This is an interim intervention report. We have spent limited time to form an initial view of this program and, at this point, our views are preliminary. We plan to consider undertaking additional work on this program in the future.
- What is the program? Mortality rates are high among preterm newborns. Two causes of death in this population are necrotizing enterocolitis (NEC) and sepsis. Supplementation with probiotics (bacteria that are thought to confer health benefits) has been hypothesized to prevent NEC and sepsis and to reduce all-cause mortality among preterm newborns.
- What is the evidence of effectiveness? A meta-analysis of randomized controlled trials from low- and middle-income countries provides strong evidence that probiotics lower all-cause mortality among preterm newborns and decrease the risk of NEC and sepsis.
- How cost-effective is it? Our current best guess is that the cost-effectiveness of probiotic supplementation for preterm newborns is within range of programs we would consider recommending funding in the future. While probiotic supplementation is more expensive than some other death-averting interventions we have reviewed, it is likely to have a large impact on mortality for preterm newborns. However, we have high uncertainty about the cost of probiotic supplementation and the appropriate value to place on averting the death of a newborn. Updates to our estimates for these parameters could substantially affect our cost-effectiveness estimate for this intervention.
- Is there room for more funding? We have not yet reviewed room for more funding or spoken to potential implementers of this intervention.
- Bottom line: We will likely continue to consider probiotic supplementation for preterm newborns as we prioritize programs to recommend funding in the future. We may reach out to organizations implementing this intervention to learn more about costs and funding opportunities.
Published: December 2019
- What is the problem?
- What is the program?
- What is the evidence of effectiveness?
- Is it cost-effective?
- Is there room for more funding?
- Key questions for further investigation
What is the problem?
Southern Asia and sub-Saharan Africa are the two regions with the highest rates of preterm birth.3
While we have not found systematic data on the mortality rate among preterm newborns across countries or regions, the average mortality rate among preterm newborns from a recent meta-analysis of the effect of probiotic supplementation on mortality among preterm newborns in low- and middle-income countries was 8.6%, or 86 per 1,000 preterm newborns, for newborns not receiving the probiotic supplementation intervention.4
Neonatal sepsis and necrotizing enterocolitis (NEC) are two conditions that can cause deaths among newborns.5
Sepsis occurs when the body's response to an infection injures its own tissues and organs.6 Sepsis is most commonly caused by a bacterial infection but can be caused by viral, parasitic, or fungal infection as well.7 In 2018, sepsis was the third leading cause of death among newborns (15% of all neonatal deaths), with the greatest burden being in developing countries.8 Low birth weight and preterm babies have the highest susceptibility to sepsis.9
NEC occurs when a portion of the bowel dies and is especially common among preterm newborns.10 There does not seem to be a clear understanding among medical researchers about what causes NEC, though bacterial colonization of the intestine is a prerequisite for its development.11
What is the program?
The World Health Organization (WHO) recommends antibiotics as part of a treatment strategy for newborns with neonatal sepsis or NEC.12 However, some have proposed prevention of sepsis by providing probiotics to preterm newborns who may be at risk of these conditions.13
Probiotics are bacteria that are thought to confer health benefits.14 Probiotics are hypothesized to prevent sepsis and NEC by suppressing the growth of disease-causing bacteria, limiting the spread of these bacteria outside of the intestine and enhancing immune function.15
In studies to date with preterm newborns in low- and middle-income countries, probiotics are administered enterally (i.e., via the gastrointestinal tract)16 once or twice per day for one week to several weeks.17 These studies use many different strains, though the two most common classes in one meta-analysis were Lactobacillus and Bifidobacterium.18 In studies, probiotics have generally been delivered in hospital settings.19 However, we found one study that provided probiotics to households through community health workers, though this study involved full-term (rather than preterm) newborns.20
Our impression, based on notes from an unpublished conversation with two researchers who study probiotic supplementation for newborns and two studies that mention the prevalence of this practice, is that probiotic supplementation for low-birth-weight or preterm newborns is becoming more common in the United States and other developed countries but is still not widely used.21 We have not independently vetted these claims.
What is the evidence of effectiveness?
There have been a large number of small-scale randomized controlled trials (RCTs) examining the impact of probiotic supplementation for preterm newborns on all-cause mortality in low- and middle-income countries, as well as higher-income countries.22 These provide strong evidence that probiotic supplementation lowers all-cause mortality among preterm newborns and that this may occur through its impact on NEC and sepsis.
In our cost-effectiveness analysis, we model the primary benefits of probiotic supplementation as coming through its effect on all-cause mortality, based on the meta-analysis by Deshpande et al. 2017, which focuses on low- and middle-income countries.
Meta-analysis of RCTs on probiotics in low- and middle-income countries
We view this as a high-quality meta-analysis. The search strategy used to identify papers appears consistent with other high-quality meta-analyses, and the authors provide extensive detail on their search strategy.24 The meta-analysis is limited to RCTs,25 and results are robust to excluding studies that may have a higher risk of bias according to the GRADE (Grading of Recommendations Assessment, Development and Evaluation) framework, a commonly used framework for assessing evidence quality.26
We also place high weight on this review because it is limited to non-high-income countries, which we think are more relevant as potential settings of interest to us, and because it is relatively recent and therefore more likely to capture the current research available on probiotics, since it is our impression this is an active area of research.
