Sykes 2009

Methods	Trial design: Single‐centre, open‐label, randomized, controlled trial with two treatment regimens. Follow up: Evaluation visits and blood smears, filter paper sample and haemoglobin tests on day 0 to 2, 7, 14, 28 and 42, or on any day if the child was unwell. Adverse event monitoring: Mothers or guardians were asked about possible adverse effects, which were classified according to WHO criteria. Follow‐up visits on day 0 to 2, 7,14, 28, and 42, or if child unwell. Blood count and liver function tests on day 0 and 14. Number available for adverse events not reported.
Participants	Number enrolled: 261 Number available for analysis: 236 Inclusion criteria:  Children aged six months to 59 months with fever (axillary temperature > 37.5 ºC or a history of fever within the previous 48 hours); positive rapid diagnosis test (Parahit); Plasmodium falciparum asexual parasite density on blood smear of > 2000 and < 200,000 parasites/µl of blood; haemoglobin > 7 g/dl; living in the catchment area; willingness to attend study follow‐up visits and to provide informed consent. Exclusion criteria: Children with symptoms suggestive of severe febrile disease (based on the modified WHO criteria for severe malaria); unable to tolerate drugs orally; inability to feed; an obvious alternative cause of fever; use of an effective antimalarial drug in the previous seven days; mixed plasmodial infection; known hypersensitivity to a study drug. Age range: Not reported. Median age: 30 months (interquartal range 18 to 42) in arm 1, and 27 months (interquartal range 16 to 42) in arm 2 Proportion male:  124/261 (47.5%) (estimated from percentages in table 1 and enrolment in figure 1) Pregnant women:  Not applicable
Interventions	arm 1) Azithromycin 20 mg/kg body weight, and artesunate 4 mg/kg once a day for three days arm 2) Artemether‐lumefantrine (fixed dose combination tablets 20 mg artemether and 120 mg lumefantrine), one tablet two times per day for those with body weight < 15 kg and 2 tablets two times per day for a body weight > 15 kg for three days Supervision of intake: Complete observed intake for arm 1, only observed intake for morning dose for arm 2 Other (symptomatic) medication: Not reported
Outcomes	Outcomes used in this review: Total failure by day 28 (primary outcome). PCR‐adjusted failure. Adverse events
Notes	Location: Muheza Designated District Hospital, Muheza, Tanzania Date: June 2008 to February 2009 Malaria endemicity: Perennial, stable, hyper‐holoendemic Drug resistance: Reported to be > 70% for sulfadoxine‐pyrimethamine, chloroquine, and amodiaquine in the district Method of determining parasite density: Giemsa‐stained thick and thin smears. Parasite density calculated from thick smears, assuming a white blood cell count of 8000 cells/µl. No definition for negative smear. Read by two readers, discordant results solved by third reader. Drug source: Not reported Source of funding: The Gates Malaria Partnership Other notes: None
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	"Randomization occurred in blocks of random sizes using Stata software"
Allocation concealment?	Low risk	"Slips with study allocation were placed in sealed opaque envelopes and were opened in front of parents or guardians."
Blinding? All outcomes	High risk	Open‐label study; microscopists were blinded
Incomplete outcome data addressed? All outcomes	Low risk	Low loss of participants and equally distributed across arms: 129 and 132 randomised in respective groups, 119 (92.2%) and 120 (90.9%) results at day 28. Of the 25 participants who did not complete, 16 withdrew (seven in arm 1 and nine in arm 2), and nine were lost to follow‐up (six in arm 1 and three in arm 2)
Free of selective reporting?	High risk	Study stopped after planned interim analysis of DSMB. Drug intake in arm 1 completely observed; only half of drug intake in arm 2 observed. However, results are in the opposite direction of the potential effect of this bias. No follow‐up visit on day 21.
Free of other bias?	Unclear risk	Not sufficient information