ABSTRACT
The treatment of vulvovaginal candidiasis (VVC) due to Candida glabrata is challenging, with limited therapeutic options. Unexplained disappointing clinical efficacy has been reported with systemic and topical azole antifungal agents in spite of in vitro susceptibility. Given that the vaginal pH of patients with VVC is unchanged at 4 to 4.5, we studied the effect of pH on the in vitro activity of 11 antifungal agents against 40 C. glabrata isolates and compared activity against 15 fluconazole-sensitive and 10 reduced-fluconazole-susceptibility C. albicans strains. In vitro susceptibility to flucytosine, fluconazole, voriconazole, posaconazole, itraconazole, ketoconazole, clotrimazole, miconazole, ciclopirox olamine, amphotericin B, and caspofungin was determined using the CLSI method for yeast susceptibility testing. Test media were buffered to pHs of 7, 6, 5, and 4. Under conditions of reduced pH, C. glabrata isolates remained susceptible to caspofungin and flucytosine; however, there was a dramatic increase in the MIC90 for amphotericin B and every azole drug tested. Although susceptible to other azole drugs tested at pH 7, C. albicans strains with reduced fluconazole susceptibility also demonstrated reduced susceptibility to amphotericin B and all azoles at pH 4. In contrast, fluconazole-sensitive C. albicans isolates remained susceptible at low pH to azoles, in keeping with clinical observations. In selecting agents for treatment of recurrent C. glabrata vaginitis, clinicians should recognize the limitations of in vitro susceptibility testing utilizing pH 7.0.
INTRODUCTION
Vulvovaginal candidiasis (VVC) accounts for up to one third of all vaginitis cases presenting to gynecologists (1, 2, 3). VVC is most commonly caused by Candida albicans but can also be caused by non-albicans Candida species, with Candida glabrata being the most common (24, 26). Symptomatic C. glabrata vaginitis poses a significant problem for clinicians because effective treatment and eradication of C. glabrata from the vagina have proven difficult (23, 25, 26). The organism has variable intrinsic resistance to azole drugs (23, 25, 26). C. glabrata vaginitis has been moderately successfully treated with boric acid, but this is not curative in one third of patients (8, 23, 25, 26). Other therapies have been advocated, such as topical flucytosine, oral itraconazole, and nystatin suppositories (8, 18). Amphotericin B suppositories in patients with non-albicans Candida resistant to azoles were studied by Phillips and found to be promising; however, symptomatic C. glabrata vaginitis is often unresponsive to these regimens (11, 18). VVC is also occasionally caused by fluconazole-resistant C. albicans, posing a similar treatment dilemma in that susceptibility of these organisms to other azole and non-azole drugs is not clinically predictive (26).
Drug treatment of vaginal infections may be unique in that the normal pH of the vagina is 4 to 4.5, which remains unchanged during VVC (13). Previous studies have found that the test medium pH in in vitro susceptibility testing can alter the azole MIC for Candida species and that an acidic pH tends to increase the MICs of fluconazole for selected Candida species (16). However, it was concluded that more acidic conditions did not change the designation of the isolates from susceptible to resistant, neither were clinical implications evident. The purpose of this study was to determine whether a change in test medium pH had an effect on in vitro susceptibility of C. glabrata and both fluconazole-susceptible and reduced-susceptibility C. albicans to seven azole and four non-azole antifungal agents, in order to explain the frequent in vivo failure of these agents in women with vaginitis caused by C. glabrata.
