Fungal species in endodontic infections: A systematic review and meta-analysis

Fungal infections are common on oral mucosae, but their role in other oral sites is ill defined. Over the last few decades, numerous studies have reported the presence of fungi, particularly Candida species in endodontic infections, albeit in relatively small numbers in comparison to its predominant anaerobic bacteriome. Here, we review the fungal biome of primary and secondary endodontic infections, with particular reference to the prevalence and behavior of Candida species. Meta-analysis of the available data from a total of 39 studies fitting the inclusion criteria, indicate the overall weighted mean prevalence (WMP) of fungal species in endodontic infections to be 9.11% (from a cumulative total of 2003 samples), with 9.0% in primary (n = 1341), and 9.3% in secondary infections (n = 662). Nevertheless, WMP for fungi in primary and secondary infections which were 6.3% and 7.5% for culture-based studies, increased to 12.5% and 16.0% in molecular studies, respectively. The most prevalent fungal species was Candida spp. The high heterogeneity in the reported fungal prevalence suggests the need for standardized sampling, and speciation methods. The advent of the new molecular biological analytical platforms, such as the next generation sequencing (NGS), and matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF), that enables identification and quantitation of a broad spectrum of hitherto unknown organisms in endodontic infections should radically alter our understanding of the endodontic mycobiome in the future. Candida spp. appear to be co-pathogens with bacteria in approximately one in ten patients with endodontic infections. Hence, clinicians should comprehend the importance and the role of fungi in endodontic infections and be cognizant of the need to eradicate both bacteria and fungi for successful therapy.


Introduction
Fungi are common opportunistic pathogens, and they comprise a substantial proportion of the human microbiome [1]. The fungal human microbiome, termed the mycobiome has numerous constituents, but the predominant of which belongs to the genus Candida [2]. Of such as short-chain carboxylic acids, lactate and acetates that demineralize enamel. The acidification is also due to the activation of H+-ATPase-dependent sugar uptake system of C. albicans plasma membrane, and to the generation of carbon dioxide from glucose metabolism [31]. Such acid production in a localized, focal biofilm niche causes a rapid drop of the pH to 5.5 or below leading to enamel and dentine demineralization [14,19,32]. The reduction of pH also leads to the activation of acid proteases, collagenases, and phospholipases of Candida spp., that assist degradation of the dentinal collagen [25,26,33,34]. In conclusion, Candida could have a deleterious cariogenic effect on dental hard tissues through independent, dichotomous mechanisms: first, by dissolving the inorganic matrix of either enamel or dentine with its metabolic acids, and second, by disassembling the organic matrix by its abundant collagenolytic enzymes.
There are also a number of studies that points to a critical role of Candida in the pathogenesis of endodontic infections. The yeast, being a microaerophilic eukaryote possesses the metabolic armory necessary to survive within the harsh and barren ecosystem of the root canal. Several workers have shown that C. albicans could use dentin itself as a source of nutrition in vitro, in the absence of other extraneous food supplements, and colonize the canal walls as well as the dentinal tubules [29,35,36]. In one experiment, the penetration of C. albicans into dentinal tubules in vivo was shown to be facilitated by the presence of a smear layer, produced by instrumentation [37]. The sheltered life of the yeast within the tubules, out of reach of instruments, and disinfectant irrigants, is also likely to perpetuate chronic endodontic infections [6,38].
With regard to the localization of fungi within the root canal system, Nair et al. [39] elegantly demonstrated through ultrastructural imaging, the presence of yeast-like organisms in root canals, and at the apical foramen. Persoon et al. [40] also observed dense masses of yeast cells in root canals, while dentinal tubules were filled with hyphae. Siqueira et al. [41] investigated the patterns of microbial colonization in primary root canal infections through scanning electron microscopy, and observed single or budding yeast cells in root canal systems.
Nevertheless, since then, there has been an explosive advancement in microbial identification techniques, and new data are available on the fungi in this oral eco-niche. Hence, the aim of this systematic review and meta-analysis was to review the recent literature on the prevalence of fungal species in primary and secondary endodontic infections.

Data sources
Two investigators (AA and SC) performed an electronic search of MEDLINE/PubMed, SCOPUS, ISI Web of Science, and Cochrane Central without any language restrictions published since 2000. Grey literature was searched for pertinent papers by using OpenGrey and Greylist. A manual search was performed of all the issue published since 2000 of Journal of Endodontics, International Endodontic Journal and Oral Surgery, Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology. Moreover, reference lists of all included papers were screened for articles that need to be evaluated for inclusion in the study. The final electronic search was performed on 10 th March 2021. Our investigation was focused on relatively recent literature, in view of the fact that older, conventional microbiological analytical techniques, mainly depending on phenotypic analyses, have undergone a radical revolution with the advent of molecular biological screening techniques, such as next generation sequencing (NGS) platforms.
The specific review question formulated was as follows: What is the prevalence of fungal species in primary and secondary endodontic infection in human permanent teeth?.

