Abstract
Background/Aim: To maximize the effect of perioperative chemotherapy in breast cancer, it is critical to keep the relative dose intensity (RDI) high. While bi-weekly doxorubicin and cyclophosphamide, dose-dense AC (ddAC), instead of tri-weekly conventional AC (cAC) followed by a taxane has been adopted as standard perioperative chemotherapy, postponement or discontinuation are sometimes experienced during ddAC or subsequent taxane phase. This study aimed at evaluating whether ddAC, compared to cAC, was associated with reduced RDI. Patients and Methods: We compared ddAC and cAC, both followed by a taxane, for perioperative breast cancer regarding the proportion of completion of planned treatment (%completion), defined as an RDI ≥85% for both AC and taxane phases. Results: There was no remarkable difference between the groups in patient characteristics after propensity score matching (n=46 in ddAC, and n=86 in cAC). The %completion was similar between the groups (67.4% vs. 65.1%). Most other endpoints related to RDI were similar between groups. The incidence of pneumonia was higher in the ddAC group (13% vs. 3%) including one Pneumocystis jiroveci pneumonia. Conclusion: ddAC followed by a taxane can be considered with sufficient supportive measures and precautions for pneumonia.
Breast cancer is the most common cancer in women worldwide (1). The survival benefit of perioperative, adjuvant or neoadjuvant, chemotherapy has been established in patients with early-stage and locally-advanced breast cancer (2). The standard and one of the most commonly employed perioperative chemotherapy regimens is doxorubicin plus cyclophosphamide (AC) followed by a taxane (docetaxel or paclitaxel) (3, 4).
The effect of chemotherapy is generally correlated with dose intensity, defined as the total amount of drug given in a fixed unit of time, and most commonly expressed as milligrams per meter squared per week (mg/m2/week) (5). Relative dose intensity (RDI) is defined as the ratio of the actual dose intensity over the standard or planned dose intensity. The RDI of perioperative chemotherapy for breast cancer has been reported to affect disease-free survival (DFS) and overall survival (OS) (5). A randomized controlled trial suggested that early-stage breast cancer patients treated with a cyclophosphamide, methotrexate, and 5-fluorouracil (CMF) regimen delivered with an RDI of ≥85% had a significantly better survival than those treated with the same regimen but delivered with an RDI of <85% (6). The survival benefits of receiving a higher RDI have been confirmed subsequently in several observational studies (5, 7), and clinical guidelines recommend keeping an RDI of at least 85% for optimal treatment (8).
To increase dose intensity, a dose-dense (dd)AC regimen was developed; it shortens the interval of AC therapy from the standard 21 to 14 days by using granulocyte colony stimulating factor (G-CSF) support. In a randomized trial, ddAC followed by bi-weekly paclitaxel was shown to improve DFS and OS compared to tri-weekly conventional AC (cAC) followed by tri-weekly paclitaxel (9). A meta-analysis that included 26 randomized trials also demonstrated that increasing the dose intensity by shortening the interval between treatment cycles significantly reduced the risk of recurrence and death from breast cancer without increasing mortality from other causes (10). Based on these data, ddAC was adopted as the standard perioperative chemotherapy regimen for breast cancer (8).
Since 2014, when a long-acting granulocyte-colony stimulating factor (G-CSF) preparation (pegfilgrastim) became available in Japan, our hospital has adopted the use of ddAC followed by a taxane for the treatment of high-risk breast cancer patients. However, for several patients we had to reduce the dose of drugs or postpone or discontinue treatment in either the ddAC or subsequent taxane phase. This raised a concern that in a non-negligible proportion of patients, shortening the treatment interval may lead to a reduced dose intensity and potentially worsen the patient’s prognosis. Therefore, we decided to suspend the use of ddAC and adopt only cAC. However, the clinical question remains as to whether ddAC followed by a taxane is in fact associated with reduced RDI as compared with cAC followed by a taxane.
Therefore, in this study, we compared cAC followed by a taxane with ddAC followed by a taxane concerning the proportion of completion of planned treatment (%completion), defined as the proportion of patients who achieved an RDI ≥85% for both AC and taxane phases.
Patients and Methods
Patient and data collection. Data were collected retrospectively from electronic medical charts of clinical stage I-IIIC breast cancer patients who were treated with either cAC or ddAC followed by a taxane, docetaxel or paclitaxel, as perioperative chemotherapy at the National Cancer Center Hospital East. After pegfilgrastim became available in 2014, the dosing schedule of AC in our daily practice shifted from cAC to ddAC almost fully by 2016. However, the experience of several patients who had to reduce the dose of drugs and/or discontinue the treatment in either the ddAC or subsequent taxane phase because of adverse events made us suspend the use of this regimen at the end of 2017. Therefore, we enrolled patients in a cAC group from the list of patients who started cAC from January 1, 2014 to December 31, 2015, and from January 1, 2018 to March 31, 2019, and we enrolled patients in a ddAC group from the list of patients who started ddAC from January 1, 2016 to December 31, 2017. Patients who were treated with upfront taxanes followed by a cAC or ddAC regimen, who were switched from paclitaxel to docetaxel and vice versa during the taxane phase, or who had a concomitant cancer were excluded (Figure 1).
CONSORT diagram of selection process for study subjects.
Pretreatment demographics and clinical characteristics collected included sex, age at diagnosis, menstruation status, height, weight, body surface area (BSA), smoking status, comorbidities, breast cancer subtype determined by estrogen receptor (ER), progesterone receptor (PgR), and human epidermal growth factor receptor 2 (HER2) statuses, timing of chemotherapy (preoperative or postoperative), clinical stage (TNM), and planned chemotherapy dose and schedule. Treatment data collected included the actual chemotherapy dose and schedule, presence or absence of dose reduction of anticancer drugs, postponement or discontinuation of treatment, and reasons for them if present. Blood test results collected included total bilirubin, aspartate aminotransferase, alanine aminotransferase, alkaline phosphatase, serum creatinine, creatinine clearance, absolute neutrophil count, and absolute lymphocyte count at the start of the first cycle of cAC or ddAC, and absolute lymphocyte count at the start of a taxane.
Ethics. The Institutional Review Board (IRB) of the National Cancer Center approved this study (IRB number 2019-072), which was conducted in accordance with the principles stated in Japan’s Ethics Guidelines for Epidemiological Research. The IRB waived the requirement for obtaining written informed consent from the study’s subjects.
Outcomes. The endpoint of primary interest was %completion, defined as the proportion of patients who achieved an RDI of 85% or greater for both cAC or ddAC and a taxane. The planned dosing schedules applied to AC were 4 cycles of doxorubicin 60 mg/m2 plus cyclophosphamide 600 mg/m2 with 21-day and 14-day intervals for cAC and ddAC, respectively. Pegfilgrastim was administered typically on day 2 of each cycle of ddAC. The subsequent taxane phase consisted of either 4 cycles of docetaxel 75 mg/m2 with a 21-day interval or 12 cycles of paclitaxel 80 mg/m2 with a 7-day interval. While G-CSF was not given as a primary prophylaxis for regimens other than ddAC, secondary prophylaxis with G-CSF was considered when the patient experienced a febrile neutropenia during a previous cycle.
The other endpoints included the RDI of doxorubicin, the RDI of a taxane, the proportion of treatment discontinuation or postponement at the AC and/or taxane phases, and absolute lymphocyte count at the start of a taxane.
Statistical methods. Our primary interest was to estimate the relative risk when comparing %completion between AC and ddAC followed by a taxane. Confounding variables were adjusted using propensity score (PS) analyses. The PS was estimated using a multivariable logistic regression model that included the following covariates: age at diagnosis, body mass index (≤24 vs. >24 kg/m2), smoking history (never smoker vs. current or past smoker), Charlson Comorbidity Index (CCI) (2 vs. ≥3), heart disease (coexisting vs. not), chemotherapy administration schedule (neoadjuvant vs. adjuvant), hormone receptor statuses (ER- and/or PgR-positive vs. negative), HER2 status (positive vs. negative), clinical lymph node metastasis (positive vs. negative), and clinical tumor size (≤2.0 vs. 2.1-5.0 vs. ≥5.1 cm). A greedy matching with a ddAC:cAC of 1:2 was performed setting a standardized deviation width of 0.20 for the logit transformation of PS as the caliper. To assess the imbalance of patient characteristics, standardized differences were calculated. For the PS-matched analysis set, risk ratios and their 95% confidence intervals (CIs) were estimated using a univariable conditional log-linear model.
