Drug-induced QT interval prolongation: mechanisms and clinical management

Senthil Nachimuthu, Manish D. Assar, and Jeffrey M. Schussler

Additional article information

Abstract

article-meta

The prolonged QT interval is both widely seen and associated with the potentially deadly rhythm, Torsades de Pointes (TdP). While it can occur spontaneously in the congenital form, there is a wide array of drugs that have been implicated in the prolongation of the QT interval. Some of these drugs have either been restricted or withdrawn from the market due to the increased incidence of fatal polymorphic ventricular tachycardia. The list of drugs that cause QT prolongation continues to grow, and an updated list of specific drugs that prolong the QT interval can be found at http://www.qtdrugs.org. This review focuses on the mechanism of drug-induced QT prolongation, risk factors for TdP, culprit drugs, prevention and monitoring of prolonged drug-induced QT prolongation and treatment strategies.

Keywords: drugs, QT interval, Torsades de Pointes

QT interval physiology and mechanism of QT drug-induced prolongation

The QT interval on the surface EKG represents the summation of action potential (AP) of ventricular myocytes. The action potential reflects the flow of ion currents across a cell membrane through specialized channels made of protein complexes (Figure 1, Titier et al. 2005). Malfunction of these protein channels can lead to either increased inward current or reduced outward current. This subsequently increases the action potential duration and hence QT interval prolongation.

fig ft0fig mode=article f1

Figure 1.

caption a4

Five phases of cardiac depolarization and repolarization. Phase 0: Large inward current of sodium ions (INa). Phase 1: Inactivation of INa and the transient efflux of potassium ions (It0). Phase 2: Plateau phase with influx of calcium ions through L-type …

Mutations of the genes that encode the protein channels (IKr, IKs and Na) result in congenital long QT syndrome (LQTS) [Ching and Tan, 2006]. In acquired LQTS, the mechanism is almost always due to blockage of the inward potassium rectifier (IKr) channel, also known as the hERG (ether a go go) channel. It conducts a rapid delayed rectifier potassium current (Ikr), a critical current in the phase 3 repolarization of the cardiac action potential [Roden and Viswanathan, 2005]. Inherited mutations (loss of function) of the hERG gene lead to type 2 LQTS. Medications that prolong QT interval act on the same hERG channel. The distinct molecular structure of the hERG channel makes it more susceptible to medications.

The structure of the hERG channel

The structure of the hERG channel is well understood from the structure of bacterial and mammalian K channels [Swartz, 2004]. The hERG channel is essentially formed by the co-assembly of four alpha subunits, each of which has six transmembrane spanning alpha-helical segments (S1–S6) [Sanguinetti and Tristani-Firouzi, 2006; Swartz, 2004] (Figure 2). The first four helices (S1–S4) in each segment form a voltage sensor domain (VSD) that senses the transmembrane potential. The next two helices (S5 and S6) form the pore domain that contains a short alpha helix (pore helix) and selectivity filter. Four of these (one from each subunit) come together to form a central pore that is responsible for the movement of potassium current. Below the selectivity filter, the pore widens to form a central cavity. It is lined by many unique aromatic residues that are absent in most other K channels. These optimally positioned aromatic residues and the polar residues are an integral part of the unique binding sites for diverse pharmacologic agents [Perry et al. 2010; Perrin et al. 2008; Kannankeril, 2008].

fig ft0fig mode=article f1

Figure 2.

caption a4

(a) A single hERG subunit containing six α-helical transmembrane domains, S1–S6. (b) Structure of a KcsA K+ channel crystallized in the closed state. Only two of the four subunits are shown. White spheres are K+ ions located within the …

The other mechanisms by which drugs induce prolonged QT interval are: disruption of KCNH2 protein trafficking leading to loss of K channels by drugs such as arsenic oxide, pentamidine and fluoxetine [Ficker et al. 2004; Kuryshev et al. 2005; Rajamani et al. 2006], rescue of SCN5A channel causing increased inward sodium current by cisapride [Kannankeril, 2008] and an increase inward calcium current by antimony [Kuryshev et al. 2006].

The congenital long QT syndrome

More than 10 different types of congenital LQTS have been recognized [Hedley et al. 2009; Modell and Lehmann, 2006; Roden, 2008]. LQT1, LQT2, and LQT3 account for the majority of the cases of congenital LQTS. LQT1 accounts for 40–55% of cases of the LQTS [Schwartz et al. 2001; Splawski et al. 2000]. It is caused by mutations in the KVLQT1 (also called KCNQ1). LQT1 is characterized by events that are induced by exercise.

