Racecadotril for acute diarrhoea in children.

BACKGROUND: Acute diarrhoea is a leading cause of death for children under five years of age. Most deaths are caused by excessive fluid and electrolyte losses. Racecadotril is an anti-secretory drug that has been used for acute diarrhoea in children as an adjunct to oral rehydration therapy.
OBJECTIVES: To assess the efficacy and safety of racecadotril for treating acute diarrhoea in children under five years of age. SEARCH
METHODS: We searched the Cochrane Infectious Diseases Group Specialized Register; the Cochrane Central Register of Controlled Trials (CENTRAL, published in the Cochrane Library Issue 3, March 2019); MEDLINE; Embase; LILACS; ClinicalTrials.gov; and the World Health Organization International Clinical Trials Registry Platform (WHO ICTRP), up to 4 March 2019, for clinical trials regardless of publication language or status. SELECTION CRITERIA: Randomized controlled trials (RCTs) that compared racecadotril to placebo or no intervention in addition to standard care (oral rehydration therapy) in children under five with acute diarrhoea. The primary outcomes were failure of oral rehydration, duration of diarrhoea, and number of stools. The secondary outcomes were stool output, length of the hospital stay, and adverse events. DATA COLLECTION AND ANALYSIS: Two review authors independently assessed trial eligibility, extracted the data and assessed risk of bias. We presented dichotomous data with risk ratios (RR) and continuous data with mean difference (MD) or standardized mean difference (SMD). Where appropriate, we combined trials with meta-analysis and used a random-effects model if there was significant heterogeneity (I² ≥ 50%). We assessed the certainty of the evidence using the GRADE approach. MAIN
RESULTS: Seven RCTs with a total of 1140 participants met the inclusion criteria. The trials were carried out on children aged three months to five years, in outpatient and inpatient facilities from France, Spain, Peru, India, Kenya, and Ecuador. The efficacy and safety of racecadotril were compared to placebo or no treatment. Racecadotril may reduce the risk of rehydration failure (RR 0.41, 95% CI 0.13 to 1.23; 2 RCTs, 192 participants; low-certainty evidence). Data on duration of diarrhoea, number of stools in the first 48 hours are insufficient to reach a conclusion; stool output in the first 48 hours appears to be lower in the two trials measuring this, although the data is not combinable. Length of hospital stay was similar in two studies measuring this, and overall there was no evidence that racecadotril increased overall rate of adverse events (RR 0.90, 95% CI 0.66 to 1.22; 5 RCTs, 688 participants; low-certainty evidence). Most adverse events in the racecadotril group were mild or moderate. AUTHORS'
CONCLUSIONS: Racecadotril seems to be a safe drug but has little benefit in improving acute diarrhoea in children under five years of age. Current evidence does not support routine use of racecadotril in management of acute diarrhoea in children under five outside of the context of placebo controlled RCTs. 18 December 2019 Up to date All studies incorporated from most recent search All studies identified during the most recent search (4 Mar, 2019) have been incorporated in the review, and no ongoing studies identified.

Background

Description of the condition

Diarrhoea is defined as the passage of three or more loose or liquid stools per day (WHO 2017). Acute diarrhoea lasts usually several hours or days, while persistent diarrhoea has a duration of 14 days or longer. Acute diarrhoea remains a leading cause of death for children under five years of age. Each year, an estimated 1.7 billion cases of acute diarrhoea occur worldwide among children, leading to around 525,000 deaths (WHO 2017). Acute diarrhoea is a common symptom of gastrointestinal infections mostly caused by a wide range of pathogens, including bacteria, viruses, and protozoa. However, just a handful of organisms are responsible for most acute cases of childhood diarrhoea, with rotavirus being the leading cause of acute diarrhoea among children under five (WHO 1999; WHO 2008). Most deaths from acute diarrhoea are caused by excessive fluid and electrolyte losses (Tormo 2008), and most are therefore avoidable provided losses are replaced promptly by rehydration. Oral rehydration therapy has been recommended as the essential treatment for acute diarrhoea in children (NICE 2009; UNICEF/WHO 2004). A network meta‐analysis reported that 27 different types of treatments or treatment combinations for children with acute diarrhoea were studied in clinical trials; and it concluded that zinc, Saccharomyces boulardii combined with zinc, and smectite combined with zinc showed the best combination of evidence quality and magnitude of effect, especially in low‐ and middle‐income countries (Florez 2018).

Description of the intervention

Racecadotril is an anti‐secretory drug that exerts its anti‐diarrhoeal effects by rapidly hydrolyzing to the active metabolite thiorphan, which is a powerful and selective inhibitor of the enzyme enkephalinase (Schwartz 2000). It prevents the degradation of endogenous enkephalins and thus reduces hypersecretion of water and electrolytes into the intestinal lumen without interfering with motility (Primi 1999). Unlike loperamide, racecadotril does not act on µ‐opioid receptors and therefore does not prolong intestinal transit, nor does it encourage bacterial growth in the small intestine (Maldonado 2006). Also, unlike the antisecretory drug bismuth salicylate, racecadotril is not associated with a risk of developing Reye's syndrome (Abramson 1993).

