help button home button The Oncologist http://theoncologist.alphamedpress.org/misc/eLetters.shtml
HOME HELP CONTACT US SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The Oncologist, Vol. 10, No. 6, 438-448, June 2005; doi:10.1634/theoncologist.10-6-438
© 2005 AlphaMed Press

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow eLetters: Submit a response to this article
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow E-mail this article link to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Reprints/Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Duh, M. S.
Right arrow Articles by Waltzman, R. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Duh, M. S.
Right arrow Articles by Waltzman, R. J.

Assessing the Clinical Benefits of Erythropoietic Agents Using Area Under the Hemoglobin Change Curve

Mei Sheng Duha, Patrick Lefebvreb, John Fastenauc, Catherine Tak Piechc, Roger J. Waltzmand

a Analysis Group, Inc., Boston, Massachusetts, USA; b Groupe d’Analyse, Ltée., Montréal, Canada; c Ortho Biotech Clinical Affairs, LLC, Bridgewater, New Jersey, USA; d St. Vincent’s Comprehensive Cancer Center, New York, New York, USA

Correspondence: Correspondence: Mei Sheng Duh, M.P.H., Sc.D., Analysis Group, Inc., 111 Huntington Avenue, Tenth Floor, Boston, Massachusetts 02199, USA. Telephone: 617-425-8131; Fax: 617-425-8001; e-mail: mduh{at}analysisgroup.com


   ABSTRACT
Top
 Abstract
Introduction
 Methods
 Results
 Discussion
 Conclusion
 Disclosure of Potential...
 References
 
Introduction. In assessing erythropoietic agents for chemotherapy-induced anemia, traditional single time–point end points (e.g., hematopoietic response [HR]) fail to reflect clinical benefits over the entire therapy course. Area under the hemoglobin change curve (Hb AUC) is introduced as an alternative measure, and its reliability, clinical significance, and superiority are assessed.

Methods. Using data from a phase IV open-label epoetin alfa (EPO) trial, we tested Hb AUC reliability by comparing its values derived from primary patient data with those derived from aggregated data. Clinical significance of the Hb AUC was investigated in three phase IV EPO trials by examining the linear relationship between Hb AUC quartiles and established clinical end points. The superiority of the Hb AUC over HR in its association with blood transfusion was tested through logistic regressions and area under the receiver operating characteristic (ROC) curve analysis.

Results. The Hb AUC values derived from patient and aggregated data were similar. Strong and statistically significant linear trends of decreasing transfusion requirements, increasing quality-of-life improvements, and decreasing time to HR were found across Hb AUC quartiles. The Hb AUC rendered the HR variable insignificant when both were present in the same model. Area under the ROC curve analysis supported the superior performance of the Hb AUC.

Conclusions. We found that the Hb AUC is an objective, reliable, clinically meaningful, and comprehensive summary statistic that may be used to quantify clinical benefits for patients receiving erythropoietic agents. Further prospective validation of the Hb AUC metric is recommended.

Key Words. Anemia • Area under the curve • erythropoietin • chemotherapy-induced anemia


   INTRODUCTION
Top
 Abstract
 Introduction
Methods
 Results
 Discussion
 Conclusion
 Disclosure of Potential...
 References
 
Cancer patients with chemotherapy-induced anemia (CIA) typically experience fatigue, shortness of breath, diminished functional status, and suboptimal quality of life (QOL) [1]. Fatigue, one of the most debilitating effects of cancer and its treatment [2], can develop relatively quickly after the onset of chemotherapy [3,4]; it has been recognized that prompt and sustained relief of this symptom of anemia is an important goal of supportive treatment [5–10]. There is little doubt that a relationship exists among hemoglobin (Hb) level, fatigue, and QOL, and that raising Hb levels can relieve the symptom of fatigue and improve QOL [2, 11, 12]. However, investigations of erythropoietic agents have only just begun to consider the clinical importance of prompt Hb improvement. In a recent retrospective analysis of data from two large, open-label clinical studies [13,14], Rosberg and colleagues [15] observed that clinical outcomes—including patient QOL, transfusion requirements, and average weekly dose—improved in patients who responded early to epoetin alfa (EPO; Procrit®; Ortho Biotech Products, L.P., Bridgewater, NJ, http://www.orthobiotech.com), by week 4, compared with late or nonresponders. In addition, patients with CIA receive erythropoietic therapies for relatively short periods [16–19] and are sensitive to the time constraints imposed on their lives as a result of their cancer diagnosis. Glaspy et al. [9] recognized a front-loading dosing strategy as one that could potentially enhance the benefit of treatment with an erythropoietic agent, minimizing the time to achieve a hematopoietic response so as to avoid transfusion risks and gain symptom relief as quickly as possible. With these observations in mind, the rate of Hb response is an important consideration in caring for the patient with cancer, and it is a clinically relevant parameter with which to judge the effectiveness of an erythropoietic agent.

Existing studies of erythropoietic agents for CIA have been conducted with different patient populations and have used different methodologies, most frequently single time point-based and threshold-based end points of clinical efficacy [6, 7, 14, 20, 21–23]. These end points have included: A) Hb response, defined as an increase from baseline of ≥2 g/dl occurring at any time point in the study in the absence of transfusions, usually within the preceding 28 days; B) hematopoietic response (HR), defined as an Hb increase from baseline of ≥2 g/dl or an Hb level ≥12 g/dl occurring at any time point in the study in the absence of transfusions, usually within the preceding 28 days; C) mean change in Hb from baseline measured at specified time points, usually at monthly intervals; or D) RBC transfusion requirements over specified time intervals.