LQT2 accounts for 35–45% of cases of congenital LQTS [Schwartz et al. 2001; Splawski et al. 2000]. It is caused by a variety of mutations in the hERG (also known as KCNH2) potassium channel gene, located on chromosome 7. The mutations may involve the pore or the nonpore region of the hERG channel. Pore mutations carry high risk for cardiac events and may affect young patients [Moss et al. 2002] whereas nonpore mutations often lead to Torsades de Pointes (TdP) in the presence of hypokalemia [Berthet et al. 1999]

LQT3 accounts for 8–10% of cases [Schwartz et al. 2001; Splawski et al. 2000]. It is caused by mutations in the sodium channel gene (SCN5A) located on chromosome 3 at location 21–24. It is characterized by events occurring at rest or during sleep.

Measurement of the QT interval

On a 12-lead ECG, the QT interval is measured from the beginning of the QRS complex to the end of T wave as it returns to baseline. Manual measurements of the QT interval should be taken from leads II and V5 or V6 with the longest value being used. Measurements taken from these leads have the greatest positive and negative predictive value in detecting abnormal QT intervals [Monnig et al. 2006]. A mean value should be derived from at least 3–4 cardiac cycles. The end of T wave can be determined reliably by the slope method where it is defined by the intersection point between the tangent drawn at the maximum downslope of the T wave and the iso-electric line. If the T wave is notched, the tangent should be applied to the maximum slope. Smaller U waves (<0.1 mV) should be excluded whereas larger U waves merging with T waves should be included in the measurements [Anderson et al. 2002]. There are no standards for interpreting prolonged QT intervals from Holter or 24/48 h ambulatory monitoring records available. As result, ambulatory monitoring of QT assessment is not recommended.

Several factors such as gender, heart rate, underlying rhythm and conduction defects influence the QT interval. It is also influenced by the physiologic and metabolic state of the patients. Numerous methodologies for correcting QT intervals for heart rate have been proposed, and each has its own benefits and shortcomings. There is no consensus as to which one is the most effective. However, the most universally adopted method is Bazett’s formula (QTc = QT/√RR in seconds) that provides an adequate correction for heart rate ranging anywhere between 60 and 100 beats/min. Nonetheless, it underestimates and overestimates the QT interval at low and high heart rates, respectively.

For heart rates outside the normal range, other correction methods such as Fredericia (QTc = QT/(RR)1/3) or Framingham (QTc = QT + 0.154(1 – RR) should be utilized (Table 1) [Aytemir et al. 1999]. However, these correction methods are based on population mean correction factor and do not address intra- or inter-individual variability. As there is now a strong evidence for significant inter-individual variability, the best HR correction for QT should be estimated for each individual [Batchvarov et al. 2002; Malik et al. 2002; Couderc et al. 2005].

table ft1table-wrap mode=article t1

Table 1.

caption a4

Methodologies for correcting QT intervals for heart rate.

Estimation of individual correction factor is a cumbersome, time-consuming process. It is preferred in clinical studies but not applicable in clinical practice [Piotrovsky, 2005]. Fossa and colleagues proposed a QT-HR nomogram based on a QT-RR cloud diagram developed from human preclinical studies (Figure 3) [Chan et al. 2007] that can readily be used in the clinical setting. The nomogram incorporates HR rather than RR interval and is found to be safe, with excellent sensitivity, and at the same time is specific enough to allow the assessment of many patients as ‘not at risk’ for drug-induced TdP and therefore not requiring cardiac monitoring [Chan et al. 2007]. The performance of the nomogram in patients who incurred an antidepressant overdose but did not develop arrhythmia was studied by Bateman and colleagues [Waring et al. 2010].

fig ft0fig mode=article f1

Figure 3.

caption a4

QT interval nomogram for determining ‘at risk’ QT–HR pairs from a single 12-lead ECG. Use: The QT interval should be measured manually on a 12-lead ECG from the beginning of the Q wave until to the end of the T wave in multiple …

The QT nomogram was associated with a lower false-positive rate than other widely accepted QTc criteria. The greatest discrepancy between the nomogram and QTc methods was amongst patients with heart rates between 30 and 60 beats/min. For example, patients with a citalopram overdose had QT values above the nomogram compared with venlafaxine and mirtazapine overdose, predicting its higher risk for developing arrhythmia and thus the need for close cardiac monitoring [Waring et al. 2010].

Based on Bazett’s corrected QTc value, in adult males a QT interval greater than 450 ms is considered prolonged and between 430 and 450 ms is considered borderline. For females, a QT interval greater than 470 ms is considered prolonged and between 450 and 470 ms is considered borderline [Goldenberg et al. 2006].