Purpose: The purpose of our study was to review and determine the cardiovascular safety profile of combretastatin A4 phosphate (CA4P) in a Phase I study in 25 patients with advanced solid tumors.
Experimental Design: CA4P was administered in a dose-escalating fashion starting at 18 mg/m2 i.v. every 21 days, and the maximal dosage was 90 mg/m2. Continuous evaluation included bedside blood pressure and pulse monitoring, 12-lead electrocardiogram (ECG) at fixed time points for measured QT interval determination, determination of the corrected QT interval (QTc) using Bazett’s formula QTc = QT/(R-R interval)1/2, and chart review. Pharmacodynamic correlations of CA4P dose, CA4P/CA4 area under the curve, and Cmax versus heart rate (HR), blood pressure, QT, and QTc intervals, over the first 4 h postdosing were analyzed.
Results: After CA4P administration, there were significant increases in QTc interval at the 3-h and 4-h time points [27.2 ms (P < 0.0001) and 30.8 ms (P < 0.0001), respectively] and HR at the 3- and 4-h time points [13.2 beats per minute (bpm; P < 0.01) and 15.1 bpm (P < 0.001), respectively]. Three of 25 patients had prolonged QTc intervals at baseline, whereas 15 (60%) of 25 and 18 (75%) of 24 patients had prolonged QTc intervals at 3 and 4 h. The slope of HR and QTc increasing as a function of time during the first 4 h was correlated to dose (in milligrams) of CA4P (P = 0.01 and r = 0.49 for HR, P = 0.005 and r = 0.55 for QTc) and to CA4 area under the curve (P = 0.04 and r = 0.41 for HR, P = 0.02 and r = 0.44 for QTc); blood pressure and uncorrected QTc interval dose-response correlations were not significant. Two patients had ECG changes consistent with an acute coronary syndrome within 24 h of CA4P infusion.
Conclusions: CA4P prolongs the QTc interval. There was a temporal relationship with the CA4P infusion and with ECG changes consistent with an acute coronary syndrome in two patients. It is advisable that future trials with CA4P have eligibility guidelines limiting patients with known coronary artery disease or those with multiple coronary artery disease risk factors until more experience is gained regarding potential cardiovascular toxicity with this agent.
Combretastatin A4 phosphate (CA4P) is a novel, natural product tubulin inhibitor that was isolated from the South African tree Combretum caffrum (1) . Preclinical studies have demonstrated selective in vitro activity of CA4P against proliferating endothelial cells as well as in vivo effects yielding shutdown of tumor vasculature with resultant central tumor necrosis (2 , 3) . Our group reported the first Phase I trial of this agent to determine the maximum-tolerated dose (MTD), safety and pharmacokinetic profile, its effect on tumor blood flow using dynamic contrast enhanced magnetic resonance imaging (DCE-MRI), effect on levels of cell adhesion molecules, and preliminary data on clinical activity of CA4P administered on a single i.v. dose schedule at 3-week intervals (4) .
The clinical toxicity we encountered over the course of this study was consistent with an agent that is “vascularly active” and was devoid of traditional cytotoxic side effects, namely myelosuppression, stomatitis, and alopecia. Vascular side effects included a variable symptom complex of hot flash/flushing, headache, and abdominal cramping pain over all dose levels; infusion-related tumor pain; hemodynamic changes in blood pressure and heart rate (HR); electrocardiographic changes consisting of prolongation of QTc interval; and two episodes of acute coronary syndrome that resolved without sequelae.
We also observed clinical activity with this agent, including a patient with a pathological complete remission who survives more than 3 years disease free from his last dose of CA4P (4) . Cardiovascular side effects, including QTc interval prolongation, have been encountered and present a challenge in the clinical development of antineoplastic agents (5 , 6) . This report focuses on our independent evaluation of the cardiovascular side-effect profile that we encountered and on determining whether CA4P had an effect on the QTc interval.
PATIENTS AND METHODS
Patients with advanced solid tumors for whom no standard effective therapy was available were entered into the protocol between November 1998 and May 2000. CA4P was administered in a dose-escalating fashion starting at 18 mg/m2 i.v. every 21 days and the maximal dosage was 90 mg/m2. Patient selection, consent, inclusion and exclusion criteria; CA4P dose escalation and administration; clinical monitoring for toxicity and response; and pharmacokinetic/pharmacodynamic methods have been previously reported (4) . Study eligibility criteria required a normal 12-lead electrocardiogram (ECG), reviewed by a cardiologist within 2 weeks of entry.
