Abstract
There is a paucity of information about the factors associated with oral colonization with Candida species and the changes associated with antiretroviral (ARV) therapy. This study investigated the role of ARV therapy and other factors in a study population. Relevant clinical and laboratory information was obtained and oral rinse specimens were tested for yeast identification. The findings were compared with previous data from the same clinic before ARV therapy was available. Of 197 patients, 117 (59.4 %) were colonized. Candida albicans was the dominant species (71 %) and Candida dubliniensis was the most frequent non-albicans Candida. The colonized group had a higher rate of concurrent tuberculosis (TB) infection (77.4 % compared with 56 % in the non-colonized patients, P = 0.03) and a lower median CD4+ count (346.5 cells mm−3) compared with the non-colonized group (418 cells mm−3). Participants not on ARV therapy and those having oral prosthesis were all colonized (P = 0.003 and P = 0.022, respectively). The oral Candida count was negatively correlated with the CD4+ count in participants on ARV therapy (P = 0.006). Associated factors using logistic regression were dental caries (odds ratio = 1.30; 95 % confidence interval = 1.07–1.60] and diabetes mellitus (odds ratio = 5.52; 95 % confidence interval = 1.68–18.12). The colonization rate was higher (81.3 %) as well as the yeast count before ARV therapy was available, while the prevalence of C. dubliniensis was found to have increased from 6.3 to 11 %. Dental caries, diabetes mellitus, oral prostheses and TB infection were associated with oral colonization. The colonization rate, variety and yeast counts declined with ARV therapy.
Introduction
Several Candida species are known to colonize the human oropharyngeal tract, with Candida albicans colonization occurring most frequently. These yeasts are usually commensals but may act as opportunistic pathogens (Cannon et al., 1995). Colonization does not always lead to infection; however, it is a prelude to infection when host immunity is compromised and the risk of a disseminated infection is high (Lott et al., 1999). Such infections remain a major healthcare challenge (van der Meer et al., 2010). The prevalence of yeast colonization in the mouth has differed widely depending on the age, location and immune status of the population sampled. A prevalence of 80 % was found in human immunodeficiency virus (HIV)-positive children compared with 57.5 % in their seronegative siblings (Cerqueira et al., 2010) A previous study in our hospital found a prevalence of 81.3 % in HIV-positive adults compared with 63 % in their HIV-negative counterparts (Patel et al., 2006).
In HIV infection, higher rates of yeast colonization have been a consistent finding, thus predicting subsequent infection with oropharyngeal candidiasis (Vargas & Joly, 2002). With impaired host defence mechanisms, C. albicans assumes a more pathogenic and virulent capacity, which can lead to significant morbidity (Mercante et al., 2006). Many factors have been suggested as risk factors for oral colonization; these include diabetes mellitus, head and neck cancer, smoking, the use of oral prostheses, age, race and poor nutritional status. Others are reduced salivary flow, the use of antibiotics, immunosuppressive states and even the presence of other locally available microflora (Ellepola et al., 2011; Jabra-Rizk et al., 2001; Schelenz et al., 2011). In HIV-positive individuals, the CD4+ count, HIV viral load, non-availability or non-usage of highly active antiretroviral therapy (HAART) and the type of antiretroviral (ARV) medication have been explored as possible factors that influence oral colonization with yeasts (Back-Brito et al., 2009; Delgado et al., 2009; Wu et al., 20121). Although Wu et al. (2012) showed that the initiation of HAART reduced the rate of oral colonization by Candida, Yang et al. (2006) showed no decrease in colonization of the oral cavity. Factors influencing oral colonization could be attributed to differences in study populations, sampling techniques, use and type of ARV therapy and several other factors. Risk factors may be peculiar to geographical settings (Cerqueira et al., 2010). A previous study in South Africa described the prevalence of yeast carriage in paediatric HIV subjects (Blignaut, 2007), but none has investigated the risk factors for colonization globally and none has compared the prevalence rates of Candida colonization in the oral cavity before and after initiation of HAART.
The aim of this study therefore was to investigate the factors associated with oral colonization of Candida spp. in a cohort of HIV-positive women in South Africa and to determine the effect of the introduction of HAART on prevalence rates.
Methods
Study setting, design and subjects.
The study was conducted from January to December 2011 at the HIV clinic at the Charlotte Maxeke Johannesburg Academic Hospital, Johannesburg, South Africa. Ethical clearance was obtained from the Human Research Ethics Committee of the University of the Witwatersrand, Johannesburg. Confirmed HIV-positive women who were willing to participate were enrolled after obtaining written consent. A detailed questionnaire was administered to collect data on demographics, personal hygiene and habits, HIV treatment, and previous history of oral candidiasis. The most recent CD4+ count and HIV viral load (within 3 months) were obtained from the clinical records.
Detailed oral examination was conducted on all participants by trained and standardized oral healthcare personnel to screen for oral candidiasis and other HIV-related oral conditions. The status of oral hygiene was documented using plaque and gingival indices (Löe, 1967), and caries status was assessed using the number of decayed, missing and filled surfaces (DMFS) (WHO, 1997).
