Lack of Candida africana in Ugandan pregnant women: results from a pilot study using MALDI-ToF

Abstract

Background

Candida africana is an emergent variant that has been listed as a new species or variety within the Candida albicans complex since 2001. It has a worldwide intra-albicans complex pooled prevalence of 1.67% and varies between 0 and 8% depending on geographical region. We present the results of a pilot study on its prevalence in Uganda.

Methodology

We conducted a cross-sectional study between March and June 2023. We recruited 4 pregnant women from Mulago Specialized Women and Neonatal Hospital, 102 from Kawempe National Referral Hospital, and 48 from Sebbi Hospital. Vaginal swabs were tested using microscopy, culture and matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF).

Results

The prevalence of C. africana was zero. Out of the 103 isolates, the majority (81.553%) were identified as Candida albicans, followed by Nakeseomyces glabrata (13.592%) and Pichia kudriavzevii (1.942%). Cyberlindnera jadinii, Candida tropicalis, and Candida parapsilosis each accounted for 0.971% of the isolates.

Conclusion

The prevalence of C. africana in Uganda is zero. However, large-scale cross-sectional studies, including studies involving the collection of vaginal samples from both urban and rural settings in Uganda and the use of both MALDI-TOF- and PCR-based laboratory methods, are needed to fully describe the public health burden of C. africana infections.

Background

Candida africana is a recently described opportunistic yeast pathogen that is associated with vaginal candidiasis. It was first described in a study as an atypical chlamydospore-negative Candida albicans strain in 1995 [1] and was proposed as a new Candida species that is clearly different from typical Candia albicans isolates in 2001 [2] based on morphological, biochemical, and physiological characteristics.

In Africa, C. africana has been identified in multiple countries, including Senegal [3], Angola, Madagascar [1], Cameroon [4], Tunisia [5], Gabon [6], Nigeria [7], and Algeria [8]. Among these, Angola and Madagascar reported the highest prevalence of C. africana, with the highest prevalence of 40% reported in Madagascar and 23% in Angola. However, to the best of our knowledge, no baseline study on C. africana has been conducted in Uganda. With available published data, female genital tract specimens have been the primary source of C. africana isolates, accounting for 92.81% of all C. africana isolates [9]. C. africana has been classified as susceptible to fluconazole, voriconazole, anidulafungin, caspofungin, or micafungin based on clinical breakpoints established for closely related C. albicans by the Clinical & Laboratory Standards Institute (CLSI). According to the new breakpoints published by the European Committee on Antimicrobial Susceptibility Testing, it has also been deemed resistant to itraconazole [9].

Variability in the antifungal susceptibility profiles of C. africana and other candida species is a complex phenomenon influenced by multiple factors, including geographic location, genetic diversity, patient population types, and differences in laboratory methodologies. Moreover, differences in the prevalence of risk factors such as prior antifungal exposure, age, gestation period, comorbidities, and immunosuppression among the study populations across regions can significantly influence the susceptibility patterns of candida species. These collective factors underscore the intricate interplay of genetic, environmental, and clinical elements that contribute to the variation in susceptibility profiles of candida species. Although studies from other regions have reported overall low levels of resistance to antifungal agents among Candida africana isolates, the unique genetic and environmental factors in Uganda may influence the susceptibility profile differently.

Notably, in Uganda, 60% of pregnant women [10] and 45.4% of women of reproductive age experience at least one episode of vulvovaginal candidiasis (VVC), with approximately 72% of cases attributed to C. albicans, followed by other nonalbicans species, such as Pichia krusei, Nakaseomyces glabrata, C. parapsilosis, and C. tropicalis [11, 12]. Candida africana can cause vaginitis alone or in combination with other albicans or nonalbicans species. However, because the same treatment regimens can be used to treat both C. africana and C. albicans infections, they have rarely been studied in routine laboratory testing, implying that they could be misidentified in less suspicious cases owing to their low frequency in clinical settings. The risk of misidentification/misdiagnosis is increased by the fact that C. africana can form germ tubes and produce green colonies on Chromagar Candida, which are the phenotypic features recommended for the presumptive identification of C. albicans. This could mean that C. africana is silently misdiagnosed as C. albicans, limiting targeted treatment and increasing the risk of antifungal resistance. Thus, using culture and matrix-assisted laser desorption/ionization–time of flight (MALDI-ToF) methods, this pilot study aimed to determine the prevalence of C. africana vaginitis among pregnant women who presented to prenatal clinics at three hospitals in central Uganda.

Methods and materials

Study design

We conducted a cross-sectional study between March and June 2023.

