A retrospective analysis of nadir‐neutropenia directed pegylated granulocyte‐colony stimulating factor on febrile neutropenia rates in (neo)adjuvant breast cancer chemotherapy regimens

Abstract

Background

Pegfilgrastim, a pegylated granulocyte colony‐stimulating‐factor (GCSF), reduces chemotherapy morbidity and mortality in early stage breast cancer. The optimal approach to individual patient selection for GCSF is unknown, in particular whether secondary GCSF should be given after asymptomatic neutropenia, or only after febrile neutropenia (FN).

Aims

To determine if preplanned nadir blood counts and subsequent nadir‐neutropenia directed GCSF was effective to reduce rates of FN associated with (neo)adjuvant breast cancer chemotherapy. We also aimed to describe (neo)adjuvant chemotherapy and GCSF prescribing practices at our institution.

Methods

This was a retrospective electronic medical record review. The rate of FN with secondary GCSF after cycle 1 nadir‐neutropenia <1.0 × 10⁹ cells/L was compared with the rate of FN in patients who did not have cycle 1 nadir blood counts or secondary GCSF, and analyzed according to patient and treatment data.

Results

Between 11/4/2011 and 22/4/2018, 584 patients received (neo)adjuvant chemotherapy. Over all rates of FN were 17%, compared with 9% in patients who received primary GCSF. Rates of FN were highest in docetaxel/carboplatin/trastuzumab (TCH, 27%) and docetaxel/cyclophosphamide (TC, 27%). There were 268 patients (46%) who received primary GCSF and 238 patients (41%) who received secondary GCSF. Rates of FN did not differ between patients who received nadir‐neutropenia directed secondary GCSF (6/125, 5%, 95% CI 1.1‐8.6), and those who did not have nadir blood counts or secondary GCSF (0/49, 0%, 95% CI not calculable) (P = .1186). GCSF use was associated with lower rates of non‐FN hospital admissions. Patients ≥65 years were more likely to have FN and non‐FN admissions. Two of three treatment related deaths occurred due to FN.

Conclusions

In this population, nadir‐neutropenia directed secondary GCSF was not associated with reduced rates of FN. Observed rates of FN >20% in TC and TCH in routine clinical practice should guide primary GCSF use in accordance with international guidelines.

1. INTRODUCTION

(Neo)adjuvant breast cancer chemotherapy regimens are associated with high rates of febrile neutropenia (FN). This is a major cause of early, and potentially preventable, treatment related mortality and morbidity. ¹ It is also a cause of significant health care resource usage, ¹ as well as a limiting factor for chemotherapy dose intensity and cancer outcomes in patients being treated with curative intent. ²

Published rates of FN vary between regimens, from essentially 0% with weekly paclitaxel ³ , ⁴ ; to approximately 10% with anthracycline, ⁵ , ⁶ , ⁷ docetaxel/cyclophosphamide (TC), ⁵ , ⁶ docetaxel/carboplatin/trastuzumab (TCH), ⁸ and cyclophosphamide/methotrexate/fluorouracil (CMF) ⁹ ; and >20% with fluorouracil/epirubicin/cyclophosphamide‐docetaxel (FEC‐D), ⁷ and docetaxel/doxorubicin/cyclophosphamide (TAC). ¹⁰ , ¹¹ However, these rates differ from those described in routine clinical practice, with registry series often showing higher rates of FN, potentially due to inclusion of patients with comorbidities and reduced functional status. ¹² , ¹³ , ¹⁴ More than 50% of cases of FN in (neo)adjuvant breast cancer treatment occur during cycle 1. ¹⁵

Granulocyte colony‐stimulating‐factor (GCSF) stimulates production and differentiation of neutrophils from blood precursor cells to reduce the depth and duration of chemotherapy induced neutropenia, reducing rates of FN, FN related deaths, and other infective complications during myelosuppressive chemotherapy for solid tumors or lymphoma. ¹⁶ , ¹⁷ , ¹⁸ GCSF also improves chemotherapy relative dose intensity ¹⁸ , ¹⁹ and assists the delivery of dose dense regimens, ¹⁷ however, it may not be required before all dose dense treatments. ²⁰

