Breast Cancer Res Treat (2014) 144:143–152 DOI 10.1007/s10549-014-2843-8

CLINICAL TRIAL

Body fat composition impacts the hematologic toxicities and pharmacokinetics of doxorubicin in Asian breast cancer patients A. L. Wong • K. Y. Seng • E. M. Ong • L. Z. Wang • H. Oscar • M. T. Cordero R. Copones • L. Fan • S. H. Tan • B. C. Goh • S. C. Lee



Received: 4 December 2013 / Accepted: 15 January 2014 / Published online: 31 January 2014 Ó Springer Science+Business Media New York 2014

Abstract Body surface area (BSA)-based dosing leads to wide inter-individual variations in drug pharmacokinetics and pharmacodynamics, whereas body composition has been shown to be a more robust determinant of efficacy and toxicity of certain chemotherapeutic agents. We correlated various parameters of body composition with doxorubicin pharmacokinetics and hematologic toxicities in Asian patients with locally advanced or metastatic breast cancer. Our analysis included 84 patients from two studies who received pre- or post-operative single-agent doxorubicin; pharmacokinetic parameters were available for 44 patients. Body composition parameters were derived from CT crosssectional images and population pharmacokinetic analysis was conducted using mixed-effects modeling. Higher intraabdominal fat volume and fat ratio (intra-abdominal:total A. L. Wong  S. H. Tan  B. C. Goh  S. C. Lee (&) Department of Haematology-Oncology, National University Cancer Institute, National University Health System, Level 7, NUHS Tower Block, 1E Kent Ridge Road, Singapore 119228, Singapore e-mail: [email protected] A. L. Wong  L. Z. Wang  B. C. Goh  S. C. Lee Cancer Science Institute, Singapore, Singapore A. L. Wong  M. T. Cordero  S. H. Tan  B. C. Goh  S. C. Lee Haematology Oncology Research Group, National University Cancer Institute, National University Health System, Singapore, Singapore K. Y. Seng  L. Fan  B. C. Goh Department of Pharmacology, Yong Loo Lin School of Medicine, National University Health System, Singapore, Singapore E. M. Ong  H. Oscar  R. Copones Department of Diagnostic Imaging, National University Health System, Singapore, Singapore

abdominal fat volume) correlated with greater incidence of grade 4 leukopenia on cycle 1 day 15 (mean intraabdominal fat volume: 97.4 ± 46.5 cm3 vs 63.4 ± 30.9 cm3, p = 0.014; mean fat ratio: 0.43 ± 0.11 vs 0.33 ± 0.09, p = 0.012, grade 4 vs grade 0-3 leukopenia). On subset analysis, this relationship was maintained even in underweight patients. Concordantly, there were positive correlations between doxorubicin AUC and intra-abdominal fat volume as well as total abdominal fat volume (r2 = 0.324 and 0.262, respectively, all p \ 0.001). BSA and muscle volume did not predict for doxorubicin pharmacokinetics or toxicities. High-intra-abdominal fat volume but not BSA predicted for greater doxorubicin exposure and hematologic toxicities, suggesting that body composition is superior to BSA in determining doxorubicin pharmacokinetics and pharmacodynamics. Body composition has an emerging role in chemotherapy dose determination. Keywords Body composition  Doxorubicin toxicities  Doxorubicin pharmacokinetics  Inter-ethnic variability  Breast cancer

Introduction Although body surface area (BSA) has traditionally been employed as the main variable in determining the prescribed dose of cytotoxic agents, its use is not substantiated by rigorous scientific investigation [1, 2]. In fact, a study of 33 anti-neoplastic agents undergoing evaluation in phase I clinical trials demonstrated that BSA-based dosing failed to reduce inter-patient pharmacokinetic variability in the majority (*85 %) of agents, suggesting severe limitations in its ability to predict drug efficacy and toxicity, and highlighting the need to re-evaluate its routine use [2–4].

123

144

Doxorubicin is a commonly used cytotoxic agent in both early and advanced stage breast cancer, which undergoes extensive hepatic metabolism and subsequent biliary excretion [5]. Its pharmacokinetic profile and dose-limiting myelosuppressive effects are known to be highly variable despite dose normalization to BSA [6–9]. Available evidence suggests that abnormalities in liver function, age differences and genetic polymorphisms of metabolizing enzymes may contribute to inter-individual variation in doxorubicin disposition and toxicities [5, 10]. Interestingly, prior studies have also demonstrated that body composition may determine doxorubicin pharmacokinetics more accurately than BSA [5, 6, 11]. Body composition refers to the proportions of fat and fat-free mass in the body. The latter encompasses all body proteins including metabolic tissues, such as muscle, liver, and kidney, as well as bone and total body water [12]. There is an emerging body of evidence suggesting that body composition is more robust than BSA in predicting the toxicities of certain chemotherapeutic agents, such as 5-fluorouracil and capecitabine, but data are lacking for the vast majority of cytotoxics, including doxorubicin [2, 13]. The optimal method of body composition evaluation remains controversial. Traditional anthropometric measures including body weight and BMI lack sensitivity, therefore, body composition is typically measured using dual-energy X-ray absorptiometry or bioelectrical impendence analysis [14]. More recently, the estimation of body composition from single abdominal cross-sectional computerized tomography (CT) image analysis has been widely adopted due to its precision and the general availability of CT scans among cancer patients [15]. We sought to investigate the association of body composition with doxorubicin pharmacokinetics and hematologic toxicities in locally advanced or metastatic breast cancer patients from two neoadjuvant chemotherapy studies. Body composition parameters (fat and muscle volume) were obtained from single-slice CT analysis and doxorubicin pharmacokinetic analysis was conducted using a mixed-effects modeling approach.

