Journal of Oncology Pharmacy Practice

Journal of Oncology Pharmacy Practice http://opp.sagepub.com/

Effect of the tyrosine kinase inhibitors (sunitinib, sorafenib, dasatinib, and imatinib) on blood glucose levels in diabetic and nondiabetic patients in general clinical practice Nicole Agostino, Vernon M Chinchilli, Christopher J Lynch, Anita Koszyk-Szewczyk, Rebecca Gingrich, Jeffrey Sivik and Joseph J. Drabick J Oncol Pharm Pract published online 4 August 2010 DOI: 10.1177/1078155210378913 The online version of this article can be found at: http://opp.sagepub.com/content/early/2010/08/03/1078155210378913

Published by: http://www.sagepublications.com

Additional services and information for Journal of Oncology Pharmacy Practice can be found at: Email Alerts: http://opp.sagepub.com/cgi/alerts Subscriptions: http://opp.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav

Downloaded from opp.sagepub.com at PENNSYLVANIA STATE UNIV on August 5, 2010

J Oncol Pharm Pract OnlineFirst, published on August 4, 2010 as doi:10.1177/1078155210378913

Journal of

Oncology Pharmacy Practice

Original Article

Effect of the tyrosine kinase inhibitors (sunitinib, sorafenib, dasatinib, and imatinib) on blood glucose levels in diabetic and nondiabetic patients in general clinical practice

J Oncol Pharm Practice 0(0) 1–6 ! The Author(s) 2010 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1078155210378913 opp.sagepub.com

Nicole M Agostino Division of Hematology–Oncology, Department of Medicine, Penn State Milton S Hershey Medical Center, Hershey, PA 17033-0850, USA

Vernon M Chinchilli Department of Public Health Sciences, Penn State Milton S Hershey Medical Center, Hershey, PA 17033-0850, USA

Christopher J Lynch Department of Cellular and Molecular Physiology, Penn State Milton S Hershey Medical Center, Hershey, PA 17033-0850, USA

Anita Koszyk-Szewczyk Division of Hematology–Oncology, Department of Medicine, Penn State Milton S Hershey Medical Center, Hershey, PA 17033-0850, USA

Rebecca Gingrich Division of Hematology–Oncology, Department of Medicine, Penn State Milton S Hershey Medical Center, Hershey, PA 17033-0850, USA

Jeffrey Sivik Department of Pharmacy, Penn State Milton S Hershey Medical Center, Hershey, PA 17033-0850, USA

Joseph J Drabick Division of Hematology–Oncology, Department of Medicine, Penn State Milton S Hershey Medical Center, Hershey, PA 17033-0850, USA

Abstract Tyrosine kinase is a key enzyme activity utilized in many intracellular messaging pathways. Understanding the role of particular tyrosine kinases in malignancies has allowed for the design of tyrosine kinase inhibitors (TKIs), which can target these enzymes and interfere with downstream signaling. TKIs have proven to be successful in the treatment of chronic myeloid leukemia, renal cell carcinoma and gastrointestinal stromal tumor, and other malignancies. Scattered reports have suggested that these agents appear to affect blood glucose (BG). We retrospectively studied the BG concentrations in diabetic (17) and nondiabetic (61) patients treated with dasatinib (8), imatinib (39), sorafenib (23), and sunitinib (30) in our clinical practice. Mean declines of BG were dasatinib (53 mg/dL), imatinib (9 mg/dL), sorafenib (12 mg/dL), and sunitinib (14 mg/dL). All these declines in BG were statistically significant. Of note, 47% (8/17) of the patients with diabetes were able to discontinue their medications, including insulin in some patients. Only one diabetic patient developed symptomatic hypoglycemia while on sunitinib. The mechanism for the hypoglycemic effect of these drugs is unclear, but of the four agents tested, c-kit and PDGFRb are the common target kinases. Clinicians should keep the potential hypoglycemic effects of these agents in mind; modification of hypoglycemic agents may be required in diabetic patients. These results also suggest that inhibition of a tyrosine kinase, be it c-kit, PDGFRb or some other undefined target, may improve diabetes mellitus BG control and it deserves further study as a potential novel therapeutic option.

