Probiotic Supplementation in Extremely Preterm Infants in Scandinavia

Overview

The primary aim of this research is to determine whether supplementation with probiotics during the first weeks of life reduces the risk of necrotizing enterocolitis (NEC) and neonatal mortality and is safe to use among extremely preterm (EPT) infants born before gestational week 28. P: The study population include EPT infants (n= 1620) born at six tertiary neonatal units in Sweden and four units in Denmark. I: This is a double-blinded multicenter randomized controlled trial where infants in the intervention group will as soon as they tolerate 3 mL breastmilk per feed receive a probiotic combination of Bifidobacterium infantis, Bifidobacterium lactis, and Streptococcus thermophilus diluted in 3 mL breastmilk and given once daily until gestational week 34. C: The control group will receive 3 mL breastmilk without probiotic supplementation (blinded) daily. O: Primary outcome variables is a composite endpoint of incidence of NEC and mortality. Secondary outcomes include incidence of sepsis, duration of hospital stay, use of antibiotics, feeding tolerance, growth, and body composition after hospital discharge. Patient benefit: To provide evidence on the usage of probiotics among EPT infants that are not currently covered by clinical recommendations. As EPT infants have the highest risk for NEC and mortality our results have the potential to change current recommendations and improve patient outcomes, decrease mortality, shorten hospitalization, and decrease overall health-care costs.

Study plan PURPOSE AND AIMS Studies have shown conclusive results on reduced incidence of necrotizing enterocolitis (NEC) in preterm infants born after 28 gestational weeks (1, 2). However, previous research has been underpowered for extremely preterm (EPT) infants and thus, the efficacy in this group of infants is not fully eluted. Therefore, the Swedish guidelines for probiotic supplementation in neonatal care currently excludes EPT infants from being eligible to receive probiotic supplementation (3). Notably, EPT infants is the patient group with highest mortality and NEC rate, and thus most at need for strategies that can reduce this incidence (4). We believe our study design will enable to answer if probiotic supplementation can be expanded to also include EPT infants. Overall aim The overall aim is to determine whether early probiotic supplementation can be used as a preventable nutritional intervention to reduce the risk of NEC and neonatal mortality among EPT infants (born less than 28 gestational weeks) using a double-blinded randomized control design. Probiotic supplement effect on other common neonatal morbidities will also be studied. The study will be carried out at the neonatal intensive care unit (NICU) at Karolinska University Hospital and collaborating tertiary neonatal centres in Sweden and in Denmark. Primary aim 1. To determine whether the usage of probiotic supplementation will reduce the incidence of a composite endpoint of NEC and neonatal mortality among EPT infants. Secondary aims 2. To determine whether probiotic supplementation will reduce the incidence of NEC. 3. To determine whether probiotic supplementation will reduce the incidence of neonatal mortality. 3\. To determine whether probiotic supplementation will reduce the incidence of neonatal sepsis 4. To determine whether probiotic supplementation will reduce the rate of prescription and duration of antibiotic use during hospital stay. 5\. To determine whether probiotic supplementation will reduce duration of hospital stay. 6\. To determine whether probiotic supplementation affect development of microbiome. Tertiary aims 1. To assess whether probiotic supplementation will reduce the incidence of enteral feeding intolerance. 2. To assess postnatal growth and body composition by following participants post-NICU hospitalization at Karolinska University Hospital. SURVEY OF THE FIELD Care of preterm infants has improved considerably during the last decade. Despite this, about 20% of EPT infants (born \<28 gestational weeks) in Sweden still dies (4). In EPT infants, NEC is a common cause of death (5). NEC is a gastrointestinal disease that causes an inflammatory response and is the most severe gastrointestinal-related morbidity in preterm infants (5-7). The pathogenesis of NEC is multifactorial and not fully understood, but the principle underlying cause believes to be immature gastrointestinal system. The current conservative treatment for NEC is exclusive parental nutrition and antibiotics. In up to 50% of cases, gastrointestinal surgery is necessary. In a systematic review and meta-analysis from 2021 summarizing 45 trials including 12 320 infants, it