Comparing the impact of different regimens on long-term prognosis Background Moderate differences in efficacy between adjuvant chemotherapy regimens for breast cancer are acceptable and can influence treatment choice. We sought to understand these differences. Methods We conducted an independent study raw data meta-analysis of randomized trials comparing: any regimen based on paclitaxel plus an anthracycline antibiotic versus a regimen of the same drug, more drugs, or chemotherapy without paclitaxel (n=44,000); a regimen based on an anthracycline antibiotic versus a regimen based on another anthracycline antibiotic (n=7,000) or a regimen of cyclophosphamide, methotrexate, and fluorouracil, and combination chemotherapy versus chemotherapy-free regimens (n=32,000). Standard CMF, standard 4AC, CAF, and CEF were defined using predefined doses of these three drugs and anthracycline doxorubicin (A) and epirubicin (E). The study reported a log-rank test for breast cancer mortality ratio (RR). Results Over the course of the trial, breast cancer mortality was reduced with the addition of four separate courses of paclitaxel for extended treatment duration in a fixed regimen control group based on anthracycline antibiotics (RR 0.86, SE 0.04, significance level of two-sided test [2p]=0.0005). There was no significant difference between the two groups after the addition of four additional courses of other species of cytotoxic drugs (roughly double the dose of non-paclitaxel-based drugs) to offset the effect of four additional courses of paclitaxel in the control group during the trial. In all meta-analyses involving paclitaxel-based or anthracycline-based antibiotic regimens, age, lymph node status, tumor diameter or level of differentiation (moderate or poor; rarely well-differentiated), estrogen receptor status, or tamoxifen use had little effect on proportional risk reduction. Thus, certain regimens based on paclitaxel plus anthracycline antibiotics or based on higher cumulative doses of anthracycline antibiotics (without stem cells) reduce breast cancer mortality by an average of approximately one-third, an effect largely independent of the age (at least 70 years) and currently understood tumor characteristics of the patients we selected for these trials. The 10-year difference in overall mortality was consistent with the difference in breast cancer mortality regardless of the use of paclitaxel, anthracycline antibiotics, and other toxicants. Interpretation of results The 10-year benefit of a one-third reduction in breast cancer mortality depends on the absolute risk of not receiving chemotherapy (for estrogen receptor-positive disease, a risk that remains after receiving appropriate endocrine therapy). A low absolute risk implies a low absolute benefit, but information on tumor gene expression markers or quantitative immunohistochemistry that would help predict risk, chemotherapy sensitivity, or both, is lacking.