Annals of Epidemiology 27 (2017) 575e582

Contents lists available at ScienceDirect

Annals of Epidemiology
journal homepage: www.annalsofepidemiology.org

Review article

Recurrent vulvovaginal candidiasis

Freida Blostein a, Elizabeth Levin-Sparenberg MPH, PhD a, Julian Wagner PhD b,
Betsy Foxman PhD a, *
a
Center for Molecular and Clinical Epidemiology, University of Michigan, School of Public Health, Ann Arbor
b
Independent, Lausanne, Switzerland

a r t i c l e i n f o a b s t r a c t

Article history: Purpose: Recurrent vulvovaginal candidiasis (RVVC), multiple episodes of vulvovaginal candidiasis (VVC;
Received 3 April 2017 vaginal yeast infection) within a 12-month period, adversely affects quality of life, mental health, and
Accepted 4 August 2017 sexual activity. Diagnosis is not straightforward, as VVC is defined by the combination of often
Available online 15 August 2017
nonspecific vaginal symptoms and the presence of yeastdwhich is a common vaginal commensal.
Estimating the incidence and prevalence is challenging: most VVC is diagnosed and treated empirically,
Keywords:
the availability for purchase of effective therapies over the counter enables self-diagnosis and treatment,
Vulvovaginal candidiasis
and the duration of the relatively benign VVC symptoms is short, introducing errors into any estimates
Recurrent
Prevalence
relying on medical records or patient recall.
Incidence Methods: We evaluate current estimates of VVC and RVVC and provide new prevalence estimates using
Diagnosis data from a 2011 seven-country (n ¼ 7345) internet panel survey on VVC conducted by Ipsos Health
(https://www.ipsos.com/en). We also evaluate information on VVC-associated visits using the National
Ambulatory Medical Care Survey.
Results: The estimated probability of VVC by age 50 varied widely by country (from 23% to 49%, mean
39%), as did the estimated probability of RVVC after VVC (from 14% to 28%, mean 23%).
Conclusions: However estimated, the probability of RVVC was high suggesting RVVC is a common
condition.
Ó 2017 Elsevier Inc. All rights reserved.

