A double blind randomized community-based trial of amoxicillin versus placebo for non-severe pneumonia in children aged 2-59 months in Pakistan

Pneumonia is a major cause of illness and death in children in low-income countries. With a view to decreasing death from pneumonia, the World Health Organization and UNICEF developed the the Integrated Management of Childhood Illness (IMCI) algorithm which simplifies management of common childhood illnesses such as pneumonia and diarrhoea into different levels of severity for determining the most appropriate case management by primary healthcare providers. Many pneumonia cases are categorized as non-severe pneumonia (defined as fast breathing above the specified age cut-off for respiratory rates). As there is incomplete information regarding the cause of this type of "pneumonia" from primary care settings, treatment guidelines by WHO are dictated by culture information from hospital pneumonia cases which are different in severity and cause. Current WHO guidelines advocate the use of oral antibiotics for non-severe pneumonia. However, it is postulated that most non severe pneumonia not requiring hospitalization is of viral aetiology, thus does not require antibiotic treatment. The cost of antibiotic treatment for all children with pneumonia is high; an estimated US$ 200 million in South Asia & sub Saharan Africa alone. Since more than 60% of pneumonia is classified as non-severe, this puts a strain on already under-sourced programmes in low-income countries. Giving antibiotics where they confer no benefit also puts the child at risk of side effects and increases the risk of antimicrobial resistance in the community. This uncertainty forms the basis of the proposed study. We propose to show in a clinical trial that the outcome of children diagnosed with WHO defined non severe pneumonia is similar regardless of whether they receive antibiotics or not. This study will be conducted in five primary health care centres located in low income communities of Karachi, Pakistan, with extensive trial experience. Children identified to have fast breathing without any danger signs will be randomized to receive either three days of the WHO recommended oral antibiotic (Amoxicillin 45mg/kg/day divided twice daily) or matching placebo (a drug that will taste and look like the amoxicillin but will not have an active ingredient) by a study physician working at the primary health centre. The assignment of the antibiotic amoxicillin or placebo to a child will be done using a computer generated randomization list in a manner that at the end of the trial, there are equal numbers of children in both arms of the trial. Based on the statistical calculations for sample size, we will need to assign 521 children to receive amoxicillin and the same number of children to receive placebo. All children will receive the antibiotic or placebo under supervision of the primary health care physician in the morning. Evening doses will be delivered by locally hired Community Health Workers (CHWs) visiting the children at their home. All children will be assessed again on day 3 by a study physician to see if the child's presenting sign of high respiratory rate has resolved or not. All children with persistently high respiratory rate and/or development of a new clinical sign indicating illness progression will be labelled a treatment failure. There will invariably be some children with treatment failure in both the treatment arms; we hypothesize that there will be equal number of treatment failures in both the groups i.e. around 7%. If we are able to show with the help of this trial that there is no added advantage of prescribing antibiotics to children with non-severe pneumonia we will develop an evidence base to revise the current WHO guidelines and thus reduce the financial burden on an already resource constrained health system and also decrease out of pocket expenses for families. In the long term this will have implications for decreasing global antimicrobial resistance to antibiotics.

The WHO IMCI algorithm is designed for use in primary healthcare facilities in low-income countries to improve case management of common childhood illnesses, including pneumonia. It defines non-severe pneumonia as follows- a respiratory rate of 50 breaths/ min or more in infants, or respiratory rate of 40 breaths/ min or more in children 12-59 months irrespective of presence of wheeze and without any danger signs present. The current treatment guidance in low HIV-endemic areas is of a three day course of oral amoxicillin. This recommendation for empiric antibiotic use in non-severe lower respiratory tract illness, which is usually viral, has been questioned in the context of the poor specificity of pneumonia diagnosis in IMCI and the increasing use of vaccines against common bacterial respiratory pathogens. The main objective of the trial is to evaluate if amoxicillin therapy in children aged 2-59 months with non-severe pneumonia at primary health centres, is equivalent to placebo. We also aim to evaluate predictors of treatment failure and the association with particular aetiologies. We will conduct a two arm double blinded randomised trial in which children in reference arm will receive oral Amoxicillin 45 mg/kg/day in two divided doses for three days while children in the comparison arm will receive an inert placebo. Those at greater risk of bacterial pneumonia and asthmatics will be excluded. Both groups will be followed for treatment failure up to day 3 (primary outcome). Per protocol analysis (to avoid a beta error) of outcomes will be undertaken to compare proportional hazards of treatment failure by allocation group. If equivalence (a difference in failure rates of < 5% between groups) is proven, a basis for revision of the current guidelines can be proposed, thus reducing financial burden on both an already resource constrained health system and poor caregivers. In longer term this will have implications for decreasing antimicrobial resistance.

The proposed study is projected to have both academic and societal/economic impact.
Direct Beneficiaries: The local population will be direct beneficiaries of this research. These include locally resident community health workers. Employment of these women will improve their common knowledge regarding identification of childhood illnesses and their management which will help inculcate awareness among the community in general. It will also add to their monetary well-being. With economic prosperity they will be able to contribute meaningfully to their societies thus adding to their own social welfare. These benefits can be realized in a very short term but will also have long term impact with regards to welfare and empowerment of the locally hired Community Health Workers and enable them to contribute to the health and well-being of their community.
The primary health care physicians will be trained in the management of children in conformity with the World Health Organization's Integrated Management of Childhood Illnesses (IMCI). This will result in wider implementation of these WHO guidelines and will systematically address the most important causes of death and illness among children in these communities. In addition to focus on treatment of pneumonia, it will also provide an opportunity for, and emphasize important preventive interventions such as immunization and improve infant and child nutrition, including breastfeeding.
The results of this trial will help to rationalize the use of antibiotics and thus decrease the out of pocket expenses for the caregivers of these children. The children will also be saved from the undesirable side effects of the antibiotics including the long-term deleterious effects on the native gut micro-biota that may impair growth and nutrition subsequently resulting in altered immune processing that precipitates a vicious cycle of recurrent infections.
The communities in the study area will also benefit from the research as awareness will be fostered regarding signs of disease in children and their prevention and care.
Indirect beneficiaries: This study will help to fill the knowledge gap that exists regarding the usefulness of antibiotics for management of non-severe pneumonia in children two months to five year old, which is postulated to have a viral aetiology in majority of the cases. Rationalizing antibiotic use will reduce the burden on the health care system. It becomes even more important in the context of low income developing countries which have limited and scarce resources with competing priorities. There will be significant economic implications and consequences for the global problem of emerging antibiotic resistance. The results of this study can be used to inform a revision of the policy of empiric antibiotic use for children with isolated fast breathing, classified as non-severe pneumonia by World Health Organization (WHO), residing in HIV non endemic countries. This will increase the effectiveness of the public health care system in the developing countries. These benefits can be realized in a very short duration after the successful implementation of the policy at local, regional and international level. The use of naso-pharyngeal aspirates to detect aetiology of non-severe pneumonia within a trial and their relation with treatment outcomes will provide new insights into the role of organisms especially common viruses in the outcomes of non-severe pneumonia.