However, a study by Cremieux et al. [24], comparing the cost-effectiveness of EPO with that of the standard-of-care in patients with CIA, supported the idea that cumulative Hb gains measured over the entire course of therapy may be a more useful measure, as cumulative gains did not mask differences in trends between the two interventions over the course of the study. In addition, if one accepts that early symptom relief is an important therapeutic goal, threshold-based measurements such as HR provide little indication of the onset of patients’ responses, and so insufficiently distinguish between agents or regimens that provide early benefit. Thus, the ideal method would reveal strong correlations between objective laboratory measures of efficacy (e.g., Hb change) and clinical benefits (e.g., reduced transfusion needs and improvements in QOL), and capture the concept of time-to-response. The area under the Hb change curve (Hb AUC) method is one such approach, in which the cumulative effects of erythropoietic agents can be assessed from the beginning to the end of therapy. An AUC measure enables clinicians to evaluate erythropoietic agents based on the benefits to patients over the entire course of therapy and improves the detection of early benefit, particularly if threshold-based responses are equivalent at the end of therapy. In fact, Egrie and colleagues [25] indicated that the AUC provided the most robust measure of biological activity of erythropoietic agents, as compared with mean Hb values at single time points or imposing linearity in log dose-response curves.

Pharmacokinetic studies have been the traditional avenue for the application of the principles associated with AUC measurements, in which AUC is defined as the total area under the curve describing drug concentration in the systemic circulation as a function of time [26]. In broad applications, studies have used AUC extensively to monitor, assess, and compare the effects of various therapeutic entities [27–31]. Regulatory agencies are also accepting of AUC methods. For example, a guidance issued by the Center for Drug Evaluation and Research (CDER) directing the clinical evaluation of analgesic drugs suggested that the cumulative effect, as well as peak effects and time-to-peak effects, of such drugs could be assessed using the AUC [32]. Other examples exist in the literature that confirm the utility of AUC methods in evaluating the clinical effectiveness of drugs and biologics [33–36].

With specific regard to erythropoietic agents, studies have used AUC as a metric for evaluating equivalence [37], hematologic benefits [38], and dose conversion ratios [39], among other variables [40,41]. Illustrated with Hb change data following treatment with erythropoietic agents, this manuscript introduces the concepts and rationale for the use of Hb AUC as an efficacy measure. Additionally, it is herein shown that the Hb AUC can also be used to facilitate an appropriate comparison of erythropoietic agents. However, readers should bear in mind that Hb-based end points, whether the continuous Hb AUC or the discrete HR, need to be interpreted with caution due to other confounding host factors (e.g., nutritional deficiencies [42,43], patient or tumor characteristics [42, 44–46], depression or other psychiatric diagnoses [42, 46, 47], insomnia [42, 43, 48], and infections [42,43]) that may correlate with the multifactorial causes of CIA or its associated fatigue.

The objectives of this study were fourfold: A) to demonstrate how the Hb AUC can be calculated; B) to demonstrate that the Hb AUC can be reliably calculated using either published aggregated or primary patient-level data; C) to demonstrate that the Hb AUC correlates with accepted measures of clinical benefit; and D) to demonstrate that the Hb AUC, as a comprehensive continuous variable, is more informative than a discrete single time-point measure.


   METHODS
Top
 Abstract
 Introduction
 Methods
Results
 Discussion
 Conclusion
 Disclosure of Potential...
 References
 
Calculating the Hb AUC
The trapezoidal rule is the simplest and most common approach used to calculate AUC [49], in which the total area under the Hb change versus time curve is obtained by summation of each individual area between two consecutive time intervals. This AUC value approximates the integral of Hb change values over time obtained from values of the ordinate and the abscissa [50]. The slope, or the first derivative of the Hb AUC curve, represents the rate of rise of Hb.

Figure 1Go illustrates the Hb AUC calculation using the trapezoidal rule by aggregating data from a phase IV, multicenter, open-label study [14], in which 2,964 patients received EPO once weekly (QW) at a dose of 40,000 units over the course of 16 weeks. Dose escalations to 60,000 units QW were made if, at study week 4, the patient’s Hb level had not increased by at least 1 g/dl over baseline. We calculated the 16-week Hb AUC value as follows:



View larger version (12K):
[in this window]
[in a new window]
  		Figure 1. Area under the hemoglobin change curve (Hb AUC) calculation. Calculated from the mean change in Hb of the on-treatment population (i.e., missing values were not imputed), where Hb measurements recorded within 28 days of an RBC transfusion were considered missing (337 patients who had RBC transfusions within 28 days of baseline were excluded). In this illustration, the Hb AUC was 24.0 g/dl over 16 weeks, computed from {sum} (areas A + ... + G), where:
A = 4 x 1.0 x 1/2 = 2.0 g x week/dl
B = (8 – 4) x (1.8 – 1.0) x 1/2 = 1.6 g x week/dl
C = (8 – 4) x 1.0 = 4.0 g x week/dl
D = (12 – 8) x (2.1 – 1.8) x 1/2 = 0.6 g x week/dl
E = (12 – 8) x 1.8 = 7.2 g x week/dl
F = (16 – 12) x (2.2 – 2.1) x 1/2 = 0.2g x week/dl
G = (16 – 12) x 2.1 = 8.4 g x week/dl

