EpiCandIn: An open online resource for epidemiology of Candida infections in India

Screening publications for data extraction

PubMed and ScienceDirect databases were searched to obtain articles pertaining to Candida research in India published till December 31, 2022. The search query used for PubMed was ‘((Candida [Title]) OR (Candidemia [Title]) OR (Candidiasis [Title]) OR (Candiduria [Title])) AND ((India [Title/Abstract]) OR (Indian [Title/Abstract]) OR (India [Affiliation]))’. The search query used for ScienceDirect was ‘Title: Candida OR Candidemia OR Candidiasis OR Candiduria, Author Affiliation: India’.

The abstracts of all articles were manually curated to include original scientific reports of Candida infections in the Indian population. Publications on Candida infections for which full text was not accessible were rejected. Publications with inconsistencies in the data reported in different article sections were rejected. Publications reporting non-epidemiology data, epidemiology data from non-Indian patients, data obtained from in silico, in vitro or in vivo models were excluded. The workflow adopted for this study is shown in Figure.

Close

Figure.

Flowchart of the literature screening process. Manual review of 2103 records resulted in the selection of 210 eligible publications having clinical data on candidiasis in the Indian population.

Export to PPT

Manual curation and data extraction

All the eligible publications were manually curated to extract the following information: author names, publication year, study duration, study design, clinical setup, geographical location, comorbidities, risk factors, species, type of candidiasis, niche infected, symptoms, antifungal drugs prescribed, age and sex of participants, number of patients, number of Candida isolates, identification methods, antifungal susceptibility profiles, methods of antifungal susceptibility testing, prevalence, and mortality.

Rules adopted for redundancy elimination in the database are listed in Supplementary Table I. The organs infected with Candida spp. and the type of candidiasis was determined as per the collected culture-positive biological samples (Supplementary Table II). The spellings of species names have been corrected as per NCBI Taxonomy database¹⁶. If a species name was not available in NCBI Taxonomy database, the most frequently used spelling was selected (Supplementary Table III).

Data analysis

Due to a lack of uniformity in the sample collection and reporting methodology, statistical analysis could not be performed on the collated data. Data visualization packages such as ggplot2 library in R software (cran.r-project.org/web/packages/ggplot2/index.html) were employed to gain insights from the database.

Database architecture

EpiCandIn is built on Apache HTTP Server 2.4.51 with MySQL Server 5.7.36 as the back-end and PHP 7.4.26, HTML and JavaScript as the front-end. D3.js JavaScript library was used for data visualization.

Homepage

The Homepage of EpiCandIn contains an interactive map of India. The geographical sites contributing the data in EpiCandIn are highlighted on the map. Clicking on these sites will provide a tabular view of epidemiology studies and associated details from the selected locations. Quick links to access studies related to individual species, niches and antifungal drugs are also available here.

Search & download

Users can avail the simple or advanced search options to query the database. Simple search is through keyword check on all fields and advanced search functions through query builder wherein keywords are searched based on the specific fields defined by the user. Users can also download the data in a tabular format.

Submit

Through this portal, clinicians and researchers can submit epidemiological data on Candida infections from India along with the associated publications. This data will be manually curated and verified for inclusion and subsequent data extraction for EpiCandIn updates.

The literature search yielded 2103 results, comprising 1541 articles from PubMed and 562 articles from ScienceDirect. After removing 301 duplicate articles, the remaining 1802 publications were manually reviewed, and 210 eligible papers were included in the EpiCandIn database. It contains data from 51 sites in 16 States and four Union Territories of India. These studies provide information on 38 unique Candida species isolated from 25 different host niches (Supplementary Table II) reported over a span of five decades.

The major challenge during data curation and extraction was the incomplete or missing information in the retrieved publications. The data on geographical location, gender, treatment modalities and outcome were missing in several papers. In the case of studies spanning multiple years, the observations were not segregated year-wise. With respect to age as a parameter, several articles used terms such as ‘neonates’, and ‘reproductive-aged women’ without indicating the exact age. A few reports mentioned a broad range of 1–88 yr¹⁷, 0–80 yr¹⁸ and 0–70 yr¹⁹ for the study participants. Notably, we did not find any published clinical reports on candidiasis from Himachal Pradesh, Bihar, Jharkhand and the northeastern States (except Assam).

A limitation of this study is the absence of statistical analysis because of the lack of uniformity in sample collection and data reporting across the publications. These issues have restricted the analytical scope of the data present in EpiCandIn. Therefore, a simple analysis of the data was performed, which revealed the following: (a) C. albicans is the most reported species for candidiasis followed by C. tropicalis, C. glabrata and C. parapsilosis (Supplementary Fig. 1 and 2); (b) clinical reports on C. auris and other NACs increased since 2011 (Supplementary Fig. 1); (c) blood was the most well-studied niche followed by the oropharynx and genitourinary tract (Supplementary Fig. 3 and 4); and (d) fluconazole was the most widely prescribed antifungal drug, followed by amphotericin B (Supplementary Fig. 5).

We observed a shift from simple microscopic and culture-based methods to more sophisticated molecular techniques such as polymerase chain reaction (PCR), DNA sequencing and matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry for the detection of Candida species over the last two decades. This may have contributed to improved species-level identification, leading to a rise in reports on NACs such as C. auris since 2011²⁰. These species were identified mostly from the blood, oral cavity and genitourinary tracts (Supplementary Fig. 4), probably due to the ease of sample collection. Candidiasis management heavily depended on fluconazole and amphotericin B (Supplementary Fig. 5). Despite the rise in azole-resistant Candida isolates, fluconazole has remained the most widely prescribed antifungal drug for candidiasis in India over the years (Supplementary Fig. 5). This indicates the urgent need for expanding the antifungal therapeutic options through new drugs and targets. The World Health Organization has stressed upon the need for reports and evidence to understand the burden of infections with regard to the prevalence and antifungal resistance to strengthen global health policies²¹.

This study is the first step towards developing a unified platform for studying the epidemiological landscape of candidiasis in India. EpiCandIn will enable researchers, clinicians and policymakers to browse the data on Indian clinical reports and draw insights on the increasing burden of Candida infections. Periodic updates of this database with clinical data will enable us to strengthen the candidiasis management strategies in India.