Children with severe malnutrition who are admitted sick to hospitals have a high mortality(death rate), usually because of infection. All children with severe malnutrition admitted to hospitals are treated with antibiotics(medication used to kill bacteria). However, the current antibiotics used in hospitals may not be the most effective. It is possible that the antibiotics that are currently used after initial antibiotics should be used first. No studies have been carried out to determine if the current antibiotics used for treating malnourished children who are sick and admitted in hospital are the most appropriate. The aim of this study is to find out if a changed antibiotic system for children with malnutrition is safe, reduces the risk of death and improves nutritional recovery.

Children with complicated severe acute malnutrition (SAM) admitted to hospital in sub-Saharan Africa have a case fatality(death rate) between 12% and more than 20%. Because children with SAM may not exhibit the usual signs of infection, WHO guidelines recommend routine antibiotics(medication used to kill bacteria). However, this is based on "low quality evidence". There is evidence that because of bacterial resistance to the currently recommended first-line antibiotics (gentamicin plus ampicillin or penicillin) could be less effective than potential alternatives. Some hospitals in Africa are already increasing use of ceftriaxone as a first-line treatment. However, this is not based on any data that ceftriaxone actually improves outcomes. Of concern is that ceftriaxone use may also lead to increased antimicrobial resistance, including inducing extended spectrum beta-lactamase (ESBL) and other classes of resistance.

A further area where evidence for policy is lacking is the use of metronidazole in severely malnourished children. The WHO guidelines recommend "Metronidazole 7.5 mg/kg every 8 h for 7 days may be given in addition to broad-spectrum antibiotics; however, the efficacy of this treatment has not been established in clinical trials." Metronidazole is effective against anaerobic bacteria, small bowel bacterial overgrowth, Clostridium difficile colitis and also Giardia, which is common amongst children with SAM. Small cohort studies of metronidazole usage suggest there may be benefits for nutritional recovery in malnourished children. However, metronidazole can cause nausea and anorexia, potentially impairing recovery from malnutrition and may also rarely cause liver and neurological toxicity.

This multi-centre clinical trial will assess the efficacy of two interventions, ceftriaxone and metronidazole, on mortality and nutritional recovery in sick, severely malnourished children in a 2x2 factorial design. There will also be an analysis of antimicrobial resistance and an economic analysis. To extend our understanding of metronidazole and ceftriaxone pharmacokinetics, additional pharmacokinetic data for the dosing schedule used in the trial will be collected from 120 participants in a sub-study. The trial will be conducted at Kilifi County Hospital, Coast General Hospital, Mbagathi Hospital in Kenya and Mbale Regional Referral Hospital in Uganda. The trial will assess antimicrobial resistance that is carried by children in their intestines and in invasive bacterial isolates. A further sub-study will examine the relative costs of care for SAM for health facilities and for families, including antimicrobial usage will also be assessed. Clear data on the benefits, risks and costs of these antimicrobials will influence policy on case management and antimicrobial stewardship in this vulnerable population.

Inclusion:

• Age 2 months to 13 years inclusive
• Severe malnutrition defined as:
• kwashiorkor at any age or:
• for children between 2 to 5 months: MUAC <11cm or weight-for length Z score <-3
• for children between 6 to 59 months: MUAC <11.5cm or weight-for length Z score <-3
• for children between 5 to 13 years: MUAC <11.5cm or BMI-for-age Z score <-3
• Admitted to hospital and eligible for intravenous antibiotics according to WHO guidelines
• Planning to remain within the hospital catchment area and willing to come for specified visits during the 90 day follow up period
• Informed consent provided by the parents/guardian

Exclusion:

• Known allergy or contraindication to penicillin, gentamicin, ceftriaxone or metronidazole
• A specific and documented clinical indication for another class of antibiotic
• Previously enrolled in this study

Data sharing will be considered by applying to the Data Governance Committee at CGMR,C Kilifi who will manage the process and ensure that appropriate ethical approval is in place and consent has been obtained for uses outside those given in this protocol:DGC@kemri-wellcome.org. Explanation of this eventuality will be included in the participant information and consent form. We intend to share anonymised data with the CHAIN network cohort study (KEMRI/SERU/CGMR-C /054/3318, OxTREC 34-16) which consists of institutions doing studies in malnourished children in Africa and South Asia; to upload anonymised bacterial and resistance sequences on recognised international repositories; and share anonymised bacterial sequence data relating to bacterial resistance with groups working an antimicrobial resistance in Africa and elsewhere.

2 years after study end, no end date foreseen

Managed access. Apply to the Data Governance Committee: dgc@kemri-wellcome.org

Metronidazole and its placebo will be masked. They will both be contained in similar bottles labelled with sequential study numbers according to a prepared blocked randomisation list before the trial begins. They will also be of similar appearance.

Ceftriaxone and penicillin + gentamicin will not be masked.