The World Health Organization (WHO) has recommended oral rehydration solution and zinc for the routine treatment of acute diarrhoea in children (WHO 2005). Oral rehydration therapy does not, however, provide relief of diarrhoeal symptoms, which can be distressing for both children and their caregivers (Szajewska 2007). Racecadotril has been used for acute diarrhoea in children as an adjunct to oral rehydration therapy, especially in Europe, South America and several Asian countries (Bioprojet Pharma 2007). It might help improve the success of rehydration by reducing fluid loss and provide relief on the diarrhoeal symptoms by reducing stool frequency or shortening the duration of diarrhoea, or both. A meta‐analysis of randomized controlled studies in adults showed that racecadotril can shorten diarrhoea duration and reduce number of stools, compared to placebo (Coffin 2014). According to published studies, it also has good tolerability both in children and adults, with adverse effects similar to placebo groups (Alam 2003; Cézard 2001). A cost‐utility analysis from the perspective of a UK payer found racecadotril as adjuvant therapy was more effective and less costly compared to oral rehydration therapy alone for children's diarrhoea (Rautenberg 2012).

Why it is important to do this review

Although racecadotril has been used in many countries, WHO and National Institute for Health and Clinical Excellence (NICE) guidelines have not recommended use of any antidiarrhoeal drugs in children under five (NICE 2009; WHO 2017). The ESPGHAN evidence‐based guideline, Guarino 2014, mentioned that racecadotril may be considered in the management of acute diarrhoea in children but well‐designed prospective studies should be carried out in outpatient children, since outpatients account for the majority of rotavirus‐associated cases of diarrhoea in children (Parashar 2003). A systematic review covering a small number of earlier published trials found that racecadotril was effective in reducing weight and frequency of stool output as well as shortening the duration of diarrhoea in children (Szajewska 2007). However, stool output weight, as the primary outcome, has limited practical value and tends to be influenced by factors such as food intake and to be contaminated by urine. Two systematic reviews or meta‐analyses concluded that racecadotril was effective in children with acute diarrhoea; risk of bias was not assessed for included trials in those reviews, however (Eberlin 2018; Lehert 2011). In addition, one of the main purposes of using racecadotril is to prompt correction of water‐electrolyte disturbances, and thus outcomes for evaluating the success of rehydration should also be included. Previous reviews did not include this outcome when evaluating the efficacy of racecadotril (Gordon 2016; Lehert 2011; Szajewska 2007). Although several systematic reviews have been conducted in order to assess the efficacy of racecadotril for children with acute diarrhoea, the benefit of using racecadotril is still controversial as newly published trials have inconsistent conclusions on its efficacy (García 2016; Gharial 2017; Kang 2016a; Kang 2016b; Sreenivas 2017). An updated systematic review that carefully assesses risk of bias of available trials and addresses outcomes both on symptomatic relief and success of rehydration was needed.

Objectives

To assess the efficacy and safety of racecadotril for treating acute diarrhoea in children under five years of age.

Methods

Criteria for considering studies for this review

Types of studies

Randomized controlled trials (RCTs).

Types of participants

Children less than five years old with uncomplicated acute diarrhoea, defined as passing loose or watery stools at least three times in a 24‐hour period and lasting less than seven days.

Types of interventions

Intervention

Orally administered racecadotril irrespective of dose, duration, and interval of administration.

Control

Placebo, or no treatment.

Oral rehydration therapy as the standard care should be given to both intervention and control groups. We also included trials in which both groups received zinc as an adjunct to oral rehydration therapy, as it is recommended by the WHO. (UNICEF/WHO 2004)

Types of outcome measures

Primary outcomes

Failure of oral rehydration, defined as any of the following: worsening or no improvement of dehydration; recurring dehydration; need for unscheduled intravenous fluid infusion; participants initially without dehydration subsequently developing dehydration; or need for hospital admission in outpatients.

Duration of diarrhoea, defined as the time period in days from randomization or start of treatment to the resolution of diarrhoea.

Number of stools, defined as the number of stools or bowel movements within 48 hours after administration of interventions.

Secondary outcomes

Stool output: the accumulated weight of stool output within 48 hours, measured by using diapers or others ways.

Length of hospital stay for inpatients: the time period in days from randomization or the start of treatment to hospital discharge for hospitalized patients.

Adverse events: number of all types of adverse events.

Search methods for identification of studies

We attempted to find all relevant trials regardless of language or publication status (published, unpublished, in press, and in progress).

Electronic searches

We searched the following databases: the Cochrane Infectious Diseases Group Specialized Register; the Cochrane Central Register of Controlled Trials (CENTRAL, published in the Cochrane Library issue 3, March 2019); MEDLINE (via OVID, 1966 to 4 March 2019); Embase (via OVID, 1980 to 4 March 2019); and LILACS (1982 to 4 March 2019), using the search terms detailed in Appendix 1. We also searched Clinicaltrials.gov and the search portal of the World Health Organization International Clinical Trials Registry Platform (WHO ICTRP), to identify ongoing or unpublished trials, using “diarrh*”, “racecadotril” and “child*” as search terms. All the search work was up to 4 March 2019.

Searching other resources

We reviewed the citations of existing reviews and clinical trials obtained by the above methods. We also contacted individual researchers and pharmaceutical companies for data of unpublished trials.

Data collection and analysis

Selection of studies

Two review authors (YL and LZ) independently screened the titles and abstracts of all studies identified by the search strategy and obtained the full articles for all potentially relevant trials. We independently assessed the full text using an eligibility form based on the inclusion criteria. We resolved disagreements by discussion until consensus was achieved. We scrutinized each of the articles to ensure that each trial was included only once. We documented the reasons for excluding trials in the Characteristics of excluded studies table. We constructed a PRISMA diagram to illustrate the study selection process.