However, we have not reviewed individual studies that are part of this meta-analysis, and it is possible that closer inspection would lead us to downgrade our assessment of this evidence. In addition, several of the included studies are small-scale, which may raise concerns about the generalizability of their findings or inflation of effect sizes due to p-hacking or publication bias (though the funnel plot provided does not indicate any evidence of publication bias).27
Deshpande et al. 2017 find probiotic-supplemented preterm newborns had 27% lower all-cause mortality (95% CI 10%-41%, 19 trials, sample size 4196).28 Removing studies with low risk of bias for random sequence generation and allocation concealment gives a similar effect size.29
Deshpande et al. 2017 also report effects separately for other subgroups, based on newborns' gestational age, use of different probiotics, and use of single-strain vs. multistrain probiotics.30 Effects on all-cause mortality are similar across these subgroups.31
In addition to effects on all-cause mortality, the meta-analysis finds probiotic supplementation reduces risk of NEC by 54% (95% CI 39%-66%)32 and late-onset sepsis by 20% (95% CI 9%-29%).33 As with the effects on all-cause mortality, these impacts are still statistically significant, though smaller in magnitude, when the analysis is limited to studies with low risk of bias for random sequence generation and allocation concealment.34
Meta-analyses of RCTs on probiotics that include developed countries
The findings of Deshpande et al. 2017 are broadly consistent with recent meta-analyses that are not limited to low-and middle-income countries:
- AlFaleh and Anabrees 2014 is a Cochrane Review that includes studies published through 201335 that use RCTs and quasi-experiments and involve preterm newborns.36 They find preterm newborns who received probiotic supplementation had 35% lower all-cause mortality (95% CI 19%-48%, 17 studies, sample size 5112),37 57% lower incidence of NEC (95% CI 44%-67%, 20 studies, sample size 5529),38 and 9% lower incidence of sepsis (95% CI -3%-20%, 19 studies, sample size 5338).
- Sawh et al. 2016 updated the review in AlFaleh and Anabrees 2014 to include studies on the effect of probiotics published from 2013 to 2016.39 Included studies were RCTs40 involving preterm newborns.41 They find preterm newborns who received probiotic supplementation had 21% lower all-cause mortality (95% CI, 7%-32%, 29 trials, sample size 9507),42 47% lower incidence of NEC (95% CI 34%-58%, 38 trials, sample size 10520),43 and 12% lower incidence of sepsis (95% CI, 0%-23%, 31 trials, sample size 8707).44
The research we reviewed also discusses the following additional benefits of probiotic supplementation for newborns, beyond reducing all-cause mortality and likelihood of NEC and sepsis:46
- Neurodevelopmental impacts. Some of the research we have reviewed indicates that NEC and sepsis may affect long-term brain development.47 This suggests that preventing NEC and sepsis among preterm newborns through probiotic supplementation may benefit newborns through improvements in brain development. We have not reviewed these claims and so do not know how reliable they are or how large the effects might be.
- Time to full enteral feeding. The meta-analysis by Deshpande et al. 2017 also includes as one of its primary outcomes time to full enteral feeding (i.e., days elapsed from after birth until the baby stops receiving food intravenously and starts receiving it through its gastrointestinal tract). That analysis finds probiotics led to a decrease of 3.09 days (95% CI 2.69, 3.49) in time to full enteral feeding.48 We have not prioritized understanding the benefits of reduced time to full enteral feeding and how we might model them in our cost-effectiveness analysis.
We do not currently factor these additional benefits into our assessment of the cost-effectiveness of probiotic supplementation.49
Potential offsetting/negative impacts
Some have expressed concerns that widespread use of probiotics could have potentially serious adverse negative effects, including probiotic sepsis, antibiotic resistance and altered immune responses.50
Our impression is that the available evidence suggests these events are rare,51 but we have not reviewed this evidence in depth.
Is it cost-effective?
A preliminary cost-effectiveness model for this intervention is available here. We estimate that the cost-effectiveness of probiotic supplementation for preterm newborns is within range of programs we would consider recommending funding in the future. Probiotic supplementation has a relatively large effect on mortality for preterm newborns, who have a high baseline mortality rate. This drives the cost-effectiveness of this intervention, despite its relatively high cost compared to other death-averting interventions we have reviewed. However, this estimate relies on several assumptions, about which we have a high degree of uncertainty.
Note that our cost-effectiveness analyses are simplified models that do not take into account a number of factors. There are limitations to this kind of cost-effectiveness analysis, and we believe that cost-effectiveness estimates such as these should not be taken literally due to the significant uncertainty around them. We provide these estimates (a) for comparative purposes and (b) because working on them helps us ensure that we are thinking through as many of the relevant issues as possible.
Key uncertainties in our cost-effectiveness analysis, which we would guess have a high potential to change our bottom line on probiotic supplementation for preterm newborns, are:
- Cost of probiotic supplementation. The cost of probiotic supplementation per newborn is estimated to be $28, based on an estimate of $1 per day for the cost of the probiotic supplement and an estimate of roughly four weeks of duration for probiotic supplementation.52 However, we have high uncertainty about this parameter, as well as any additional costs of administering probiotic supplements.53 We estimate that reasonable deviations in this parameter could substantially affect our cost-effectiveness estimate. We may further refine our cost estimate if we decide to reach out to organizations implementing this intervention.
- Valuation of a newborn life. Our current cost-effectiveness analysis does not assign moral weights for averting the death of a newborn. Our best guess is that this weight could range from 1% to 100% of the weight assigned to averting a death of a non-newborn child, so our current cost-effectiveness model for probiotic supplementation assumes a value of 50%. However, we are highly uncertain about this parameter. Updating the moral weight assigned to averting the death of newborn could substantially affect our cost-effectiveness estimate.
Additional uncertainties, which may also influence our cost-effectiveness estimate, are:
- Healthcare costs. By reducing the likelihood of NEC and sepsis, probiotic supplementation may lower costs of treatment of these conditions. Our cost-effectiveness analysis currently excludes healthcare cost savings due to averted costs of treating NEC and sepsis. To the extent that households pay a large share of the costs of treating these conditions, incorporating averted treatment costs could meaningfully improve the cost-effectiveness of probiotic supplementation.
- Adverse events. As discussed above, some research we reviewed expressed concerns that widespread use of probiotics can have potentially serious adverse negative effects, including probiotic sepsis, antibiotic resistance and altered immune responses. We have not thoroughly reviewed the evidence for these claims. If these effects have sound evidence and are large, they would worsen our assessment of the cost-effectiveness of probiotic supplementation, though our current impression is that these events are rare.
- Additional benefits. As discussed above, some of the research we reviewed cited evidence that those newborns who survive NEC and sepsis may experience long-term impacts on brain development and that there may also be additional benefits due to shorter time to full enteral feeds. We have not reviewed the evidence for these claims. If these effects have sound evidence and are large, they could improve our assessment of the cost-effectiveness of probiotic supplementation.
- Development benefits. In our cost-effectiveness analysis for probiotic supplementation, we have included potential development benefits (i.e., increases in later-life income as a result of early-life health interventions) to be consistent with other death-averting interventions implemented by our top charities. However, while some have suggested averting NEC and sepsis can have long-term impacts on brain development, which could potentially improve adult income, we have not identified any direct evidence for these income effects, and as a result, we are highly uncertain about this parameter.