Search terms
A specific search string was formulated for each of the databases, which included each of the following group of search terms, combined by means of the Boolean operator AND: • (Dental pulp disease?) OR (apical periodont � ) OR (periapical periodont � ) OR (apical disease?) OR (periapical disease?) OR (periapical lesion?) OR (apical lesion?) OR pulpitis OR (dental pulp necrosis) OR (tooth necrosis) OR (endodontic lesion?) OR (endodontic pathosis).
• Fung � OR mycet � OR mycos � OR candid � OR yeast? OR microb � • Exp dental pulp diseases OR exp periapical diseases OR exp periapical periodontitis OR Exp fungi Study selection Inclusion criteria. Human endodontic studies; studies on the detection of fungal species within infected root canal of either carious teeth or after failed endodontic treatment; studies clearly relating method of preparation of the sampling site and sample collection; clear description of the methodology employed for fungal detection; use of standardized methods for sample collection, transportation, and analysis (e.g., isolation with a rubber dam, use of sterile paper point; studies on permanent teeth with complete root formation.
Exclusion criteria. Exclusion criteria were ex vivo and in vitro studies and in vivo studies in teeth with endo-periodontal lesions.

Electronic data-search and assessment
To answer the review question, a four-step approach of evidence-based analysis was employed. According to the pre-set inclusion and exclusion criteria, the title and abstract were independently screened by two authors (AA and SC) at stage-one; full texts of the included records were assessed for eligibility by the same authors (AA and SC). Disagreements in article selection processes were solved by consulting a third reviewer (LS). At stage-two of the data extraction, the author (AA) meticulously screened the included full texts and extract information on sample size, primary or secondary infection after endodontic treatment. Each study's characteristics were recorded employing the Cochrane pattern determining the study design, setting, methods sample collection, culture medium, yeast species recovered. Additionally, the microbiological evaluation and outcomes about the prevalence of fungal species were recorded in a spreadsheet. The authors were contacted via emails if insufficient information were available from the included studies.
Quality of the evaluated studies. At the third stage of the review, a critical appraisal of the included studies was performed adopting a modification of the checklist proposed by Hoy and colleagues [42] using the following criteria: i) was the target population close representation of the national population in relation to relevant variables? ii) was the sampling frame a true or close representation of the target population? iii) were data collected directly from the subjects? iv) was the study instrument that measured the parameter of interest shown to have reliability and validity? v) was the same mode of data collection used for all subjects? Based on the above criteria a score was assigned for each study. A score of 0 is low risk bias, while (1-2) is considered as moderate risk and a score between (3-5) reflects high risk of bias, respectively.

Summary measures, synthesis of results, and other analysis
The fourth and final stage involved the analysis of the results, performed by an author (SC) by using R (R Core Team (2013). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/).
Data on the prevalence of fungal species were extrapolated from the studies and pooled together by calculating the weighted mean prevalence (based on the number of samples in each study). Separate analyses were performed for i) primary and the secondary endodontic treatment; ii) methods used for evaluating the presence of fungal species (molecular, cultural, or both), iii) the collection method (e.g., paper points and endodontic hand files), and iv) studies with periapical lesions in systemically healthy subjects. The meta-analysis was performed using the DerSimonian-Laird random-effects method (assuming heterogeneity among studies), calculating pooled proportion.
Cochran's test served to measure the results' consistency, with a significant level set at P <0.1. The heterogeneity of the outcomes was also evaluated as follows: <40% the heterogeneity was considered negligible, 41% to 60% moderately heterogeneous, if 61% to 90% substantially heterogeneous [43].