Data extraction and management

Two review authors (YL and LZ) independently extracted the data using a pre‐designed data collection form. The extracted items included: 1) authors; 2) year of publication; 3) the country where the trial was conducted; 4) number of participants randomized into each group; 5) inclusion and exclusion criteria; 6) interventions in both groups; 7) dose and duration of racecadotril; 8) outcomes; 9) source of funding. For each treatment arm, we extracted the number of participants randomized to each group and the number of analysed participants for each outcome. For dichotomous outcomes, we extracted the number of participants experiencing the event in each group of the trial; for continuous data, we extracted means and standard deviations (median and interquartile for skewed data). Studies that were published in languages other than English were translated with the help of Cochrane Infectious Diseases Group.

Assessment of risk of bias in included studies

Two review authors (YL and LZ) independently assessed the risk of bias using the Cochrane ‘Risk of bias' tool, which addresses six specific domains: sequence generation, allocation concealment, blinding, incomplete outcome data, selective outcome reporting, and other bias (Higgins 2011). We categorized the judgements as low, high or unclear risk of bias, and documented reasons for each judgement. We resolved disagreements by discussion or consulting methodologists from the Chinese Cochrane Center. We created plots of risk of bias assessment in Review Manager 5 (RevMan 5) (Review Manager 5). We contacted authors of included trials for additional information if needed for risk of bias assessment.

Measures of treatment effect

We used risk ratio for dichotomous data and mean difference (standardized mean difference if the outcome was reported on different scales) for continuous data, presenting with 95% confidence intervals.

Dealing with missing data

We attempted to contact the trial authors when data were insufficient or missing. We noted proportions of participants for whom we could obtain no outcome data, and reported them in a ‘Risk of bias’ table. We dealt with missing data at individual study and outcome level. We checked the number of children lost to follow‐up and the reasons for missing data in each group. If, in accordance with the guidance in Guyatt 2011b, the proportion was less than 20% and meanwhile the reasons and numbers of missing were balanced in the two groups, we graded the risk of attrition bias as low; otherwise, we graded it as high. We performed sensitivity analyses based on participants originally assigned to two groups or reported results from intention‐to‐treat (ITT) analyses in the original clinical trials if no meta‐analysis was conducted.

Assessment of heterogeneity

We judged clinical heterogeneity based on our knowledge at patient, intervention, control and outcome level. We used the I² statistic to check statistical heterogeneity, considering an I² value of 50% or greater to denote significant heterogeneity. We used a fixed‐effect model to synthesize data when heterogeneity was not significant (I² < 50%) and a random‐effects model when heterogeneity was significant (I² ≥ 50%).

Assessment of reporting biases

We planned to examine publication bias using a funnel plot, and to use a regression approach to assess funnel plot asymmetry if more than nine trials reported the same outcome (Egger 1997). However the number of included trials failed to meet the requirements, so we did not conduct analysis of reporting bias.

Data synthesis

We conducted meta‐analyses using Review Manager 5 (RevMan 5), and present the results with 95% confidence intervals (CI) (Deeks 2011). For failure of rehydration and adverse effects, we calculated the pooled estimate of the risk ratio (RR) between two groups. For duration of diarrhoea, number of stools, stool output, and length of hospital stay, we planned to pool the results by calculating the mean difference or standardized mean difference. If included trials reported results using median instead of means, this might indicate that the data were skewed so we did not combine those results in meta‐analysis (Deeks 2011). We used either a fixed‐effect or random‐effects model according to the heterogeneity in outcomes across trials assessed by the I² statistic test.

Subgroup analysis and investigation of heterogeneity

We expected to conduct subgroup analysis on inpatient versus outpatient groups, different disease severity groups, rotavirus versus non‐rotavirus groups and different dosage of racecadotril.

Sensitivity analysis

We conducted sensitivity analyses by limiting the analyses to those with lower overall risk of bias, especially with adequate randomization, allocation concealment, and blinding. We also performed sensitivity analyses on binary outcomes based on the number of participants originally randomized into two groups or reported results of ITT analyses from original clinical trials for both continuous and dichotomous outcomes for sensitivity purposes.

Assessment of the certainty of the evidence

We assessed the certainty of the evidence using the GRADE approach and used the GRADEpro Guideline Development Tool (GDT) to create the ‘Summary of findings' tables (GRADEpro GDT). The certainty rating of the studies has four levels: high, moderate, low, and very low. RCTs are initially categorized as high certainty but can be downgraded after assessment of five aspects: risk of bias, consistency, directness, imprecision, and publication bias (Guyatt 2011a). In the GRADE system, high‐certainty evidence denotes that further research is unlikely to change our confidence in the estimate of effect; moderate‐certainty evidence denotes that further research is likely to have an important impact on confidence in the estimate of effect and may change the estimate; low‐certainty evidence denotes that further research is very likely to have an important impact on confidence in the estimate of effect and is likely to change the estimate; very low certainty evidence denotes a serious lack of confidence in the effect estimate (Schünemann 2011). For outcomes for which we could not obtain a single pooled estimate, we assessed the certainty of evidence based on all trials that reported the same outcome and presented estimates with narrative description in the ‘Summary of findings' table (Murad 2017).

Results

Description of studies

Results of the search

We identified 213 citations by initial searching of databases and 9 records from clinical trial registry platforms. During study selection we encountered three non‐English articles and obtained help from the Cochrane Infectious Diseases Group (CIDG) with translating them into English (Calatayud 2009; Cojocaru 2002; García 2016). After selecting according to inclusion criteria, we included seven RCTs, involving 1140 participants (Cézard 2001; García 2016; Gharial 2017; [Kang 2016a; Kang 2016b]; Salazar‐Lindo 2000; Santos 2009; Sreenivas 2017). See Figure 1 for a flow diagram of the selection process.