ECGs were performed hourly during the first 4 h after CA4P dosing and repeated 24 h after dosing. An ECG was not available for review in one patient at the 1-h time point, in one patient at the 2-h time point, in one patient at the 4-h time point, and in one patient at the 24-h time point. These patients were included in the QT analysis. ECGs were available at the 3-h time point for all patients. Hourly 12-lead ECGs beyond 4 h and continuous single-lead ECG telemetry monitoring were performed for patients with QTc interval prolongation of ≥480 ms. [grade 1 electrophysiological cardiac toxicity by definition, per revised Common Toxicity Criteria (7)] and/or a QTc interval increase of ≥25% from baseline demonstrated within the first 4 h after dosing. Patients were transferred to a telemetry unit if they had a QTc prolongation ≥500 ms. QT intervals were manually measured by one author of this study (B. S. S.) and were corrected for HR using Bazett’s formula QTc = QT/(R-R interval)1/2. A QTc interval of 440 ms was defined as the upper limit of normal value (i.e., above the 95% upper limit of normal in men and women in the Framingham Heart Study; Ref. 8 ). QTc values >440 ms were considered prolonged. QTc intervals >500 ms, a >25% increase in the QTc from baseline, and ventricular arrhythmias noted on the 12-lead ECG were considered clinically significant. The protocol was amended to enhance cardiac monitoring for the final three patients entered onto the study and the patient remaining under active treatment. These patients underwent three ECGs on the day of treatment to obtain an average baseline QTc interval before dosing as well as continuous bedside and Holter monitoring.
We tested the significance of correlation between five measures of CA4P dose, namely, the dose in mg administered per cycle, the CA4P area under the curve and Cmax values, and the CA4 area under the curve and Cmax values, and five measures of cardiac response, namely, the slopes of HR, QT, and QTc as linear functions of time fitted over the first 4 h, and the curvatures of the systolic and diastolic blood pressures fitted to quadratic functions of time over the first 4 h, Pearson’s correlation test was used uncorrected for multiple comparisons because the dose measures were all positively correlated. Two-sided exact binomial tests for the signs of the QT, HR, and QTc slopes, and the signs of the systolic and diastolic blood pressure curvatures, were used to test time course trends, assuming a null hypothesis that the signs of these estimated parameters are equally likely to be positive or negative. Paired t tests were used between the 0 (baseline) and 3- and 4-h time points.
Twenty-five patients (12 men and 13 women) with a median age of 55 years (range, 31–77 years) were enrolled on this study as reported previously (4) .
QT Interval and Pharmacodynamics.
The QT interval, the HR, and the corrected QT interval (using Bazett’s formula) for all patients in this study are shown in Table 1⇓ .
The mean QT interval was 366.4 ms at baseline among all evaluable patients (n = 25). At h 1, 2, 3, 4, and 24 after CA4P administration, the mean QT was 384.0, 376.0, 362.6, 361.0, and 373.1 ms, respectively. Fifteen of 25 QT slopes over the first 4 h were negative and 10 of 25 were positive, suggesting no significant trend over time (P = 0.42).
The mean QTc interval was 421.8 ms at baseline among all evaluable patients (n = 25). At h 1, 2, 3, 4, and 24 after CA4P administration, the mean QTc intervals were 420.4, 442.4, 449.0, 452.8, and 425.2 ms, respectively. The increases in mean QTc interval at the 3- and 4-h time points after CA4P administration (mean changes, 27.2 and 30.8 ms, respectively) were statistically significant (P < 0.001; paired t test), and corresponded to a 7–8% increase in mean QTc over baseline. At baseline, 3 (12%) of 25 patients had a QTc interval >440 ms compared with 7 (29%) of 24 patients at 1 h, 14 (58%) of 24 patients at 2 h, 15 (60%) of 25 patients at 3 h, 18 (75%) of 24 patients at 4 h, and 5 (21%) of 24 patients at 24 h after CA4P. Twenty-two of 25 QTc slopes fitted over the first four time points were positive (P = 0.0002), further supporting an increase in QTc in response to treatment.
No patient had a QT or a QTc interval >500 ms or had a percentage increase in QTc interval over baseline that was >25% at any of the prespecified time points; the longest QTc measured in any patient was 498 ms, and the largest percentage increase in QTc over baseline was 21.6%. The changes in mean QTc at the 4-h time points were 3.4 ± 14.3 ms (18 mg/m2, n = 3), 12.2 ± 20.8 ms (36 mg/m2, n = 3), 34.9 ± 14.9 ms (60 mg/m2 over 10 min, n = 9), 44.8 ± 27.5 ms (60 mg/m2 over 1 h, n = 7), and 37.4 ± 15.3 ms (90 mg/m2, n = 3) There was a single, isolated QTc interval prolongation of 559 ms observed on the 12-lead ECG that occurred beyond the specified 24-h time point at 44 h after the CA4P infusion.
