Figure 6.

Proinflammatory biomarkers are significantly lower in preterm infants fed B. infantis EVC001. (A) Multivariate analysis by linear models (MaAsLin2) was used to assess the effect that each clinical variable had, independently, on proinflammatory biomarker concentrations. Probiotic feeding (B. infantis EVC001) was significantly associated with lower TNFα concentrations (-1.06 pg/mL per kg per day; FDR-corrected P value = 0.0011). TPN (mL/kg * day) correlated with increased IL-8 production (1.25 pg/mL per unit increase; FDR-corrected P value = 0.02) and a decrease of IL-1β and TNFα (0.064 and 0.1 pg/mL * day; FDR-corrected P value = 0.02, 0.02, respectively). Day of life at sampling associated with decreased calprotectin, IL-1β, IL-8, and TNFα production at a rate of 0.04, 0.07, 0.1, and 0.08 pg/mL * day, respectively (FDR-corrected P value = 0.0001, 0.0001, 0.002, and 0.0001, respectively). No significant associations between inhaled steroid exposure, diaper rash and treatment, antibiotic exposure, formula feeding, human milk consumption, birth weight, or gestational age at birth were identified. (B) Significantly decreased proinflammatory cytokine, IL-8, correlated with B. infantis EVC001 absolute abundance (P = 0.001, ρ = −0.24). (C) The mean relative abundance of bacterial families in samples analyzed at the ~34-week time point, colored by family, is shown. Mean and standard deviations are available in Supplementary Table 7. (D–I) Box plots represent fecal calprotectin [ug/g] and proinflammatory cytokine concentrations [pg/mg] from infants not fed B. infantis EVC001 (n = 28) and EVC001-fed infants (n = 29) at 34 weeks corrected gestational age. Calprotectin and cytokine concentrations were measured in duplicate using MesoScale Discoveries R and U-plex, respectively. Statistical analysis was completed using Wilcoxon rank-sum test. P-values were adjusted using the FDR method and considered to be statistically significantly decreased if ****P < 0.0001; ***P < 0.001; **P < 0.01; *P < 0.05.