1

Study flow diagram.

Included studies

See: Characteristics of included studies

Setting

Two trials were conducted in India ([Kang 2016a; Kang 2016b]; Sreenivas 2017); the rest were conducted in France (Cézard 2001), Spain (Santos 2009), Peru (Salazar‐Lindo 2000), Kenya (Gharial 2017), and Ecuador (García 2016).

Participants

Participants were aged from three months to 5 years. All trials excluded patients who received antibiotics or other anti‐diarrhoeal drugs during this episode of diarrhoea. Two did not exclude participants with comorbid diseases or conditions like HIV and severe malnutrition (Gharial 2017; Santos 2009). Salazar‐Lindo 2000 studied only boys to minimize the contamination of stool by urine in assessment of stool output. Gharial 2017 and [Kang 2016a; Kang 2016b] graded the severity of diarrhoea with the Vesikari score (Lewis 2011); and Gharial 2017 included participants with a score greater than 11 (severe) only. Two trials included outpatients only (García 2016; Santos 2009); while four included inpatients only (Cézard 2001; Gharial 2017; Salazar‐Lindo 2000; Sreenivas 2017). One trial included both inpatients and outpatients but reported their results separately (Kang 2016a; Kang 2016b). Sreenivas 2017 included paediatric ward inpatients but children with severe diarrhoea were excluded.

Interventions

In Santos 2009 and García 2016, 10 mg of racecadotril was given every 8 hours for children weighing less than 9 kg, 20 mg every 8 hours for children weighing 9 kg to 13 kg, and 30 mg every 8 hours for those weighing more than 13 kg. In Gharial 2017, racecadotril was administered at 10 mg per dose for children below 12 months of age and 30 mg for those over 12 months of age. In the other four trials, racecadotril was given three times daily at a dose of 1.5 mg/kg (Cézard 2001; [Kang 2016a; Kang 2016b]; Salazar‐Lindo 2000; Sreenivas 2017).

Comparison

Five trials used placebo as control compared to racecadotril (Cézard 2001; García 2016; Gharial 2017; [Kang 2016a; Kang 2016b]; Salazar‐Lindo 2000), while in the other two, racecadotril was compared to no treatment in addition to standard therapy (Santos 2009; Sreenivas 2017).

Outcomes

Two trials reported the primary outcome ‘failure of rehydration' (García 2016; Sreenivas 2017). All except one —García 2016— reported the outcome ‘duration of diarrhoea'. Two trials reported the number of outpatients' stools during the first and second day after treatment (García 2016; Santos 2009); and two reported the number of stools within 48 hours in inpatients (Gharial 2017; Sreenivas 2017). Three trials reported the weight of stool output (Cézard 2001; Kang 2016a; Salazar‐Lindo 2000). Two reported the outcome ‘length of hospital stay' (Gharial 2017; Kang 2016a). All except one reported adverse events (Kang 2016a; Kang 2016b).

Funding

Three trials were supported by the pharmaceutical industry (Gharial 2017; Salazar‐Lindo 2000; Santos 2009). García 2016 did not clearly report source of funding. One failed to report the source of funding but several authors are from pharmaceutical industry (Cézard 2001). Kang 2016a and Kang 2016b received funding from a governmental organization (Swedish International Development Agency) and Sreenivas 2017 did not receive any funding for conducting the trial.

Excluded studies

See: Characteristics of excluded studies

Risk of bias in included studies

See Figure 2 and Figure 3 for a summary of risk of bias assessment. We emailed the authors of included trials for further information regarding risk of bias but only one replied (Gharial 2017). We were not able to contact authors from two trials as the email addresses listed in the articles are no longer valid (Cézard 2001; Salazar‐Lindo 2000).

2

'Risk of bias' graph: review authors' judgements about each 'Risk of bias' item presented as percentages across all included trials.

3

'Risk of bias' summary: review authors' judgements about each 'Risk of bias' item for each included trial.

Allocation

Five trials were at high risk of selection bias: two failed to report how the random sequence was created (Cézard 2001; Salazar‐Lindo 2000); and five did not report if allocation concealment was conducted (Cézard 2001; García 2016; Salazar‐Lindo 2000; Santos 2009; Sreenivas 2017). Two trials reported that allocation concealment was achieved by sealed envelopes with randomization codes given directly to the hospital pharmacy ([Kang 2016a; Kang 2016b]; Gharial 2017).

Blinding

Two trials were at high risk of performance bias and detection bias for no blinding (Santos 2009; Sreenivas 2017). Five were reported as double‐blinded placebo‐controlled trials (Cézard 2001; García 2016; Gharial 2017; [Kang 2016a; Kang 2016b]; Salazar‐Lindo 2000).

Incomplete outcome data

Three trials were at high risk of attrition bias, two due to high loss to follow‐up (> 20%) (Cézard 2001; Santos 2009) and one due to imbalance in loss to follow‐up (García 2016). Four trials were at low risk of attrition bias (Gharial 2017; [Kang 2016a; Kang 2016b]; Salazar‐Lindo 2000; Sreenivas 2017).

Selective reporting

Five trials had unclear risk of reporting bias as it was not reported if their research protocols were registered or accessible (Cézard 2001; García 2016; Salazar‐Lindo 2000; Santos 2009; Sreenivas 2017). It was unknown whether all the pre‐designed outcomes in those trials had been reported. Two trials have research protocols registered on clinical trial registration platforms and reported all pre‐designed outcomes (Gharial 2017; [Kang 2016a; Kang 2016b]).

