Impact of non-invasive ventilation on microbial colonisation patterns in preterm infants: a single-centre study

Study design, setting and participant selection

This prospective analysis was conducted at Zhejiang University Children’s Hospital’s NICU, a leading tertiary perinatal care centre. Over a 12-month period (March 2022–March 2023), data were collected to account for seasonal variations and potential environmental influences on microbial colonisation. Eligible participants were preterm infants (gestational age: 28–35 weeks) recruited within 24 hours of birth. Exclusion criteria comprised infants with congenital anomalies, intrauterine infections, early-onset sepsis, born to mothers with Group B Streptococcus colonisation or requiring immediate invasive mechanical ventilation to control for confounding factors.

Exposures and outcomes

NIV methods, including NCPAP, nasal intermittent positive pressure ventilation, or high-flow nasal cannula (HFNC), were initiated within 24 hours of birth. For continuous positive airway pressure (CPAP), long nasal prongs were used, while the Infant Optiflow HFNC was employed for HFNC. The application was standardised across the NIV group to ensure consistency in treatment. NIV was administered via non-invasive ventilators, and humidified oxygen was provided to ensure optimal delivery of respiratory support. The fraction of inspired oxygen (FiO₂) to which the infants were exposed varied, typically ranging from 21% to 60%, depending on the clinical needs of each infant.

Infants not requiring respiratory support were categorised as either on room air or receiving low-flow nasal cannula with a flow rate ≤2 L/min. The inclusion of infants on low-flow nasal cannula in the ‘no respiratory support’ category is based on clinical guidelines indicating that this flow rate is typically considered supportive only for maintaining adequate oxygenation in stable infants, rather than reflecting significant respiratory intervention. This classification aims to clearly delineate infants who are stable and not in need of more intensive respiratory therapies from those requiring active respiratory support.

Regarding nutritional support, all infants received feeding according to the institution’s protocols to ensure appropriate nutritional intake. Specifically, infants above 28 weeks gestational age receiving NIV or with no respiratory support were fed in accordance with established guidelines.11

Both groups of infants received standardised oral care protocols, which included routine suctioning and the application of antiseptic solutions. Nursing staff administered scheduled oral care, using rinsing with hydrogen peroxide and sodium bicarbonate solutions to maintain oral hygiene and reduce bacterial colonisation.

All infants were permitted to have variations in respiratory support within the NIV group and changes in the no respiratory support group. However, any requirement for escalation or de-escalation of respiratory support between groups, or the need for intubation, would result in the exclusion of the infant from the study. The primary outcome measured the colonisation of oral and nasal microbiota at 5 days post birth, quantified as colony-forming units per millilitre (CFU/mL).

Sample collection

Samples were collected by research nurses at admission and on the fifth day. Prior to collection, oral care was conducted using saline solution to remove residual food particles and ensure uncontaminated samples. Swabs were used to collect samples separately from the oral cavity and nasal passages. After collection, swabs were placed in sterile containers and transported to the microbiology lab within 30 min. Each sample was diluted 1:100, then 100 μl diluted samples were spread on nutrient agar, and incubated aerobically at 37°C for 24–48 hours to facilitate bacterial growth. Negative plates were discarded after 3 days. Positive cultures were reported as CFU/mL. Based on this dilution, a threshold of <10³ CFU/mL was established to define significant microbial colonisation, representing the minimum detectable level of microbes to accurately classify colonisation status.

Microbial analysis

Colony morphology, including size, shape, colour and texture, was examined after incubation. Initial identification of bacterial isolates was based on these characteristics. Matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry was used for discrimination.

