Cardiac Safety in Clinical Trials

Less than 15 years ago the CPMP released a points-to-consider memorandum describing steps to be taken to study the ventricular repolarization liability on the 12-lead ECG. Since that time, many steps have been undertaken by regulatory bodies and the industry (ICH S7A, S7B, and ICHE14 Guidance) as a whole to ensure that drugs that make it to market are safe and without ventricular repolarization liabilities as seen on the surface ECG, as evidenced by a prolongation of the QT interval. Despite these key steps on assessment of the risk of cardiac arrhythmias, drugs such as Vioxx and Avandia have shown a propensity for cardiac toxicity outside the risk of torsades de pointes.

It is now realized that cardiac safety involves more than just measuring the QT interval on the surface ECG. This article reviews, briefly, the current state of affairs in the area of cardiac safety in the world of drug development.

Non-cardiovascular compounds

The ICH has released three guidelines that are important to cardiac safety; S7A, S7B, and E14. The S7A deals with a variety of preclinical issues while the S7B and E14 address the need for dedicated experiments and clinical trials for assessing the QT interval. Since that time these guidelines have been useful for the sponsor to determine what is needed to prove to regulatory agencies that the compound has no overt cardiac safety liability.

S7A

The ICH S7A (July 2001) was constructed to help describe a rational approach for conducting preclinical safety pharmacology studies. Among other evaluations, the S7A recommends certain tests to be conducted for cardiovascular safety. This "core battery" of tests includes primary and follow-up assessments for the central nervous system, respiratory system, and the cardiovascular system.

The cardiovascular system tests can include investigations for heart rate, blood pressure, and electrocardiograms (to include an assessment for ventricular repolarization). Additional tests may include assessments for cardiac output, ventricular contractility, vascular resistance, and effects for endogenous and/or exogenous substances. These preclinical studies, in combination with short-term and long-term toxicology studies will help the sponsor determine the direction of cardiovascular testing as the compound advances through different phases of clinical testing.

S7B

The ICH S7B (October 2005) specifically addresses the non-clinical evaluation of delayed ventricular repolarization in human pharmaceuticals. It is an extension of the ICH S7A and the non-clinical correlate to the ICH E14. The S7B calls for an integrated risk assessment using both in-vivo and in-vitro testing.

Section 2.3.1 of the S7B mentions the hERG (human ether-go-go related gene). The so-called "hERG-assay" gives a relatively reliable indication as to whether or not the compound has the ability to prolong the QT-interval in human subjects via a block of the potassium outward current ("loss of function"). Caution is required, as the correlation from the hERG assay to the QT-interval on the ECG is not perfect. In addition, the hERG testing could also evidence a potential for a drug-induced QT shortening when the drug is associated with a gain of function of the repolarizing potassium currents.

Other testing methodology is addressed including measuring of ionic currents in isolated myocytes, measurement of ECGs in animals, action potential duration, and proarrhythmic effects.

E14

The ICH E14 calls for a dedicated trial in humans to establish the propensity of a non-antiarrhythmic compound to prolong the QT interval. Anti-arrhythmic compounds are exempt from this requirement as their mechanism of action sometimes prolongs the QT interval.

The endpoint of an E14 trial will be to exclude a QTcF prolongation of 10 msecs as measured by the upper-bound of the one-sided 95% confidence interval. This does not mean, however, that a 9.9 millisecond change will be deemed safe. The overall profile of the drug will be taken into account when looking at approvability.

The design of an E14 trial will usually be either a parallel or crossover trial. The study is generally four arms: placebo, therapeutic, supratherapeutic, and positive control. Some sponsors have chosen to reduce the trial size by eliminating the therapeutic arm. This design should be discussed with the regulatory bodies before starting the study.

The QT dilemma. In 1981, approval was granted to market Seldane (terfenadine), a non-sedating anti-histamine for the treatment of allergic rhinitis. Reportedly, the number of prescriptions for terfenadine products reached 16 million in 1991.

Among the mounting evidence of terfenadine-related cardiac-toxicity, in a 1993 article Honig showed terfenadine had a modest effect on the Bazett's corrected QTc interval (QTcB) of the surface ECG. The effect, a modest prolongation of the QTcB of approximately 8 msecs only was increased to approximately 82 msecs with the administration of ketoconazole, an anti-fungal agent that blocked the CYP450 enzyme, the same pathway that cleared terfenadine from the body.

In 1998, the US FDA requested the withdrawal of terfenadine from the market. This helped spur the regulators to demand increased scrutiny when reading ECGs. Regulations came in two forms: technical and study requirements.