was observed that a combination of Bifidobacterium and Lactobacillus was associated with a 54% lower rate of NEC and a 44% lower rate of total mortality (1). The ProPrems trial, conducted in Australia and New Zealand randomized 1099 infants born \<32 gestational weeks. In the ProPrems trial, a probiotic combination B. infantis Bb-02, B. lactis Bb-12, and Str. thermophilus TH4 was tested (2). However, this trial was underpowered for EPT infants and thus, the efficacy and safety cannot be generalized to this vulnerably patient group. The recent position paper from year 2020 on "Probiotics and Preterm Infants" by the European Society for Paediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN) has investigated the results from existing randomized controlled trials, meta-analysis, and systematic reviews (8). They concluded that two probiotic supplements can be used on infants with birth weight \<1500 grams. The committee recommends providing either L. rhamnosus GG ATCC53103 or the combination of B. infantis Bb-02, B. lactis Bb-12, and Str. thermophilus TH4 which will be used in our study. Because of uncertainty regarding efficacy for EPT infants, the Swedish recommendations only include infants born above 28 gestational weeks (4). There is, to our knowledge, no ongoing randomized controlled trials investigating probiotic supplements to EPT infants. We believe that our research will provide high-quality evidence and that study results will contribute to the knowledge on whether current recommendations on probiotics supplementation can be expanded to also include EPT infants in Sweden, Denmark and that the results can be generalized globally. STUDY DESIGN We will conduct a double-blinded randomized trial testing the effect of probiotic supplementation on the incidence of NEC and neonatal mortality. The study will be performed at Karolinska University Hospital in Solna and Huddinge and 10 other tertiary neonatal units in Sweden and Denmark. The trial will be designed in accordance with recommendations for interventional trials by the SPIRIT Statement (9) and Consolidated Standards for Reporting of Trials CONSORT guidelines (10). POPULATION For all studies in the current proposal, we will adapt a double-blinded multicenter randomized controlled and registry-based trial examining the effect of probiotic supplementation on EPT infants. We will recruit EPT infants, born between 22 0/7 and 27 6/7 weeks of gestation. The recruitment period will be between August 2022 and June 2026 or until enough patients have been included. Infants will be recruited within 72 hours from birth, legal guardians will be informed about the study and asked to participate. Informed written consent will be retrieved prior to inclusion. Patients with severe complications and low chance of survival or major congenital abnormalities or patients participating in other trial with the same or overlapping outcome measures will be excluded. The inclusion of patients in the trial will be performed in Sweden and Denmark at ten different hospitals, the patient may however be relocated to other hospitals during the trial. In case of relocation prior to completion (34w GA) continuation of the intervention or control will be granted under the supervision of the referring hospital. Included hospitals in Sweden are Karolinska University Hospital in Solna and Huddinge, Skåne University Hospital, Sahlgrenska University Hospital, Linköping University Hospital, Umeå University Hospital and Uppsala University Hospital. Included hospitals in Denmark are Rigshospitalet Copenhagen, Rigshospitalet Glostrup, Aarhus University Hospital and Odense University Hospital. The inclusion of patient will start in August of 2022 and prolong approximately until December of 2026 or until sufficient number of patients has been included. Enrollment, randomization and blinding Clinicians will approach the parents for enrollment during the first 48 hours postpartum or antenatally if appropriate. After the parents have given written consent and patient given a Study-ID, the legal guardian will be randomly assigned to intervention- or control group in a 1:1 fashion. Randomization will be done within 24 hours after parents have given written consent. The study will be double-blinded. The preparation for intervention- and control group will be made by assistant nurses working in the nutrition kitchen where the enteral nutrition is being prepared for all patients at NICUs in Sweden. Thus, the medical and nursing teams caring for the infants, as well as participating researchers will be unaware of the type of supplementation and will not have knowledge on whether group A or B is receiving the probiotic supplement. See detailed description about