Table 4.

GRADE evidence profile 1: Cotrimoxazole versus amoxicillin for non-severe pneumonia^†

Quality assessment					Summary of findings
					Number of patients		Effect	Quality of evidence
No of studies/Design	Limitations	Inconsistency	Indirectness	Imprecision	Amoxicillin	Cotrimoxazole	Odds ratio (95% CI)
Outcome 1: Treatment failure based on clinical signs. Follow-up 2 - 5 days. Importance: Critical
2 RCTs [22,23]	no serious limitations	no serious inconsistency	serious1	no serious imprecision3	147/922 (15.9%)	231/1132 (20.4%)	0.83 (0.65 - 1.07)	Moderate
Outcome 2: Treatment failure based on clinical signs. Follow-up 4-6 days. Importance: Critical
1 cluster randomised trial [24]	no serious limitations	no serious inconsistency	serious2	no serious imprecision3	137/993 (13.8%)	97/1016 (9.5%)	1.52 (0.73 - 3.17)	Moderate
Outcome 3: Mortality Importance: Critical
No studies	-	-	-	-	-	-	-	-
Outcome 4: Cost Importance: Important
No studies	-	-	-	-	-	-	-	-
Overall quality of evidence: Moderate quality evidence suggests no difference between standard and proposed treatments for outcomes assessed
Benefits or desired effects:	Amoxicillin may be also effective for treatment of severe pneumonia, potentially simplifying treatment by reducing severity classes to two.
Risks or undesired effects	Potential for increased bacterial resistance to amoxicillin with widespread use. Reduced options for second line treatment in case of treatment failure - a whole new class of antibiotic might have to be provided as second line treatment.
Values and preferences:	Cotrimoxazole is formulated as a commonly used tablet in adults too that can be divided for children if pediatric formulations are missing - the same is not true for amoxicillin that is often distributed as capsules if syrups are not available
Costs:	Amoxicillin is more costly than cotrimoxazole (US$ 0.12 and US$ 0.21 for cotrimoxazole and amoxicillin syrups respectively for a course appropriate for a child weighing approximately 10 kg [KEMSA* July 2009])
Feasibility	Both drugs are widely available and in use

^† Clinical question: For Kenyan children aged 2 - 59 months who meet WHO criteria for non-severe pneumonia, should cotrimoxazole be replaced by amoxicillin?

¹ Indirectness of population (studies conducted in Pakistan)

² Indirectness of population (study conducted in India)

³ Although studies failed to show a difference between the two treatments, narrow confidence intervals and power calculation described by authors supported decision against downgrading the quality for imprecision.

KEMSA - Kenya Medical Supplies Agency