Real-world adherence and discontinuation among Medicare beneficiaries initiating venetoclax vs. BTKis in relapsed/refractory chronic lymphocytic leukemia

Abstract

The treatment landscape for chronic lymphocytic leukemia (CLL) has been transformed by the availability of Bruton’s tyrosine kinase inhibitors (BTKis) and the B-cell lymphoma 2 (BCL-2) inhibitor venetoclax. Despite clinical trial data supporting these novel oral agents, evidence evaluating real-world adherence is limited. This study used 2015–2019 Medicare claims data for elderly patients with relapsed/refractory CLL to assess differences in real-world adherence and discontinuation in the 12 months after treatment initiation. In the final sample of 711 venetoclax patients and 1,566 BTKi patients, we found that those initiating venetoclax tended to be younger (mean age 75.6 [SD 6.0] vs 77.6 [SD 6.9] years, p < .001) but had poorer clinical characteristics. After risk-adjustment, the venetoclax group had higher adherence (61.9% vs. 45.4%, p < .0001) and lower discontinuation when compared to the BTKi group (28.5% vs. 47.4%, p < .001). These favorable real-world findings underscore the importance of developing well-tolerated novel combinations for older adults.

Keywords:

• CLL
• venetoclax
• adherence
• discontinuation
• BTKi

Introduction

Chronic lymphocytic leukemia (CLL) is a malignancy of auto-reactive mature B cells and represents the most common leukemia among adults in Western countries [1]. CLL is most prevalent in elderly patients with a median age at diagnosis of 72 years [2]. While asymptomatic patients are managed with active surveillance, patients presenting with symptomatic disease or compromised bone marrow function are placed on therapy [3,4]. Until 2014, treatment options were predominantly cytotoxic chemotherapy (CT) agents utilized as monotherapy or in various combinations with anti-CD20 monoclonal antibodies (chemoimmunotherapy, CIT) [5,6].

The 2014 approval of the first BTK inhibitor (BTKi) ibrutinib changed the treatment paradigm and improved outcomes for CLL patients [7]. Second-generation BTKis such as acalabrutinib have also become available on the market, with many other BTKis in the pipeline. Recently, the new first-in-class BCL-2 inhibitor, venetoclax, has offered patients an alternative to BTKis as a fixed-duration treatment providing deep, durable responses [8,9]. Further, in those patients progressing or intolerant to BTKis, venetoclax-based therapy remains highly efficacious [10]. In addition to improved outcomes relative to traditional chemotherapy, both BTKis and venetoclax are oral self-administered therapies that may offer greater convenience for patients. However, real-world adherence to self-administered oral anticancer agents can be challenging [11,12]. This is particularly worrying given premature discontinuation of therapy has been associated with suboptimal clinical outcomes for individuals with CLL [13,14].

Evidence evaluating real-world adherence of novel therapies in CLL is limited, particularly those comparing rates of adherence and discontinuation in patients receiving venetoclax vs. BTKis. Evaluating real-world adherence to these CLL agents is especially important for older individuals covered through Medicare, a federally-funded insurance program for elderly and disabled adults in the United States. This is because CLL is most prevalent in the older population and individuals on Medicare may face several challenges (multiple comorbidities, polypharmacy, out-of-pocket costs) that in turn impact adherence. The objective of this study was to examine real-world adherence and discontinuation of venetoclax vs. BTKis in the first year after initiation in older adults with relapsed/refractory (R/R) CLL.

Materials and methods

Study design and data Source

This retrospective cohort study utilized the 2015–2019 100% Medicare claims data extract available from the Centers for Medicare and Medicaid Services (CMS). The data extract included Medicare Part A and Part B medical claims as well as Part D prescription claims for all fee-for-service Medicare beneficiaries with CLL. The Medicare claims files were linked to personal summary files that contain patient demographics, eligibility, and date of death information.

Sample selection

The study sample consisted of patients receiving treatment with venetoclax or BTKis for CLL. Since the population of venetoclax users was expected to be smaller due to its recent approval in June 2018 in the R/R setting, the sample of venetoclax patients was selected first to maximize the sample size. BTKi users were then identified from the remaining population. Given our data period ended in December 2019, we anticipated most patients in our BTKi sample would be treated with ibrutinib rather than second-generation agents (i.e. acalabrutinib or zanubrutinib). A sample selection schematic is presented in Figure 1.