Introduction The availability of over-the-counter treatment, challenges of

Vulvovaginal candidiasis (VVC) affects as many as one of every
two women during their life but is perceived as a common or
nuisance condition because it is easily treated, often with medicines
available over the counter [1e5]. However, on a population scale
VVC’s impact is large, costing an estimated $2.84 billion in the
United States alone (adjusted to 2017 dollars from [6]). Most
women will experience only one or two episodes of VVC, but there
is a large, albeit poorly defined, subset that experiences multiple
recurrences [6e9]. Recurrent vulvovaginal candidiasis (RVVC)
adversely affects quality of life, mental health, and sexual activity
[10e12]. Diagnosis is not straightforward, as it is defined by the
combination of often nonspecific vaginal symptoms and the pres-
ence of yeastdwhich is a common vaginal commensal [2, 13]. In
clinical practice, VVC and RVVC are usually treated based solely on
signs and symptoms.
Drs. Wagner and Foxman have consulted for NovaDigm Therapeutics.
* Corresponding author. Department of Epidemiology, 1415 Washington Heights,
Ann Arbor, MI 48109-2029.
E-mail address: bfoxman@umich.edu (B. Foxman).
http://dx.doi.org/10.1016/j.annepidem.2017.08.010
1047-2797/Ó 2017 Elsevier Inc. All rights reserved.
diagnosis, and perceived benign nature, pose significant challenges
for accurately estimating the occurrence and risk of recurrence. In
this review, we evaluate current estimates of VVC and RVVC, pro-
vide new estimates using data from the National Ambulatory
Medical Care Survey (NAMCS) and a 2011 multicountry internet
panel survey on VVC conducted by Ipsos Health (https://www.
ipsos.com/en), and describe the challenges in obtaining accurate
estimates of the occurrence of VVC and RVVC.
Overview of condition
VVC, more generally referred to as a vaginal yeast infection, is
characterized by vulvar erythema, excoriation, pruritus, and an
abnormal “cheese-like” or watery vaginal discharge. VVC also may
be accompanied by a change in vaginal odor [2, 4, 14e16]. Symp-
toms found among patients with vaginal cultures positive for
Candida range from none to many [4]. Candida albicans most
commonly causes VVC, accounting for 85%e90% of all cases, with
the remainder attributed to C. glabrata, C. krusei, C. famata, and
C. tropicalis [2, 13, 17]. VVC is rare before menarche, peaks during
576 F. Blostein et al. / Annals of Epidemiology 27 (2017) 575e582
the reproductive years, then declines after menopause [2, 4, 18].
Risk factors for sporadic VVC include sexual activity, contraceptive
use, antibiotic use, carbohydrate intake, and diabetes [4, 18e20].
Few descriptions of the natural history of VVC exist in the
medical or scientific literature, but clinical impression is that the
initial acute infection is often followed by one or more episodes.
This was confirmed by the few studies that directly estimate
recurrence. Yue et al. [21] followed 400 Chinese women seen at
the Sichuan University of China Hospital with culture-confirmed
VVC for up to two years. After effective treatment that cured all
symptoms, 53% (212/400) had a second episode within two years.
Amouri et al. [22] identified 231 culture-confirmed Tunisian VVC
patients, and followed 71 of those patients for one year. Sixty
percent (49/71) of those followed experienced a recurrence.
Richter et al. [9] obtained yeast isolates from 429 culture-
confirmed Iowan VVC patients over a 39-month period. Almost
one-fifth (19.6% [84/429]) had multiple positive cultures during
the study period.
RVVC has been defined as 3 or more, and as 4 or more, episodes of
VVC within a 12-month period [18, 23]. Behavioral risk factors asso-
ciated with VVC have not been associated with RVVC, but there is
some evidence of an association between polymorphisms in genes
coding for innate immunity and RVVC. The most complete evidence is
for a polymorphism in mannose-binding lectin (MBL2), which binds
fungi as part of host innate immune response. In a meta-analysis,
Nedovic et al. [24] estimated that heterogeneity in the MBL2B allele
increased the risk of RVVC by as much as four-fold and of VVC by 2.5
times. Studies by Rosentul et al. [25] identified a nonsynonymous
polymorphism in toll-like receptor 2 (Pro631His, rs5743704) that
occurred almost three times more frequently among RVVC patients
(n ¼ 119) than 263 healthy controls. Finally, in a multicountry study of
270 RVVC patients and 583 healthy controls, presence of the 12/9
genotype of the NLRP3 gene (which encodes the component of the
inflammasome that processes inflammatory cytokines IL-1band IL-
18) occurred more frequently in cases than controls [26].
Since 1967, when azoles were approved for treatment of VVC
and RVVC [27], VVC has been treated using antifungal therapies.
Resistance to antifungal therapies remains rare [28, 29] but
prolonged treatment and the widespread availability of over-the-
counter azole agents since 1990 increases the potential for
development of resistant strains [30]. In our analysis of NAMCS1
data from 2006 to 2011, antifungals were prescribed at 27% of
visits where a VVC diagnosis was made [31]. This is likely an un-
derestimate of use, as women may have been directed by their
health care provider to purchase over-the-counter medication;
further, many women may self-diagnose and treat.
RVVC imposes a significant health burden
In addition to physical discomfort, RVVC imposes significant
psychological and monetary costs [10, 11, 32]. Compared with
women presenting for family planning services with no history of
RVVC, women with RVVC had lower self-esteem, felt under greater
stress, were more likely to suffer from clinical depression, and were
1
NAMCS is a national probability-based sample survey of patient visits to
nonfederal, office-based physicians during a 1-week reporting period. In our anal-
ysis, data were weighted to the 2006e2011 age-specific female populations ac-
cording to the United States census. A diagnosis code of 1121 (candidiasis of vulva
and vagina) was used to define VVC. The diagnosis code of 61,610 (vaginitis un-
specified) was also included if there was a prescription for fluconazole (drug entry
code 93,215). VVC was also considered for diagnosis code 1129 (candidiasis un-
specified) where vaginal symptoms were present (reason for visit 17,650, 17,701,
17,600). Incidence and prevalence rates cannot be calculated from these data
because the estimates are in terms of visits and not persons.
less satisfied with their lives in general. RVVC also interfered with
emotional and sexual relationships [33]. In our analysis of the
NAMCS, psychotherapeutic agents were prescribed twice as
frequently during visits where a VVC diagnosis was made than
during visits for family planning (6% vs. 3%; p-value ¼ 0.1.) [31].
Further, in an internet panel survey of 620 women with RVVC in six
countries (USA, France, Germany, Spain, Italy, and the UK), 53%
reported anxiety/depression compared with less than 20% in the
general population [12].
In qualitative interviews of 127 women with RVVC in the United
States, nearly 80% of women reported that yeast infections affected
their personal lives including the ability to socialize and exercise,
and half reported fear of social interaction and dating. Approxi-
mately 40% of women surveyed considered VVC a very significant
burden on their lives; the most extreme burdens were discomfort
from symptoms, avoidance of sexual activity, and out-of-pocket
costs due to prescription or over-the-counter drugs and office
visit copays [34].
At a population level, out-of-pocket costs due to treatment and
office visits are significant. During 2006e2011, 5 of 1000 women
visits were for VVC among women aged 15 and over during the one-
week NAMCS reporting period [31].
VVC also reduces productivity: in qualitative interviews of 127
women with RVVC in the United States, 1/3 of respondents had
missed work due to their yeast infection [34]. Aballéa et al. [12]
estimated productivity loss due to RVVC as an average of w33
work hours per year, or approximately $1261 in 2011 dollars.
VVC and RVVC diagnosis
Symptoms of VVCditching, inflammation, and dis-
chargedoverlap with those of other common vaginal infections,
and Candida can be found even when it is not the cause of symp-
toms. For example, Candida can be isolated from the vaginal cavity
of an estimated 20% to 30% of women with bacterial vaginosis [35].
Moreover, an estimated 12%e30% of nonpregnant [36e38], and
9%e20% of pregnant women [36, 39e41], carry C. albicans asymp-
tomatically in the vaginal cavity. Uncertainty in diagnosis is
somewhat problematic clinically but poses a significant challenge
for estimating VVC and RVVC incidence. Most estimates of VVC and
RVVC rely on self-report of physician diagnosis [6e8, 34].
To evaluate the extent that physician misdiagnosis might bias
incidence and prevalence estimates based on physician visits or
patient self-report (ignoring issues of recall), we conducted a
review of recent medical textbooks, current CDC guidelines, and
scholarly journal articles (presented visually in Tables 1 and 2).
Information from 10 articles, nine medical texts, and CDC guide-
lines were used to construct graphics comparing diagnostic
technique recommendations (see Table 3 for search terms). Green
dots represent diagnostic techniques or symptoms recommended
by the reviewed sources as primary, reliable, and encouraged
ways to diagnose a VVC infection without any caveats as to
sensitivity, specificity, predictive value, or clinician access.
Although medical texts have fairly uniform recommendations
(Table 1), they are not always consistent with the scientific liter-
ature (Table 2). Further, while KOH microscopy or some other
confirmatory objective laboratory measure is consistently rec-
ommended in texts and the scientific literature, it is rarely used in
medical practice, with clinicians instead relying on signs and
symptoms [58, 59].
Unfortunately, the sensitivity and specificity of symptom-based
diagnosis of VVC is low [59e62], so both overdiagnosis and un-
derdiagnosis are possible. In a review of studies from 1966 to
2003, approximately 30% of symptomatic women experiencing
VVC remained undiagnosed after a clinical evaluation [62]. Even
 F. Blostein et al. / Annals of Epidemiology 27 (2017) 575e582 577

