Abstract
The neonatal skin microbiome consists of all the genomes and genetic products of micro-organisms harbouring the skin of babies. Host and the microbiota develop a harmonious environment resulting in symbiosis. Any disruption of this environment could lead to pathological disease. Our study was conducted to explore the neonatal skin fungal microbiome of very preterm neonates admitted to Neonatal Intensive Care Unit at a tertiary health care setting using Next Generation Sequencing of the18S rRNA gene. The most abundant genera found in 22/30 samples were Candida followed by Bipolaris & Cladosporium on the skin microbiome of these neonates. The presence of these fungi, whether just as commensals or as potential pathogens, is currently under research, owing to the risk of early exposure and incidence of infection right from birth.
Introduction
Skin colonization in preterm babies commences in utero and continues after delivery. It is predominantly influenced by maternal genital flora, mode of delivery, type of feeding, level of sickness, and antibiotic exposure. Symbiosis between host and microbiome is critical for a stable environment in vivo. The role of bacterial colonization in neonates and the diversity of flora has being greatly studied over the past decade. 1 Studies on cutaneous mycobiome in preterm babies are, however, sparse. Our prospective study was conducted to understand and bridge this gap regarding the mycobiome of preterm babies <32 weeks old in the early neonatal period.
Materials and methods
Skin swabs of 30 preterm neonates < 32 weeks of gestation admitted to a Level III NICU were collected in the first five days of life after admission following obtaining of consent over a six month period. The swab was taken from the skin of the neonate in a single sweep pattern starting from right axilla > right abdomen > groin > left abdomen > left axilla in order (Figure 1). Samples were sent to our genomics laboratory in sterile water for Next-generation sequencing (NGS). Detection was done using Credence Rapid Infection Detection (credence RIDTM) for Next Generation Sequencing after Polymerase Chain Reaction (PCR). ITS1 (internal transcribed spacer 1) of the 18S rRNA gene was used for the identification of fungal species. The reads obtained were then compared to standard library matching barcoded bacterial and fungal libraries that were multiplexed on a single chip on a 400 bp run to obtain sequencing data. Specimens were run in batches of 20 on an Ion 318TMv2 chip (Thermo Fisher Scientific). Data were analyzed using Credence Genomics proprietary bioinformatics pipeline for analysis of clinical isolates. Statistical analysis was done using IBM-SPSS v.23, New York, USA. As these were preliminary findings, a narrative description of the fungal mycobiome has been expressed as a graphical format and demographic description has been tabulated. Heat Map for demonstrating relative abundance at the species level has been carried out for the same. Ethical approval was obtained for the study from Institutional Ethics Committee (IEC- BMCRI/PS/254/2020-21)

Heat Map for demonstrating relative abundance at the species level. (X-axis : Sample Number, Y-axis: Right : Species type, Left : Relative abundance in %, Bottom right image inset: The swab was taken from the skin of the neonate in a single sweep pattern starting from right axilla > right abdomen > groin > left abdomen > left axilla in order).
Results
Out of 30 preterm babies, 13 (43.33%) were males with a median (IQR) gestation of 31(30–32) weeks and a mean (SD) birth weight of 1.34 (0.21) kg. Other neonatal and maternal baseline characteristics are mentioned in Table 1. Using the credence RIDTM, the NGS results were analyzed as percentages, i.e. relative abundances. The most abundant genera found in 22/30 samples were Candida followed by Bipolaris & Cladosporium. The hierarchical clustering based on species was carried out for the ten most common fungal abundances that were found. These include Candida albicans, Candida Africana, Candida tropicalis, Candida parapsilosis, Candida orthopsilosis, Cladosporium endophytica, Cladosporium halotolerans, Candida hyderabadensis, Bipolaris subramanianii, and Bipolaris variabilis. (Figure 1)
Baseline characteristics y(mother /infant).
Discussion
Premature babies are at risk of skin colonization by various bacteria or fungi during the perinatal or neonatal period from maternal vaginal flora, NICU environment, or health care staff. Bacteria or fungal colonization predisposes these babies to invasive sepsis. 2 In our study, we found colonization was mostly by commonly seen candida species and two uncommon species, Cladosporium and Bipolaris. Cladosporium, although ubiquitous, is rarely associated with invasive sepsis. 3 Bipolaris species are plant pathogens and detection of these species, although intriguing, could be due to environmental contamination. 4 A previous study 5 have found Malassezia as the most common cutaneous mycobiome in premature babies. This disparity could be explained by the geographical location of increased heat and humidity, population differences of low socio-economic status, high incidence of fungal sepsis in NICU, and a 56.6% antibiotic usage in the first week of life. No vaginal swab was done in any mother before delivery. Some 30% of mothers had a vaginal leak and 10% had a vaginal discharge before delivery. Colonization by Candida species in high abundance increases the risk of candidaemia according to previous studies, though none of our babies had any invasive fungal sepsis.5,6 One baby died due to multi-organ dysfunction secondary to bacterial sepsis. The mere presence of these species in premature babies is alarming for the treating physician concerned about the possible hazards of invasive fungal sepsis, cross-infection, and outbreaks.
The strengths of our study included proper technique in taking the sample including transport and processing methods; therefore making the possibility of cross-contamination minimal.
Footnotes
Acknowledgments
We thank all parents for their consent and dissemination of study results. The authors would like to thank DrSneha K Chunchanur and DrSathyanarayan Muthur Shankar for their guidance and expert advice on the entire analysis and proceedings of the study. Furthermore, the authors would like to thank Subhadeep Das for all statistical analysis and graphical representation of the study's findings.
Ethics
Ethical approval was obtained for the study from Institutional Ethics Committee, Bangalore Medical College, and Research Institute. (IEC - BMCRI/PS/254/2020-21: 04.01.2021).
Funding
This study has received a grant from Rajiv Gandhi University of Health Sciences, Karnataka, India (Ref: RGU/ADV-Res/UG Research/16/2018-19) for the project Code. : UG20MED303
Author contributions
ShDi: Conceived the idea, designed the study, supervised the data collection, critically appraised the data analysis, drafted the first manuscript, and approved the final manuscript
