Lactobacillus reuteri Colonisation of Extremely Preterm Infants in a Randomised Placebo-Controlled Trial

Abstract

Lactobacillus reuteri DSM 17938 supplementation reduces morbidities in very low birth weight infants (<1500 g), while the effect on extremely low birth weight infants (ELBW, <1000 g) is still questioned. In a randomised placebo-controlled trial (ClinicalTrials.gov ID NCT01603368), head growth, but not feeding tolerance or morbidities, improved in L. reuteri-supplemented preterm ELBW infants. Here, we investigate colonisation with the probiotic strain in preterm ELBW infants who received L. reuteri DSM 17938 or a placebo from birth to postmenstrual week (PMW) 36. Quantitative PCR was used on 582 faecal DNA samples collected from 132 ELBW infants at one, two, three, and four weeks, at PMW 36, and at two years of age. Human milk oligosaccharides were measured in 31 milk samples at two weeks postpartum. At least 86% of the ELBW infants in the L. reuteri group were colonised with the probiotic strain during the neonatal period, despite low gestational age, high antibiotic pressure, and independent of infant feeding mode. Higher concentrations of lacto-N-tetraose, sialyl-lacto-N-neotetraose c, and 6'-sialyllactose in mother's milk weakly correlated with lower L. reuteri abundance. Within the L. reuteri group, higher L. reuteri abundance weakly correlated with a shorter time to reach full enteral feeding. Female sex and L. reuteri colonisation improved head growth from birth to four weeks of age. In conclusion, L. reuteri DSM 17938 supplementation leads to successful colonisation in ELBW infants.

Keywords: Lactobacillus reuteri; antibiotic; extremely low birth weight; feeding intolerance; human milk oligosaccharide; premature; probiotic; randomised controlled trial.

Figure 1

Flow chart of the PROPEL trial. PMW 36 = postmenstrual week 36. ^a Study product was discontinued by mistake after transfer to another hospital (n = 1). ^b Study product was not administered by mistake after temporarily being withheld during nil oral (n = 3). ^c Study product ran out temporarily at the study site (n = 3).

Figure 2

L. reuteri in faeces of the L. reuteri-supplemented and placebo group. L. reuteri prevalence (a) and abundance (b) in faeces of extremely preterm ELBW infants. Prevalence is expressed as the percentage of infants with a L. reuteri-positive faecal sample. Abundance is shown as the number of L. reuteri bacteria per 1 g wet faeces. Boxplots show median with 25% and 75% percentiles and 1.5× the interquartile range. Numbers (n / N) indicate the number of infants with a L. reuteri-positive faecal sample and the total number of infants per group. For graphical display, the number of L. reuteri bacteria per 1 g wet faeces in (b) was set to 10⁰ for infants with L. reuteri-negative faeces. Statistics: Fisher’s exact test with Benjamini–Hochberg correction (a), Mann–Whitney U test with Benjamini–Hochberg correction (comparisons between supplementation groups) and Kruskal–Wallis test with Dunn’s post hoc test with Benjamini–Hochberg correction (comparison of L. reuteri abundance in the L. reuteri-supplemented group over time) (b). ** adjusted p < 0.01, *** adjusted p < 0.001. ELBW = extremely low birth weight, PMW = postmenstrual week.

Figure 3

Antibiotic treatment in the L. reuteri-supplemented and placebo group. Bars show antibiotic treatment rates (in %) for placebo (yellow) and L. reuteri-supplemented (blue) extremely preterm ELBW infants, who were treated with the specified antibiotics for at least one day during the first, second, third, and fourth week of life. ELBW = extremely low birth weight.

Figure 4

Antibiotic treatment and L. reuteri abundance in faeces of the L. reuteri-supplemented group. L. reuteri abundance in faeces of L. reuteri-supplemented extremely preterm ELBW infants at one (a), two (b), three (c), and four (d) weeks of age, who were not (grey dots) or were (red dots) treated with antibiotics for at least one day during the week preceding faecal sampling. Abundance is expressed as L. reuteri bacteria per 1 g wet faeces. Boxplots show median with 25% and 75% percentiles and 1.5× the interquartile range. For graphical display, the number of L. reuteri bacteria per 1 g wet faeces was set to 10⁰ for infants with L. reuteri-negative faeces. Numbers below the panels indicate the number of infants in the respective group. This figure only shows antibiotics, for which the sample size at any week was at least 10, while data on less common antibiotics are presented in Table S5. Statistics: Mann–Whitney U test with Benjamini–Hochberg correction. * adjusted p < 0.05, ** adjusted p < 0.01. ELBW = extremely low birth weight.

