Hassanjani Roushan 2010.

Methods	Randomized controlled trial Duration: April 2005 to September 2008
Participants	Number: 164 patients enrolled [2 patients in DS and 1 in DG groups excluded due to spondylitis] Inclusion criteria: Patients > 10 yrs. The diagnosis of brucellosis was made by using standard tube agglutination (STA) titre >= 1:320 and 2‐mercaptoethanol (2ME) titre >= 1:160, together with compatible clinical findings (fever, sweating, arthralgias, peripheral arthritis, sacroiliitis, and epididymo‐orchitis). Exclusion criteria: age of <10 years, spondylitis, endocarditis, neurobrucellosis, pregnancy, and the history of >2 days of antibiotic treatment before enrolment.
Interventions	1‐doxycycline 100 mg orally twice daily for 45 days, plus streptomycin 1g intramuscularly daily for 14 days 2‐doxycycline 100 mg orally twice daily for 8 weeks, plus gentamicin 5 mg/kg/day intramuscularly for 5 days
Outcomes	1‐relapse 2‐therapeutic failure 3‐adverse drug reactions
Notes	Location: Babol, Northern Iran Setting: inpatient and outpatient Source of funding: the Infectious Diseases Research Center and Tropical Diseases of Babol Medical University.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Shuffling envelopes
Allocation concealment (selection bias)	Low risk	"For each patient, a record was drawn and the therapy regimen, which was noted on it, was administered. We could not predict the therapy regimen for any patient."
Blinding (performance bias and detection bias) All outcomes	High risk
Incomplete outcome data (attrition bias) All outcomes	High risk	82 patients were randomized to each group. Three were excluded due to the diagnosis of spondylitis. Five patients in DS and 2 in DG lost to follow‐ups. The proportion of missing outcomes compared with observed events was considerable.
Selective reporting (reporting bias)	Low risk
Other bias	Low risk