Pb1932 predictive powers of various cardiovascular examinations for cardiovascular occlusion events in cml patients receiving tyrosine kinase inhibitors

Background: The prognosis of chronic myelogenous leukemia (CML) has been dramatically improved by tyrosine kinase inhibitors (TKI). Therefore, it has become very important to control off‐target adverse effects accompanying the long‐term administration of TKI. Among them, cardiovascular occlusion events (CVE) are becoming a problem. Although cardiovascular examinations such as electrocardiography and measurement of ankle‐brachial blood pressure index (ABI) are recommended, the clinical usefulness and predictive value of CVE have been not sufficiently evaluated. Furthermore, few reports have evaluated cardiac ankle vascular index (CAVI) and cervical echo. Aims: Here, we comprehensively investigated the usefulness of various cardiovascular evaluation tests for monitoring CVE and the predictive value of CVE in TKI treatment and performed a risk analysis of CVE. Methods: We performed cardiovascular evaluation tests including measurement of ABI and CAVI, cervical echo, and electrocardiography in CML patients receiving TKI treatment (imatinib, nilotinib, dasatinib, bosutinib, or ponatinib) between April 2017 and December 2018 in our institute. Results: A total of 74 patients aged 21–91 years (median age, 63.5 years) were evaluated. Comparison with TKI given at the time of the examination revealed no significant difference in frequency of abnormalities of cervical echo, CAVI abnormality rate, or QTc prolongation depending on TKI type. On the other hand, 11 of 22 (50.0%) patients who were taking nilotinib had an ABI < 1.0, which was significantly more prevalent than for other TKIs. Ten patients developed CVE during TKI treatment. The TKIs that were taken when CVE occurred were nilotinib in 6 patients, imatinib in 2, and bosutinib in 2. Comparisons of the CVE and non‐CVE groups revealed a significant difference in age, and the proportion of renal dysfunction cases and dyslipidemia tended to be higher in the CVE group. The proportion of patients with an ABI < 1.0 in the CVE group was significantly higher (50% vs. 18.8%, p = 0.044). However, the proportion of patients with an ABI < 1.0 and those with an ABI ≤0.9 were only 50% and 30% in the CVE group, respectively. The QTc value was significantly prolonged (median 440 ms vs. 425 ms; p = 0.028). CAVI values tended to be high (median 8.5 vs. 8.0; p = 0.13). On the other hand, various parameters of cervical echo and the presence or absence of plaque showed no significant difference. To evaluate the predictive power of CVE for each examination, the sensitivity and specificity of CVE for each were evaluated. Regarding echo plaque, the area under the receiver operating characteristic curve was 0.5, suggesting absolutely no predictive power. The results of the remaining indicators are summarized in the Table. Using any index alone or a composite model consisting of ABI/CAVI, the predictive power of CVE was insufficient (Table). Summary/Conclusion: Although we found that measuring ABI was a better method for monitoring and predicting CVE, the sensitivity and positive predictive value of ABI were insufficient, probably because it is an indicator of lower‐limb artery occlusion and stenosis, and thus, does not reflect systemic vascular damage. image As monitoring of CVE associated with TKI treatment might be inadequate using existing routine methods and CVE can be a risk of sudden death, it is necessary to establish a more useful monitoring and/or predictive method for the early detection of CVE in CML patients under TKI treatment.