Medication adherence among adults prescribed imatinib,dasatinib,or nilotinib for the treatment of chronic myeloid leukemia

Abstract

Background

Oral tyrosine kinase inhibitors are the standard of care for chronic myeloid leukemia. Tyrosine kinase inhibitors are administered in an outpatient setting for an indefinite period which may negatively impact adherence. Non-adherence to tyrosine kinase inhibitors is associated with disease progression.

Objectives

To evaluate the need for adherence-enhancing interventions, this study was designed to determine the proportion of chronic myeloid leukemia patients non-adherent to their tyrosine kinase inhibitor regimen. The secondary objective was to identify the influence of patient characteristics on tyrosine kinase inhibitor adherence.

Methods

Cross-sectional retrospective chart and dispensing record reviews were performed to identify patients receiving a tyrosine kinase inhibitor from 1 June 2010 to 31 January 2012. Adherence was evaluated using the medication possession ratio.

Results

A total of 124 patients were included. Thirty-eight (31%) patients were non-adherent to their tyrosine kinase inhibitor regimen. Patients not receiving concurrent medications were more likely to be non-adherent (odds ratio (OR) 2.33, 95% confidence interval (CI) 1.05–5.13, p = 0.04). The median medication possession ratio was 0.95 (IQR = 0.83–1.07). Median medication possession ratio was lower in patients receiving imatinib compared to dasatinib or nilotinib (0.95 vs. 1.00, p = 0.01) and in those less than 50 years old compared to those greater than 50 years old (0.92 vs. 0.97, p = 0.02).

Conclusions

Optimal tyrosine kinase inhibitor adherence in chronic myeloid leukemia patients poses a significant obstacle in achieving best possible outcomes while reducing healthcare costs. In this study, one in three chronic myeloid leukemia patients treated with a tyrosine kinase inhibitor were non-adherent to their regimen. Those at higher risk of non-adherence were on no concurrent medications, less than 50 years old, and those treated with imatinib. Active intervention to improve tyrosine kinase inhibitor adherence should be developed, implemented, and evaluated to improve patient outcomes at our center.

Keywords

Cancer oncology patient adherence chronic myeloid leukemia oral chemotherapy imatinib dasatinib nilotinib

Introduction

Chronic myeloid leukemia (CML) is a myeloproliferative disease characterized by the Philadelphia chromosome, a reciprocal translocation between chromosomes 9 and 22. The translocation results in the formation of a chimeric oncogene, BCR-ABL. BCR-ABL encodes a constitutively active tyrosine kinase oncoprotein, which promotes uncontrolled proliferation, discordant maturation, and a lack of apoptosis in CML cells.^1,2

Imatinib, dasatinib, and nilotinib are potent and selective BCR-ABL tyrosine kinase inhibitors (TKIs). TKIs prevent tyrosine phosphorylation, thereby interfering with signal transduction and disrupting the pathway to CML cell survival.¹ Both first- (imatinib) and second-generation (dasatinib and nilotinib) TKIs have dramatically improved the prognosis for patients with CML and are, therefore, the first-line treatment in this population.³

Appropriate imatinib adherence results in improved patient outcomes and a reduction in healthcare costs. In patients treated with imatinib, those who achieve a complete molecular response (CMR), or whose levels of BCR-ABL transcripts have achieved a three-log reduction (known as a major molecular response [MMR]), have a significantly lower risk of disease progression than those who do not.⁴ The six-year probability of achieving these markers is strongly correlated with imatinib adherence. Studies show that MMRs are not observed when imatinib adherence is less than 80%, while CMRs are not observed when imatinib adherence is less than 90%.^5–7 Achievement of a MMR has also been inversely associated with healthcare expenses,^8–11 suggesting adequate imatinib adherence is associated with lower resource utilization and costs.