Keywords Dasatinib, diabetes, glucose, imatinib, sorafenib, sunitinib

Corresponding author: Nicole Agostino, Division of Hematology–Oncology, Department of Medicine, Penn State Milton S Hershey Medical Center, 500 University Drive, CH46, PO Box 850, Hershey, PA 17033-0850, USA Email: [email protected]


2

J Oncol Pharm Practice 0(0)

Introduction Over the past decade, approval for many new drugs in the tyrosine kinase inhibitor (TKI) family has been obtained by the Food and Drug Administration (FDA) for the treatment of a variety of malignant diseases. These drugs work as a class, inhibiting the phosphorylation of target proteins by occupying the ATP-binding site of the tyrosine kinase.1 Sunitinib and sorafenib inhibit multiple tyrosine kinases including vascular endothelial growth factor (VEGF), c-kit, platelet-derived growth factor-beta (PDGFRb), FLT3, and BRAF. They have been currently approved for the treatment of renal cell carcinoma (RCC). Sorafenib is also FDA approved for unresectable hepatocellular carcinoma. Imatinib and dasatinib are primarily utilized for their inhibition of the Bcr-abl tyrosine kinase, but like the other drugs are somewhat promiscuous, with activity against PDGFRb and c-kit. Both dasatinib and imatinib are used for the treatment of chronic myelogenous leukemia (CML) and Philadelphia chromosome positive acute lymphoblastic leukemia (PhCr+ALL), with dasatinib being approved for imatinib resistance or intolerance. Imatinib also has activity in gastrointestinal stromal tumors (GIST), as does sunitinib, due to c-kit inhibition. In addition, imatinib has been approved for the treatment of the more obscure condition, dematofibrosarcoma protuberans, via its effect on PDGFRb. In our clinical practice, we had anecdotally noticed the effects of these drugs on plasma glucose levels, particularly in diabetics. Recently, scattered reports have appeared in the literature documenting improvements in the glycemic control following administration of imatinib, dasatinib, and sunitinib in diabetic patients.2–5 In addition, hypoglycemia has been reported in nondiabetic patients receiving imatinib.6 Therefore, in an effort to characterize this phenomenon better, we retrospectively reviewed the blood glucose values in patients before, during, and after treatment with TKIs in both diabetic and nondiabetic patients.

Materials and methods A retrospective chart review of 80 patients was performed with data obtained from private physician databases. These patients were being treated with TKIs for one of the four diagnoses: CML, PhCr+ALL, RCC, or GIST. Imatanib, dasatinib, sunitinib, and sorafenib were the drugs evaluated in this review and some patients received two of these medications if progression occurred on one. Too few patients had received nilotinib to comment on that drug. PhlCr+ALL and

CML patients who developed resistance to imatinib were then treated with dasatinib. RCC patients who progressed on sunitinib could have their treatment changed to sorafenib. Demographic data for age, race, sex, hematologic/oncologic diagnosis, diabetic status, and mean blood glucose levels before, during, and after the usage of TKIs were collected. Plasma blood glucose values were collected from basic metabolic panels and no finger stick blood glucose levels were included in the data. The fasting status of the patient at the time of collection was not available for all blood glucose values that were collected, but were likely a mixture of predominantly nonfasting determinations, since the specimens were taken during clinic times and oncologists rarely request fasting determinations. A diagnosis of diabetes mellitus had been made in our patients by their primary care physicians. All of the patients in this study had type II diabetes mellitus.

Statistical analysis A linear model of mixed-effects was applied to the blood glucose measurements. The model contained the following: (1) mean parameters for the phase of the drug administration (before, during, and after) for each of the four drugs (dasatinib, imatinib, sorafenib, and sunitinib) and (2) a covariate for diabetic status for each drug. The initial model included a covariate for age for each drug, but the estimated age covariates were not statistically significant (p > 0.05); so they were excluded from the final version of the model. The model also imposed a separate equi-correlation and equi-variance structure for each of the four drugs to account for repeated measurements within each patient. Linear contrasts, resulting in approximate t-tests, were constructed from the model parameter estimates to compare the before versus during phase and the after versus during phase for each drug. The software package PROC MIXED of SAS 9.1 was used for this statistical analysis.

Results See Table 1 for demographic information for patients in this study. Mean ages ranged from 53 to 66. The percentage of those with diabetes ranged widely from 13% to 23% in each of the drug treatment groups. Six patients in the imatinib group were eventually transitioned to dasatinib. A total of 14 patients in the sunitinib group were eventually treated with sorafenib also. The majority of our patient population were Caucasian. Mean blood glucose levels were lower while nondiabetic patients were on TKIs (Figure 1). In nondiabetic


Agostino et al.