The mean HR was 80.8 bpm at baseline among all evaluable patients (n = 25). At hours 1, 2, 3, 4, and 24 after CA4P administration, the mean HRs were 73.6, 84.7, 94.1, 95.7, and 80.0 bpm, respectively. The increases over baseline in mean HRs at the 3- and 4-h time points after CA4P administration (mean changes, 13.2 and 15.1 bpm, respectively) were statistically significant using the paired t test (P < 0.001). As with QTc, 22 of 25 HR slopes fitted over the first 4 h were positive (P = 0.0002). Of 18 patients that had more than three blood pressure measurements over the first 4 h (data not shown), 18 of 18 had negative systolic blood pressure time course curvatures (P < 0.0001), and 17 of 18 had negative diastolic blood pressure curvatures (P = 0.0001), i.e., the time courses were almost all concave down (slopes decreasing) over time. Paired t tests between baseline measurements and the mean of the 3- and 4-h time point measurements were not statistically significant, however, suggesting that any initial increases in blood pressure had returned to baseline within 4 h.
Pharmacodynamic correlations of CA4P dose, CA4P/CA4 area under the curve and Cmax versus HR, blood pressure, QT, and QTc were analyzed across patients. The slopes of the HRs and QTc intervals during the first 4 h were correlated to dose (in mg) of CA4P (P = 0.01 and r = 0.49 for HR, P = 0.005 and r = 0.55 for QTc) and to CA4 area under the curve (P = 0.04 and r = 0.41 for HR, P = 0.02 and r = 0.44 for QTc). No other dose-response correlations tested were statistically significant.
No ventricular arrhythmias were noted on any of the 25 ECG tracings, including the two patients who had ECG changes consistent with acute coronary events.
Acute Coronary Syndromes.
Two of the 25 patients had ECG changes consistent with an acute coronary syndrome within 24 h of the CA4P infusion. The statistical conclusions described above are independent of the inclusion or exclusion of these two patients (patients 11 and 25) with acute coronary syndromes.
Patient 11 was a 57-year-old white male with regionally advanced pancreatic cancer. The patient’s only known coronary artery disease risk factor was distant tobacco use. Three weeks after cycle 1, he received cycle 2 of combretastatin at 60 mg/m2. Ninety-five min after starting cycle 2 of CA4P the patient complained of “squeezing” chest pain that lasted ∼2 h. The ECG revealed acute ST elevations in leads II, III, AVF, V4, V5, and V6, and ST depression in leads V1, V2, and V3 (Fig. 1)⇓ . An urgent cardiac catheterization was significant for a subtotal stenosis of the distal left anterior descending artery with left-to-left collaterals. Serial cardiac enzymes including troponin were normal. An ECG performed the same day as the CA4P infusion and after cardiac catheterization revealed normal myocardial wall motion and left ventricular ejection fraction. The patient completely recovered without further cardiac sequelae and was not treated further.
Patient 25 was a 77-year-old white male with a diagnosis of anaplastic thyroid cancer with cardiac risk factors of hypertension, hyperlipidemia, and tobacco abuse. The pretreatment ECG was normal, and there was no antecedent history of angina. Twenty-one h after the CA4P infusion, the patient’s ECG demonstrated changes of infero-lateral ST segment depression followed by deep and symmetric T wave inversion across the anterior precordial leads (Fig. 2)⇓ . Cardiac enzymes revealed a peak level of troponin I to 3.77 ng/ml with an upper limit of normal of <0.50 ng/ml. The creatinine kinase and myocardial subfraction levels were normal.
Two days after cycle 1 of CA4P, a persantine nuclear stress test revealed an area of irreversible decrease in myocardial perfusion in the antero-apical portion of the left ventricle, suggestive of myocardial infarction. The left ventricular ejection fraction, previously normal at 60%, was moderately reduced to 42%.
Subsequently cardiac catheterization revealed a 99% stenosis in the proximal portion of the left anterior descending artery and 70–80% narrowing in the right coronary artery. The patient underwent intracoronary angioplasty and stent deployment in both coronary arteries with no residual stenoses detected. The patient was taken off study and discharged to home in good condition 8 days after receiving CA4P. The patient had no further cardiac sequelae for the next 11 months after removal from the study.