Table 1
Diagnostic recommendations in medical textbooks; 2006e2016 [14e16, 42e48]*

* Green dots represent diagnostic techniques or symptoms recommended by the authors as primary, reliable, and encouraged ways to diagnose a VVC infection without any
caveats as to sensitivity, specificity, predictive value, or clinician access. A diagnostic technique was coded as yellow if it was described as being a secondary diagnosis
technique, dependent on the failure of the first (i.e., “may be obtained but seldom necessary”) or if mention was made as to its unreliability due to issues with sensitivity,
specificity, predictive value, or clinician access, without a provision being made for the technique still being the best available. An asterisk indicates a diagnostic technique was
explicitly named the “gold standard”.

when augmented with a recommended laboratory technique such
as KOH microscopy and culture, problems with sensitivity and
specificity remain. As summarized in Figures A1 and A2, KOH
microscopy has a low sensitivity, and although culture has a much
better sensitivity, it has low specificity, due in part to asymptom-
atic carriage of Candida. Therefore, even if there were accurate
reporting of physician diagnosis it is likely that some diagnoses
were in error.
Further, since 1990 antifungals for VVC treatment have been
available over the counter [30]. As a result, women have been able
to purchase treatments for VVC without consulting a physician first.
Thus VVC prevalence estimates based on physician diagnosis likely
underestimate the true prevalence [30, 63e65].
Incidence and prevalence estimates based on sales of anti-
fungals are also problematic. Although women often feel confi-
dent in their ability to diagnose a yeast infection, they often
mistake nonspecific symptoms of other vaginal infections for VVC
[66e68]. Only 11% of women completing a questionnaire
designed to test their knowledge of different vaginal conditions
accurately diagnosed common causes of vaginitis after reading
578 F. Blostein et al. / Annals of Epidemiology 27 (2017) 575e582

Table 2
Diagnostic techniques in scientific literature; 2006e2016 [13, 18, 42, 49e57]*

* Green dots represent diagnostic techniques or symptoms recommended by the authors as primary, reliable, and encouraged ways to diagnose a VVC infection without any
caveats as to sensitivity, specificity, predictive value, or clinician access. A diagnostic technique was coded as yellow if it was described as being a secondary diagnosis
technique, dependent on the failure of the first (i.e., “may be obtained but seldom necessary”) or if mention was made as to its unreliability due to issues with sensitivity,
specificity, predictive value, or clinician access, without a provision being made for the technique still being the best available. An asterisk indicates a diagnostic technique was
explicitly named the “gold standard”.

classic case scenarios. Women with a prior VVC diagnosis did
somewhat better, but still poor (34.5%) [67]. In a study of women
purchasing over-the-counter medication for a self-diagnosed VVC
infection, only one-third of women actually had VVC, and an
additional 20% who did have VVC infection also had another type
of vaginitis, which may have contributed to symptoms and
required a different treatment. Prior clinician recommendation to
use over-the-counter medication did not increase accuracy of
self-diagnosis [68].
Estimated prevalence of VVC and RVVC: results of a systematic
literature review
To estimate the prevalence of VVC and RVVC, we conducted a
systematic literature review of papers published between 1980 and
2016 using the search terms shown in Table 3, which also shows the
number of articles identified per term. After limiting to articles from
scholarly journals and removing reviews, we found nine articles
estimating VVC estimates (Figures A3 Appendix). We compare the
 F. Blostein et al. / Annals of Epidemiology 27 (2017) 575e582 579