Sample collection and laboratory techniques.
Before oral examination, the concentrated oral rinse technique (Samaranayake et al., 1986) was used to collect samples for analysis. A 100 µl sample of the oral rinse was cultured on CHROMagar Candida plates (CHROMagar Microbiology) and incubated at 37 °C for 48 h. Where there was growth, colonies were counted and subcultured on Sabouraud dextrose agar for 48 h. Yeasts were then identified using the API 20C AUX system (bioMérieux) and germ-tube tests.
Statistical tests.
Continuous variables that were normally distributed were described as means±sd, whilst the non-normal variables were described using the median and interquartile range. A t test was used to compare means in the two groups for parametric data, whilst a Mann–Whitney test was applied to non-parametric data. Categorical variables were compared using χ2 statistics or Fisher’s exact test, as appropriate.
A multivariable logistic regression model was constructed with colonization as the outcome variable. The best-fit option was used for predictor selection. Non-parametric predictors such as CD4+ count and viral load were dichotomized, and collinear predictors were dropped from the model. The DMFS number was skewed but was retained as a continuous predictor. The variable was normalized with a square root transformation as suggested by the ladder of powers (Tukey, 1977). Standard procedures for model checking were adopted and statistical significance was inferred at P<0.05. Statistical analysis was performed using Stata version 12 (StataCorp).
Data from a previous study carried out in the same department (Patel et al., 2006) and in a similar group of patients attending the same HIV clinic before ARV therapy was introduced were extracted and compared with the present data. The same protocols were adopted, although the time and sample sizes were different.
Results
Colonization rate, characteristics of participants and the role of associated factors
Of 197 women analysed, 117 (59.4 %) were found to be colonized with yeasts, whilst 18 (9.14 %) had active Candida infection. The mean age (±sd) for all participants was 38.3±8.4 years. The colonized group was younger, with a mean age of 37.9±8.1 years, but this was not statistically significant (Table 1). The groups were similar when comparing level of education and history of cigarette smoking. Colonization rate was significantly higher (77.4 %) in those with concurrent or recent infection with tuberculosis (TB) compared with 56 % colonization in those without TB (P = 0.03). All those who had not started ARV therapy and all those who had oral prostheses were colonized (P = 0.003 and P = 0.022, respectively), but the period of commencement of ARV therapy and the use of protease inhibitors were not associated with the carriage of yeasts (P = 0.39 and P = 0.742, respectively). The subjective assessment of oral dryness and the plaque and gingival indices were not statistically different in the groups. The median CD4+ count (IQR) was significantly lower in the colonized group (346.5 cells mm−3; IQR, 311 cells mm−3) compared with the non-colonized group (418 cells mm−3; IQR, 357 cells mm−3; P = 0.024). However, the difference in the median viral load was not significant (P = 0.19). Candida counts were significantly higher among the colonized patients who had low CD4+ counts (Spearman’s ρ −0.3; P = 0.006) (Fig. 1). Although the median DMFS (IQR) was higher in the colonized group (19/27) compared with the non-colonized group (16/24), this was not statistically significant with bivariate analysis (P = 0.07).
IQR, Interquartile range.
Correlation of CD4+ count (cells mm−3) with C. albicans counts in patients on ARV therapy.
Characteristics of the yeasts identified
In total, 120 yeasts were isolated from 117 colonized participants. Three (2.6 %) of the colonized participants had more than one yeast species. C. albicans was the most common species, accounting for 71 %, whilst Candida dubliniensis was the most frequent non-albicans Candida species (11 %).
Multivariate analysis and comparison with data from the pre-HAART period
The multivariate logistic regression showed a younger age, the dental caries status (as measured by the DMFS number) and diabetes mellitus to be associated with colonization (Table 2). This was adjusting for CD4+ count, viral load, use of protease inhibitors, oral hygiene status and the presence of TB. Age and dental caries status, which showed no association in the bivariate analysis, showed significance in the logistic regression. CD4+ count and TB were significant in the bivariate analysis but were no longer significant after adjusting for other covariates in the regression output.
When the present data were compared with a previous study carried out before ARV therapy was introduced to our hospital, the prevalence rate was found to be much higher in the previous study (81 %). Similarly, the counts of yeasts and the variety of yeasts were also higher. However, the prevalence of C. dubliniensis (6.3 %) was lower (Table 3).
Results are shown as the percentage of patients.
Discussion
The innate immune system (polymorphonuclear leukocytes and macrophages, antimicrobial peptides) and the acquired response (an increase in circulating and mucosal IgG and mucosal IgA antibodies) work together to keep Candida in check in a colonized healthy host (Fidel, 2006). However, very early infection with HIV has been demonstrated to influence colonization with oral Candida and the development of opportunistic candidiasis (Owotade et al., 2008). This is a consequence of defective T helper 1-type CD4+ T cells and changes in saliva composition and function (Fidel, 2006).