The CA4P analogue, combretastatin B1 (CB1) is extracted from the South African plant Combretum krausii and contains the hiccup nut toxin (9) . CB1 is a K+ channel blocker and accordingly delays the action potential by blocking repolarizing K+ currents and reversibly increases the duration of the action potential in rat sensory neurons by 300% (9) . K+ delayed rectifier currents (Ikr) are inhibited by CB1 in neurons and human myotubules; the HERG-type inward rectifier K+ channels in tumor cells are inhibited to a greater degree. The HERG channels are present in the heart, and the inhibitory action of CB1 on HERG is similar to the blockade produced by several class III antiarrhythmic drugs. The effects of CB1 are selective and reversible.
There is the potential for cross-reactivity in pharmacological action between CA4P and CB1. It is possible that CA4P may prolong repolarization in cardiac tissues similarly to CB1. Prolongation of cardiac repolarization has the potential to cause potentially serious ventricular arrhythmias, in particular torsades de pointes. QTc prolongation is an important issue in drug development.
The data from the ECGs of the 25 evaluable patients demonstrate statistically significant increases in QTc interval and HR after CA4P administration. None of the patients in this study, however, were judged to have “clinically significant” increases in QTc interval, defined as any QTc interval >500 ms, a >25% increase in QTc over baseline, or ventricular arrhythmias observed on the 12-lead ECGs. The effects on QTc interval and HR returned to baseline values by at least the 24-h time point.
Because this study did not include a control or placebo group and it was nonrandomized, open-label, and nonblinded, no definitive conclusions can be drawn from this study on the effect of CA4P on QTc interval. The data also must be viewed with caution because of the small numbers of patients studied at a single institution. Finally, there are recognized limitations related to the ECG measurements including the use of manual QT measurements, the inability on some ECG tracings to accurately define the end of the QT interval at higher HRs, and most importantly, the use of the Bazett formula for correction of QTc interval for changes in HR, particularly at higher HRs as was seen in this study.
Despite these and other limitations, the data do suggest that CA4P, at the doses administered in this study, may produce a modest prolongation of QTc interval. Furthermore, the increase in HR that we observed in this study may fortuitously afford protection against proarrhythmia associated with prolongation of ventricular repolarization.
Two patients had ECG evidence of acute coronary syndromes. Patient 11 had chest pain and transient ECG changes less than 2 h after CA4P infusion. Although there was no laboratory evidence of cardiac ischemia, there was ECG evidence of an acute coronary syndrome and catheterization revealed single distal vessel coronary artery disease. Patient 25 had multiple risk factors for coronary artery disease and had laboratory evidence of myocardial ischemia 21 h after receiving CA4P. The cardiac catheterization demonstrated severe two-vessel coronary artery disease. It is plausible that this episode may have been precipitated by the marked increase in blood pressure after the administration of CA4P. Given the temporal relationships between the administration of CA4P and the development of acute myocardial ischemia, it is likely that there exists a causal relationship between drug administration and coronary artery effects.
In summary, CA4P may affect action potential repolarization and K+ currents in the ventricular myocardium, possibly in a similar manner to combretastatin B1. In light of this, observation and ECG monitoring of future patients treated with this drug is warranted. Because there has been some demonstrable clinical efficacy in early Phase I trials of this agent (4 , 10 , 11) , we think that the cardiovascular safety profile observed in this study should not preclude further clinical development of CA4P. Safety monitoring for patients treated with this drug should include avoidance of other drugs or agents at the time of administration of CA4P that may prolong the QTc interval (see www.torsades.org); avoidance of hypokalemia (K+ <4.0 mEq/liter) or hypomagnesemia, (Mg <1.8); and monitoring of patients to further characterize the electrocardiographic effects of this agent with access to serial ECG recordings and/or continuous monitoring. Carefully designed studies, with eligibility criteria that exclude patients with significant preexisting cardiovascular disease, should provide additional insight into the safety profile of CA4P.
Grant support: Supported in part by a clinical research grant from OXiGENE, Inc., Waltham, MA and NIH Grant M01 RR-00080.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Notes: Presented in part at the American Society of Clinical Oncology’s Molecular Therapeutics Symposium, Coronado, California, November 7–10, 2002.
Requests for reprints: Scot C. Remick, M.D., University Hospitals of Cleveland, 11100 Euclid Avenue, Cleveland, OH 44106.
- Received March 9, 2003.
- Revision received August 25, 2003.
- Accepted August 28, 2003.