Aim
The aim of this study will be to investigate whether an HMO-based DHM “matching” strategy can improve the gut microbiota composition of preterm infants and provide pilot data for a full-scale trial to establish clinical benefits. The primary objective of this study is to evaluate the effect of secretor matched DHM on gut microbiota composition in premature infants < 34 weeks gestational age (GA). We hypothesize that preterm infants receiving matched DHM (compared to standard unmatched DHM) will have gut microbiota profiles more similar to that of preterm infants exclusively receiving their MOM.
The secondary objectives are a) to measure infant growth and observe clinical outcomes such as necrotizing enterocolitis, infection and feeding intolerance, and b) to evaluate the implementation of this trial in the clinical setting using the Consolidated Framework for Implementation Research (CFIR)[29].
Ethics and registry
This research involves human participants and is performed in accordance with the Declaration of Helsinki. This trial was approved by the University of Calgary, Conjoint Health Research Ethics Board approval number REB19-0542 (Protocol Version 1.5, July 25, 2022). Registration on ClinicalTrials.gov was completed on October 17, 2019, and updated on February 11, 2022, with the Identifier: NCT04130165. Written informed consent is required to participate in the trial. Study staff will maintain appropriate medical and research records for this study, in compliance with International Conference on Harmonization E6, Section 4.9 and regulatory and institutional requirements for the protection of confidentiality of subjects. The study coordinator will maintain an up-to-date list of study staff who have access to research records and data for this study. This manuscript was prepared in accordance with The SPIRIT (Standard Protocol Items: Recommendations for Interventional Trials) guidelines [Figure 2; Supplementary File 30].
Design
We will conduct a pilot randomized controlled, open label design to compare two groups of preterm infants (< 34 weeks gestation): 1) infants receiving DHM matched to their mother’s secretor status, 2) infants receiving standard issue (i.e. unmatched) DHM (Fig. 2). In Alberta, approximately 30% of infants admitted to the NICU exclusively receive their mothers’ own milk (MOM) throughout their hospitalization and do not require DHM; these infants will be retained as a reference group. Thus, we anticipate that approximately 21 infants from each randomized group will receive DHM per their group allocation (control and intervention DHM groups), and 18 infants will not require DHM supplementation (MOM reference group). Full blinding is not feasible because the nurses providing DHM need to be aware of group allocation. However, investigators will be blinded during outcome assessments (i.e. microbiome profiling).
Setting
Patients will be enrolled at four neonatal intensive care units (NICUs) in the Alberta Health Services (AHS) Calgary Zone. In Alberta, Level II NICUs mainly admit infants born after 32 weeks gestation. DHM is routinely used at these centres and is currently provided to the majority of infants within the first 2 weeks of life. In Calgary, between 2015–2018, 700–720 infants born < 34 weeks gestation were admitted to level II NICUs annually. Thus, it is feasible to enrol 60 mothers in the proposed 12-month recruitment period of our study.
Participant Mothers and Infants
We will recruit mothers who present at the antenatal unit or labour and delivery unit who are less than 34 weeks gestation, planning on breastfeeding and who are in imminent preterm labour. Participant mothers must consent to donor human milk supplementation, if indicated, rather than preterm infant formula. Infants of enrolled mothers, who are born < 34 weeks gestation (up to 33 6/7 days) will be included in the study. Infants who are diagnosed with a clinically significant major congenital malformation, intestinal perforation, stage-2 NEC, receiving an extended course of antibiotics (> 48 hours) and those who are unlikely to survive the study period, will be excluded from the study.
Randomization Procedures
Allocation will occur once a mother has provided consent to participate in the study and their secretor status is determined. Enrolled subjects will be stratified based on their secretor status and block randomized to either the intervention or control group according to a computer-generated randomization sequence developed by an independent staff member of the clinical research unit who is unaffiliated with the clinical investigators and the NICU staff. Allocations will be sealed in opaque envelopes labelled with the participant recruitment order (secretor 1, non-secretor 1, etc.). Once maternal secretor status is determined, the research assistant will select the envelope for the corresponding participant recruitment order to determine their allocation group. There will be no blinding of assignment for clinicians.
Intervention - Donor Human Milk
Infants who are randomized to the intervention group will receive DHM that is matched to their mother’s secretor status. Infants who are randomized to the control group will receive DHM that is pooled as per standard practice (described below) and has no matching criteria.
The Northern Star Mother’s Milk Bank (NMMB) is the only human milk bank in Alberta and is the sole provider of DHM to all NICU sites in Alberta. All milk donors who have met Human Milk Banking Association of North America (HMBANA) screening criteria and who present at the NMMB to donate milk, will be invited to participate in the study by milk bank staff. Consenting donors will provide 2–3 mL of saliva using a passive drool technique and donor secretor status will be determined using the Weiner Agglutination technique [31].