Figure 1. Sample Selection schematic. BTKi: Bruton tyrosine kinase inhibitor; Dx: diagnosis; CLL: chronic lymphocytic leukemia; SLL: small lymphocytic leukemia; Rx: prescription.

Patients in the venetoclax sample were required to meet the following criteria: (1) ≥1 claim for venetoclax starter pack (NDC: 00074057928) between January 1, 2016 and December 31, 2018 from a standalone Medicare Part D plan (index date = date of the first venetoclax starter pack claim); (2) continuous fee-for-service Medicare Part A, B, and D coverage for at least 12 months pre- and post-index date (or until death); (3) ≥66 years old on index date; (4) ≥1 diagnoses of CLL/SLL (ICD-9-CM: 200.8x or 204.1x; ICD-10-CM: C91.1x or C83.0x) in the 12-month pre- and post-index period; (5) no evidence of another indication for venetoclax (i.e. acute myeloid leukemia [ICD-10-CM: C92.0x or C92.4x or C92.5x, ICD-9-CM: 205.0x]) during 12-month pre- or post-index period; (6) no evidence of a claim for venetoclax in the 12-month pre-index period (i.e. to identify ‘new users’); (7) evidence of ≥1 prior CLL treatment in the 12-month pre-index period and (8) evidence of ≥1 fills of index agent in the post-index period but not including index date (to ensure patients were actually treated with the agent rather than just having a single claim).

After identifying all possible venetoclax patients, the sample of BTKi patients was selected from the remaining population. Patients were included in the BTKi sample if they had evidence of ≥1 claim for a BTKI treatment (ibrutinib, acalabrutinib, zanubrutinib) between January 1, 2016 and December 31, 2018 from a standalone Medicare Part D plan (index date = date of the first BTKI prescription claim). Patients in the BTKi sample were required to meet identical continuous coverage, age, and CLL/SLL diagnostic criteria as the venetoclax sample, in addition to the following criteria: (1) no evidence of other indications for BTKis (i.e. mantle cell lymphoma [MCL], marginal zone lymphoma [MZL], Waldenström macroglobulinemia, or chronic graft versus host disease) during 12-month pre- or post-index period; (2) no evidence of a claim for any venetoclax (starter or maintenance pack – all NDCs) prescription in the 12-month pre-index period including index date; (3) no evidence of a claim for the index BTKi agent in the 12-month pre-index period (i.e. to identify ‘new users’); (4) evidence of ≥1 prior CLL treatment in the 12-month pre-index period and (5) evidence of ≥1 fills of index agent in the post-index period but not including index date (to ensure patients were actually treated with the agent rather than just having a single claim).

Outcomes

The primary study outcomes included rates of adherence to and discontinuation of the index CLL/SLL agent in the first year after venetoclax or BTKi treatment initiation. Although the FDA-approved duration of venetoclax is 24-months in the relapsed/refractory setting and BTKi treatment is to be received indefinitely, we opted for a 12-month follow-up period given that the latest available data at the time of the analysis ended on December 31st, 2019 (approximately 19 months from the FDA approval of venetoclax for relapsed/refractory CLL). Additionally, prior studies of BTKi treatment patterns have shown that the majority of discontinuations due to adverse effects occur within 12-months [15,16]. Hence, we would expect to observe differences between the two agents, if any, within 12 months of initiation.

Adherence was defined using the proportion of days covered (PDC) method, measured as the number of days covered with the index treatment divided by a fixed time interval from the date of index treatment initiation. For this analysis, we used a 365-day fixed time interval to measure adherence for both study samples. A patient with a PDC ≥0.80 during the 12-month post-index period was considered adherent to their index treatment. Additional details on the PDC calculation can be found in the Appendix.

Discontinuation of the index treatment was captured as a dichotomous measure indicating the presence of a fixed period of at least 90 consecutive days with no supply of the index treatment after the days’ supply of its most recent prescription is exhausted. We conducted sensitivity analyses using at least 60-day and 120-day continuous gaps to define discontinuation.