GCSF is a low risk intervention most commonly associated with mild‐moderate bone pain. ²¹ There is however evidence of increased risk of secondary malignancies ²² and early pre‐clinical data suggestive that GCSF stimulates breast cancer progenitor cells. ²³

International guidelines clearly recommend the use of secondary GCSF following an episode of chemotherapy induced FN. ²¹ , ²⁴ , ²⁵ Guidelines also recommend the use of primary GCSF for patients with >20% risk FN, ²¹ , ²⁴ , ²⁵ however the optimal approach to individual patient selection is unknown, especially for patients receiving regimens traditionally considered to be of intermediate risk of FN (ie, risk 10‐20% FN). Rivera et al ²⁶ demonstrated that for patients receiving adjuvant fluorouracil/doxorubicin/cyclophosphamide for early breast cancer, a cycle 1 nadir neutrophil count <0.5 × 10⁹ cells/L taken on either day 8 or 15 was predictive of episodes of FN, neutropenia related chemotherapy dose reduction >15%, or chemotherapy dose delays >7 days. Thus, some clinicians perform a preplanned cycle 1 nadir neutrophil count and instigate secondary GCSF for neutropenia <1.0 × 10⁹ cells/L with the assumption this may identify at risk patients and reduce rates of FN and associated morbidity and mortality. We aimed to review GCSF prescribing practices and rates of FN in our institution and the impact of GCSF use following asymptomatic nadir‐neutropenia <1.0 × 10⁹ cells/L for standard (neo)adjuvant breast cancer regimens.

2. METHODS

A retrospective analysis was performed of patients treated with (neo)adjuvant chemotherapy for stage I‐III breast cancer between April 11, 2011 and April 22, 2018 at a tertiary cancer center and associated regional cancer center in Australia. Data were collected from patient electronic medical records in the context of hybrid paper‐electronic medical records. Approval was granted by a recognized human research ethics committee.

Neutropenia was defined as neutrophil count <1.0x10⁹ cells/L (ie, Common Terminology Criteria for Adverse Events grade 3/4 neutropenia). FN was defined as neutrophil count <1.0 × 10⁹ cells/L plus any temperature > 38.0°C, or neutrophil count <1.0 × 10⁹ cells/L plus other clinical evidence of sepsis such as clinically significant hypotension or positive microbiology as per local clinical practice. A nadir blood count was a full blood count performed with the intention of determining mid cycle blood counts during cycle 1 in a patient who did not receive primary GCSF that was unrelated to a hospital presentation or admission, however the specific timing was at clinician's discretion. A hospital admission was defined as any unplanned admission to hospital including outpatient blood transfusions. Emergency Department presentations that did not result in hospital admission and pre‐planned presentations for breast surgery or insertion of a port‐a‐cath were not considered hospital admissions.

The primary objective was to ascertain the rate of FN in patients who received secondary GCSF after cycle 1 nadir‐neutropenia and to compare it with the rate of FN in patients who did not have a cycle 1 nadir blood count or secondary GCSF. Other variables collected included demographics, chemotherapy regimen, tumor data, non‐FN hospital admissions and treatment related mortality.

Data were recorded in Microsoft Excel and analyzed using Microsoft Excel (Microsoft Office 365) and Stata 13.1 (IC, StataCorp, College Station, Texas). Demographics, treatment and tumor data are presented using descriptive statistics. Continuous data are presented using mean and SD. The primary objective was statistically compared using a two‐sample test of proportions. Lower‐order comparisons have not been tested formally and P‐values are not provided due to the risks inherent in multiple statistical testing of a falsely significant P‐value. The comparisons where there is no P‐value included can be considered exploratory.