Our cost-effectiveness analysis explores how sensitive results are to adjustments in some key parameters that would make probiotic supplementation more or less cost-effective.
Is there room for more funding?
We have not prioritized reviewing room for more funding for probiotic supplementation to preterm newborns or identified organizations who deliver this intervention. In the future, we may reach out to organizations implementing this intervention to learn more about room for more funding, as well as costs of probiotic supplementation for preterm newborns and specific funding opportunities.
Key questions for further investigation
- What are the costs of this intervention in new settings where it might be implemented?
- How might a funder support implementation of this intervention?
- What is the appropriate moral weight for averting the death of a newborn, relative to averting the death of older children or adults?
- How would incorporating healthcare costs savings due to averted treatment for NEC and sepsis influence the cost-effectiveness of probiotic supplementation?
- How would incorporating additional benefits of probiotic supplementation (i.e., long-term developmental effects and reduced time to full enteral feeding) influence cost-effectiveness?
- How likely are adverse events associated with probiotic supplementation, and how would incorporating these adverse events influence cost-effectiveness?
- Will a forthcoming Cochrane Review by Imdad et al. 2019 update our assessment of the impact of probiotic supplementation for preterm newborns on all-cause mortality?54
"The first 28 days of life – the neonatal period – are the most vulnerable time for a child's survival. Children face the highest risk of dying in their first month of life, at a global rate of 18 (17, 19) deaths per 1,000 live births." WHO, Levels & Trends in Child Mortality, 2019, p. 16.
"The highest risks of death in utero, in the neonatal period and throughout infancy and early childhood, are faced by small and low-birth-weight babies, that is, those who are born preterm or small for gestational age, or both. More than 80% of all newborn deaths occur among small babies in southern Asia and sub-Saharan Africa." WHO and UNICEF 2014, p. 13.
See Figure 1, WHO, Born Too Soon: The Global Action Report on Preterm Birth, 2012, p. 2.
"Data from 19 trials (n=4196) showed reduced risk of death due to all causes in the probiotic versus control group (137/2148 (6.37%) vs 176/2048 (8.59%))." Deshpande et al. 2017, p. 4.
- "Neonatal sepsis and necrotizing enterocolitis (NEC) are neonatal morbidities that can be fatal (Oza, Lawn, Hogan, Mathers, & Cousens, 2015; WHO, 2017b)." Imdad et al. 2019, p. 2.
- "Preterm birth is associated with increased risk of mortality and morbidity including late-onset sepsis (LOS), necrotising enterocolitis (NEC), feeding difficulties and long-term neurodevelopmental impairment. Although survival of preterm neonates has improved in some LMICs, morbidities such as NEC and LOS are still a major issue." Deshpande et al. 2017, p. 1.
"Sepsis arises when the body's response to any infection injures its own tissues and organs. If not recognized early and managed promptly, it can lead to septic shock, multiple organ failure and death." WHO, Sepsis.
"Bacterial infections are the most common cause of sepsis. Sepsis can also be caused by fungal, parasitic, or viral infections." Cleveland Clinic, Sepsis.
See Figure 8, WHO, Levels & Trends in Child Mortality, 2019, p. 16. The top three causes of neonatal mortality globally are: preterm birth complications (35%), intrapartum-related complications (24%) and sepsis (15%).
"Low birth weight and preterm birth predispose the infant to sepsis, and even in the United States, one-quarter of very low birth weight infants (less than 1,500 g) experience culture-positive sepsis, adding to the considerable total burden of morbidity, extended hospital stay, and mortality." Panigrahi et al. 2017, p. 407.
- "Necrotizing enterocolitis: A medical condition where a portion of the bowel dies. It typically occurs in newborns who may be premature, small and sick, and are not fed human milk. Symptoms may include poor feeding, abdominal distension, decreased activity, blood in the stool or vomiting of bile." WHO and UNICEF 2019, p. 129.
- "NEC is a condition that occurs in newborns and can lead to injury to bowel. The extent of injury may vary from mucosal injury to full thickness bowel wall injury. It happens most commonly in preterm babies especially extremely preterm babies (AlFaleh & Anabrees, 2014; Patel & Denning, 2015)." Imdad et al. 2019, p. 2.
- "The pathogenesis of NEC remains incompletely understood. NEC most likely represents a complex interaction of factors causing mucosal injury (Neu 1996). It is speculated that NEC occurs with the coincidence of two of the three pathologic events of intestinal ischemia, colonization of the intestine by pathologic bacteria, and excess protein substrate in the intestinal lumen (Kosloske 1984; La Gamma 1994). Bacterial colonization is necessary for the development of NEC (Kosloske 1990; Musemeche 1986)." AlFaleh and Anabrees 2014, p. 3.
- "The pathophysiology of classic necrotizing enterocolitis is incompletely understood. However, epidemiologic observations strongly suggest a multifactorial cause. The combination of a genetic predisposition, intestinal immaturity, and an imbalance in microvascular tone, accompanied by a strong likelihood of abnormal microbial colonization in the intestine and a highly immunoreactive intestinal mucosa, leads to a confluence of predisposing factors (Figure 2)." Neu and Walker 2011, p. 258.
- "What Causes It? Doctors aren't sure. They do know that premature infants have lungs and intestines that are weak and less mature than those of full-term babies. That means their bodies don't move blood and oxygen around like they should. They also have problems breaking down their food and fighting infection." WebMD, What is necrotizing enterocolitis.
- "What causes NEC? Doctors don't know what causes NEC. It may happen if not enough blood and oxygen reach your baby's immature intestinal tissues. Bacteria from the environment can damage the tender tissues. This can harm the tissues and cause them to die. When this happens, a hole forms in the intestine. This can cause a severe infection in your baby's belly (abdomen)." Stanford Children's Health.
- "Young neonates with suspected necrotizing enterocolitis should be treated with IV or IM ampicillin (or penicillin) and gentamicin as first line antibiotic treatment for 10 days. (Strong recommendation, low quality evidence)." P. 17.