Patients and methods The cohort comprised newly diagnosed Asian locally advanced/metastatic breast cancer patients who were treatment-naı¨ve and who were enrolled into two prospective, open-label phase II pre-operative chemotherapy studies conducted in a single institution from April 2002 till July 2005 and March 2007 till July 2010, respectively. In study 1, participants were randomized to one of the two alternating sequences of doxorubicin (A) at 75 mg/m2 and docetaxel (T) at 75 mg/m2 every 3 weeks for 6 cycles

123

Breast Cancer Res Treat (2014) 144:143–152

(Arm A: A-T-A-T-A-T, n = 50; Arm B: T-A-T-A-T-A, n = 51). Only subjects from Arm A (n = 50) were included in our analysis to avoid possible confounding effects of prior docetaxel administration on hematologic and pharmacokinetic parameters in Arm B subjects. Study 2 randomized patients to treatment with either 4 cycles of 3-weekly pre-operative doxorubicin (75 mg/m2) followed by 4 cycles of 3-weekly post-operative docetaxel (75 mg/m2) or the same chemotherapy in reverse sequence pre- and post-operatively, respectively (n = 50). The study protocols were approved by the institution’s ethics review board and all participants provided written informed consent. Eligibility criteria of the studies were broadly similar and included female patients aged C18 years with histologically confirmed breast cancer, measurable primary tumor where the two longest diameters in perpendicular dimensions were C2 cm, Karnofsky performance score C70, and adequate bone marrow and organ function. Baseline laboratory investigations, including complete blood counts, liver and renal function tests, were performed within 2 weeks of study entry. Hematologic parameters were monitored on days 8 and 15 of cycle 1. All toxicities were graded according to the National Cancer Institute Common Toxicity Criteria for Adverse Events version 2.0. Body composition measurements Anthropometric measurements Baseline clinical assessments were performed within 2 weeks of study enrollment. Weight and height were measured according to standard methods. Body mass index (BMI) [weight (kg)/height (m2)] and BSA [H(weight (kg) 9 height (cm)/3,600)] were calculated for each participant. Radiologic analysis All CT scan images were obtained within 4 weeks prior to study entry as part of routine-screening investigations and were acquired on a Siemens Volume Zoom (4 slice) CT scanner. The third lumbar vertebra (L3) was chosen as a standard landmark and two consecutive images extending from L3 to the iliac crest were selected to measure total muscle, total abdominal fat, and intra-abdominal fat crosssectional area [16]. The first image was chosen when both transverse processes were clearly visible, followed by one 7 mm slice below. On both slices, the muscles of the abdominal wall, paraspinal region, and back were manually drawn as the region of interest (ROI), and the total area of the ROI for each slice was then computed. The average area of the two slices was then multiplied by the slice thickness to obtain the muscle volume (in cm3) between the

Breast Cancer Res Treat (2014) 144:143–152

slices. The whole abdomen was selected as the ROI for total abdominal fat measurement. The area within the abdominal wall was marked as the ROI for intra-abdominal fat measurement. The fat content was identified using preestablished thresholds of Hounsfield units (HU) (-10 to -300). The average area of the two slices was then multiplied by the slice thickness to obtain the volume of fat (in cm3) between the slices. Fat ratio was defined as the ratio of intra-abdominal fat volume to total abdominal fat volume. Images were analyzed using Siemens Syngo (VA40C) software. Liver attenuation in HU was derived using the average attenuation values of two ROIs obtained from a homogenous area of parenchyma located in the right lobe of the liver [17]. Pharmacokinetics Pharmacokinetic data were available for patients in Study 1 but not Study 2. Five milliliters whole plasma was collected for analysis of doxorubicin and doxorubicinol levels at 0, 1, 2, 4, 7, and 24 h after the first doxorubicin treatment. Plasma concentrations of doxorubicin and its major metabolite doxorubicinol were determined by a modified method of reverse-phase high-performance liquid chromatography with fluorescence detection, according to methods previously described [10, 18]. The pharmacokinetics of doxorubicin-doxorubicinol were assessed with a compartmental approach by means of population pharmacokinetic analysis using the nonlinear mixed-effects modeling program NONMEM (Version 7.2, NONMEM Project Group, San Francisco, CA, USA) interfaced with PDx-POP (Version 5.0, Icon Development Solutions, Ellicott City, MD, USA). Additional calculation, data manipulation, and graphics were performed using R (Version 2.15.0). A basic model structure was established, using the full doxorubicin-doxorubicinol pharmacokinetic profile data. The individual parameters were assumed to be log-normally distributed, and proportional error was employed for description of residual variability. Nineteen covariates were considered in the covariate analysis. Both continuous (age, BMI, BSA, height, weight, intra-abdominal fat volume, total abdominal fat volume, total muscle volume, intra-abdominal:total abdominal fat ratio, total bilirubin, serum albumin, serum creatinine, calculated creatinine clearance, serum alanine transaminase, and serum aspartate transaminase), and categorical variables (race, baseline clinical tumor stage, baseline clinical nodal stage, and presence or absence of metastatic disease at baseline) were included. The covariate relationships were modeled proportional to the parameter, as the fractional change in the parameter with the covariate. Continuous covariate variables were centered to their median values in a nonlinear