Other potential sources of bias

Three trials were supported by the pharmaceutical industry (Gharial 2017; Salazar‐Lindo 2000; Santos 2009), and one failed to report the role of the pharmaceutical company in the trial, leading to high risk of bias for potential conflict of interest (Salazar‐Lindo 2000). Santos 2009 clearly reported that the pharmaceutical company was not involved in data collection, analysis and result interpretation; Gharial 2017 received part of its funding from the pharmaceutical industry but, as the conclusion did not favour the company, there was low risk of bias caused by conflict of interest. One trial failed to report source of funding, but some authors were from the pharmaceutical industry, causing potentially high risk of bias (Cézard 2001).

Effects of interventions

See: Table 1

Summary of findings for the main comparison

Summary of findings table 1

Racecadotril compared to placebo/no treatment for acute diarrhoea in children
Patient or population: acute diarrhoea in children Setting: hospital and outpatients Intervention: racecadotril Comparison: placebo/no treatment
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	Number of participants (trials)	Certainty of the evidence (GRADE)	Comments
	Risk with placebo/no treatment	Risk with racecadotril
Failure of rehydration	106 per 1000	44 per 1000 (14 to 131)	RR 0.41 (0.13 to 1.23)	192 (2 RCTs)	⊕⊕⊝⊝ LOWa,bDue to risk of bias and imprecision	Racecadotril may reduce the risk of failure of rehydration
Duration of diarrhoea	Not pooled: mean and median were inconsistently reported	2 trials concluded that racecadotril might shorten duration of diarrhoea while 3 with lower overall risk of bias reported no difference between two groups on this outcome	_	867(5 RCTs)	⊕⊝⊝⊝ VERY LOWc,dDue to risk of bias and inconsistency	We are uncertain whether or not racecadotril shortens duration of diarrhoea.
Number of stools: measured in the first 24 hours after introduction of racecadotril	4.8 stools/person	In the racecadrotril group, patients on average had 0.24 stools less than the control group (95% CI ‐0.90 to 0.43)		258(2 RCTs)	⊕⊝⊝⊝ VERY LOWe,fDue to risk of bias and inconsistency	We are uncertain whether racecadotril influences the number of stools
Stool output:measured in the first 48 hours after introduction of racecadotril	The mean stool output was 331 g/kg	The mean stool output in the racecadotril group was 25.35 g/kg lower (34.32 lower to 15.99 lower)	_	301(2 RCTs)	⊕⊕⊝⊝ LOW g,hDue to risk of bias and imprecision	Racecadotril may reduce stool output; it was estimated using a standardised mean difference of ‐0.65 (‐0.88 to ‐0.41)
Length of hospital stay for inpatients	Not pooled: median was reported. The median length of hospital stay 3 days in 1 trial and 5 days in another.	The median length of hospital stay is similar compared to the control group	‐	224(2 RCTs)	⊕⊕⊝⊝ LOWiDue to imprecision	Racecadotril may make little or no difference to length of hospital stay for inpatients
Adverse events:number of all types of adverse events	184 per 1000	166 per 1000 (121 to 224)	RR 0.90 (0.66 to 1.22)	688 (5 RCTs)	⊕⊕⊝⊝LOW j,kDue to risk of bias and imprecision	Racecadotril may make little or no difference to rates of adverse events

aDowngraded by 1 for serious risk of bias: both trials have unclear allocation concealment; one study has > 20% loss to follow‐up.
 bDowngraded by 1 for serious imprecision: a wide 95% CI that includes appreciable benefit (RR < 0.75).
 cDowngraded by 1 for serous risk of bias: 1 trial has unclear method of random sequence generation; 3 trials have unclear allocation concealment; 1 trial has > 20% loss to follow‐up; 2 trials have no blinding.
 dDowngraded by 2 for very serous inconsistency: inconsistent conclusions on the effect of the intervention (2 trials concluded that racecadotril might shorten duration of diarrhoea while 3 with lower overall risk of bias reported no difference between 2 groups on this outcome).
 eThe certainty of evidence was graded based on 4 studies reporting this outcome. Downgraded by 1 for serious risk of bias: 3 trials have unclear allocation concealment; 2 trials have no blinding; 2 trials have > 20% loss to follow‐up.
 fDowngraded by 2 for very serous inconsistency: inconsistent conclusions on the effect of the intervention (3 trials reported no statistically difference between 2 groups in number of stools within 48 hours while 1 trial reported racecadotril might be effective for this outcome).
 gDowngraded by 1 for serious risk of bias: both trials have unclear method of random sequence generation and unclear allocation concealment; 1 trial has > 20% loss to follow‐up.
 hDowngraded by 1 for serious imprecision; the total sample size was small (less than 400).
 iDowngraded by 2 for serious imprecision; the total sample size was small (less than 400) and no variance is given.
 jDowngraded by 1 for serious risk of bias: 2 trials have unclear method of random sequence generation; 4 trials have unclear allocation concealment; 1 trial has no blinding; 3 trials have > 20% loss to follow‐up.
 kDowngraded by 1 for serious imprecision: a wide 95% CI that includes appreciable benefit (RR < 0.75).