For the entire analysis set before PS-matching, we also performed multivariable analysis using the log-linear model including the same covariates as in the PS estimation, multivariable analysis using the log-linear model including PS as the covariate, inverse probability treatment weighting analyses with two types of weights (so-called standard and exposed weight), and stratified analysis with five equal-sized strata according to percentiles of the PS.
SAS software, v 9.4 (SAS Institute, Cary, NC, USA), and the IBM SPSS Statistics 22.0 (SPSS Japan, Tokyo, Japan) for Windows were used for all analyses. All p-Values are reported as two-sided.
Results
PS matching. In the baseline cohort of 238 patients, 179 patients (75%) received cAC and a taxane, and 59 patients (25%) received ddAC and a taxane (Figure 1). Patients in the ddAC group were matched with patients in the cAC group at approximately 1:2. The PS-matched cohort consisted of 46 patients in the ddAC group and 86 patients in the cAC group (Figure 1). Patient characteristics before and after PS matching are presented in Table I. In the baseline cohort (n=238), the ddAC group had a lower frequency of CCI ≥3 (p=0.03) compared to the cAC group. After PS matching (n=132), there were no noteworthy differences in the variables used to construct the PS, resulting in a decrease in standardized differences to less than 0.2 (Table I).
Patient characteristics.
Percent completion rate. The overall %completion was 59.8% (128/214) in the baseline cohort and 65.9% (87/132) in the PS-matched cohort (Table II). The %completion between the ddAC group and the cAC group was not significantly different in the baseline cohort [34/54, 63.0% (95% CI=48.7-75.7%) vs. 94/160, 58.8% (95% CI=50.7-66.5%), p=0.6325 (Fisher’s exact test)] or in the PS-matched cohort [31/46, 67.4% (95% CI=52.0-80.5%) vs. 56/86, 65.1% (95% CI=54.1-75.1%), p=0.8489 (Fisher’s exact test)].
Percent completion rate.
In the univariable log-linear regression analysis for the PS-matched cohort, the ddAC regimen did not significantly reduce %completion [risk ratio (RR) for ddAC to cAC=1.035; 95% CI=0.803-1.334, p=0.5753] (Table III). The same tendency was shown using the other analyses (Table III).
Comparison of several logistic regression analyses for %completion as the outcome variable.
Secondary endpoints. The proportion of patients who achieved an RDI of 85% or greater for doxorubicin or a taxane was not different between the ddAC group and the cAC group in the PS-matched cohort (cAC vs. ddAC, proportion of RDI ≥85% for doxorubicin, 93% vs. 91%, p=0.723; cAC vs. ddAC, proportion of RDI ≥85% for a taxane, 70% vs. 71%, p=0.303) (Table IV). The proportion of treatment discontinuation or postponement of AC and/or a taxane was also not significantly different between the ddAC group and the cAC group in the PS-matched cohort, with the exception that postponement of AC was more frequent in the cAC group (22% vs. 4%, p=0.017) (Table IV). The most common reason for postponement of cAC was reduced neutrophil count on the cycle’s due date of cAC. While a numerically greater proportion of patients in the ddAC group had absolute lymphocyte counts fewer than 500 mm3 at the start of a taxane (cAC vs. ddAC, 20% vs. 35%), this difference did not achieve statistical significance (p=0.06) (Table IV).
Secondary endpoints.
The incidence of pneumonia was numerically higher in the ddAC group than in the cAC group in the PS-matched cohort [13% (6/46) vs. 3% (3/86), p=0.052] (Table V). Notably, all these pneumonias took place during the taxane phase (Table V). Although all patients recovered from pneumonia with cessation of the causative taxane with or without systemic corticosteroids and/or antibiotics (Table V), these patients had to discontinue the taxane. One case in the ddAC group was clinically diagnosed with Pneumocystis jiroveci pneumonia (PJP) and successfully treated with a trimethoprim-sulfamethoxazole combination (Table V).
Incidence of pneumonia.
