Contemporary "Real World" Molecular Testing and Tyrosine Kinase Inhibitor Adherence Patterns Among Older Pts with Chronic Myeloid Leukemia in the United States

Introduction:

Patients (pts) with chronic-phase chronic myeloid leukemia (CML) are recommended to have quantitative BCR-ABL1 polymerase chain reaction (qPCR) testing every 3 months (mo) during the 1 ^st year of tyrosine kinase inhibitor (TKI) treatment to assure achievement of milestone response goals. Adherence to TKI therapy is also critical to achieving responses and ultimately functional cure. Prior evaluations of qPCR testing/TKI adherence have only focused on imatinib-treated and/or younger pts. We sought to evaluate older CML pts in the United States using the Surveillance, Epidemiology, and End Results (SEER)-Medicare dataset.

Methods:

Using the SEER-Medicare database, we assembled a population-based cohort of CML pts diagnosed in 2007-17 who: 1) were age 66-99 years at diagnosis, 2) had continuous Medicare Parts A/B coverage from 1 year before diagnosis to end of follow-up, 3) Medicare Part D coverage from 3 mo before diagnosis to end of follow-up, 4) did not receive a TKI within 3 mo before diagnosis, and 5) received a TKI after diagnosis. To ensure adequate time to assess 1 ^st-year adherence/monitoring, we only included pts followed for ≥13 mo from TKI initiation. As suggested by guidelines, we assessed qPCR tests at 3, 6, 9, an 12 mo (all ±30 days) after TKI initiation. Optimal monitoring was defined as having a qPCR test at ≥3 milestones during the 1 ^st year of treatment. First-year TKI adherence was measured by the proportion of days covered (PDC); adherence was defined as a PDC >80%. Pearson's χ ² tests and Wilcoxon rank tests were used to compare patient characteristics in a bivariate manner, and multivariable logistic regression was used to assess factors associated with optimal monitoring and its impact on TKI adherence. All statistical tests were two-sided and conducted with SAS (Version 9.4).

Results: We identified 1188 newly-diagnosed CML pts with a median age at diagnosis of 74.5 (interquartile range [IQR]: 70-80) years. Median time from diagnosis to TKI initiation was 42 (IQR: 27-72) days with 970 pts (81.7%) initiating TKI within 90 days of diagnosis. Of 745 pts starting frontline imatinib, 17.9% switched to a 2 ^nd generation TKI; of 443 pts starting a frontline 2 ^nd generation TKI, 127 (30.9%) received a subsequent TKI with 89 (70.0%) switching to imatinib.

In the 1 ^st year after TKI initiation, 962 pts (81.0%) had a qPCR test with a median of 3 (IQR: 2-4) tests. Among pts who had ≥2 tests, the median days between two consecutive tests was 89 (IQR: 63-105). Only 876 pts (73.7%) had at least one test around the 4 milestones with 329 (27.7%) and 547 (46.0%) of pts had tests at ≥3 and 1-2 milestones, respectively. The most recently diagnosed pts were more likely to receive a test at a milestone (Figure 1a), but even among pts diagnosed in the latest study period of 2015-17, only 126 (32.0%) had optimal monitoring. Compared with less frequently monitored pts, those with optimal monitoring were more likely to be non-Hispanic white (p=0.01), diagnosed in more recent years (p<0.01), not have low-income subsidy (p<0.01), reside in a high socioeconomic (SES) neighborhood (p<0.01) and have received an influenza vaccination within 12 mo preceding CML diagnosis (p=0.01). Second generation TKI users who did not switch during the 1 ^st year had significantly more testing at the 9-mo milestone (p=0.02) and were more likely to have optimal monitoring when compared with imatinib users (p<0.01)(Figure 1b). In the multivaraible model, only year of diagnosis increased the odds of having optimal monitoring (2011-14 odds ratio [OR]=1.66, 95% confidence interval [CI]: 1.16-2.37, p=0.01; 2015-19 OR=1.78, 95% CI: 1.22-2.58, p<0.01).

The median 1 ^st-year PDC was 90.3% (IQR: 70.6-98.3%) with 795 (66.9%) adherent pts. The median PDC among pts with qPCR tests at 0, 1-2 and ≥3 milestones was 86.4%, 90.1% and 93.1%, respectively (p<0.01, Figure 2). Compared with less monitored pts, those with optimal qPCR monitoring were more adherent (73.9% vs 64.3%, p<0.01). After adjusting for demographic, SES and comorbidities, pts with optimal monitoring were more likely to be adherent (OR=1.44, 95% CI: 1.08-1.94, p=0.01).

Conclusions:

We describe "real world" CML management patterns of 1 ^st-year molecular testing and TKI adherence using a large dataset. Our analyses showed that many older pts do not have recommended molecular monitoring, which was associated with decreased TKI adherence. This work was supported by the Frederick A. DeLuca Foundation.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal