Introduction
This article discusses the important role of clinicians in the anticipation, identification, and treatment of one particular cardiac occurrence in the practice of pharmaceutical medicine, namely the polymorphic ventricular tachycardia Torsade de Pointes (Torsade). Although drugs from diverse pharmacological classes can lead to multiple types of arrhythmias, the focus is on Torsade as many drugs prolong the cardiac repolarization period, a phenomenon associated with Torsade. The mechanism of action of repolarization prolongation is discussed in due course.
Sudden death caused by drug-induced arrhythmia has been described as one of the most feared complications in medicine (Link, Yan, & Kowey, 2010). Torsade was first described by Dessertenne (Dessertenne, 1966: see also Bubien, 1999, for discussion in the nursing literature). It is a rare polymorphic ventricular arrhythmia that typically occurs in self-limiting bursts that can lead to symptoms of dizziness, palpitations, syncope, and seizures, but the one that can occasionally progress to ventricular fibrillation and sudden cardiac death. As shown in Figure 1, the electrocardiogram (ECG) waveform of Torsade is very different from that seen during normal sinus rhythm. It is characterized by rapid, irregular QRS complexes that appear to twist around the isoelectric baseline.
Torsade can result from inherited condition called long QT syndrome (LQTS), and of importance in the practice of pharmaceutical medicine, it can also be drug induced by a wide range of drugs and drug classes. In both cases, Torsade is associated with prolongation of the QT interval, the time between the onset of the QRS complex and the offset of the T wave, as seen on the surface ECG. Figure 2 presents a stylistic representation of the QT interval and QT-interval prolongation. For each cardiac cycle, ventricular depolarization is reflected in the QRS waveform, whereas ventricular repolarization is reflected in the T wave. Delayed ventricular repolarization is manifest as prolongation of the QT interval (Turner et al., 2018).
In recent years, a dedicated clinical pharmacology study has been conducted for each new drug under development to assess its propensity to induce Torsade using QT-interval prolongation as a biomarker. When submitting a New Drug Application to the US Food and Drug Administration and similar documents to regulatory agencies in other geographic jurisdictions, the results of this study are included in the sponsor's application dossier. Identification of QT-interval prolongation does not necessarily prevent a drug from receiving marketing approval if its overall benefit-risk balance is favorable, but, if approved, a warning is placed in the drug's Prescribing Information (Turner, Karnad, Cabell, & Kothari, 2017).
When prescribing, dispensing, and administering such a drug to a patient, health care professionals need to consider this information in conjunction with the patient's likelihood of receiving required therapeutic benefit and his or her susceptibility for Torsade. This susceptibility is influenced by biological risk factors including advanced age and female sex, concurrent clinical characteristics including hypocalcemia, hypokalemia, and hypomagnesemia, and concomitant medications that also prolong the QT interval. Monitoring patients' ECGs and calcium, sodium, and magnesium electrolyte levels in hospital settings is also important (Turner et al., 2018).
This article explains why drugs can have a proarrhythmic propensity, discusses the origins and nature of the regulatory landscape that governs current premarketing assessments, provides examples of resulting Prescribing Information warnings, and presents a scenario illustrating clinical practice implications for nurse practitioners and physician assistants. Table 1 provides a summary of key take-home messages.
Inherited long QT syndrome
Although drug-induced (acquired) QT prolongation is the primary focus of this article, it is important to be familiar with inherited (congenital) LQTS. In clinical medicine, QT prolongation is of concern because it is the defining phenotypic characteristic of a constellation of LQTSs that have been associated with Torsade and sudden death (Fernandez-Falgueras, Sarquella-Brugada, Brugada, Brugada, & Campuzano, 2017). Of specific interest in the present context is long QT syndrome 2 (LQT2), discussed shortly.
As noted previously, QT-interval prolongation is a manifestation of delayed ventricular repolarization. Both depolarization and repolarization are governed by the flow of ionic currents through cardiac ion channels. Ion channels, which occur in multiple biological systems, are structured transmembrane protein complexes that are embedded in cell plasma membranes. Multiple subunits within these complexes organize to form a transmembrane (central) pore through which ionic currents flow across cell membranes. Multiple cardiac ionic currents influence both depolarization and repolarization. Ikr, a repolarizing current, is of direct relevance here.
LQT2 is seen when an individual's genetic inheritance includes an abnormal variant of the human ether-a-go-go-related gene (hERG) that encodes a protein comprising the subunit forming the central pore of the cardiac potassium ion channel (known as the hERG channel) through which Ikr flows. Abnormal variants of hERG lead to a cascade of consequences as follows: loss of function of the expressed ion channels, reduced Ikr flowing through them, reduced net repolarizing influence, delayed repolarization, QT-interval prolongation, and sometimes proarrhythmia, that is, the potential generation of a new arrhythmia or the worsening of an existing arrhythmia (Turner & Durham, 2009).
Drug-induced QT prolongation
Ion channels have various defining characteristics. One, just discussed, is that abnormal variants of genes encoding them can cause specific arrhythmias. Another is that their activities can be affected by drugs. (Murray & Granner, 2006). In the case of the hERG channel, drugs can lead to a reduction in function. This results in a cascade of consequences-reduced Ikr, reduced net repolarizing influence, delayed repolarization, QT-interval prolongation, and sometimes proarrhythmia-that are in many ways similar to those resulting from abnormal variants of hERG and the resultant inherited LQT2. Given this similarity, hERG channel blockade has become a significant scientific, clinical, and regulatory concern during the past three decades (Turner et al., 2018). A brief summary is provided in the following section.
Origins of the proarrhythmic cardiac regulatory landscape and the thorough QT/QTc study
Multiple drugs were removed from the United States and other markets in the late 1980s to the early 2000s because of proarrhythmic concerns, including astemizole, cisapride, grepafloxacin, prenylamine, sparfloxacin, sertindole, terfenadine, and terodiline (Satin, Durham, & Turner, 2011; Talbot & Waller, 2004). Concerns regarding these marketing withdrawals sparked a chain of events that culminated in the design of the dedicated clinical pharmacology study mentioned in the Introduction section of this article. This study, the thorough QT/QTc (TQT) study, was introduced in a guideline released by the International Council for Harmonisation (ICH). This guideline, ICH E14 (ICH, 2005), has been adopted by multiple regulatory agencies worldwide, which recommend that sponsors conduct a TQT study for a new drug under development to assess its propensity to prolong the QT interval. A TQT study is typically conducted at a clinical pharmacology unit in which study participants reside for the length of the study to allow very tight control of their daily activities. It is also a safe environment should any untoward drug reactions occur. Electrocardiograms are collected digitally after dosing and then transmitted to specialized laboratories where any drug-induced QT-interval prolongation is measured by expert ECG readers. Before the data are analyzed, the QT-interval measurements are "corrected" for concurrent heart rate, yielding a measurement called QTc. This is done because, regardless of the administration of a drug, heart rate affects the QT interval: higher heart rates are associated with shorter QT intervals and vice versa.
Study participants receive two doses of the test drug: the intended clinical dose should the drug be approved for marketing, and a supratherapeutic dose that, if tolerable, is several multiples of the intended clinical dose. The purpose of administering the supratherapeutic dose is to evaluate drug-induced changes in ECG parameters under "worst case scenarios," that is, the highest exposures that would likely be attained in patients due to effect modifiers including pharmacokinetic variability, drug-drug interactions, alterations in metabolism or elimination, and/or underlying heart disease.
The ICH E14 guideline threshold of concern is a mean of TQT study participants' increases in QTc from a placebo baseline of "around 5 ms." The guideline also recommends categorical descriptive statistical summarization of data from outliers, that is, individual participants who have notably larger increases. The number of participants showing increases in QTc of greater than 30 and 60 ms is presented in the study report. In addition, description of absolute values is recommended. The number of participants showing values greater than 450, 480, and 500 ms is reported.
These QT values were chosen for specific reasons. First, drugs that produce mean changes of less than 5 ms at high supratherapeutic exposures in healthy participants have rarely been associated with significant cardiac risk in clinical use. This criterion can therefore be reasonably used to exclude risk. Second, proarrhythmic events have usually been associated with individuals having QTc values greater than 480 ms or having changes from baseline in QTc greater than 60 ms (Turner et al., 2018).
Risk factors for Torsade
It is important to emphasize that QTc interval prolongation by itself does not necessarily lead to Torsade. Drug-induced Torsade typically requires multiple factors to be present at the same time. Therefore, the most important risk-reducing intervention clinicians can make is undertaking a careful review of other risk factors when prescribing medications (Beach et al., 2018). Various authors have discussed about risk factors (Heemskerk et al., 2018; Vandael, Vandenberk, Vandenberghe, Willems, & Foulon, 2017; Vlachos, Georgopoulos, Efremidis, Sideris, & Letsas, 2016), and Table 1 provides a summary of this information. These factors can act as effect amplifiers that can make an otherwise relatively safe drug potentially unsafe with regard to the risk for Torsade in the setting of QT-interval prolongation (Vlachos et al., 2016).
