:1929–33.
4. Rothenberg RB, Simon R, Chipperfield E, Catterall RD. Efficacy of selected diagnostic tests for sexually transmitted diseases. JAMA 1976;235:49–51.
5. Bergman JJ, Berg AO, Schneeweiss R, Heidrich FE. Clinical comparison of microscopic and culture techniques in the diagnosis of Candida
vaginitis. J Fam Pract 1984;18:549–52.
6. Reed BD, Huck W, Zazove P. Differentiation of Gardnerella vaginalis
, Candida albicans
, and Trichomonas vaginalis
infections of the vagina. J Fam Pract 1989;28:673–9.
7. Lipsky MS, Waters T, Sharp LK. Impact of vaginal antifungal products on utilization of health care services: Evidence from physician visits. J Am Board Fam Pract 2000; 13:178–82.
8. Gurwitz JH, McLaughlin TJ, Fish LS. The effect of an Rx-to-over-the-counter switch on medication prescribing patterns and utilization of physician services: The case of vaginal antifungal products. Health Services Res 1995;30:672–85.
9. Lipsky MS, Waters T. The “prescription-to-over-the-counter switch” movement: Its effects on antifungal vaginitis preparations. Arch Fam Med 1999;8:297–300.
10. Ferris DG, Dekle C, Litaker MS. Women's use of over-the-counter antifungal pharmaceutical products for gynecologic symptoms. J Fam Pract 1996;42:595–600.
11. Nyirjesy P, Weitz V, Grody MHT, Lorber B. Over-the-counter and alternative medicines in the treatment of chronic vaginal symptoms. Obstet Gynecol 1997;90:50–3.
12. Fertility and Maternal Health Drugs Advisory Committee. Proceedings of the FDA panel on over-the-counter vaginal fungicides. Washington, DC: Miller Reporting Company; 1991:138–98.
13. Chaponis R, Bresnick P, Weiss R, Edwards L. Candida
vaginitis: Signs and symptoms aid women's self recognition. J Clin Res Drug Dev 1993;7:17–23.
14. Amsel R, Totten PA, Spiegel CA, Chen KC, Eschenbach D, Holmes KK, et al. Nonspecific vaginitis diagnostic criteria and microbial and epidemiologic association. Am J Med 1983;74:14–22.
15. Nugent RP, Krohen MA, Hillier SL. The reliability of diagnosing vaginosis is improved by a standardized method of Gram stain interpretation. J Clin Microbiol 1991;29:297–301.
16. Sobel JD, Faro S, Force RW, Foxman B, Ledger WJ, Nyirjesy PR, et al. Vulvovaginal candidiasis: Epidemiologic diagnostic, and therapeutic considerations. Am J Obstet Gynecol 1998;178:203–11.
17. Fischer PM, Richards JW Jr, Berman EJ, Krugman DM. Recall and eye tracking study of adolescents viewing tobacco advertisements. JAMA 1989;261:84–9.
18. Martin HL, Richardson BA, Nyange PM, Lavreys L, Hillier SL, Chohan B, et al. Vaginal lactobacilli, microbial flora, and risk of Human Immunodeficiency Virus type 1 and sexually transmitted disease acquisition. J Infect Dis 1999;180:1863–8.
19. Hillier SL. The vaginal microbial ecosystem and resistance to HIV. AIDS Res Hum Retrovir 1998;14:S17–21.
20. Soper DE, Bump R, Hurt WG. Bacterial vaginosis and trichomoniasis vaginitis are risk factors for cuff cellulitis after abdominal hysterectomy. Am J Obstet Gynecol 1990; 163:1016–23.
21. Lin L, Song J, Kimber N, Shott S, Tangora J, Aroutcheva A, et al. The role of bacterial vaginosis in infection after major gynecologic surgery. Infect Dis Obstet Gynecol 1999;7:169–74.
22. Watts DW, Krohn M, Hillier SL, Eschenbach DA. Bacterial vaginosis as a risk factor for post cesarean endometritis. Obstet Gynecol 1990;75:52–8.
23. Gravett MG, Nelson PH, DeRoven T, Critchlow C, Eschenbach DA, Holmes KK. Independent associations of bacterial vaginosis and Chlamydia trachomatis
infection with adverse pregnancy outcome. JAMA 1986;256:1899–903.
24. Hillier SL, Martins J, Krohn M, Kiviat N, Holmes KK, Eschenbach DA. A case-controlled study of chorioamniotic infection and histologic chorioamniotis in prematurity. N Engl J Med 1988;319:972–8.
25. Lipsky MS, Taylor CA. The use of over-the-counter antifungal vaginitis preparations by college students. Fam Med 1996;29:493–5.