- "Neonates with signs of sepsis should be treated with ampicillin (or penicillin) and gentamicin as the first line antibiotic treatment for at least 10 days. (Strong recommendation, low quality of evidence)." P. 9.
"Necrotizing enterocolitis (NEC) and nosocomial sepsis are associated with increased morbidity and mortality in preterm infants. Through prevention of bacterial migration across the mucosa, competitive exclusion of pathogenic bacteria, and enhancing the immune responses of the host, prophylactic enteral probiotics (live microbial supplements) may play a role in reducing NEC and the associated morbidity." AlFaleh and Anabrees 2014, p. 1.
"Probiotics are live bacteria and yeasts that are good for you, especially your digestive system. We usually think of these as germs that cause diseases. But your body is full of bacteria, both good and bad. Probiotics are often called 'good' or 'helpful' bacteria because they help keep your gut healthy." WebMD, What are probiotics.
- "Necrotizing enterocolitis (NEC) and nosocomial sepsis are associated with increased morbidity and mortality in preterm infants. Through prevention of bacterial migration across the mucosa, competitive exclusion of pathogenic bacteria, and enhancing the immune responses of the host, prophylactic enteral probiotics (live microbial supplements) may play a role in reducing NEC and the associated morbidity." AlFaleh and Anabrees 2014.
- "How the intervention might work? Potential mechanisms by which probiotics may protect high risk infants from developing NEC or sepsis, or both, include an increased barrier to migration bacteria and their products across the mucosa (Mattar 2001; Orrhage 1999), competitive exclusion of potential pathogens (Reid 2001), modification of host response to microbial products (Duffy 2000), augmentation of immunoglobulin A (IGA) mucosal responses, enhancement of enteral nutrition that inhibits the growth of pathogens, and up‐regulation of immune responses (Link‐Amster 1994)." AlFaleh and Anabrees 2014, p. 3.
- "Newborns and preterm babies have immature intestines free of normal commensal bacteria and are more likely to develop NEC and sepsis due to growth of pathogenic bacteria in the intestines (AlFaleh & Anabrees, 2014; Patel & Denning, 2015; Rao et al, 2016). Probiotics are used to proactively colonize the intestines with bacteria like lactobacillus which are known to be beneficial (Millar et al, 2003; Patel & Denning, 2015). Probiotics therefore reduce the growth of pathogenic bacteria which leads to NEC and sepsis. It also increases gut immunity by increasing IgA levels with the help of normal flora which help maintain the mucosal barrier as well (Patel & Denning, 2015). These protective mechanisms also reduce intestinal permeability producing a protective mucosal barrier against bacteria and increase the production of anti‐inflammatory cytokines (Deshpande, Jape, Rao, & Patole, 2017; Millar et al, 2003). Probiotics are especially protective in preterm babies with immature guts and neonates on antibiotics which affects the normal flora of the intestines allowing for colonization by pathogenic bacteria causing NEC. Prebiotics and probiotics can be given together in the form of a synbiotic to improve the gut flora and it can potentially reduce all‐cause neonatal mortality (Johnson‐Henry et al, 2016; Panigrahi et al, 2017)." Imdad et al. 2019, p. 3.
In the meta-analysis by Deshpande et al. 2017, which forms the bulk of our review on the effects of probiotic supplementation for newborns, studies are limited to those with enteral administration: "Intervention and comparison: Enteral administration of probiotic supplement versus control (placebo/no probiotic)." Deshpande et al. 2017, p. 2.
We have not done a thorough review of each study to understand method of implementation detail, but our impression, based on a review of a subset of studies, is that probiotics may be administered as a powder added to breastmilk or formula or through a liquid placed directly into the baby's mouth. These appear to be able to be delivered orally or through feeding tubes. Two examples below:
- "The intervention was initiated on the second day of life of all infants and was maintained until 30 d of life, a diagnosis of NEC, discharge from the hospital, or death, whichever occurred first. At 0900, the infants from the probiotics group received 3 mL human milk from the bank milk to which L. casei and B. breve had been added [half an envelope of Yakult LB (São Paulo, Brazil) providing 3.5 × 107 to 3.5× 109 CFU; the infants in the control group received the same volume of human milk without probiotics]. The probiotics and/or human milk were offered in glass receptacles, identified with the infants' respective names, between 30 and 60 min after preparation, even if the diet had been suspended. In infants using an orogastric tube, the tube was kept closed for 1 h after the probiotics plus human milk or human milk only were administered. Neither the medical and nursing staff responsible for monitoring the infants nor the researchers were aware of which group the infants were allocated to. The primary outcome was the occurrence of NEC, as defined by Bell's criteria and modified by Walsh and Kliegman as stage ≥2 (33)." Braga et al. 2011, pp. 82-83.
- "Study participants were randomly assigned to probiotic or placebo by the use of a computer-generated balanced block randomization scheme. Infants were stratified by institution and by birth weight (≤1500 g and 1501–2000 g). Treatment assignment was performed by using sealed, sequentially numbered, opaque envelopes, color-coded for strata, available in each NICU pharmacy. The pharmacist was in charge of assignment to ensure concealment allocation. Both probiotic and placebo were packaged in identical vials of an oil-based suspension labeled with an individual number indicating the randomization sequence. Infants were administered probiotic or placebo regardless of whether enteric feeds were started. Infants in the probiotic group received 5 drops of an oil-based suspension containing 108 colony-forming units of L reuteri DSM 17938 (BioGaia AB, Stockholm, Sweden) once a day. To maintain stability, the oil-based suspension was stored in a specified refrigerator at 2°C to 8°C. Each 5-mL vial was suitable for 25 five-drop doses. For infants without per oral feeds, the 5 drops were administered through a feeding tube followed by a flush of 0.5 mL of sterile water. For infants on oral feeds, 5 drops were placed in the posterior oropharynx after suctioning oral secretions. For infants in the placebo group, an equal number of drops from an identical vial containing only the oil base were administered following the same protocol as that described for the probiotic group. If during the hospitalization enteric feeds were stopped because of feeding intolerance or NEC, the administration of the probiotic or placebo was also stopped and restarted only when the clinician deemed enteric feeds could be reinstated. Assigned treatment was administered daily until death or discharge from the hospital." Rojas et al. 2012, p. E1114.