Medication adherence is defined as “the extent to which the patient’s behavior matches agreed recommendations from the prescriber.”¹² Adherence rates are less than ideal in some CML patients prescribed a TKI. One study found that after five years of therapy, more than one-third of patients had abandoned their imatinib regimen.^13,14 Of the patients still taking imatinib, only 14.2% of them followed the regimen exactly as prescribed.¹⁴ Barriers to patient adherence include duration of therapy, side effects, lack of efficacy, progression of disease, and reassurance from a healthcare professional that poor adherence does not alter clinical response.^15,16

Evaluation of medication adherence is important as intervention could result in improved patient outcomes. The primary objective of this study was to determine the proportion of CML patients who are ≥90% adherent to their prescribed regimen of imatinib, dasatinib, or nilotinib. The secondary objective was to identify patient subpopulations that are less adherent to these agents. This information can be used to determine the need for development of interventions to improve adherence, such as a pharmacist-conducted education session. Efforts can also be made to target subpopulations at higher risk of non-adherence.

Methods

This was a descriptive study using a retrospective cross-sectional analysis of patients at the Tom Baker Cancer Centre (TBCC). Data were collected from the patient’s electronic chart and pharmacy dispensing record. Approval to conduct this study was granted from the Conjoint Health Research Ethics Board at the University of Calgary on 1 December 2011. Due to the retrospective design, as well as the lack of patient contact and intervention, the study was considered minimal risk. Therefore, patient consent was not required nor obtained.

Patients were included in the study if they had a diagnosis of CML; had received imatinib, dasatinib, or nilotinib from the TBCC at any point between 1 June 2010 and 31 January 2012; had been treated with their current TKI regimen for at least six months (the initial six months was not included in the data collection); and they had at least three months of refill data in addition to the six-month initiation period. Patients were excluded if they were enrolled in a clinical trial during the time of the study, were not Alberta residents, did not meet the Alberta Cancer Board Drug Benefit List criteria for use of a TKI, or were younger than 18 years.

Patients who were initiated on a TKI after the start date of the study were included in the analysis. However, no data were collected until the patient had received the TKI for six months. If a patient switched TKIs during the study, the most recent eligible data were used. The same process applied to those patients whose doses were adjusted during the study. Data eligibility only began after the patient had received the medication at the new dose for six months. Due to these changes, some patients had less than 18 months of data.

There are a variety of calculations, which have been used to estimate adherence in retrospective studies.¹⁷ We chose to use the medication possession ratio (MPR) which determines adherence rate by dividing the number of days the medication was supplied for by the number of days between the first and last refills¹⁷ (see the Appendix). A patient was considered adherent if they took ≥90% (MPR ≥ 0.9) of prescribed doses. Patients with an MPR > 1, meaning they had filled their medications before the expected refill date, were considered 100% adherent (MPR = 1). The MPR was used to determine adherence in this study because it is one of the most commonly used equations in estimating adherence from a retrospective database.¹⁷

The data were not normally distributed, requiring the use of non-parametric statistical tests. Medians and interquartile ranges (IQR) were reported for continuous variables, while frequencies and percentages were reported for categorical data. To analyze differences in adherence between subpopulations, the following variables were compared between those who were adherent and those who were not: the TKI prescribed, age, sex, years of treatment, home address (within or outside of Calgary), number of concurrent medications, side effects, and if the patient was previously treated with interferon. Univariate logistic regression analysis (Chi-square test) was used to determine the relationship of TKI adherence with potential risk factors. Associations between subpopulations and adherence were also investigated by comparing median MPRs between the groups. Median MPRs were compared using the Mann–Whitney U test. Significance was determined based on a p value <0.05. SAS software version 9.2 (SAS Institute Inc., Cary, NC) was used to conduct all statistical analyses.

Results

Two hundred and nine patients received imatinib, dasatinib, or nilotinib from the TBCC between 1 June 2010 and 31 January 2012. All patients were screened for the study and assigned a unique study number to protect confidentiality. Eighty-five patients were excluded: 51 because they had less than three months of refill data, four because they did not have CML, three because they were in a clinical trial, and 27 because the patient was duplicated on the list (prescribed two different TKIs). In total, 124 patients were included in the study. Patient demographics are shown in Table 1. The study population was 63% male. Patients were older than 18 years, 95% of which were between 30 and 90 years old. Seventy-four percent of patients lived within Calgary. Ninety percent of patients were taking imatinib. Seven percent of patients had previously been treated with interferon. In those patients who had side effects listed in their chart, 42% said they had experienced at least one side effect while taking the TKI.

