Gardin 2007.

Methods	Design: RCT with two arms: IDA versus DNR Recruitment period: November 1999 to March 2006 Median follow‐up: 34 months (range not stated)
Participants	Eligibility criteria: Inclusion criteria: ≧ 65 years newly diagnosed, previously untreated AML defined by the WHO criteria in the absence of CNS involvement Exclusion criteria: AML evolving from a prior myeloproliferative disorder according to the WHO classification M3 AML subtype prior exposure to chemotherapeutic agents and/or radiotherapy prior congestive heart failure requiring treatment and/or left ventricular systolic ejection fraction below the normal range a creatinine or bilirubin level more than twice the upper limit of normal, except if AML‐related a PS score of 4 uncontrolled severe infection. Patients randomised (n = 416): IDA arm: n = 207 DNR arm: n = 209 Median age: IDA arm: 72 years (range not stated) DNR arm: 72 years (range not stated) Gender (male, female): IDA arm: not stated DNR arm: not stated Country: France, 24 centres
Interventions	IDA arm: IA regimen, 1 cycle IDA: 9 mg/m²/d, days 1 to 4, IV Ara‐C: 200 mg/m²/d, days 1 to 7, IV DNR arm: DA regimen, 1 cycle DNR: 45 mg/m²/d, days 1 to 4, IV Ara‐C: 200 mg/m²/d, days 1 to 7, IV
Outcomes	Outcomes and time‐points from the study that are considered in the review: reported: OS, CR, death on induction therapy not reported: DFS, relapse, AEs, quality of life
Notes	Published as a journal article Funded in part by Assistance Publique‐Hopitaux de Paris and Programme Hospitalier de Recherche Clinique The authors declared no potential conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were randomized (...) to" Comment: the study probably had an adequate sequence generation
Allocation concealment (selection bias)	Unclear risk	No information provided
Blinding (performance bias and detection bias) Overall survival	Low risk	Comment: the review authors judge that the outcome OS is unlikely to be influenced by whether or not blinding
Blinding (performance bias and detection bias) All other outcomes	High risk	Quote (from registered protocol): "Open Label" Comment: patient and physician unblinded
Incomplete outcome data (attrition bias) OS and DFS	Low risk	Comment: all randomised patients were assessed in the analyses
Selective reporting (reporting bias)	Low risk	Comment: protocol is available (ClinicalTrials.gov: NCT00363025) Pre‐defined outcomes (relevant for the review) that were reported: OS CR death on induction therapy Pre‐defined outcomes (relevant for the review) that were not reported: none
Other bias	Unclear risk	No information provided