In addition to adherence and discontinuation, we also assessed the number of 30-day supply equivalent prescription fills for the index drug. This represents the total number of venetoclax or BTKi prescription refills in the 12-month pre-index period. The counts for all prescriptions were standardized into 30-day supply equivalents by dividing the days’ supply on each prescription claim by 30. Only fills for the index agent were counted in each study group.

Analytic approach

Patient sociodemographic and clinical characteristics were reported for both the venetoclax and BTKi study samples. Descriptive analyses were also conducted to compare rates of adherence, discontinuation, and the number of 30-day supply equivalent prescription fills between the venetoclax and BTKi samples. Logistic regressions were estimated for the two binary outcomes of adherence (PDC ≥0.80) and discontinuation (treatment gap of ≥90 d). The regressions controlled for sociodemographic and clinical characteristics, including age, sex, race, census region, metropolitan status, Part D low-income subsidy status, Part D drug benefit type, number of Elixhauser comorbidities in the 12-month pre-index period, use of concomitant CD20 therapy (4-week pre-index and 8-weeks post-index drug initiation), number of prior CLL agents in the 12-month pre-index period, evidence of any hospitalization in the 12-month pre-index period, death during the 12-month post-index period, and index year (indicators for the year when BTKi or venetoclax was initiated).

This study was deemed exempt from review by Pearl IRB according to FDA 21 CFR 56.104 and 45CFR46.104(b)(4): (4)Secondary Research Uses of Data or Specimens on 03/01/2021. A Waiver of Individual Authorization under HIPAA pursuant to 45 CFR 164.512 (i)(2)(i)–(v) exempt status as specified in 45 CFR 164.512 was approved.

Results

The final sample contained two mutually exclusive groups of beneficiaries initiating venetoclax (n = 711) or BTKis (n = 1566). The vast majority of BTKi patients (>99%) in our sample were treated with ibrutinib, though a small number had evidence of treatment with acalabrutinib (n = 12, <1%). Patient sociodemographic and clinical characteristics are outlined in Table 1. In the final sample of venetoclax patients, the median (IQR) age was 75.0 (71.0, 80.0) years; the BTKi sample was slightly older with a median (IQR) age of 77.0 (72.0, 83.0). The number of patients ≥80 years was significantly lower for the venetoclax group than the BTKi group (25.9% vs. 38.2%, p < .0001). Both samples did not significantly differ on other sociodemographic characteristics and were largely male (56.9% [venetoclax], 59.5% [BTKi]), white (>88.7% [venetoclax], 88.8% [BTKi]), and residing in the South (37.6% [venetoclax], 37.4% [BTKi]). Venetoclax patients were more likely to live in an urban area than BTKi patients (84.5% vs. 78.9%, p < .01). The venetoclax group had a higher comorbidity burden as measured by the Elixhauser comorbidities compared to the BTKi group (39.4% vs. 30.4% with ≥8 comorbidities, p < .001). Specifically, the venetoclax group was more likely to have congestive heart failure (21% vs. 13.5%, p < .001), valvular disease (30.9% vs. 22.8%. p < .001), hypothyroidism (28.4% vs. 24.1%, p < .05), coagulopathy (50.6% vs. 43.4%, p < .001), obesity (19.8% vs. 13.4%, p < .001), blood loss anemia (6.5% vs. 4.5%, p < .05), and fluid/electrolyte disorders (46.7% vs. 37.5%, p < .001). The venetoclax group had evidence of receiving a greater number of CLL/SLL treatments within the previous 12 months (15.8% vs. 10.7% with ≥3 prior CLL treatments, p < .01) and a higher rate of all-cause hospitalization (51.5% vs. 42.5%, p < .001) in the 12-month pre-index period compared to the BTKi group.

Table 1. Patient characteristics.