See Table 1, "Study characteristics" column, Deshpande et al. 2017, pp. 5-8.
"Out of the 23 included studies, single-strain probiotics were used in 11 studies, whereas 12 used multiple strains. Lactobacillus was part of the supplementation in 13 studies; Bifidobacterium was part of the supplementation in 11 studies and saccharomyces in 3 studies (table 1)." Deshpande et al. 2017.
See Table 1, "Study characteristics" column, Deshpande et al. 2017, pp. 5-8..
This example is from Panigrahi et al. 2017. We exclude this study from our main analysis because it includes full-term newborns, rather than preterm newborns. However, it provides an example of implementation at a large-scale in a community setting. Details of the intervention are below:
- "We enrolled 4,556 infants that were at least 2,000 g at birth, at least 35 weeks of gestation, and with no signs of sepsis or other morbidity, and monitored them for 60 days." Panigrahi et al. 2017, p. 407
- "Written informed consent was obtained during the third trimester of pregnancy by the CHV. She identified the participating newborn after birth, collected maternal and newborn demographic data and prepared and administered synbiotics/placebo in the infant's home under the supervision of the field supervisor (who supplied the clinical trial material from the field office) on the first day of administration. After this, the CHV administered the remaining six daily doses (second to seventh dose). Infants were breastfed and burped before receiving the dose, and watched for 30 min by the CHV. In the case of vomiting, the dose was repeated an hour later, and the child was evaluated at home by a study physician before the next dosing. Infants were followed daily by the CHV until 60 days of age. CHVs and mid-level managers recorded signs or symptoms suggestive of sepsis daily following the pSBI criteria. In the event of suspect sepsis, infants were referred to study hospitals by field staff. Dedicated operators entered data into a web-based system. Data transmission to the University of Nebraska Medical Center server was done after data lock." Panigrahi et al. 2017, p. 413.
- "Probiotics were routinely given to all VLBW infants in 8.8% (44/500) NICUs, while it was given in selected VLBW infants in 5.2% (26/500) of NICUs." Viswanathan et al. 2016, p. 1106.
- "A growing body of data from observational studies and randomized controlled trials indicate that prophylactic probiotic supplementation in preterm infants can reduce the incidence of NEC. However, probiotics have not been widely adopted as a standard clinical practice in all NICUs." Pell et al. 2019, p. 196.
We exclude from our review trials focusing on non-preterm newborns. While our general impression is probiotic supplementation is primarily targeted to preterm newborns, there is at least one large-scale trial we are aware of that focuses implementation on full-term newborns in India (Panigrahi et al. 2017).
Deshpande et al. 2017, p. 1:
- "Eligibility criteria: RCTs comparing probiotics versus placebo/no probiotic in preterm neonates (gestation<37 weeks) conducted in LMICs."
- "Results: Total 23 (n=4783) RCTs from 4 continents and 10 LMICs were eligible for inclusion in the meta-analysis using fixed effect model."
- "Guidelines from the Cochrane Neonatal Review Group (http://neonatal.cochrane.org/resources-review-authors), Centre for Reviews and Dissemination (http://www.york.ac.uk/crd/guidance/) and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement were followed for undertaking and reporting this systematic review and meta-analysis." P. 1.
- "The databases Medline searched via PubMed (https://www.ncbi.nlm.nih.gov 1966–2017), Embase (Excerpta Medica dataBASE) via Ovid (http://ovidsp.tx.ovid.com, 1980–2017), Cochrane Central Register of Controlled Trials (http://www.thecochranelibrary.com, through January 2017), Cumulative Index of Nursing and Allied Health Literature via OVID (http://ovidsp.tx.ovid.com, 1980–January 2017) and E-abstracts from the Pediatric Academic Society meetings (https://www.pas-meeting.org/about/#past, 2000–January 2017) were searched in January 2017. Abstracts of other conference proceedings such as European Academy of Paediatric Societies and the British Maternal and Fetal Medicine Society were searched in Embase. 'Google Scholar' was searched for articles that might not have been cited in the standard medical databases. Grey literature was searched using the national technical information services (http://www.ntis.gov/), Open Grey (http://www.opengrey.eu/), and Trove (http://trove.nla.gov.au/). We have also searched Literatura Latino-Americana e do Caribe em Ciências da Saúde (LILACS) and Caribmed via the BIREME/PAHO/WHO—Latin American and Caribbean Center on Health Sciences Information; PAHO, Pan American Health Organization (http://lilacs.bvsalud.org/en/) using broad terminologies Probiotics OR Probiotic Or Bifidobacterium OR Bifidobacteria OR Lactobacillus OR Lactobacilli OR Saccharomyces. We also searched ClinicalTrials.gov (https://clinicaltrials.gov), International Clinical Trials Registry Platform (http://www.who.int/ictrp/en/) and BioPortfolio (https://www.bioportfolio.com) for ongoing RCTs. The reference lists of eligible studies and review articles were searched to identify additional studies. Reviewers SR, GJ and GD conducted the literature search independently. No language restriction was applied. The non-English studies were identified by reading the recent systematic reviews of probiotic supplementation for reducing the risk of NEC and from cross references of individual studies. Full texts of all non-English studies were obtained via University of Sydney and Department of New South Wales (NSW) health library. A research officer from the NSW Health, University of Sydney translated the articles. Attempts were made to contact the authors for additional data and clarification of methods. Only published data were used for those studies where available.
"PubMed was searched using the following terminology: ((('Infant, Newborn' [Mesh]) OR ('Infant, Extremely Premature' [Mesh] OR 'Infant, Premature' [Mesh])) OR ('Infant, Low Birth Weight' [Mesh] OR 'Infant, Extremely Low Birth Weight' [Mesh] OR 'Infant, Very Low Birth Weight' [Mesh])) AND 'Probiotics' [Majr]. It was also searched using (('Infant, Extremely Premature' [Mesh] OR 'Infant, Extremely Low Birth Weight' [Mesh] OR 'Infant, Very Low Birth Weight' [Mesh] OR 'Infant, Small for Gestational Age' [Mesh] OR 'Infant, Premature, Diseases' [Mesh] OR 'Infant, Premature' [Mesh] OR 'Infant, Newborn, Diseases' [Mesh] OR 'Infant, Newborn' [Mesh] OR 'Infant, Low Birth Weight' [Mesh])) AND ((('Bifidobacterium' [Mesh]) OR 'Lactobacillus' [Mesh]) OR 'Saccharomyces' [Mesh]). The other databases were searched using similar terminologies. The detailed search terminology is given in online supplementary appendix 1." Pp. 2-3.