Discussion
Our study demonstrated that ddAC followed by a taxane did not reduce the %completion compared to cAC followed by a taxane using PS matching (Table II and Table III). We also did not find significant differences between the cAC and ddAC groups for the other secondary endpoints related to RDI with the exception that postponement of AC was more frequent in the cAC than in the ddAC group (Table IV). Additionally, we found that the incidence of pneumonia was numerically higher in the ddAC than in the cAC group including one PJP case in the ddAC group (Table V). To our knowledge, this is the first study to compare adherence to scheduled treatments between ddAC and cAC in a setting outside of clinical trials, not only in the AC but also in the subsequent taxane phase.
It is always a big challenge in retrospective studies to know how to reduce selection bias. For example, in our study, because the ddAC regimen was employed generally for breast cancer patients with a higher risk of recurrence, there could have been an imbalance in tumor burden between the ddAC and cAC groups. We, therefore, decided to perform PS matching to reduce the bias because of confounding variables. A previous report has shown that age ≥65 years, BSA >2 m2, negative lymph nodes, and comorbidities, particularly renal disease, were predictors of reduced dose intensity in early-stage breast cancer (11). We included those factors as covariates along with additional factors such as body mass index and smoking status. However, two covariates, BSA and renal dysfunction, were not taken into account in PS matching because there were only two patients who had a BSA >2 m2 and none who had renal dysfunction in our baseline cohort. Nonetheless, PS matching enabled us to compare ddAC and cAC in well-balanced patient populations (Table I).
In our study, the proportion of patients who achieved an RDI of 85% or greater for doxorubicin in the ddAC group and the cAC group was 89% (41/46) vs. 92% (79/86), respectively, in the PS-matched cohorts. In a previous report, the same indicator was reported to be 79% (698/883) in patients who received cAC followed by a taxane (11), which was somewhat lower than that in our study. Reasons for a reduced RDI in patients who received perioperative anthracycline and a taxane could be related to the incidence of febrile neutropenia and hypersensitivity reactions to taxanes that have been reported (12). In our study, there was no discontinuation caused by febrile neutropenia during treatment with ddAC or cAC in the PS-matched cohort. In patients in the ddAC and cAC groups, dose reduction because of febrile neutropenia occurred in 2/46 (4.3%) and 4/86 (4.6%) in the AC phase, respectively, and 1/46 (2.2%) and 3/86 (3.5%) in the taxane phase, respectively. There were no patients who experienced hypersensitivity reactions to a taxane in our cohort because those who switched from one taxane to another were excluded from the study (Figure 1). A recent study from Japan found that in breast cancer patients who received FEC (5-fluorouracil, epirubicin, and cyclophosphamide) followed by weekly paclitaxel as neoadjuvant chemotherapy events causing reduced RDI occurred more frequently in the paclitaxel phase rather than more myelotoxic FEC phase (13). The study reported that the main reason for reducing RDI were grade ≥3 adverse events such as pneumonia, liver dysfunction, and fatigue (13). Our current result was consistent with the study, in that a smaller proportion of patients could maintain RDI ≥85% in the taxane phase than in the AC phase (Table IV) and that the leading causes of reducing RDI in the taxane phase were pneumonia (n=9) (Table V), neutropenia (n=6), and liver dysfunction (n=4).
We found more cases of pneumonia in the ddAC than in the cAC group. Notably four out of six pneumonias observed in the ddAC group were diagnosed as drug-induced interstitial lung disease (Table V). There has been no report indicating that ddAC is a risk factor for interstitial lung disease thus far, and we cannot exclude the possibility that the high incidence observed in our study was by chance. Conversely, the ddAC regimen has been reported to be a risk factor for PJP (14, 15). One retrospective study reported that the incidence of PJP was 13/2,057 (0.63%) in patients who underwent perioperative chemotherapy and were on the ddAC regimen, but there was no incidence of PJP (0/1001) in those who were on the cAC regimen (14). In our cohort, one patient in the ddAC group was clinically diagnosed with PJP during cycle 1 on day 21 of docetaxel treatment and was successfully treated with a trimethoprim-sulfamethoxazole combination (Table V). Although administering the trimethoprim-sulfamethoxazole combination was reported to prevent PJP during dose-dense epirubicin and cyclophosphamide (15), we did not employ this prophylaxis routinely. Because having a decreased lymphocyte count is considered a risk factor for PJP (16, 17), we compared the proportion of patients between the cAC and ddAC groups who had reduced absolute lymphocyte counts <500/mm3 at the start of taxane treatment (i.e., after AC). Although there was no statistically significant difference, a numerically greater proportion of patients in the ddAC group had an absolute lymphocyte count <500 mm3 (Table IV). Therefore, prophylactic use of the trimethoprim-sulfamethoxazole combination during not only ddAC, but also subsequent taxane treatment appears to be reasonable practice.