145

(Eq. 1) manner so that the population estimates would represent those of an average subject. For example,   BILj hBIL P ¼ hpop  ð1Þ BILm where hpop is the parameter value in individuals with the median total bilirubin level of the population, BILj is the bilirubin level of the jth individual, BILm is the median bilirubin level of the population, and hBIL is an exponent describing the correlation function. For categorical covariates with more than two categories (e.g., race), a change in the parameter was evaluated using Eq. 2 (illustrated for the race covariate): 8 if Chinese < h1 P ¼ h1  h2 if Malay ; ð2Þ : h1  h3 if Indian where h1 is the population average parameter for Chinese subjects, and h2 and h3 are the fractional changes in h1 associated with the Malay and Indian ethnicity groups, respectively. The categorical covariate presence or absence of metastatic disease at baseline was modeled as shown in Eq. 3: P ¼ hpop  ðhmeta Þmeta ;

ð3Þ

where hpop is the population typical value for the parameter, and hmeta is the fractional difference between patients with metastatic disease at baseline (meta = 1) and without (meta = 0). Where applicable, multiplicative equations were used to describe the combined effect of multiple covariates on the same parameter. The covariate search was performed using a stepwise covariate model-building procedure [19]. The procedure included a forward inclusion step, with parameter–covariate relationships being added to the model in a stepwise manner until no further relationship was statistically significant (p \ 0.05). Backward elimination steps were then carried out, with the identified relationships being excluded from the model if they failed to achieve stricter statistical significance (p \ 0.01), in order to adjust for multiple testing. The model-building procedure was guided by the likelihood ratio test, diagnostic plots, and visual predictive checks (VPC). The area under the curve (AUC) was calculated from the relationship as follows: AUC = drug dose/(empirical Bayes’ estimate of clearance). Statistical analysis Quantitative variables were described using mean and standard deviation, whereas qualitative variables were described using frequencies and percentages. The various clinical parameters were compared with pharmacokinetic

123

146 Table 1 Baseline patient characteristics and hematologic parameters (n = 84)

Breast Cancer Res Treat (2014) 144:143–152

Characteristics

Number (Percentage)

Race Chinese

43 (51.2)

Malay

35 (41.7)

Indian

6 (7.1)

Clinical tumor stage T2

5 (6.0)

T3

34 (40.5)

T4

45 (53.6)

Clinical nodal stage N0

35 (41.7)

N1 N2

25 (29.8) 11 (13.1)

N3

13 (15.5)

Metastatic disease

28 (33.3) n (%)

Age (years)

Mean ± standard deviation 50.4 ± 10.1

Body mass index (BMI) (kg/m2) Underweight (BMI \18.5) Normal range (BMI 18.5–25)

24.5 ± 4.7 8 (9.5) 39 (46.4)

Overweight (BMI 25–30)

25 (29.8)

Obese (BMI [ 30)

12 (14.3)

Body surface area (BSA) (m2)

1.57 ± 0.15

Intra-abdominal fat volume (m3) 3

Total abdominal fat volume (m )

65.9 ± 33.0 195.7 ± 76.7

Fat ratio (intra-abdominal fat volume: total abdominal fat volume)

0.33 ± 0.10

Total muscle volumea (m3)

69.9 ± 21.0

Liver attenuation (Hounsfield units)

111.7 ± 25.5

Total bilirubin (lmol/l)

8.6 ± 5.1

Serum albumin (g/l)

40.9 ± 5.6

Serum alanine transaminase (U/l)

26.0 ± 14.8

Serum aspartate transaminase (U/l)

27.5 ± 15.4

Calculated creatinine clearance (ml/min)

87.4 ± 29.6

Cycle 1 doxorubicin hematologic parameters

123

D8 leukocyte count (9109/l)

5.14 ± 2.00

D15 leukocyte count (9109/l)

3.39 ± 7.74

D8 neutrophil count (9109/l)

3.58 ± 1.88

D15 neutrophil count (9109/l)

0.90 ± 1.67

D8 hemoglobin (g/dl)

11.7 ± 1.8

D15 hemoglobin (g/dl)

10.9 ± 1.7

D8 platelet count (9109/l)

242 ± 85

D15 platelet count (9109/l)

293 ± 161

Breast Cancer Res Treat (2014) 144:143–152 Table 1 continued

Incidence of doxorubicin-induced grade 4 hematologic toxicities (cycle 1 day 15) Grade 4 leukopenia Grade 4 neutropenia

a

Data available for 15 patients

147

6 (7.1) 47 (56)

Grade 4 anemia

1 (1.2)

Grade 4 thrombocytopenia

0 (0)

and toxicity parameters using student’s t test, Pearson’s correlation, or their respective nonparametric tests as appropriate. All p values were two-sided, and level of significance was p B 0.05. Statistical analysis was performed using SPSS v13.0 (SPSS Inc., Chicago, IL).