3. RESULTS

Between April 11, 2011 and April 22, 2018, 584 patients were treated with standard (neo)adjuvant breast cancer chemotherapy regimens: anthracycline (doxorubicin‐ or epirubicin‐based, 4‐6 cycles), anthracycline‐taxane (6 cycles), anthracycline‐taxane (8 cycles), docetaxel/doxorubicin/cyclophosphamide (TAC, 6 cycles), docetaxel/cyclophosphamide (TC, 4 cycles), docetaxel/carboplatin/trastuzumab (TCH, 6 cycles), weekly paclitaxel (12 weeks) and other (cyclophosphamide/methotrexate/5‐FU [CMF], hybrids and changes of protocol). The most commonly prescribed regimen was TC (32%) (Table 1).

TABLE 1.

Patient and treatment characteristics

	Number (all = 584)	Percentage
Age (mean, [SD])	55 (11.2)
Sex
Female	579	99%
Male	5	1%
Age
<65 years	453	78%
≥65 years	131	22%
Nodal involvement
Node positive	255	45%
Node negative	329	56%
Regimen
TC	186	32%
Anthracycline‐taxane (6 cycles)	175	30%
Anthracycline‐taxane (8 cycles)	120	21%
TCH	55	9%
Weekly paclitaxel	22	4%
Anthracycline	12	2%
Other	9	2%
TAC	5	1%
Treatment timing
Neoadjuvant treatment	68	12%
Adjuvant treatment	516	88%
Dose intensity
Dose‐dense	10	2%
Non dose‐dense	574	98%
Prior chemotherapy or radiotherapy	7	1%

Abbreviations: TC, docetaxel/cyclophosphamide; TCH, docetaxel/carboplatin/trastuzumab; TAC, docetaxel/doxorubicin/cyclophosphamide.

Of 584 patients, 507 (87%) received GCSF during treatment, which was given subcutaneously the day after chemotherapy. The GCSF preparation used in all patients was pegfilgrastim 6.0 mg/0.6 mL. Primary GCSF was used in 269 patients and secondary GCSF was used in 238 patients. Secondary GCSF was given for cycle 1 FN in 64 patients and for cycle 1 nadir‐neutropenia in 124 patients. Other uses of secondary GCSF included FN and asymptomatic neutropenia occurring after cycle 1, and postoperative infections (Table 2).

TABLE 2.

GCSF usage

GCSF use	Number (%)
All patients	584
Any GCSF	507 (87%)
Primary GCSF	269 (46%)
Secondary GCSF	238 (41%)
Cycle 1 FN	64 (11%)
Cycle 1 nadir‐neutropenia	125 (21%)
Other	50 (9%)
GCSF source
Government reimbursed	466 (80%)
Primary	229 (39%)
Secondary	237 (41%)
Compassionate	41 (7%)
Primary	40 (7%)
Secondary	1 (0%)
Days before GCSF (mean, SD)	12.4 (SD 14.4)

Abbreviations: FN, febrile neutropenia; GCSF, granulocyte colony‐stimulating‐factor.

The majority of GCSF was government subsidized (466 patients) and was used similarly for both primary and secondary prophylaxis of FN. Compassionate GCSF via a pharmaceutical company access program (41 patients) was almost exclusively used as primary prophylaxis (Table 2).

Primary GCSF use ranged between 0% and 100% and secondary GCSF use ranged between 0% and 69% depending on the chemotherapy regimen. The overall rate of FN in patients was 17%, with rates highest for TCH (27%) and TC (27%) (Figure 1). The rate of FN in patients who received primary GCSF was lower (9%) (Table 3).

FIGURE 1.

Rates of primary GCSF use, secondary GCSF use, nadir blood counts, nadir neutropenia and FN by regimen. FN, febrile neutropenia; GCSF, granulocyte colony‐stimulating‐factor; TAC, docetaxel/doxorubicin/cyclophosphamide; TC, docetaxel/cyclophosphamide; TCH, docetaxel/carboplatin/trastuzumab

TABLE 3.

