The results of our innovative RCT in very preterm infants showed that the heat-inactivated probiotic was safe, well tolerated, with an intestinal anti-inflammatory effect comparable to that of the live probiotic. The FCP, fecal SCFA and microbiota were comparable between the HP and P group infants. At T2 compared with T1, Actinobacteria significantly increased while Bacteroidetes decreased (p < 0.05 for both groups); Bifidobacteriacae increased (p < 0.001) and Staphylococcaceae decreased for both groups (p < 0.05). There were no adverse events including probiotic sepsis. Clinical outcomes were comparable between the groups.
Our results are significant considering the advantages of heat-inactivated probiotics in alleviating the concerns about supplementation with live probiotics in preterm infants. Apart from the risk of probiotic sepsis, and development of antibiotic resistance, the concerns include the possibility that persistent gut colonization with live probiotic microorganisms may prevent/delay the establishment of core microbiome [4–7, 18]. Heat-inactivated probiotics could mitigate such concerns [19, 20]. Other advantages of heat-inactivated probiotics include longer shelf-life, ease of transport and storage without the need for cold-chain maintenance, and lower cost [7, 21].
Heat-inactivation is inexpensive, simple, and feasible across different set ups (e.g., tropical climate, limited resources) compared with other methods of inactivating live probiotics. These include γ or UV rays, formalin, mechanical disruption, pressure, sonication, lyophilisation and acid deactivation [8]. Heat-inactivation is known to release beneficial bacterial components such as lipoteichoic acids, peptidoglycans or exopolysaccharides with immunomodulatory and pathogen antagonistic effects [7]. Other pathways for their benefits include IgA production, increased expression of sIgA receptor, suppression of pro-inflammatory cytokine production, maintenance the ability to adhere to the gut epithelial surface, down-regulation of TNF-alpha expression, reduction of translocation, bacteriocin release and interference with biofilm formation [7, 22].
Campeotto et al have evaluated the effects of feeding with a standard (SF; n = 24) or fermented preterm formula (FF) containing heat-inactivated B. breve C50 and Streptococcus thermophilus 065; n = 34) in 58 preterm infants with mean (SD) gestation 33+ 5 (1–3) weeks. They reported no difference in digestive tolerance, except for reduced abdominal distension in FF group during week 3 and 4. Only one infant in SF group had suspected NEC. Gain in birth weight, and increase in weight, length, and head circumference during the study period were comparable. There was no difference in fecal microbiota, particularly bifidobacteria, however, FCP levels were lower in FF group and secretory IgA was increased with mother’s milk and FF [23].
Compared to Campeotto et al, our trial enrolled very preterm infants < 32 weeks of gestation exclusively fed with breastmilk with a pre-stated subgroup of infants born < 28 weeks. The method of heat-inactivation and the choice of the probiotic strains differed. Our trial assessed a combination of Bifidobacteria derived from healthy infant stools. Furthermore, there were methodological differences in assessment of fecal microbiota and inflammatory markers. Importantly, our trial was adequately powered for the primary outcome of FCP.
The limitations and strengths of our trial need to be discussed. The limitations include the choice of FCP as the sole marker of intestinal inflammation. We believe that fecal FCP was a pragmatic choice considering the limited information on other markers of intestinal inflammation (e.g., intestinal fatty acid binding protein, trefoil factor, volatile organic compounds) in very preterm infants [13, 24, 25]. Although FCP levels are considered as a reliable marker of intestinal inflammation, elevated levels are also reported with acute and cumulative pain or stress, feeding with mother’s own milk, non-white race, and increasing severity of illness [24]. Others have reported decreasing FCP levels with increasing postnatal age and with exclusive human milk feeding [26]. Ho et al reported positive correlation between relative abundance of Klebsiella, but not Gammaproteobacteria and increasing FCP levels [27]. It was difficult to explain whether the increased fecal bifidobacteria at T2 related to continuation of breastmilk feeding or trial supplementation. The benefits of heat-inactivated strains cannot be ruled out in this context considering their ability to retain probiotic/bifidogenic effects. Given the unavailability of strain-specific primers we were unable to assess cross-contamination. However, the fact that cross-contamination was not a significant issue in our previous probiotic trial is assuring [15]. Our unit policy of strict hand hygiene and preparing individual probiotic doses before administration is important in this context. Finally, it would have been ideal to distinguish between live and dead probiotic bacterial cells using techniques such as flow cytometry [28], however, this was beyond the scope of our project.
The strengths of our study include its robust design including masking of trial investigators and outcome assessors, and adequate power for the primary outcome as FCP. The comparison between heat-inactivated and live probiotic was robust as their composition and dose (3 ×109 CFU/ day) were identical. The validity of our results is supported by the comparable baseline characteristics of participants in the heat-inactivated vs live probiotic group, indicating successful randomization. Except for the trial registered in 2016 (https://clinicaltrials.gov/study/NCT02796703), which never commenced recruitment, we did not find similar studies in this field.