MATERIALS AND METHODS
Vaginal isolates of C. glabrata and C. albicans were chosen from the Wayne State Vaginitis Clinic microbiology laboratory organism bank. The definition of fluconazole-susceptible C. albicans was the presence of an MIC of ≤2 μg/ml, and reduced susceptibility was defined as an MIC of ≥4 μg/ml (4). Vaginal isolates were randomly chosen from the years 2000 to 2010 and plated on CHROMagar to verify purity of culture. These plates were incubated for 48 h at 37°C in ambient air. Susceptibility testing was then performed using a broth microdilution method, according to CLSI document M27-A3 (2008) guidelines utilizing pH 7 (4). Antifungals and concentrations tested were flucytosine and fluconazole (at MIC ranges of 0.125 to 64 μg/ml), and voriconazole, posaconazole, itraconazole, ketoconazole, clotrimazole, miconazole, ciclopirox olamine, amphotericin B, and caspofungin (all with MIC ranges of 0.03 to 16 μg/ml). C. albicans isolates known to be fluconazole susceptible (MIC, ≤ 2 μg/ml) were not tested against itraconazole, ketoconazole, clotrimazole, and miconazole. A 0.1-ml yeast inoculum of 1.5 (± 1.0) ×103 cells/ml in RPMI 1640 medium was added to each microdilution well. The trays were then incubated at 35°C for 48 h in ambient air. The MICs were read as the lowest antifungal concentration with substantially lower turbidity (80% growth reduction) compared to growth in the antifungal-free growth well for all agents. Testing known ATCC strains of Candida parapsilosis and Candida krusei ensured quality control. Antifungal susceptibility testing was carried out for each isolate at pH 6, 5, and 4 using a MOPS (morpholinepropanesulfonic acid) (Sigma-Aldrich) buffer solution, and MIC ranges, medians, MIC50s, and MIC90s were compared.
RESULTS
A total of 40 vaginal strains of C. glabrata and 15 fluconazole-sensitive and 10 reduced-fluconazole-susceptibility C. albicans strains were studied, and MICs were recorded at pH levels 7, 6, 5, and 4 for each antifungal tested. Table 1 outlines MIC50 and MIC90 susceptibility results, including ranges of antifungal agents tested for each pH value for both C. glabrata and C. albicans.
MIC50 and MIC90 susceptibility results
C. glabrata.At pH 7, all C. glabrata isolates were susceptible to flucytosine, amphotericin B, caspofungin, and ciclopirox olamine. In contrast, a range of in vitro activity was present for the various azole agents. The MIC90 for fluconazole was 32 μg/ml (range 2 to >64 μg/ml) with considerably lower MICs for all other azoles tested. Notably low MICs were documented for posaconazole and voriconazole at 0.5 μg/ml. Itraconazole and ketoconazole were highly active at pH 7 and the topical agents clotrimazole and miconazole were similarly active.
With progressive reduction in pH, MIC90 values for 5-fluconazole and caspofungin were unchanged; however, an increase in MIC was evident for amphotericin B and to a lesser extent ciclopirox olamine. A dramatic increase in MIC90 was evident for all azoles tested to drug levels achievable in the vagina with systemic azole use, although pharmacologic data are not available. The trends observed for MIC90 were also reflected in MIC50 values.
C. albicans.At pH 7, fluconazole-susceptible strains of C. albicans were predictably susceptible to all antifungal agents tested. With a decrease in pH, a significant increase in MIC was evident only with amphotericin B and ciclopirox olamine. Azole activity at the lower pH was maintained in the fluconazole-susceptible isolates.
At pH 7, 10 vaginal isolates of fluconazole-reduced-susceptibility C. albicans were evaluated. The MIC range for fluconazole activity was 4 to >64 μg/ml, with MIC90 being 4 μg/ml. These isolates remained susceptible to all other azole drugs tested but demonstrated a moderately higher MIC to flucytosine (MIC, 2 μg/ml). In contrast, when a lower pH was tested, dramatic increases in MIC were seen for flucytosine, amphotericin B, fluconazole, posaconazole, voriconazole, itraconazole, and ketoconazole.
DISCUSSION
The results of this study reveal that different classes of antifungals and the two species of Candida studied in vitro behaved differently with decrease in pH. The results confirm the susceptibility of fluconazole-sensitive C. albicans isolates to all azoles and the variable resistance of C. glabrata to fluconazole, and they may also offer insight as to why some antifungal medications may not be as effective in vivo with a more acidic physiologic vaginal pH. Previous studies similarly found that the medium pH can alter azole MICs for Candida species, and specifically an acidic pH was reported to increase the MICs of fluconazole for selected Candida species (16). The clinical implications of this observation were not, however, recognized.