Failure of rehydration

Two trials reported outcomes that can be categorized into failure of rehydration (García 2016; Sreenivas 2017). Participants in these two trials all have mild to moderate diarrhoea, even though Sreenivas 2017 was conducted in an inpatient setting and García 2016 was conducted in outpatients from the emergency room. One (113 participants) reported that three children from the racecadotril group and seven from the placebo group received intravenous fluid therapy due to worsening of dehydration (Sreenivas 2017). The other trial (79 participants) found that one child from the racecadotril group and seven from the placebo group were hospitalized after treatment in the emergency room (García 2016). The pooled analysis showed that there was no significant difference between racecadotril and control group on the rate of rehydration failure (RR 0.41, 95% CI 0.13 to 1.23, I² = 0%; 2 trials, 192 participants; Analysis 1.1; Figure 4). The sensitivity analysis based on the number of participants originally randomized in two groups does not change the significance of the result.

1.1

Analysis

Comparison 1 Racecadotril versus placebo/no treatment, Outcome 1 Failure of rehydration.

4

Comparison 1 Racecadotril versus placebo/no treatment, Outcome 1.1 Failure of rehydration.

Duration of diarrhoea

Five trials (867 participants) reported the average duration of diarrhoea in two groups. Three trials included only inpatients (Gharial 2017; Salazar‐Lindo 2000; Sreenivas 2017); one included only outpatients (Santos 2009); and one was conducted on both inpatients and outpatients (Kang 2016a; Kang 2016b). No pooled analysis was conducted as the outcome was reported inconsistently with mean or median (Table 2). Overall, three trials (619 participants) with lower risk of bias reported no significant difference on duration of diarrhoea between two groups (Gharial 2017; [Kang 2016a; Kang 2016b]; Santos 2009); while the other two (248 participants) showed duration of diarrhoea was shorter in those receiving racecadotril (Salazar‐Lindo 2000; Sreenivas 2017).

1

Average duration of diarrhoea in two groups

Group	Trial	Number of participants	Duration of diarrhoea (days)		P value
			Racecadotril	Control
Outpatients	Santos 2009	179	4.0 (2.1)a	4.7 (2.2)a	0.15
	Kang 2016b	196	2 (2, 4)b	2 (2, 4)b	0.88
Inpatients	Salazar‐Lindo 2000(rotavirus‐positive)	73	1.2c	2.2c	< 0.001
	Salazar‐Lindo 2000(rotavirus‐negative)	62	1.2c	3c	< 0.001
	Gharial 2017	120	3 (2, 4)b	2 (1, 3)b	0.77
	Kang 2016a	124	1.1 (0.6, 1.8)b	1.0 (0.7, 1.9)b	0.57
	Sreenivas 2017	113	1.6 (0.3)a	2.1 (0.6)a	< 0.001

aMean (standard deviation).
 bMedian (interquartile range).
 cMedian without reporting interquartile range.

Number of stools

Four trials (491 participants) reported number of stools in two groups. No pooled analysis was conducted across all four trials as the outcomes were measured with different time period and were reported inconsistently with mean or median (García 2016; Gharial 2017; Santos 2009; Sreenivas 2017). Two trials reported the daily mean number of stools and the pooled analysis shows that the daily number of stools was not significantly different between two groups, both at 24 hours (MD −0.24, 95% CI −0.90 to 0.43, I² = 36%; 2 trials, 258 outpatients; Analysis 1.2; Figure 5) and 48 hours after treatment (MD −0.80, 95% CI −1.87 to 0.27, I² = 67%; 2 trials, 258 outpatients; Analysis 1.3; Figure 6) (García 2016; Santos 2009). Santos 2009 compared the number of stools at different time points according to microbiological results (bacterial and viral) of the stools cultures and found no significant difference between groups at any time of the visits. The trial also conducted a Poisson linear regression which showed a rapid decline in the number of stools in both groups without any statistically significant difference by 48 hours after the beginning of therapy. Similar results were obtained in both ITT and per protocol analyses.

1.2

Analysis

Comparison 1 Racecadotril versus placebo/no treatment, Outcome 2 Number of stools within 24 hours.

5

Forest plot of comparison: 1 Racecadotril versus placebo/no treatment, outcome: 1.2 Number of stools within 24 hours.

1.3

Analysis

Comparison 1 Racecadotril versus placebo/no treatment, Outcome 3 Number of stools between 24 to 48 hours.

6

Forest plot of comparison: 1 Racecadotril versus placebo/no treatment, outcome: 1.3 Number of stools between 24 to 48 hours.

Gharial 2017 (120 inpatients) reported the median number of stools in the first 48 hours after introduction of racecadotril or placebo and it was five (IQR 3 to 7) in the racecadotril group and five (IQR 2.5 to 7.5) in the placebo group, and the difference was not significant (P = 0.63). However another trial (113 inpatients), with higher overall risk of bias, reported the mean number of stools within 48 hours was 10.5 ± 3.2 (SD) in racecadotril group and 15.9 ± 4.6 (SD) in the control group and the difference was significant (P < 0.001) (Sreenivas 2017).

Stool output

Two trials reported the stool output within 48 hours after treatment (Cézard 2001; Salazar‐Lindo 2000).The pooled estimates showed that stool output was significantly lower in the racecadotril group than in the placebo group (SMD −0.65, 95% CI −0.88 to −0.41, I² = 0%; 2 trials, 301 inpatients; Analysis 1.5; Figure 7).

1.5

Analysis

Comparison 1 Racecadotril versus placebo/no treatment, Outcome 5 Stool output within 48h.

7

Comparison 1 Racecadotril versus placebo/no treatment, Outcome 1.5 Stool output within 48 hours.