Table 3 episode of VVC estimated 5% of women had 4 or more VVC episodes

Search terms used in the systematic literature review by section, years 1980e2016 in the past 12 months [12]. In a survey used to identify RVVC pa-
Section Search term used Number results tients for qualitative interviews, 6.3% of 2391 U.S. women reported
before four or more VVC episodes in the past year [34]. A 2010 Brazilian
refinement study of 669 nondiabetic women in a cervical cancer screening
VVC prevalence VVC prevalence 409 program reported 8.2% of participants with culture-confirmed
graphic Vulvovaginal candidiasis prevalence 1348 acute VVC with 20% of these (1.6% overall) reporting 4 or more
VVC lifetime prevalence 92
episodes in the past year [70]. Although not specifically/precisely
Studies of VVC prevalence 392
RVVC RVVC prevalence 122 assessing 1-year period prevalence, these three surveys provide
prevalence Recurrent vulvovaginal candidiasis 1614 relatively consistent initial indications of a period prevalence on the
graphic prevalence order of 5%e8%. The inconsistency in estimates of current and
RVVC lifetime prevalence 38
lifetime RVVC prevalence needs further study. Some of the in-
Dot Graphic Vulvovaginal candidiasis diagnosis 1782
VVC diagnostic techniques 197
consistencies may be attributable to the different age structures of
Problems with VVC diagnosis 255 the populations. It is also possible, as we show below, that some of
RVVC diagnosis 123 the inconsistency is attributable to a recall effect. Further, most
Vulvovaginal candidiasis guidelines 766 studies do not include women in the oldest age groups or have
VVC (using Taubman Health Sciences 49
limited precision to estimate their VVC prevalence (even if recall
Library)
Yeast infection (using Taubman Health 39,452 were perfect).
Sciences Library The proportion of VVC patients who suffer from or report re-
Introduction Vulvovaginal candidiasis 3702 currences range from 6% to 20%, depending on the study population
Vulvovaginal candidiasis risk factors 1557 and methods. Active surveillance of women at an Iowan VVC clinic
Vulvovaginal candidiasis age 1517
Vulvovaginal candidiasis genetics 882
between January 1998 and March 2001 found 19.6% (84/429)
Vulvovaginal candidiasis blood 1259 experienced multiple culture-confirmed VVC episodes [9]. An Ital-
ian prospective survey of women at hospital and gynecology clinics
between October 1999 and March 2001 found 10% reported a prior
history of RVVC [8]. A 2011 Tunisian study of 231 women with
results of this review with an analysis of the cumulative probability culture-confirmed VVC found 6.1% that had a VVC episode in the
of diagnosis by age 50 estimated using data from an online seven- previous year went on to have three episodes in the year following
country omnibus opinion poll conducted by Ipsos Health, France enrollment [22]. A 2006 VVC study of 576 Greek women found that
(https://www.ipsos.com/en; epidemiologic survey 11-030783-01; 8.5% (49/576) had four or more documented positive cultures of
results of six of these countries were reported in Foxman et al. 2013 symptomatic VVC in the prior year to enrollment [71].
and Aballea et al. 2013). Incidence and prevalence cannot be esti- We used data from the 2011 Ipsos survey of 7345 participants in
mated using NAMCS data because of how the data are collected. seven countries to estimate the probability of VVC and of RVVC
given VVC by age 50 using the method of Kaplan-Meier [72]. We
Estimated prevalence of VVC and RVVC report the cumulative probability by age 50 because at older ages
the sample sizes were relatively small and estimates less stable.
The annual incidence of VVC is high. In a random digit dialing Further, it was roughly 50 years ago that the regular treatment of
survey of 1698 white Americans aged 18 and older, 26.4% reported a VVC with antifungals became commonplace [27].
physician-diagnosed yeast infection during the past 5 years or 5.2% VVC occurrence was defined by self-report of physician diag-
per year. Among the 144 black Americans participating in the sur- nosis; and RVVC occurrence was by self-report of any 12-month
vey, 46.5% reported a physician-diagnosed yeast infection during period with 4 or more yeast infections. Since age at time of first
the past 5 years or 9.3% per year [69]. Self-reported 5-year incidence VVC was not recorded, age at response to survey was used as a
decreased with age, from 38% among 18e29 year olds to 8% among proxy variable in the analysis. To estimate the cumulative proba-
those 65 and older. bility of RVVC in the total population, we multiplied cumulative
Lifetime VVC prevalence has been estimated via self- probabilities of VVC occurrence by the cumulative probability of
administered questionnaires, random digit dialing surveys, and RVVC occurrence given VVC occurrence. To assess the effects of
internet panels in different populations: U.S. college students [20], a recall, we analyzed the proportions of RVVC patients in the study by
random sample of telephone numbers in the United States [69], and age category and by self-reported age at first RVVC. There were
two different selections of participants in internet surveys in six w1000 participants per country, and sampling was stratified by age
countries [7, 64]. Regardless of method and study population, the for each country.
percent of women reporting VVC infection over their lifetimes is The cumulative probability of self-reported physician-diagnosed
high, ranging from 29% to 49% [4, 7, 20, 64]. Although these surveys VVC by age 50 averaged over the seven countries was 39%, similar to
are subject to errors, as noted previously, there can be little doubt previous estimates of lifetime VVC prevalence (Table 4). Given a
that VVC is a common condition. VVC episode, the average overall probability of experiencing RVVC
There is similar heterogeneity in methods used to estimate by age 50 was 23%; which translates to a lifetime RVVC prevalence
RVVC prevalence [6, 7, 34, 35]. Eight percent of 2000 U.S. women by age 50 of 9%. This estimate is also remarkably close to other
aged 18 years of age or older participating in a random digit dialing estimates for RVVC, which range from 6% to 10% [6, 7]. Estimates
survey reported experiencing four or more infections over a 1-year were very similar across the seven countries with the exception of
period sometime in their lives [6]. A 2013 web-based survey of France and Japan, which were lower. This difference may stem from
women from six countries estimated lifetime prevalence (defined cultural differences in response to disease, access to care, risk
as a 1-year period with 4 or more VVC episodes) at 9% [7]. Estimates behavior, or the study questions [7].
of RVVC prevalence in the past year (period prevalence) are higher If memories were perfect, and the occurrence of RVVC has not
than one might expect given estimates of lifetime prevalence. A changed with time, we would expect that those at the oldest age
cross sectional internet panel survey of U.S. and European women groups would have the most RVVC. This is not the case; the group
aged 18 to 65 with at least one health care professionalediagnosed reporting the most RVVC was aged 19 to 25 (Table 5). Further
580 F. Blostein et al. / Annals of Epidemiology 27 (2017) 575e582

Table 4 Table 5
Estimated cumulative probability by age 50 of vulvovaginal candidiasis (P(VVC)), Estimated prevalence of recurrent vulvovaginal candidiasis (RVVC) given a prior
cumulative probability by age 50 of recurrent vulvovaginal candidiasis given prior VVC, by self-reported age at first RVVC. Data from 2011 Ipsos survey (n ¼ 7345)
VVC (P(RVVCjVVC)), and estimated cumulative probability by age 50 of RVVC in the (Ipsos Health, France [https://www.ipsos.com/en]depidemiology survey 11-
total population* Data from 2011 Ipsos survey (n ¼ 7345) (Ipsos Health, France 030783-01 [June 2011]don behalf of Pevion Biotech, Switzerland
[https://www.ipsos.com/en]dEpidemiology survey 11-030783-01 [June 2011]don
behalf of Pevion Biotech, Switzerland) Age at First Current age
RVVC
12e18 19e25 26e35 36e45 46e55 56e65
Country P(VVC) P(RVVCjVVC) P(VVC)*P(RVVCjVVC)
12e18 5 (100%) 20 (25%) 13 (8%) 10 (7%) 3 (3%) 1 (1%)
France 0.32 0.14 0.04
19e25 0 61 (75%) 66 (40%) 40 (27%) 28 (28%) 11 (15%)
Japan 0.23 0.17 0.04
26e35 0 0 84 (52%) 55 (38%) 22 (22%) 24 (32%)
US 0.40 0.24 0.10
36e45 0 0 0 41 (28%) 22 (22%) 6 (8%)
Italy 0.48 0.25 0.12
46e55 0 0 0 0 25 (25%) 16 (22%)
UK 0.36 0.25 0.09
56e65 0 0 0 0 0 16 (22%)
Germany 0.43 0.27 0.12
Ever VVC of 19% 29% 25% 20% 18% 17%
Spain 0.49 0.28 0.13
ever RVV
Averages 0.39 0.23 0.09