C. glabrata vaginal infection is by no means infrequent, but case numbers are insufficient to perform a randomized controlled trial in order to establish optimal treatment (23, 25, 26). The resistance of C. glabrata to fluconazole, at all pH levels, observed in the present study is consistent with numerous in vitro studies (19, 20) and reflects experience when treating vulvovaginal candidiasis (7, 20, 25) and bloodstream infections (12, 19). Posaconazole and voriconazole are frequently but not invariably active against fluconazole-resistant C. glabrata. The Candida surveillance study demonstrated that resistance to fluconazole was highly predictive for resistance to voriconazole (14). Sabatelli et al. studied 1,218 C. glabrata isolates and their resistance to different azoles and amphotericin B, concluding that isolates with elevated MICs to one azole were generally less susceptible to all azoles (22). An important new finding in the present study reveals that C. glabrata isolates resistant to fluconazole but susceptible to posaconazole and voriconazole at pH 7 are unlikely to be effectively treated in vivo given the dramatic increases in MIC to these drugs at pH 4 and 5.
The topical agents miconazole and clotrimazole, which achieve high local concentrations, are similarly likely to be ineffective. This conclusion is strongly supported by clinical experience (27). In contrast, flucytosine maintained activity at low pH, a finding that supports experience in successfully treating C. glabrata-affected woman with symptomatic vaginitis (25).
Topical amphotericin B has in small studies demonstrated effectiveness for treatment of non-albicans Candida vaginitis, but emerging resistance to this antifungal has been documented (11, 18). Topical amphotericin B has also been used in combination with other antifungals, such as flucytosine. If flucytosine is as effective as previously described and stable at a low pH, perhaps it is contributing much more than amphotericin B to successful treatment (25). It was found in the present study that for both Candida species, amphotericin B activity was profoundly affected by pH, with at least a 16-fold increase in MIC90 with decrease in pH.
Ciclopirox olamine is an agent applied topically and well known for its potency against dermatophytes, and it has been suggested as an antifungal for resistant VVC. It is a synthetic topical agent, widely used to treat onychomycosis, tinea pedis, pityriasis versicolor, and seborrheic dermatitis. Its use in treating vaginal candidiasis has also been studied, with limited success (9, 17, 21), and it has shown clinical promise against azole-resistant Candida species, including C. glabrata. It has demonstrated good topical and systemic tolerance in rats and rabbits when vaginal tissue was examined (5, 15) and has been studied in settings with a lower pH (9, 10). However, in this study, a 4-fold rise in MIC90 from 0.5 to 2 μg/ml with a decrease in pH was seen. One factor that limits the clinical application of these data is that the breakpoint of ciclopirox olamine is unknown and the clinical relevance of increased MICs is questionable.
Caspofungin is an echinocandin that has demonstrated activity against Candida species both in vitro and in vivo for systemic infections (6). None of the echinocandins are available as topical agents, and they have not yet been studied for vulvovaginal candidiasis or at decreased pH levels. The results of this study demonstrated stable MICs with a decrease in pH, with all C. albicans isolates having an MIC90 of less than 2 μg/ml, and continued activity against C. glabrata isolates at lower pH. Additional studies would need to be performed to evaluate echinocandin response in vivo as a topical compound.
This in vitro study demonstrates the potential limitations of conventional in vitro testing in predicting antifungal clinical success when faced with the challenge of treating recurrent vulvovaginal C. glabrata infections as well as fluconazole-refractory C. albicans vaginitis. Although the importance of medium pH in standardizing susceptibility testing is widely recognized in recommending routine testing at pH 7, the profound effect of pH on C. glabrata susceptibility has not been appreciated but is probably relevant only to patients with yeast vaginitis. The exact mechanism of pH-induced reduced susceptibility has not been established. In contrast, fluconazole-susceptible C. albicans strains responsible for the majority of vaginitis episodes are less vulnerable to the pH influence. Finally, C. albicans vaginal isolates already demonstrating reduced azole sensitivity at pH 7 are further compromised by lowering pH, resembling the effect seen with C. glabrata. This study also emphasizes the need for new alternate agents for treatment of C. glabrata vaginitis as well as to consider measuring C. glabrata drug susceptibility in vitro at pH 4 to 5 before recommending antimycotic therapy; however, validation studies are essential.
FOOTNOTES
- Received 7 June 2011.
- Returned for modification 26 July 2011.
- Accepted 24 December 2011.
- Accepted manuscript posted online 9 January 2012.
- Copyright © 2012, American Society for Microbiology. All Rights Reserved.