Data collection

In order to establish a comprehensive understanding of the clinical context for each participant, standardised data collection was performed. The following information was meticulously recorded: general information encompassing gender, gestational age, birth weight, type of delivery (vaginal vs caesarean section) and postnatal age. Prenatal information consists of premature rupture of membranes, maternal hypertension, maternal diabetes and prenatal steroid therapy. Delivery room management involved evaluating Apgar scores at 5 min after birth, with a specific focus on scores below 7 indicating the potential necessity for resuscitation, along with gathering data on endotracheal intubation and the usage of vasoactive drugs. Additionally, antibiotic use details included information about the administration of antibiotics during the sampling period, treatment duration and whether broad-spectrum or narrow-spectrum antibiotics were employed. Particular emphasis was placed on early-onset sepsis diagnosis within 72 hours of birth. Lastly, respiratory support records incorporated comprehensive descriptions of the mode of respiratory support provided, as well as specific parameters such as oxygen concentration, flow rate and pressure, in addition to the type of ventilator and tubing used.

Statistical analysis

Descriptive statistics provided an overview of the demographic, clinical features and patterns of oral and nasal microbiota colonisation in preterm infants. Categorical variables were analysed using the Pearson χ² test, while the Mann-Whitney U test was used for continuous variables. Binary logistic regression was used to evaluate the relationship between NIV and microbial colonisation in the nasal and oral cavities. The model adjusted for potential confounders such as gestational age at birth and the initial colonisation status of the nasal and oral microbiota. Adjusted ORs with 95% CIs were calculated to determine the strength and direction of the association between NIV and colonisation outcomes. Statistical analyses were performed using SPSS Statistics V.27 (IBM), with a significance level of 0.05.

Patient and public involvement

Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Enrolment and dropout

A total of 206 preterm infants (<24 hours old, gestational ages: 28–35 weeks) initially enrolled in the NICU study. After applying inclusion criteria, 160 infants remained eligible, while 31 were later excluded due to conditions (congenital anomalies and maternal infections). The final cohort consisted of 129 infants, divided into a control group (n=60, receiving room air or low-flow nasal cannula as per respiratory support needs) and an intervention group (n=69, undergoing NIV based on respiratory support needs). However, 19 infants from the intervention group and 10 infants from the control group dropped out due to non-compliance with treatment protocols. Nonetheless, each group completed the study with 50 infants, ensuring robust analysis. The final analysis included data from 100 infants, providing a thorough examination of the effect of NIV on microbial colonisation in preterm infants (figure 1).

Request permission

Demographic and clinical characteristics

The demographic and clinical characteristics of the enrolled infants are provided in table 1. The study population was evenly divided between the two groups, each consisting of 50 infants. There were no significant differences in baseline characteristics such as gender, birth weight, prenatal conditions and initial colonisation status of oral and nasal microbiota between the groups. However, a difference in gestational age at birth was observed, with the NIV group having a lower median gestational age (30.5 weeks) compared with the no respiratory support group (32.2 weeks), with a p value <0.05.

Analysis of microbial growth

The assessment of microbial colonisation, measured by colony-forming units per millilitre (CFU/mL), revealed distinct differences in bacterial growth concentrations between the two groups, as outlined in table 2. Notably, infants in the NIV group exhibited a higher median CFU/mL in nasal samples, indicating a potentially greater microbial load in this group compared with the group without respiratory support. Further analysis, presented in table 3, used a binary logistic regression model to control for confounding factors such as gestational age at birth and the initial colonisation status of both nasal and oral microbiota. This adjusted analysis confirmed a significant association between NIV and an increased likelihood of nasal bacterial colonisation in preterm infants. The adjusted OR of 2.91, with a 95% CI of 1.12 to 7.58, and a p value of 0.028, suggests that NIV is associated with approximately 2.91 times higher odds of nasal bacterial colonisation in preterm infants after 5 days of hospital stay, when accounting for other variables.

Microbial colonisation patterns

The patterns of microbial colonisation in the nasal and oral cavities for each study group are shown in online supplemental table 1 (day 1) and table 4 (day 5). Notably, the NIV group exhibited a significantly higher rate of nasal bacterial colonisation after 5 days, with Staphylococcus species being the most prevalent, followed by Klebsiella pneumoniae and Acinetobacter baumannii. In contrast, the no respiratory support group showed a lower rate of nasal bacterial colonisation, with Staphylococcus species also being the most common isolate. The oral cavity colonisation was comparable between the two groups, with no significant differences in the types or quantities of microbial growth observed.