Efficacy and safety of darbepoetin alfa initiated at hemoglobin ≤10 g/dL in patients with stage IV cancer and chemotherapy‐induced anemia

Abstract Data on efficacy and safety of darbepoetin alfa (DA) administered at hemoglobin (Hb) ≤10 g/dL are limited. In this analysis, we examined DA's efficacy and safety in patients with Stage IV cancers and chemotherapy‐induced anemia (CIA) initiated on DA at Hb ≤10 g/dL. Data for patients with Stage IV cancers and CIA and who initiated DA at Hb ≤10 g/dL were extracted from three phase 3 trials identified in a central database of Amgen‐sponsored DA studies in CIA. Efficacy outcomes were assessed by achievement of Hb increases of ≥1 g/dL and ≥2 g/dL and red blood cell (RBC) or whole blood transfusion requirements. Data were analyzed for all patients with baseline Hb ≤10 g/dL, and by the subgroups of patients with baseline Hb ≥9 to ≤10 g/dL versus <9 g/dL. Crude and Kaplan–Meier proportions of patients who experienced each outcome and time (days) to each outcome were summarized by treatment. Meta‐analysis (fixed‐effects inverse‐variance model) was performed to compare outcomes for DA versus placebo. Safety was assessed by occurrence of adverse events. Data from 213 patients were analyzed: DA, n = 115; placebo, n = 98. More patients in the DA versus the placebo subgroup achieved Hb increase of ≥1 g/dL (72% vs. 36%; HR: 2.92, 95% CI: 1.95, 4.39) and ≥2 g/dL (44% vs. 18%; HR: 2.98, 95% CI: 1.71, 5.21) during the first 12 treatment weeks. Median times to Hb increase of ≥1 g/dL and ≥2 g/dL were 36 days and 78 days for DA, respectively. RBC or whole blood transfusions were less common in patients in the DA versus the placebo subgroup (24% vs. 45%; HR: 0.44, 95% CI: 0.27, 0.73). No new safety issues were reported. Our results confirm that DA use in patients with Stage IV cancer and CIA is more effective than placebo at increasing Hb levels and at reducing transfusion needs when DA treatment is initiated at Hb ≤10 g/dL.


Introduction
Patients with cancer who are treated with myelosuppressive chemotherapy often develop chemotherapy-induced anemia (CIA) [1,2]. CIA often leads to the need for blood transfusion (red blood cells [RBCs] or whole blood), which is associated with a range of challenges that include ensuring safety of the blood supply (e.g., to avoid transfusion-transmitted diseases), limited availability of suitable blood, inconvenience to both patients and healthcare professionals, and associated costs [3]. Erythropoiesisstimulating agents (ESAs) such as darbepoetin alfa (DA) are among treatments that can increase serum hemoglobin (Hb) concentrations and thereby reduce the need for blood transfusions [3], providing benefit to patients and healthcare systems.
In 2008, the US Food and Drug Administration (US FDA) and the European Medicines Agency (EMA) added

ORIGINAL RESEARCH
Efficacy and safety of darbepoetin alfa initiated at hemoglobin ≤10 g/dL in patients with stage IV cancer and chemotherapy-induced anemia a boxed warning to the labels for ESAs, stating that ESAs may increase the risk of death, myocardial infarction, stroke, venous thromboembolism (VTE), thrombosis of vascular access, and tumor progression or recurrence [4,5]. At the same time, the US FDA updated DA's prescribing information, decreasing the Hb treatment initiation threshold to <10 g/dL and also added limitations of use, stating that DA was not indicated for patients receiving myelosuppressive chemotherapy when the anticipated outcome is cure [4]. The EMA made similar changes to DA's summary of product characteristics, decreasing the Hb treatment initiation and discontinuation thresholds to ≤10 g/dL and >12 g/dL, respectively [5].
Since the updates to the DA labels to add the boxed warning and change Hb initiation and discontinuation thresholds, no clinical trials have been conducted to evaluate DA's efficacy when administered per the current labels. As an example, the key registration phase 3 placebo-controlled trials of DA had included patients when their Hb was ≤11.0 g/dL, and the DA dose was withheld if a patient's Hb increased to >15.0 g/dL for men or >14.0 g/dL for women [6,7]. In these registration studies, DA administration was reinstated at 50% of the previous dose when the patient's Hb fell to ≤13.0 g/dL.
Several meta-analyses and systematic reviews have assessed DA's safety and efficacy profile in the CIA setting [8][9][10][11][12], and most have reported that DA is safe and effective in this setting, and leads to improvements in Hb levels and/or reductions in the need for blood transfusions. The recent meta-analysis by Pirker et al. [12] evaluated patient-level data from four studies in patients with disease burden varying from Stage II or lower/limited to Stage III or higher/extensive and who had initiated DA at Hb ≤10 g/dL per EMA's summary of product characteristics for DA [5]. However, no studies have evaluated the efficacy of DA when initiated at Hb ≤10 g/dL in patients with advanced disease and CIA, to meet the limitations of use for DA in the US; specifically that DA is not for use in patients with cancer receiving myelosuppressive chemotherapy when the anticipated outcome is cure [4].
In this retrospective subgroup analysis of clinical trial data, we examined the efficacy and safety of DA when administered at baseline Hb ≤10 g/dL in patients with Stage IV cancer.