	Venetoclax Group		BTKi Group		p value
Characteristic	N	%	N	%
N	711	100%	1566	100%
Age, median (IQR)	75.0 (71.0, 80.0)		77.0 (72.0, 83.0)		<.0001
Age categories^a, years					<.0001
65–69	125	17.6%	212	13.5%
70–74	210	29.5%	364	23.2%
75–79	192	27.0%	392	25.0%
80+	184	25.9%	598	38.2%
Sex^a					.957
Male	424	59.6%	932	59.5%
Female	287	40.4%	634	40.5%
Race^a					.8368
White	>631	>88.7%	1390	88.8%
Black	47	6.6%	119	7.6%
Hispanic	n < 11	<1.5%	11	0.7%
Other	22	3.1%	46	2.9%
Census Region^a					.6252
Northeast	161	22.6%	326	20.8%
Midwest	161	22.6%	389	24.8%
South	267	37.6%	585	37.4%
West	122	17.2%	266	17.0%
Metropolitan Status^a					.0017
Urban	601	84.5%	1236	78.9%
Rural	110	15.5%	330	21.1%
Part D Low-Income Subsidy (LIS) Status^a					.0001
Full or Partial LIS	63	8.9%	229	14.6%
Non-LIS	648	91.1%	1337	85.4%
Part D Drug Benefit Type^a					.4316
Basic alternative	142	20.0%	308	19.7%
Enhanced alternative	361	50.8%	748	47.8%
Defined standard benefit	107	15.0%	255	16.3%
Other	101	14.2%	255	16.3%
Number of Elixhauser comorbidities (12-month pre-index)*					.0003
0	n < 11	<1.5%	20	1.3%
1–2	69	9.7%	170	10.9%
3–4	128	18.0%	343	21.9%
5–7	>223	>31.4%	557	35.6%
8–10	185	26.0%	346	22.1%
11+	95	13.4%	130	8.3%
Anti-CD20 treatments around index drug initiation (4 week pre-index and 8-weeks post-index)^a
No Anti-CD20	502	70.6%	1196	76.4%	.0034
Obinutuzumab	49	6.9%	66	4.2%	.0069
Rituximab	164	23.1%	304	19.4%	.0456
Type of CLL treatments around index drug initiation (4-week pre-index and 8-weeks post-index)
None (i.e. venetoclax or BTK only and none of the drugs below)	325	45.7%	1079	68.9%
Obinutuzumab	49	6.9%	66	4.2%	.0069
Rituximab	164	23.1%	304	19.4%	.0456
BTK Inhibitors	181	25.5%	N/A	N/A
Other CLL treatments (not listed above)	79	11.1%	213	13.6%	.0995
Number of prior CLL agents (in 12-month pre-index period)^a					.0021
1–2	599	84.2%	1399	89.3%
3–4	93	13.1%	144	9.2%
5+	19	2.7%	23	1.5%
Type of prior CLL treatments (in 12-month pre-index period)
Chemotherapy ± CD20	155	21.8%	814	52.0%	<.0001
CD20 monotherapy (no other flag for chemo, PI3K, BTKi, or other classes not shown in the rows here)	74	10.4%	677	43.2%	<.0001
PI3K ± CD20	82	11.5%	58	3.7%	<.0001
BTKI ± CD20	493	69.3%	0	0.0%	<.0001
All-cause hospitalization in the 12-month pre-index period^a	363	51.1%	665	42.5%	.0001
Death during 12-month post-index period^a	155	21.8%	315	20.1%	.3571
Index year^a					<.0001
2016	72	10.1%	601	38.4%
2017	206	29.0%	524	33.5%
2018	433	60.9%	441	28.2%

Cell sizes <11 cannot be reported due to CMS policy.

^aCharacteristic included as a covariate in the multivariable logistic regression models.

Table 2 displays the number of 30-day refills and unadjusted adherence/discontinuation results. The mean number of 30-day refills was 8.5 (SD 4.0) for venetoclax patients and 7.8 (SD 4.0) for BTKi patients. Less than half (47.3%) of patients treated with a BTKi were adherent over the 12-month follow up period compared to 57.9% of venetoclax patients. The venetoclax group had a higher unadjusted rate of adherence compared to the BTKi group (57.9% vs. 47.3%, p < .001). Additionally, unadjusted rates of discontinuation were higher for the BTKi patients compared to the venetoclax patients (45.1% vs. 32.8%, p < .001). Sensitivity analyses consistently showed higher discontinuation within 12 months for the BTKi group when using a 60-day gap (50.5% vs. 38.5%, p < .001) and a 120-day gap (40.2% vs. 27.8%), p < .001).

