Retrospective analysis of the use of G-CSF and its impact on dose response for anthracycline plus taxane-based schedules in early breast cancer

Adjuvant chemotherapy has shown increased survival in early breast cancer. About 6 months of adjuvant anthracycline-based polychemotherapy schedules (FAC or FEC) reduced the annual breast cancer death rate by about 38 % for women younger than 50 years and by about 20 % for those diagnosed at 50–69 years of age [1] compared to non-anthracycline-containing schedules.

The benefit of addition of taxanes to anthracycline-based schedules was confirmed in a meta-analysis of thirteen studies including 22,903 patients [2]. The pooled hazard ratio (HR) estimated was 0.83 for disease-free survival (DFS) and 0.85 for overall survival (OS). The risk reduction was not influenced by the type of taxane, by estrogen receptor (ER) expression, by the number of axillary metastases (N1 to 3 vs. N4), or by the patient’s age/menopausal status.

Several trials have studied the relationship between dose intensity and survival. Muss et al. [3] observed in a retrospective study with 6,487 older and younger patients with lymph node-positive breast cancer that the dose level and the dose intensity of chemotherapy were significantly related to OS and DFS, independently of the patient’s age. The observed reduction in the hazard of failure of relapse was 22 % for patients who received more intensive chemotherapy in comparison with those who received less intensive chemotherapy (HR: 0.78; 95 % confidence interval (CI), 0.72–0.85).

In another recent retrospective study reported by our group, the dose response effect was identified as a crucial factor in the administration of anthracycline-based non-taxane schedules for the adjuvant treatment of early breast cancer [4]. Optimal delivery of the programmed chemotherapy improved OS and DFS. Thus, delays and/or reductions of chemotherapy should be avoided whenever possible to achieve the maximal benefit for the patient.

Breast cancer treatment is not exempt from adverse events, with neutropenia and febrile neutropenia being two of the most common and life-threatening side effects of adjuvant chemotherapy. Although the incidence of these adverse events may be minimized with dose reductions and delays in treatment, the most common strategy to maintain dose intensity is to administer granulocyte colony-stimulating factors (G-CSF) during chemotherapy administration [5]. According to the recommendations of the American Society of Clinical Oncology, the use of G-CSF as primary prophylaxis is justified in patients who are at high risk of febrile neutropenia based on age, medical history, disease characteristics, and the myelotoxicity risk of the chemotherapy regimen given. Moreover, it is recognized that G-CSF may allow a modest to moderate increase in the dose intensity of chemotherapy [6].

Studies with dose-dense taxane-containing regimens followed with G-CSF support have shown contradictory results on survival in early breast cancer in different studies [7, 8]. However, little is known about the impact of maintaining relative dose intensity (RDI) in conventional chemotherapy schedules including taxanes.

The aim of our study was to assess the impact on DFS and OS of dose-density of anthracyclines plus taxane-based schedules in the adjuvant setting of patients with early breast cancer. A secondary objective of this study was to evaluate the dose intensity achieved in a non-selected population treated with adjuvant anthracyclines and taxanes in whom G-CSF was administered at clinician’s discretion.

Our results showed that DFS and OS were not affected by delays in scheduled chemotherapy. Only those patients who received a reduced RDI had a lower probability of survival without recurrence of disease at 2 years. Also, patients who received reduced RDI had a lower probability of being alive at 4 and 5 years. However, the impact on DFS and OS is not consistent along years.

Low RDI is a common issue in clinical practice. In a population study of 1,243 community oncology practice in the USA with data from 20,799 early breast cancer patients showed that 36.5 % presented dose reductions of ≥15 days and 24.9 % had treatment delays ≥7 days resulting in 55.5 % of patients receiving RDI <85 %. Multivariate analysis identified the absence of primary prophylaxis with G-CSF as one of the independent predictors for low RDI [9]. A later study by the same group including 2,280 women with early breast cancer from community practice, low dose intensity in conventional schedules was reduced, probably due to the introduction of G-CSF, but the frequency of inappropriate chemotherapy delivery was 31, 24 and 26 % for dose delays, dose reductions and low RDI, respectively [10]. Strikingly, in our series the proportion of patients with low RDI was very small (2 %). This is probably as a consequence of a wide use of prophylactic G-CSF in our practice. This difference in sample size between patients with normal versus low RDI is the main limitation of our study and this issue precludes definitive conclusions.