"Eligibility criteria: RCTs comparing probiotics versus placebo/no probiotic in preterm neonates (gestation<37 weeks) conducted in LMICs." Deshpande et al. 2017, p. 1.
Our assessment of the evidence is based on the following considerations:
- Across all studies, only one study had "high risk" of bias on any dimension, and this was presumably excluded from the sub-group analysis that included only studies with low risk of bias on allocation concealment, since its risk of bias for this dimension is "unclear." See Table 2, Deshpande et al. 2017, p. 9.
- The authors deemed the overall evidence quality as "high" based on the following factors: "The overall evidence according to GRADE guidelines...was deemed high in view of the large sample size, low risk of bias in majority (14/20) of the included studies, narrow CIs around the effect size estimate, very low P value for effect size estimate and mild statistical heterogeneity." Deshpande et al. 2017, p. 10.
- Effect sizes in the meta-analysis are also similar using a random effects model, instead of a fixed effects model. See Table 3, Deshpande et al. 2017, p. 12.
See Figure 6, Deshpande et al. 2017, p. 13.
- "Data from 19 trials (n=4196), showed reduced risk of death due to all causes in the probiotic versus control group (137/2148 (6.37%) vs 176/2048 (8.59%)). Meta-analysis using a FEM estimated a lower risk (RR 0.73 (95% CI 0.59 to 0.90), P=0.003) of death in the probiotic group. " Deshpande et al. 2017, p .4.
- A risk ratio of 0.73 (95% CI 0.59 to .90) is equivalent to a 27% (95% CI 10% to 41%) reduction.
- See also Figure 4, Deshpande et al. 2017, p. 11.
- Results that limiting the analysis to studies with low risk of bias for random sequence generation and allocation concealment are 28% (95% CI 9%-43%, 14 trials, sample size 3366) and 24% (95% CI 4%-40%, 13 trials, sample size 3073), respectively. See Table 3, Deshpande et al. 2017, p. 12.
- "All-cause mortality: studies with low ROB on random sequence generation, RR (95%CI) (FEM): 0.72 (0.57 to 0.91)". A risk ratio of 0.72 (95% CI 0.57 to 0.91) is equivalent to a 28% (95% CI 9% to 43%) reduction.
- "All-cause mortality: studies with low ROB on allocation concealment, RR (95%CI) (FEM): 0.76 (0.60 to 0.96)". A risk ratio of 0.76 (95% CI 0.60 to 0.96) is equivalent to a 28% (95% CI 4% to 40%) reduction.
- The authors note that a limitation is "the fact that nearly 40% of the included trials carried a high risk of bias in many domains of assessment." Deshpande et al. 2017, p. 10. However, the findings from the GRADE framework review note show that only one study has a "high" risk of bias, though approximately 40% have "unclear" risk of bias.
- Results that limiting the analysis to studies with low risk of bias for random sequence generation and allocation concealment are 28% (95% CI 9%-43%, 14 trials, sample size 3366) and 24% (95% CI 4%-40%, 13 trials, sample size 3073), respectively. See Table 3, Deshpande et al. 2017, p. 12.
Results for subgroups are below:
- Newborns with gestational age below 32 weeks or birth weight below 1500 g: reduction in all-cause mortality is 25% (95% CI 7%-39%, 12 studies, sample size 2591).
- Use of probiotic Lactobacillus as part of supplementation: reduction in all-cause mortality is 30% (95% CI 11%-44%, 16 studies, sample size 3473).
- Use of probiotic Bifidobacterium as part of supplementation: reduction in all-cause mortality is 30% (95% CI 7%-48%, 12 studies, sample size 2173).
- Use of single-strain probiotic: reduction in all-cause mortality is 30% (9 studies, sample size 2444).
- Use of multistrain probiotic: reduction in all-cause mortality is 24% (95% CI -3%-44%, 10 studies, sample size 1752).
Deshpande et al. 2017, Table 4, p. 12.
Deshpande et al. 2017 do not report results for newborns fed breastmilk vs. formula. Some researchers have noted that results may be larger for newborns fed breastmilk, but we have not reviewed this source of heterogeneity in any depth: "Probiotics in breastmilk vs. formula‐fed infants: Many believe that probiotics are not required if the infant is fed breast milk – the ideal food provided by nature that contains many bioactive elements including probiotics, human milk oligosaccharides, and lactoferrin. The results of two non‐RCTs are important in this context (Repa et al., 2015; Samuels et al., 2016). Repa et al. (2015) reported overall no significant impact of probiotics on NEC. However, NEC was significantly reduced in probiotic group infants fed any breastmilk [20/179 (11.2%) vs. 10/183 (5.5%); P = 0.027]. No benefits were noted in exclusively formula‐fed infants [4/54 (7.4%) vs. 6/44 (13.6%); P = 0.345] (Repa et al., 2015). Samuels et al. (2016) reported that introduction of probiotics was associated with reduced adjusted odds for 'NEC or sepsis or death' only in exclusively breastmilk‐fed infants [OR: 0.43, 95% CI: 0.21–0.93, P = 0.03]. Our non‐RCT supports the benefits of probiotics in breastmilk‐fed preterm infants (Patole et al., 2016). The reasons why probiotics may not benefit formula‐fed infants to the same extent as those fed breastmilk are easy to understand; no formula could ever replicate breastmilk with its many bioactive components." Athalye-Jape and Patole 2019, p. 251.
Deshpande et al. 2017 do not report formal statistical tests of whether results are statistically different from the main effect of 27% on all-cause mortality. However, based on confidence intervals provided, they do not appear to be statistically different from each other, and differences are small in magnitude (effect sizes ranging from 24%-30%).
See Figure 2, Deshpande et al. 2017, p. 10: "Risk Ratio M-H, Fixed, 95% CI: 0.46 [0.34, 0.61]". A risk ratio of 0.46 (95% CI 0.34 to 0.61) is equivalent to a 54% (95% CI 39% to 66%) reduction.