Table 2. Unadjusted treatment pattern outcomes.

	Venetoclax Group		BTKi Group
	N	%	N	%
N	711	100%	1566	100%
Number of 30-day refills, mean (SD)	8.5 (4.0)		7.8 (4.0)
2 fills	75	10.5%	172	11.0%
3–5 fills	127	17.9%	402	25.7%
6–8 fills	99	13.9%	236	15.1%
9–11 fills	134	18.8%	287	18.3%
12 fills	135	19.0%	233	14.9%
13 fills	120	16.9%	209	13.3%
14 or more fills	21	3.0%	27	1.7%
Adherent (PDC ≥0.80)	412	57.9%	741	47.3%
Discontinuation (≥90-day gap) - primary definition	233	32.8%	706	45.1%
Discontinuation (≥60-day gap) - sensitivity analysis definition	274	38.5%	791	50.5%
Discontinuation (≥120-day gap) - sensitivity analysis definition	198	27.8%	629	40.2%

Risk-adjusted rates of adherence and discontinuation are reported in Figures 2 and 3, respectively. After risk-adjustment, the venetoclax group had an even higher adherence rate than the BTKi group (61.9% vs. 45.4%, p < .0001, Figure 2). Similarly, the risk-adjusted rate of discontinuation (90-day gap) was even lower for the venetoclax group compared to the BTKi group (28.5% vs. 47.4%, p < .001).

Figure 2. Risk-Adjusted* Rate of adherence among Medicare patients treated with venetoclax vs. BTKi. *Risk-adjusted rates estimated from multivariable logistic regressions. BTKi: Bruton tyrosine kinase inhibitor.

Figure 3. Risk-Adjusted Rate of discontinuation among Medicare patients treated with venetoclax vs. BTKi. * Risk-adjusted rates estimated from multivariable logistic regressions. BTKi: Bruton tyrosine kinase inhibitor.

Discussion

This population-based study represents the first real-world, direct comparison between venetoclax and BTKi patients with respect to adherence and discontinuation. Given the scarcity of data comparing novel agents, our study offers several important insights in a high-risk, high-prevalence CLL population.

Our study found that Medicare patients with R/R CLL who initiated venetoclax had several key differences from patients who initiated a BTKi. While venetoclax patients tended to be younger than BTKi patients, they also tended to have a greater number of comorbidities in the preceding 12 months. Rates of congestive heart failure, coagulopathy, valvular disease, and blood loss anemia were all higher in patients receiving venetoclax. Finally, venetoclax patients were more likely to have a history of being hospitalized relative to BTKi patients, potentially indicating a higher level of frailty before initiation of treatment.

Although a higher prevalence of comorbidities and recent hospitalizations could be predicted to lower adherence among venetoclax patients relative to the BTKi patients, our study found the opposite. Both unadjusted and adjusted measures of adherence and discontinuation favored venetoclax. In fact, in adjusted analyses, nearly two-thirds of venetoclax patients were adherent, compared to less than half of BTKi patients. Similarly, we found that risk-adjusted rates of discontinuation were lower for venetoclax patients relative to BTKi patients. While a little more than a quarter of venetoclax patients discontinued within 12 months of initiation, nearly half of BTKi patients had evidence of discontinuation during this period. These findings persisted in sensitivity analyses.

While few prior studies have compared patients treated with venetoclax vs. BTKis, it should be noted that our findings are similar to those reported by Eyre et al. [17]. This multicenter, international study conducted an indirect comparison of 48 venetoclax patients and 385 ibrutinib patients using previously published data. The authors found that patients initiating ibrutinib had a higher rate of discontinuation compared to those initiating venetoclax (41% vs. 25%, p = .06). Although clinical trials of R/R CLL have typically reported low rates of discontinuation within a year of treatment initiation, our findings add to a growing body of literature which suggests that real-world tolerability of BTKis may be worse than that reported in clinical trials and limited multicenter studies, particularly in the elderly population [15,17–19].