Results Patient characteristics A total of 100 patients were enrolled into Study 1 (Arm A) and Study 2, but only 84 patients were included in the final analysis because 9 patients did not have available CT scans, and 7 patients in Study 2 did not receive post-operative doxorubicin. Their baseline clinical parameters, radiologic measurements of body composition, and laboratory values are summarized in Table 1. The cohort was predominantly Chinese or Malay and had a mean age of 50.4 ± 10.1 years. Clinical T3-4 primary breast lesions were present in 94.1 % and metastatic disease in 33.3 %. The mean BSA was 1.57 ± 0.15 m2, and mean BMI was 24.5 ± 4.7 kg/m2, with 10 % who were underweight (BMI \18.5 kg/m2), 46 % who were in the normal range (BMI 18.5–25 kg/m2), 30 % who were overweight (BMI 25–30 kg/m2), and 14 % who were obese (BMI [30 kg/m2) according to the WHO classification. Mean intra-abdominal fat volume was 65.9 ± 33.0 cm3, mean total abdominal fat volume was 195.7 ± 76.7 cm3, and mean intra-abdominal:total abdominal fat ratio was 0.33 ± 0.10. Wide variations in volumes of intra-abdominal fat, with a 30-fold range (6.0–189.1 cm3), and total abdominal fat, with a 14-fold range (28.1–392.7 cm3), were observed within the cohort. There was a three-fold variation in both mean total muscle volume 69.9 ± 21.0 cm3 (36.1–103.1) and mean liver attenuation 111.7 ± 25.5HU (47.8–166.6). Hematologic toxicities Hematologic parameters following the first treatment cycle of doxorubicin were available for all 84 patients (Table 1). Nadir myelosuppression was observed on day 15 following doxorubicin administration. Leukopenia and neutropenia were the predominant dose-limiting toxicities, with grade 4 leukopenia and grade 4 neutropenia affecting 7.1 and 56 % of patients, respectively. Grade 4 anemia (1.2 %) was rare,

and no patient developed grade 4 thrombocytopenia. In order to identify potential determinants of doxorubicin hematologic toxicity, baseline patient characteristics, body composition parameters, and laboratory values were correlated with the incidences of grade 4 leukopenia and neutropenia. The effects of body composition on cycle 1 day 15 doxorubicin hematologic toxicities are summarized in Table 2. Two radiologic parameters of body composition were found to correlate significantly with doxorubicininduced leukopenia; mean intra-abdominal fat volume as well as fat ratio were higher in those who developed grade 4 leukopenia compared to those who did not (mean intraabdominal fat volume: 97.4 ± 46.5 vs 63.4 ± 30.9 cm3, p = 0.014; mean fat ratio: 0.43 ± 0.11 vs 0.33 ± 0.09, p = 0.012, grade 4 vs grade 0–3 leukopenia) (Table 2). Correlative analyses did not reveal any significant determinants of doxorubicin-induced grade 4 neutropenia (Table 2). While BSA and BMI correlated positively with intraabdominal fat volume (r2 = 0.200, p \ 0.001; r2 = 0.425, p \ 0.001, respectively), neither these parameters nor any of the remaining clinical, radiologic, and laboratory variables, including total muscle volume, were independently associated with doxorubicin-induced grade 4 leukopenia and neutropenia. We explored the association between body fat composition and doxorubicin-induced grade 4 leukopenia across the various BMI categories and found that significant correlations were maintained in both underweight and overweight patients (Table 2). Of note, in underweight patients, higher intra-abdominal and total abdominal fat volumes were observed in the individual who developed grade 4 leukopenia compared to those who did not (mean intra-abdominal fat volume: 84.6 cm3 vs 20.4 ± 20.6 cm3, p = 0.027; mean total abdominal fat volume: 262.3 vs 72.5 ± 60.7 cm3, p = 0.026, grade 4 vs grade 0–3 leukopenia) (Table 2). Similarly, in overweight patients, higher intra-abdominal fat volumes and fat ratios were observed in patients who developed grade 4 leukopenia compared to those who did not (mean intra-abdominal fat volume: 113.9 ± 67.4 vs 68.2 ± 27.4 cm3, p = 0.034; mean fat ratio: 0.48 ± 0.09 vs 0.29 ± 0.008 cm3, p = 0.002, grade 4 vs grade 0–3 leukopenia) (Table 2). Subset analysis was not performed in obese patients since none developed grade 4 leukopenia, possibly due to the use of non-protocol empiric dose reductions. Overall, 22.5 % of the cohort had empiric dose reductions

123

148

Breast Cancer Res Treat (2014) 144:143–152

Table 2 Correlation of mean parameters of body composition with grade of doxorubicin-induced neutropenia and leukopenia on cycle 1 day 15, with subset analysis according to WHO BMI categories Grade 4 Leukopenia (n = 6)

Grade 0-3 Leukopenia (n = 78)

p valueb

Grade 4 Neutropenia (n = 47)

Grade 3 Neutropenia (n = 37)

p valueb

Intra-abdominal fat volume (cm3)

97.4 – 46.5

63.4 – 30.9

0.014

70.6 – 33.8

59.8 – 31.5

0.139

Underweight (BMI \ 18.5)

84.6 (n = 1)

20.4 ± 20.6 (n = 7)

0.027

39.1 ± 33.8 (n = 4)

12.6 ± 6.0 (n = 4)

