1.1.1. The effectiveness of various antibiotics in community-acquired childhood pneumonia: a systematic review

Kabra et al (16) systematically reviewed literature comparing various antibiotics, in order to identify effective drug therapy for community-acquired pneumonia in children. All randomized controlled trials (RCTs) that compared at least two antibiotics for community-acquired pneumonia in children in hospital or ambulatory settings were selected.

The review showed that for treatment of pneumonia, cotrimoxazole was inferior in efficacy to both amoxicillin (clinical failure rates odds ratio (OR) 1.33; 95% CI 1.05 to 1.67) and procaine penicillin (clinical cure rates OR 2.64; 95% CI 1.57 to 4.45). Injectable penicillin in conjunction with injectable gentamicin was more effective than injectable chloramphenicol alone (re-hospitalization rates OR 1.61; 95% CI 1.02 to 2.55). Co-amoxyclavulanic acid was more effective than amoxicillin alone (cure rates OR 10.44; 95% CI 2.85 to 38.21). There were no differences between injectable penicillin and oral amoxicillin (failure rates OR 1.03; 95% CI 0.81 to 1.31); between azithromycin and erythromycin; between cefodoxime and amoxicillin; or between azithromycin and co-amoxyclavulanic acid.

The authors concluded that for treatment of ambulatory patients with community-acquired pneumonia, amoxicillin was more effective than cotrimoxazole. Moreover, in hospitalized patients, penicillin was found to be more effective than cotrimoxazole; and the combination of penicillin and gentamicin was superior to chloramphenicol. Injectable penicillin and oral amoxicillin had similar clinical failure rates.

Another systematic review (17) was conducted by nine clinicians and researchers with extensive experience in the field of childhood pneumonia. The review aimed to identify appropriate first- and second-line antimicrobial agents for the empirical treatment of fast breathing pneumonia in children by first-level health care providers, in addition to defining treatment failure and criteria for referral. For the first-line treatment oral amoxicillin was found to be most effective in the treatment of fast breathing pneumonia. The review suggested that treatment failure for fast breathing pneumonia should be redefined as “clinical deterioration” only, instead of changing therapy when the child's condition is the “same” on follow-up after two or three days of treatment. It was suggested that once classified as “treatment failure” each child should be assessed independently for referral as not all cases will mandate immediate referral. The review recommended the development of an algorithm to determine the best course of action when treatment failure occurs.

1.1.2. A three-day course of antibiotics is as effective as a five-day course in treating children with fast breathing pneumonia

A systematic review by Haider et al (18) reviewed three randomized controlled trials evaluating the efficacy of short-course versus long-course treatment of fast breathing pneumonia in children age 2–59 months. Two of the studies (19,20) examined treatment with oral amoxicillin, while the third (21) examined treatment with oral cotrimoxazole. All studies compared the duration of two courses of treatment, one lasting three days and the other five days, while holding the antibiotic constant. The three studies were double blind, randomized, placebo controlled trials with a total of 6210 participants. Data was available for 5763 cases. The primary outcome was clinical cure at the end of the treatment. The studies found no significant differences in clinical cure rates between the two groups (RR 0.99; 95% CI 0.97 to 1.01). Nor were there significant differences in either the treatment failure or relapse rates between groups. The review recommended a shorter course of antibiotic therapy, keeping in mind the benefits to the individual as well as to the health system, especially in settings with limited resources.

1.2.1. Oral amoxicillin is as effective as injectable penicillin in the treatment of chest indrawing pneumonia in children in low-resource settings

A multicentre, randomized, open-label trial by Addo-Yobo et al (24) (APPIS Study) compared the efficacy of oral amoxicillin and injectable penicillin. The study was conducted at nine tertiary health facilities in eight developing countries and enrolled 1702 children age from 3 to 59 months with chest indrawing. Children were randomly assigned to receive either a five-day course of oral amoxicillin (n=857) or parenteral penicillin (n=845). Evaluations were carried out at 48 hours, five days and 14 days. The primary outcome of the study was clinical treatment failure at 48 hours. The study found treatment failure rates of 19% in each treatment group (n=161 penicillin; n=167 amoxicillin; risk difference -0.4%; 95% CI -4.2 to 3.3). The study described a number of advantages of oral treatment over parenteral treatment, including reduced risk of injection-related morbidity, and fewer needs for medical supplies such as needles, and suggested that oral treatment be considered as an equally effective alternative to parenteral treatment. Although this study proved the effectiveness of oral amoxicillin, it did not fully address the issue of safety, as all enrolled children were hospitalized for 48 hours and kept under close supervision. More research was needed to test the hypothesis under a wider range of circumstances.

1.2.2. Oral amoxicillin is equally effective for pneumonia of various severities in a high-resource setting

A study in England in 2007 (PIVOT Trial) (25) compared oral amoxicillin and intravenous (IV) benzyl penicillin in the management of severe pneumonia. The randomized, controlled, non-blinded equivalence trial was conducted in eight paediatric centres (district general and tertiary hospitals) and enrolled children with all but the most severe cases of pneumonia. Additional exclusion criteria were: wheeze, oxygen saturation less than 85%, shock, immunodeficiency, pleural effusion at presentation requiring drainage, chronic lung condition (excluding asthma), penicillin allergy, and age less than 6 months. Children were randomly assigned to a 7-day treatment of either oral amoxicillin or IV benzyl penicillin. The primary outcome was the time required for temperature to be below 38 °C for 24 continuous hours. The study found the two treatments to be equivalent, each having a median time of 1.3 days to achieve the primary outcome. While equivalence between oral and parenteral antibiotics itself was not a new development, this study was unique in the scope of varying severities of pneumonia treated. It was also the first study of its kind for children in a high-resource setting with radiologically confirmed pneumonia. The study recommended that children be treated with oral amoxicillin instead of IV benzyl penicillin, as oral treatment was both painless and non-invasive.

1.2.3. It is safe to treat chest indrawing pneumonia at home with oral amoxicillin

Hazir et al (26) (NO-SHOTS Study) carried out a randomized, open-label equivalency trial at seven sites in Pakistan, comparing hospitalization with parenteral ampicillin to home treatment with oral amoxicillin. Children with chest indrawing pneumonia were randomized to treatment in hospital (n=1012) with two days of injectable ampicillin followed by three days of oral amoxicillin (80–90 mg/kg/day), or were sent home with a five-day twice daily course of oral amoxicillin (n=1025). The primary outcome was clinical treatment failure. Follow-up assessments were conducted at 1, 3, 6, and 14 days after enrolment. At day 6 there were 87 (8.6%) treatment failures in the hospitalized group and 77 (7.5%) treatment failures in the home-treatment group (risk difference 1.1%, 95% CI − 1.3–3.5). In addition to finding equivalence between the two treatments, the authors suggested that in cases of chest indrawing pneumonia without underlying complications, home treatment with a short course of high-dose oral amoxicillin was preferable to parenteral treatment because of the associated reduction in referral, admission, and treatment costs as well as the reduced invasiveness of oral treatment.

1.2.4. Home therapy with oral amoxicillin is effective in a range of settings

To test whether the interpretation that chest indrawing pneumonia can be treated safely and effectively at home was generalizable across communities and geographic regions, a multicentre observational study was conducted in Bangladesh, Egypt, Ghana and Viet Nam (27). Outcome was ascertained on a total of 823 children age 3 to 59 months old, with chest indrawing pneumonia, who were given oral amoxicillin (80–90 mg/kg/day) twice daily for five days. Follow-up at home was carried out on days 1, 2, 3 and 6 to assess treatment failure, and at the health facility on day 14 to assess relapse. Overall treatment failure was 9.2%, varying from 6.4% in Ghana to 13.2% in Vietnam. The common reasons for failure were persistence of chest indrawing at day 6 (3.8%), abnormally sleepy or difficult to wake (1.3%) and central cyanosis (1.3%). The authors concluded that among children with chest indrawing pneumonia treated at home with oral amoxicillin, clinical treatment failure and adverse event rates did not differ across geographic areas. This argues in favour of the option of home-based therapy of chest indrawing pneumonia in a wide variety of settings.

In these studies, the children with chest indrawing pneumonia were identified and treated on an outpatient basis by qualified doctors, and were followed up either in an outpatient facility or at home. This made it difficult to generalize to situations where pneumonia is treated within the community, by CHWs who may have little or no formal education. Data were needed to assess the effectiveness of community case management of chest indrawing pneumonia.

1.2.5. Efficacy of higher dose (80–90 mg/kg/day) vs. lower dose (45 mg/kg/day) of amoxicillin

1.3.1. Penicillin/gentamicin vs chloramphenicol at high altitude

The first of these RCTs (41), conducted in the highlands of Papua New Guinea (1600–1800 m above sea level), included 1116 children aged 1–59 months with WHO-defined very severe pneumonia (modified to include heart failure as a danger sign). Enrolled children had a median oxygen saturation of 71%. Five hundred and fifty-nine (559) children were treated with 100 mg/kg/day chloramphenicol; 557 children were treated with penicillin (200 mg/kg/day) plus gentamicin (7.5 mg/kg/day). Duration of treatment was 14 days. Measured outcomes were: death, treatment failure by day 5, and readmission. More children in the penicillin/gentamicin group required a change of antibiotic (60 versus 49), while 147 (26%) children treated with chloramphenicol and 123 (22%) treated with penicillin and gentamicin had adverse outcomes (p = 0.11, not significant). Thirty-six children treated with chloramphenicol and 29 treated with penicillin and gentamicin died (difference not significant). More children treated with chloramphenicol presented again with severe pneumonia within one month of hospital discharge (p = 0.03), as compared to those treated with penicillin and gentamicin.

1.3.2. Penicillin/gentamicin vs chloramphenicol at low altitude

The second of the RCTs (42) was a multi-country study, with 80% of children residing at sea level. Children aged 2–59 months were enrolled: 479 were randomized to receive chloramphenicol (75 mg/kg/day), and 479 to receive ampicillin (200 mg/kg/day) and gentamicin (7.5 mg/kg/day). Median oxygen saturation on admission was higher than for the Papua New Guinea study (88%) (41). Duration of treatment was 10 days. More children in the chloramphenicol group required a change in antibiotic (45 versus 26). Measured outcomes were death or treatment failure by days 5, 10, and 21. More children failed treatment with chloramphenicol at day 5 (16% versus 11%; relative risk 1.43, 95% CI 1.03 to 1.97) as well as at days 10 and 21. There was a trend towards reduced death rates in children treated with ampicillin plus gentamicin, but this was not significant.

1.3.3. Ceftriaxone as second-line treatment

Although there were no data on the use of ceftriaxone in the treatment of pneumonia with general danger signs, the WHO Guidelines Development Group recognized the need to include ceftriaxone as a second-line treatment for children with severe pneumonia with general danger signs, especially for hospital care.

1.6.1. Management of pneumonia at community level

According to the WHO/UNICEF joint statement on management of pneumonia in community settings (52), an important strategy to increase access to quality care for pneumonia is to train and deploy CHWs to assess and treat children with pneumonia. There is strong scientific and programmatic evidence to support the effectiveness of this approach.

Two large-scale studies, in Bangladesh and Nepal, assessed the ability of volunteer CHWs with intensive basic training and close supervision to properly diagnose and treat pneumonia. Hadi et al (53) showed that the sensitivity, specificity, and overall agreement rates in pneumonia diagnosis and treatment were significantly higher among those health volunteers in Bangladesh who had intensive basic training and routine supervision than among those who had not. Ghimire et al (54) showed a significant trend towards a decrease in the proportion of pneumonia and severe pneumonia cases from 2004 to 2006 in districts of Nepal where the volunteers were given special training to manage pneumonia.