Our claims-based study is unable to determine the reason for the differences in discontinuation between BTKis and venetoclax. Hence, future work is needed to investigate whether differences in CLL progression, side effects, or other factors are associated with the observed differences in discontinuation between these agents. For instance, Eyre et al. [17] reports that the most common reason for discontinuation of ibrutinib was adverse events (22%) while the most common reason for discontinuation of venetoclax was an allogenic stem cell transplant (10%). Prior studies have also shown that adverse events and toxicities are the primary reason for discontinuation among BTKi-treated patients [16]. While a more favorable safety profile likely contributed to lower discontinuation and better adherence in the venetoclax group despite a greater number of comorbidities and higher pre-index hospitalizations in our study, it should be noted that nearly 70% of individuals in this group received a BTKi within the previous 12 months. Hence, it is possible that these patients may have had a greater clinical incentive to remain on venetoclax due to lack of other treatment options (after having already tried a BTKi). On the other hand, one might argue that these patients are also more likely to have refractory CLL (given they failed a BTKi in the past 12 months) and hence are more likely to be refractory to venetoclax as well, which would increase their discontinuation of venetoclax. Hence, this remains an area for future investigation. Further research is also needed to evaluate the differences in longer-term adherence and discontinuation since BTKis are to be taken daily unless the patient experiences an adverse event or disease progression. However, venetoclax is also a once-daily pill, but unlike a BTKi, it is a fixed-duration treatment, given for 12 or 24 months (frontline and R/R, respectively), inducing durable remission with prolonged treatment-free intervals [9,20]. Furthermore, future investigation is also needed to examine how differences in adherence and discontinuation between venetoclax and BTKIs translate into differences in clinical outcomes and health care costs in the real-world setting.

Our study has several limitations. First, as with any administrative claims-based study, coding errors are possible in the medical claims data. Second, medical claims data do not possess detailed clinical data such as disease stage that would allow for direct identification of patients as relapsed/refractory; hence, we had to use the evidence of a claim for any CLL treatment in the past 12 months as an alternative method of identifying relapsed/refractory patients. Given our 12-month window to identify prior CLL treatment, our sample is more likely to have included refractory patients since patients can go several years without CLL treatment prior to experiencing a relapse. It should also be noted that our claims-based measures of adherence and discontinuation only captured whether or not the BTKi or venetoclax prescription was filled by the patient at the pharmacy; we cannot know if the medications were actually consumed. However, these limitations are a necessary tradeoff that allow for more timely, efficient, and generalizable population-level analyses than is permitted by prospective follow-up and limited multicenter trials. Third, our measures of adherence and discontinuation are based on the days’ supply (e.g. 28 d, 30 d) recorded on the prescription claim. In instances where the physician may have asked the patient to lower their dose, the same prescription could have lasted for a longer duration. Since claims data do not include providers’ instructions, our measures are unable to account for the possibility that patients may have been switched to a lower dose thus stretching their prescription’s days’ supply (e.g. halving the dose which doubles the days’ supply of the prescription) and hence may underestimate adherence and overestimate discontinuation. However, our discontinuation findings were consistent across sensitivity analyses varying the treatment gap from 60 d to up to 120 d, and hence unlikely to be explained by dose reductions recommended by the doctor. Fourth, our study is generalizable only to the fee-for-service Medicare population rather than the Medicare Advantage population. However, 60–70% of the U.S. Medicare population was covered under the fee-for-service Medicare program during our study period [21]. Finally, there may be systematic differences between patients receiving BTKis or venetoclax. While multivariate analysis will control for observed confounders, there could still be unobserved confounders between the two groups that could bias the results of our real-world observational study.

This real-world study represents the first comparison of adherence and discontinuation among patients treated with BTKis or venetoclax. We found that elderly adults with R/R CLL treated with venetoclax had higher adherence and lower discontinuation rates relative to BTKi-treated patients over 12 months after initiation, despite having greater comorbidities and pretreatment hospitalizations. The differences we observed may reflect better tolerability or effectiveness (i.e. lack of progression) of venetoclax relative to BTKis, particularly in elderly patients in real-world clinical practice. These favorable results for venetoclax in this large real-world study are encouraging and support the goal of developing well-tolerated novel combinations for elderly adults in the relapsed/refractory setting. As data continues to accumulate on venetoclax (and newer BTKis), future research should be conducted over longer timeframes and must also examine differences in clinical outcomes and healthcare costs.

