Cardiotoxicity Monitoring in Patients Treated with Tyrosine Kinase Inhibitors

Tyrosine kinase inhibitors (TKI) can cause cardiotoxicity. This brief communication describes two cases of TKI‐induced heart failure and suggests monitoring exercise tolerance with a simple daily activity at home may be more useful than intermittent measurement of laboratory tests such as LVEF.

Abstract

Tyrosine kinase inhibitors (TKIs) can cause cardiotoxicity, and some suggest routine monitoring of cardiac function during TKI use. We describe two cases of TKI‐induced heart failure (HF) that suggest the utility of monitoring with laboratory tests is questionable. One patient developed HF 5 days after starting pazopanib. The other developed HF while receiving 25 mg per day sunitinib, and had previously received a higher dose (50 mg per day) with no symptoms of cardiotoxicy. In addition, she later received 5 cycles of sunitinib (25 mg per day) without developing an abnormal left ventricular ejection fraction (LVEF) value by echocardiography or cardiac symptoms. Although the LVEF is commonly performed to monitor TKI cardiotoxicity, evidence for its predictive utility is limited. These cases raise questions regarding the practical utility of sequential measurement of LVEF in adults treated with TKIs. We suggest a simple daily activity such as stair climbing to monitor exercise tolerance.

Introduction

Tyrosine kinase inhibitors (TKIs) can cause serious cardiotoxicity [1], [2]. Left ventricular ejection fraction (LVEF) has been found to progressively fall in groups of patients treated with free doxorubicin, presumably reflecting a continuum of cardiac damage [3]. Although the LVEF is commonly performed to monitor TKI cardiotoxicity, evidence for its predictive utility is limited [4], [5]. Here we describe two cases of TKI‐induced heart failure (HF) that raise further questions about the utility of routine monitoring of cardiotoxicity of TKIs by LVEF or other laboratory methods in the absence of studies demonstrating clear utility of such measures. One case also provides additional evidence that HF does not necessarily preclude from using sunitinib again after adequate recovery if indicated.

Case 1

A 67‐year‐old woman with a recurrent leiomyosarcoma had an echocardiogram (ECHO) showing a LVEF of 60% with normal cardiac function. She began doxorubicin and olaratumab but progressed 6 months later. Tumor progressed on trabectedin and pazopanib was begun; she was admitted 5 days later for shortness of breath and found to have nonischemic cardiomyopathy and HF with a traced LVEF of 12.9%. Pazopanib was stopped and she was treated medically. Five days later, she felt better, and a magnetic resonance imaging (MRI) stress test showed a LVEF of 21% but no valve disease, infarction, fibrosis, or ischemia. Eight days after stopping pazopanib, she had no dyspnea on exertion; 4 days later she began treatment with ifosfamide with no recurrence of cardiac symptoms; five months later, LVEF was 47% by ECHO, and a month later LVEF was 51% with normal systolic function on cardiac MRI.

Case 2

A 58‐year‐old woman with metastatic renal cell carcinoma presented with the syndrome of inappropriate antidiuretic hormone (SIADH). Treatment with demeclocycline and sunitinib (50 mg p.o. per day) was begun. Sunitinib was stopped after 3 weeks because of thrombocytopenia, anemia, and fatigue; her demeclocycline was stopped, hydrochlorothiazide was changed to triamterene 50 mg day, and baseline lisinopril and atenolol for hypertension were continued. By week 4, platelets had risen to 70,000 × 10⁶ per L and SIADH resolved. A second cycle of sunitinib (25 mg per day) was started at week 6. Dyspnea on exertion gradually developed, and after 18 days of treatment she was admitted with shortness of breath and HF. An ECHO showed LV hypokinesis, mitral, tricuspid, and aortic regurgitation and an LVEF of 35% (Fig. 1). Sunitinib was discontinued, and furosemide was transiently added. She recovered rapidly. Seven weeks after starting cycle 2, an ECHO showed normal LV systolic function and EF, and a week later a dobutamine stress ECHO was normal. Lisinopril was increased to 40 mg per day. She later received four cycles at the same dose of sunitinib using a 2 week on 2 week off schedule without developing an abnormal LVEF or cardiac symptoms, however, her disease progressed; throughout this period she was able to climb a floor of stairs without dyspnea.