The subgroup analysis from data of rotavirus‐positive children showed that the stool output was also significantly lower in the racecadotril group than in the placebo group (SMD −0.83, 95% CI −1.19 to −0.47, I² = 0%, 2 trials, 128 inpatients; Analysis 1.5). One trial also reported that weight of stool output within 24 hours after drug treatment was significantly lower in the racecadotril group compared to placebo (MD = −5.50, 95% CI −7.71 to −3.29; 1 trial, 167 inpatients; Analysis 1.4; Figure 8). Stool output in the racecadotril group was significantly lower both in ITT (P = 0.025) and per protocol analysis (P = 0.015) (Cézard 2001).

1.4

Analysis

Comparison 1 Racecadotril versus placebo/no treatment, Outcome 4 Stool output within 24h.

8

Comparison 1 Racecadotril versus placebo/no treatment, Outcome 1.4 Stool output within 24 hours.

Length of hospital stay for inpatients

Two trials (244 participants) reported the median length of hospital stay in racecadotril and placebo group. Kang 2016a reported that length of hospital stay was three (IQR 2.4) days in both groups (P = 0.96). Gharial 2017 also found no significant difference between two groups and the duration was 4 (IQR 1.5 to 6) days in the racecadotril group and 4.5 (1.8 to 6.3) days in the placebo group (P = 0.71).

Adverse effects

Five trials (688 participants) reported the overall incidence of adverse events in two groups (Cézard 2001; García 2016; Gharial 2017; Salazar‐Lindo 2000; Santos 2009). The pooled overall rate of adverse events was 17.4% in the racecadotril group and 18.4% in the control group (RR 0.90, 95% CI 0.66 to 1.22, I² = 0%; 5 trials, 688 participants; Analysis 1.6; Figure 9), showing no significant difference between two groups. The sensitivity analysis based on the number of participants originally randomized in two groups does not change the significance of the result.

1.6

Analysis

Comparison 1 Racecadotril versus placebo/no treatment, Outcome 6 Adverse events.

9

Comparison 1 Racecadotril versus placebo/no treatment, Outcome 1.6 Adverse events.

The commonly reported adverse events were urticaria and vomiting, but the latter may be related to the underlying diarrhoea aetiology. Santos 2009 reported three serious adverse effects. One child in the racecadotril group had an elevation in transaminases (ALT, 957 UI/L; AST, 1357 UI/L) causing one week of hospitalization, and the transaminases became normalized 3 months later. However, the causality is not clear as a viral infection was suspected. Two children in the control group were admitted to hospital for vomiting and dehydration. Salazar‐Lindo 2000 reported four cases of adverse events (two hypokalaemia, one ileus, and one mild fever) that were considered to be possibly related to racecadotril, but none of them were considered serious. Sreenivas 2017 (113 participants) reported that six children from the control group and five from the trial group had serum K+ values of less than 3.5 mmol/L, but none in either group had any electrocardiogram changes or symptoms and all recovered without any treatment and had normal serum K+ when repeated after 24 hours.

Discussion

Summary of main results

We identified seven RCTs (1140 children under five years old) that compared the efficacy and safety of racecadotril to placebo or no treatment. Trials were conducted in both inpatient and outpatient settings in countries with different levels of income. Compared to placebo or no treatment, racecadotril may reduce the risk of rehydration failure (low‐certainty evidence). Data on duration of diarrhoea, number of stools in the first 48 hours are insufficient to reach a conclusion; stool output in the first 48 hours appears to be lower in the two trials measuring this, although the data is not combinable. Length of hospital stay was similar in two studies measuring this, and overall there was no evidence that racecadotril increased overall rate of adverse events (low‐certainty evidence). Most adverse events in the racecadotril group were mild or moderate. See Table 1.

Overall completeness and applicability of evidence

Although racecadotril is widely used in Europe, Asia, and South America, only seven trials met the inclusion criteria and were included in this review. More trials, especially on children from developing countries, are needed. A key issue that limits the application of the evidence is the heterogeneity in context, with trials either examining the inpatient or outpatient setting and as such focusing on very different outcomes.

Failure of rehydration is an important outcome that evaluates the effect of racecadotril on improving the success of rehydration and could directly reflect the clinical value of this drug. Other Cochrane Reviews on children's diarrhoea made failure of rehydration one of the primary outcomes in their efficacy assessment (Hahn 2002; Hartling 2006). However, only two included trials addressed this outcome (García 2016; Sreenivas 2017). Future trials on this topic may consider it as an outcome for efficacy assessment.

The efficacy of racecadotril might be influenced by the severity of the disease, and children with more severe diarrhoea may have better response to treatment. Only two trials measured the severity of diarrhoea using a score system (Gharial 2017; Kang 2016a), and only Gharial 2017 applied it to the inclusion criteria. We roughly stratified participants into two level of severity based on being inpatients or outpatients. Future trials might use a score system to measure the severity of diarrhoea and investigate the impact of disease severity on the efficacy of racecadotril.

This review only included RCTs for safety assessment, but observational studies could also provide important data on safety, especially on reporting serious rare adverse events. Nucera 2006 reported a case of a 3‐year‐old child (weight: 20 kg) with generalized oedema, itching, and aphonia after a two‐day treatment with racecadotril. He recovered after discontinuation of racecadotril and treatment with methylprednisolone, hydroxyzine, and adrenaline. Moreno 2009 reported a 12‐year‐old boy who developed cephalic pruritus 10 minutes after administration of 100 mg of racecadotril. Simple‐blinded controlled oral challenge was conducted with increasing dose of racecadotril; 15 minutes later the child suffered a facialis erythema and a generalized urticaria, which illustrated that racecadotril is related to hypersensitivity. The small sample size and short duration of follow‐up within the included RCTs limits this review's ability to analyse adverse effects. Further reviews may consider including observational studies for safety assessment.