The CredibleMeds web site
CredibleMeds (Woosley, Black, Heise, & Romero, 2018) maintains a list of drugs known to prolong QTc. To assess the risk of harm from medicines scientifically, this organization has developed a risk-stratification process, the Adverse Drug Event Causality Analysis, that includes monitoring and analysis of multiple related sources to yield a list of several hundred drugs of interest (Woosley et al., 2017). This list is divided into categories based on a drug's likelihood to cause QTc prolongation or Torsade. List 3 contains those with a conditional risk; list 2 contains those with a possible risk; and list 1 contains drugs with a known risk of Torsade. Examples are provided in Table 2.
There is also a fourth category comprising drugs that pose a high risk of Torsade de Pointes for patients with inherited LQTS. Drugs in this category include all those in the conditional, possible, and known categories plus additional drugs that do not prolong the QT interval per se but which have a "Special Risk" because of other associated drug-induced effects that could lead to proarrhythmia not necessarily linked with QT prolongation.
There are three sections to the web site as follows: information for everyone, information for health care providers, and information for research scientists. Readers of this article are encouraged to become familiar with all sections. It is also appropriate to direct patients prescribed a drug with a proarrhythmic liability to the "information for everyone" section (Table 3).
Treatment strategies
Interventions should be made when episodes of Torsade (rather than just QTc prolongation) occur. The first intervention to consider is termination, or dose reduction, of the drug. If this is not possible or desirable, for example, if there is no other drug that can provide the desired therapeutic benefit, other interventions are possible.
When Torsade occurs in recurrent self-terminating bursts, the more common occurrence, treatment has two aims: stabilize the myocardium using magnesium sulfate; and shorten repolarization (the QTc interval) by increasing heart rate using either chronotropic drugs such as isoproterenol and atropine or cardiac pacing (Thomas & Behr, 2016). Magnesium sulfate is recommended as an immediate first-line treatment and is simple and relatively safe to administer. That said, care should be taken not to induce hypermagnesemia. Higher doses can lead to conditions such as nausea, vomiting, and drowsiness, whereas substantially higher doses can lead to a variety of more serious outcomes including cardiac arrhythmias, coma, and cardiac arrest (Thomas & Behr, 2016).
Prolonged episodes of continuous TdP associated with severe hypotension or cardiac arrest should be terminated by electrical cardioversion (Drew et al., 2010).
Case presentation
The following scenario, developed by the authors, is presented to discuss practical considerations when prescribing a medication with proarrhythmic potential in the clinical setting. Specifically, the use of a fluoroquinolone antibiotic is discussed, as these drugs are frequently prescribed and have (to a greater or lesser extent, depending on the specific fluoroquinolone) demonstrated proarrhythmic potential (Gorelik et al., 2019; Mehrzad & Barza, 2015). Much of the current literature has discussed and provided clinical considerations for the use of fluoroquinolones as they relate to risk of tendon rupture and central nervous system complications, particularly among the older person. Although these continue to be important considerations in the judicious use of these medications, the purpose of this scenario is to focus on the use of fluoroquinolones as they relate to risk of QT prolongation and possible dysrhythmia.
Patient 1 is a 65-year-old woman. Her medical history is as follows: diabetes mellitus type 2, for which she had been prescribed metformin; hypertension, which has been well controlled on hydrochlorothiazide and lisinopril; and hypercholesterolemia, for which she is taking atorvastatin, which is well tolerated. Her last annual check-up was 6 months before today's acute visit. At that time, her body mass index was calculated at 29.0, and her laboratory test results were unremarkable with the exception of hemoglobin A1c at 7.6, a total cholesterol of 238, and a creatinine clearance of 62 ml/min. Based on her annual laboratory test results, adjustments in her medications were made to include the addition of sitagliptin to her metformin and an increased dose adjustment in her atorvastatin. In addition to changes made to her medications, L.E. had also agreed to follow-up with the in-house diabetes educator in an effort to reduce her blood glucose levels. She was to have followed up in 3 months' time, but due to transportation issues missed the appointment date and did not reschedule.
L.E. presents to her primary care clinic for an acute visit. Her chief complaint is consistent with a urinary tract infection (UTI), including a persistent urge to urinate with burning sensation. Her symptoms have been going on for the past three days, and she reports that "I maybe felt a little warm" the night before today's visit but denies any other symptoms. L.E. is recently widowed and denies any recent sexual activity. She states that she thinks "I have another urinary tract infection." A previous infection was treated successfully 6 months before this visit at an urgent care clinic, but she cannot recall the medication she was given. There is no record from this visit in her current medical record.
