Systematic Review With Meta-Analysis

Lactobacillus Rhamnosus GG for Treating Acute Gastroenteritis in Children

A 2019 Update

Hania Szajewska; Maciej Kołodziej; Dorota Gieruszczak-Białek; Agata Skórka; Marek Ruszczyński; Raanan Shamir

Disclosures

Aliment Pharmacol Ther. 2019;49(11):1376-1384. 

In This Article

Results

For a flow diagram documenting the identification process for eligible trials, see Figure S1. Detailed characteristics of the included RCTs are presented in Table S2, and characteristics of the excluded trials are presented in Table S3. Eighteen RCTs that randomised 4208 participants (2201 in the experimental group and 2007 in the control group) who ranged in age from 1 month to 7 years (majority [12 trials] in children aged 36 months or less) were included[6,11,14–16,18–26,13,27–29] Among them, there were three RCTs that randomised 1295 participants[6,15,16] identified since our previous systematic review. The sample size in all included trials ranged from 36 to 971 participants. Included trials were carried out in countries such as Australia (in Aboriginal children), Brazil, Finland, India, Italy, Pakistan, Peru, Poland, Russia, Uruguay, and USA. One study (Guandalini et al[19]) was a multicentre study; however, it was performed predominantly in European countries. Except for two multi-centre trials,[6,19] the included studies were single-centre trials. There was also clinical heterogeneity among the trials in regard to type of treatment (inpatients and/or outpatients): 10 RCTs were carried out in inpatients; six, in outpatients; and two, in both inpatients and outpatients. The daily doses of L rhamnosus GG ranged from 1.2 x 108 CFU to 2 x 1012 CFU. The comparator treatment was placebo in nine trials and no L rhamnosus GG (oral rehydration solution or no intervention) in nine trials. In all studies, L rhamnosus GG was used in addition to rehydration therapy consisting of an oral rehydration solution and/or intravenous rehydration.

Risk of Bias in Included Studies

Most trials were at risk of bias for at least one of the domains (see Figure S2). Only 72% of trials adequately generated their randomisation sequence, only 56% adequately concealed allocation, and only 61% blinded participants, study personnel, and outcome assessors. There were incomplete outcome data in almost 28% of trials. Overall, only five studies were considered to be at low risk of bias with regard to adequate randomisation, allocation concealment, blinding, and follow-up.[6,11,23,24,28]

Heterogeneity and Publication Bias

Significant heterogeneity (I2 ≥ 50%) was found for both primary outcomes (ie, total stool volume [I 2 = 75%], duration of diarrhoea [I 2 = 99%]). Publication bias was formally only evaluated in the analysis of the duration of diarrhoea. There was no evidence of publication bias for this outcome (Egger test's test P = 0.51; see Figure S3). For other outcomes, publication bias was not formally assessed using a funnel plot due to the small number of studies (<10) included in the analyses.

Effects

A summary of all of the results is presented in Table 1.

Stool volume. Compared with controls, L rhamnosus GG had no effect on the total stool volume (2 RCTs, n = 303) (Figure S4).

Duration of diarrhoea. A meta-analysis of 15 trials (3820 participants) showed a reduction in the duration of diarrhoea for those treated with L rhamnosus GG compared with placebo or no treatment (MD −0.85 days, 95% CI −1.15 to −0.56; high heterogeneity (I2 = 99%); Figure 1). Additionally, data on the duration of diarrhoea were reported in two other trials; however, the data were reported in a format that did not allow pooling of data. Nixon et al[26] reported that among US children, there was no significant difference in the median time until normal stool between the L rhamnosus GG and placebo groups (60 hours [interquartile range: 37-111] vs 74 hours [43-120], respectively, P = 0.37). However, among children who presented with more than 2 days of diarrhoea, the L rhamnosus GG group compared with the placebo group returned to normal stool earlier (51 hours [32-78] vs 74 hours [45-120], respectively, P = 0.02). Moreover, one study (Salazar-Lindo et al[23]) reported the duration of diarrhoea only in children who responded to the treatment within 5 days of admission. Children with ongoing diarrhoea were not further considered. For this reason, the data from this study are not included in our analysis.

Figure 1.

Forest plot of randomised controlled trials of Lactobacillus rhamnosus GG (high dose and low dose) vs control in acute gastroenteritis. Effect on duration of diarrhoea

As intended, a number of pre-planned subgroup analyses were performed.

Dose. L rhamnosus GG was effective when used at a daily dose ≥1010 CFU (11 RCTs, n = 2764, MD −0.83, 95% CI −1.17 to −0.49) and <1010 CFU (4 RCTs, n = 1056, MD −0.92, 95% CI −1.83 to −0.02); however, the latter dose produced results of borderline significance (Figure 1). The test for subgroup differences suggests that there is no significant difference (P = 0.90), meaning that dose does not modify the treatment effect. However, a smaller group of trials and participants contributed to the <1010 CFU subgroup (4 RCTs, n = 1056) than to the ≥1010 CFU subgroup (11 RCTs, n = 2764), indicating that the analysis may not be able to detect subgroup differences.

