Intestinal failure (IF) has been defined as the critical reduction of functional gut mass below the minimal amount necessary for adequate digestion and absorption to satisfy body nutrient and fluid requirements for maintenance in adults or growth in children [1]. Short-bowel syndrome (SBS) is the most common cause of IF in infants; other causes being motility disorders (aganglionosis), chronic intestinal pseudo-obstruction syndrome (myopathic and neuropathic) and congenital diseases of enterocyte development [1]. SBS results from surgical resection, congenital defect or disease-associated loss of absorption capacity of the gut and is characterized by the inability to maintain protein-energy, fluid, electrolyte or micronutrient balances when on a conventionally accepted, normal diet [2]. These patients are therefore dependent on parenteral nutrition (PN). The duration of PN significantly correlates with the length of residual gut [3]. SBS has also been defined as the need for PN greater than 42 days or 2 mo after bowel resection of ≥70% or a residual small bowel length of less than 25% of that expected for gestational age [4,5].

The most common cause of SBS in the neonatal period is necrotizing enterocolitis (NEC), accounting for 35%–50% of cases [6,7]. The other causes include abdominal wall defects (gastroschisis, omphalocele), midgut volvulus, intestinal atresia, meconium ileus, Hirschsprung’s disease and superior mesenteric artery abnormalities [6,8]. The contribution of NEC to SBS appears to be decreasing in some centres, due to advances in perinatal care and antenatal steroids, resulting in the decreased incidence of NEC [5,9]. Neonatal research network hospitals in the US have reported an incidence of 7/1000 in very low birth weight (VLBW) infants and 11/1000 in extremely low birth weight (ELBW) infants [10]. Similar to NEC, birth weight and gestational age were inversely related to the incidence of SBS. NEC was responsible for 96% of SBS cases. In a Canadian study, the incidence was estimated to be 22.1 per 1000 NICU admissions at a tertiary centre, whereas population-based incidence was 24.5 per 100,000 live births; only three out of 40 SBS infants were of term gestation [11]. An Italian study reported an incidence of 5/1000 NICU admissions and 1/1000 live births [9]. Approximately 80% of SBS in the paediatric population occurs in the neonatal period [8].

The health burden of SBS is significant. A case fatality rate of 27.5%–37.5% has been reported within 1.5–5 year follow-up periods in four retrospective studies, and hepatic failure accounted for 60% and sepsis for 10%–20% of deaths [5,6,7,12]. Incidence of sepsis is high and is the most common cause for readmission in patients with SBS, increasing the length of hospitalization and the cost of care [5,10,13]. Growth deficits (weight, length and head circumference) were prevalent in 74% of VLBW infants with SBS at 18–22 month age [10]. Failure to thrive (body weight < fifth percentile) was seen in 76.5% of patients at 6 mo and in 47.6% at 2.5 year in a retrospective study [7]. SBS imposes disproportionately high healthcare costs on tax payers. In the United States, the mean total cost of care per child over a five-year period after onset of SBS was estimated to be over 1.62 million (range 1.3–2 million) USD, of which hospitalization accounted for the maximum cost [5]. Shorter residual bowel length could incur higher costs. PN dependence ranged from 2.4 months to 12.6 years, with a median of 1.5 years. In the Netherlands, the average total cost was 355,000 USD, with a maximum of 600,000 USD [14].

Our search revealed no RCTs/Q-RCTs on the use of probiotics in children with SBS. However, we found one small cross-over RCT, one case control study and nine case reports on the use of probiotics in children with SBS. The nine case reports included five reporting beneficial effects (Table 4) and four reporting adverse affects of probiotics.

Crossover RCT: Sentongo et al. [77] used the design of a double-blind, placebo-controlled randomised crossover clinical trial to assess the effects of Lactobacillus rhamnosus (LGG) treatment on intestinal permeability (IP) in children with SBS. Baseline IP was measured by the urinary lactose-mannitol ratio in nine children with SBS (cases) and 12 healthy children (controls). The median (range) age of the 21 enrolled children was 4.5 (1.6–16.4) years. SBS patients received LGG or placebo for four weeks, followed by a three-week washout period before treatment was crossed over for another four weeks. IP, quantitative faecal cultures for Lactobacillus species and the hydrogen breath test (HBT) were performed during LGG and placebo phases of treatment. Baseline IP (mean ± SD) was comparable in cases with SBS and healthy control: 0.08 ± 0.06 versus 0.07 ± 0.05 (p = 1.0). IP was significantly correlated with age in controls (r = −0.83, p = 0.001), but not in children with SBS (r = −0.55, p = 0.16). Faecal colonization [median (range)] with Lactobacillus species did not differ during LGG versus placebo treatment (1.4 × 10⁹ (4.0 × 10⁵ to 4.0 × 10⁹) cfu/g) versus (6.0 × 10⁹ (1.0 × 10³ to 1.0 × 10¹⁰) CFU/g), respectively; (p = 0.83). LGG treatment had no consistent effects on IP (p = 0.58) or its relationship with age (r = −0.40, p = 0.29) and was associated with conversion to positive HBT results in one subject. Considering the inconsistent effects of probiotic treatment, it was concluded that empiric LGG therapy to enhance IP in children with SBS was not justified.

Case-control study: Uchida et al. [71] have evaluated the changes in immunonutritional parameters before and after treatment with a synbiotic (Bifidobacterium breve, Lactobacillus casei, galactooligosaccharides) in four children with SBS (cases) and four normal, healthy, age-matched children enrolled as controls. Faecal samples were analysed for bacterial flora and organic acid (OA) contents. Levels of short chain fatty acids (SCFA), such as butyrate, propionate and acetate, increased in one patient, and SCFA/total OA levels increased in three patients. Serum lymphocyte counts and pre-albumin levels increased after commencing synbiotic treatment, reaching a statistically significant level at the ninth month compared to the baseline level. There was an increasing trend in height and weight gain velocity during the study versus the baseline period. The faecal bacterial flora improved in SBS patients after synbiotic therapy.