Our study had certain limitations. First, we could not exclude the possibility that some unknown factors that affect RDI and %completion could exist despite usage of PS matching. For instance, the choice of taxane, docetaxel or paclitaxel, could have been influenced by the toxicity observed during the preceding AC phase. Although our primary interest was to adjust baseline confounding variables using PS matching, adjustment of time-dependent confounders (e.g., the toxicity during AC) may be needed. In addition, because pegfilgrastim was not available in Japan until 2016, secondary prophylaxis for febrile neutropenia in some cAC group patients had to be done using regular filgrastim requiring daily injections, which physicians and/or patients could have chosen to avoid and instead selected dose reduction or prolongation of the treatment cycle interval. In fact, postponement of AC was more frequent in the cAC group. Second, because this study included a relatively small number of patients from a single institute, 95% of CIs of %completion had relatively large ranges, and results from other databases would vary within these ranges. Third, while the dose of docetaxel employed, 75 mg/m2, was the approved dose by the Japanese regulatory agency for breast cancer irrespective of HER2 status, it was lower than that commonly used for HER2-negative breast cancers in Western countries, which is 100 mg/m2. Forth, bi-weekly 175 mg/m2 of paclitaxel, which was adopted in the Cancer and Leukemia Group B trial 9741 as the taxane regimen following ddAC (9), was not employed in our cohort. However, The National Comprehensive Cancer Network® guidelines equally recommend ddAC followed by weekly 80 mg/m2 of paclitaxel and ddAC followed by bi-weekly 175 mg/m2 of paclitaxel (8), referring to the study (18), which showed superiority of weekly paclitaxel compared with conventional tri-weekly paclitaxel. Therefore, we believe our current findings are clinically relevant.
In conclusion, because ddAC followed by a taxane did not reduce %completion compared to cAC followed by a taxane, the ddAC regimen can be considered for high-risk breast cancer patients. Sufficient supportive measures and precautions for pneumonia including the prophylactic trimethoprim-sulfamethoxazole combination should be considered when starting treatment with ddAC followed by a taxane.
Acknowledgements
We thank Mark Abramovitz, PhD, from Edanz (https:/jp.edanz.com/ac), for editing a draft of this manuscript.
Footnotes
Authors’ Contributions
Conceptualization: Yoichi Naito, Takahiro Kogawa, and Toru Mukohara; Methodology: Kanako Mamishin, Yoichi Naito, Shogo Nomura, Gakuto Ogawa, and Toru Mukohara; Formal analysis and investigation: Kanako Mamishin, Shogo Nomura, and Gakuto Ogawa; Writing - original draft preparation: Kanako Mamishin, Shogo Nomura, Gakuto Ogawa, and Toru Mukohara; Writing - review and editing: Yoichi Naito, Kumi Niguma, Kaede Baba, Saeko Sakaeda, Hiromichi Nakajima, Shota Kusuhara, Chikako Funasaka, Takehiro Nakao, Yoko Fukasawa, Chihiro Kondoh, Kenichi Harano, Takahiro Kogawa, Nobuaki Matsubara, Ako Hosono, and Toshikatsu Kawasaki; Supervision: Toshikatsu Kawasaki and Toru Mukohara.
Conflicts of Interest
Toru Mukohara received research funding from Daiichi-Sankyo, Sysmex, MSD, Pfizer, Sanofi, and Chugai Pharmaceuticals. The other Authors declare that they have no conflicts of interest.
- Received October 7, 2021.
- Revision received November 8, 2021.
- Accepted November 10, 2021.
- Copyright © 2021 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.