ACOSOG Z1041 2013.

Methods	Accrual: 15 September 2007 to 15 December 2011 Open‐label, phase III randomised controlled trial Neoadjuvant therapy Multicentre (36 centres) across the USA and Puerto Rico Follow‐up: yearly for a maximum of 5 years
Participants	Age: 25–39 years: 21% (both arms); 40–49 years: approximately 33% (both arms); 50–59 years: 38.7% (arm 1) and 35.5% (arm 2); 60–69 years: 10.6% (arm 1) and 14.5% (arm 2); > 70 years: approximately 3% (both arms) HER2‐positive disease Stage described as clinical T stage and N stage: generally 7% T1, 55% T2, 29% T3, 9% T4; 36% N0, 51–57% N1, 5% N2, 0.7% to 8% N3 ER‐positive, PR‐positive: 41.5% (arm 1), 35.5% (arm 2); ER‐positive, PR‐negative: 16.2% (arm 1), 22.5% (arm 2); ER‐negative and PR‐negative: 40.8% (arm 1) and 39.1% (arm 2) Excluded: any current breast cancer treatment except hormonal therapy (taken for up to 28 days after diagnosis but had to be stopped before study registration)
Interventions	Arm 1 ('concurrent' in the trial publication): paclitaxel (80 mg/m²) and trastuzumab 4 mg/kg first dose (2 mg/kg after days 1, 8, 15) once a week for 12 weeks followed by fluorouracil (500 mg/m²), epirubicin (75 mg/m²), cyclophosphamide (500 mg/m²) every 3 weeks for 4 cycles and trastuzumab (2 mg/kg after a 4 mg/kg loading dose) once a week for 12 weeks Arm 2 ('sequential' in the trial publication): fluorouracil (500 mg/m²), epirubicin (75 mg/m²), cyclophosphamide (500 mg/m²) every 3 weeks for 4 cycles followed by paclitaxel (80 mg/m²) and trastuzumab (2 mg/kg after a 4 mg/kg loading dose) once a week for 12 weeks
Outcomes	Primary outcome pCR in the breast Secondary outcomes pCR in the breast and axillary nodes Cardiac outcomes Adverse events OS Progression‐free survival
Notes	Clinical trial registration number: NCT00430001 Funding considerations: supported by a National Cancer Institute grant, National Cancer Institute to the Alliance for Clinical Trials in Oncology, and Alliance Statistics and Data Center. Data quality and analysis: completed by Alliance Statistics and Data Center.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were randomly assigned to treatment (1:1) with a biased coin minimisation algorithm so that marginal distributions of stratification factors would be similar in each treatment group" (p. 1318).
Allocation concealment (selection bias)	Low risk	Central allocation. Quote: "...patients were, unstratified, randomly assigned centrally by the Danish Breast Cancer Cooperative Group Secretariat."
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "Neither patients nor investigators, except for a cardiac review panel, were masked to treatment assignment." Comment: participants in both groups received the same drugs though in different order. Therefore, it was unlikely that this would lead to a material bias in participants' and physicians' behaviours even when unblinded.
Blinding of outcome assessment (detection bias) Overall survival	Low risk	Outcome collected but not yet reported. Unblinding unlikely to have had an impact on outcome assessment.
Blinding of outcome assessment (detection bias) DFS	Low risk	Unblinding unlikely to have had an impact on outcome assessment for this review question where trial participants received both treatment regimens. DFS was determined each year for a maximum of 5 years (data not provided).
Blinding of outcome assessment (detection bias) Toxicity and treatment adherence	Unclear risk	Laboratory tests (blood counts) repeated before each cycle. LVEF measured at completion of first 12‐week regimen and second 12‐week regimen. Cardiac review panel reviewed multigated acquisition scan and echocardiography results for all participants each month. Judged at unclear risk of bias because assessors were not blinded to treatment allocation for assessment of other toxicities. Dose reductions/delays were based on toxicities. Knowledge of treatment allocation may have had some influence on physicians' assessments.
Blinding of outcome assessment (detection bias) Neoadjuvant studies only: pCR	Low risk	Before each cycle of treatment, assessments for tumour size were completed. Imaging and tumour evaluation was done every 3 months. Mammogram of ipsilateral breast taken at completion of neoadjuvant chemotherapy. pCR was viewed to be an objective outcome and a review by a pathologist on whether pCR had been achieved or not was unlikely to be influenced by unblinding.
Incomplete outcome data (attrition bias) All outcomes	Low risk	138/140 participants in the comparator group and 142/142 participants in intervention group proceeded to treatment. 4/142 participants in the intervention arm discontinued treatment (2 refused, 1 allergic reaction, 1 disease progression) and 8/138 participants discontinued in the comparator arm (4 refused treatment, 2 based on physicians' discretion, 1 second primary cancer and 1 death unrelated to treatment). All participants who began study treatment were included in the efficacy analysis.
Selective reporting (reporting bias)	Low risk	Prespecified outcomes in Clinical Trials.gov record (NCT00430001) and the Methods section of the trial publication are the same. Most outcomes except OS and DFS (due to immature data), were reported in the results section. Some additional toxicity data were reported in the trial publication that were not included in the trial registry record.
Other bias	Low risk	None identified