Rates of FN by regimen and GCSF use

	All patients		Primary GCSF		Secondary GCSF		No GCSF
Regimen	All	FN (number, %)	All (number, %)	FN (number, %)	All (number, %)	FN (number, %)	All (number, %)	FN (number, %)
Anthracycline	12	2 (17%)	7 (58%)	0 (0%)	5 (42%)	2 (40%)	0 (0%)	0 (NA)
Anthracycline‐taxane (6 cycles)	175	14 (8%)	173 (99%)	13 (8%)	2 (1%)	0 (0%)	0 (0%)	0 (NA)
Anthracycline‐taxane (8 cycles)	120	17 (14%)	32 (27%)	5 (16%)	59 (49%)	11 (19%)	29 (24%)	1 (3%)
Other	9	3 (33%)	2 (22%)	1 (50%)	4 (56%)	0 (0%)	3 (22%)	1 (33%)
TAC	5	1 (20%)	5 (100%)	1 (20%)	0 (0%)	0 (NA)	0 (0%)	0 (NA)
TC	186	50 (27%)	40 (22%)	4 (10%)	129 (69%)	43 (33%)	17 (9%)	3 (18%)
TCH	55	15 (27%)	10 (18%)	0 (0%)	38 (69%)	14 (37%)	7 (13%)	1 (14%)
Weekly paclitaxel	22	0 (0%)	0 (0%)	0 (NA)	0 (0%)	0 (NA)	22 (100%)	0 (0%)
Total	584	100 (17%)	269 (46%)	24 (9%)	237 (41%)	70 (30%)	78 (13%)	6 (8%)

Abbreviations: FN, febrile neutropenia; GCSF, granulocyte colony‐stimulating‐factor; NA, not applicable; TAC, docetaxel/doxorubicin/cyclophosphamide; TC, docetaxel/cyclophosphamide; TCH, docetaxel/carboplatin/trastuzumab.

Of the 316 patients who did not receive primary GCSF, 239 patients (76%) had cycle 1 nadir blood counts: anthracycline‐taxane (6 cycles) 2/2 patients (100%), other 6/7 patients (86%), TC 122/146 patients (84%), anthracycline 5/6 patients (83%), anthracycline‐taxane (8 cycles) 70/88 patients (80%), TCH 34/45 patients (76%), TAC 0/0 patients (not applicable), and weekly paclitaxel 0/22 (0%). Planned nadir blood counts were not done in 5 (2%) patients due to omission and 18 (6%) patients due to FN prior to the planned nadir count. Some 133 patients had nadir‐neutropenia: TCH 23/34 patients (68%), other 4/6 patients (67%), anthracycline 3/5 patients (60%), TC 72/122 patients (59%), anthracycline‐taxane (8 cycles) 31/70 patients (44%), anthracycline‐taxane (6 cycles) 0/2 patients (0%), weekly paclitaxel 0/0 patients (0%), TAC 0/0 patients (not applicable). The mean nadir‐neutropenia neutrophil count was 0.35 × 10⁹ cells/L. Timing of nadir blood counts varied between day 7 and 15 with mean timing day 12.8 (Figures 1 and 2; Table 4). Some clinicians performed multiple nadir blood counts within a cycle (eg, day 7 and 14).

FIGURE 2.

Nadir blood counts, secondary GCSF use and episodes of FN in patients who did not receive primary GCSF. Flow chart showing possible events for patients who did not receive primary GCSF for cycle 1; specifically nadir blood counts, FN events and use of secondary GCSF in cycle 2 for either FN or nadir‐neutropenia. Some patients ceased treatment after cycle 1. FN, febrile neutropenia; GCSF, granulocyte colony‐stimulating‐factor; G3/G4 neutropenia, common terminology criteria for adverse events grade 3/4 neutropenia

TABLE 4.

Outcome of nadir blood counts in patients who did not receive primary GCSF

Nadir blood count	Number (%)
Total patients not receiving primary GCSF	316
Nadir blood count	239 (76%)
Nadir‐neutropenia	133 (42%)
No neutropenia	64 (20%)
FN post nadir	42 (13%)
No nadir blood count	77 (24%)
Planned, not done due to FN prior to planned nadir	18 (6%)
Planned, not done	5 (2%)
Not planned	54 (17%)
Mean nadir‐neutropenia neutrophil count	0.35
Days between first chemotherapy and G3/4 neutropenia	12.8 (SD 13.04, range 5‐77)

Abbreviations: FN, febrile neutropenia; GCSF, granulocyte colony‐stimulating‐factor; G3/G4 neutropenia, common terminology criteria for adverse events grade 3/4 neutropenia.

