Annals of Oncology Advance Access originally published online on August 13, 2008
Annals of Oncology 2009 20(1):34-40; doi:10.1093/annonc/mdn560
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
breast cancer |
Pretreatment haematological laboratory values predict for excessive myelosuppression in patients receiving adjuvant FEC chemotherapy for breast cancer
1 Gloucestershire Oncology Centre, Cheltenham General Hospital, Cheltenham
2 Department of Haematology, University of Birmingham Medical School, Birmingham, UK
* Correspondence to: Dr P. Jenkins, Gloucestershire Oncology Centre, Cheltenham General Hospital, Cheltenham, GL50 2LR, UK. Tel: +44-8454-224019; Fax: +44-8454-223506; E-mail: peter.jenkins{at}glos.nhs.uk
 |
Abstract |
|---|
 Top Abstract introductionpatients and methodsresultsdiscussionAcknowledgementsReferences |
|---|
Background: A predictive model that identifies patients at risk of excessive neutropenia following chemotherapy would be valuable in guiding the use of supportive therapies.
Patients and methods: We conducted a retrospective analysis of 741 patients who had received adjuvant 5-fluorouracil, epirubicin, cyclophosphamide (FEC) chemotherapy for breast cancer. The cause of every schedule alteration was identified. The ability of pretreatment haematological indices to predict for excessive myelosuppression was assessed.
Results: Pretreatment absolute neutrophil count (ANC) and absolute lymphocyte count (ALC) were strongly associated with the risk of neutropenic events (NEs), febrile neutropenia (FN) or receiving suboptimal chemotherapy dose intensity (DI < 85%). The timing and pattern of NEs suggest that they reflect intrinsic chemosensitivity rather than cumulative toxicity and that FN results from chance infection rather than protracted myelosuppression. By combining quintiles on the basis of the rank order of ANC and ALC, we defined five groups of patients with variable risks of NE (18%–52%), DI <85% (9%–36%) and FN (4%–21%).
Conclusions: Pretreatment differential white blood cell count can be used to identify patients at increased risk of significant myelosuppression with FEC chemotherapy. Patients in the highest risk group have a risk of FN >20% and would qualify for primary prophylaxis with granulocyte colony-stimulating factor support under current guidelines.
Key words: breast cancer, chemotherapy, myeloid growth factors, neutropenia, risk models
|
introduction |
|---|
|
TopAbstractintroductionpatients and methodsresultsdiscussionAcknowledgementsReferences |
|---|
Myelosuppression, particularly neutropenia, is the most common toxicity seen in patients undergoing systemic cancer chemotherapy. In addition to predisposing to life-threatening infections, it often leads to treatment delays and dose reductions that reduce the intensity of chemotherapy. The importance of chemotherapy dose intensity for survival with adjuvant therapy in breast cancer is being increasingly recognised [1–3]. It appears from these studies that lower dose intensity is associated with reduced clinical benefit. There may even be a threshold level of dose intensity below which patients derive no advantage from treatment [1]. Patterns of care studies have shown that, for patients treated outside clinical trials, the schedule of chemotherapy is frequently interrupted due to haematological toxicity [4–6]. Identifying individuals who are at risk of significant myelosuppression is therefore of considerable importance to medical oncologists treating patients with breast cancer. A predictive instrument that could quantify the risk of neutropenic events (NEs) might be valuable in helping to reduce the complications arising from myelosuppression [6–9]. The focus of much of the research in this area has been to better target supportive strategies and thereby avoid either serious infections or compromised dose intensity. However, previous studies are potentially confounded by the multiple chemotherapy regimens employed, different thresholds for treatment delay, a priori dose reductions, failure to discriminate between haematological and non-haematological toxicity and variable use of myeloid growth factor support. We have analysed a large cohort of breast cancer patients treated with a single-chemotherapy regimen that is not confounded by these aforementioned factors. Previously we have reported that obesity and high body surface area (BSA) are not associated with increased myelosuppression in this group [10]. The present study describes the development of a simple model, based on pretreatment laboratory haematology values, which can be used to identify groups of patients who are at increased risk of chemotherapy-associated neutropenia.
|
patients and methods |
|---|
|
TopAbstractintroductionpatients and methodsresultsdiscussionAcknowledgementsReferences |
|---|
Using our patient management information system, an electronic database through which chemotherapy is prescribed, we identified all patients treated with FEC chemotherapy for breast cancer in the period 2001–2005. The regimen employed (5-fluorouracil 600 mg/m2, epirubicin 60 mg/m2, cyclophosphamide 600 mg/m2 administered i.v. every 3 weeks) has been reported previously [11]. A total of 741 patients received neo-adjuvant or adjuvant FEC. All patients were planned to receive a minimum of six cycles of treatment. Neither a priori dose reductions nor BSA capping were employed for any of the patients. A dose-banding algorithm was used to facilitate the reconstitution of drugs in our cytotoxic pharmacy. The actual doses prescribed were within 5% of the calculated doses based on BSA.