3

Table 1. Demographic information Drug

Total patients

Mean age

Race

Gender

Diagnosis

Diabetes

Imatinib

39

53

Female, 23 Male, 16

8

56

CML, 25 GIST, 13 ALL, 1 CML, 8

Yes, 6 No, 33

Dasatinib Sorafenib

23

65

Sunitinib

30

66

White, 35 Black, 3 Asian, 1 White, 7 Black, 1 White, 22 Black, 1 White, 30

Female, 3 Male, 5 Female, 5 Male, 18 Female, 8 Male, 22

RCC, 23 RCC, 28 GIST, 2

Yes, No, Yes, No, Yes, No,

1a 7 5 18 7b 23

a

One patient in this group overlapped in the diabetic imatinib group. One patient in this group overlapped with the diabetic sorafenib group.

b

patients, mean blood glucose concentrations declined by 8.4–37.9% during TKI therapy. Mean blood glucose declined 52 mg/dL (37.9%) in patients treated with dasatinib and increased significantly, +29 mg/dL (34.1%), after discontinuing treatment. Mean blood glucose levels prior to treatment in the imatinib group were already low and exhibited a more modest decline after treatment of 9 mg/dL (8.4%). However, no significant increase was observed after discontinuation of treatment. Sorafenib treatment was associated with a 12 mg/dL (12.2%) decrease during treatment and blood glucose increased significantly, +10 mg/dL (10.0%), after discontinuation. Sunitinib therapy was associated with a 15 mg/dL (12.8%) reduction in blood glucose on average followed by a +14 mg/dL (13.7%) return to normal after treatment. Figure 1 also illustrates lower mean blood glucose levels for diabetic patients on TKIs. Blood glucose concentrations for diabetic patients treated with dasatinib declined by 52 mg/dL (23.9%) on average, but rose to +30 mg/dL (18%) on average after discontinuing dasatinib. As in the case of nondiabetic subjects, imatinib treatment caused only a modest decline in diabetic patients of 9 mg/dL (4.7%), with no statistically significant change after imatinib was discontinued. Sorafenib treatment was associated with a 12 mg/dL (7.5%) decrease during treatment and increased +11 mg/dL (7.9%) after discontinuation. Mean blood glucose levels in diabetic patients treated with sunitinib fell 14 mg/dL (9.4%), and this was completely reversed after treatment was finished. These data seem to indicate that the magnitude of the effect of the TKIs on glucose was similar in diabetic and nondiabetic subjects, suggesting that the percent decrease was lower in the diabetic patients who had higher starting blood glucose concentrations. Table 2 provides an overview of the changes that were made to diabetes medication regimens in the diabetic patients.

Three diabetic patients had no changes to their diabetic regimens. One patient required an increase in the dose of the diabetes medication and one patient required initiation of diabetes therapy with metformin. Two patients required a decrease in the medication doses. Two patients required changes to different classes of diabetes medications. A total of 47%percent of patients with diabetes were able to stop diabetes medications completely on TKI therapy. Of interest, patient number 80 had his/her diabetes medication stopped after presentation to the emergency department due to an episode of severe and symptomatic hypoglycemia with a blood glucose value of 22.

Discussion We have shown that dasatinib, imatinib, sorafenib, and sunitinib all reduce the blood glucose levels in both diabetic and nondiabetic patients. This inhibition appears to reverse with discontinuation of the drug at least for dasatinib, sorafenib, and sunitinib. Blood glucose levels remained somewhat suppressed in imatinibtreated patients. Whether this was a statistical artifact or real effect remains to be determined in a larger series. Patients on imatinib were transitioned to dasatinib and this may be a reflection of that change. The effects on blood glucose were for the most part modest. Only one diabetic patient developed symptomatic hypoglycemia requiring a discontinuation of hypoglycemic therapy. Of note, however, 8 of 17 (47%) patients with diabetes were able to stop hypoglycemic medication completely suggesting a real therapeutic effect in diabetes patients in terms of BG control. Our study was limited by its retrospective nature, but the results are compelling and suggest that BG and other glycemic parameters should be studied in prospective trials of these agents. Since this was a retrospective study of real oncology clinical practice, few data on factors such as Hemoglobin A-1 c were


4

J Oncol Pharm Practice 0(0)

Dasatinib p=0.03

p=0.007

p=0.95

160 140

Blood glucose (mg/dl)

80

100

120

200 150 50

100

Blood glucose (mg/dl)

180

200

250

p