In addition, numerous countries implement community case management of childhood pneumonia. For example, the Gambia has a nationwide programme addressing pneumonia in the community (52). In Honduras, pneumonia treatment has been incorporated into the national integrated community child care programme where community volunteers – in addition to their other responsibilities – provide treatment for pneumonia and diarrhoea in more than 1800 communities (52). In 2008, the government of Malawi initiated activities to deliver community-level treatment of common childhood illnesses including suspected pneumonia. By September 2011, 3296 health surveillance assistants (HSAs) had been trained, and 2709 village health clinics were functional. Evaluation has shown that HSAs are able to treat sick children with a quality similar to that provided by professional health personnel in fixed facilities. Monitoring data also showed that communities were utilising the HSA services (55). Large-scale, national-level programmes supporting CHWs to treat illness including childhood pneumonia are also implemented in Ethiopia and Pakistan.

The current WHO/UNICEF tools for CHWs recommend oral amoxicillin in two daily doses. Because the CHW is not expected to treat chest indrawing pneumonia, these guidelines will retain two age bands and will not be revised at the present time. Some data from Asia (56,57) shows that CHWs can manage chest indrawing pneumonia with oral amoxicillin. However, current CHW guidelines will not be changed until more evidence becomes available from additional regions and countries.

1.6.2. Community management of chest indrawing pneumonia

Two studies have shown that CHWs, when properly trained and supported, can effectively and safely treat chest indrawing pneumonia at home with oral amoxicillin.

Bari et al (56) aimed to compare community management of children with chest indrawing pneumonia by lady health workers (LHWs), using oral amoxicillin, to the standard of care: a first dose of cotrimoxazole followed by referral. Twenty-eight clusters in the Haripur district of Pakistan were randomly assigned equally in a 1:1 ratio to either intervention or control. Children age 2 to 59 months with chest indrawing pneumonia were included. In the intervention clusters, LHWs provided mothers with oral amoxicillin (80–90 mg/kg/day in two divided doses) for five days, and gave specific guidance on its use. In the control clusters, LHWs gave the first dose of oral cotrimoxazole and referred the child for standard care to the nearest health facility. The primary outcome was treatment failure by day 6. A total of 1995 children participated in the 14 intervention clusters and 1477 in the 14 control clusters. Treatment failure rates were significantly reduced in the intervention clusters (165 [9%] vs241 [18%], risk difference −8.9%). Most of the risk reduction was in the occurrence of fever and chest indrawing on day 3 (−6.7%, −10.0 to −3.3). Two deaths were reported in the control clusters and one in the intervention cluster. It was concluded that community case management by LHWs could result in appropriate treatment for children with chest indrawing pneumonia, reduce delays in initiating treatment, and reduce costs for families and health systems.

Similar evidence on the management of severe pneumonia by LHWs was generated in a cluster-randomized controlled trial in Matiari district of rural Sindh, Pakistan (57). Public-sector LHWs undertook community management of chest indrawing pneumonia. In the intervention clusters children with suspected pneumonia were screened by LHWs, and identified cases of chest indrawing pneumonia were prescribed oral amoxicillin (90 mg/kg per day in two doses) for five days at home. Children in control clusters were given one dose of oral cotrimoxazole and were referred to their nearest health facility for further management. In both groups, follow-up visits at home were conducted at days 2, 3, 6, and 14 by the LHWs. The primary outcome was treatment failure by day 6 after enrolment. 2341 children in the intervention clusters and 2069 children in the control clusters participated in the study. Treatment failure by day 6 was 8% in the intervention group and 13% in the control group. After adjusting for clustering, the risk difference for treatment failure was −5.2% (95% CI −13.7% to 3.3%). Two deaths by day 6 and one between days 7 and 14 were recorded; no serious adverse events were witnessed. The authors concluded that properly trained LHWs were able to satisfactorily diagnose and treat chest indrawing pneumonia at home in rural Pakistan. This strategy could effectively increase access to care for pneumonia in settings where referral is difficult, and could become a key component of community detection and management strategies for childhood pneumonia.