Physical assessment for L.E. reveals vital signs within normal limits. She is not currently febrile, has no palpable bladder distension, and has no costovertebral angle tenderness. She does have mild suprapubic discomfort exhibited on palpation. Urine void is cloudy, without discernibly foul odor. Urine dipstick reveals a moderate amount of leukocytes, positive for nitrates, and a small amount of blood with a trace amount of protein, no glucose, and no ketones. She denies having had anything to eat before her appointment, and her finger stick glucose is 128 mg/dl. Based on her history of present illness, physical examination, and urinalysis, the clinician's working diagnosis is cystitis versus pyelonephritis. The clinician decides to treat for cystitis rather than pyelonephritis because of lack of costovertebral angle pain or tenderness and her vitals at this visit, all being within normal limits.
Although current research suggests that UTIs among patients diagnosed with diabetes mellitus need not immediately be considered complicated, the clinician decides to send her urine for a urine culture test (Vinken et al., 2018). The clinician bases this decision on her history of suboptimal hemoglobin A1c levels combined with an UTI 8 months prior. Although this past diagnosis does not reflect the accepted definition of a recurrent UTI (i.e., three or more UTI in 12 months), the clinician considers this history as additional cautious reasoning for sending her urine for a urine culture test (Arnold, Hehn, & Klein, 2016; Malik, Wu, Christie, Alhalabi, & Zimmern, 2017).
L.E. verifies she has not taken any antibiotics since her prior diagnosis of an UTI 8 months ago. She also verifies having no known drug allergies. To begin treatment, the clinician prescribes nitrofurantoin ER based on current evidence and treatment recommendations (Arnold et al., 2016; Malik et al., 2017; Zhanel et al., 2006) for uncomplicated UTIs, as well as pyridium to alleviate associated urinary symptoms.
Another first-line agent commonly prescribed in this scenario might be trimethoprim-sulfamethoxazole. However, the clinician avoids this option because L.E. is also on an angiotensin-converting enzyme inhibitor (ACEI) which could potentially increase the risk of hyperkalemia (Merel & Paauw, 2017). L.E. is educated regarding proper use of current medications and instructed to call the clinic if symptoms have not resolved within 48 hours or if symptoms worsen.
Urine culture and sensitivity results are received by the clinician 48 hours later and reveal the infective organism as Escherichia coli. Sensitivity analysis reveals multiple drug resistance to include L.E.'s currently prescribed antibiotic. On further review of the results, the highest susceptibility demonstrated for appropriate outpatient antibiotic treatment at this time is levofloxacin.
In the meantime, L.E. has not contacted the clinic to report any lack of improvement or worsening of symptoms. Based on the culture and sensitivity results; however, the clinician contacts her to discuss the results of the laboratory test report and to inquire as to her current state of health. Her symptoms have essentially worsened with continued subjective fevers as well as recent onset of nausea and loose stools within the last 8-12 hours. Based on her report and the culture and sensitivity results, the clinician decides to have her come back that same day for an assessment.
Repeat assessment reflects symptoms consistent with pyelonephritis, including mild fever, costovertebral angle tenderness, nausea, and recent onset of loose stools/diarrhea (Nitzan, Elias, Bibiana, & Saliba, 2015). Moreover, the provider's diagnosis is further informed by a known infectious organism and antibiotic sensitivities for said organism. Although fluoroquinolones are known to the clinician as carrying a risk of sudden death due to QT prolongation, the clinician must consider the specific risk(s) related to this individual patient. In addition, based on the high sensitivity of the infectious organism to fluoroquinolone and the serious complications associated with an improperly treated UTI, the clinician must consider the benefit of this particular therapy weighed against this patient's specific risk(s). This is of particular importance for this patient because the clinician is aware that diabetic patients have worse UTI outcomes than nondiabetic patients. These comparatively worse outcomes include the need for inpatient management with longer hospitalizations, bacteremia, and septic shock. In addition, the clinician is also aware that related mortality rates from UTI are five times higher in diabetic patients aged 65 years and older (Nitzan et al., 2015).