Table 1.

Patient demographics (N = 124).

Baseline characteristics	Number (%)
Sex
Male	78 (63)
Female	46 (37)
Age
18–30	8 (7)
30–50	34 (27)
50–70	52 (42)
70–90	29 (23)
>90	1 (<1)
Residence
Within Calgary	90 (74)
Outside of Calgary	32 (26)
TKI
Imatinib	111 (90)
Dasatinib	7 (6)
Nilotinib	6 (4)
Interferon
Previously treated	9 (7)
Not previously treated	114 (93)
Side effects
Yes	33 (42)
No	46 (58)
Years of treatment
<1	5 (4)
1–3	23 (19)
3–5	32 (26)
5–7	30 (24)
>7	34 (27)
Concurrent medications
0	42 (34)
1–2	29 (23)
3–5	25 (20)
6–9	20 (16)
>10	8 (6)

Thirty-eight of the 124 patients (31%) were non-adherent to their TKI regimen (Figure 1). Patients on no concurrent medications were found to be less adherent than those taking medications in addition to the TKI (OR 2.33; 95% CI 1.053–5.134; p = 0.04). No relationship was found between adherence and other patient characteristics.

The median MPR for the population was 0.95 (IQR 0.83–1.07) (Figure 2). While there was no significant difference in adherence rate, median MPR was significantly lower in patients less than 50 years old compared to those aged 50 years or more (0.92 vs. 0.97, p = 0.02) (Table 2). Median MPR was significantly lower in patients treated with imatinib than those treated with dasatinib or nilotinib (0.95 vs. 1.00, p = 0.01) (Table 3). No significant differences were found in median MPRs among the other patient categories.

Table 2.

Univariate analysis of patient characteristics correlating with adherence.

	Univariate analysis
	Odds ratio	95% CI	p
Female (male = 1)	0.86	0.39–1.90	0.72
<50 years old (>50 years old = 1)	0.51	0.23–1.12	0.09
Resides in Calgary (resides outside of Calgary = 1)	0.82	0.34–2.00	0.67
Treated with imatinib (treated with nilotinib or dasatinib = 1)	0.17	0.02–1.33	0.09
<1 year of treatment with TKI (>1 year of treatment with TKI = 1)	1.81	0.20–16.71	0.60
Concurrent medications (no concurrent medications = 1)	2.33	1.05–5.13	0.04
Side effects present (side effects not present = 1)	0.84	0.36–1.98	0.70
Not previously treated with interferon (previously treated with interferon = 1)	0.65	0.13–3.26	0.60

TKI: tyrosine kinase inhibitor; CI: confidence interval.

Table 3.

Median MPR by patient characteristics.

Categorization	Number (%)	Median MPR (IQR)	p
Sex (N = 124)
Male	78 (63)	0.95 (0.12)	0.80
Female	46 (37)	0.97 (0.13)	–
Age (N = 124)
<50	42 (34)	0.92 (0.20)	0.02
>50	82 (66)	0.97 (0.10)	–
Residence (N = 122)
Inside Calgary	90 (74)	0.95 (0.12)	0.39
Outside Calgary	32 (26)	0.98 (0.15)	–
TKI (N = 124)
Imatinib	111 (90)	0.95 (0.13)	0.01
Dasatinib or nilotinib	13 (10)	1.00 (0.04)	–
Years of treatment (N = 124)
<1 year	5 (4)	0.95 (0.04)	0.76
>1 year	119 (96)	0.95 (0.12)	–
Concurrent medications (N = 124)
Yes	82 (66)	0.95 (0.09)	0.39
No	42 (34)	0.95 (0.14)	–
Side effects (N = 79)
Yes	33 (42)	0.97 (0.13)	0.69
No	46 (58)	0.95 (0.12)	–
Previously on interferon (N = 123)
Yes	9 (7)	0.98 (0.09)	0.61
No	114 (93)	0.95 (0.12)	–

TKI: tyrosine kinase inhibitor; MPR: medication possession ratio; IQR: interquartile ranges.