Throughout the period of this audit, an institutional policy for the prescription of chemotherapy was in operation. Automated full blood counts were carried out on the day of chemotherapy. Each cycle of treatment was delivered if the absolute neutrophil count (ANC) was 
1.5 x 109/l and the platelet count 100 x 109/l. If not, chemotherapy was delayed for 1 week. If chemotherapy had to be delayed on two occasions due to inadequate blood counts, a dose reduction of 20% was made for all agents and maintained for the rest of the treatment course. Greater dose reductions were instituted if the patient experienced further myelosuppression despite these modifications. The Gloucestershire Oncology Centre operates a 24-h telephone helpline and a dedicated neutropenic ward for the admission of all patients with febrile neutropenia (FN). After an episode of FN, doses were also reduced by 20%. Due to funding restrictions, granulocyte colony-stimulating factors (G-CSFs) were not employed as either primary or secondary prophylaxis during the period of this audit. No patient received concomitant radiotherapy and endocrine treatment was only commenced after completion of chemotherapy.
This retrospective study was approved by the departmental research and audit committee. The actual doses delivered, timing of each cycle and occurrence of FN were obtained for each patient by searching the electronic database. In any case where there was a dose reduction or cycle delay, the oncology records were reviewed to determine the cause for the alteration in treatment schedule.
A NE was defined as either an episode of FN or a dose delay 1 week caused by prolonged myelosuppression. Dose intensity was defined as (dose received/dose planned)/(overall treatment time/planned treatment time) over six cycles of treatment [12]. This was considered to be suboptimal if it fell <0.85 (DI < 85%) [1].
Group comparisons for continuous blood count variables were carried out using Student's t-test. Chi-square tests were used to determine the significance of differences between categorical variables. All tests of significance were two tailed. Statistical analyses were carried out using SPSS version 10.0 (SPSS, Chicago, IL).
|
results |
|---|
|
TopAbstractintroductionpatients and methodsresultsdiscussionAcknowledgementsReferences |
|---|
In total, 741 consecutively treated patients were identified by this retrospective study. Three hundred and twenty-five patients had a schedule interruption. The cause for this was defined as haematological in 192, non-haematological in 59, social in 15, unrelated medical conditions in 18, progressive disease in seven, patient request in seven or combined in 27. Eight patients had no retrievable blood counts. As we were interested in determining whether pretreatment blood count parameters predict for future myelotoxicity, we excluded those 141 patients with non-haematological toxicity or absent blood counts. The mean dose intensity for the remaining 600 patients was 94.6%. Of the 192 patients with a NE, 53 had an episode of FN (three patients had two episodes) and 160 experienced one or more treatment delays due to myelosuppression. In total, 104 (17.3%) patients received a DI <85% and 10 (1.6%) received a DI <65% due to myelosuppression. Patient demographics, tumour characteristics and pretreatment haematological values for this group are shown in Table 1.
|
Figure 1 shows the pattern of NEs over the six-cycle course of chemotherapy. While treatment delays were evenly spread throughout the course of chemotherapy, FN was more common in cycle 1 (25 of 53, 47%). Of relevance to the use of secondary G-CSF prophylaxis is the fact that 83% of patients who develop a NE in cycle 1 go onto receive a DI <85%. Importantly, however, delay of a previous cycle does not seem to correlate with the risk of subsequent FN. Of the 29 patients with FN in cycles 2–6, only two (7%) had experienced a previous treatment delay due to protracted myelosuppression (Figure 1). In total, 100 patients had a 20% dose reduction following either two or more delays or an episode of FN. This was only of limited effectiveness in preventing subsequent events as 47 (47%) went on to develop a further NE despite this reduction in chemotherapy doses.
|
Table 2 summarises the association between pretreatment haematological parameters and excessive neutropenia. Total white cell count (WCC), ANC and absolute lymphocyte count (ALC) are all associated with NEs, DI <85% and FN. Although the differences in blood count indices are statistically highly significant, their absolute values are small (Figure 2). Figure 3 shows receiver-operator characteristic curves for ANC and ALC with relation to any NE and DI <85% and FN.
