Patients with chronic myeloid leukemia (CML) who received tyrosine kinase inhibitors (TKIs) had 1.7 times the rate of arterial or venous vascular events of population-based controls in a large retrospective cohort study.
In addition, second-generation TKIs were associated with higher rates of myocardial infarction than was first-generation imatinib, Dr. Torsten Dahlén of Karolinska University Hospital Solna, Stockholm, and his associates reported. Although absolute numbers of cardiovascular events were low, physicians “should be aware of these risk factors when initiating TKI therapy in patients with CML,” the authors wrote in a study published online June 13 in the Annals of Internal Medicine .
Tyrosine kinase inhibitors have “revolutionized” the prognosis of CML and are generally well tolerated, the researchers noted. But case reports and follow-up studies of clinical trial participants have raised concerns about cardiovascular toxicities with second-generation TKIs, such as nilotinib, they added.
To further study the issue, the investigators compared 896 patients in Sweden who were diagnosed with CML between 2002 and 2012 with 4,438 age- and sex-matched controls from the national population register. By crosschecking both groups against a national patient database, the investigators calculated rates of venous thrombosis, pulmonary embolism, myocardial infarction, cerebrovascular ischemia, and other arterial thromboses (Ann. Intern. Med. 2016 Jun 13. doi: 10.7326/M15-2306).
A total of 846 CML patients (94%) received a TKI during a median of 4.2 years of follow-up, the investigators reported. First-line therapy usually consisted of imatinib (89%), followed by nilotinib (9%) and dasatinib (1%).
The TKI cohort had 78 arterial and venous events during 3,969 person-years of follow-up, compared with 250 events during 21,917 person-years of follow-up for controls, for a statistically significant incidence rate ratio (IRR) of 1.7 (95% confidence interval, 1.3-2.2). Individual IRRs for arterial and venous events also reached statistical significance at 1.5 (95% CI, 1.1-2.1) and 2.0 (95% CI, 1.2-3.3), respectively. Deep venous thrombosis and myocardial infarction accounted for most of the excess risk, with IRRs of 2.2 (95% CI, 1.1-4.4) and 1.9 (95% CI, 1.3-2.7), respectively.
When investigators looked only at the TKI cohort, they found that the rates of arterial thromboembolic events were highest for nilotinib (29 events per 1,000 person-years), followed by dasatinib (19 events per 1,000 person-years) and imatinib (13 events per 1,000 person-years). Nilotinib also was associated with a substantially higher rate of all arterial and venous events (42/1,000 person-years) than dasatinib (20/1,000 person-years) and imatinib (16/1,000 person-years).
Furthermore, nilotinib and dasatinib were associated with higher rates of myocardial infarctions (29 and 19 per 1,000 person-years, respectively) and cerebrovascular ischemic events (11 and 4 events per 1,000 person-years, respectively) than was imatinib (8 events per 1,000 person-years and 4 events per 1,000 person-years, respectively). However, the absolute numbers of events were too small to allow for statistical comparisons, the researchers said.
“The observed increase in thrombotic events may be related to CML itself, the treatment administered, or both,” they noted, but “the prevalence of myocardial infarction in patients with CML before diagnosis was similar to that of the control population, [which] might indicate a treatment-related association.”
Among the 31 patients on TKIs who had a myocardial infarction, 26 (84%) had been previously diagnosed with at least one risk factor for cardiovascular disease, including diabetes (19%), atrial fibrillation (26%), angina pectoris (39%), hypertension (55%), and hyperlipidemia (23%).
Most patients who received nilotinib or dasatinib had previously received imatinib, meaning that they could have had more advanced disease that increased their risk of adverse events, according to the researchers.
“The small number of events also leads us to exercise caution in drawing any strong conclusions,” they added. “Future data from the Swedish CML register will provide more robust evidence regarding the risks of individual drugs as exposure time increases.”
The researchers received no funding for the work. Dr. Dahlén disclosed grant support from Merck outside the submitted work. Two coinvestigators disclosed ties to Ariad, Bristol-Myers, Novartis, and Squibb.