Patients and study design
A review of a central database of Amgen-sponsored DA trials in CIA described in Pirker et al. [12] had identified three Amgen-sponsored randomized, placebo-controlled, double-blind, phase 3 clinical trials [7,13,14] that had included patients with Stage IV cancers and CIA who had initiated DA at Hb ≤10 g/dL. The three original studies had been approved by the relevant institutional review boards, and all patients who had participated in those studies had provided written, informed consent. . For the current analysis, data for patients with Stage IV cancer who had been initiated on DA therapy or placebo at baseline Hb ≤10 g/dL were extracted from the original studies and pooled. Data from patients with hematologic cancers were excluded, and then data from the remaining subgroup of patients were analyzed.

Study outcomes
The proportions of patients with an Hb increase of ≥1 g/dL and ≥2 g/dL and time to first Hb increase of ≥1 g/dL and ≥2 g/dL were assessed from initiation of treatment through 12 weeks. The proportions of patients receiving blood transfusions, either RBCs or whole blood, and time to first transfusion were assessed from the start of week 5 (day 29) through to week 12. Efficacy outcomes were also assessed by subgroups of patients who initiated DA treatment at Hb ≥9 to ≤10 g/dL versus those who initiated DA at Hb <9 g/dL. Safety was assessed by occurrence of adverse events (AEs).

Statistical analysis
Crude and Kaplan-Meier (K-M) proportions of patients who experienced each outcome were determined and 95% confidence intervals (CIs) were calculated; Wilson's Exact method was used for the crude CIs [15], and Greenwood's formula [16] of taking the 1 minus the survivor function at the last noncensored time point was used for timeto-event data. For the meta-analysis, hazard ratios (HRs) from the fixed-effects inverse-variance model were used to compare results for patients who received DA versus patients who received placebo, and heterogeneity between studies was reported using the I 2 statistic.

Patients
A total of 213 patients met the eligibility criteria for data extraction and analysis for this study: 100 from the Hernandez et al. study [13] that had enrolled patients with nonmyeloid malignancies (DA, n = 57; placebo, n = 43); 83 from the Vansteenkiste et al. study [7] that had enrolled patients with lung cancer (DA, n = 46; placebo, n = 37), and 30 from the Pirker et al. study [14] that had enrolled patients with extensive-stage SCLC (DA, n = 12; placebo, n = 18). Table 2 summarizes baseline demographics and clinical characteristics of the identified patients. In total, 115 patients had received DA and 98 had received placebo. Of these, 157 had baseline Hb ≥9 to ≤10 g/dL (DA, n = 82; placebo, n = 75) and 56 had baseline Hb <9 g/dL (DA, n = 33; placebo, n = 23).
Over half of the patients in both groups were men (DA, 58.3%; placebo, 58.2%). Mean (SD) age was similar in the DA and placebo subgroups (61.7 [11.0]

Hb response
In the first 12 treatment weeks, more patients in the DA subgroup versus the placebo subgroup achieved Hb increases of ≥1 g/dL and ≥2 g/dL for all patients ( Fig. 1A and B, Table 3). Similar results were observed when data were analyzed by the subgroups of baseline Hb of ≥9 to ≤10 g/ dL or <9 g/dL ( Fig. 1A and B, Table 3). Overall, fixedeffects model HRs for DA versus placebo for an Hb increase of ≥1 g/dL and ≥2 g/dL during the first 12 treatment weeks were 2.92, 95% CI: 1.95, 4.39 and 2.98, 95% CI: 1.71, 5.21, respectively ( Fig. 1C and D). By study, Hb increases of ≥1 g/dL and ≥2 g/dL were significantly different for DA versus placebo in the subsets of patients from the Hernandez et al. study [13] (nonmyeloid malignancy; n = 100) and the Vansteenkiste et al. study [7] (lung cancer; n = 83) ( Fig. 1C and D). However, this difference did not reach statistical significance in the subset of patients from the Pirker et al. study [14] (SCLC; n = 30) ( Fig. 1C and D), probably due to the small population size.
Median time to Hb increase of ≥1 g/dL was shorter with DA (36 days) versus placebo (not evaluable, since <50% of patients achieved Hb increase of ≥1 g/dL) in all patients (Fig. 1E, Table 4). Similar results were observed for the subgroup of patients who initiated DA at baseline Hb ≥9 to ≤10 g/dL or <9 g/dL (Table 4). Similarly, median time to Hb increase of ≥2 g/dL was shorter for DA versus placebo (78 days vs. not evaluable) (Fig. 1F, Table 4).

