Risk of Selected Cardiovascular Toxicities in Patients With Cancer Treated With MEK Inhibitors: A Comparative Systematic Review and Meta-Analysis

Purpose We conducted a literature-based meta-analysis of the risk of cardiovascular toxicities associated with MEK inhibitors. Methods Eligible trials included randomized phase II and III trials of patients with cancer who were given a mitogen activated protein (MAP)/extracellular signal-regulated kinase (ERK) kinase (MEK) inhibitor (trametinib, selumetinib, or cobimetinib) and that described events of hypertension and decreased ejection fraction. Results Our search strategy yielded 300 potentially relevant citations from PubMed/MEDLINE, Google Scholar, and Cochrane Central Register of Controlled Trials. After ineligible studies were excluded, a total of 10 clinical trials were considered eligible for the meta-analysis. The relative risk for all grades of hypertension was 1.54 (95% CI, 1.02 to 2.32; P = .05), 1.85 (95% CI, 1.01 to 3.40; P = .05) for high-grade hypertension, and 4.92 (95% CI, 2.93 to 8.25; P < .001) for decreased ejection fraction. Subgroup analysis revealed no difference between trametinib and selumetinib for risk of hypertension. Conclusion Our meta-analysis demonstrated that MEK inhibitor–based treatment is associated with an increased risk of all-grade and high-grade hypertension and asymptomatic decrease in ejection fraction. Clinicians should be aware of this risk and perform regular assessment.


INTRODUCTION
The mitogen activated protein (MAP)/extracellular signal-regulated kinase (ERK) kinase (MEK) inhibitors are an emerging group of active anticancer agents that have recently been presented as reliable alternatives in the management of a variety of solid tumors, including advanced melanoma and advanced non-small-cell lung cancer (NSCLC). [1][2][3][4] Trametinib, selumetinib, and cobimetinib are the most common clinically studied agents in this group, and they are essentially tyrosine kinase inhibitors that have shown a wide spectrum of preclinical and clinical activity against many solid tumors. 5 Trametinib has been approved for the treatment of BRAF-mutant advanced melanoma. 6 In addition, both trametinib and selumetinib have been assessed in KRAS-mutant NSCLC. 7,8 Several ongoing phase II and III studies are assessing the three agents in other solid tumor indications.
The peculiar mechanism of action of MEK inhibitors has been linked to a characteristic pattern of mechanism-driven adverse events, including GI, cutaneous, cardiac, and vascular events. 6,9,10 This may be directly related to the inhibition of the mitogen-activated protein kinase (MAPK) pathway. However, there has been no systematic attempt to synthesize the data regarding cardiovascular (CV) toxicities of these agents, and the overall risk of CV toxicities induced by these agents needs to be further assessed. The range of CV toxicities assessed in our analysis includes hypertension and asymptomatic decline in ejection fraction (EF). by using "selumetinib" OR "trametinib" OR "cobimetinib" OR "MEK inhibitor" as search keywords. Our search was restricted to RCTs that recruited patients with cancer and were published in English. Trials were systemically reviewed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement. 11 Quality of the included studies was assessed by using the Jadad scale. 12

Study Selection
Clinical trials that met the following criteria were included: the study was a randomized phase II or III study in patients with cancer, participants were randomly assigned to treatment with MEK inhibitor-based therapy versus controls, and data regarding sample size and events were available for hypertension and decreased EF.
Independent reviewers (H.E. and H.A.) audited reports that incorporated the search terms initially by their titles and abstracts for relevance; later, the full texts were scanned for eligibility.

Data Extraction and Clinical End Points
Two review authors (O.A.-R. and H.E.) independently extracted the data. A checklist of data to be extracted from each study included the name of the first author, the trial phase, the type of cancer, the treatment regimens in different arms, and the number of each type of adverse event. Any discrepancies between the authors were resolved by consulting a third author. Most of the included trials used the Common Terminology Criteria for Adverse Events version 4.0 for grading the relevant adverse events.