Figure 1.

Timeline of sunitinib treatment and EF of case 2. Sunitinib (50 mg p.o. per day) was begun and stopped after 3 weeks because of thrombocytopenia (platelets 27,000 × 10⁶ per L), anemia, and fatigue. By week 4, platelets had risen to 70,000 × 10⁶ per L and Na levels to 130 mmol/L. At week 6, a second cycle of sunitinib (25 mg per day) was started. Dyspnea on exertion gradually developed, and after 18 days of treatment she was admitted with shortness of breath and found to have HF. An echocardiogram (ECHO) showed left ventricular (LV) hypokinesis, mitral, tricuspid, and aortic regurgitation and an left ventricular ejection fraction (LVEF) of 35%. Sunitinib was discontinued. She recovered rapidly. At 7 weeks into cycle 2 she felt fine; an ECHO showed normal LV systolic function and EF with moderate mitral and tricuspid regurgitation and pulmonary hypertension. At week 8 of cycle 2, a dobutamine stress ECHO was normal with normal LV function, no valve disease, and an LVEF of 55%–60%. Imaging revealed stable disease, and there was no evidence of syndrome of inappropriate antidiuretic hormone (SIADH). One week later, sunitinib was begun at 25 mg per day for 2 weeks on and 2 weeks off, and lisinopril was increased to 40 mg per day. Fourteen days after starting sunitinib, she climbed stairs easily; an ECHO showed normal LV function with an EF of 55%. Two weeks later (28 days into cycle 3) another cycle of sunitinib (25 mg per day) was begun. Two weeks later she was going up stairs easily; an ECHO showed an EF of 55% with normal LV function. Sunitinib was stopped. One week later, a 2‐week course of sunitinib (25 mg per day) was started with the intention of pausing treatment for a 1‐week interval before beginning the next cycle. Eleven days later an ECHO showed mild diffuse LV hypokinesis with an EF of 50%. She continued to feel well. The 2‐week course of sunitinib was completed. A week after stopping sunitinib she felt well and began another cycle of sunitinib (25 mg per day). Two weeks later an ECHO showed a low normal EF of 50% with no wall motion abnormality. Sunitinib was stopped after 2 weeks. One week later she began another 2‐week course of sunitinib (25 mg per day). Na was 128 mmol/L; an ECHO showed an LVEF of 50%–55%. Two weeks later she was active but bothered by dysphagia. An ECHO showed an LVEF of 50%; imaging showed disease progression with recurrence of SIADH.

Abbreviations: EF, ejection fraction; HF, heart failure; mg/d, mg per day.

Discussion

TKIs, like some other anticancer drugs, can induce cardiac damage. In one report, 6 of 75 patients treated with sunitinib had HF; sunitinib was restarted in 5 patients without subsequent HF, although 4 had episodic reductions in EF [1]. In another report, HF resulting in serious morbidity and/or death was seen in 6 of 224 patients receiving sunitinib and occurred between 4–44 days after starting drug; HF was not reversible in all cases [2]. A meta‐analysis found that ∼4.1% of 6,935 patients developed HF on treatment with sunitinib compared with 1.8% of those on placebo [6]. A retrospective analysis of phase III trials found decreases in LVEF more common with sunitinib than the comparator arm, but HF did not differ [7]; notably, many patients with a decrease in LVEF continued sunitinib [7]. In a worldwide treatment‐use trial of sunitinib in GIST, HF, myocardial infarction, or reduced LVEF occurred in ≤1% each of 1,124 patients [8].

The difficulty of predicting long term effects of cardiotoxicity of drugs may lead to a failure to prevent adverse effects or inappropriate treatment, including limiting use of a life‐saving drug. Several guidelines recommend “close” or “regular” monitoring for the development of cardiotoxicity with some chemotherapy drugs [9], [10], [11] including monitoring by determining LVEF “periodically” [10]. However, it is well recognized that specific evidence‐based recommendations are lacking.

