The review concluded that taxane-based adjuvant chemotherapy could reduce the risk of cancer recurrence and death in patients with early or operable breast cancer but drug-related toxicities should be considered. This was a generally well-conducted review and the authors' conclusion is likely to be reliable.

To determine the efficacy and safety of adjuvant chemotherapy, with or without taxanes, in the treatment of patients with early-stage or operable breast cancer.

PubMed, EMBASE and Cochrane Central Register of Controlled Trials (CENTRAL) were searched from January 1980 and April 2011. There were no restrictions on language and publication status. Search terms were reported. ClinicalTrials.gov and proceedings from American Society of Clinical Oncology annual scientific meetings and San Antonio Breast Cancer Symposium (1995 to 2011) were searched. Reference lists from relevant articles were checked.

High quality randomised controlled trials (RCTs) that compared a taxane-based adjuvant chemotherapy with a taxane-free adjuvant chemotherapy in patients with early or operable breast cancer were eligible for inclusion. Reported primary outcomes were required to include either disease-free survival or overall survival. Trials that included locally advanced breast cancer or where taxane was not the target drug or where both arms included a taxane were excluded.

The primary outcome of interest in this review was disease-free survival. This was defined as time from randomisation to any recurrence of breast cancer (local or distant), new primary breast cancer, a second cancer or death.

Taxanes included docetaxel and paclitaxel. Other chemotherapy drugs included epirubicin, doxorubicin, cyclophosphamide, fluorouracil and methotrexate. Treatment regimens varied substantially across trials; details were reported in an online supplement. Most trials reported a sequential regimen; five trials conducted concurrent regimens. One trial recruited only patients with node-negative breast cancer and another only patients with tumour size of 2cm or more. Median age of participants ranged from 48 to 57 years. Median follow-up ranged from 35 to 102 months.

Two reviewers independently selected trials for inclusion in the review.

Two reviewers independently assessed the quality of the included trials using the five-point Jadad scale. This assessment was checked independently by two additional reviewers.

Two investigators independently extracted data using a standardised data recording form. Two additional reviewers independently checked the extracted data. Hazard ratios (HRs) were derived for survival outcomes and odds ratios (ORs) were extracted for toxicity outcomes.

Pooled odds ratios and hazard ratios and associated 95% confidence intervals (CIs) were calculated using both fixed-effect and random-effects models (estimates from the random-effects model were reported). Heterogeneity was investigated using the Χ² test and the Ι² statistic (this was considered significant where p<0.10 for the Χ² test or Ι²>50%).

Planned subgroup analyses included: node status, drug dosage, schedule, observation period, menopausal status, hormone receptor status and tumour size. Interaction tests between the estimates were performed. Adverse events were analysed as drug-related toxicities (World Health Organisation grades ≥3). Sensitivity analyses excluded low-scoring trials (Jadad score ≤2) for hazard ratio estimates. Publication bias was evaluated with a visual examination of funnel plots and the Begg-Mazumdar test.

Nineteen open label phase III RCTs (30,698 women) were included. Fourteen trials had a Jadad score of 3 and five trials had a score of 2.

A significant improvement in disease-free survival (HR 0.83, 95% CI 0.79 to 0.88; 19 RCTs, 8,426 events) and overall survival (0.83, 95% CI 0.77 to 0.90; 17 RCTs) was found with taxane-based adjuvant chemotherapy compared with taxane-free adjuvant chemotherapy. There was no evidence of significant statistical heterogeneity.

A significantly lower risk was also found in concurrent (five RCTs) and sequential regimens (14 RCTs) in the taxane-based treatment arm for disease-free survival. Other subgroup analyses according to menopausal status (six RCTs), oestrogen receptor status (11 RCTs), tumour sizes less than 2cm (four RCTs) and 2cm or more (five RCTs) and node status (node-negative patients, docetaxel-based treatment only) (four RCTs) also found a statistically significant benefit in favour of taxane-based adjuvant chemotherapy. The exceptions were for HER-2 status (positive and negative) and paclitaxel (every two weeks) where no significant between-group differences were found.

A significant increase in toxicity was associated with taxane-based adjuvant chemotherapy for neutropenia, febrile neutropenia, fatigue, diarrhoea, stomatitis and oedema; significant heterogeneity was found in these analyses.

There was no evidence of significant publication bias.

Taxane-based adjuvant chemotherapy could reduce the risk of cancer recurrence and death in patients with early or operable breast cancer but drug-related toxicities should be considered. There was some evidence that docetaxel-based therapy was superior to taxane-free therapy for high risk node-negative breast cancer.

The review question was supported by largely well-defined inclusion criteria. Several relevant sources were searched without restrictions on language or publication status. The methods used in the review process were likely to minimise the possibility of bias. The quality of the included trials was assessed using a standardised scale, although this scale did not include allocation concealment. The meta-analyses seemed appropriate. Prespecified subgroup analyses and sensitivity analyses were performed. Substantial heterogeneity was found for the drug-related toxicity analyses.

The authors' cautious conclusion is likely to be reliable.

Practice: The authors stated that the benefit of taxanes should be balanced against their toxicity.

Research: The authors stated that further well-designed RCTs with varying drug schedules for both operable and node negative breast cancer were warranted. Additional data on quality of life should be provided in future analyses.

National Nature Science Foundation of China; grant from the leading talents of science in Shanghai 2010; Program of Shanghai Subject Chief Scientist.