Of the 133 patients who had nadir‐neutropenia, 125 proceeded to cycle 2 and received nadir‐neutropenia directed secondary GCSF, of which 6 developed FN: anthracycline 1/3 patients (33%), other 1/4 patients (25%), TCH 1/20 patients (5%), TC 3/71 patients (4%), anthracycline‐taxane (8 cycles) 0/27 patients (0%). There were 77 patients who did not receive primary GCSF nor have nadir blood counts, of which 49 proceeded to cycle 2 without secondary GCSF. None of these patients developed FN: anthracycline 0/1 patients (0%), anthracycline‐taxane (8 cycles) 0/13 patients (0%), TC 0/10 patients (0%), TCH 0/3 patients (0%), weekly paclitaxel 0/22 patients (0%). There was no significant difference in rates of FN in patients who received nadir‐neutropenia directed secondary GCSF (6/125, 5%, 95% CI 1.1‐8.6) compared with those who did not have a nadir blood count (0/49, 0%, 95% CI not calculable) (P = .1186) (Figure 2).

The overall rate of non‐FN admissions was 29%. Twenty‐eight percent of patients who received GCSF had a non‐FN admission compared with 34% of those who did not receive GCSF.

Patients ≥65 years comprised 22% of the cohort. Anthracycline‐taxane (6 cycles) was the mostly commonly used regimen in patients <65 years (32%) whereas TC was the most commonly used regimen in patients ≥65 years (43%). Episodes of nadir blood counts (44% vs 40%), FN (21% vs 17%) and non‐FN admissions (33% vs 27%) occurred in patients ≥65 years and <65 years respectively. GCSF use was lower in patients ≥65 years (81% vs 89%).

Seven patients had received prior chemotherapy or chemo‐radiotherapy for antecedent malignancies, and one continued to receive targeted therapy for chronic myeloid leukemia during breast cancer chemotherapy. Five of these patients had primary GCSF, and the two remaining patients had nadir blood counts. There were two cases of FN in this group (29%).

There were three deaths related to chemotherapy (0.51%). Two deaths were due to FN and occurred with TC (0.54%) and anthracycline‐taxane (6 cycles) (1.14%). One death was due to pneumonitis with anthracycline‐taxane (6 cycles) (1.14%). Both FN deaths occurred in patients who had received prior chemotherapy or chemo‐radiotherapy, with one death occurring despite primary GCSF. All three patients who died were aged <65 years.

4. DISCUSSION

This analysis provides a snapshot of current routine clinical practice with (neo)adjuvant chemotherapy for breast cancer in Australia. In particular, it demonstrates the high rates of FN associated with standard (neo)adjuvant breast cancer chemotherapy regimens, with an average rate of 17% and rates of up to 33% in some regimens.

As previously reported in the literature, this analysis found chemotherapy regimen, age ≥65 years and prior chemotherapy were risk factors for FN. ²¹ , ²⁴ , ²⁵ TC and TCH had particularly high rates of FN (27%), which is of importance given TC was the most commonly prescribed regimen. These rates are much higher than rates of 4% to 10% reported in the seminal trials. ⁵ , ⁶

GCSF was a very frequently used intervention and, as expected, its use was associated with lower rates of FN. Interestingly, non‐FN hospital admissions were also lower with GCSF use. Possible reasons for this include reduction in other clinically significant infections.

Whilst there is limited data on the incidence of nadir blood count measurements in the literature, it was a common practice at our institution, occurring in 75% patients who did not receive primary GCSF. Nadir blood counts were performed at similar rates across regimens (excluding weekly paclitaxel), and in patients aged <65 years and ≥65 years. Nonadherence with planned nadir blood counts did not appear to be a significant problem.