Thirty-three patient charts confirmed the presence of TKI adverse effects, 46 charts confirmed no TKI adverse effects, while 45 charts did not confirm the presence or absence of TKI adverse effects. The most common adverse effect was nausea (n = 10, 13%) followed by periorbital edema (n = 8, 10%), muscle aches (n = 8, 10%), rash (n = 6, 8%), myelosuppression (n = 6, 8%), diarrhea (n = 6, 8%), fatigue (n = 3, 4%), and chest pain (n = 1, 1%) (Table 4).

Table 4.

TKI adverse effects.

Adverse effect	N	%
None	46	58
Nausea	10	13
Muscle aches	8	10
Periorbital edema	8	10
Rash	6	8
Myleosuppresion	6	8
Diarrhea	6	8
Fatigue	3	4
Chest pain	1	1

TKI: tyrosine kinase inhibitor.

Discussion

The utilization of TKIs in the treatment of CML has enhanced survival rates among patients. However, a positive outcome is correlated with the patient’s ability and willingness to take the medication as prescribed.^5–7 According to the health belief model, a patient’s medication taking behavior is contingent on their evaluation of the benefits versus the costs associated with taking it.¹⁸ Therefore, in those patients who believe that the costs outweigh the benefits of TKI adherence, intervention may alter their perception and improve their outcomes.

Our study aimed to evaluate TKI adherence among patients treated at a tertiary cancer care center over an 18-month period. This population was found to have an overall median MPR of 0.95 (IQR 0.83–1.07). Median/mean MPRs have been reported in other studies in the range of 0.64 and 0.96.^9,11,19,20 Thirty-one percent of patients in our study were found to have a MPR < 0.9. Past studies have reported 30% to 46% of patients to be non-adherent to their TKI regimen.^4,9,11

Despite the relatively high rate of adherence in our study, a significant number of patients remain non-adherent to their TKI regimen. Therefore, intervention to improve patient adherence would be beneficial at our center.

Subpopulation analyses were performed to identify patients at higher risk of non-adherence who are most likely to benefit from active engagement in adherence-enhancing interventions. No association was found between TKI adherence and patient sex, location of residence, years of treatment, side effects, or previous treatment with interferon.

Darkow et al. found that TKI adherence decreased as the number of medications a patient was taking increased. The authors attributed this to pill burden.¹¹ However, in our study, patients on concurrent medications were more likely to be adherent than patients on no medications in addition to the TKI (OR 2.33, CI 1.05–1.53, p = 0.04). Patients on other medications may have established routines or medication reminders to prevent missed doses. Patients prescribed multiple medications may also consider themselves to be in poorer health, providing incentive to take their medications as prescribed.¹¹

Patients treated with imatinib had a lower median MPR than those treated with nilotinib or dasatinib (0.95 vs. 1.00, p = 0.01). This may be a result of different adverse effect profiles. Although similar adverse effects occur, the incidence of each adverse effect differs among the TKIs. For instance, nausea is more common with imatinib than with dasatinib or nilotinib.^21,22 Variation in efficacy may also account for differences in TKI adherence. The DASISION and ENESTnd trials illustrate that patients treated with dasatinib and nilotinib have faster and superior results compared to those treated with imatinib, although survival benefit is yet to be determined.^21,22 No past studies have compared adherence rates between the first- and second-generation TKIs to verify this result. Our result needs to be interpreted with caution as the majority of our patients were prescribed imatinib (90%), leaving a small sample size of patients treated with the second-generation TKIs. In addition, practice at our center during this study was to treat a patient first-line with imatinib. In those who develop imatinib resistance or intolerance, second-generation TKIs were then prescribed. These practices may contribute to the differences in adherence seen among the TKIs.