0.171

Normal range (BMI 18.5–25)

78.9 ± 8.9 (n = 2)

57.5 ± 22.2 (n = 37)

0.187

53.7 ± 23.4 (n = 21)

54.1 ± 22.7 (n = 18)

0.952

Overweight (BMI 25-30)

113.9 ± 67.4 (n = 3)

68.2 ± 27.4 (n = 22)

0.034

78.9 ± 37.1 (n = 16)

58.6 ± 31.6 (n = 9)

0.215

102.1 ± 25.3 (n = 6)

103.1 ± 21.7 (n = 6)

0.942

Obese (BMI [ 30)

98.3 ± 26.7 (n = 12) 229.1 – 82.3

193.2 – 76.3

0.271

203.1 – 72.7

186.4 – 81.6

0.324

262.3

72.5 ± 60.7

0.026

124.5 ± 99.8

50.5 ± 31.9

0.201

Normal range (BMI 18.5–25)

206.1 ± 52.4

167.5 ± 47.7

0.274

155.9 ± 43.4

170.5 ± 50.2

0.335

Overweight (BMI 25–30)

233.5 ± 120.2

229.4 ± 54.7

0.918

248.4 ± 55.6

209.6 ± 69.8

0.158

288.8 ± 55.3

294.0 ± 61.9

0.877

0.43 – 0.11

0.33 – 0.09

0.012

0.35 – 0.11

0.31 – 0.06

0.120

Underweight (BMI \ 18.5)

0.32

0.27 ± 0.06

0.438

0.28 ± 0.05

0.25 ± 0.07

0.824

Normal range (BMI 18.5–25)

0.40 ± 0.15

0.34 ± 0.10

0.447

0.34 ± 0.10

0.31 ± 0.07

0.457

Overweight (BMI 25-30)

0.48 ± 0.09

0.29 ± 0.08

0.002

0.29 ± 0.12

0.28 ± 0.07

0.786

Obese (BMI [ 30) Liver attenuation (HU)

123.7 – 23.1

0.36 ± 0.06 110.8 – 25.6

0.236

0.38 ± 0.09 111.5 – 24.3

0.35 ± 0.01 112.0 – 81.2

0.666 0.938

Muscle volumea

0

69.9 – 21.0



76.1 – 19.1

52.6 – 17.5

0.051

Body surface area (m2)

1.56 – 0.15

1.57 – 0.15

0.898

1.58 – 0.15

1.57 – 0.15

0.757

Body mass index (kg/m2)

24.4 – 3.8

24.5 – 4.7

0.966

24.6 – 4.8

24.4 – 4.6

0.843

Total abdominal fat volume (cm3) Underweight (BMI \ 18.5)

Obese (BMI [ 30) Fat ratio

a

276.2 ± 62.5

b

Data available for 15 patients; Student’s t test comparing mean ± SD of various parameters of body composition (intra-abdominal fat volume, total abdominal fat volume, fat ratio, liver attenuation, muscle volume, body surface area, and body mass index) between patients who developed grade 4 leukopenia/neutropenia versus those who developed grade 0–3 leukopenia/neutropenia Statistically significant p values (p \ 0.05) are indicated in bold italics

exceeding 5 % of their calculated BSA-based doxorubicin dose in cycle 1. This proportion was significantly higher in obese patients compared to non-obese patients (45.5 vs 20.0 %, p = 0.011); the mean dose reduction among obese patients compared to non-obese patients was 16.4 ± 10.5 vs 2.3 ± 12.6 %, p B 0.01. Pharmacokinetics Population pharmacokinetic model for doxorubicindoxorubicinol Population pharmacokinetic analysis was performed for 44 patients with available pharmacokinetic data. A two-

123

compartment model for doxorubicin and one sequential compartment representing doxorubicinol best described the data. All doxorubicin was assumed to be converted to doxorubicinol. Population estimates of the following pharmacokinetic parameters were obtained: the volume of distribution of doxorubicin in the central compartment (Vc); volume of distribution of doxorubicin in the peripheral compartment (Vp); intercompartmental clearance of the parent drug (Q); clearance of doxorubicin to doxorubicinol (CL); clearance of doxorubicinol (CLm); and volume of distribution of doxorubicinol (Vm). The covariate search revealed a statistically significant (p \ 0.01) effect of age on CL. The typical value for CL is 53.5 9 (Age in years/50)-0.393. The parameter estimates from the final

Breast Cancer Res Treat (2014) 144:143–152

149

Table 3 Population pharmacokinetic parameter estimates of doxorubicin and doxorubicinol Parameter

Final model estimate (SE)

CV%

Doxorubicin CL (l/h)

53.5 (7.95)

Age * CL

-0.393 (0.13)

Vc (l)

25.6 (9.87)

Vp (l)

446 (125)

Q (l)

55.9 (24.2)

2

x for CL x2 for Vc

0.0237 (0.00814) 0.0248 (0.0212)

15.4 15.8

x2 for Vp

0.0205 (0.0098)

14.3

Correlation CL, Vc

0.229 (0.0831)

r2 (proportional)

0.186 (0.022)

43.2

Doxorubicinol CLm (l/h)

92.7 (5.39)

Vm (l)

1,700 (244)

x2 for CLm

0.0564 (0.0166)