See Figure 3, Deshpande et al. 2017, p. 10: "Risk Ratio M-H, Fixed, 95% CI: 0.80 [0.71, 0.91]". A risk ratio of 0.80 (95% CI 0.71 to 0.91) is equivalent to a 20% (95% CI 9% to 29%) reduction.
See Table 3, Deshpande et al. 2017, p. 12. These effects are:
- "Definite NEC: studies with low ROB on random sequence generation, RR (95%CI) (FEM): 0.55 (0.40 to 0.74)". A risk ratio of 0.55 is equivalent to a 45% reduction.
- "Definite NEC: studies with low ROB on allocation concealment, RR (95%CI) (FEM): 0.48 (0.34 to 0.66)". A risk ratio of 0.48 is equivalent to a 52% reduction.
- "LOS: studies with low ROB on random sequence generation, RR (95%CI) (FEM): 0.85 (0.74 to 0.97)". A risk ratio of 0.85 is equivalent to a 15% reduction.
- "LOS: studies with low ROB on allocation concealment, RR (95%CI) (FEM): 0.86 (0.75 to 0.99)". A risk ratio of .86 is equivalent to a 14% reduction.
Deshpande et al. 2017 do not report formal statistical tests of whether results are statistically different from the main effect of 27% on all-cause mortality.
"For this update, searches were made of MEDLINE (1966 to October 2013), EMBASE (1980 to October 2013), the Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library (2013, Issue 10), and abstracts of annual meetings of the Society for Pediatric Research (1995 to 2013)." AlFaleh and Anabrees 2014, p. 1.
"Only randomized or quasi‐randomized controlled trials that enrolled preterm infants < 37 weeks gestational age or < 2500 g birth weight, or both, were considered." AlFaleh and Anabrees 2014, p. 1.
"Twenty‐four eligible trials were included. Included trials were highly variable with regard to enrolment criteria (that is birth weight and gestational age), baseline risk of NEC in the control groups, timing, dose, formulation of the probiotics, and feeding regimens. In a meta‐analysis of trial data, enteral probiotics supplementation significantly reduced the incidence of severe NEC (stage II or more) (typical relative risk (RR) 0.43, 95% confidence interval (CI) 0.33 to 0.56; 20 studies, 5529 infants) and mortality (typical RR 0.65, 95% CI 0.52 to 0.81; 17 studies, 5112 infants). There was no evidence of significant reduction of nosocomial sepsis (typical RR 0.91, 95% CI 0.80 to 1.03; 19 studies, 5338 infants)." AlFaleh and Anabrees 2014, p.1.
"Twenty‐four eligible trials were included. Included trials were highly variable with regard to enrolment criteria (that is birth weight and gestational age), baseline risk of NEC in the control groups, timing, dose, formulation of the probiotics, and feeding regimens. In a meta‐analysis of trial data, enteral probiotics supplementation significantly reduced the incidence of severe NEC (stage II or more) (typical relative risk (RR) 0.43, 95% confidence interval (CI) 0.33 to 0.56; 20 studies, 5529 infants) and mortality (typical RR 0.65, 95% CI 0.52 to 0.81; 17 studies, 5112 infants). There was no evidence of significant reduction of nosocomial sepsis (typical RR 0.91, 95% CI 0.80 to 1.03; 19 studies, 5338 infants)." AlFaleh and Anabrees 2014, p. 1.
"Updated searches were conducted January 19, 2015. Clinical trial registries were searched on January 14, 2015. Abstracts and conference proceedings were searched on January 15, 2015. On June 3, 2016 another full update of our search strategy was conducted." Sawh et al. 2016, p. 4.
"All randomized clinical trials were considered for inclusion. No language restrictions were applied." Sawh et al. 2016, p. 3.
"Participants: Infants of less than 37 weeks gestation or weighing less than 2,500 g at birth." Sawh et al. 2016, p. 3.
"The incidence of all-cause mortality was significantly reduced in infants receiving probiotics in 29 trials (9,507 patients)—RR 0.79 95% CI [0.68–0.93] (Fig. 4)." Sawh et al. 2016, p. 14. A risk ratio of 0.79 (95% CI 0.68 to 0.93) is equivalent to a reduction of 21% (95% CI 7% to 32%).
"The primary outcome, severe NEC, was significantly reduced in infants who received probiotics compared to placebo with 38 trials (10,520 patients) reporting on this outcome—RR 0.53 95% CI [0.42–0.66]—see Fig. 2." Sawh et al. 2016, p. 14. A risk ratio of 0.53 (95% CI 0.42 to 0.66) is equivalent to a reduction of 47% (95% CI 34% to 58%).
"The incidence of culture-proven sepsis was not different between the probiotics and control—RR 0.88 95% CI [0.77–1.00] in 31 trials comprising 8,707 patients, see Fig. 3." Sawh et al. 2016, p. 14. A risk ratio of 0.88 (95% CI 0.77 to 1.00) is equivalent to a reduction of 12% (95% CI 0% to 23%).
Two other meta-analyses that were cited in research we reviewed are Rao et al. 2016 and Dermyshi et al. 2017. Our impression is that the findings of these meta-analyses are consistent with the meta-analyses in the main text, though this is based on a superficial review and we have not fully investigated the extent to which these meta-analyses may differ from those in the main text.
The study by Panigrahi et al. 2017, which focuses on full-term newborns, find the synbiotic intervention led to an 80% decrease in risk of diarrhea (95% CI 62%-89%) and 34% lower incidence of lower respiratory tract infection (95% CI 12%-49%). See Table 2, Panigrahi et al. 2017, p. 409. We have not explored whether there is evidence for these same benefits among preterm newborns.
- "The excessive inflammatory process initiated in the highly immunoreactive intestine in necrotizing enterocolitis extends the effects of the disease systemically, affecting distant organs such as the brain and placing affected infants at substantially increased risk for neurodevelopmental delays. Indeed, an infant recovering from necrotizing enterocolitis may have nearly a 25% chance of microcephaly and serious neurodevelopmental delays that will transcend concerns that pertain to the gastrointestinal tract." Neu and Walker 2011, p. 255.
