Increasing the dose intensity of chemotherapy by more frequent administration or sequential scheduling: a patient-level meta-analysis of 37 298 women with early breast cancer in 26 randomised trials

Abstract

Background: Increasing the dose intensity of cytotoxic therapy by shortening the intervals between cycles, or by giving individual drugs sequentially at full dose rather than in lower-dose concurrent treatment schedules, might enhance efficacy.

Methods: To clarify the relative benefits and risks of dose-intense and standard-schedule chemotherapy in early breast cancer, we did an individual patient-level meta-analysis of trials comparing 2-weekly versus standard 3-weekly schedules, and of trials comparing sequential versus concurrent administration of anthracycline and taxane chemotherapy. The primary outcomes were recurrence and breast cancer mortality. Standard intention-to-treat log-rank analyses, stratified by age, nodal status, and trial, yielded dose-intense versus standard-schedule first-event rate ratios (RRs).

Findings: Individual patient data were provided for 26 of 33 relevant trials identified, comprising 37 298 (93%) of 40 070 women randomised. Most women were aged younger than 70 years and had node-positive disease. Total cytotoxic drug usage was broadly comparable in the two treatment arms; colony-stimulating factor was generally used in the more dose-intense arm. Combining data from all 26 trials, fewer breast cancer recurrences were seen with dose-intense than with standard-schedule chemotherapy (10-year recurrence risk 28·0% vs 31·4%; RR 0·86, 95% CI 0·82-0·89; p<0·0001). 10-year breast cancer mortality was similarly reduced (18·9% vs 21·3%; RR 0·87, 95% CI 0·83-0·92; p<0·0001), as was all-cause mortality (22·1% vs 24·8%; RR 0·87, 95% CI 0·83-0·91; p<0·0001). Death without recurrence was, if anything, lower with dose-intense than with standard-schedule chemotherapy (10-year risk 4·1% vs 4·6%; RR 0·88, 95% CI 0·78-0·99; p=0·034). Recurrence reductions were similar in the seven trials (n=10 004) that compared 2-weekly chemotherapy with the same chemotherapy given 3-weekly (10-year risk 24·0% vs 28·3%; RR 0·83, 95% CI 0·76-0·91; p<0·0001), in the six trials (n=11 028) of sequential versus concurrent anthracycline plus taxane chemotherapy (28·1% vs 31·3%; RR 0·87, 95% CI 0·80-0·94; p=0·0006), and in the six trials (n=6532) testing both shorter intervals and sequential administration (30·4% vs 35·0%; RR 0·82, 95% CI 0·74-0·90; p<0·0001). The proportional reductions in recurrence with dose-intense chemotherapy were similar and highly significant (p<0·0001) in oestrogen receptor (ER)-positive and ER-negative disease and did not differ significantly by other patient or tumour characteristics.

Interpretation: Increasing the dose intensity of adjuvant chemotherapy by shortening the interval between treatment cycles, or by giving individual drugs sequentially rather than giving the same drugs concurrently, moderately reduces the 10-year risk of recurrence and death from breast cancer without increasing mortality from other causes.

Funding: Cancer Research UK, Medical Research Council.

Figure 1

Recurrence in trials testing dose-intense strategies versus standard scheduling Taxanes: D=docetaxel. P=paclitaxel. Anthracyclines: A=doxorubicin. E=epirubicin. Other: C=cyclophosphamide. Cap=capecitabine. F=fluorouracil. M=methotrexate. Bev=bevacizumab (mg/kg). AC=A60C600. EC=E90C600. ADC=A50D75C500.CMF=classical CMFd1d8. Chemotherapy doses are in mg/m². q2=2-weekly. q3=3-weekly. q4=4-weekly. × 14=days 1–14 orally. × 2=day 1, day 8. 2p=two-sided p value. *For balance, the 75 control patients in the two 3-way trials count twice in subtotal (C) and in final total of events/patients. NSABP B-38 trial assumes a 2:1 dose equivalence ratio for P to D. †Primary prophylaxis with colony-stimulating growth factors. ‡Pre-operative chemotherapy: patients in these trials were analysed as having unknown nodal status. §Seven trials with no data do not contribute to subtotals or to the overall total. Semicolon indicates treatment sequence. First column shows study name and year started. χ² tests in section headers are for heterogeneity between trials.

Figure 2

Dose-dense (2-weekly) chemotherapy versus the same chemotherapy given 3-weekly 10-year cumulative risk of any recurrence (A), breast cancer mortality (B), death without recurrence (C), and all-cause mortality (D). Of the 10 004 women, 71% are N+. RR=rate ratio.

Figure 3

10-year risk of any recurrence (left) and breast cancer death (right) for all trials (confounded and unconfounded) (A) Dose-dense (2-weekly) versus standard schedule (3-weekly) chemotherapy. Of the 15 512 women, 65% are N+. (B) Sequential versus concurrent chemotherapy (both 3-weekly). Of the 11 028 women, 91% are N+. (C) Sequential (2-weekly) versus concurrent (3-weekly) therapy. Of the 6532 women, 90% are N+.

Figure 4

Pooled analysis of all dose-intensification trials 10-year cumulative risk of any recurrence (A), breast cancer mortality (B), death without recurrence (C), all-cause mortality (D), for dose-intense versus standard schedule arm. Of the 37 298 women, 77% are N+. RR=rate ratio.

Figure 5

Subset analyses of pooled data from all dose-intensification trials; any first recurrence (including locoregional recurrence, distant recurrence, and new contralateral disease)* ER=oestrogen receptor. PR=progesterone receptor. NS=not significant. *For balance, the 75 control patients in 3-way trials count twice in numbers of events/patients. Unless otherwise specified, patients whose category is unknown are omitted. Unless specified as for trend, χ² tests are for heterogeneity.

Figure 6

10-year recurrence (A) and breast cancer mortality (B) by oestrogen receptor status Pooled data from all trials of dose intense versus standard schedule chemotherapy. Of the 10 900 women who are oestrogen receptor (ER)-negative, 66% are N+; and of the 25 029 women who are ER-positive, 84% are N+.