Supportive care and antibiotics for severe pneumonia among hospitalized children: A pragmatic randomised controlled trial

Pneumonia, an infection of the lungs, is the leading cause of deaths among young children. The World Health Organization (WHO) have developed recommendations for the diagnosis and treatment of pneumonia in low and middle income countries using simple clinical features and low cost, widely available antibiotics. The recommend treatment for children at the highest risk of death (severe pneumonia) is injectable benzylpenicillin or ampicillin and gentamicin. Following the introduction of vaccines against the main causes of pneumonia to national immunization programmes in many low-income countries, there has been growing debate over the appropriateness of the currently recommended treatments. Many clinicians already believe that the recommended treatment is ineffective and frequently opt to use other antibiotics such as amoxicillin-clavulanic acid and ceftriaxone instead. The first key question in this study seeks to compare two antibiotics against the current recommended treatment. We will investigate is whether either (i) intravenous amoxicillin-clavulanic acid or (ii) ceftriaxone is superior to benzylpenicillin plus gentamicin (standard care) for the treatment of children admitted to hospital with severe pneumonia.

Some authorities advise against feeding through a tube inserted into the stomach through the nose in severely ill children. The main reason for this is are the potential for compromising the ability to breath in a patient already experiencing difficulty breathing and an increased risk of choking on feeds given through the tube. However, the alternative, providing fluids through an intravenous drip requires careful monitoring by a nurse to ensure the fluid is given at a safe rate over the desired duration. This a common challenge in many low resource settings where a limited number of nursing staff are required to attend to several duties. Fluids provided through a drip are also lack the necessary nutrients to match the increased demands of the body during a serious illness. The second key question for this study is therefore whether feeding through a tube inserted into the stomach through the nose is superior to providing fluids through an intravenous drip for the management of children with severe pneumonia.

We will recruit 4392 children at 12 hospitals in Kenya. Children who meet the criteria for recruitment will be allocated to the study treatment groups through a balanced process pre-determined process that ensures each participant has a fair chance of receiving any given study treatments. Each of the two questions will be studied in the same set of patients. Thus, a child recruited in the study will receive any one of the three antibiotic treatments and either of the two fluid treatments. For each of the study questions, we will compare the percentage of children who die within the first five day of recruitment in the alternative treatment groups. We will also compare the length of hospitalisation, and the percentage of children who die within 30 days of recruitment in the alternative treatment groups. Chest X-rays and blood samples will be collected in a smaller group of patients to examine possible explanations for differences in responses to the treatments. Finally, we will compare the costs of receiving the alternative study treatments against the outcomes we observe among children assigned to the respective study groups and explore the social perceptions of caregivers and health workers towards the treatments.

Pneumonia is the leading cause of deaths among young children, with inpatient mortality exceeding 10% for the most severe form of the syndrome (severe pneumonia). WHO pneumonia case management guidelines recommend empirical treatment with injectable benzylpenicillin or ampicillin and gentamicin for children at the highest risk of mortality (severe pneumonia). Following the introduction of vaccines against the dominant causes of pneumonia to national immunization programmes in many low-income countries, there has been growing debate over the appropriateness of the current recommendations for empirical treatments. The first key question is whether either intravenous amoxicillin-clavulanic acid or ceftriaxone is superior to benzylpenicillin plus gentamicin (standard care) for the treatment of children hospitalized with severe pneumonia.
Some authorities, including the British Thoracic Society, advise against nasogastric hydration in severely ill children, citing concerns of compromised respiratory status and increased risk of aspiration with nasogastric feeding. However, administration of intravenous fluids demands careful monitoring - a common challenge in many low resource settings. Intravenous fluids are also deficient in nutritive value to match the raised metabolic demands associated with acute illness. The second key question is therefore whether nasogastric feeding is superior to intravenous fluid therapy for supportive management of children with severe pneumonia.
This 3x2 factorial pragmatic trial will enrol 4392 children at 12 hospitals in Kenya. The primary endpoint will be mortality at Day 5 post-enrolment. Secondary outcomes will be length of hospitalisation, and mortality at Day 30. Chest radiographs and blood cultures will be performed on a representative sample of patients to provide data on aetiology. We will assess cost effectiveness of the interventions as well as social perceptions of caregivers and health workers towards the study treatments.

We will seek to maximise economic and societal impact of the proposed study through targeting three interrelated audiences.

1. Policy makers
The investigators are active members of national and international technical panels supporting the development policy recommendations in child health. Locally, the PI sits on a national technical working group (TWG) convened by the Ministry of Health to specifically tackle childhood pneumonia. We will use this channel of communication to convene quarterly meetings to share study-related updates as well as coordinate public engagement activities on child health focusing on pneumonia. These meetings will also provide an opportunity for the investigators to receive feedback from the members of the TWG on opportunities that the study team may use to complement the wider Ministry of Health communication strategy. Ultimately, the stakeholders represented on this TWG will play an important role in the dissemination of the study findings and their incorporation into local, regional and global evidence-based clinical practice guidelines for implementation.

2. Health workers
Since 2013, KWTRP has maintained a collaboration with the Ministry of Health in 14 public hospitals to help develop approaches to optimise use of routine hospital data to improve quality of clinical documentation and clinical care. This ongoing project, supported by funding awarded to the Prof. Mike English (Co-PI) involves all 12 of the proposed study sites. Clinical staff at the proposed study hospitals have been directly engaged in the development of the study protocol. We will continue to engage the hospital teams by supporting regular meetings (three annual visits at each site during patient recruitment) at the study hospitals. Such meetings will seek to develop capacity for conducting high quality, ethical research at site level through providing training for clinical staff on areas including basic research methods and GCP. These meetings will also provide the investigators useful feedback on the local experiences of those involved in the study. We will also convene three meetings each year with the hospital paediatricians (site principal investigators) to share study-related updates, and where possible, support further development of their research skills including protocol development for study questions of their own and ultimately, facilitate their involvement in the writing of manuscripts for the trial.

3. General public
Engagement with the wider community will be integrated with the activities of KWTRP's Community Liaison Group. Currently, information on the research activities of KEMRI is given to the community through the activities of the programme's Community Liaison Group, including regular meetings with the KEMRI-Community Representatives, and meeting with County Health Management Teams, hospital and peripheral health staff and local administrative leaders.
Building on experience gained from activities undertaken with support of an advocacy grant awarded to the PI during a previous clinical trial, we will host annual activities on World Pneumonia Day (12th November) at the study sites. Here, caregivers, particularly those of study participants will be provided an opportunity to interact with the study investigators to share their experiences and understanding of the study and the wider implications of research on public health. The information gathered from these events will inform future approaches to designing and conducting similar future studies.
Material communicating lay information on pneumonia and the importance of this and other studies to the general public will be printed on flyers and posters for distribution to the study hospitals. We will also develop briefs for radio and print media stations. A regularly updated interactive public web page hosted by KWTRP will provide additional information on the study while allowing readers to share their views with the study team.