As discussed earlier and outlined in Table 1, risks for this patient include female sex and age 65 years, which, according to the World Health Organization, is generally considered the age at which one is considered older person (WHO, 2010). In addition, this patient's medical history has demonstrated a potential risk of increased drug bioavailability based on her renal function. Her renal function could be due to age-related changes and/or previously poorly controlled blood glucose levels. She was unable to return for follow-up after appropriate medication adjustments were made at her last annual visit over 8 months ago. It is reasonable to predict an improvement in renal function because of effective changes in her antidiabetic medication. At this time, her preprandial glucose level is in an appropriate range, and there were no glucose or ketones present in her urine. In addition, age-related changes in kidney function might be a bit premature, although reasonable to consider.
In terms of concomitant medication use that might place this patient at increased risk, she is taking a thiazide-type diuretic, which could potentially cause decreased potassium levels. However, she is also taking an ACEI, which may minimize her risk of secondary hypokalemia caused by diuretics (Palmer, 2008). It is important to still consider electrolyte imbalances for this patient, however, as she is currently febrile and has reported subjective fevers since her first office visit. In addition, she is now experiencing loose stool/diarrhea. For these reasons, the clinician orders a comprehensive metabolic panel and a serum magnesium level to assess for any electrolyte imbalance and to re-evaluate her renal function. Cardiac characteristics for the patient are negative at this time. Her vital signs have not evidenced any bradycardia. She denies congenital heart disease and has no current medical history to suggest any structural heart disease. It is important that, however, there is no baseline ECG for this patient to determine any conduction abnormalities as an added potential risk. Although her past visits have not elicited any clinical suspicion for possible conduction abnormalities, the clinician finds it important to establish a baseline ECG before prescribing a fluoroquinolone for her UTI. In this scenario, the patient's ECG demonstrates no evidence of QT prolongation.
Given the serious complications associated with medications that may induce QT prolongation, the clinician has made themselves familiar with these medications. In this scenario, the use of fluoroquinolone is determined necessary as evidenced by the sensitivity results and the patient being an appropriate candidate for outpatient management. The clinician considered current recommendations addressing monitoring parameters including obtaining an ECG before treatment with any drug with proarrhythmic propensity in an effort to detect baseline prolonged QT intervals. The clinician considered obtaining this patient's electrolytes to correct (if indicated) any abnormalities, particularly hypokalemia or hypomagnesemia (Blancett, Flynn, Akers, & Smith, 2006). The clinician also recommended that this patient discontinues her hydrochlorothiazide during fluoroquinolone treatment because of possible fluid loss related to fever and recent onset of diarrhea, as well as to reduce her risk related to concomitant medication use. In addition, the clinician understands that medications with proarrhythmic propensity may vary within a given class of medication. In this patient, for example, the fluoroquinolone levofloxacin differs from ciprofloxacin in terms of proarrhythmic propensity from improbable to very improbable, respectively (Al-Khatib, Allen LaPointe, Kramer, & Califf, 2003; Liu, 2010).
In a further attempt to decrease risk of possible cardiac complications in this patient due to fluoroquinolone use, the clinician opts to treat her infection with ciprofloxacin, given its very improbable status for causing QT prolongation. Finally, the clinician takes the opportunity to educate her about the use of ciprofloxacin and instructs her to immediately report if she is unable to take the medication because of worsening nausea or vomiting. The provider makes a note to follow-up with her the following day through telephone as the patient is reluctant to return to the clinic for follow-up. In addition, she is educated about symptoms of QT prolongation including any shortness of breath, tachycardia, or any episodes of syncope and to seek emergency care if any of these symptoms develop.
The provider has realized that when drugs with proarrhythmic potential are used, the potential benefits are an essential consideration and that individual patient risks are recognized and minimized. Although in-clinic follow-up to include a repeat ECG would be a reasonable consideration for this patient, the provider has carefully informed her of potential risks to assist the patient in making an informed decision. Following-up with the patient through telephone revealed the patient tolerating the medication with no symptoms of QT prolongation and improvement of symptoms overall.
Practical implications and conclusions
When considering the use of medications that carry a Prescribing Information warning of proarrhythmic potential, providers must consider individual patient risk (or risks) for developing QT prolongation and the potential for Torsades. When prescribing such medications, health care providers must understand that these risks are influenced by biological factors including advanced age and female sex, current clinical characteristics including hypocalcemia, hypokalemia, and hypomagnesemia, structural heart disease, and/or conduction irregularities that affect normal cardiac electrophysiology, as well as concomitant medications that might also potentially prolong the QT interval. It is important that physician assistants, nurse practitioners, and indeed all health care providers are aware of all these considerations, as optional medication treatments may be few or impractical. In addition, despite the potential for QT prolongation among certain medications, judicious use is often appropriate and necessary based on the potential therapeutic benefits of these drugs.
References