enrollment, randomization- and blinding process under heading "Material: Patient selection - population, sample". INTERVENTION The intervention product, commercially known as ProPrems®, is a combination of (one billion freeze-dried bacteria per 0.5 g in a maltodextrin base powder: Bifidobacterium infantis Bb-02 (DSM 33361) 300 million, Bifidobacterium lactis (BB-12®) 350 million, and Streptococcus thermophilus (TH-4®) 350 million. The supplementation of probiotics or pure breastmilk will be administered by gastric tube and starts when the infant tolerates at least 3 mL breastmilk per meal. For the intervention group, the standard dose of 0.5 grams will be mixed with 3 mL breastmilk. If the mother's breastmilk is available this will be used as first choice, if not, donated breastmilk will be used. The dose will be administered once daily until 34 weeks of gestational age and will not change with increasing age and weight. If enteral feeds and probiotic supplementation feeds are not tolerated after initiation and have to be discontinued or reduced below 3ml/meal, supplementation will be paused and only started again when the infant is tolerating 3 mL per meal again. CONTROL The control group will receive 3 mL of mothers' own breastmilk or donated breastmilk without the probiotic supplement and administered by gastric tube. This may however be administered as bolus or continues feeds according to local feeding strategy. OUTCOME Primary outcome variables are NEC rate and overall mortality. Secondary outcomes are incidence of sepsis, duration of hospital stay, use of antibiotics. Tertiary outcomes include feeding tolerance, and growth and body composition after hospital discharge. RESEARCH QUESTIONS Primary aim and secondary aim 1 and 2: Do probiotic supplementation reduce the incidence of a composite endpoint of NEC and neonatal mortality among EPT infants? Participants: EPT infants, 22 0/7 to 26 6/7 gestational weeks, at all sites. Data collection: Incidence of NEC and mortality will be registered and retrieved from the Swedish Neonatal Quality (SNQ) Register. Incidence of NEC will be defined according to Bell's stage II and above. Bell stage II includes clinical signs with abdominal symptoms and radiological findings. Stage III includes findings in stage II in addition to severe clinical condition such as need of intubation and/or need for surgery. Registered incidence of NEC in SNQ will be assessed by a neonatologist before inclusion in the analyses and must include the following: 1. Abdominal symptoms in connection to NEC episode 2. Radiological findings 3. Possible differential diagnoses 4. Abdominal surgery Data analysis: Relative risk (RR) and 95% confidence intervals (CI) using standardized logistic regression (see "STATISTICAL METHODS" for more information. Secondary aim 3: Does probiotic supplementation reduce the incidence of neonatal sepsis? Participants: EPT infants, 22 0/7 to 26 6/7 gestational weeks, at all sites. Data collection: Incidence of neonatal sepsis will be registered and retrieved from SNQ. Incidence of sepsis will be defined as blood culture proven sepsis and/or laboratory inflammatory response were at least one of the following must be confirmed: 1. LPK \< 5 or \> 20 (x109 cells/L) 2. TPK \< 100 (x109 cells/L) 3. CRP \> 15 mg/L Data analysis: RR and 95% CI using standardized logistic regression. Secondary aim 4: Does probiotic supplementation reduce the rate of prescription and duration of antibiotic use during hospital stay? Participants: EPT infants, 22 0/7 to 26 6/7 gestational weeks, at all sites. Data collection: Prescription and duration of antibiotics will be registered and retrieved from SNQ. Data analysis: RR and 95% CI using standardized logistic regression. Secondary aim 5: Does probiotic supplementation reduce duration of hospital stay? Participants: EPT infants, 22 0/7 to 26 6/7 gestational weeks, at all sites. Data collection: Duration (number of days) of hospital stay will be registered and retrieved from SNQ. Data analysis: Change in mean duration and 95% CI using linear regression. Secondary aim 6: Does probiotic supplementation influence the development of gut microbiome? Participants: EPT infants, 22 0/7 to 26 6/7 gestational weeks, first 100 included infants, at all sites. Data collection: Fecal samples collected directly after inclusion, 14 days of life, 34 week of gestational age and 12 months corrected age. Data analysis: Differences in gut microbiome between intervention and control group. Tertiary aim 1: Does probiotic supplementation reduce the incidence of enteral feeding intolerance during intervention period? Participants: EPT infants, 22 0/7 to 26 6/7 gestational weeks, Karolinska University Hospital in Solna and Huddinge. Data collection: Daily assessment of nutritional intake including macro-and micronutrients content and type of nutritional product will be calculated and retrieved from nutritional calculation program (Nutrium.se). Primary effect measurement is days until full enteral feeds. Data analysis: RR and 95% CI using standardized logistic regression. Tertiary aim 2: Does probiotic supplementation improve growth (increase of weight, length, and head circumference) during intervention period? Participants: EPT infants, 22 0/7 to 26 6/7 gestational weeks, Karolinska University Hospital in Solna and Huddinge. Data collection: Weakly measurement of weight, length, and head circumference during ongoing intervention, including standard deviation score (SDS) in growth curve. Data analysis: Change in mean SDS and 95% CI using linear regression. Tertiary aim 3: Does probiotic supplementation improve postnatal growth after hospital discharge? Participants: EPT infants, 22 0/7 to 26 6/7 gestational weeks, at all sites in Sweden. Data collection: Weight, length, and head circumference, including Z-score in growth curve, at 40 weeks of gestational age, 1-year corrected age, 2.5 years corrected age, and 5.5 years of age. This will be registered and retrieved from SNQ. All growth measurements are taken according to clinical routine when patients attend the national neonatal follow-up program at above given timepoints. Data analysis: Change in mean SDS and 95% CI using linear regression. Tertiary aim 4: Does probiotic supplementation affect body composition at follow-up during the first 6 months of life? Participants: EPT infants, 22 0/7 to 26 6/7 gestational weeks, Karolinska University Hospital in Solna and Huddinge. Data collection: Body composition (using PeaPod) measured at 40 gestational weeks and at three months corrected age. Data analysis: Change in mean fat and fat-free mass and 95% CI using linear regression. All variables collected from medical records will be recorded in an electronic case report form. Fecal samples Fecal samples to identify and analyze possible differences in intestinal flora between intervention and control group will be collected at all Swedish hospitals at the following timepoints: * Directly after inclusion ± 2 days * 14 ± 2 days gestational age * 34 weeks ± 2 days gestational age * 12 months corrected age ± 2 weeks The samples will be collected by nurse and stored immediately in -20º C or -80º C. As soon as possible samples should be transferred frozen to a -80º C freezer before transfer to Biobank. Microbial composition will be analyzed using next-generation sequencing (NGS). In addition to traditional microbiota profiling using a 16S seq approach we will also apply shotgun metagenomic sequencing on a selection of the samples to achieve information about functional genes in the microbiome. We have access to the Illumina Novaseq platform for this purpose and have developed a pipeline for large-scale analyses of shotgun data. DNA extraction, sequencing library preparation and sequencing will take place at the Centre for Translational Microbiome Research (CTMR) MATERIAL: PATIENT SELECTION - POPULATION, SAMPLE Population and study period The study population will include extremely preterm infants (born between 22+0 - 27+6 weeks gestational age) approximately between December 2022 and December 2026 or until sufficient number of patients have been included. Management of differences in treatment routine/clinical practice To avoid the risk of differences in treatment routine, the randomization will be stratified by center. Furthermore, the data collection from registers will include important variables in neonatal patient care that may vary between centers, such as fortification strategies, fluid regime, approach to hemodynamic instability, antibiotic strategies, strategies for PDA closure and ventilator support, more? Inclusion and randomization Clinicians will approach the legal guardians for enrollment during the first 48 hours postpartum or antenatally. The legal guardians will be given verbal and written information about the study. After the legal guardians have given written consent, the patient will be randomly assigned to intervention or control group. Randomization will be done within 24 hours after written consent. Allocation to either intervention or control group will be made using the randomization software Randomize.net. The randomization will be done by the principal investigator of each hospital. The randomization will be stratified by center and gender. Data collection and analysis will be made using a binary treatment code to remain group allocation of blinding until results are finalized. The patient will be allocated to either group A or group B. Blinding The study