Effects of oral arsenic trioxide therapy on QT intervals in patients with acute promyelocytic leukemia: implications for long-term cardiac safety

Arsenic trioxide (As₂O₃) is efficacious in acute promyelocytic leukemia (APL).^1-4  It blocks potassium currents I_Kr and I_Ks,⁵  causing QT prolongation. Ventricular tachyarrhythmias were reported in about 30% of patients treated with intravenous As₂O₃^6-9 .

As₂O₃ may be effective in other neoplasms,¹⁰  making clinical trials of As₂O₃ in these diseases pressing. However, reliance on intravenous As₂O₃ hampers these efforts. Long-term intravenous As₂O₃ is resource-demanding. Moreover, a potentially fatal cardiac toxicity is worrisome, especially in clinical trials.

We have developed an oral formulation of As₂O₃^4,11 . Details of the preparation and pharmacokinetics of oral As₂O₃ have been previously reported.¹¹  Oral As₂O₃ gives a similar bioavailability, but lower peak plasma arsenic concentrations, as compared with intravenous As₂O₃¹¹ . Oral As₂O₃ represents an advance in arsenic therapy. Patients take oral As₂O₃ at home, rendering long-term therapy feasible, thus facilitating clinical trials. Most importantly, in more than 600 patient weeks of oral As₂O₃ administered during 5 years, no ventricular tachyarrhythmias were observed.^4,11 

Prolongation of QT or corrected QT (QTc) increases the risks of ventricular tachyarrhythmias.^12,13  QT-interval dispersion measures regional nonhomogeneities of ventricular repolarization.¹⁴  Greater dispersions increase ventricular arrhythmias.¹⁵  Heart rate variability (HRV) measures the beat-to-beat heart rate variations, correlating with changes in autonomic tone. HRV changes increase cardiac arrhythmias.¹⁶ 

To explain the favorable cardiac side-effect profile of oral As₂O₃, we studied a cohort of patients on long-term oral As₂O₃,to determine the cardiac safety and changes of QT intervals and HVR.

We studied 17 consecutive patients with relapsed APL (Table 1). All had normal left ventricular ejection fraction and blood biochemistry.

Patients received oral As₂O₃ (10 mg/d) for 2 weeks, followed by a drug-free period of 6 to 8 weeks before the next course. Patients took oral As₂O₃ at 14:00 every day to match the timing for electrocardiography (ECG) and Holter recording, and blood arsenic assay. Treatment protocol was approved by the institutional review board of Queen Mary Hospital, and all patients gave informed consent in accordance with the Declaration of Helsinki.

Data were collected at day 10 of oral As₂O₃ (As₂O₃-ON) and 4 weeks after stopping oral As₂O₃ (As₂O₃-OFF). Resting 12-lead surface ECGs were recorded (paper-speed: 50 mm/second) for measuring QT intervals. QT intervals at each lead and the corresponding RR-interval at lead II were measured to calculate the QTc (Bazett formula: QTc = QT/square-root of RR interval). QT dispersion was the difference between the maximum and minimal QT measured from each of 12 leads. During As₂O₃-ON and As₂O₃-OFF, 24-hour Holter monitoring (Zymed DigiTrak Plus; Zymed 1810, Philips, The Netherlands) was performed to assess circadian QT variations.

Holter recordings were reviewed and edited manually. Recordings must exceed 20 hours and be of good quality to be analyzed. The time domain (standard deviation of normal RR interval [SDNN]) and frequency domain (ratio of low-frequency to high-frequency power [LF/HF]) of HVR was measured to assess parasympathetic and sympathetic activities, respectively.¹⁶  ECG and Holter were analyzed by a cardiologist blinded to patient treatment.

After venepuncture, EDTA-anticoagulated blood was immediately separated into plasma and cell fractions. Arsenic levels were measured by inductively coupled plasma mass spectrometry.¹⁷ 

Continuous variables were expressed as mean plus or minus standard error of the mean. Comparisons were performed with the Student t test or Fisher exact test (SPSS software, version 10.0). P values less than .05 were considered significant.

Sixteen patients completed the study (one died of sepsis). Consistently, cellular arsenic levels were significantly higher than plasma arsenic (Figure 1A), similar to previous observations during long-term oral As₂O₃.¹³ 

Both QT and QTc were significantly longer during As₂O₃-ON than in As₂O₃-OFF (Table 2, P < .01). However, the QT and QTc dispersion were comparable during As₂O₃-ON and As₂O₃-OFF (Table 2, P > .05). Holter monitoring showed similar circadian heart rate variations during As₂O₃-ON and As₂O₃-OFF, but 24-hour heart rates were consistently higher during As₂O₃-ON than in As₂O₃-OFF, being significantly different during late evening (21: 00, 22:00, 23:00), early morning (02:00, 03:00, 04;00), and afternoon (12:00, 13:00, 14:00; P < .05; Figure 1B).