23.8

x2 for Vm

0.126 (0.0245)

35.4

r2 (proportional)

0.151 (0.0136)

38.9

SE standard error of parameter estimate, x2 inter-individual variability variance for model parameter, r2 residual error variance for parent drug or metabolite, CV% percentage coefficient of variation

model are shown in Table 3. The VPC for the final model, stratified for doxorubicin and doxorubicinol, is shown in Fig. 1. Doxorubicin AUC was investigated as a secondary doxorubicin pharmacokinetic parameter. Univariate linear regression analyses showed that there were statistically significant positive correlations between doxorubicin AUC and intra-abdominal fat volume (r2 = 0.324, p \ 0.001) as well as between doxorubicin AUC and total abdominal fat volume (r2 = 0.262, p \ 0.001) (Fig. 2). Neither BSA nor BMI correlated with doxorubicin AUC.

Discussion We describe the first study evaluating the impact of various parameters of body composition on doxorubicin pharmacokinetics as well as hematologic toxicities in a population of newly diagnosed, treatment-naive Asian breast cancer patients receiving single-agent doxorubicin as initial therapy. Most notably, we found that body fat composition, particularly intra-abdominal fat, was highly variable in our population, and that higher intra-abdominal fat content correlated significantly with increased doxorubicin exposure and rates of grade 4 leukopenia. Subset analysis according to BMI category demonstrated a significant association even in underweight patients, implying that individuals with excess body fat relative to LBM are at

Fig. 1 Visual predictive check for the final a doxorubicin and b doxorubicinol models. Circles represent the observed concentration–time points

greater risk of toxicity regardless of BMI. Importantly, BSA was not a determinant of the pharmacokinetics or pharmacodynamics of either agent. This, together with the highly variable drug pharmacokinetics and toxicities observed, is consistent with previous data suggesting that BSA-based dosing frequently fails to reduce interpatient variability in drug effect and underscores the importance of further evaluating the role of body composition in chemotherapy dose determination [1, 2, 20]. Data from a clinical study of breast cancer patients receiving adjuvant doxorubicin-based chemotherapy suggested that full-weight BSA-based dosing in obese patients did not lead to increased toxicities and was important in achieving optimal outcomes [21]. However, several publications which have explored the impact of body composition on doxorubicin disposition, and the use of alternative parameters in dose calculation have provided evidence to the contrary. For example, pharmacokinetic studies have demonstrated that obese patients receiving BSA-based dosing of doxorubicin are subject to higher systemic drug

123

150

Fig. 2 Scatterplots and linear correlations between doxorubicin AUC and a intra-abdominal fat volume (cm3) and b total abdominal fat volume (cm3)

exposure compared to lean controls due to a reduction in drug clearance, and that the use of alternative measures of body composition in dose calculation may eliminate this risk of over-exposure [5, 11]. Another study conducted in a pediatric population showed that doxorubicinol volume of distribution, and clearance were significantly reduced in individuals with a high proportion of body fat exceeding 30 %, regardless of BMI [6]. The authors postulated that the excess fat in such individuals is not available for the distribution of hydrophilic compounds, hence the volume of distribution normalized to BSA might decrease, predisposing to greater toxicities. Doxorubicin has been shown to exhibit a hydrophilic pharmacokinetic profile in

123

Breast Cancer Res Treat (2014) 144:143–152

patients and physiologic models, first distributing rapidly into highly perfused organs (lung, kidney, spleen, and liver) and subsequently into muscles, but with minimal distribution into adipose tissue [5, 22]. Hence, our results not only support previous data suggesting that increased body fat content alters the pharmacokinetic profile of doxorubicin, but also demonstrate for the first time that an increased proportion of body fat results in greater doxorubicin-induced myelosuppression. Racial disparities in doxorubicin-related hematologic toxicities have been reported by us and other investigators in breast cancer patients receiving adjuvant doxorubicin/ cyclophosphamide (AC). Although the incidence of grade 4 neutropenia in our present cohort was higher than that observed in Western populations [23], it is concordant with several previous reports, where Asian patients were found to be significantly more susceptible than Caucasians to hematologic toxicities following AC administration. One prospective study demonstrated grade 4 neutropenia rates of 53.9 versus 18.9 % in East Asians compared to Caucasians, whereas the corresponding figures were 25 versus 0.3 % in a retrospective review of Chinese patients compared to a historic Western cohort [8, 9]. We postulated that this could be attributed in part to morphometric differences and indeed, several studies have shown that although Asians may have lower BMI, they have paradoxically higher proportions of body fat compared to Caucasians [24, 25]. More specifically, Chinese and South Asians were found to have higher proportions of intraabdominal adipose tissue than Caucasians in a study evaluating body composition in a large multicultural cohort [24, 26]. These interethnic differences in body composition may in turn contribute to higher doxorubicin-related toxicities observed in Asian patients receiving BSA-based chemotherapy doses, due to the proportionately lower fatfree mass available for drug distribution in Asians compared to Caucasians. An alternative measure of body composition, lean body mass (LBM), has been shown to influence the pharmacokinetics and toxicities of several chemotherapeutic agents, including 5-fluorouracil, capecitabine, and epirubicin [13, 27, 28]. It was postulated that patients with lower LBM may be relatively overdosed due to a disproportionately small volume of distribution available to hydrophilic cytotoxic agents [13]. In these studies, the parameters of body composition examined were LBM and total fat mass. In comparison, ours is the first study to specifically examine the correlation between the various patterns of regional fat deposition with cytotoxic drug pharmacokinetics and toxicities. This analysis was deemed worthwhile for several reasons; (i) the previously reported variability in body fat composition between Asians and Caucasians [24, 25], (ii) data suggesting that intra-abdominal fat determines