will be double-blinded. The preparation for intervention and control group will be made by assistant nurses working in the "nutrition kitchen" where the enteral nutrition is being prepared for all patients at the NICUs. Thus, the medical and nursing teams caring for the infants, as well as all researchers will be blinded and not have knowledge on weather group A or B is receiving the probiotic supplement. A decrypted code with information if group A and B are receiving the probiotic supplement will be kept locked in the nutrition kitchen. The assistant nurse will use this code to know which milk to prepare. Before study start, all assistant nurses working in the nutrition kitchen must read an information brochure and sign a contract stating they have read the information and will not spread information about group allocation. The same regulations will apply for the fraction of patients relocated before 34 weeks of gestational age to other hospitals than the primary study center. ESTIMATED SAMPLE SIZE AND POWER The sample size calculation is based on the Swedish average incidence of NEC and mortality in EPT infants. Denmark has a comparable incidence of NEC and mortality among EPT infants. As most EPT born infants die during the first 2-3 days of life we calculate mortality rate from the third day of life as this is the estimated time the patients will be exposed to the intervention treatment. Based on pooled incidence rate data from SNQ data from the years of 2019, 2020 and 2021, we considered a 18% background rate of composite outcome of NEC and mortality, with an estimated 9% overlap. We estimate a 30% reduction rate as this was considered clinically relevant and was also the risk reduction rate of NEC that was seen in the ProPrems trial (6). With analysis of relative risk for the power analysis, assuming a significant level of 0.05 and 30% reduction rate with interim analysis based on a Pocock boundary at 1/3, 2/3s and 100% of the information. The proposed enrollment of 1620 patients will have sufficient power (80%) to test our hypotheses. The inclusion of patients will prolong until this number of patients has been included. Assuming an 80% consent rate, we would expect to enroll 1620 subjects in three years and eight months. STATISTICAL METHODS Primary aim 1 and secondary aims 1-3. In our initial analysis, we will compare baseline characteristics of the study population by randomized intervention assignment to ensure that balance was achieved by the randomization. We will report the mean (standard deviation \[SD\]) or median (first quartile, third quartile) for continuous variables, and count and percentages for categorical variables. The primary analyses the intention-to-treat (ITT) sample, including all subjects that were randomized in the study. A crude and a multivariable-adjusted adj standardized logistic regression will be used to calculate the relative risk (RR) and 95% confidence interval (CI) to estimate the effect of probiotic supplementation on the risk of NEC and total mortality (primary aim) as well as NEC, total mortality and neonatal sepsis separately (secondary aims 1-3). The randomization will be stratified on site and gender and thus the analysis will adjust for at least these variables before standardization, although other variables may be included. The decision to include other adjustment variables will be based on expert knowledge and will be decided and finalized prior to any data collection regression model. All interim analyses will use the same method. A per-protocol analysis accounting for non-compliance will also be performed. All data analyses will be using the STATA and R software. A two-sided p-value of 0.05 will be considered significant evidence of the effect of the intervention. Secondary aim 4 and 5: The effect of probiotic supplementation on the rate of prescription and duration of antibiotic use during hospital stay (secondary aim 4) and duration of hospital stay (secondary aim 5) will be analyzed using standardized logistic regression and linear regression model. Secondary aim 6: The effect of probiotic supplementation on gut microbiome will be investigated by comparing the alpha and beta diversity of the microbiota samples. Alpha diversity measures the mean diversity within a sample; beta diversity involves a comparison of diversities between samples, and will be displayed as a principle coordinate or component analysis plot. Tertiary aim 1-3: We will use standardized logistic regression to estimate the risk of enteral feeding intolerance (tertiary aim 1) and linear regression model to estimate changes in growth (weight, length, and head circumference) during hospital stay (tertiary aim 2), growth after hospital discharge (tertiary aim 