There is also limited data on the effects of GCSF following nadir‐neutropenia. In this review, the use of secondary GCSF for nadir‐neutropenia <1.0 × 10⁹ cells/L was not associated with significantly lower rates of FN and thus the use of nadir blood counts and nadir‐neutropenia directed GCSF is not recommended to reduce rates of FN. However, we hypothesize that in regimens with true intermediate risk of FN including anthracycline, anthracycline‐taxane (6 cycles) and anthracycline‐taxane (8 cycles), there may be a role for nadir blood counts and nadir‐neutropenia directed GCSF to identify patients at increased risk of FN and reduce rates of FN as well as improve chemotherapy relative dose delivery.

Whilst GCSF is a relatively low risk intervention, it is not without adverse effects and its use is associated with increased initial treatment costs. The observed rates of FN >20% in TC and TCH, and of 20% in TAC despite 100% GCSF use, can guide the use of primary GCSF in these regimens in accordance with international guidelines. ²¹ , ²⁴ , ²⁵ In Australia, patients receiving these regimens are currently eligible for primary GCSF under government reimbursement arrangements, ²⁷ and in alternatively resourced settings these findings allow better identification of patients at increased risk of FN who may benefit from primary GCSF pending access. This review also supports a potential role for primary GCSF in patients ≥65 years as well as in patients who have previously received chemotherapy or chemo‐radiotherapy to reduce rates of FN and non‐FN hospital admissions (excluding weekly paclitaxel).

Limitations regarding these findings include being a single center, non‐randomized, retrospective study influenced by local protocols and individual clinician practices. This included electronic chemotherapy prescribing which resulted in standardization of regimens and inclusion of primary GCSF for TAC and anthracycline‐taxane (6 cycles) as per Australian government reimbursement criteria.

Secondly, a high proportion of patients received primary GCSF or had an episode of FN in cycle 1, reducing the number of patients who were considered for nadir‐neutropenia directed secondary GCSF (125/584 patients, 21%). This was influenced by government reimbursement criteria and a pharmaceutical access program that facilitated primary GCSF for 40 patients that were perceived to be at increased risk for FN who might otherwise have been considered for nadir blood counts.

Thirdly, the definitions of neutropenia and FN used in this study were broader than Australian and international guidelines and likely captured some cases of low risk or non‐FN. Additionally, the study only considered the neutrophil value at the time of clinician specified nadir measurement, which did not capture the depth or duration of neutropenic episodes, nor determine the ideal time for nadir blood count measures.

Opportunities for further analysis of this cohort include review of dose modifications and cessation, disease relapse and survival, and Multidisciplinary Team involvement in decision making and outcome.

5. CONCLUSION

In this population, secondary GCSF following nadir‐neutropenia was not associated with reduced rates of FN, and thus nadir blood counts and nadir‐neutropenia directed GCSF is not recommended to reduce rates of FN in patients receiving (neo)adjuvant breast cancer chemotherapy. However, observed rates of FN >20% in TC and TCH in routine clinical practice demonstrates the high risk of FN with these regimens and should guide primary GCSF use according with international guidelines. Older patients and patients who have previously received chemotherapy should also be considered for primary GCSF due to higher observed rates of FN.

AUTHOR CONTRIBUTIONS

Sarah Zardawi: Conceptualization; formal analysis; investigation; methodology; visualization; writing‐original draft; writing‐review and editing. Ina Nordman: Conceptualization; methodology; supervision; writing‐review and editing. Nicholas Zdenkowski: Conceptualization; formal analysis; methodology; resources; supervision; writing‐review and editing.

CONFLICT OF INTEREST

The authors report no conflict of interest regarding the content of this manuscript.

ETHICAL STATEMENT

This study, as a retrospective review, received an exemption from the Hunter New England Human Research Ethics Committee to proceed.

Zardawi SJ, Nordman I, Zdenkowski N. A retrospective analysis of nadir‐neutropenia directed pegylated granulocyte‐colony stimulating factor on febrile neutropenia rates in (neo)adjuvant breast cancer chemotherapy regimens. Cancer Reports. 2020;3:e1266. 10.1002/cnr2.1266

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.