Blood transfusions
The proportion of patients who received RBC or whole blood transfusions between week 5 and week 12 was lower in the DA subgroup compared with the placebo subgroup, regardless of baseline Hb at which DA was initiated (≤10 g/ dL [all patients], ≥9 to ≤10 g/dL, or <9 g/dL) ( Fig. 2A, Table 5). Fixed-effects model HR for DA versus placebo for receiving RBC or whole blood transfusions between week 5 and week 12 was 0.44, 95% CI: 0.27, 0.73 (Fig. 2B). The K-M plot of time to receiving first RBC or whole blood transfusion between week 5 and week 12 is shown in Figure 2C. A lower proportion of patients in the DA subgroup than in the placebo subgroup received RBC or whole blood transfusions with time. Median time to receiving first transfusion was not evaluable as <50% of patients in each subgroup received transfusions (DA, 24%; placebo, 45%; Fig. 2A). For the AEs of interest, deep vein thrombosis occurred in 1 (0.9%) patient with DA versus 2 (2.0%) patients with placebo, arterial thrombosis occurred in 1 (0.9%) patient with DA versus none with placebo, venous thrombosis occurred in none of the patients with DA and 1 (1.0%) patient with placebo, and thrombosis occurred in 1 (0.9%) patient with DA and none with placebo. Darbepoetin Alfa Initiated at Hemoglobin ≤10 g/dL R. V. Boccia et al.

Discussion
To the best of our knowledge, our study is the first to evaluate the efficacy and safety of DA when treatment is initiated at baseline Hb ≤10 g/dL in patients with Stage IV cancer and CIA. Results from our study confirm that DA is more effective than placebo at increasing Hb levels and at reducing the need for RBC or whole blood transfusions in patients with Stage IV cancer and CIA when DA treatment is initiated at baseline Hb ≤10 g/dL. Similar results were observed when patients were evaluated by        subgroups of those who initiated DA at baseline Hb ≥9 to ≤10 g/dL versus <9 g/dL.
In this study, analysis of efficacy outcomes was limited to data up to 12 weeks as the DA treatment duration in one of the studies, Vansteenkiste et al. [7], was 12 weeks, with DA treatment of up to 15 weeks and 18 weeks in the studies in Hernandez et al. [13] and Pirker et al. [14], respectively ( Table 1). As demonstrated in earlier studies, DA treatment for up to 12 weeks can result in achievement of efficacy outcomes including Hb increases and reductions in the need for blood transfusions [6,7,[17][18][19]. RBC or whole blood transfusion needs were assessed starting from week 5 (day 29) to allow adequate time for DA-stimulated erythropoiesis and production of sufficient RBCs to avoid the need for transfusion [17][18][19].
In this study, we evaluated data from subsets of patients with Stage IV cancer who had participated in the studies reported in Hernandez et al. [13], Vansteenkiste et al. [7], and Pirker et al. [14]. A recent meta-analysis by Pirker et al. [12] evaluated data from patients who had participated in the same three studies [7,13,14] and an additional fourth study by Hedenus et al. [6], focusing on a more heterogeneous patient population that included patients with early-stage to advanced disease. Consistent with the findings from this study, Pirker et al. [12] reported that more patients treated with DA versus placebo achieved Hb increases of ≥1 g/dL (HR: 2.07, 95% CI: 1.62, 2.63) and ≥2 g/dL (HR: 2.91, 95% CI: 2.09, 4.06), and that less patients treated with DA versus placebo received blood transfusions (HR: 0.58, 95% CI: 0.44, 0.77). Also, that study demonstrated similar efficacy outcomes regardless of whether DA was initiated at baseline Hb 9 to ≤10 g/ dL or <9 g/dL. Similar findings were reported from a post hoc analysis of a phase 3 study in patients with nonmyeloid malignancies and CIA who initiated DA treatment at baseline Hb <10 g/dL, with K-M transfusion incidences of 36% for DA administered every 3 weeks and 41% for DA administered every week [20]. These observations demonstrate DA's effectiveness in improving Hb levels and decreasing transfusion needs in patients with both early-stage as well as advanced disease. Data from our study are also consistent with data reported in studies of DA in patients where treatment was initiated at a broader range of baseline Hb, including those in which DA treatment was initiated at baseline Hb >10 g/dL [7,13,[21][22][23].
Overall, no new safety issues associated with DA use were identified in this study, with most AEs occurring with similar frequencies in the DA and placebo subgroups or at higher frequencies in the placebo subgroup versus the DA subgroup. AEs that occurred at a slightly higher frequency in the DA subgroup versus the placebo subgroup include back pain, peripheral edema, and pyrexia.
Of note, deep vein thrombosis occurred in two patients in the placebo subgroup and one patient in the DA subgroup, and venous thrombosis occurred in one patient with placebo and in none of the patients with DA.  1 Hb measurements within 28 days after a blood transfusion were excluded from the analysis. 2 95% CIs were calculated based on Wilson's Exact method.
CI, confidence interval; DA, darbepoetin alfa; Hb, hemoglobin; K-M, Kaplan-Meier; RBC, red blood cell. DA, darbepoetin alfa. 1 One patient randomized to DA did not receive treatment, and was included with the placebo group for safety analyses.