Figure 5

Infant feeding and L. reuteri colonisation in the L. reuteri-supplemented group. L. reuteri prevalence (a) and abundance (b) in faeces of L. reuteri-supplemented extremely preterm ELBW infants by feeding type. Prevalence is expressed as the percentage of infants with a L. reuteri-positive faecal sample and the total number of infants in each feeding category. Abundance is shown as the number of L. reuteri bacteria per 1 g wet faeces. Boxplots show median with 25% and 75% percentiles and 1.5× the interquartile range. For graphical display, the number of L. reuteri bacteria per 1 g wet faeces in (b) was set to 10⁰ for infants with L. reuteri-negative faeces. Infants in the partially breast milk-fed (‘mixed’) feeding group received 1–50% breast milk. Statistics: Fisher’s exact test (a), Kruskal–Wallis test (b). ELBW = extremely low birth weight, PMW = postmenstrual week.

Figure 6

Human milk oligosaccharides and L. reuteri abundance in faeces of the L. reuteri-supplemented group. (a–o) Spearman correlation between human milk oligosaccharide concentrations (µmol/L) in milk collected at two weeks postpartum from exclusively breastfeeding mothers (n = 31) and L. reuteri abundance in faeces at three weeks of age (n = 36) from L. reuteri-supplemented infants. Abundance is shown as the number of L. reuteri bacteria per 1 g wet faeces. For infants with a faecal sample negative for L. reuteri, the number of L. reuteri bacteria per 1 g wet faeces was set to 10⁰. p values were adjusted with the method from Benjamini and Hochberg. DSLNT = disialyl-lacto-N-tetraose, 2′ FL = 2′-fucosyllactose, 3′ FL = 3′-fucosyllactose, LDFT = lacto-difucotetraose, LNDH I = lacto-N-difucohexaose I, LNFP I = lacto-N-fucopentaose I, LNFP II = lacto-N-fucopentaose II, LNFP III = lacto-N-fucopentaose III, LNnT = lacto-N-neotetraose, LNT = lacto-N-tetraose, LSTa = sialyl-lacto-N-tetraose a, LSTb = sialyl-lacto-N-tetraose b, LSTc = sialyl-lacto-N-neotetraose c, 3′ SL = 3′-sialyllactose, 6′ SL = 6′-sialyllactose.

Figure 7

Growth and clinical outcomes in the L. reuteri-colonised L. reuteri-supplemented and non-colonised placebo group. (a) Criteria for selection of infants from the placebo and L. reuteri-supplemented groups for inclusion in the comparisons. (b) Boxplots (median with 25% and 75% percentiles and 1.5× the interquartile range) show the time to full enteral feeding of placebo and L. reuteri-supplemented infants with a faecal sample negative or positive for L. reuteri at one week of age, respectively. (c–e) Line graphs show the change in weight, length, and head circumference z-scores (mean with 95% confidence interval) from birth to two and four weeks of age in placebo and L. reuteri- supplemented infants with a faecal sample negative or positive for L. reuteri one week prior to the growth measurement, respectively. (f–k) Bars show prevalence of clinical outcomes of placebo and L. reuteri-supplemented infants with a faecal sample negative or positive for L. reuteri at one week of age, respectively. There were no significant differences in the prevalence of the clinical outcomes if infants were grouped based on L. reuteri colonisation at two, three, or four weeks of age. Infants were excluded if data on the specified outcome were missing or if outcomes occurred before or on the day of faecal sampling. One infant from the placebo group did not achieve full enteral feeding by PMW 36. This infant was included in the statistical analysis but is not shown in the plot in (b). Statistics: Mann–Whitney U test (b), Student’s t-test with Benjamini–Hochberg correction (c–e), Fisher’s exact test (f–k). ** adjusted p < 0.01. BPD = bronchopulmonary dysplasia, NEC = necrotising enterocolitis, PMW = postmenstrual week, ROP = retinopathy of prematurity.

Figure 8

L. reuteri abundance in faeces of the L. reuteri-supplemented group at one week of age and time to full enteral feeding and sepsis. (a) Bars show percentage of infants with a time to full enteral feeding (TEF) below or equal to the group’s median of 14 days in infants of the L. reuteri-supplemented group, who had a faecal sample positive (L. reuteri-colonised) or negative (non-colonised) for L. reuteri at one week of age. (b) Spearman correlation between L. reuteri abundance at one week of age and TEF (expressed in days after birth) in L. reuteri-supplemented infants. (c) L. reuteri abundance in faeces of L. reuteri-supplemented infants at one week of age who later do or do not develop sepsis. Boxplots show median with 25% and 75% percentiles and 1.5× the interquartile range. Infants were excluded if information on TEF was missing (n = 3) or if infants reached full enteral feeding before faecal sampling at one week of age (n = 3) (a,b). For one infant who did not reach full enteral feeding by PMW 36, for graphical display, TEF was set to the day on which the infant reached PMW 36 (74 days) (b). Three infants were excluded because sepsis occurred before faecal sampling at one week of age (c). L. reuteri abundance is shown as the number of L. reuteri bacteria per 1 g wet faeces and, for graphical display, the number of L. reuteri bacteria per 1 g wet faeces was set to 10⁰ for infants with L. reuteri-negative faeces (b,c). Statistics: Fisher’s exact test (a), Mann–Whitney U test (c). * p < 0.05.