Results
The initial search identified 5,255 articles. Among them, 55 were duplicates and hence excluded. Of the remaining 5,200 papers, 5,110 were excluded after perusal of the titles and abstracts as they did not fit the inclusion criteria. Further screening for full texts resulted in the exclusion of 52 articles, as the outcomes were incongruent with the study objectives. Finally, a total of 39 studies using culture and molecular fungal assessment methods were included in the current review . Twenty-three included studies used the culture method for fungal evaluation. Among them, eight studies examined samples from primary endodontic infection, while ten, assessed the presence of fungal species in teeth with secondary endodontic infection. Another five studies evaluated the fungal presence in either primary or secondary endodontic infections. Of nineteen molecular studies reviewed, 12 studies examined the fungal presence in primary endodontic infection, while six appertained to secondary infection, and a single study in either a primary or a secondary infection. Fig 1 represents the articles selection process. Tables 1-6 present the characteristics of the included studies which investigated the yeast colonization of root canals in primary and secondary pulp infections through cultural methods and molecular analytical techniques, respectively. The prevalence of fungi in primary endodontic infections was investigated in 25 studies, with a total of 1341 analyzed samples and a weighted mean prevalence (WMP) of 9.0% (CI 95%: 6.0-12.0) (Fig 2). We found a significant heterogeneity among papers with regard to the reported values, as the prevalence of fungal species varied between 0 and 46.8% [78].
The studies that reported higher yeast prevalence [78,79] were based only on a small sample size (36 and 32 samples, respectively), while three of the studies with the highest number of clinical samples were based on PCR techniques and reported a prevalence of 19.66% [71], 22.62% [64] and 28.33% [80].
Considering all endodontic infections evaluated, the mean prevalence of fungi was found to be 9.11% (CI 95%: 6.8-11.4) in a total of 2003 specimens. The foregoing clearly implies that fungi contribute to up to one in ten to one in eleven root canal infections and re-infections. Interestingly, no differences in WMP were found among the detecting methods for each lesion type (either primary or secondary).

Discussion
Our meta-analysis of the literature comparing the fungal biome of primary and secondary endodontic infections, comprised selectively curated 39 studies, performed over a period of approximately two decades. These indicate the presence of fungal infection in approximately one in ten endodontic infections, with similar prevalence rates in both the primary (9.0%) and secondary infections (9.3%). A great heterogeneity was observed in the analyzed data, which represents a limitation of the present study. There are a number of reasons for the heterogeneity of the data that were noted, and they are essentially dichotomous in nature and falls into two broad categories, the health and demographic characteristics of the studied subjects, and the microbiological sampling and evaluation technology.
The vast majority of the cases included in the review were from systemically healthy and immunocompetent adults, not undergoing antibiotic or antifungal therapy over 1-6 months prior to sampling. However, a minority of studies were in children [62], and some were in human immunodeficiency virus (HIV) infected [70], and diabetic subjects [80]. These accounted for a limited number of cases (146 samples, respectively from 66 children, 20 HIV positive and 60 diabetic patients). Interestingly, the diabetic patients presented with a high prevalence of C. albicans (38.3%) in the infected root canals [80].      The microbiological sampling and evaluation technology appear to have played a critical role in the disparate outcomes reported here irrespective of whether they were based on traditional microbial culture technology, or molecular biological studies. Culture or molecular technology outcomes are essentially a direct reflection of the sample quality [83], which can be significantly impacted by the poor isolation of the sampling site, inadequate asepsis and disinfection of the entry portal (of the sampling material), poor access cavity design, missed or obtuse canals, inadequate instrumentation and debridement to remove superficial debris and/ or saliva contamination, and prior contamination of the root canal system due to seepage of the contaminant oral flora through poorly sealed temporary or permanent restorations (in the case of secondary infections).

Fungi in endodontic infections
As regards the sample collection protocols, multiple differences were found among studies. Paper point absorption method was the most commonly used, and proved to be the more sensitive according to one group [45], although endodontic instruments such as K-type or Headstrom files were used for sample collection by a large number of workers [45,50,55,61,62,70,[80][81][82]. There were inherent deficiencies in some studies as chemicals, that may affect fungal viability, were used during the mechanical instrumentation by the researchers, prior to sample collection [80]. As for persistent chronic infections, no chemical solvents were used by any of the investigators, to remove the root filling except in one study by Egan et al. [49].
Varying protocols for field decontamination were used prior to the endodontic access by different workers, including the use of iodine, hydrogen peroxide, chlorhexidine gluconate, sodium hypochlorite, or isopropyl alcohol of varying dilutions [45,49,50,55,57,61,63,64,[67][68][69][71][72][73]75,78,81,82]. It must be noted that liquid disinfectants may flow into the pulp    chamber and interfere with the sampling if applied after endodontic access or when the endodontic space communicates with the oral cavity. Sodium thiosulfate was also used in some studies to inactivate iodine, so as not to impact the microbiological sampling quality. Some of the protocols also employed sterility checks using a control samples to ascertain the integrity of the sampling technique [45,46,48,[57][58][59]64,[67][68][69]71,72,75,76,81,82]. Another noteworthy point is the microbiological culture methods used by different investigators which may lead to discrepant results. Due to the quantitatively low numbers of fungi populating the infected root canal systems their growth may take upto 72 hr to be discernible as colony forming units (CFU) unless appropriate culture media are used and observed for at least three days. If the laboratory workers do not use fungal culture media (such as Sabourauds agar or chromogenic agar) and/or evaluates the culture results only upto 24 hr, then there is a high likelihood that the yeast prevalence will be under reported. However, this deficiency could be overcome through molecular mycological analyses using fungal specific universal primers, when the total culturable as well as unculturable fungal flora could be detected with confidence, irrespective of their population size within the endodontic ecosystem.