Patients less than 50 years old were found to have a lower median MPR than those greater than 50 years old (0.92 vs. 0.97, p = 0.02). This suggests that patient perception of disease and medication taking may change throughout their lifetime. Several other studies have also reported lower TKI adherence with younger age.^14,20

Our results suggest that a significant number of our patients may benefit from intervention to improve patient TKI adherence. They also suggest we should actively engage patients less than 50 years old, those treated with imatinib or those on no other medications in adherence-enhancing interventions. Prior studies show that general medication adherence can be improved by increasing patient instruction (verbal, written, and visual materials), telephone follow-up/monitoring, patient education (about disease pathology, side effect management, and importance of therapy), and daily medication reminders.²³ Strategies proven to be associated with improved TKI adherence include establishing a daily routine, using prompts to take the medication, and managing side effects.¹⁶ Intervention to improve TKI adherence should focus on these strategies.

Future studies that focus on the creation, implementation, and evaluation of interventions to improve adherence would be beneficial to our patient population. Research to explore the rationale for patient non-adherence in high-risk populations including those on no concurrent medications, those less than 50 years old, and those treated with imatinib would also be useful.

This study had limitations inherent to its retrospective design. Retrospective data can determine association, but not causation of endpoints. Retrospective data collection are also completely reliant on the computer refill record, which can be misleading. For example, the calculation of MPR would suggest a patient was adherent if medication was refilled at the proper time, regardless of the amount actually taken. The reverse is also true. For example, despite the six-month wait period, it is possible that some patients had agreed with their healthcare team to take a short break from their medications, for side effects or personal reasons, which would make the patient appear less adherent in the dispensing history. Retrospective data also resulted in missing information for subpopulation analysis. For instance, some patients saw their oncologist at a separate tertiary center even though their medications were filled at the TBCC. For these patients, refill data were available but the patient notes were not accessible and could not be reviewed for side effects. At the time of this study, nilotinib and dasatinib were relatively new medications used second-line following imatinib resistance or intolerance. Therefore, the sample size of patients taking nilotinib or dasatinib was relatively small. Additionally, patients who are intolerant to imatinib or have become resistant to its benefits may be more inclined to be adherent to their regimen. Because of these factors, results for differences among TKI median MPRs must be interpreted with caution.

Conclusion

Optimal TKI adherence in CML patients poses a significant obstacle in achieving best possible outcomes while reducing healthcare costs. In this study, one in three CML patients treated with a TKI were non-adherent to their regimen. Patients at higher risk of non-adherence were on no concurrent medications, less than 50 years old, or treated with imatinib. Active intervention to improve TKI adherence should be created, implemented, and evaluated to improve patient outcomes at our center.

Footnotes

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Conflict of interest

ML Savoie has received honoraria from and participated in advisory boards on behalf of Novartis and BMS. The other authors declare that there are no other conflicts of interest.

Acknowledgments

This work has been awarded with the Alberta Hospital Residency Project Award for 2012. The authors thank Lauren Bresee, Project coordinator, and Sunita Ghosh, Project statistician, for their help with this work.

Appendix

References

Cortes

Silver

Kantarjian

. Holland-Frei cancer medicine, 8th ed. CT: Shelton, 2010.

Van Etten R. Clinical manifestation and diagnosis of 3 chronic myeloid leukemia. In: Larson R (ed.) Uptodate, Waltham, MA: www.uptodate.com (accessed 29 May 2012).

De Almeida

Barbosa Pagnano

Sanches Souza

. Adherence to tyrosine kinase inhibitors (TKI) in chronic myeloid leukemia (CML) seems to be related to duration of treatment and type of TKI. Haematologica 2010; 95: 343–343. [In: 15th congress of the European Hematology Association, conference publication].

Druker

Guilhot

O’Brien

. Five-year follow-up of patients receiving imatinib for chronic myeloid leukemia. N Engl J Med 2006; 355: 23–23.