 
  1. We constructed the curve by plotting the mean change in Hb from baseline at each time point and drawing straight lines to connect the time points. In plotting these data, Hb measurements recorded within 28 days of an RBC transfusion were considered missing to avoid attributing the Hb effect of transfusions to that of EPO therapy.
  2. We divided the entire area under the curve into rectangles and triangles of measurable dimensions (A through G), as shown in Figure 1Go.
  3. We calculated each area by using simple geometric formulas, and summed as follows:
    1. Area of each triangle (A, B, D, F) = base x height x 1/2
    2. Area of each rectangle (C, E, G) = base x height
    3. Summation of areas A through G = Hb AUC

In the aforementioned study [14], the calculated Hb AUC was 24.0 g/dl, based on monthly average Hb changes, independent of blood transfusion, observed over the 16-week course of treatment in 2,627 on-treatment patients who had no RBC transfusions within 28 days prior to baseline.

The accuracy of the trapezoidal method in calculating the Hb AUC hinges on the number of time points available; the more time points, the more accurate the measure. However, if the time points are spaced widely, meaning that there are few of them, the imprecise curvature of the line will cause a greater error in the area estimate [49].

In addition to this example, we presented the Hb AUC values using the trapezoidal rule based on aggregated-level data from other published studies of EPO [6, 7, 13, 51] and darbepoetin alfa (DARB; Aranesp®; Amgen Inc., Thousand Oaks, CA, http://www.amgen.com) [52,53] as well.

The Reliability of the Hb AUC
We determined the reliability of the Hb AUC using data from the aforementioned study [14], by comparing the Hb AUC values derived from patient-level data with those derived from aggregated-level data using two statistical approaches for the treatment of missing Hb values. In the first analysis, we used an on-treatment approach without imputing any missing Hb values. In the second analysis, we used the last-value-carry-forward (LVCF) principle to impute missing Hb measurements. For the purpose of calculating the Hb AUC, all Hb measurements recorded within 28 days of an RBC transfusion were considered missing. For aggregated-level data, we determined Hb AUC values using aggregated mean changes in Hb over baseline at weeks 4, 8, 12, and 16. For patient-level data, we calculated the Hb AUC for each patient and summarized the mean Hb AUC across all patients.

To determine how the frequency of Hb values could affect Hb AUC, we used three samples of patients that had different frequencies of Hb measurements during the study. First, we analyzed all patients eligible for the study (n = 2,964). Second, we calculated the Hb AUC for the subset of patients that had nonmissing Hb values at the study end (n = 1,622). Finally, we examined patients with nonmissing Hb measurements for all visits (i.e., baseline and weeks 4, 8, 12, and 16; n = 1,266).

The Clinical Significance of the Hb AUC
We investigated the clinical significance of the Hb AUC by examining the correlations between the Hb AUC and several established measures of clinical benefit. The clinical parameters with which the Hb AUC were benchmarked included RBC transfusion rates, patient-reported QOL, and time to achieve an HR. Patient QOL was evaluated based on the overall Linear Analogue Scale Assessment (LASA) and the Functional Assessment of Cancer Treatment subscales for the anemia (FACT-An) and fatigue (FACT-F) parameters.

We used data from multiple EPO studies to ensure the replicability of the findings. Individual patient data from three phase IV, multicenter, single-arm, open-label, 16-week EPO trials were used; two used an initial EPO dose of 40,000 units QW [13,14], and one used an initial dose of 10,000 units thrice weekly (TIW) [7]. For each patient, we calculated the Hb AUC by using the trapezoidal method that accumulated the Hb AUC from baseline to week 16 of the EPO therapy, with Hb values recorded within 28 days of an RBC transfusion considered missing. We used the LVCF principle to impute missing Hb values, and the Hb AUC was formulated as a categorical variable based on quartile distribution (0–25th, 26th–50th, 51st–75th, and 76th–100th percentiles).

For binary outcomes (i.e., RBC transfusion rate), the Cochron-Armitage trend test was used to determine whether there was a significant decreasing linear trend of percentages of patients transfused across Hb AUC quartiles. For continuous variables (i.e., QOL change scores, time to HR), the t-statistic was used to test whether there was a significant increasing/decreasing linear trend of the means of the outcome across Hb AUC quartiles.

Potential Advantages of Hb AUC Methodology Over Traditional Discrete Time-Point Hb Measures
To demonstrate that Hb AUC is a more informative measure than HR, we compared Hb AUC with HR in the degree of association with the blood transfusion rate, using patient-level data from an aforementioned study that used an initial EPO dose of 40,000 units QW (n = 2,964) [14].

Two different analyses were undertaken. First, the dependent variable of interest, the RBC transfusion rate, a binary variable indicating whether a patient received RBCs during the 16-week trial, was logistically regressed on both the continuous Hb AUC and the binary HR variables, along with other covariates for adjustment (i.e., age, gender, baseline Eastern Cooperative Oncology Group [ECOG] performance status score, baseline FACT-An score, primary site of malignancy, chemotherapeutic agent, and radiation therapy dummy). The simultaneous presence of the Hb AUC and the HR variables in the model allowed the two variables to compete with each other to explain the transfusion outcome. That is, the variable that was stronger in the association with blood transfusion would retain its statistical significance, while the other would be rendered insignificant.