In applying the evidence to clinical practice, the types of patients should be considered. We only included children under five in the efficacy and safety assessment for a single clinical presentation, leading to a homogeneous data set. However, racecadotril is widely used in other age groups of children and adults. The results of this review might not be applicable to other age groups. In addition, racecadotril has been used for other causes of loose stools, such as in bacterial infectious diarrhoea, and diarrhoea secondary to other conditions like chemotherapy, HIV, malnutrition, etc. The results of this review may not be applicable to those groups of patients.

Certainty of the evidence

See Table 1. The certainty of evidence obtained from current trials ranged from very low to low. Generally, the quality of trial reporting is higher in recently published studies than older ones. There has been improvement with regard to reporting methods of randomization in recent studies but many still did not report how they performed allocation concealment. Future studies should also improve transparency on reporting roles of pharmaceutical companies during the trial if they receive funding from the industry.

Potential biases in the review process

We used a well‐designed search strategy to find clinical studies on this topic. There were, nevertheless, unpublished data that were not available to us; and there might also be existing published studies that were not found by our search strategies. There is consequently the possibility of bias resulting from incomplete data for analysis.

We did not find enough trials to test publication bias using funnel plots. There is, however, a potential for publication bias as some trials were funded by the pharmaceutical industry and the conflict of interest might influence published research results.

Agreements and disagreements with other studies or reviews

Several recently published systematic reviews have evaluated the efficacy and safety of racecadotril in children (Eberlin 2018; Gordon 2016; Lehert 2011). This trial found that racecadotril might reduce the weight of stool output, which is consistent with previous reviews, but there are more discrepancies. Our review found that there is not sufficient evidence to show racecadotril shortens the duration of diarrhoea or reduces stool frequency, which is inconsistent with other published reviews (Eberlin 2018; Gordon 2016; Lehert 2011). This might mainly be due to different trials being included in the analysis. Lehert 2011 included mostly earlier trials which tend to report significant results on these outcomes. Also, Lehert 2011 was an analysis of individual patient data and included data from five additional studies compared to this review (Calatayud 2009; Cojocaru 2002; Gutiérrez‐Castrellón 2010; Melendez‐Garcia 2007; Savitha 2006). We excluded Calatayud 2009 and Cojocaru 2002 from this review for being non‐randomized studies. We excluded Melendez‐Garcia 2007 as the intervention in the control group was kaolin/pectin. We also excluded the unpublished study, Gutiérrez‐Castrellón 2010, because some information in the study report we obtained might not be accurate and we failed to obtain further explanation from the author. For example, it reported that non‐dehydrated children have much longer disease duration than dehydrated children in the results, which is contrary to common sense; and Chi² test was used for stool frequency which was, in fact, reported as a continuous variable. Also, the effect size reported in this study was much bigger than other studies, which brings great heterogeneity and might seriously bias the pooled estimates. We do not have access to data of the unpublished study Savitha 2006, which was included by another meta‐analysis (Lehert 2011). Compared to Gordon 2016, we did not include Gutiérrez‐Castrellón 2010 for the same reasons above. Compared to Eberlin 2018, we did not search for and include RCTs published in Chinese medical journals, as many of them were found not to be RCTs (Wu 2009).

Authors' conclusions

Racecadotril seems to be a safe drug but has little benefit in improving acute diarrhoea in children under five years of age. Current evidence does not support routine use of racecadotril in management of acute diarrhoea in children under five outside of the context of placebo‐controlled RCTs.

More RCTs with good study design, reporting quality, and larger sample size are still needed if this drug is to be further investigated as an option. Further trials may consider failure of rehydration as an outcome for evaluating efficacy of racecadotril. They might also stratify participants based on different level of disease severity using more direct approaches, such as a severity scoring system, and investigate the impact of disease severity on efficacy.

Savitha 2006

Methods	RCT
Participants	Children aged 3 to 60 months with acute diarrhoea
Interventions	Racecadotril + rehydration therapy versus placebo + rehydration therapy
Outcomes	Diarrhoea duration
Notes	Data was not found

Abbreviations: RCT: randomized controlled trial.

Comparison 1

Racecadotril versus placebo/no treatment

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Failure of rehydration	2	192	Risk Ratio (M‐H, Fixed, 95% CI)	0.41 [0.13, 1.23]
2 Number of stools within 24 hours	2	258	Mean Difference (IV, Fixed, 95% CI)	‐0.24 [‐0.90, 0.43]
3 Number of stools between 24 to 48 hours	2	258	Mean Difference (IV, Random, 95% CI)	‐0.80 [‐1.87, 0.27]
4 Stool output within 24h	1	167	Mean Difference (IV, Fixed, 95% CI)	‐5.5 [‐7.71, ‐3.29]
5 Stool output within 48h	2		Std. Mean Difference (IV, Fixed, 95% CI)	Subtotals only
5.1 Stool output within 48 hours (all patients)	2	301	Std. Mean Difference (IV, Fixed, 95% CI)	‐0.65 [‐0.88, ‐0.41]
5.2 Stool output within 48 hours (rotavirus‐positive patients)	2	128	Std. Mean Difference (IV, Fixed, 95% CI)	‐0.83 [‐1.19, ‐0.47]
6 Adverse events	5	688	Risk Ratio (M‐H, Random, 95% CI)	0.90 [0.66, 1.22]