Because the development of cardiotoxicity is difficult to predict, some suggest intermittent monitoring of cardiac function in patients on TKIs; the LVEF is commonly used for this purpose. The LVEF was originally determined by cardiac catheterization, and later by MUGA; ECHO has gradually become the preferred approach because of the ease of performing the test and the absence of radiation exposure. Studies have reported that an early decline in LVEF can predict cardiotoxicity in adults receiving free doxorubicin [12]. However, the utility of monitoring serial LVEFs in patients on TKIs has not been clearly demonstrated. The utility of EF to detect cardiotoxicity depends on the accuracy of changes in EF, the ability of EF changes to predict cardiotoxicity, and the time between initiation of drug and the development of relevant cardiotoxicity. There is substantial variation in LVEF measurement [4], [5], [7], [13]. Patients may also have HF with a preserved LVEF. Nevertheless, EF is widely used to monitor cardiotoxicity in patients receiving anticancer chemotherapy.

In the cases presented here, the sudden onset of HF after only 5 days of treatment in one case, and 18 days into a dose half that of the previous cycle in which no cardiac symptoms were noted in another case, coupled with the ability to subsequently give further drug at the same dose without development of HF, casts some doubt on the utility of routine monitoring of cardiac function in asymptomatic patients receiving TKIs, other than history and physical examination. Other reports also note the onset of severe HF within days of starting a TKI [2], emphasizing that cardiotoxicity can develop very soon after starting a TKI. Thus, even if the LVEF were an accurate predictor of cardiotoxicity, given the short time interval during which toxicity can develop, it would be necessary to monitor LVEF quite frequently.

In conclusion, HF can develop rapidly after starting a TKI, and the development of HF related to a TKI does not necessarily mean the drug cannot be used again if the risk/benefit seems appropriate. Nevertheless, one must be aware of the potential serious adverse cardiac events with TKIs; close attention to history, physical examination, and BP is important. These cases also question the utility of monitoring cardiotoxicity with sequential measurement of LVEF in adults treated with TKIs, as the time to development of cardiotoxicity differs greatly from that of free doxorubicin [3], [12]. The low incidence of HF in patients treated with sunitinib [6], [7], [8], coupled with the variability in the measurement of EF, further complicate the utility of monitoring cardiotoxicity in patients treated with sunitinib by EF, and monitoring EF adds significant costs to treatment (reviewed in [5]). In addition, given the prognosis of many patients receiving TKIs, erroneous interpretation of EF changes might inappropriately limit their access to a useful drug or lead to initiation of additional inappropriate treatment.

Thus, there are questions about the efficacy of “cardiac monitoring” during TKI use by methods other than history and physical examination. Indeed, the recent ASCO guidelines emphasize three important considerations: (a) careful history and physical examination should be done to detect symptoms of cardiotoxicity of anticancer drugs, (b) the optimal screening method and interval are unclear, and (c) no recommendations can be made regarding changes in cancer treatment based on screening results [14], [15]. It is well known that changes in exercise capacity and resting heart rate may provide some evidence of cardiac function [16]. Perhaps a more useful and inexpensive monitoring system for TKIs would be performing a daily activity such as stair climbing or other simple exercise at home to detect changes in exercise tolerance.

There is a clear need for trials that define the benefit of a specific monitoring regimen [9]. Given the reversibility of cardiotoxicity in some cases, a trial demonstrating that treatment of specific asymptomatic changes in measurements of cardiac function is beneficial is also needed. Unfortunately, the results may depend on the drug in question. Future randomized trials should examine whether monitoring LVEF or another variable can reliably predict clinically significant drug‐induced cardiotoxicity, and whether the medical treatment of asymptomatic changes in LVEF or other measure of cardiac function affects later functional outcome. A useful trial would directly compare a daily activity such as stair climbing with at least one alternative approach to truly determine the relative cost and benefit of at least those two methods. The addition of a daily activity measurement would add little cost to a trial and could easily be incorporated into trials of new TKIs and other drugs.

Disclosures

The author indicated no financial relationships.