In the second analysis, we conducted two separate logistic regression models for the RBC transfusion outcome. Thus, while the same adjusting covariates as defined above were included for both models, one model included the Hb AUC while the other included the HR as an independent variable. We then determined the accuracy of the Hb AUC-based versus the HR-based model in predicting RBC transfusion using the area under the receiver operating characteristic (ROC) curve [54]. We constructed ROC curves by computing the sensitivities and specificities of the Hb AUC-based and HR-based models in predicting RBC transfusion rate. The area under the ROC curve is a measure of the probability that the model (Hb AUC-based or HR-based) will correctly classify whether a patient received a blood transfusion. ROC curves show the discriminatory ability of a model by the position of the curve; the farther upward and to the left the curve lies in the graph, the greater the area, and the better the performance of the model. Figure 2Go shows that a perfect model has an area under the ROC curve equal to one; a classification model with no predictive power, as it is equal to randomly assigning patients to receiving blood transfusion, has an area under the curve equal to 0.5, and is described by the 45-degree identity line. Swets suggested that areas of 0.5 to <0.7 reflect low accuracy, 0.7 to <0.9 reflect moderate accuracy, and 0.9 or above reflect high accuracy [55]. The areas under the ROC curves from the two models were compared using the nonparametric Mann-Whitney statistic [56].



View larger version (17K):
[in this window]
[in a new window]
  		Figure 2. Example of receiver operating characteristic (ROC) graph. The "perfect," "worthless," and "intermediate" classification models are depicted. A "perfect" model correctly predicts the outcome status for every subject, resulting in the area under the ROC curve equal to 1. A "worthless" model is analogous to tossing a coin in predicting the outcome status for each subject, resulting in an area under the ROC curve equal to 0.5. The "intermediate" classification model falls between the "perfect" and "worthless" models.

 

   RESULTS
Top
 Abstract
 Introduction
 Methods
 Results
Discussion
 Conclusion
 Disclosure of Potential...
 References
 
Calculating the Hb AUC
The Hb AUC values calculated from available published studies of EPO and DARB are presented in Table 1Go. The front-loading dosing regimens of DARB reached a similar early Hb improvement at week 4 to those in the EPO studies, resulting in similar Hb AUC values. In contrast, the DARB Q2W study had a lower week 4 Hb change than the other studies, leading to a lower Hb AUC. This suggests that the Hb AUC is more sensitive than the end of study Hb measurement in detecting early Hb response and reflects the comprehensive Hb effect over the entire course of treatment.


View this table:
[in this window]
[in a new window]
  		Table 1. Hematologic benefita with erythropoietic agents from studies of varying design
 
The Reliability of the Hb AUC
Table 2Go presents the comparison of aggregated-level versus patient-level Hb AUC values from a previously mentioned phase IV, multicentered, open-label study [14] for three samples of patients with different frequencies of Hb measurements (i.e., all patients, patients with nonmissing Hb measurement at study end [completers], and patients with nonmissing Hb measurements for all visits). In addition, we show the data across the two different statistical approaches in treating missing Hb values—on-treatment refers to the approach in which missing measurements were not imputed, while LVCF refers to the approach in which the LVCF principle was used to impute the missing values.


View this table:
[in this window]
[in a new window]
  		Table 2. Reliability of the Hb AUC using aggregated-level versus patient-level dataa
 
Aggregated-level and patient-level Hb AUC values were mathematically identical when missing Hb readings were imputed using the LVCF method. Moreover, on-treatment Hb AUC values calculated from aggregated-level data were systematically equal to or higher than LVCF Hb AUC values, indicating that the LVCF principle is a conservative approach.

Comparisons of the Hb AUC values across the three different samples showed that the Hb AUC value tended to be higher when more Hb points were available, while the variability of Hb AUC values was smaller when on-treatment data were used compared with LVCF data.

This information suggests that, when LVCF data are used, the Hb AUC can be reliably derived using published data, as its value is identical to that using primary patient data. If published data report Hb changes for on-treatment patients, the estimated Hb AUC value would tend to be higher than that based on a LVCF approach.

The Clinical Significance of the Hb AUC
The relationships between the Hb AUC quartiles and the RBC transfusion rate, patient-reported QOL, and time to achieve HR across the three EPO studies [7, 13, 14] are shown in Figure 3AGo–3E, respectively.







View larger version (57K):
[in this window]
[in a new window]
  		Figure 3. Relationship between Hb AUC values and clinical outcomes. In three studies in which EPO was administered either at a starting dose of 40,000 units once weekly (study 1 [14] and study 3 [13]) or 10,000 units thrice weekly (study 2 [7]), the observed linear trends of decreasing transfusion rate (A), increasing patient QOL (B–D), and decreasing time to hematopoietic response (E) with increasing Hb AUC quartile were statistically significant (p < .0001, p < .0001, p < .0001, respectively). LASA scores ranged from 0–100 (mm). FACT-An and FACT-F scores have a maximum of 80 and 52 points, respectively. Data were not available from study 3 to calculate FACT-An and FACT-F scores. Abbreviations: Hb AUC = area under the hemoglobin change curve; LASA = Linear Analogue Scale Assessment; FACT = Functional Assessment of Cancer Therapy (-An = Anemia; -F = Fatigue); EPO = epoetin alfa.