3), and body composition (tertiary aim 4). Responsible statistician for the project is Erin Gabriel, Associate Professor in Biostatistics, University of Copenhagen and Karolinska Institutet. PROJECT ORGANIZATION The following regions and researchers will collaborate in this trial: 1. Region Stockholm (Karolinska University Hospital Solna/Huddinge) and Karolinska Institutet. Alexander Rakow, MD, PhD, is the principal investigator (PI), developer of the project and main supervisor for the PhD-student Sofia Söderquist Kruth working on the project. His role is to coordinate the trial, ensure a safe and friendly research environment, secure financing for the trial, structure data management, finalizing trials results, manuscript preparation of primary and secondary endpoints and facilitate progress. Alexander Rakow is currently involved in two large multicenter trial and is PI for the Stockholm sites. He is a senior consultant neonatologist and clinical director at Medical Unit Neonatology and thereby contributing as the main clinical expert on the field. He will dedicate at least 20% of his time to this project which agrees with his organization. Susanne Rautiainen Lagerström, MD, PhD, is a senior researcher at Astrid Lindgren's Children's Hospital (ALB), Dept of Global Public Health, Karolinska Institutet, and Brigham and Women's Hospital and Harvard Medical School in Boston using randomized controlled trials to provide high-quality evidence for complex nutritional and disease relationships. Co-supervisor of the PhD-student Sofia Söderquist Kruth. Elisabeth Stoltz Sjöström, PhD, registered Dietitian, senior lecturer, researcher within the field of neonatal nutrition and Director of Doctoral Studies at the Department of Food, Nutrition and Culinary Sciences, Umeå University. Co-supervisor of the PhD-student Sofia Söderquist Kruth. Eric Herlenius, MD, PhD, Professor of Pediatrics, Head of pediatric emergency and infectious disease. Head of the Swedish CAP-trial group a double-blind placebo-controlled study, involving 35 neonatal centres worldwide. Co-supervisor of the PhD-student Sofia Söderquist Kruth. 2. Region Västerbotten (Umeå University Hospital). Magnus Domellöf, MD, PhD is a Professor of Neonatology with expertise on research in neonatal nutrition including RCT:s and population-based studies. He is the chair for the Swedish Neonatal Societies working group for national nutritional guidelines. He will be PI for Region Västerbotten. 3. Region Östergötland (Linköping University Hospital). Thomas Abrahamsson, MD, PhD is Associate Professor and neonatologist with experience conducting RCT:s as well as research on another probiotic supplement among preterm infants. He is a member of the Swedish Neonatal Societies working group for national nutritional guidelines. He will be PI for Region Östergötland. 4. Region Västra Götaland (Sahlgrenska University Hospital). Anders Elfvin, MD, PhD Associate professor and neonatologist. His researched has mainly focused on development of NEC and sepsis. He is a member of the Swedish Neonatal Societies working group for national nutritional guidelines. He will be PI for Region Västra Götaland. 5. Region Skåne (Lund University Hospital). Ingrid Pupp, MD, PhD Associate Professor and neonatologist. She has great experience working on RCT:s and cohort studies with nutritional focus. She is a member of the Swedish Neonatal Societies working group for national nutritional guidelines. She will be PI for Region Skåne. 6. Region Uppsala (Uppsala University Hospital) Fredrik Ahlson, MD, PhD Associate Professor and neonatologist with many years' experience conducting research with nutritional focus on preterm infants. He is a member of the Swedish Neonatal Societies working group for national nutritional guidelines. He will be PI for Region Uppsala. 7. Rigshospitalet Copenhagen and Glostrup. Lise Aunsholt, MD, PhD, is a neonatologist and senior lecturer and has mainly focused her research on neonatal nutrition. She will be the PI for Rigshospitalet Copenhagen and Glostrup. 8. Aarhus University Hospital Jesper Padkaer Petersem, MD, PhD, is a senior neonatologist. He is the chair for the Danish Neonatal Quality Register. He will be the PI for Aarhus University Hospital. 9. Odense University Hospital Gitte Zacharissen MD, PhD is Professor of Neonatology with experience of RCT:s and population-based research with nutritional focus. She will be the PI for Odense University Hospital. Coordination and project management. Department of Women and Children´s Health and Centrum för Klinisk Barnforskning (CKB) will coordinate project coordinator 100% dedicated to project will help and coordinate Regional and national member sites, CRF collection, biobanking and analysis.