Data Analysis
The principal measure analyzed was relative risk (RR) and its corresponding 95% CIs for all-grade (grades 1 to 4) and high-grade (grades 3 to 4) selected CV toxicities. Cochrane's Q statistic was applied to assess statistical heterogeneity in adverse events between trials, and inconsistency was quantified by using the I 2 statistic. A P value threshold of .1 was regarded as a threshold between homogeneous and heterogeneous calculation, and P Ͻ .1 was considered suggestive of heterogeneity. Whenever no substantial heterogeneity existed, the pooled estimate was determined on the basis of the fixed-effects model. Along the same lines, the pooled estimate was calculated on the basis of the random-effects model by using the DerSimonian method whenever a substantial heterogeneity prevailed. 13,14 Data analyses were performed by using Review Manager 5.3 (Nordic Cochrane Centre, Copenhagen, Denmark).

Patient Characteristics
A total of 2,704 patients were included in this analysis. Most patients had an Eastern Cooperative Oncology Group performance score between 0 and 2 and competent liver, kidney, and bone marrow function. Overall, age and sex were equally distributed in the majority of studies. Several other prognostic variables were reported in some studies and were generally comparable among randomized groups. The baseline patient characteristics and the number of adverse events in each trial are described in Table 2.

Overall Incidence of Relevant Adverse Events
All-grade hypertension ranging from 2.5% to 15% was reported in seven studies; high-grade hypertension ranging from 2% to 12% was reported in only four trametinib studies. Decreased EF ranging from 4% to 23% was reported in nine studies, whereas clinical congestive heart failure (CHF) was reported in only one trametinib study. 16

RR of All-Grade Relevant Adverse Events
A meta-analysis of the RR of all-grade adverse events was performed on the RCTs that contained direct comparison between MEK inhibitors and control treatment.  (Fig 2A, Fig 2B, and Fig 3). Because CHF was reported in only one study, RR could not be assessed for this toxicity.
Thus, patients treated with MEK inhibitors have an increased risk of all-grade and high-grade hypertension and asymptomatic decrease in EF. The fixed-effects model was used for high-grade hypertension and decreased EF, and the random-effects model was used for all-grade hypertension. Funnel plot analysis gave the impression of a publication bias (only three studies were within the left side of the inverted V shown in Figure 4 and Figure 5). We have used simple sensitivity tests to address this point.

Subgroup Analysis According to the Type of Agent Used
We conducted subgroup analyses for the risk of hypertension and decreased EF according to the type of agent used but we did not find a statistically significant difference between these subgroups for hypertension (P ϭ .39) despite a nearly doubled RR for selumetinib (2.58) versus 1.40 for trametinib. For decreased EF, a subgroup analysis could not be performed because selumetinib and cobimetinib were represented by only one study each (Fig 2A and Fig 3).

Sensitivity Analyses
The RR of all-grade and high-grade hypertension was analyzed after excluding the study by Zaman et al 18  For decreased EF, a second assessment of the RR was performed by using a random-effects model instead of a fixed-effects model to obtain more conservative results, and again, it revealed an increased RR of 3.88 (95% CI, 1.93 to 7.82; P Ͻ .001).

DISCUSSION
To the best of our knowledge, this is the most updated meta-analysis that provides an evaluation of the incidence and risk of selected CV toxicities in patients with cancer who were treated with MEK inhibitors. Our analysis of data demonstrated an increased risk of all-grade and high-grade hypertension and asymptomatic decrease in EF with MEK inhibitor-based treatment compared with controls. The three MEK inhibitors evaluated in our analysis were trametinib, selumetinib, and cobimetinib. Subgroup analysis revealed no difference between trametinib and selumetinib for risk of hypertension.  The MAPK pathway (also known as the RAS-RAF-MEK-ERK pathway) is frequently mutated in many solid malignancies. Activation of this pathway regulates a range of biologic processes including proliferation and survival. 24 CV adverse events are considered an important cause for treatment interruption or discontinuation in studies of MEK inhibitors; however, we do not currently have reliable methods to predict patients at highest risk, and thus regular monitoring of patients receiving these agents should be considered standard practice.

