Evidence generated by controlled clinical trials over the past 4 decades indicated that combination chemotherapy produced superior results to single-agent cytotoxic therapy. Response rates and times to progression were superior with combination chemotherapy, and survival was also favorably altered by this approach. This paradigm has been recently challenged on the basis of the Norton-Simon hypothesis and by the development of more effective, new cytotoxic agents, specifically the taxanes.
A few pre-1990 studies suggested that single-agent doxorubicin provided results equivalent to non–anthracycline-containing combinations in terms of response rates and times to progression. Other trials suggested that sequential utilization of full-dose single-agent therapy provided equivalent survival to that obtained by simultaneous combinations, even though the combinations resulted in higher response rates. With the introduction of the taxanes, the drugs considered today as the most effective agents against breast cancer, the Norton-Simon hypothesis received additional fuel for investigation.
The abstract by Alba et al (abstract #349) describes the limited experience of the authors with sequential administration of full-dose single-agent doxorubicin and docetaxel (Taxotere) as first-line chemotherapy for metastatic breast cancer. The results suggest that this strategy produces response rates and times to progression similar to those expected from standard doxorubicin- or taxane-containing combination therapy. The toxicity, while manageable, was similar to that expected after combination chemotherapy. While the results of sequential and simultaneous administration of available agents might provide similar survival rates, alternative directions in research must be pursued to improve the results of treatment.
A second abstract, by Paridaens and collaborators (abstract #1259), describes the early results of a randomized phase II trial comparing alternating and sequential administration of doxorubicin and docetaxel. These results show no difference at the time of this analysis, thus failing to support the Norton-Simon hypothesis; however, the sample size is too small to exclude more modest differences in outcome.