 
Consistently observed across all three EPO trials, increasing Hb AUC values were well-correlated with statistically significantly improved clinical outcomes, including a decreasing linear trend of transfusion (p < .0001), an increasing linear trend of QOL improvement (p < .0001), and a decreasing linear trend of time to HR (p < .0001).

Potential Advantages of Hb AUC Methodology Over Traditional Discrete Time-Point Hb Measures
Table 3Go presents the results from the logistic regression model that permitted Hb AUC and HR to compete simultaneously to explain the risk of RBC transfusion during the study. The regression resulted in an insignificant coefficient on HR (p = .5181), but a significant coefficient on Hb AUC (p < .0001). Thus, the model revealed that the Hb AUC variable could explain the risk of transfusion with statistical significance even in the presence of an HR variable. In contrast, when controlling for Hb AUC, the binary HR variable failed to explain the risk of RBC transfusion with statistical significance. These results indicate that the association between Hb AUC and RBC transfusion is stronger than the corresponding association between HR and RBC transfusion.


View this table:
[in this window]
[in a new window]
  		Table 3. Multivariate analysis of independent risk factors for RBC transfusion
 
In a further confirmatory analysis, ROC curves derived from Hb AUC-based and HR-based logistic regressions were compared (Fig. 4Go). The area under the ROC curve for the Hb AUC-based model was 0.74, compared with 0.69 for the HR-based model. These two areas were found to be statistically significantly different (p < .0001), indicating that the Hb AUC-based model was more accurate in predicting RBC transfusion than the HR-based model.



View larger version (14K):
[in this window]
[in a new window]
  		Figure 4. ROC curve analysis: comparison of the Hb AUC-based versus HR-based models in predicting the probability of RBC transfusion. The ROC curves for the Hb AUC-based and HR-based models are pictured, along with an ROC curve from a hypothetical "worthless" model that discriminates the probability of blood transfusion by chance alone (described by the 45-degree identity line). The greater the area under the ROC curve, the more accurate a model is in predicting the probability of RBC transfusion. More details on the interpretation of the ROC curves are presented in Methods and in Figure 2Go. Abbreviations: Hb AUC = area under the hemoglobin change curve; Hb = hemoglobin; HR = hematopoietic response.

 

   DISCUSSION
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
Conclusion
 Disclosure of Potential...
 References
 
In this retrospective study using data from multiple phase IV EPO trials, we have demonstrated that the Hb AUC is an objective, comprehensive, reliable, and clinically meaningful alternative measure for hematological efficacy of erythropoietic agents. We found that the Hb AUC reflected early Hb response as well as the comprehensive Hb effect from the entire course of the erythropoietic therapy better than the end of study Hb measurement. The Hb AUC can be reliably approximated using either published or primary data. Its validity as a laboratory Hb end point is supported by its strong correlation with clinical end points (i.e., RBC transfusion rate, patient-reported QOL, and time to achieve HR). Moreover, the Hb AUC is potentially more advantageous than the traditional discrete single time-point measure of HR, as it provides more information as a continuous variable. We showed that when the Hb AUC and HR variables were present in a regression model simultaneously, the Hb AUC overwhelmed the HR in explaining the probability of an RBC transfusion. Furthermore, when the Hb AUC and HR were separately inserted into a regression model, the model with the Hb AUC, all else equal, outperformed the model with the HR in the degree of association with RBC transfusion.

The AUC methodology that we introduced in this paper can be further developed and refined in its application to the assessment of erythropoietic agents through selecting the optimal number of time points and replicates per time point to obtain the most accurate and precise estimate of the Hb AUC [57], as well as through penalizing overly high Hb readings that may be associated with safety concerns in the construction of the Hb AUC. In addition, corrected Hb AUC, defined as the total Hb AUC per unit of time, can be used to adjust for different observation durations when comparing across studies of different lengths. Furthermore, methods for Hb AUC calculation in an observational study setting, in which the extent of missing values in Hb data is usually greater than that in a clinical trial setting, need to be enhanced as well.

When comparing EPO with DARB from available published aggregated data, the Hb AUC of EPO was higher than that of DARB. As shown in Table 1Go, the Hb AUC of DARB approached the Hb AUC of EPO studies only when front-loading dosing strategies of DARB were used. However, it is important to note that the studies compared in Table 1Go were conducted at different times, with populations that had different baseline characteristics, tumor types, and treatment regimens, for which adjustments were not made.

The Hb AUC measure introduced herein needs to be further validated in a prospective randomized trial that compares EPO and DARB users arising from the same source population. However, to the extent that it is valid to compare Hb AUC results across different studies, inspection of representative curves that illustrate Hb effect over time is instructive and demonstrates a worthwhile concept to be incorporated into future study designs.


