Aim

To evaluate the effectiveness of amoxicillin in children who presented with pneumonia as per WHO definition of respiratory symptoms plus fast breathing. 

Rationale

Severe pneumonia kills many children each year. However, there is also concern about antibiotic over- prescribing in milder cases where the infection may not be bacterial in origin.The WHO has a simplified protocol; children with cough and fast breathing (respiratory rate over 50 if <11 months or over 40> if 12-59 months) and no chest indrawing are defined as non-severe pneumonia. This enables less experienced healthcare professionals to treat this common condition. The authors are questioning the rationale of this approach

Design

Mulit-centre double blind placebo controlled trial.

Setting

Hospitals in 3 cities in Pakistan 

Population

Children aged 2-59 months with fast breathing, or cough or difficulty breathing. If fulfilled criteria of non-severe pneumonia as per WHO algorithm were included. Excluded children with underlying chronic illness, history of >3 episodes of wheeze or recent antibiotic use. Respiratory rate measured twice and average taken.900 children enrolled

Intervention

Intervention; 3 days of amoxicillin 15mg/kg every 8 hours 

Control; equivalent placebo

Salbutamol and paracetamol given to both groups as required

Primary outcome

Therapy failure at day 3; defined as lower chest indrawing or danger signs.  

Secondary outcomes

Treatment failure at day 5; fast breathing, chest indrawing or danger signs. 

Relapse; signs of pneumonia at 6-14 days.

Results

In per-protocol analysis at day 3, 31 (7.2%) of the 431 children in the amoxicillin arm and 37 (8.3%) of the 442 in the placebo group had therapy failure. This difference was not statistically significant (odds ratio [OR], .85; 95%CI, .50–1.43; P = .60).

Author's Discussion

The clinical outcome of those with WHO defined non-severe pneumonia is unchanged by amoxicillin administration. Further studies are needed in areas with higher incidence of pneumonia.

Strengths

I like defintion of therapy failure; the reason we treat these milder cases with antibiotics is to try and prevent more serious pneumonia

.Double blinded large trial Treatment failure rate of 7-8% feels about right.

Pragmatic trial in a low-resource clinic setting, no use of imaging or blood work.

Weakness

I could not see a power calculation

Large difference of treatment failure at day 5 between amoxicillin ( 13.5% [58 children]) and in the placebo group (17.6% [78 children]) and although not statistically significant this could be due to being underpowered for the outcome. 

It would have been nice to have some idea what percentage of cases were genuine cases as per gold standard

My humble opinion

This study was an interesting use of a placebo controlled trial to examine a definition of pneumonia. It is effective as instead of comparing to a (potentially flawed) gold standard, it uses a measure of effectiveness of antibiotics to prevent severe pneumonia as a marker. Something we as clinicians should be much more interested in than if XR changes or the like. Interestingly, the more symptoms and signs the more likely they were to have treatment failure, but no one sign was predictive. My personal practice of using a large number of factors to differentiate a well child with likely viral illness and potentially slightly more unwell child with pneumonia will not change, I will also have a lower threshold still in patients who are less able to return to a healthcare facility. However, when putting systems in place for less trained healthcare staff to screen for pneumonia I would be more cautious about using respiratory rate as the sole qualifier for antibiotics.