Atherosclerotic Arterial Disease-Related Mortality in Chronic Myeloid Patients Treated with Tyrosine Kinase Inhibitors

Introduction:Vascular events is an emerging issue in patients with chronic myeloid leukemia (CML) treated with tyrosine kinase inhibitors (TKIs). As these patients achieve near-normal life expectancies on TKIs, there is clinical concern that TKI-associated vascular events may adversely affect survival. Cardiovascular disease-related mortality (CVDRM), including heart disease and stroke, is the leading cause of death in the US and disproportionately affects African-Americans (AA) whose age-adjusted CVDRM rates are 33% higher than the US population. Whether use of TKIs in AA patients with CML further increases the risk of CVDRM is not known. In this population level study, we compared the risk of CVDRM in CML patients between pre-TKI and TKI treatment eras and between ethnicities.

Methods: All 18 Surveillance Epidemiology and End Results (SEER) registries were queried for CML cases, information on ethnicity, survival and cause of death using the custom-built SEERaBomb program. Treatment periods were defined based on the year of introduction of imatinib (2001). Pre-TKI and TKI treatment eras therefore included CML cases diagnosed before and after 2002, respectively. Pharmacotherapy (PhT) receipt in SEER includes both chemotherapy and targeted anti-cancer agents and is recorded only if the initial treatment was initiated within the first 3 months of diagnosis. For this study, we assumed PhT in the TKI era to be predominantly TKIs. CVDRM included mortality from cardiac and cerebrovascular diseases. Absolute CVDRM risks were calculated based on observed deaths and number of patients at risk for years 2-10 after cancer diagnosis, because mortality in the first year is likely to be influenced by predisposing conditions. Hazard ratios (HR) for CVDRM were calculated in multivariate Cox regression analyses with backwards selection procedures. Relative risks (RR) for CVDRM, adjusted for age, sex and year, were calculated by comparing CVDRM rates in CML cohort to the general population using US background population data from the Centers of Disease Control's Wonder dataset.

Results: Of 21,997 CML cases, 10,326 were diagnosed in the pre-TKI era and 11,671 in the TKI era. 17,697 patients where white and 2,380 were of AA ethnicity. In the TKI era, 74% and 78% of white and AA CML patients received PhT, respectively. When compared to the referent background US population rates, CVDRM rates were higher in the pre-TKI era than in the TKI era (Fig 1). In the pre-TKI era, there were no differences in CVDRM risks (heart disease or cerebrovascular disease) based on receipt of PhT (Fig 1a+c). In contrast, in the TKI era, receipt of PhT was associated with significantly lower risks of CVDRM compared to no PhT in years 2-10 after diagnosis (RR for PhT vs. no PhT: 0.69 [95% CI 0.55-0.88] P= 0.002; Fig 1b+d). Absolute risk for CVDRM in years 2-10 after CML diagnosis was higher in the pre-TKI era than in the TKI era (8.1% vs. 3.8%). In the TKI era, absolute risk for CVDRM in years 2-10 was lower in those who received PhT compared to untreated patients (2.9% vs. 7.0%). In multivariable regression analyses, receipt of PhT was significantly associated with lower hazard of CVDRM, while AA ethnicity was associated with significantly higher CVDRM hazards in the TKI era but not in the pre-TKI era (Table 1). While CVDRM risks remained stable in Caucasian CML patients and was not influenced by PhT, in AA CML patients on PhT (compared to no PhT) the CVDRM risks continue to show a steady upward trend even a decade out from diagnosis (Fig 2).

Conclusion: This study is the first to shed light on differences in CVDRM risks in CML patients by ethnicity and receipt of PhT. A worrisome finding was the observed increased risk of CVDRM in AA CML treated with PhT, suggesting that these patients are particularly vulnerable to TKI-associated fatal atherosclerotic complications. This data argues for aggressive management of CVD risk factors in AA CML patients on PhT. The reason for why risks of CVDRM in CML patients treated with PhT is lower than those untreated in the TKI era is not clear. One plausible explanation is that although PhT (TKIs) increase risks of CVD (as shown in randomized clinical trials), majority of these atherosclerotic events may not be necessarily fatal. Future research should be directed towards understanding the biology behind ethnic differences in risk of CVDRM with TKI use and its potential interaction with other atherosclerotic risk factors.

View largeDownload slide

View largeDownload slide

Close modal