In 2009, our group, Chirivella et al. [4] assessed how the suboptimal delivery of an anthracycline-based non-taxane adjuvant chemotherapy may impact the outcome of patients with early breast cancer. Delays of >2 cycles, ≥15 delayed days or RDI <85 % had a statistical impact on DFS. Moreover, when clinically relevant disease characteristics were controlled, such as the number of lymph nodes affected and the hormonal receptor status, the significance on DFS remained. However, in contrast with our study, prophylactic G-CSF was not widely used, as a result, the percentage of patients who had delayed chemotherapy administration or who received a reduced dose of chemotherapy was much higher than in our series. Patients who delayed chemotherapy >2 cycles or ≥15 days were 27 and 29 %, respectively, and 12 % received <85 % of RDI. In the current analysis, in which 36 % of patients received G-CSF and primary or secondary prophylaxis at clinician’s discretion, these percentages went down to 6, 2 and 2 %, respectively. Likely as a result of this, the impact observed in the outcome of patients in terms of DFS and OS was also smaller.

The role of G-CSF on maintaining an appropriate RDI in taxanes-containing schedules for breast cancer patients has been previously analyzed. Martin et al. [11] assessed the toxicity and the health-related quality of life of patients with breast cancer treated with anthracyclines, with or without taxanes, and with or without primary prophylactic G-CSF. In the group of patients treated with taxanes, 96 % of patients who received prophylaxis with G-CSF completed the six-cycle schedule, in comparison with 90 % of patients without G-CSF prophylaxis (p = 0.019). In comparison with patients who received secondary prophylaxis with G-CSF, primary prophylaxis was associated with a significant reduction in the number of cases of febrile neutropenia (26 vs. 7 %; p < 0.001) and grade 2/3 anemia (47 vs. 28 %; p < 0.001); and fewer patients required red cell transfusions (7 vs. 2 %; p < 0.010). Primary prophylaxis with G-CSF was also associated with a significant reduction in the incidence of asthenia, anorexia, myalgia, nail disorders and stomatitis compared with secondary prophylactic G-CSF. The reduction of these adverse events would facilitate compliance of the treatment as we have observed in our study.

It has been demonstrated that TAC schedules are more toxic than FAC schedules, not only with respect to neutropenic fever events, but also with respect to many extrahematological side-effects such as asthenia, stomatitis, diarrhea and myalgia [12]. In our study, those patients who received more aggressive treatment schedules, primary prophylaxis with G-CSF has been widely used in highly toxic regimens such as TAC. Overall, 78 % of patients treated with TAC and 46 % of patients treated with ET → X received primary prophylaxis with G-CSF. However, when less toxic schedules were given, only 2 % received primary prophylaxis and 8 % received secondary prophylaxis. Interestingly, none of the patients treated with EC → T received G-CSF as primary prophylaxis, but in 44 % of them it had to be introduced later on as secondary prophylaxis. Hence, it may be more beneficial for certain patients treated with EC → T to be given G-CSF from the beginning to avoid the occurrence of adverse events later on.

According to clinical guidelines, use of G-CSF is recommended to maintain chemotherapy if reduction of dose-density is associated to poor prognosis or if the risk of febrile neutropenia is high (≥20 %) [13]. The important rate of prophylactic G-CSF administration in our study, although in accordance to clinical guidelines, may be a relevant factor to explain the high rate of RDI >85 % observed in our study.

Strikingly, a Cochrane Database Systematic Review including eight randomized clinical trials assessing the effect of prophylactic G-CSF showed evidence of prevention of febrile neutropenia but failed to confirm any effect on maintaining dose density in breast cancer patients [14]. However, in most of the trials analyzed the chemotherapy regimens used had a risk of febrile neutropenia that was below the threshold at which current guidelines recommend routine primary prophylaxis. Moreover the small number of evaluable patients in some trials and the variability of definitions may strongly bias these findings.

Of note, at 5 and 10 years, the rate of DFS in our study was 88 and 80 %, respectively; and the rate of OS was 97 and 82 %, respectively. These data are outstanding taking into account, that more than a half of our patients were diagnosed with stage III breast cancer. Although it is impossible to identify which factors were responsible for these results, it seems reasonable to suppose that both the addition of taxanes to an anthracycline-based chemotherapy in the adjuvant setting, together with the optimal administration of chemotherapy, may have played an important role.

In fact, it cannot be discarded that reduced survival in many treatment schedules in the adjuvant setting may be due, at least in part, to a reduced compliance of the RDI [15]. Also, the high frequencies of adverse events that occur as a result of more aggressive therapies further hinder compliance with therapy. In this respect, the administration of prophylactic G-CSF at clinician’s discretion in those patients receiving more aggressive schedules could help to improve this compliance. On the other hand, it is important to note that nodes positivity and hormone receptors are key variables related with outcome in patients with breast cancer. However, in our study the number of patients with RDI <85 % is too small to conduct a multivariate analysis on this relevant issue.

In summary, our results are inconclusive for the primary endpoint. However, despite the previously mentioned limitations, our results suggest that, with the adequate use of G-CSF in patients with breast cancer treated with anthracyclines plus taxane-based schedules in the adjuvant setting, optimal chemotherapy administration could be achieved in almost all patients.