QT intervals over 24 hours showed marked variations, although the pattern and measurement at each hour were comparable between As₂O₃-ON and As₂O₃-OFF (Figure 1C). Circadian variations of mean QTc intervals (Figure 1D) were comparable between As₂O₃-ON and As₂O₃-OFF, but measurements at each hour were significantly longer during As₂O₃-ON (P < .05, except at 06:00, 19:00, 23:00). The mean differences of QTc between As₂O₃-ON and As₂O₃-OFF are shown in Figure 1E. QTc prolongation of more than 30 milliseconds was observed at only one time point (16:00, 2 hours after oral As₂O₃). No case had QTc prolongation of more than 50 milliseconds. This resulted in a QTc of more than 500 milliseconds in only 3 of 16 patients, all within 4 hours of oral As₂O₃. Although the SDNN was significantly lower during As₂O₃-ON than in As₂O₃-OFF, there was no significantly difference in LF/HF between As₂O₃-ON and As₂O₃-OFF (Table 2).

Ventricular premature beats were comparably frequent between As₂O₃-ON and As₂O₃-OFF (0.03% ± 0.02% versus 0.03% ± 0.03%; P = .8). No ventricular tachyarrhythmia was observed.

Long-term oral As₂O₃ significantly increased the mean heart rate and QTc interval, and reduced SDNN. However, QT-interval, QT and QTc dispersions, and LF/HF were not changed. Furthermore, for indicators of proarrhythmic risks,^12,13  significant QTc prolongation of more than 30 milliseconds was observed at only a single time point (2 hours after oral As₂O₃), QTc prolongation never exceeded 50 milliseconds, and a QTc interval of more than 500 milliseconds was observed in only 3 patients within 4 hours of oral As₂O₃. Importantly, these observations translated into absent ventricular proarrhythmia in all patients. These results were superior to intravenous As₂O₃, where 26% of patients had QT intervals more than or equal to 500 milliseconds, with QTc interval prolonged by 30 to 60 milliseconds in 36.6% of treatment courses, and by more than 60 milliseconds in 35.4% of patients, resulting in torsades de pointes in 1% of cases.⁸ 

The cardiac safety of oral As₂O₃ may be due to several reasons. First, in ventricular proarrhythmias during intravenous As₂O₃, As₂O₃ was infused over 1 to 3 hours.^6-9  Pharmacokinetic studies showed that peak plasma arsenic levels after intravenous As₂O₃ reached 0.34 μM to 2.0 μM.¹¹  However, after oral As₂O₃, peak arsenic levels were much lower at 0.25 μM to 0.55 μM (although gradual intestinal absorption led to a total area-under-the-curve availability comparable to intravenous As₂O₃¹¹ ). Because of lower peak arsenic levels, QTc prolongations after oral As₂O₃ were transient and lasted only 4 hours.

Second, arsenic shows a concentration-dependent blockade of I_Kr and I_Ks currents (IC₅₀, I_Kr: 0.14 μM ± 0.01 μM; I_Ks: 1.13 μM ± 0.06 μM).⁵  Hence, peak plasma arsenic concentrations reached after intravenous As₂O₃, but not oral As₂O₃, are high enough to block I_Kr and I_Ks, thereby severely compromising repolarization and increasing susceptibility to ventricular proarrhythmias.

Finally, heart rates were significantly increased during oral As₂O₃. HVR analysis demonstrated significant decreases in parasympathetic activity without changes in sympathetic activity. The change in sympathetic/parasympathetic balance increases the heart rate, alleviating risks of ventricular tachyarrhythmias during QT prolongation. Hence, despite significant increases in QTc interval, overall QT intervals were unchanged during oral As₂O₃.

A limitation of this study is absence of an intravenous As₂O₃ control group. Since formulation of oral As₂O₃ in 2000, we have not used intravenous As₂O₃, making it difficult to include an intravenous As₂O₃ group just for this study. Previous reports of intravenous As₂O₃ have not measured QT dispersion, HRV, and circadian QT and HR variations, so how they relate to arrhythmias during intravenous As₂O₃ is difficult to judge. Finally, a 1% frequency of torsades de pointes in intravenous As₂O₃ might not have been detected in our study size.

Our results provide insights into how an oral formulation improves the cardiac safety of As₂O₃. Oral As₂O₃ is convenient and safe for outpatients, and may be the preferred formulation for prolonged arsenic treatment.

The authors thank the S. K. Yee Medical Foundation for provision of free oral arsenic trioxide to patients.