Breast Cancer Res Treat (2014) 144:143–152

the pharmacokinetics of some compounds [29, 30], and (iii) studies demonstrating that excess intra-abdominal fat is highly discriminatory in predicting for certain disease states, such as the metabolic syndrome [31]. Of the various parameters explored in our study, intra-abdominal fat demonstrated the most consistent correlation with doxorubicin pharmacokinetics and pharmacodynamics. Since this differs from previous studies demonstrating the correlation between LBM and effect from other chemotherapeutic agents, we conclude that the optimal parameter of body composition for dosing of different chemotherapy drugs has yet to be determined and is an issue which warrants further evaluation in subsequent studies. Our study is one of few that examined the clinical effects of single-agent doxorubicin, since the drug is more commonly administered in combination therapy, making it difficult to attribute observed pharmacokinetic and pharmacodynamic effects to a single drug. However, there are several limitations. First, we were able to establish a positive correlation between body fat composition and doxorubicin AUC, but not with other pharmacokinetic parameters, including clearance and volume of distribution. However, it can be argued that drug exposure is a composite of these parameters, and it is possible that the discovery of any statistically significant relationships was hindered by the relatively small sample size available for pharmacokinetic correlative analysis. Second, although the robustness of the population pharmacokinetic model developed in this study was validated internally using VPC, the developed model requires further prospective external evaluation. Importantly, due to the limited sample size and exploratory nature of the study, our findings should be regarded as hypothesis-generating and are by no means conclusive. In particular, we are unable to draw conclusions regarding dose selection in the obese because the use of non-protocol dose reductions led to the exclusion of obese patients from our subset analysis. Overall, these data contribute to the mounting body of evidence that tailoring chemotherapy doses to body composition has the potential to greatly improve tolerability and possibly efficacy [12]. However, the implementation of this in clinical practice can only be realized through the determination of the optimal measure of body composition, systematic investigation of how each body composition parameter impacts different chemotherapeutic agents, followed by rigorous prospective evaluation in doseoptimization studies.

Conclusions We present novel data demonstrating that body fat composition measured by single-slice CT analysis was

151

predictive of doxorubicin-related hematologic toxicities and pharmacokinetics in Asian breast cancer patients, whereas BSA-based dosing and muscle volume were not. These data support the emerging role of body composition in chemotherapy dosing, and may in part explain the greater susceptibility of Asian patients, who have been reported to have proportionately higher levels of body fat than Caucasians, to doxorubicin-related hematologic toxicities. This has wide-ranging clinical implications and further studies are indicated to incorporate body composition evaluation into chemotherapy dose determination. Acknowledgments The authors are grateful to all patients who participated in the studies. All authors express no conflicts of interest in contributing to this work. This work was supported by the National Medical Research Council of Singapore [NMRC/CSA/015/2009 (S.C. Lee)] and the NUHS Clinician Scientist Program Clinician Scientist Unit, Yong Loo Lin School of Medicine, National University of Singapore (A.L. Wong). The study was approved by the National Healthcare Group Domain-Specific Ethics Review Board.

References 1. Ratain MJ (1998) Body-surface area as a basis for dosing of anticancer agents: science, myth, or habit? J Clin Oncol 16:2297–2298 2. Gurney H (1996) Dose calculation of anticancer drugs: a review of the current practice and introduction of an alternative. J Clin Oncol 14:2590–2611 3. Miller AA (2002) Body surface area in dosing anticancer agents: scratch the surface! J Natl Cancer Inst 94:1822–1823 4. Baker SD, Verweij J, Rowinsky EK et al (2002) Role of body surface area in dosing of investigational anticancer agents in adults, 1991–2001. J Natl Cancer Inst 94:1883–1888 5. Rodvold KA, Rushing DA, Tewksbury DA (1988) Doxorubicin clearance in the obese. J Clin Oncol 6:1321–1327 6. Thompson PA, Rosner GL, Matthay KK et al (2009) Impact of body composition on pharmacokinetics of doxorubicin in children: a Glaser pediatric research network study. Cancer Chemother Pharmacol 64:243–251 7. Frost BM, Eksborg S, Bjork O et al (2002) Pharmacokinetics of doxorubicin in children with acute lymphoblastic leukemia: multiinstitutional collaborative study. Med Pediatr Oncol 38:329–337 8. Beith J, Goh BC, Yeo W et al (2002) Inter-ethnic differences in the myelotoxicity of adriamycin/cyclophosphamide (AC) for adjuvant breast cancer. Proc Am Soc Clin Oncol 21:252 9. Ma B, Yeo W, Hui P et al (2002) Acute toxicity of adjuvant doxorubicin and cyclophosphamide for early breast cancer: a retrospective review of Chinese patients and comparison with an historic Western series. Radiother Oncol 62:185–189 10. Fan L, Goh BC, Wong CI et al (2008) Genotype of human carbonyl reductase CBR3 correlates with doxorubicin disposition and toxicity. Pharmacogenet Genomics 18:621–631 11. Sparreboom A, Wolff AC, Mathijssen RH et al (2007) Evaluation of alternate size descriptors for dose calculation of anticancer drugs in the obese. J Clin Oncol 25:4707–4713 12. Thibault R, Pichard C (2012) The evaluation of body composition: a useful tool for clinical practice. Ann Nutr Metab 60: 6–16 13. Prado CM, Baracos VE, McCargar LJ et al (2007) Body composition as an independent determinant of 5-fluorouracil-based chemotherapy toxicity. Clin Cancer Res 13:3264–3268