Upon receipt of donor milk, DHM will be labeled according to the donor mother’s secretor status and frozen at -25⁰C. NMMB standard practice is to pool milk from 3–4 donors and batch pasteurize using the Holder pasteurization technique [14]. Holder pasteurization eliminates bacterial, fungal, and viral pathogens from DHM, yet predominantly maintains the integrity of HMOs [14]. Each bottle of pooled DHM is bar coded and linked to infant charts within NICU sites and the milk bank. As such, secretor and non-secretor batches of DHM will be easily trackable through the healthcare system. Ten millilitres (10 mL) of each batch will be stored for future HMO analysis (as described below). Matched DHM will be transported to the NICU as per standard practice. DHM arrives at the site frozen in individual bottles and is thawed each day based on the needs of the unit. Matched DHM will be packaged in colour coded bottles. While DHM is not labelled based on caloric or nutritional content, we will assess the nutritional content of each batch of matched DHM to ensure it has similar nutritional content to the standard DHM.
At the hospital, milk allocations are requested the day prior and the corresponding amount of DHM is thawed each day. One nurse (usually the charge nurse or lactation consultant) oversees thawing and aliquoting daily DHM feeds for each infant admitted to the NICU. Infants who are enrolled in the intervention arm of the study will be flagged and the milk distribution nurse will ensure that the appropriate matched DHM is prepared. Infants who are assigned to the control arm of the study will receive standard DHM as per the usual unit protocol. Administration of the study and control product will be done by staff nurses who are experienced in providing DHM for preterm infants.
Outcomes
The primary outcome of this study is to compare microbiome signatures (richness, diversity, and functional taxonomic classification) between control and intervention groups. Although this proof-of-concept study will not be powered to assess clinical outcomes, we will monitor infant growth, length of stay, and adverse health events as secondary outcomes. Finally, a third outcome will be to assess barriers and facilitators to implementation and scale-up of the new DHM strategy using the Consolidated Framework for Implementation Research [CFIR; 29].
Process and timeline
Potential mothers (< 34 weeks GA) admitted to the antenatal unit or to labour and delivery units at each of the sites will be screened by clinical staff for eligibility upon admission to the unit and approached by the clinical staff for permission to be contacted by a research assistant (RA). Study staff will discuss the study, the consent procedure and forms and obtain consent for the study and authorization to use protected health information. The informed consent form will be completed prior to any study-specific procedures. After consenting, maternal secretor status will be determined using the Weiner Agglutination technique [31] from 2–3 mL of non-stimulated saliva (passive drool) collected at recruitment (Fig. 2). It is vital the secretor status be determined prior to delivery of the infant as it is standard practice for preterm infants to receive their first dose of DHM within hours following birth. Once secretor status is determined, mothers will be block randomized into the control or intervention group based on their secretor status as per the aforementioned allocation strategy. Baseline demographic and clinical data will be collected from the mother’s chart and all concomitant and preceding medications the mother and the infants receive will be recorded on the study data capture sheet.
Upon delivery and transfer to the NICU, infants will be assessed by a neonatologist and an oral feeding plan with be established, as per standard practice. Once established on oral feeds, and if supplementation with DHM is indicated, infants who are allocated to the intervention group will receive DHM matched to their mother’s secretor status. Infants who are allocated to the control group will receive standard DHM. DHM supplementation will continue until the infant is entirely transitioned to MOM or the infant is discharged or transferred from the NICU.
Data Collection
Three infant fecal samples will be collected from soiled diapers, at birth, day seven and day 14 of life (Fig. 1). Fecal samples will be collected and stored at -80°C until analysis. Infant feeding records will be monitored daily to ensure adherence to study protocol, and clinical data will be collected daily for safety monitoring. Additionally, one urine sample, using absorbent cotton stored in a diaper, will be collected from each infant at week one to determine infant secretor status using the Weiner agglutination technique [32].
To assess HMO profiles of mother’s own milk, NICU staff nurses will collect breastmilk (10 mL) on Day 7–14 post-partum from unused/ leftover milk. To assess HMO profiles of DHM, 10 mL from each batch will be captured prior to shipment to the NICU from the NMMB.
Infant demographic data, mode of birth, growth, feeding records (including proportions of DHM and MOM), length of stay and adverse health events will be collected from electronic chart data.
Sample Analysis
Fecal microbiome analysis
Microbial DNA will be extracted from stool using QIAamp PowerFecal DNA kit (Qiagen). Shotgun metagenomics will be used to identify microbial population structures and functional capacity. Short reads will be functionally annotated by mBLASTx alignment to a peptide database containing peptide sequences from all KEGG-annotated species [33 34]. Hits with an E-value < 1 will be counted and assigned a KEGG Orthology group (KO) using the ‘top gene’ approach [35]. Gene abundances will be determined using MUSiCC, a marker-genes based normalization scheme for microbiome studies [36].