Marin

Bazeos

Mahon

. Adherence is the critical factor for achieving molecular responses in patients with chronic myeloid leukemia who achieve complete cytogenetic responses on imatinib. J Clin Oncol 2010; 28: 2381–2388.

Ganesan

Sagar

Dubashi

. Nonadherence to imatinib adversely affects event free survival in chronic phase chronic myeloid leukemia. Am J Hematol 2011; 86: 471–474.

Ibrahim

Eliasson

Apperley

. Poor adherence is the main reason for loss of CCyR and imatinib failure for chronic myeloid leukemia patients on long-term therapy. Blood 2011; 117: 3733–3736.

Breccia

Efficace

Alimena

. Imatinib treatment in chronic myelogenous leukemia: what have we learned so far? Cancer Lett 2011; 300: 115–121.

Johnson

Beaulieu

. Healthcare resource utilization and costs associated with non-adherence to imatinib treatment in chronic myeloid leukemia patients. Curr Med Res Opin 2010; 26: 61–69.

10.

Bollu

Guo

. Non-adherence to imatinib in chronic myeloid leukemia patients is associated with a short term and long term negative impact on healthcare utilization and costs. Blood 2009; 114: abstract 4270–abstract 4270. [In: 51st annual meeting of the American Society of Hematology, conference publication].

11.

Darkow

Henk

Thomas

. Treatment interruptions and non-adherence with imatinib and associated healthcare costs: a retrospective analysis among managed care patients with chronic myelogenous leukaemia. Pharmacoeconomics 2007; 25: 481–496.

12.

Horne R, Weinman J, Barber N, et al. Concordance, adherence and compliance in medicine taking. Report for the National Co-ordinating Centre for NHS Service Delivery and Organisation R & D (NCCSDO). December 2005, http://www.netscc.ac.uk/hsdr/files/project/SDO_FR_08-1412-076_V01.pdf (2005, accessed 30 May 2012).

13.

Wei

Rafiyath

Liu

. First-line treatment for chronic myeloid leukemia: dasatinib, nilotinib, or imatinib. J Hematol Oncol 2010; 3: 47–47.

14.

Noens

Van Lierde

De Bock

. Prevalence, determinants, and outcomes of nonadherence to imatinib therapy in patients with chronic myeloid leukemia: the ADAGIO study. Blood 2009; 113: 5401–5411.

15.

Marin

. Current status of imatinib as frontline therapy for chronic myeloid leukemia. Semin Hematol 2010; 47: 312–318.

16.

Eliasson

Clifford

Barber

. Exploring chronic myeloid leukemia patients' reasons for not adhering to the oral anticancer drug imatinib as prescribed. Leuk Res 2011; 35: 626–630.

17.

Peterson

Nau

Cramer

. A checklist for medication compliance and persistence studies using retrospective databases. Value Health 2007; 10: 3–12.

18.

Vermeire

Hearnshaw

Royen

. Patient adherence to treatment: three decades of research. A comprehensive review. J Clin Pharm Ther 2001; 26: 331–342.

19.

Doti

Stemmelin

Shanley

. Cytogenetic response in relation to the adherence to treatment with imatinib mesylate: a case control study. Blood 2007; 110: 4553–4553.

20.

St Charles

Bollu

Hornyak

. Predictors of treatment non-adherence in patients treated with imatinib mesylate for chronic myeloid leukemia. Blood 2009; 114: 2209–2209.

21.

Cortes

Jones

O’Brien

. Results of dasatinib therapy in patients with early chronic-phase chronic myeloid leukemia. J Clin Oncol 2010; 28: 398–404.

22.

Saglio

Kim

Issaragrisil

. Nilotinib versus imatinib for newly diagnosed chronic myeloid leukemia. N Engl J Med 2010; 362: 2251–2259.

23.

Haynes

Ackloo

Sahota

. Interventions for enhancing medication adherence. Cochrane Rev 2008; 4: 1–159.