Setting. There appeared to be a reduction in the duration of diarrhoea in children treated both in Europe (6 RCTs, n = 844, MD −1.22 d, 95% CI −1.73, −0.71) and in non-European settings (9 RCTs, n = 2976; MD −0.61 days, 95% CI −0.98, −0.25); however, the latter effect was smaller (see Figure 2; for a more detailed analysis by region, see Figure S5). The test for subgroup differences suggests that there is a significant difference (P = 0.06), meaning that geographic setting may modify the treatment effect. However, a smaller group of trials and participants contributed to the European subgroup than to the non-European subgroup. There was also heterogeneity between the trials within each subgroup. Overall, whether there is a true significant subgroup effect requires further investigation.

Figure 2.

Forest plot of randomised controlled trials of Lactobacillus rhamnosus GG vs control in acute gastroenteritis. Effect on duration of diarrhoea Europe vs other setting

Inpatients/outpatients. There appeared to be a reduction in the duration of diarrhoea in studies carried out in inpatients (8 RCTs, n = 1868, MD −0.66 days, 95% CI −1.27 to −0.05) and outpatients (4 RCTs, n = 1368, MD −0.95, 95% CI −1.72 to −0.17) (Figure S6). The test for subgroup differences suggests that there is no significant difference (P = 0.47). However, a smaller number of trials and participants contributed to the outpatient group than to the inpatient group, indicating that the analysis may not be able to detect subgroup differences.

Aetiology. With regard to the aetiology, limited data indicate that L rhamnosus GG was effective in treating diarrhoea due to rotavirus (5 RCTs, n = 324; MD −1.38, 95% CI −2.14 to −0.63) or of unknown cause (3 RCTs, n = 253; MD −1.10, 95% CI −1.87 to −0.33), but not due to invasive pathogens (2 RCTs, n = 95, MD 0.33, 95% CI −0.09 to 0.76) (Figure S7). The test for subgroup differences suggests that there is a significant difference between groups (P < 0.0001). However, the number of trials and participants contributing to each subgroup was small, precluding meaningful conclusions.

Rotavirus vaccination status. One RCT evaluated the effectiveness of L rhamnosus GG in a population at least partially vaccinated against rotavirus.[6] However, data for vaccinated and non-vaccinated populations were not presented separately.

A post hoc analysis based on the duration of diarrhoea prior to the randomisation found that L rhamnosus GG was more effective when used in children enrolled with diarrhoea lasting ≤ 5 days (6 RCTs, n = 923, MD −1.4 days, 95% CI −2.32 to −0.47) than when used in children with diarrhoea lasting > 7 days (4 RCTs, n = 1258, MD −0.44 days, 95% CI −0.98 to 0.09). However, it is to be noted that five RCTs did not provide adequate data to be included in the analysis (Figure S8).

Another post hoc analysis of four RCTs (n = 2280) considered to be at low risk of bias with regard to adequate randomisation, allocation concealment, blinding, and follow-up found that compared with controls, L rhamnosus had no effect on the duration of diarrhoea (MD −0.77 days, 95% CI −2.01 to 0.47). High heterogeneity (I2 = 99%) was found. Of note, the findings of the study by Salazar-Lindo et al[23] could not be included (the authors reported the duration of diarrhoea only in children who responded to the treatment within 5 days of admission) (Figure S9).

Pre-planned subgroup analyses based on trial methodological quality were performed. Statistically significant between-study heterogeneity persisted in subgroup analyses, suggesting that the differences in outcomes between studies were caused by factors other than differences in methodological quality (Figures S10-S12).

Duration of Hospitalisation

A meta-analysis of five RCTs (n = 1790) showed a reduction in the duration of hospitalisation for those treated with L rhamnosus GG compared with the control group (MD −1.22 days, 95% CI −2.33 to −0.10; high heterogeneity: I 2 = 99%) (Figure 3).

Figure 3.

Forest plot of randomised controlled trials of Lactobacillus rhamnosus GG vs control in acute gastroenteritis. Effect on duration of hospitalization

Presence of Diarrhoea

Limited data showed that compared with placebo, L rhamnosus GG reduced the risk of diarrhoea on day 2 (1 RCT, n = 40, RR 0.38, 95% CI 0.16-0.87) and diarrhoea > 7 days (1 RCT, n = 287, RR 0.25, 95% CI 0.09-0.75) (Figure S13).

Clinical Severity Score

One RCT (n = 943) found that the modified Vesikari scale score for the 14-day period after enrolment was similar in the L rhamnosus GG and the placebo groups (RR 0.96, 95% CI 0.68-1.35).[6]

Adverse Effects

Data regarding therapy-related adverse events were available from eight studies.[6,11,15,16,23,26,28,29] In these trials, adverse effects rates were similar in the experimental and control groups.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....