Fungi in endodontic infections
Another limitation of the study is that we were not able to perform a patient-level analysis of the prevalence of fungi in endodontic infections. Although the majority of the included studies collected and analyzed only one sample per patient, some of them included more samples from the same subjects [49,57,60,64,71,76,79], while in some other papers this was not specified [44,46,[50][51][52]56,59,61,63,65,73].
Microorganisms may still persist within the root canal system even after adequate debridement and disinfection and related clinical procedures. This is due to the complex anatomy and the architecture of the root canal system that may include the presence of accessory canals, anatomical lacunae, and delta spaces. A number of workers have compared the prevalence of fungi in primary endodontic infections before and after endodontic therapy, and some have noted the increased prevalence of yeasts and other fungal species, post treatment, even though, in a few cases, they were undetectable prior to treatment [46,51,58]. The same findings had been already reported in previous studies [84,85]. This implies that fungi have the potential to often survive a harsh, nutritionally depleted and chemically infused ecosystem consequential to endodontic treatment procedures [86] and may gain access into the root canal system during or after the endodontic treatment.
Various studies have analyzed the inter kingdom co-habitation of yeasts and bacteria in the infected endodontic ecosystems. Although C. albicans is typically isolated from persistent endodontic infections in mixed cultures of cohabitant oral bacteria, at least four groups have reported identification of pure cultures of Candida from root canal infections [36,46,87,88]. Thus, Hancock et al. in a very early report, over two decades ago, described the mono-infection of root canals with C. albicans [45]. Also, in one of the largest studies to date, Waltimo and coworkers reported isolation of fungi in pure cultures from six of 967 endodontic samples, and yeast/bacterial mixed cultures in another 41 cases [86].
Another major reason for the heterogeneity of microbiological results we noted, could be the shift from cultural studies to molecular biological studies for microbiome evaluation, over the last few decades. Together with the increased sophistication of the molecular technologies there has been a parallel increase in the knowledge of the oral microbiome, leading to the elucidation of a hitherto unknown universe of unculturable flora. Accordingly, the cumulative data

PLOS ONE
from our review clearly indicate that PCR analytical techniques yield higher fungal prevalence values as compared with the culture methods. This is clearly seen in WMP for fungi in primary and secondary infections which were 6.3% and 7.5% for pure culture-based studies, that increased to 12.5% and 16.0% in PCR studies, respectively. Adding to this confusion, some workers used both cultural and molecular techniques for mycological analyses, and additional positive cases were detected through PCR analyses [69,72,75]. This was also the case in a recent study by Al-Sakati et al. [88], who compared culture-dependent and culture-independent methods to detect bacteria and fungi in reinfected root-filled teeth, and found a higher detection rate of fungi using the latter techniques.
Apart from the low population density of fungi in infected root canal systems, there may be other reasons for heterogeneous results even within molecular mycological analyses. These include the use of dissimilar primers by different investigators, with varying degrees of specificity. Al-Sakati et al. [88] noted, for instance, an increased rate of fungal detection by improving the specificity of the primers used. One disadvantage of the use of multiple specific primers instead of a single universal primer is the increased costs and time associated with the former approach. Recently, matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-ToF MS) has been proposed for fungal identification [89]. The absence of a comprehensive and well-curated spectral database of fungi, is the major limitation of this technique, although they are becoming increasingly robust and complete as time progresses with input from ongoing studies. MALDI-ToF MS analyses being inexpensive, easy, fast and accurate, and the fact that it can be applied to detect a comprehensive array of microorganism implies that the method would be very useful for future workers investigating the endodontic mycobiome. We reviewed, two studies which compared MALDI-ToF MS analysis and conventional cultural tests, but these workers failed to detect any fungal isolates from the endodontium [69,75].
Another exciting development in the quantitative and qualitative microbiological analyses is the advent of the Next Generation Sequencing (NGS) platforms. These, based on various platforms such as Illumina and Ion-torrent, as well as novel inexpensive, miniaturized platforms such as Minion technology, hold much promise in the demystification of the fungal flora of infected root canal systems.

Conclusions
The role of fungi in the pathogenesis of endodontic infections is unclear, as yet. The significant heterogeneity of results in the reviewed studies reflect the variety of the techniques used for sample collection, culture and identification. On the other hand, novel molecular techniques such as MALDI-TOF and NGS analyses should elucidate the true mycobiome as the new technology could redefine both the culturable and unculturable mycobiota in endodontic infections. Finally, as almost one in ten endodontic infections are associated with fungal infestation, clinicians must pay heed the need to eradicate both bacteria and fungi from infected root canal systems during treatment procedures.