Favors Experimental Favors Control
CV toxicities (particularly hypertension and left ventricular dysfunction) have been reported with many other cytotoxic and targeted therapeutics used in cancer management, and likewise they have been linked to dose reduction and/or interruption. 25,26 Several pathogenetic mechanisms have been proposed to explain the development of cardiac dysfunction secondary to MAPK pathway targeting. For example, in a neonatal rat myocyte model, treatment of myocytes with sorafenib caused dose-dependent damage at therapeutically relevant concentrations. This has been ascribed to the inhibition of RAF1 and BRAF kinases by sorafenib. 27 Similar results were also found by sunitinib in a neonatal rat model (although the main mechanism of action of sunitinib is on the vascular endothelial growth factor pathway rather than on the MAPK pathway). 28 Contrary to the cardiotoxic effects of MEK inhibitors in patients with cancer discussed earlier, some preclinical studies have shown an interesting cardioprotective effect of these agents in mice with mutations in the A-type nuclear lamins (LMNA) gene, which leads to a dilated cardiomyopathy. 29 Moreover, a synergistic cardioprotective effect has been shown in combination with angiotensin IIconverting enzyme inhibitors in mice diagnosed with the same condition. 30 How to exploit these findings in the best way to ameliorate the potential cardiotoxicity of MEK inhibitors should be an active area of research, and this remains to be clarified.
Several agents have been shown to have a cardioprotective effect in the setting of anticancer agents. In particular, angiotensin-converting enzyme inhibitors and angiotensin-receptor blockers have been studied with anthracyclines and trastuzumab. 31 Whether these agents play a protective role with MEK inhibitors (especially because they are also potent antihypertensives) remains to be evaluated and clarified.
Earlier detection of cardiac toxicity secondary to anticancer drugs has been an interesting target for researchers for a long time. Several radiologic and laboratory biomarkers have been proposed as good candidates to achieve this. For example, by using a murine model of bevacizumab-and sunitinib-mediated cardiotoxicity, Bordun

Continued
can detect early changes in cardiac function before a decrease in EF can be identified. They found that although serum cardiac biomarkers were not predictive of early left ventricular systolic dysfunction, tissue velocity imaging confirmed early left ventricular systolic dysfunction 5 days before the echocardiographic documentation of decreased EF.
Despite the high risk of decrease in EF with these agents, clinical CHF was not reported in most of the studies. This may be ascribed to the fact that in the setting of controlled clinical trials, meticulous follow-up with early referral to appropriate cardiooncology services to institute proper treatments prevents deterioration into frank, clinically apparent, heart failure.
Regarding hypertension, we suggest that patients receiving MEK inhibitors should have baseline measurement of blood pressure and then regular monitoring throughout the treatment period.
A practical approach to the management of hypertension induced by these agents starts with proper grading, and we recommend the use of the threegrade system commonly used with antiangiogenic drugs. With grade I hypertension, which is defined as an asymptomatic increase to Ͼ 150/100 mmHg in a previously normal patient, no intervention is needed. With grade 2 hypertension, which is defined as a symptomatic increase to Ͼ 150/100 mmHg in a previously normal patient, initiating an antihypertensive (eg, calcium channel blockers or   angiotensin-converting-enzyme inhibitors) is indicated. Moreover, dose reduction and/or discontinuation should be considered until proper control of blood pressure is achieved. For grade 3 hypertension, the drug should be stopped, with subsequent dose reduction on control of the blood pressure, or the drug may be permanently discontinued if satisfactory control of blood pressure has not been achieved.
Other relevant forms of CV toxicities which may be underreported in clinical trials include QTc prolon-gation. Particular care and caution has to be exercised with this toxicity because of the potential of rapidly degenerating into ventricular tachycardia with potential fatal consequences. 33

Weaknesses of Our Meta-Analysis
Several limitations in our analysis have to be mentioned, most importantly, the lack of homogeneity with regard to drugs used. This has been manifested in the wide variation of some event rates (eg, hypertension was not reported in the Larkin et al 19 study compared with 15% all-grade hypertension and 12% high-grade hypertension in the Flaherty et al study). 16 Subgroup analysis has been conducted to overcome this issue and to provide deeper understanding of the differences between individual MEK inhibitors in different CV risks. Another weakness has been evidence of publication bias in the funnel plots. We have tried to do simple sensitivity tests to overcome this issue.
In conclusion, our meta-analysis has demonstrated an increased risk of all-grade and highgrade hypertension and subclinical decreased EF with MEK inhibitor-based treatment compared with control. Clinicians should be aware of this risk and perform regular follow-up for such toxicities.