* Results from KaplaneMeier analysis predicting survival to age at first self-

reported VVC, and age when at time of first 12-month period with 4 or more VVC
episodes among those with VVC.
analysis suggests that older women might have forgotten previous
RVVC episodes, as based on the overall estimates, we would expect
RVVC to occur equally likely in each age group, but older women
tend to not report having experiencing RVVC in the youngest age
categories (Table 5).
We found only one study that prospectively followed women to
estimate the risk of RVVC. In this 2015 study of 400 women of
reproductive age with culture-confirmed symptomatic VVC
recruited through West China Second University Hospital [21], 212
of 400 women experienced one recurrence (53%, 95% CI: 48%e58%)
within the 2 years of follow-up. Thirty percent (95% CI: 26%e35%)
had two recurrences, 17.5% (95% CI: 14%e21%) three recurrences,
and 9% (95% CI: 6%e12%) four recurrences.
The estimated risk of RVVC following VVC in the Chinese study is
very similar to estimates from the literature based on the propor-
tion of women with a history of VVC that self-report a 12-month
period with three or more VVC episodes [7, 69]. It is also similar
to the cumulative probabilities of RVVC by age 50 given VVC in the
2011 Ipsos survey (Table 4). Assuming that 23 to 49% of all women
will experience at least one episode of VVC by age 50, and 14% to
28% of them will go on to experience RVVC, an estimated 4% to 13%
of women will have a 1-year period where they experience 4 or
more VVC episodes by age 50 (Fig. 1).
There is some evidence that VVC risk varies by geographic area.
VVC prevalence varied by country in the seven-country Ipsos
internet survey (from 23% to 49%). Further, in our analysis of NAMCS
data, 17.4% of visits for VVC were in the Northeast, 21.7% in the
Midwest, 38.8% in the South, and 22.1% in the West [31]. In the 2011
Ipsos Health survey of U.S. women, the percent of women who
responded that a health care provider had told them that they had a
vaginal yeast infection, vaginal thrush, or VVC was also slightly
higher in the Southern United States. How much of this variation is
attributable to differences in diagnosis, variation in modifiable risk
factors, or population variation in genetic susceptibility is
unknown.
Looking toward the future
Estimating the prevalence and incidence of common conditions
where there are not uniformly adapted criteria for diagnosis and
self-treatment is challenging. While it is clear that VVC and RVVC
impose a substantial burden, precisely quantifying that burden
using existing databases is fraught with error. For example, a major

