Heart failure was reported in initial non-controlled studies on imatinib. In an in-vitro study, imatinib affected the mitochondrial membrane potential and cellular ultrastructure, inducing apoptosis.7 Subsequent analyses of larger populations in multicenter studies found no increase in the incidence of heart failure above that expected for the age-group and associated comorbidities.5 In a review of 1276 patients receiving imatinib, most of those who developed heart failure had previous medical conditions predisposing to the disease.8 Another effect of imatinib is fluid retention, but this has little clinical repercussion and pleural effusion is very rare. Among the recommendations for patients treated with imatinib is cardiac monitoring in cardiology consultations in patients with prior heart disease.

Dasatinib is associated with pleural or pericardial effusion in 20% of patients.9,10 A different TKI should be considered in patients at risk of pleural effusion, especially those with a history of pulmonary disease or congestive heart failure. Pulmonary hypertension is another complication found with dasatinib, observed in 0.35% of cases,10,11 and so alternative TKIs should be used in patients with pre-existing pulmonary hypertension. In rare cases (<1%) QT interval prolongation may occur in patients treated with dasatinib.

Prolonged QT has also been described as being among the side effects of nilotinib.12 However, none of the published studies report prolongation of corrected QT interval (QTc) of >500 ms, which demonstrates that nilotinib is safe in this regard.13 As pointed out above, baseline and follow-up monitoring of the ECG is recommended; care should also be taken when administering antiarrhythmics or other agents that prolong QT, and hypomagnesemia and hypokalemia must be corrected.

One of the most discussed cardiovascular toxic effects of second-generation TKIs is the increase in thrombotic arterial events. An analysis of Swedish population-based registries revealed that CML itself increases the frequency of cardiovascular events.14 The second-generation TKIs, nilotinib and dasatinib, are also associated in clinical trials with a higher cardiovascular event rate than imatinib,4 as demonstrated in one study in which the incidence of thrombotic arterial events with dasatinib was 5% compared to 2% with imatinib and was 15.9% and 10% with doses of 400 mg and 300 mg of nilotinib, respectively, compared to 2.5% with imatinib. Such events are even more frequent with ponatinib,5 with their incidence reaching 17.5%. In this context, it should be pointed out that although second-generation TKIs are more often linked to thrombotic events in the above-cited trials, associated mortality was no higher in the second-generation arm of these trials than in the imatinib arm, while mortality caused by progression of CML was higher in patients receiving imatinib than in those receiving dasatinib or nilotinib, demonstrating the importance of effective antileukemic action in these patients. Large retrospective multicenter studies identified older age and metabolic disorders such as hypercholesterolemia or diabetes as major risk factors for the development of peripheral arterial disease.15 In an analysis that included data from the IRIS, TOPS, and ENESTnd trials, 92% of patients with peripheral arterial occlusive disease had cardiovascular risk factors including diabetes, dyslipidemia, hypertension, older age and smoking.16 In the ENESTnd trial, there was at least one risk factor for atherosclerosis prior to therapy in 85% of patients with acute coronary syndrome, while another study revealed a significant association between ischemic accident and cholesterol levels and other cardiovascular risk factors.17

While the known metabolic effects of nilotinib, aggravated by the presence of hyperglycemia and hypercholesterolemia, may increase the frequency of cardiovascular events, it should be borne in mind that the population of patients with CML – whose mean age is 67 years – are already at significant risk of ischemic accidents, not only because of their age, but due to the accumulation of risk factors. There is accordingly debate whether such events are merely part of these patients’ natural history or are in fact secondary to nilotinib therapy.13 Furthermore, nilotinib may act in addition to established cardiovascular risk factors or other comorbidities by direct vascular effects, due to its interaction with discoidin domain receptor 1, or by metabolic effects on blood glucose and cholesterol levels, all of which contribute to occlusive atherosclerotic disease. It has also been suggested that pre-existing subclinical vascular disease before beginning therapy may be underestimated.

Breccia et al.18 recently performed a retrospective analysis of CML patients treated with nilotinib. Retrospectively applying the SCORE chart for cardiovascular risk estimation proposed by the European Society of Cardiology (the version for low-risk European countries including Portugal),19 they classified patients before beginning nilotinib and after follow-up as low (<1%), moderate (1-5%), high (5-10%) and very high (>10%) risk, on the basis of age, gender, cholesterol level, systolic blood pressure, smoking, diabetes, renal failure, obesity, and previous cardiovascular disease. The cumulative incidence of atherosclerotic events at 48 months was 8.5%. Higher risk categories were associated with a greater incidence of ischemic events, none of which occurred in the low-risk group. This study suggests that atherosclerotic risk factors must be detected early and throughout the treatment period in order to avoid ischemic events, and that the patient should be asked about symptoms of angina and intermittent claudication at every consultation.

The majority of studies on the effect of nilotinib (and also dasatinib and imatinib) on ventricular function excluded patients with previous ventricular dysfunction, while others did not systematically examine ventricular function by echocardiography. There is thus no information currently available to suggest that TKIs significantly affect ventricular systolic function as assessed by ejection fraction. One study found changes in diastolic function, implying that there may be more subtle changes in ventricular function.20 There is at the moment insufficient evidence to conclude that ventricular function is significantly affected, but this question should be explored in further studies, which should include more sensitive indicators of ventricular function.

Most studies on bosutinib have shown no increase in the incidence of cardiovascular events.21,22 However, bosutinib inhibits signaling pathways that are also inhibited by ponatinib, such as the Src kinase family, and so until the etiopathogenic mechanisms of cardiovascular toxicity in TKIs are completely elucidated, patients taking bosutinib should be monitored for possible cardiovascular events.23 Likewise, most studies have not found that bosutinib prolongs QTc. In a study of 288 patients with CML, QTc prolongation was found in only two patients (>500 ms in one and >60 ms in the other),23 suggesting that the drug is safe but underlining the need for electrocardiographic assessment and monitoring.

Although few studies on ponatinib have been published, it appears to be associated with a high risk for cardiovascular events, particularly arterial (occlusive arterial disease in 20% of cases) but also venous.24 In additions, hypertension is a frequently cited side effect of ponatinib and therefore measures must be taken to control blood pressure.25