Study population

Pregnant women receiving antenatal care at Mulago Specialized Women and Neonatal Hospital (MSWNH), Kawempe National Referral Hospital (KNRH) and Sebbi Hospital (SH).

Inclusion criteria

Pregnant women who presented with clinical symptoms of vaginal/vulva itching, vaginal discharge, a history of previous VVC infections, or on-going antifungal treatment for VVC were included in the study.

Exclusion criteria

Pregnant women with signs of vaginal bleeding were excluded from the study. This was due to the increased potential for interference with the accuracy of test results, discomfort for the patient, and increased risk of introducing pathogens into the vaginal area.

Sample size

The sample size was calculated using the Scalex SP calculator for determining the sample size of prevalence studies [13]. This Scalex SP calculator applies the following formula:

$$\:n=\:\frac{{Z}^{2}P(1-p)}{{d}^{2}}$$

where

$$\:n\:=\:Sample\:size$$

$$\:Z\:=\:Z\:statistic\:for\:95\%\:level\:significance\:\left(1.96\right)$$

$$\:P\:=\:Expected\:prevalence\:of\:C.africana\:\left(1.67\%\right)$$

$$\:d\:=\:Precision\:\left(0.03\right)$$

Given the expected prevalence of 1.67% [9], the required sample size was calculated to be 136 for a margin of error or absolute precision of ± 3% in estimating the prevalence with 95% confidence, considering a potential loss/attrition rate of 10%. With this sample size, the anticipated 95% CI was (-1.33%, 4.67%). To improve the precision of the study estimates and account for potential sample attrition during the experimental follow-ups, 154 samples were collected from pregnant women who met the inclusion criteria. Four participants were from Mulago Specialized Women and Neonatal Hospital, 48 were from Sebbi Hospital, and 102 were from Kawempe National Referral Hospital.

Sampling procedure

Pregnant women receiving antenatal care at Mulago Specialized Women and Neonatal Hospital, Kawempe National Referral Hospital, and Sebbi Hospital were recruited through convenience sampling.

Consent of participants

Informed consent was obtained from all study participants using research ethics committee-approved forms.

Ethical considerations

Ethical approval was obtained from the School of Biomedical Sciences Ethics Review Committee (SBS REC Number 00007568). Additionally, administrative clearances were obtained from the hospitals. The microbiological testing results for each participant were promptly shared with their respective study site’s clinical team within seven to ten days and before the next participant’s antenatal visit. These results were later incorporated into the clinical evaluation of the pregnant mother, as appropriate.

Demographic data, clinical data and specimen collection

Data on age, gestation period, clinical signs and symptoms were collected from all pregnant women included in the study using a structured questionnaire approved by the Research Ethics Committee. A trained and competent midwife, wearing sterile gloves, placed the expectant mother on the examination bed in a lithotomy position. She gently inserted a sterile swab into the vagina until some slight resistance was felt. At this point, she rotated the swab 8–10 times to completely soak the swab with vaginal secretions, removed the swab and placed it in a sterile tube containing AMIES transport medium.

Two swabs were collected from each participating pregnant woman. Samples with their accompanying consent forms and questionnaires were labeled with both the unique study number and hospital patient numbers generated by the respective hospital patient data management system. This enabled the linkage of the testing reports to the correct participants for clinical use. The sample and accompanying forms were sealed in a Ziploc bag and stored at 2–8 °C until transportation to the analytical laboratory. The samples were transported on the same day under safe conditions to the Microbiology Laboratory at the College of Health Sciences, Makerere University.

10% potassium hydroxide (10% KOH) Wet Preparation Examination

A drop of 10% potassium hydroxide (10% KOH) was placed at the center of the microscope slide, and a portion of the high vaginal secretion was emulsified. A cover slip was applied, and the preparation was left to stand for 3–5 min, allowing for clearing of the preparation and increasing the visibility of fungal elements. The slides were scanned at 100× total magnification, and the nature of fungal elements, blastopores, pseudohyphea, and hyphae were discerned at 400× magnification.

The demonstration of pseudomycelia in combination with clinical symptoms was used to diagnose acute vulvovaginal candidiasis.

Culture on Sabouraud Dextrose Agar and ChromAgar Candida media

The second swab was streaked on Sabouraud dextrose agar (Oxoid) supplemented with chloramphenicol. The plates were incubated at 37 °C for 24 h, after which colony growth was determined. No fungal growth was reported if there was no growth after 10 days of incubation at 37 °C.

Gram stain

Gram staining was performed on pure 24-hour SDA colonies with safranin as the counterstain and 50% acetone alcohol as the decolorizer to confirm the growth of the yeast cells.

Germ tube test

A very small portion of pure yeast colonies from SDA was suspended in 0.5 ml of human serum and incubated for 3 h at 37 °C. After incubation, a drop of the suspension was placed on a microscope slide and examined under low-power magnification to detect germ tubes. A germ tube was defined as an appendage that was half the width and 3–4 times the length of the yeast cell from which it arose and without a constriction from the point of origin of the parent cell.

Chlamydospore production test

Polysorbent (Tween) 80 was added to corn meal agar (Oxoid) to reduce surface tension and allow for the development of pseudohyphae, hyphae, and blastoconodia. One pure colony from SDA was inoculated on a plate of cornmeal agar by making two streaks on the surface of the culture medium at approximately 90°. A sterile coverslip was placed to cover the “x” cross center of the streaked lines to reduce tension, and the samples were incubated at 37 °C for 4 days. On the fourth day, the streaked and covered areas were examined under a microscope at a magnification of 400X to detect the presence of blastoconidia, pseudohyphae, hyphae, and chlamydospores.

Phenotypic species identification using Chromagar Candida media

For the initial detection, differentiation, and presumptive identification of Candida species, pure colonies of SDA were subcultured on a split plate of ChromoBio^® Candida, a selective chromogenic culture medium. The plates were incubated under aerobic conditions at 35–37 °C for 24 h. Species identification was determined using the characteristic color of the colony as follows: Candida albicans, green; Candida tropicalis, metallic blue; Candida krusei, pink or fuzzy; Candida kefyr; Candida glabrata, Mauve/brown; and other species, white to mauve.

Matrix-assisted laser desorption ionization-time of flight (MALDI-ToF)

An 18-hour yeast colony previously subcultured on sheep blood agar was homogenized by adding a small amount of sterile distilled water and ethanol and then spotted on a stainless-steel target plate. After air drying and overlaying with formic acid and a-cyano-4-hydroxycinnamic acid (HCCA) matrix, the samples were loaded on the Bruker Biotyper platform for analysis (Instrument ID 8604832.35071, server version 4.1.100 (PYTH) 174 2019-06-158_06-16-099). The identification spectrum was produced from 40 laser shots in duplicate and compared to the equipment spectrum library. The yeast identification results were scored using the following scales: 0.00–1.69, indicating no organism identification possible; 1.70–1.99, indicating low confidence identification; and 2.00–3.00, indicating high confidence identification of the yeast organism. The consistency of identification of yeast organisms was divided into three categories; A, B and C. A For high consistency between the best matched and the second-best matched organism, B for low consistency where best match was a high or low confidence identification and the second-best match was a high or low confidence identification in which the genus was identical to best match. Category C occurred when there was no consistency between the best match and the second-best match. All samples that were classified as category C were subjected to reanalysis. The NCBI database was used for matching yeast organisms and Candida africana; NCBI: txid241526 was used as the reference strain for matching isolates in the Bruker Biotyper database.

Quality Control

SDA culture, Gram staining of the isolate, chromogenic identification on Chromagar, germ tube tests, chlamydospore production on cornmeal agar, and MALDI-ToF MS were quality controlled using C. albicans ATCC 10,231.

Results

Microscopy

Seventy-nine samples had fungal elements, namely, blastospores, hyphae or pseudohyphea, while 75 had no fungal elements under 10% KOH wet preparation, as shown in the Table 1 .

Table 1 Examination of the wet preparation of 10% potassium hydroxide

Identification of Candida species by phenotypic methods

One hundred and three of the 154 samples yielded fungal growth on Sabouraud dextrose agar, and 86 samples produced a green color on ChromAgar Candida media, as shown in Table 2 .

Table 2 Chromogenic appearance of yeast colonies

Eighty-five of the 86 green colonies from ChromAgar candida media produced pseudohyphae and chlamydospores on CMA and subsequently formed germ tubes. However, one isolate, which was green on ChromAgar Candida, failed to form germ tubes, and neither produced chlamydospores on corn mean agar as shown in Table 3 .

Table 3 Results of chlamydospores and germ tube tests for yeast isolates based on color appearance on chromagar candida media

Identification of yeasts by MALDI-TOF MS

MALDI-TOF was used to identify 98% of the isolates with a mean score of 2 and 2% with a mean score of 1.70–1.99. Out of the 103 isolates, the majority (81.553%) were identified as Candida albicans, followed by Nakeseomyces glabrata (13.592%) and Pichia kudriavzevii (1.942%). Cyberlindnera jadinii, Candida tropicalis, and Candida parapsilosis each accounted for 0.971% of the isolates. Candida africana was not isolated, and its prevalence was zero. The green colony-forming isolate on ChromAgar that was chlamydospore negative and germ tube negative was identified as C. albicans by MALDI-ToF.