The significant relationship between yeast carriage and recent or concurrent infection with TB in the bivariate analysis is an important finding. The TB burden in South Africa remains high (Wood et al., 2011) and TB further compromises the immune status. In addition, anti-TB therapy may predispose to the development of fluconazole-resistant C. albicans strains (Masiá Canuto et al., 2000). The odds of colonization were five times greater in patients with diabetes in our multivariate analysis, confirming previous findings (Gonçalves et al., 2006). The high content of glucose in the tissues of diabetics and impaired cell-mediated immunity have been suggested as contributing factors for yeast colonization (van der Meer et al., 2010). This finding should, however, be interpreted with caution, as there were only nine diabetics (4.5 %) in our study population.
Our observation that participants who were yet to commence ARV therapy were all yeast-positive is consistent with a recent report where the absence of ARV therapy and HAART increased the probability of yeast colonization (Cerqueira et al., 2010). The effect of prolonged HAART might be more beneficial in reducing opportunistic yeast infections rather than yeast colonization, as suggested by Yang et al. (2006). In our study, comparing the present data with what was observed in the same study location 10 years before, it was apparent that the introduction of HAART had also reduced the colonization rate, yeast count and distribution of Candida spp. The prevalence of C. dubliniensis also increased from 6.3 to 11 % over this period. In our study population, the prevalence of multiple yeast colonization was lower (2.6 %) compared with a similar study in Taiwan (18.7 %; Wu et al., 2012).
It has been postulated that the protease inhibitor group of ARV drugs inhibit Candida virulence by neutralizing secretory proteinase enzyme production (Migliorati et al., 2004). Although about 14.8 % of our subjects were on a lopinavir/ritonavir combination, there was no effect on yeast colonization. However, the low prevalence of our subjects on this combination may be responsible for the lack of effect on colonization. The median CD4+ count was lower in the colonized group, but the difference did not attain statistical significance with regression analysis, and neither did the median viral load, which was higher in the colonized group. The effect of these variables on oral colonization has been controversial. Some studies have found a relationship between a CD4+ count of ≤200 cells mm−3 and oral colonization (Delgado et al., 2009; Ohmit et al., 2003; Wu et al., 2012), whilst others have not (Cerqueira et al., 2010). Similarly, a high viral load was implicated in some reports (Delgado et al., 2009; Tappuni & Fleming, 2001), whilst others did not find any association (Cerqueira et al., 2010; Sánchez-Vargas et al., 2005).
Our observations with regard to CD4+ count and viral load may be related to the relatively consistent recovery of the immune system and reduced viral burden, judging by the mean CD4+ count and viral load for all the subjects and the period they had been on ARVs. However, within the colonized group, low CD4+ was significantly associated with the Candida count, suggesting the importance of a competent immune status in checking the proliferation of yeasts.
The status of oral hygiene and cigarette smoking did not influence the risk of colonization, and this was consistent with the observations of other investigators (Darwazeh et al., 2010a, b). However, a very strong association was observed with the caries status, even after adjusting for other predictors. Carious lesions are known reservoirs for yeasts and have been established as risk factors in oral colonization (Cerqueira et al., 2010; Jacob et al., 1998). Oral prostheses are similar reservoirs for oral colonization with yeasts, and our findings confirmed this, as all denture wearers were yeast-positive. Oral prostheses provide a means of adherence for Candida while isolating the area covered from the protective effects of saliva (Zadik et al., 2010). Although we could not directly relate oral colonization with having a dry mouth, this is known to act in concert with other factors to predispose to colonization. Malignant conditions and the management of such conditions with cytotoxic chemotherapy or radiotherapy compromise the cell-mediated host immunity needed to control fungal infections and reduce the protective effects of saliva by inducing hyposalivation.
ARV therapy was introduced in South Africa from 2004 to 2009, but to date, only half of the HIV-infected population has access to therapy (Webster et al., 2012). In 2006, when a similar study was conducted in our study population, HAART was not available. The prevalence of oral colonization was much higher, as well as the yeast count. The use of HAART has reduced opportunistic oral infections (Yang et al., 2006). Our study has shown that HAART has reduced the colonization rate of oral Candida.
It is apparent that many factors may predispose to oral carriage of yeast, and this may vary across geographical locations and with particular study characteristics. In our subjects, we found that non-implementation of ARV therapy, wearing of oral prostheses and the presence of caries and possibly TB infection may be associated with yeast carriage. More studies are needed to clarify the role of these risk factors in this population. There is also a need to emphasize the importance of adequate oral care in HIV infection and prompt commencement of ARV therapy. Efforts should also be made to increase the coverage of HIV medication in sub-Saharan Africa.
Acknowledgements
The authors wish to thank the Tertiary Education Tax Fund, the Federal Republic of Nigeria for funding the research project, the director and staff of the HIV Clinic (Area 516) of Charlotte Maxeke Johannesburg Academic Hospital, Johannesburg, South Africa, for their cooperation, and the staff of the Oral Microbiology Unit, School of Pathology and NHLS, University of the Witwatersrand for their support.