- "Apart from the significant economic burden, increased risk of long‐term neurodevelopmental impairment (NDI) is a serious concern; especially in survivors of surgical NEC (Neu, 2018). A policy of 'zero tolerance to NEC' is hence recommended (Swanson, 2013). Similar to NEC, LOS carries significant burden including long‐term NDI due to adverse effects of inflammation on the preterm brain during the critical phase of development (Strunk et al, 2014)." Athalye-Jape and Patole 2019, p. 249.
- "Late-onset sepsis (LOS) is a major cause of mortality and morbidity, including adverse long-term neurodevelopmental outcomes in preterm infants." Rao et al. 2016, p. 2.
"Meta-analysis of data (n=2154) from 13 trials showed significant reduction in TFEF [time to full enteral feeding] in the probiotics versus control group (MD=−3.09 days (95%CI: −3.49 to –2.69), P<0.00001)." Deshpande et al. 2017, p. 4
Our current cost-effectiveness analysis includes "development effects," designed to capture any potential increases in adult income from the health benefits of probiotic supplementation for preterm newborns. We include these to be consistent with other early-life health interventions implemented by top charities but estimate them to be a small share of the overall cost-effectiveness of probiotic supplementation. More detail is in the cost-effectiveness section.
- "Probiotic sepsis, antibiotic resistance and altered immune responses in the long run are the potential adverse effects of probiotics in preterm neonates." Deshpande et al. 2017, p. 13.
- "Probiotics are considered safe; however, there are concerns regarding probiotic supplementation in extremely premature, immunocompromised neonates and few cases of neonatal sepsis have been reported that were thought to be caused by probiotics (Dani et al, 2016). Imdad et al. 2019, p. 3.
- "Long‐term adverse effects: The results of a recent systematic review and meta‐analysis of studies assessing long‐term neurodevelopment of preterm infants enrolled in probiotic RCTs (n = 7) are reassuring in this context (Upadhyay et al, 2018). Six of the 7 RCTs enrolled preterm infants < 33 weeks. Outcomes were assessed at ≥18–22 months of corrected age in 5/7 RCTs. Probiotics had no effect on cognitive and motor impairment, cerebral palsy, visual, and hearing impairment (Upadhyay et al, 2018). Probiotics are potentially neuroprotective given their anti‐inflammatory properties, and ability to reduce NEC, LOS, feeding intolerance, and modulate the gut‐microbiota‐brain axis. Further long‐term data are important to assess this potential benefit of probiotics.
"Probiotic sepsis: The reports of probiotic sepsis and the death of one preterm infant due to fungal sepsis from a contaminated probiotic product justify the concern about probiotic supplementation in preterm infants (Centers for Disease Control and Prevention, 2014; Bertelli et al, 2015; Esaiassen et al, 2016). However, it is important to know that probiotic sepsis is easy to diagnose and treat compared to the serious hospital acquired infections they prevent. The cost‐benefit ratio is very much in favour of probiotics considering the data from over 12 000 preterm infants who have received probiotics in RCTs and non‐RCTs. Independent product quality checks, and onsite laboratory back up is important to optimize safety of probiotics (Deshpande et al, 2011)." Athalye-Jape and Patole 2019, p. 251.
- "Although none of the meta-analyses on probiotic supplementation in preterm infants reported any serious adverse events in infants receiving probiotics, there are several reports describing the occurrence of probiotic-related infections such as sepsis, pneumonia, and meningitis. In addition, the genome of probiotic bacteria is known to contain genetic elements which confer resistance to several broad-spectrum antibiotics, such as glycopeptides, aminoglycosides, mono-bactams, and fluoroquinolones; these resistant genes are themselves not harmful, but they can be transferred to other gut bacteria and eventually to opportunistic pathogens, leading to an uncontrolled increase in antibiotic-resistance, which might lead to longer-term consequences that outweigh the immediate benefits of probiotic supplementation." Aceti et al. 2018, p. 5.
- "Accumulating evidence indicates that prophylactic probiotic supplementation in preterm infants can reduce the incidence of NEC. However, substantial knowledge gaps, regulatory issues, and implementation challenges should be addressed before probiotics are introduced as standard of care for all preterm neonates. Limitations of published trial data have made it challenging to define regimens that optimize efficacy and safety in specific patient subgroups. Moreover, the current probiotic market lacks rigorous regulatory oversight, which could raise concerns about the quality and safety of probiotic products. Finally, implementation pitfalls include risks of cross-colonization and resource requirements to monitor and mitigate potential adverse events." Pell et al. 2019, p. 195.
"None of the studies reported any significant adverse effects including probiotic sepsis." Deshpande et al. 2017, p. 10.
More details of this calculation can be found here.
Our impression, based on reviewing a subset of the RCTs in the meta-analyses by Deshpande et al. 2017 and Sawh et al. 2016, is that probiotics generally do not require substantial additional costs because probiotics are either added to current methods of feeding or applied with a dropper. However, we have some uncertainty about this and anticipate we may learn more if we speak with organizations implementing this intervention.
From the protocol for this review (Imdad et al. 2019):
- "We plan to review the following three interventions: neonatal vitamin A supplementation, oral dextrose gel supplementation, and probiotic supplementation during neonatal period in LMICs. Below in this section and rest of the introduction, we describe the rationale for choosing these interventions and why it is important to do this review." P. 1.
- "The effect of probiotic supplementation for prevention of NEC and neonatal sepsis have been assessed in previous reviews (AlFaleh & Anabrees, 2014; Rao et al, 2016; van den Akker et al, 2018). Most of the included studies in these reviews were conducted in developed countries in facility based settings. A recent community based study conducted in India showed that use of synbiotics (probiotic+prebiotics) prevents neonatal sepsis/mortality (Panigrahi et al, 2017). This trial however included neonates with gestational age >35 weeks and birth weight >2000 g. The risk of sepsis might be higher in very preterm and very low birth weight babies; however, these babies might not survive in community settings without advanced care such as provided in a neonatal intensive care unit. Our objective is to include randomized and nonrandomized studies from LMICs to assess the effect of probiotic supplementation on prevention of neonatal morbidity and mortality." P. 4.
- "Participants for this study will include neonates (aged 0–28 days) from LMICs. We will include neonates regardless of their health status: this includes low birth weight and preterm babies." P. 5.