Ben Salah 2009

Methods	Study design: randomised, prospective, double‐blind trial Setting/location: Sidi Bouzid,Tunisia, and Paris, France Study period: 22 months Sample size calculation: the protocol calculated a sample size of 50 participants per group with 80 percent power and a Type I error rate of 5 percent to detect a 30% difference in the proportion of participants achieving CCR, assuming a CCR proportion of 35% in the vehicle group and 65 percent in WR279, 396 participants, with a 5% expected rate of loss to follow‐up
Participants	Type of Leishmania: L major, L tropica, L infantum Inclusion criteria: aged 5‐75 years, presence of parasitologically confirmed CL, lesions that were primarily ulcerative (i.e. not purely verrucous or nodular) and measured ≥ 1 cm² and ≤ 5 cm². Exclusion criteria: history of known or suspected hypersensitivity or idiosyncratic reactions to aminoglycoside; previous use of antileishmanial drugs (within 3 months) or nephrotoxic or ototoxic drugs; prior diagnosis of leishmaniasis; more than 5 lesions, or a lesion in the face that in the opinion of the attending dermatologist could potentially cause significant disfigurement; significant medical problems as determined by history or laboratory studies; breastfeeding and pregnancy N randomised: 92. WR279,396: 50; vehicle: 42 Withdrawals (n = 2): WR279,396: 1; vehicle: 1 N assessed (%): 90 (95.7). WR279,396: 49 (94.2), vehicle: 41 (97.6) Age < 18 years ‐ n (%): WR278,396: 47 (94), vehicle: 33 (79) Sex: male: 54, female: 38 Severity of illness: N lesions: 1: 54, 2: 16, 3: 13, 4 or 5: 9 Total lesion area (median, IQR): WR279,396: 128 (85 to 223), vehicle: 154 (70 to 264) Index lesion area (median, IQR): WR279,396: 92 (55 to 141), vehicle: 115 (50 to 172) Leishmania spp ‐ n (%): L major: WR279,396: 32 (64), vehicle: 24 (57). L Infantum: WR279,396: 1 (2), placebo: 0 (0). L tropica: WR279,396: 1 (2), vehicle: 0 (0). Unidentified: WR279,396: 16 (32), placebo: 18 (43)
Interventions	Type of interventions: Group 1: WR279,396 is an off‐white to yellowish, thick cream containing 15% (w/w) paromomycin‐sulphate (Farmitalia) and 0.5% (w/w) gentamicin‐sulphate (Schering) as active components. Group 2: the vehicle consisted of a cream without the active components and trace amounts of colorings agents to match the appearance and maintain the blind. Duration of intervention: 20 days
Outcomes	Clinical cure: defined as complete re‐epithelialisation (i.e. length 6 width of ulceration = 060) of the index lesion by day 50 or a 50% re‐epithelialisation by day 50 followed by complete re‐epithelialisation on or before day 100 with no relapse ever having occurred from day 50 through day 180. Relapse was defined as an increase in the area of ulceration relative to the previous measurement. Participants who did not complete the 180‐day period of observation were considered to have failed to achieve complete clinical response (CCR) because relapse could not be fully assessed. Complete clinical response at the index lesion according to baseline characteristics: influence of baseline factors on effect of treatment. Time to healing: time course of complete re‐epithelialisation measured at 20, 50, 100, 180 days since start of treatment. Adverse effects: immediate: observed within 30 min of application. Delayed: observed just prior to next application
Notes	Study funding sources: the Office of the Surgeon General (OTSG), Chief, Human Subjects Protection Division, U.S. Army MRMC, Fort Detrick, MD 21702‐5012. IND50,098 HSRRB Protocol#1791. Co‐sponsor: Institute Pasteur, Rue du Dr. Roux, Paris, France Possible conflicts of interest: MG has no financial competing interests
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "A sequence of genuine random numbers for the randomisation procedure was obtained from the 'fourmilab.ch/hotbits' website by a member of the Department of Chemical Information, Walter Reed Army Institute of Research, Silver Spring, Maryland and purged of duplicates. The random numbers are generated by a process which takes advantage of the inherent uncertainty in the quantum mechanical laws of nature. Specifically, they are generated by timing successive pairs of radioactive decays detected by a Geiger‐Muëller tube interfaced to a computer. This process is better than the pseudo‐random number algorithms typically used in computer programs. The randomisation of the study drugs was done by an independent group, Fischer BioServices, Rockville, Maryland a contractor to The U.S. Army Medical Research Acquisition Activity (USAMRAA), Ft. Detrick, Maryland."
Allocation concealment (selection bias)	Low risk	The randomisation of the study drugs was done by an independent group; however, allocation concealment is not fully described.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	The vehicle lacked the active components and trace amounts of colorings agents to match the appearance and maintain the blind.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"Investigators, who were blinded to whether participants received WR279,396 or placebo‐vehicle, evaluated lesions for clinical response on D20 (i.e. the end of the treatment period), D50 (i.e., 30 days after the conclusion of treatment), D100, and D180". Investigators measured all lesions in 2 perpendicular directions and took photographs at the following time points: prior to therapy, at the end of therapy
Incomplete outcome data (attrition bias) All outcomes	Low risk	49 of 50 participants randomised to WR279,396 and 41 of 42 participants randomised to placebo‐vehicle completed the study. With one exception, applications of study drugs were conducted according to the protocol
Selective reporting (reporting bias)	Low risk	ClinicalTrials.gov NCT00703924. Pre‐specified outcomes were reported
Other bias	Low risk	Other items assessed correctly reported