Discussion

This study aimed to provide baseline data on the prevalence of C. africana in Uganda using pregnant women as the reference population. No C. africana was isolated, and C. albicans was the predominant species isolated. The absence of C. africana vaginitis in this pilot study was similar to findings from other studies conducted in Argentina [14], Malaysia [15], and Turkey [16], despite these studies having used PCR methods as opposed to the MALDI-ToF method used in this study. This was the first time that MALDI-ToF has been used to identify yeast species that colonize or cause infection in the vaginal area. The choice to use MALDI-TOF mass spectrometry instead of PCR to identify the Candida species isolated in this study was based on several practical considerations and limitations, most notably the financial and infrastructural limitations faced by the majority of diagnostic laboratories in Sub-Saharan Africa. PCR of the Hyphal Wall Protein 1 gene typically requires the specific primers CR-f (5’- GCT ACC ACT TCA GAA TCA TCATC-3’) and CR-r (5’- GCA CCT TCA GTC GTA GAG ACG-3’), yeast gene extraction kits/reagents, equipment and extensive laboratory facilities for genome extraction, purification, amplification and electrophoresis which are costly [17,18,19,20,21]. In contrast, MALDI-ToF has a fast one-step turnaround time, eliminating the long DNA extraction, amplification, and analysis steps associated with PCR, making it a more cost-effective and accessible choice for this study. This method has previously demonstrated effectiveness in accurately identifying C. africana. For instance, Andeme et al. (2014) used MALDI-ToF to isolate nine C. africana isolates from 62 strains that were initially identified as C. albicans [6]. Later, MALDI-ToF MS profiling was applied to identify the first atypical Colombian C. africana clinical isolates in Columbia [22]. The use of this technique to identify the candida species isolated in this study was further supported by a review of methods for identifying clinically important cryptic candida species conducted by Criseo, Scordino, & Romeo (2015) [20], who concluded that traditional mycological techniques are not sufficient for accurately distinguishing all species belonging to a specific cryptic complex of candida, who recommended the use of MALDI-ToF and PCR techniques. This is further evidenced by its successful application in a study on the minor species C. africana, C. stellatoidea, and C. dubliniensis in the Candida albicans complex among HIV-infected patients in Yaounde (Cameroon) [4].

The results of the Chlamydospores production test showed that the majority of C. albicans isolates produced chlamydospores. This was consistent with the established characteristics of C. albicans, which is known for its ability to form chlamydospores under certain conditions [23]. In contrast, other isolated species, such as Nakeseomyces glabrata, did not produce chlamydospores, consistent with their known morphological charateteritics where they lack the dimorphic ability to switch between yeast and filamentous forms [24]. The Germ Tube test also yielded a high positivity rate for C. albicans, with 85 out of 86 isolates testing positive. This test is a critical phenotypic marker for identifying C. albicans and its level of agreement with Chromogenic agar for the identification of C.albicans seen in this study had been demonstrated before [25].

However, like phenotypic methods, MALDI-ToF is not without its limitations. The successful identification of microorganisms using MALDI-ToF MS relies heavily on a database containing the spectra of known organisms. One of the critical factors for successful MALDI-ToF MS is having a sufficient number of isolates for each species grown under various conditions to ensure that the spectral library for the organism is robust enough to account for the inherent variability expected for any organism under different environmental conditions [26]. Unfortunately, in the case of the similarities between C. albicans and C. africana and the limited number of spectra of C. africana in the NCBI database used in this study, there was a potential for poor discrimination between the two species. Furthermore, the relatively small sample size used in this study could have limited our capacity to isolate less common Candida species, such as C. africana.

Therefore, future research should address these limitations by employing a more representative sample, considering a broader demographic range (urban and rural settings), and utilizing both MALDI-ToF and PCR-based techniques or whole-genome sequencing to determine the prevalence of C. africana in Uganda.

Conclusion

The prevalence of C. africana in Uganda is zero. However, large-scale cross-sectional studies, including studies involving the collection of vaginal samples from urban rural settings, all geographical regions in Uganda, and studies involving the use of both MALDI-ToF and PCR-based laboratory methods, are needed to fully describe the local public health burden of C. africana vaginitis infections.