   CONCLUSION
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Conclusion
Disclosure of Potential...
 References
 
The Hb AUC measure is an objective, reliable, valid, and comprehensive summary statistic that may be used to quantify clinical benefits for patients receiving erythropoietic agents. Higher Hb AUC values correlate well with other measures of erythropoietic efficacy, such as lower transfusion requirements, improved patient QOL, and faster time to HR. Hb AUC has also preliminarily been found to be superior to the traditional HR measure in its association with the RBC transfusion clinical end point. The Hb AUC metric may be considered in future studies as an alternative end point with which to assess the overall efficacy of erythropoietic agents. Hb AUC measurements may provide a means to compare the efficacies of erythropoietic agents throughout the entire course of therapy. Being a more comprehensive reflection of erythropoietic efficacy than any single time point-based or threshold-based measure, the Hb AUC is expected to augment clinicians’ informed decision making when choosing a therapy most likely to improve their patients’ well-being.


   DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Conclusion
 Disclosure of Potential...
References
 
This research was sponsored by Ortho Biotech Clinical Affairs, LLC. Dr. Waltzman, Dr. Duh, and Dr. Lefebvre are consultants for Ortho Biotech. Dr. Fastenau and Dr. Tak Piech are employees of Ortho Biotech.


   REFERENCES
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Conclusion
 Disclosure of Potential...
 References
 