Disclosure statement

JTP, SKB: full-time employees of COVIA Health Solutions, a consulting firm with clients in the biotech/pharmaceutical industry; BSM, HA, DJ: employees of AbbVie Inc. and may hold stock or stock options; AR: employee of Genentech Inc. and may hold stock or stock options; SFH: consultancy for Janssen, Pharmacyclics, AbbVie, AstraZeneca, Flatiron Health Inc., Novartis, SeaGen, Genetech, Merck, TG Therapeutics, ADC Therapeutics, Epizyme, Servier, Arvinas, and Thyme Inc.; research funding from Celgene, DTRM Biopharm, and TG Therapeutics; honoraria form Pharmacyclics and AstraZeneca, Bayer; JAD: consultancy for AbbVie, Acadia, Allergan, Boehringer Ingelheim, Catabasis, Ironwood Pharmaceuticals, Janssen, Kite Pharma, MeiraGTx, Merck, Otsuka, Regeneron, Sarepta, Sage Therapeutics, Sanofi, Takeda, The Medicines Company, and Vertex; research funding from AbbVie, Biogen, Humana, Janssen, Merck, Novartis, Pfizer, PhRMA, Regeneron, Sanofi, and Valeant.

Data availability statement

AbbVie is committed to responsible data sharing regarding the clinical trials we sponsor. This includes access to anonymized, individual, and trial-level data (analysis data sets), as well as other information (e.g. protocols, clinical study reports, or analysis plans), as long as the trials are not part of an ongoing or planned regulatory submission. This includes requests for clinical trial data for unlicensed products and indications.

These clinical trial data can be requested by any qualified researchers who engage in rigorous, independent, scientific research, and will be provided following review and approval of a research proposal, Statistical Analysis Plan (SAP), and execution of a Data Sharing Agreement (DSA). Data requests can be submitted at any time after approval in the US and Europe and after acceptance of this manuscript for publication. The data will be accessible for 12 months, with possible extensions considered. For more information on the process or to submit a request, visit the following link: https://www.abbvie.com/our-science/clinical-trials/clinical-trials-data-and-information-sharing/data-and-information-sharing-with-qualified-researchers.html.

Additional information

Funding

This work was supported by AbbVie Inc. and Genentech Inc. Venetoclax is being developed in a collaboration between AbbVie Inc. and Genentech Inc. AbbVie Inc. and Genentech Inc. sponsored the study and participated in the design; study conduct; interpretation of the data; and the writing, review, and approval of the publication. No honoraria or payments were made for authorship. All authors contributed to the development of the publication and maintained control over the final content.

Appendix

Adherence to the index drug in each study group was examined using the proportion of days covered (PDC) method, measured as the number of days covered with the oral anticancer agent divided by a fixed time interval from date of index treatment initiation. For this analysis, we used a 365-day fixed time interval to measure adherence for both venetoclax and the BTKi study groups (Rationale: Venetoclax patients should be receiving treatment for 12 months in the frontline setting and 24 months in the relapsed/refractory setting; the goal with ibrutinib treatment is also to at least continue treatment for an indefinite period unless patient has adverse effects or disease progression). A patient with a PDC ≥0.80 was considered adherent to therapy.

• If a patient fills their next prescription for the index drug (e.g. venetoclax) before the end of days’ supply of the previous prescription fill for the same drug, then it was assumed that the patient will finish the prior prescription before starting the new prescription and hence the start date of the next prescription fill was pushed out to adjust for the overlapping days’ supply.

• Any day a patient spent in the hospital or in a skilled nursing facility was still counted in the measurement period as being covered by the days’ supply from the previous prescription filled for the index drug in the outpatient setting since most facilities will ask patients’ family member to bring from home their expensive cancer medications that the patient can continue taking in the hospital or skilled nursing facility.