Seventeen RCTs (n=5,689) were included.
In patients with no prior anthracycline exposure there were 7 RCTs of paclitaxel (n=2,954) and 4 RCTs of docetaxel (n=1,335). In patients with previous anthracycline exposure there were 2 RCTs of paclitaxel (n=123) and 4 RCTs of docetaxel (n=1,277).
Patients with no previous anthracycline exposure. Paclitaxel (7 RCTs).
Three RCTs assessed paclitaxel as a single agent (versus three different regimens) and none demonstrated any statistically significant benefits in terms of survival. One found significant benefits for time to progression and overall response for the control arm of doxorubicin, but there were more adverse events with doxorubicin. Five RCTs provided results for paclitaxel in combination with doxorubicin or epirubicin versus four different comparator regimens (one RCT had three arms, so was included in both the single and combined agent comparisons). Significant benefits of paclitaxel combined with doxorubicin for time to progression and response rates were found in 2 trials (versus fluorouracil-doxorubicin-cyclophosphamide (FAC) in one study and versus paclitaxel alone and doxorubicin alone in the other) and for survival in one trial (versus FAC). Higher rates of neutropenia and neurotoxicity were seen for the paclitaxel-doxorubicin combination.
Docetaxel (4 RCTs).
One RCT of docetaxel alone and 2 RCTs of docetaxel combined with other agents (doxorubicin, epirubicin or doxorubicin-cyclophosphamide) found significantly higher overall response rates for docetaxel regimens compared with doxorubicin alone, doxorubicin-cyclophosphamide and FAC. One study reported no significant differences between docetaxel and doxorubicin for time to progression or survival; the other 2 RCTs did not report progression or survival. One trial reported significantly fewer serious adverse events with docetaxel compared with doxorubicin; in general, more patients treated with docetaxel and an anthracycline experienced neutropenia, but the trials did not report if these results were statistically significant or not.
Patients with previous anthracycline exposure.
Paclitaxel. Of the 2 RCTs assessing paclitaxel as a single agent, one was stopped early and did not demonstrate any significant results. The other was a phase II trial which showed a significant benefit for paclitaxel for time to progression compared with mitomycin.
Docetaxel.
Three RCTs assessed docetaxel as a single agent, although one was still ongoing at the time of this review. The other 2 trials both reported significant benefits of docetaxel with respect to overall response and time to progression (versus mitomycin-vinblastine and versus methotrexate-5FU), with one also showing a significant improvement in survival with docetaxel versus mitomycin-vinblastine. Higher rates of adverse events were seen with docetaxel, although more patients remained on docetaxel compared with control treatment. One RCT assessed docetaxel and capecitabine in combination and reported significant benefits in response rate, time to progression and survival for the combination compared with docetaxel alone.
Docetaxel compared with paclitaxel.
No data directly comparing docetaxel with paclitaxel in either patient group were available. From observed indirect comparisons there appears to be higher response rates with single-agent docetaxel compared with paclitaxel.
Quality of life.
Eight of the 9 studies that presented quality-of-life measures reported no statistically significant differences between treatments. The one RCT reporting significant treatment differences reported improvements in scores with the experimental intervention for some domains, but a deterioration in scores for other domains.