  1. Groopman JE, Itri LM. Chemotherapy-induced anemia in adults: incidence and treatment. J Natl Cancer Inst 1999;91:1616–1634.[Abstract/Free Full Text]
  2. Sobrero A, Puglisi F, Guglielmi A et al. Fatigue: a main component of anemia symptomatology. Semin Oncol 2001;28(suppl 8):15–18.[Medline]
  3. Schwartz AL. Daily fatigue patterns and effect of exercise in women with breast cancer. Cancer Pract 2000;8:16–24.[CrossRef][Medline]
  4. Visovsky C, Schneider SM. Cancer-related fatigue. J Issues Nurs 2003;8:1–23.
  5. Cella D, Zagari MJ, Vandoros C et al. Epoetin alfa treatment results in clinically significant improvements in quality of life in anemic cancer patients when referenced to the general population. J Clin Oncol 2003;21:366–373.[Abstract/Free Full Text]
  6. Glaspy J, Bukowski R, Steinberg D et al. Impact of therapy with epoetin alfa on clinical outcomes in patients with non-myeloid malignancies during cancer chemotherapy in community oncology practice. Procrit Study Group. J Clin Oncol 1997;15:1218–1234.[Abstract/Free Full Text]
  7. Demetri GD, Kris M, Wade J et al. Quality-of-life benefit in chemotherapy patients treated with epoetin alfa is independent of disease response or tumor type: results from a prospective community oncology study. Procrit Study Group. J Clin Oncol 1998;16:3412–3425.[Abstract]
  8. Fallowfield L, Gagnon D, Zagari M et al. Multivariate regression analyses of data from a randomised, double-blind, placebo-controlled study confirm quality of life benefit of epoetin alfa in patients receiving non-platinum chemotherapy. Br J Cancer 2002;87:1341–1353.[CrossRef][Medline]
  9. Glaspy JA, Jadeja JS, Justice G et al. A randomized, active-control, pilot trial of front-loaded dosing regimens of darbepoetin-alfa for the treatment of patients with anemia during chemotherapy for malignant disease. Cancer 2003;97:1312–1320.[CrossRef][Medline]
  10. Jumbe NL, Doshi S, Yao B et al. Darbepoetin alfa rational dose/schedule evaluation based on quantitative understanding of erythropoiesis for early and sustained alleviation of anemia. Proc Am Soc Clin Oncol 2003;22:765.
  11. Sabbatini P. The relationship between anemia and quality of life in cancer patients. The Oncologist 2000;5(suppl 2):19–23.[Abstract/Free Full Text]
  12. Ross SD, Fahrbach K, Frame D et al. The effect of anemia treatment on selected health-related quality-of-life domains: a systematic review. Clin Ther 2003;25:1786–1805.[CrossRef][Medline]
  13. Shasha D, George MJ, Harrison LB. Once-weekly dosing of epoetin-alfa increases hemoglobin and improves quality of life in anemic cancer patients receiving radiation therapy either concomitantly or sequentially with chemotherapy. Cancer 2003;98:1072–1079.[CrossRef][Medline]
  14. Gabrilove JL, Cleeland CS, Livingston RB et al. Clinical evaluation of once-weekly dosing of epoetin alfa in chemotherapy patients: improvements in hemoglobin and quality of life are similar to three-times-weekly dosing. J Clin Oncol 2001;19:2875–2882.[Abstract/Free Full Text]
  15. Rosberg J, Lefebvre P, Fastenau J et al. Clinical significance of a ≥1 gram/deciliter (g/dL) rise in hemoglobin (Hb) at week 4 (early response) or at week 8 during epoetin alfa (EPO) treatment. Blood 2003;102:510b.
  16. Fahrbach KR, Frame D, Scheye R et al. Comparison of anemia treatment outcomes in patients with cancer receiving epoetin alfa and darbepoetin alfa. Blood 2003;102:510b.
  17. Mark TL, McKenzie SR, Piech CT. Better early and overall hematologic outcomes and lower drug cost with epoetin alfa (EPO) compared with darbepoetin alfa (DARB) in patients with chemotherapy-related anemia. Poster presented at the Academy of Managed Care Pharmacy’s (AMCP) 16th Annual Meeting and Showcase, San Francisco, CA, March 31–April 3, 2004.
  18. Papatheofanis FJ, Fastenau JM, Chiang TH et al. Epoetin alfa and darbepoetin alfa anemia treatment outcomes in cancer patients from a veterans administration perspective. Poster presented at the International Society for Pharmacoeconomics and Outcomes Research’s (ISPOR) 9th Annual International Meeting, Arlington, VA, May 16–19, 2004.
  19. Papatheofanis FJ, Fastenau JM, Chiang TH et al. Management of chemotherapy-related anemia: a study of epoetin alfa and darbepoetin alfa use in a self-insured managed care organization. Poster presented at the Academy of Managed Care Pharmacy’s (AMCP) 16th Annual Meetings and Showcase, San Francisco, CA, March 31–April 3, 2004.
  20. Vansteenkiste J, Pirker R, Massuti B et al. Double-blind, placebo-controlled, randomized phase III trial of darbepoetin alfa in lung cancer patients receiving chemotherapy. J Natl Cancer Inst 2002;94:1211–1220.[Abstract/Free Full Text]
  21. Littlewood TJ, Bajetta E, Nortier JW et al. Effects of epoetin alfa on hematologic parameters and quality of life in cancer patients receiving nonplatinum chemotherapy: results of a randomized, double-blind, placebo-controlled trial. J Clin Oncol 2001;19:2865–2874.[Abstract/Free Full Text]
  22. Hedenus M, Adriansson M, San Miguel J et al. Efficacy and safety of darbepoetin alfa in anaemic patients with lymphoproliferative malignancies: a randomized, double-blind, placebo-controlled study. Br J Haematol 2003;122:394–403.[CrossRef][Medline]
  23. Glaspy JA, Jadeja JS, Justice G et al. Darbepoetin alfa given every 1 or 2 weeks alleviates anaemia associated with cancer chemotherapy. Br J Cancer 2002;87:268–276.[CrossRef][Medline]
  24. Cremieux PY, Finkelstein SN, Berndt ER et al. Cost effectiveness, quality-adjusted life-years and supportive care. Recombinant human erythropoietin as a treatment of cancer-associated anaemia. Pharmacoeconomics 1999;16:459–472.[CrossRef][Medline]
  25. Egrie JC, Dwyer E, Browne JK et al. Darbepoetin alfa has a longer circulating half-life and greater in vivo potency than recombinant human erythropoietin. Exp Hematol 2003;31:290–299.[CrossRef][Medline]
  26. Wilkinson GR. Pharmacokinetics: the dynamics of drug absorption, distribution, and elimination. In: Hardman JG, Limbird LE, Gilman AG, eds. Goodman & Gilman’s The Pharmacological Basis of Therapeutics, Tenth Edition. New York: McGraw-Hill, 2001:3–29.[Medline]
  27. Kahan BD, Keown P, Levy GA et al. Therapeutic drug monitoring of immunosuppressant drugs in clinical practice. Clin Ther 2002;24:330–350.[CrossRef][Medline]
  28. Oleson FB Jr, Berman CL, Kirkpatrick JB et al. Once-daily dosing in dogs optimizes daptomycin safety. Antimicrob Agents Chemother 2000;44:2948–2953.[Abstract/Free Full Text]
  29. Kim HC, McMillan CW, White GC et al. Purified factor IX using monoclonal immunoaffinity technique: clinical trials in hemophilia B and comparison to prothrombin complex concentrates. Blood 1992;79:568–575.[Abstract/Free Full Text]
  30. Canal P, Gamelin E, Vassal G et al. Benefits of pharmacological knowledge in the design and monitoring of cancer chemotherapy. Pathol Oncol Res 1998;4:171–178.[Medline]
  31. Lee M, Kern SE, Kisicki JC et al. A pharmacokinetic study to compare two simultaneous 400 µg doses with a single 800 µg dose of oral transmucosal fentanyl citrate. J Pain Symptom Manage 2003;26:743–747.[CrossRef][Medline]
  32. U.S. Food and Drug Administration. Guideline for the Clinical Evaluation of Analgesic Drugs (publication 91D-0425). Rockville, MD: US Department of Health and Human Services, 1992.
  33. Turnidge J. Pharmacodynamics and dosing of aminoglycosides. Infect Dis Clin North Am 2003;17:503–528.[CrossRef][Medline]
  34. Choueiri TK, Hutson TE, Bukowski RM. Evolving role of pegylated interferons in metastatic renal cell carcinoma. Expert Rev Anticancer Ther 2003;3:823–829.[CrossRef][Medline]
  35. Raskin P, Klaff L, McGill J et al. Efficacy and safety of combination therapy: repaglinide plus metformin versus nateglinide plus metformin. Diabetes Care 2003;26:2063–2068.[Abstract/Free Full Text]
  36. Comfort MB, Tse AS, Tsang AC et al. A study of the comparative efficacy of three common analgesics in the control of pain after third molar surgery under local anaesthesia. Aust Dent J 2002;47:327–330.[Medline]
  37. Cheung W, Minton N, Gunawardena K. Pharmacokinetics and pharmacodynamics of epoetin alfa once weekly and three times weekly. Eur J Clin Pharmacol 2001;57:411–418.[CrossRef][Medline]
  38. Cazzola M, Beguin Y, Kloczko J et al. Once-weekly epoetin beta is highly effective in treating anaemic patients with lymphoproliferative malignancy and defective endogenous erythropoietin production. Br J Haematol 2003;122:386–393.[CrossRef][Medline]
  39. Rosberg JH, Ben-Hamadi R, Crémieux PY et al. Dose conversion and cost-effectiveness of erythropoietic therapies in chemotherapy-related anemia: a meta-analysis. Clin Drug Investig 2005;25:33–48.[CrossRef][Medline]
  40. Zagari M, Wacholtz M, Xiu L. An open-label, controlled, randomized, dose comparison study of epoetin alfa for the treatment of anemia in cancer patients receiving platinum-containing chemotherapy. Hematol J 2003;4(suppl 2):61.
  41. Reigner B, Jordan P, Pannier A et al. CERA (continuous erythropoiesis receptor activator), an innovative erythropoietic agent: dose-dependent response in phase I studies. Proc Am Soc Clin Oncol 2003;22:732.
  42. Tavio M, Milan I, Tirelli U. Cancer-related fatigue (review). Int J Oncol 2002;21:1093–1099.[Medline]
  43. Lipman AJ, Lawrence DP. The management of fatigue in cancer patients. Oncology (Huntingt) 2004;18:1527–1535; discussion 1536–1538.
  44. Abdel-Razeq HN. Cancer-related anemia. Saudi Med J 2004;25:15–20.[Medline]
  45. Rosso R, Del Mastro L, Venturini M et al. [Use of erythropoietin in oncology]. Tumori 1997;83(suppl 2):S26–S30.
  46. Lesage P, Portenoy RK. Management of fatigue in the cancer patient. Oncology (Huntingt) 2002;16:373–378, 381; discussion 381–382, 385–386, 388–389.
  47. Chen ML, Chang HK. Physical symptom profiles of depressed and nondepressed patients with cancer. Palliat Med 2004;18:712–718.[Abstract/Free Full Text]
  48. Holzner B, Kemmler G, Greil R et al. The impact of hemoglobin levels on fatigue and quality of life in cancer patients. Ann Oncol 2002;13:965–973.[Abstract/Free Full Text]
  49. Shargel L, Wu-Pong S, Yu ABC. Review of mathematical fundamentals. In: Applied Biopharmaceutics and Pharmacokinetics, Fifth Edition. New York: McGraw-Hill, 2004:21–51.
  50. Tallarida RJ, Murray RB. Procedure 25 area under a curve: trapezoidal and Simpson’s rules. In: Manual of Pharmacologic Calculations With Computer Programs, Second Edition. New York: Springer-Verlag, 1987:77–81.
  51. Witzig TE, Silberstein PT, Loprinzi CL et al. Phase III, randomized, double-blind study of epoetin alfa versus placebo in anemic patients with cancer undergoing chemotherapy. J Clin Oncol 2004 Sep 27; [Epub ahead of print].
  52. Vadhan-Raj S, Mirtsching B, Charu V et al. Assessment of hematologic effects and fatigue in cancer patients with chemotherapy-induced anemia given darbepoetin alfa every two weeks. J Support Oncol 2003;1:131–138.[Medline]
  53. Hesketh PJ, Arena F, Patel D et al. A randomized controlled trial of darbepoetin alfa administered as a fixed or weight-based dose using a front-loading schedule in patients with anemia who have nonmyeloid malignancies. Cancer 2004;100:859–868.[CrossRef][Medline]
  54. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982;143:29–36.[Abstract/Free Full Text]
  55. Swets JA. Measuring the accuracy of diagnostic systems. Science 1988;240:1285–1293.[Abstract/Free Full Text]
  56. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44:837–845.[CrossRef][Medline]
  57. Vandenhende F, Comblain M, Delsemme MH et al. Construction of an optimal destructive sampling design for noncompartmental AUC estimation. J Pharmacokinet Biopharm 1999;27:191–212.[CrossRef][Medline]
Received November 11, 2004; accepted for publication April 1, 2005.