Fig. 1. Estimated cumulative probabilities of vulvovaginal candidiasis (VVC), recurrent vulvovaginal candidiasis (RVVC) given VVC, and RVVC by age 50 in seven countries. Estimates
from a KaplaneMeier analysis. Data from 2011 Ipsos survey (n ¼ 7345) (Ipsos Health, France [https://www.ipsos.com/en]depidemiology survey 11-030783-01 [June 2011]don
behalf of Pevion Biotech, Switzerland).
 F. Blostein et al. / Annals of Epidemiology 27 (2017) 575e582
problem with using NAMCS data is that the denominator is visits,
not women so prevalence and incidence cannot be estimated.
Further, in NAMCS not all visits coded as VVC include a prescription
for an antifungal. This is probably because antifungals are available
over the counter, but we cannot be certain that indeed VVC was
diagnosed. Other databases pose different challenges, but all are
subject to vagaries in diagnosis; in practice VVC is rarely confirmed
microbiologically. Ideally, VVC and RVVC incidence would be esti-
mated by following a large cohort of women representative of the
general population using uniform criteria for diagnosis that in-
cludes laboratory confirmation. Appropriately powered cohort
studies following women with first VVC through multiple re-
currences using uniform diagnostic criteria that includes microbi-
ologic confirmation and assessment of risk factors would greatly
enhance our understanding of the natural history of RVVC.
In summary, the burden of VVC and RVVC on the population is
high, and the prevalence varies by age and geographic area sug-
gesting there may be modifiable risk factors. The burden probably
varies with age, but further primary studies are required to quantify
that burden across age groups. Although the individual impact of
VVC can be low or negligible, for women experiencing RVVC, the
physical, psychological, and monetary costs can be substantial. In-
terventions to reduce risk of VVC and of RVVC following VVC would
be most welcome.
Acknowledgments
This work was funded by NovaDigm Therapeutics, Brookline,
Massachusetts. The authors thank Anna Cronenwett for secretarial
support.
References
[1] Mitchell H. Vaginal dischargeecauses, diagnosis, and treatment. BMJ
2004;328:1306e8.
[2] Sobel J. Vaginitis. N Engl J Med 1997;337:1896e903.
[3] Hurley R, de Louvois J. Candida vaginitis. Postgrad Med J 1979;55:645e7.
[4] Sobel JD, Faro S, Force RW, Foxman B, Ledger WJ, Nyirjesy PR, et al. Vulvo-
vaginal candidiasis: epidemiologic, diagnostic, and therapeutic considerations.
Am J Obstet Gynecol 1998;178:203e11.
[5] Dovnik A, Golle A, Novak D, Arko D, Takac I. Treatment of vulvovaginal
candidiasis: a review of the literature. Acta Dermatovenerologica Alp
2015;24:5e7.
[6] Foxman B, Barlow R, D’Arcy H, Gillespie B, Sobel JD. Self reported incidence
and associated costs. Sex Transm Dis 2000;27:230e5.
[7] Foxman B, Muraglia R, Dietz J-P, Sobel JD, Wagner J. Prevalence of recurrent
vulvovaginal candidiasis in 5 European countries and the United States: re-
sults from an internet panel survey. J Low Genit Tract Dis 2013;17:340e5.
[8] Corsello S, Spinillo A, Osnego G, Penna C, Guaschino S, Beltrame A, et al.
Epidemiological survey of VVC in Italy. Eur J Obstet Gynecol Reprod Biol
2003;11:66e72.
[9] Richter SS, Galask RP, Messer SA, Hollis RJ, Diekema DJ, Pfaller MA. Antifungal
susceptibilities of Candida species causing vulvovaginitis and epidemiology of
recurrent cases. J Clin Microbiol 2005;43:2155e62.
[10] Giraldo PC, Polpeta NC, Juliato CRT, Yoshida LP, do Amaral RLG, Junior JE.
Evaluation of sexual function in brazilian women with recurrent vulvovaginal
candidiasis and localized provoked vulvodynia. J Sex Med 2012;9:805e11.
[11] Ehrström S, Kornfeld D, Rylander E. Perceived stress in women with recurrent
vulvovaginal candidiasis. J Psychosom Obstet Gynaecol 2007;28:169e76.
[12] Aballéa S, Guelfucci F, Wagner J, Khemiri A, Dietz J-P, Sobel J, et al. Subjective
health status and health-related quality of life among women with Recurrent
Vulvovaginal Candidosis (RVVC) in Europe and the USA. Health Qual Life
Outcomes 2013;11:169.
[13] Sobel JD. Vulvovaginal candidosis. Lancet 2007;369:1961e71.
[14] Goldman L. Goldman-Cecil Medicine. 25th ed. Philadelphia: Elsevier Saun-
ders; 2016.
[15] Crum CP, Laury AR, Hirsch MS, Quick CM, Peters W. Gynecologic and Obstetric
Pathology. Philadelphia: Saunders/Elsevier; 2016.
[16] Decherney A, Nathan L, Laufer N, Roman A. Current Diagnosis & Treatment
Obstetrics and Gynecology. 11th ed. Columbus, Ohio: McGraw-Hill Com-
panies; 2013.
[17] Liu H. Co-regulation of pathogenesis with dimorphism and phenotypic
switching in Candida albicans, a commensal and a pathogen. Int J Med
Microbiol 2002;292:299e311.
[18] Sobel JD. Recurrent vulvovaginal candidiasis. Am J Obs Gyn 2015;214:15e21.
581
[19] Horowitz BJ, Edelstein SW, Lippman L. Sexual transmission of Candida. Obstet
Gynecol 1987;69:883e6.
[20] Geiger A, Foxman B. Risk factors for vulvovaginal candidiasis: a case-control
study among university students. Epidemiology 1996;7:182e7.
[21] Yue XA, Chen P, Tang Y, Wu X, Hu Z. The dynamic changes of vaginal
microecosystem in patients with recurrent vulvovaginal candidiasis: a retro-
spective study of 800 patients. Arch Gynecol Obstet 2015;292:1285e94.
[22] Amouri I, Sellami H, Borji N, Abbes S, Sellami A, Cheikhrouhou F, et al.
Epidemiological survey of vulvovaginal candidosis in Sfax, Tunisia. Mycoses
2011;54:e499e505.
[23] Hong E, Dixit S, Fidel PL, Bradford J, Fischer G. Vulvovaginal candidiasis as a
chronic disease: diagnostic criteria and definition. J Low Genit Tract Dis
2014;18:31e8.
[24] Nedovic B, Posteraro B, Leoncini E, Ruggeri A, Amore R, Sanguinetti M, et al.
Polymorphism and vulvovaginal candidiasis: a systematic review and meta-
analysis. Biomed Res Int 2014;2014.
[25] Rosentul DC, Delsing CE, Jaeger M, Plantinga TS, Oosting M, Costantini I, et al.
Gene polymorphisms in pattern recognition receptors and susceptibility to
idiopathic recurrent vulvovaginal candidiasis. Front Microbiol 2014;5:483.
http://dx.doi.org/10.3389/fmicb.2014.00483.
[26] Jaeger M, Carvalho A, Cunha C, Plantinga TS, van de Veerdonk F, Puccetti M,
et al. Association of a variable number tandem repeat in the NLRP3 gene in
women with susceptibility to RVVC. Eur J Clin Microbiol Infect Dis
2016;35:797e801.
[27] Fromtling RA. Overview of medically important antifungal azole derivatives.
Clin Microbiol Rev 1988;1:187e217.
[28] MacNeill C, Weisz J, Carey JC. Clinical resistance of recurrent Candida albicans
vulvovaginitis to fluconazole in the presence and absence of in vitro resis-
tance. J Reprod Med 2003;48:63e8.
[29] Marchaim D, Lemanek L, Bheemreddy S, Kaye KS, Sobel JD. Fluconazole-
resistant Candida albicans vulvovaginitis. Obstet Gynecol 2012;120:1407e14.
[30] Lipsky MS, Waters T. The “Prescription-to-OTC switch” movement. Arch Fam
Med 1999;8:297e300.
[31] National Center for Health Statistics. National Ambulatory Medical Care Sur-
vey. 2012.
[32] Nyirjesy P, Peyton C. Causes of chronic vaginitis: analysis of a prospective