The results of our systematic review indicate that there is a paucity of clinical studies on efficacy of probiotics in SBS. We found no RCTs or Q-RCTs in this field. The literature search identified a few animal studies that reported consistent benefits of probiotics in decreasing bacterial translocation and augmenting histological features of intestinal adaptation in SBS. The crossover RCT reported only a non-clinical parameter, such as IP [77]. The baseline IP of subjects with SBS in this trial was comparable to that of controls, most likely due to exclusion of sick and clinically unstable patients, making the applicability of results to SBS patients with complications, such as SBBO, enterocolitis or intestinal failure, difficult. SBBO is known to predispose to intestinal inflammation, and various disorders involving intestinal inflammation are known to have increased IP. Increased IP has also been reported to be associated with a recent episode of sepsis and severe liver disease in patients with SBS, suggesting increased IP may have a role in predisposition to sepsis in SBS [20]. The very small sample size (n = 9) and lack of assessment of clinically important outcomes makes it difficult to agree with the author’s conclusion. Whether the three-week wash-out period was adequate to minimise/avoid carry over effects is also not clear. Moreover, it is also important to note that the effects of probiotics are strain-specific, and benefits by probiotic strains other than LGG cannot be ruled out. Analysis of crossover RCTs using paired data from the first and second period of the treatment is a complex issue [69]. Considering the small sample size and the fact that no such data was available, we did not carry a post-hoc analysis of this trial.

The positive impact of probiotic supplementation on growth (increased weight and growth velocity) and nutrition (increased levels of serum proteins) has been demonstrated in the case-control study by Uchida et al. [71] and in the case reports (Kanamori et al., Shiau et al.) [73,74,75]. It is also noteworthy that probiotics have been found to be effective in treating SBBO, enterocolitis and D-Lactic acidosis where conventional treatment modalities have failed (Vanderhoof et al., Kanamori et al.) [72,73,74]. Suppression of pathogenic bacteria/facultative anaerobes and normalization of intestinal flora with increased numbers of Bifidobacteria, Lactobacilli and other anaerobes has been associated with the resolution of these complications (Kanamori et al.) [73,74]. The clinically important outcome of treatment with probiotics is improved tolerance of enteral feeds and weaning from TPN, which has been demonstrated in the case reports by Vanderhoof et al., Kanamori et al. and Shiau et al. [72,73,74,75]. Improvement in gut motility (Kanamori et al.) [73] and intestinal absorption (Kanamori et al., Candy et al.) [74,76] could be the underlying mechanisms for tolerance of feeds and improvement in symptoms of malabsorption, such as abdominal distension and diarrhoea. A variety of probiotics were used in these studies.

Our literature search indicates that the commonly reported complications of probiotic treatment in SBS are probiotic sepsis and D-Lactic acidosis. Lactobacillus species were implicated in both these complications. Lactobacillus sepsis has also been reported in two debilitated children—one, post-cardiac surgery and the other, with cerebral palsy and gastrojejunostomy, receiving probiotic treatment for antibiotic-associated diarrhoea [82]. The tendency of Lactobacilli to cause sepsis has been suggested to be related to their increased adherence ability, which is one of the mechanisms of their action [83]. They are also the more commonly used species for probiotic supplementation. One case of Bifidobacterium sepsis has been reported in a neonate operated on for omphalocele, which was treated with ampicillin [84]. It is apparent that patients who are most likely to benefit from probiotics are also the ones susceptible to probiotic sepsis. Overall, the incidence of sepsis related to the common probiotic bacteria (Lactobacilli and Bifidobacteria) has been found to be minimal and is similar to that caused by commensal strains of these bacteria [85]. However, increased surveillance is warranted in patients with debilitating underlying conditions associated with impaired immunity and gut mucosal integrity who are receiving probiotics. This should include investigations for suspected sepsis episodes with anaerobic cultures and molecular identification of the organism if Lactobacilli or Bifidobacteria are isolated to confirm their identity with probiotic strain, as well as using appropriate empirical antibiotics covering the probiotic strain. It is prudent to select probiotic strains that do not produce D-lactic for therapy in patients with SBS. Bifidobacterium breve and Lactobacillus casei are theoretically non-D-lactate producing probiotics and, hence, may be suitable for use in neonates and children with SBS. Bifidobacterium breve and Lactobacillus casei have been used along with antibiotics to treat a SBS patient with D-Lactic acid encephalopathy, which is believed to have replaced D-lactate producing Lactobacilli in the gut with L-lactate producing non-pathogenic flora (Uchida et al.) [86]. The adverse effects of probiotics have been reported mostly in immunocompromised and debilitated patients and not in healthy individuals [87]. Infants with SBS have compromised gut barrier function and immunity. Hence, extrapolation of results of safety studies in healthy infants, fed with formula containing probiotics [88] to SBS patients is not possible.

In summary, our results indicate that there is a paucity of clinical studies of probiotic supplementation in children with SBS. However, the evidence from animal studies and clinical case reports indicate that probiotics do have a potential for benefit in this population of patients and need evaluation in large RCTs. The safety and efficacy of probiotics in SBS can only be answered by multicentre trials, considering the low incidence of this condition. Killed or inactivated probiotic strains, which intuitively cannot have adverse effects, such as probiotic sepsis, but can exert beneficial effects, like live probiotics [89], should also be evaluated in RCTs in SBS patients.