This article has been cited by other articles:


Home page
 JCOHome page
P. Pedrazzoli, A. Farris, S. Del Prete, F. Del Gaizo, D. Ferrari, C. Bianchessi, G. Colucci, A. Desogus, T. Gamucci, A. Pappalardo, et al.
Randomized Trial of Intravenous Iron Supplementation in Patients With Chemotherapy-Related Anemia Without Iron Deficiency Treated With Darbepoetin Alfa
J. Clin. Oncol., April 1, 2008; 26(10): 1619 - 1625.
[Abstract] [Full Text] [PDF]

Home page
 J Oncol Pharm PractHome page
P. Cremieux, F. Vekeman, and P. Lefebvre
Dose conversion and cost effectiveness of erythropoietic therapies in chemotherapy-related anemia: a Canadian application
Journal of Oncology Pharmacy Practice, September 1, 2006; 12(3): 165 - 178.
[Abstract] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow eLetters: Submit a response to this article
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow E-mail this article link to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Reprints/Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Duh, M. S.
Right arrow Articles by Waltzman, R. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Duh, M. S.
Right arrow Articles by Waltzman, R. J.

HOME HELP CONTACT US SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
THE ONCOLOGIST STEM CELLS CME ALPHAMED PRESS JOURNALS
http://theoncologist.alphamedpress.org/subscriptions/etoc.dtl