Milk analysis
To confirm and quantify the impact of pooling DHM based on donor secretor status, we will compare the HMO profile of different DHM batches. These will further be compared to the HMO profile of MOM from different mothers. Samples will be analyzed using high-throughput solid-phase extraction and high performance liquid Chromatography (HPLC) with fluorescence labeling and detection as described in Azad, et al. [18]. Separation of labeled HMOs on an amide column followed by mass spectrometric analysis distinguishes the 20 most abundant HMOs, which account for more than 95% of all HMOs. Adding the non-HMO raffinose as an internal standard allows for absolute quantitation.
Additionally, we will analyze the energy (calories) and macro-nutrient content (crude protein, total fat, and carbohydrates) of the matched and unmatched DHM pools. DHM samples will be analyzed using the Lactoscope, a Fourier-transformed full spectra mid-IR instrument (FTMid2; LactoScope FTA; Perten Instruments). The Lactoscope provides the most accurate measure of macronutrient content currently available in human milk analysis. Samples will be prepared and analyzed as per Perrin et al. (2019) [37].
Statistical Analysis
Sample Size Considerations
This proof-of-principal study will provide pilot data and establish feasibility for a larger clinical trial. We will recruit 60 infants, providing adequate power to detect overall microbiome differences comparable to our observations in the CHILD cohort (e.g. 0.42 lower Chao1 Index among exclusively breastfed infants of secretor vs. non-secretor mothers; p = 0.05, n = 54 infants; unpublished data), and other preliminary studies (e.g. 16.7% variance in microbiome composition explained (p = 0.028) by maternal secretor status; n = 11 exclusively breastfed infants [25]). Results of our study will provide necessary data to inform and power a full-scale trial to assess more subtle microbial differences as well as clinical outcomes.
Analysis Plan
The primary aim of this study is to compare standard DHM vs. matched DHM groups for differences in known ‘healthy microbiome signatures’ [38], such as lower diversity, higher proportions of Bifidobacteria, and lower proportions of pathogenic organisms. To address this question, we will use an intention to treat analysis and include all infants based on group allocation. Additionally, to determine if matched DHM drives the preterm gut microbiome towards a normal healthy state, we will conduct a 3-way comparison of microbiota community structures using MOM as a reference group: standard DHM (control) vs. matched DHM (intervention) vs. MOM (reference).
Microbial composition will be compared between groups to determine differences in alpha diversity (by ANOVA), beta diversity (by principal coordinates analysis and PERMANOVA), and taxonomy. Differentially abundant taxa and pathways will be identified using Wilcoxon rank-sum testing with multiple comparisons correction using a 5% false discovery rate threshold. The relative abundance of KOs associated with KEGG bacterial pathways will be summed to facilitate comparisons of community-level metabolic capacity. In addition, quality non-human reads will be assembled using the SPAdes Genome Assembler [39]. Contigs will be annotated using the IMG pipeline [40], which utilizes multiple functional database to yield comprehensive and accurate annotation of the abundant genes in the community. Exploration of functional categories will be accomplished with PCA and statistical comparisons of KEGG and MetaCyc pathways will be compared across groups as described above. We also aim to identify overall microbiota patterns that distinguish standard DHM vs. matched DHM vs. MOM study groups. Baseline clinical and demographic characteristics (eg. sex, gestational age, birth weight) will be compared between groups using appropriate parametric or nonparametric tests; if any significant differences are identified, these parameters will be considered as covariates in analysis of outcome measures. A per-protocol analysis will also be conducted to examine the influence of potential deviations from the treatment protocol.
Feasibility and Implementation Assessment
The Consolidated Framework for Implementation Research (CFIR) is a well-researched and supported implementation framework that has been demonstrated to improve success of intervention research and intervention roll-out [38]. As per the CFIR protocol, we will conduct pre- and post-intervention interviews with seven key stakeholders involved with this pilot trial: a) one staff member from the milk bank, b) two NICU nurses who provide both bedside and charge nurse care, c) the FMC NICU lactation consultant, d) one neonatologist, and e) two mothers participating in the study. These interviews will provide insight on five domains (the intervention, inner and outer settings, individuals involved, and implementation process) that influence implementation effectiveness [38] and will allow researchers to tailor the intervention protocol to better suit the culture of the NICU environment. Follow-up interviews will be conducted with each individual to assess how they perceived the intervention and will include suggestions for improvement. We will incorporate these findings into our research protocol for scale up to a larger RCT.