|
|
|
As ANC and ALC were both correlated with the total WCC (r = 0.91 and 0.55, respectively), but less well correlated with each other (r = 0.21), these two parameters were selected for further study. To develop a simple risk model, patients were divided into quintiles based on both ANC and ALC. These were then combined to generate five groups on the basis of their rank order as shown in Table 3 and Figure 4. There were highly significant differences in the risk of NEs between these separate groups. The risk of any NE varied by 2.8-fold, DI <85% varied by 4-fold and FN by 5.3-fold between the separate groups (Table 3). The distribution of NEs with relation to ANC and ALC is best appreciated in the scatter plots (Figure 4). One of the most striking features of Figure 4 is the low risk of significant myelosuppression in patients with higher WCCs. Thus, it is possible to define a subset of group I on the basis of patients with both an ANC >5.3 and an ALC >2.5. This group, comprising 6% of the total population, have an extremely low risk of NE (11%), DI >85% (3%) and FN (0%).
|
|
|
discussion |
|---|
|
TopAbstractintroductionpatients and methodsresultsdiscussionAcknowledgementsReferences |
|---|
Myelosuppression and its associated complications are the most significant morbidities encountered by patients undergoing chemotherapy for breast cancer. In addition to the risks of infection, neutropenia may result in schedule alteration, reducing the intensity of treatment. In contrast to the maximum grade of haematological toxicity, the dose intensity of chemotherapy is inconsistently reported in publications of trial results. However, a survey of oncology practices found that 56% of patients treated in community hospitals received <85% of the planned dose intensity [6]. This reflects the difference between the highly selected population eligible for clinical studies and the majority of patients not treated in trials, who are generally less healthy. As chemotherapy regimens become increasingly dose intense, schedule interruptions in this wider group of patients might potentially offset the beneficial effects of more effective agents.
This retrospective audit has provided some insight into the risk and timing of neutropenic complications. The FEC regimen employed has a low risk of causing FN so, using existing guidelines, primary G-CSF prophylaxis would not be indicated [13, 14]. Notwithstanding this, a large number of patients (26%) had schedule interruptions due to haematological toxicity. Furthermore, dose reductions were undertaken in 13% in response to an episode of FN or more than one treatment delay. We have shown that this strategy is of limited effectiveness as approximately half the patients who had a 20% dose reduction went on to develop further NEs in subsequent cycles. Clearly, more substantial dose reductions could be employed, but this will further reduce the intensity of chemotherapy and may compromise the effectiveness of treatment.
A large number of patients who had a NE in cycle 1 went on to receive a DI <85%. This is an important observation for guiding the use of secondary G-CSF prophylaxis. If secondary prophylaxis was completely effective at preventing subsequent NEs, then 61% of the cases of suboptimal dose intensity could be prevented by the targeted use of G-CSF based on cycle 1 NEs. This hypothesis is being tested by a randomised study conducted in UK (Anglo Celtic III) that has now completed accrual.
Previous studies have shown that chemotherapy progressively depletes the stem-cell pool in the bone marrow [15]. This is usually manifested as a reduction in white blood cell counts with each successive cycle of chemotherapy. However, it is noteworthy that of the 192 patients who had a NE in our study, the majority of these could be identified after just two cycles (Figure 1). Thus, intrinsic chemosensitivity rather than a cumulative toxicity appears to be more important in the development of NEs. It is also evident from Figure 1 that many patients who develop FN in cycles 2–6 have no forewarning of this such as delays of preceding cycles. This observation supports the hypothesis that episodes of FN result from chance infection in a susceptible window, rather than extreme or prolonged neutropenia as seems to be the case in acute leukaemia [16]. In support of this finding, a previous study, which employed selective administration of G-CSF on the basis of first cycle nadir counts, found that this strategy reduced treatment delays but had no effect on the risk of developing FN [17].