123

152 14. Kyle UG, Morabia A, Slosman DO et al (2001) Contribution of body composition to nutritional assessment at hospital admission in 995 patients: a controlled population study. Br J Nutr 86:725–731 15. Mitsiopoulos N, Baumgartner RN, Heymsfield SB et al (1998) Cadaver validation of skeletal muscle measurement by magnetic resonance imaging and computerized tomography. J Appl Physiol 85:115–122 16. Shen W, Punyanitya M, Wang Z et al (2004) Total body skeletal muscle and adipose tissue volumes: estimation from a single abdominal cross-sectional image. J Appl Physiol 97:2333–2338 17. Kuk JL, Katzmarzyk PT, Nichaman MZ et al (2006) Visceral fat is an independent predictor of all-cause mortality in men. Obesity (Silver Spring) 14:336–341 18. Andersen A, Warren DJ, Slordal L (1993) A sensitive and simple high-performance liquid chromatographic method for the determination of doxorubicin and its metabolites in plasma. Ther Drug Monit 15:455–461 19. Wahlby U, Jonsson EN, Karlsson MO (2001) Assessment of actual significance levels for covariate effects in NONMEM. J Pharmacokinet Pharmacodyn 28:231–252 20. Prado CM, Lieffers JR, McCargar LJ et al (2008) Prevalence and clinical implications of sarcopenic obesity in patients with solid tumours of the respiratory and gastrointestinal tracts: a population-based study. Lancet Oncol 9:629–635 21. Rosner GL, Hargis JB, Hollis DR et al (1996) Relationship between toxicity and obesity in women receiving adjuvant chemotherapy for breast cancer: results from cancer and leukemia group B study 8541. J Clin Oncol 14:3000–3008 22. Chan KK, Cohen JL, Gross JF et al (1978) Prediction of adriamycin disposition in cancer patients using a physiologic, pharmacokinetic model. Cancer Treat Rep 62:1161–1171 23. Fisher B, Brown AM, Dimitrov NV et al (1990) Two months of doxorubicin-cyclophosphamide with and without interval reinduction therapy compared with 6 months of cyclophosphamide,

123

Breast Cancer Res Treat (2014) 144:143–152

24.

25.

26.

27.

28.

29.

30.

31.

methotrexate, and fluorouracil in positive-node breast cancer patients with tamoxifen-nonresponsive tumors: results from the National Surgical Adjuvant Breast and Bowel Project B-15. J Clin Oncol 8:1483–1496 Deurenberg-Yap M, Schmidt G, van Staveren WA, Deurenberg P (2000) The paradox of low body mass index and high body fat percentage among Chinese, Malays and Indians in Singapore. Int J Obes Relat Metab Disord 24:1011–1017 Wang J, Thornton JC, Russell M et al (1994) Asians have lower body mass index (BMI) but higher percent body fat than do whites: comparisons of anthropometric measurements. Am J Clin Nutr 60:23–28 Lear SA, Humphries KH, Kohli S et al (2007) Visceral adipose tissue accumulation differs according to ethnic background: results of the Multicultural Community Health Assessment Trial (M-CHAT). Am J Clin Nutr 86:353–359 Prado CM, Baracos VE, McCargar LJ et al (2009) Sarcopenia as a determinant of chemotherapy toxicity and time to tumor progression in metastatic breast cancer patients receiving capecitabine treatment. Clin Cancer Res 15:2920–2926 Prado CM, Lima IS, Baracos VE et al (2011) An exploratory study of body composition as a determinant of epirubicin pharmacokinetics and toxicity. Cancer Chemother Pharmacol 67:93–101 Sugioka N, Haraya K, Fukushima K et al (2009) Effects of obesity induced by high-fat diet on the pharmacokinetics of nelfinavir, a HIV protease inhibitor, in laboratory rats. Biopharm Drug Dispos 30:532–541 Lottenberg SA, Giannella-Neto D, Derendorf H et al (1998) Effect of fat distribution on the pharmacokinetics of cortisol in obesity. Int J Clin Pharmacol Ther 36:501–505 Nakao YM, Miyawaki T, Yasuno S et al (2012) Intra-abdominal fat area is a predictor for new onset of individual components of metabolic syndrome: MEtabolic syndRome and abdominaL ObesiTy (MERLOT study). Proc Jpn Acad Ser B Phys Biol Sci 88:454–461

Body fat composition impacts the hematologic toxicities and pharmacokinetics of doxorubicin in Asian breast cancer patients.

Body surface area (BSA)-based dosing leads to wide inter-individual variations in drug pharmacokinetics and pharmacodynamics, whereas body composition...
594KB Sizes 3 Downloads 0 Views