database of affected women. Obstet Gynecol 2006;108:1185e91.
[33] Irving G, Miller D, Robinson A, Reynolds S, Copas AJ. Psychological factors
associated with recurrent vaginal candidiasis: a preliminary study. Sex
Transm Infect 1998;74:334e8.
[34] L.E.K. Consulting LLC. L.E.K. Consulting LLC, NDV-3 US and EU3 Opportunity
Assessment Final Presentation., 2014.
[35] Sobel JD, Subramanian C, Foxman B, Fairfax M, Gygax SE. Mixed vaginitis-
more than coinfection and with therapeutic implications. Curr Infect Dis Rep
2013;15:104e8.
[36] Lindner J, Plantema F, Hoogkamp-Korstanje JA. Quantitative studies of the
vaginal flora of healthy women and of obstetric and gynaecological patients.
J Med Microbiol 1978;11:233e41.
[37] Giraldo P, von Nowaskonski A, Gomes FA, Linhares I, Neves NA, Witkin SS.
Vaginal colonization by Candida in asymptomatic women with and without a
history of recurrent vulvovaginal candidiasis. Obstet Gynecol 2000;95:413e6.
[38] Beigi RH, Meyn LA, Moore DM, Krohn MA, Hillier SL. Vaginal yeast coloniza-
tion in nonpregnant women: a longitudinal study. Obstet Gynecol
2004;104:926e30.
[39] Cotch MF, Hillier SL, Gibbs RS, Eschenbach DA. Epidemiology and outcomes
associated with moderate to heavy Candida colonization during pregnancy.
Vaginal Infections and Prematurity Study Group. Am J Obstet Gynecol
1998;178:374e80.
[40] Akinbiyi AA, Watson R, Feyi-Waboso P. Prevalence of Candida albicans and
bacterial vaginosis in asymptomatic pregnant women in South Yorkshire,
United Kingdom. Outcome of a prospective study. Arch Gynecol Obstet
2008;278:463e6.
[41] Farr A, Kiss H, Holzer I, Husslein P, Hagmann M, Petricevic L. Effect of
asymptomatic vaginal colonization with Candida albicans on pregnancy
outcome. Acta Obstet Gynecol Scand 2015;94:989e96.
[42] Workowski KA, Bolan GA. Sexually transmitted diseases treatment guidelines.
MMWR Recomm Rep 2015;64:1e137.
[43] Bieber EJ, Danfilippo JS, Horowitz IR. Clinical gynecology. 1st ed. London, UK:
Churchill Livingstone; 2006.
[44] Klausner J, Hook E. Current diagnosis & treatment of sexually transmitted
diseases. New York: McGraw-Hill; 2007.
[45] Norwitz E, Arulkumaran S, Symonds IM, Fowlie A. Oxford american handbook
of obstetrics and gynecology. New York: Oxford University Press Inc; 2007.
[46] August EV, Fischer C. Obstetrics & gynecology correlations and clinical sce-
narios. 1st ed. New York: McGraw-Hill Education/Medical; 2013.
[47] Smith RP. Netter’s obstetrics and gynecology. 2nd ed. Philadelphia: Saunders
Elsevier; 2008.
[48] Smith MA, Shimp LA, Schrager S. Family medicine: ambulatory care and
prevention. 6th ed. McGraw-Hill Education/Medical; 2014.
[49] Engberts MK, Korporaal H, Vinkers M, van Belkum A, van Binsbergen J, Lagro-
Janssen T, et al. Vulvovaginal candidiasis: diagnostic and therapeutic ap-
proaches used by Dutch general practitioners. Eur J Gen Pract 2008;14:30e3.
[50] Esim Buyukbayrak E, Kars B, Karsidag AYK, Karadeniz BI, Kaymaz O, Gencer S,
et al. Diagnosis of vulvovaginitis: comparison of clinical and microbiological
diagnosis. Arch Gynecol Obstet 2010;282:515e9.
582 F. Blostein et al. / Annals of Epidemiology 27 (2017) 575e582
[51] Mylonas I, Bergauer F. Diagnosis of vaginal discharge by wet mount micro-
scopy: a simple and underrated method. Obstet Gynecol Surv
2011;66:359e68.
[52] Nyirjesy P. Vulvovaginal candidiasis and bacterial vaginosis. Infect Dis Clin
North Am 2008;22:637e52. vi.
[53] Ilkit M, Guzel AB. The epidemiology, pathogenesis, and diagnosis of vulvo-
vaginal candidosis: a mycological perspective. Crit Rev Microbiol
2011;37:250e61.
[54] Khosravi A, Shokri H, Savadi R, Niroumanesh S, Daieghazvini R. A comparison
study between the direct agglutination test and conventional methods in the
diagnosis of vulvovaginal candidiasis. Comp Clin Path 2011;20:639e45.
[55] Mahmoudi Rad M, Zafarghandi AS, Amel Zabihi M, Tavallaee M, Mirdamadi Y.
Identification of Candida species associated with vulvovaginal candidiasis by
multiplex PCR. Infect Dis Obstet Gynecol 2012;2012:872169. http://
dx.doi.org/10.1155/2012/872169.
[56] Mendling W, Brasch J, Cornely OA, Effendy I, Friese K, Ginter-Hanselmayer G,
et al. Guideline: vulvovaginal candidosis (AWMF 015/072), S2k (excluding
chronic mucocutaneous candidosis). Mycoses 2015;58(Suppl 1):1e15.
[57] Cartwright CP, Lembke BD, Ramachandran K, Body BA, Nye MB, Rivers CA.
Comparison of nucleic acid amplification assays with BD Affirm VPIII for
diagnosis of vaginitis in symptomatic women. J Clin Microbiol
2013;51:3694e9.
[58] Wiesenfeld HC, Macio I. The infrequent use of office-based diagnostic tests for
vaginitis. Am J Obstet Gynecol 1999;181:39e41.
[59] Schaaf VM, Perez-Stable EJ, Borchardt K. The limited value of symptoms and
signs in the diagnosis of vaginal infections. Arch Intern Med
1990;150:1929e33.
[60] Linhares IM, Witkin SS, Miranda SD, Fonseca AM, Pinotti JA, Ledger WJ. Dif-
ferentiation between women with vulvovaginal symptoms who are positive
or negative for candida species by culture. Infect Dis Obstet Gynecol
2001;9:221e5.
[61] Landers DV, Wiesenfeld HC, Heine RP, Krohn MA, Hillier SL. Predictive value of
the clinical diagnosis of lower genital tract infection in women. Am J Obstet
Gynecol 2004;190:1004e10.
[62] Anderson MR, Klink K, Cohrssen A. Evaluation of vaginal complaints. JAMA
2004;291:1368e79.
[63] Bradley C, Blenkinsopp A. Over the counter drugs. The future for self medi-
cation. BMJ 1996;312:835e7.
[64] Johnson SR, Griffiths H, Humberstone FJ. Attitudes and experience of women
to common vaginal infections. J Low Genit Tract Dis 2010;14:287e94.
[65] Gonzales E, Grillo J. Over-the-Counter histamine 2 bocker therapy. Ann Pharm
1994;28:392e5.
[66] Lipsky MS, Taylor C. The use of over-the-counter antifungal vaginitis prepa-
rations by college students. Fam Med 1996;28:493e5.
[67] Ferris DG, Dekle C, Litaker MS. Women’s use of over-the-counter antifungal
medications for gynecologic symptoms. J Fam Pract 1996;42:595e600.
[68] Ferris DG, Nyirjesy P, Sobel JD, Soper D, Pavletic A, Litaker MS. Over-the-
counter antifungal drug misuse associated with patient-diagnosed vulvova-
ginal candidiasis. Obstet Gynecol 2002;99:419e25.
[69] Foxman B, Marsh JV, Gillespie B, Sobel JD. Frequency and response to vaginal
symptoms among White and African American women: results of a random
digit dialing survey. J Womens Health 1998;7:1167e74.
[70] Gunther LSA, Martins HPR, Gimenes F, Abreu ALP, Consolaro MEL,
Svidzinski TIE. Prevalence of Candida albicans and non-albicans isolates from
vaginal secretions: comparative evaluation of colonization, vaginal candidiasis
and recurrent vaginal candidiasis in diabetic and non-diabetic women. Sao
Paulo Med J 2014;132:116e20.
[71] Grigoriou O, Baka S, Makrakis E. Prevalence of clinical vaginal candidiasis in a
university hospital and possible risk factors. Eur J Obstet Gynecol Reprod Biol
2006;126:121e5.
[72] Newman SC. Biostatistical methods in epidemiology. New York: John Wiley
and Sons; 2003.
 F. Blostein et al. / Annals of Epidemiology 27 (2017) 575e582 582.e1