In common with previous reports [18], we found that episodes of FN were more frequent in cycle 1 (Figure 1). A simple model that quantifies the risks of NE prospectively would therefore be of clinical value for guiding the selective use of primary prophylaxis strategies. Several studies have identified patient-related factors that predispose to either FN [13, 14] or excessive myelosuppression [6]. These include age >65, advanced disease, anaemia, poor performance status, prior treatment, combined chemoradiotherapy, bone marrow infiltration and medical comorbidity particularly renal disease. However, many of these pretreatment risk factors were identified in reports which included patients with haematological malignancies and are of questionable relevance to the adjuvant treatment of breast cancer. Other studies have shown that first cycle ANC nadir can be used to predict for subsequent NE [7–9]. For example, Sibler et al. [7] studied 95 patients receiving standard-dose adjuvant chemotherapy and found that the depth of the first cycle ANC nadir was a reliable predictor of NE. However, as true nadir counts are difficult to obtain reliably and inconvenient for the patient, they are not routinely carried out in clinical practice. Our data have shown that pretreatment haematological laboratory values can also be used to predict either the risk of NE, receiving a DI <85%, or FN in patients treated with FEC. Patients with lower absolute lymphocyte and neutrophil counts are more likely to experience treatment delays or FN during their chemotherapy programme (Figures 2 and 4). By combining the two principal component of total WCC (ANC and ALC), we developed a simple instrument which can be used to predict for susceptibility to excessive myelosuppression with FEC chemotherapy. Patients in the highest risk group, defined as ANC 
3.1 and ALC 1.5, comprise 6% of the total population. They have a risk of FN (21%) that is over five-fold greater than patients in the lowest risk group. Importantly, patients in this group would qualify for G-CSF support according to current guidelines even though the overall risk of FN with the FEC regimen employed is low (7%) [13, 14].
Why do pretreatment haematological indices, particularly the lymphocyte count, affect the risk of NE? Low WCCs might indicate bone marrow insufficiency but this would seem unlikely in this cohort of predominantly middle-aged women undergoing adjuvant chemotherapy. Previous studies have shown that granulopoiesis and B lymphopoiesis share a common developmental niche in the bone marrow, competing for critical growth and retention factors such as the chemokine SDF-1 [19]. Lower blood B lymphocyte and neutrophil counts may both result from relatively high bone marrow concentrations of SDF-1 that promotes retention of these cells in the marrow. During inflammation, it has been shown that G-CSF mobilises neutrophils and myeloid progenitors to the peripheral circulation by disrupting this SDF-1 retention signal. However, endogenous G-CSF is also required for granulopoiesis postchemotherapy and therefore may be present in limiting amounts. It is possible that the increased susceptibility of patients with lower lymphocyte and neutrophil blood counts to chemotherapy induced neutropenia may be explained by a relative increase in bone marrow SDF-1 levels in these patients coupled with limiting amounts of G-CSF following treatment. Further studies are required to test this hypothesis.
Can our findings be applied to other chemotherapy regimens? The FEC60 regimen employed for the duration of this audit is now considered to be inferior to more dose-intense regimens such as FEC100 and consequently it is less frequently used [20]. In addition, G-CSF is now more widely administered to maintain chemotherapy dose intensity in patients who have treatment delays or episodes of FN. Undoubtedly, many of the patients who received dose delays or dose reductions would currently be treated with G-CSF. Notwithstanding these reservations, the approach that we describe might prove useful with other regimens. On the basis of the current guidelines, primary prophylaxis with G-CSF is indicated when the risk of FN exceeds 20%. Although many adjuvant breast cancer regimens such as FEC100 are more myelosuppressive than FEC60, they still have a risk of FN <20%. We predict that patients in group V of our model will also be at >20% risk of FN with these protocols and therefore should also be considered for primary prophylaxis with G-CSF.
In conclusion, we present a simple model which provides a means of identifying a group of patients at increased risk of significant neutropenia with adjuvant FEC chemotherapy. Clearly, it will require validation in a separate dataset and we are currently working to confirm these observations with other chemotherapy regimens and disease sites. Potentially this approach will allow prophylactic G-CSF or antibiotic therapy to be targeted more effectively to patients at the highest risk of FN on the basis of the chemotherapy regimen and pretreatment laboratory parameters.
|
Acknowledgements |
|---|
|
TopAbstractintroductionpatients and methodsresultsdiscussionAcknowledgementsReferences |
|---|
The authors acknowledge Drs S Elyan, JR Owen, K Benstead, J Bowen and R Counsell for their cooperation in the completion of this project and Mrs J Whitmarsh and Mrs M Wenham for administrative assistance.
Received for publication May 16, 2008. Accepted for publication July 14, 2008.
|
References |
|---|
|
TopAbstractintroductionpatients and methodsresultsdiscussionAcknowledgementsReferences |
|---|
1. Bonadonna G, Valagussa P, Moliterni A, et al. Adjuvant cyclophosphamide, methotrexate, and fluorouracil in node-positive breast cancer: the results of 20 years of follow-up. N Engl J Med (1995) 332:901–906.
2. Budman DR, Berry DA, Cirrincione CT, et al. Dose and dose intensity as determinants of outcome in the adjuvant treatment of breast cancer. The Cancer and Leukemia Group B. J Natl Cancer Inst (1998) 90:1205–1211.