Appendix

Figure A1. Estimated sensitivity estimates (relative to culture) of vulvovaginal candidiasis diagnostic techniques from the scientific literature (see methods).

Figure A2. Estimated specificity (relative to culture) estimates of vulvovaginal candidiasis diagnostic techniques from the scientific literature (see methods).
582.e2 F. Blostein et al. / Annals of Epidemiology 27 (2017) 575e582

Figure A3. Prevalence estimates of self-reported physician-diagnosed vulvovaginal candidiasis per 100 population: results of a systematic literature review (1980e2016; see
methods).
 F. Blostein et al. / Annals of Epidemiology 27 (2017) 575e582
References for Figures A1-A3
[1] Abbott J. Clinical and microscopic diagnosis of vaginal yeast infection: a pro-
spective analysis. Ann Emerg Med 1995;25:587e91.
[2] Carlson P, Richardson M, Paavonen J. Evaluation of the Oricult-N dipslide for
laboratory diagnosis of vaginal candidiasis. J Clin Microbiol 2000;38:1063e5.
[3] Engberts MK, Goedbloed AF, van Haaften M, Boon ME, Heintz PM. Microscopic
diagnosis of vulvovaginal candidiasis in stained vaginal smears by Dutch
general practitioners. Acta Cytol 2007;51:882e5.
[4] Foxman B, Marsh JV, Gillespie B, Sobel JD. Frequency and response to vaginal
symptoms among White and African American women: results of a random
digit dialing survey. J Womens Health 1998;7:1167e74.
[5] Foxman B, Muraglia R, Dietz J-P, Sobel JD, Wagner J. Prevalence of recurrent
vulvovaginal candidiasis in 5 European countries and the United States: re-
sults from an internet panel survey. J Low Genit Tract Dis 2013;17:340e5.
582.e3
[6] Johnson SR, Griffiths H, Humberstone FJ. Attitudes and experience of women
to common vaginal infections. J Low Genit Tract Dis 2010;14:287e94.
[7] Khosravi A, Shokri H, Savadi R, Niroumanesh S, Daieghazvini R. A comparison
study between the direct agglutination test and conventional methods in the
diagnosis of vulvovaginal candidiasis. Comp Clin Path 2011;20:639e45.
[8] Liu XP, Fan SR. Methylrosaniline chloride stained vaginal smears for the
diagnosis of vulvovaginal candidiasis. Int J Gynecol Obstet 2007;99:83e6.
[9] Lowe NK, Neal JL, Ryan-Wenger NA. Accuracy of the clinical diagnosis of
vaginitis compared with a DNA probe laboratory standard. Obstet Gynecol
2009;113:89e95.
[10] Marot-Leblond A, Nail-Billaud S, Pilon F, Beucher B, Poulain D, Robert R.
Efficient diagnosis of vulvovaginal candidiasis by use of a new rapid immu-
nochromatography test. J Clin Microbiol 2009;47:3821e5.
[11] Rajakumar R, Lacey CJ, Evans EG, Carney JA. Use of slide latex agglutination
test for rapid diagnosis of vaginal candidosis. Genitourin Med
1987;63:192e5.