3. Group FAS. Benefit of high-dose epirubicin regimen in adjuvant chemotherapy for node-positive breast cancer patients with poor prognostic factors: 5-year follow-up results of French Adjuvant Study Group 05 randomized trial. J Clin Oncol (2001) 19:599–601.
4. Chang J. Chemotherapy dose reduction and delay in clinical practice: evaluating the risk to patient outcome in adjuvant chemotherapy for breast cancer. Eur J Cancer (2000) 36:S11–S14.[CrossRef][Web of Science][Medline]
5. Link BK, Budd GT, Scott S, et al. Delivering adjuvant chemotherapy to women with early-stage breast carcinoma: current patterns of care. Cancer (2001) 92:1354–1367.
6. Lyman GH, Dale DC, Crawford J. Incidence and predictors of low dose-intensity in adjuvant breast cancer chemotherapy: a nationwide study of community practices. J Clin Oncol (2003) 21:4524–4531.
7. Silber JH, Fridman M, DiPaola RS, et al. First-cycle blood counts and subsequent neutropenia, dose reduction, or delay in early-stage breast cancer therapy. J Clin Oncol (1998) 16:2392–2400.[Abstract]
8. Rivera E, Haim Erder M, Fridman M, et al. First-cycle absolute neutrophil count can be used to improve chemotherapy-dose delivery and reduce the risk of febrile neutropenia in patients receiving adjuvant therapy: a validation study. Breast Cancer Res (2003) 5:R114–R120.[CrossRef][Web of Science][Medline]
9. Savvides P, Terrin N, Erban J, et al. Development and validation of a patient-specific predictive instrument for the need for dose reduction in chemotherapy for breast cancer: a potential decision aid for the use of myeloid growth factors. Support Care Cancer (2003) 11:313–320.[Web of Science][Medline]
10. Jenkins P, Elyan S, Freeman S. Obesity is not associated with increased myelosuppression in patients receiving chemotherapy for breast cancer. Eur J Cancer (2007) 43:544–548.[CrossRef][Web of Science][Medline]
11. Mouridsen H, Andersen J, Andersson M, et al. Adjuvant anthracycline in breast cancer. Improved outcome in premenopausal patients following substitution of methotrexate in the CMF combination with epirubicin. Proc Am Soc Clin Oncol (1999) 19:68a. (Abstr 254).
12. Longo DL, Duffey PL, DeVita VT Jr, et al. The calculation of actual or received dose intensity: a comparison of published methods. J Clin Oncol (1991) 9:2042–2051.
13. Smith TJ, Khatcheressian J, Lyman GH, et al. 2006 update of recommendations for the use of white blood cell growth factors: an evidence-based clinical practice guideline. J Clin Oncol (2006) 24:3187–3205.
14. Aapro MS, Cameron DA, Pettengell R, et al. EORTC guidelines for the use of granulocyte-colony stimulating factor to reduce the incidence of chemotherapy-induced febrile neutropenia in adult patients with lymphomas and solid tumours. Eur J Cancer (2006) 42:2433–2453.[CrossRef][Web of Science][Medline]
15. Botnick LE, Hannon EC, Hellman S. A long lasting proliferative defect in the hematopoietic stem cell compartment following cytotoxic agents. Int J Radiat Oncol Biol Phys (1979) 5:1621–1625.[Web of Science][Medline]
16. Bodey GP, Buckley M, Sathe YS, et al. Quantitative relationships between circulating leukocytes and infection in patients with acute leukemia. Ann Intern Med (1966) 64:328–340.
17. Rivera E, Erder MH, Moore TD, et al. Targeted filgrastim support in patients with early-stage breast carcinoma: toward the implementation of a risk model. Cancer (2003) 98:222–228.
18. Nabholtz JM, Cantin J, Chang J, et al. Phase III trial comparing granulocyte colony-stimulating factor to leridistim in the prevention of neutropenic complications in breast cancer patients treated with docetaxel/doxorubicin/cyclophosphamide: results of the BCIRG 004 trial. Clin Breast Cancer (2002) 3:268–275.[Medline]
19. Ueda Y, Kondo M, Kelsoe G. Inflammation and the reciprocal production of granulocytes and lymphocytes in bone marrow. J Exp Med (2005) 201:1771–1780.
20. Bonneterre J, Roche H, Kerbrat P, et al. Epirubicin increases long-term survival in adjuvant chemotherapy of patients with poor-prognosis, node-positive, early breast cancer: 10-year follow-up results of the French Adjuvant Study Group 05 randomized trial. J Clin Oncol (2005) 23:2686–2693.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||