The deaths "caused" by the interaction of terfenadine and ketoconazole were due to a rare, sometimes-fatal, polymorphic ventricular tachycardia called "torsade de pointes." The name was coined in 1961 by François Dessertenne, who described this as "twisting of the pointes." On surface ECGs, because this is a rare finding, never detected during Phase I to Phase III trials that include a limited number of subjects and patients, prolongation of the QT interval still acts as a surrogate for the possibility of TdP, despite its well-known limitations. As said above, anti-arrhythmics that are developed to control cardiac arrhythmias and that have been proved to be "neutral" regarding total mortality rates in large multicenter clinical trials are QT prolongers (amiodarone, dronedarone, dofetilide).

Since the ICH E14 was released, changes in the industry standard have occurred. It is well known that other "channelopathies" (impairment of cardiac ion channels function) can occur with pharmaceutical products. To help study these effects, the initial thorough QT evaluation has been expanded to the so-called "TECG" trial. The TECG trial looks at the PR-interval and QRS duration in addition to the QT interval.

It is also well recognized using S7B core battery tests that some compounds affect more than a single ion channel in the heart, namely the potassium, the sodium, and the calcium cardiac currents. These are known as "multiple ion-channel blockers." One such compound is Ranexa (ranolazine). When being developed by CV Therapeutics, the FDA was very interested in the safety of such a drug. CV Therapeutics had to demonstrate the drug was safe given the multiple ion channel blocking properties. Extensive work was undertaken to prove cardiac safety. In 2006, CV Therapeutics was granted permission to market the compound. The label includes a warning about prolongation of the QTc.

Another multiple ion-channel blocker was recently removed from the market in November, 2010. Propoxyphene was deemed unsafe by the FDA and the NDA holder (Xanodyne) was asked to voluntarily remove the compound from the market. The rationale given by the FDA: "The results of the new study showed that when propoxyphene was taken at therapeutic doses, there were significant changes to the electrical activity of the heart: prolonged PR interval, widened QRS complex, and prolonged QT interval. These changes, which can be seen on an electrocardiogram (EKG), can increase the risk for serious abnormal heart rhythms."

Technical requirements. Before the ICH E14, "acceptable" ECG collection included using a standard print-out of the ECG wave form on thermal paper. Measurements were made by hand using either an ECG ruler or calipers on raw beats. Representative beats (median waveforms) were generally not available.

Concurrent with the development of the E14 was the development of an HL7 standard for ECG waveforms. This calls for digital collection and measurement of the ECG waveform.

The HL7 standard is the first industry-wide requirement for a common storage format for ECGs. Prior to this effort, ECGs stored using one manufacturers' format were not readable on a common platform as ECG stored in another manufacturers' format. Therefore, the US FDA could not easily compare trial data between trials. ECG core laboratories can now submit data in this format and the data is readable in widely available "viewers." This is commonly referred to and the "XML format."

The "ECG warehouse" accepts these data for review by the agency. It is hoped that the wealth of data can help the FDA and scientific community determine whether or not the surface ECG has information that can help determine whether or not a drug has a potential to cause TdP.

QT vs. plasma concentration curves. The C-QT plots the dependent variable, baseline, and placebo subtracted QT-interval against the plasma concentration of the study drug or metabolite. This type of analysis has become mandatory for TQT trials and can be used in early phase development to give the drug developer an early indication as to whether or not any repolarization liability exists.

Generally, a linear mixed-effects model is used for this analysis. The slope of the regression line can give an indication of the amount of change in the QT-interval per unit of concentration. Other variables, such as the PR-interval or QRS duration can also be used in this model. This data is not only useful for non-cardiovascular compounds, but they can yield valuable information for all compounds that have a QT-reporting requirement, including oncology compounds.

Oncology compounds

In some cases, and as recognized in the E14, a TQT trial is not feasible due to ethical or practicality reasons. For example, many neoplastic agents cannot be given to normal healthy volunteers due to toxicities. Nor can a supratherapeutic dose be given for the same reason. Many sponsors and primary investigators believe that a positive control and placebo cannot be given because it would be unethical. What is left is a single arm QT trial. This is the so-called "Intensive QT Trial."

In addition to the QT-interval, other cardiac toxicities are important in these compounds. Some of these drugs can cause cardiac muscle damage or reduce the left ventricular ejection fraction. One of the tests used to find cardiac muscle damage is a troponin assay. Other techniques, such as echocardiography and imaging can be used to look at left ventricular ejection fraction (LVEF) as a surrogate for cardiac toxicity.

Troponins. Especially in clinical trials with anti-neoplastic agents, cardiac troponins are monitored as a measure of cardiac damage. There are three main varieties of cardiac troponins: C, I, and T. Each of these measures a different damage process in the heart. In cardiac safety trials, Troponin I (TnI) and Troponin T (TnT) are the most widely used.

Echocardiography and imaging. The LVEF can be assessed using a multi-gated acquisition (MUGA) or echocardiography. This is usually done both pre-treatment and the end of every other cycle of treatment for cancer patients.

Some compounds can cause structural changes in the heart. The product, fen-phen, combined fenflouramine and phentermine as an aid for weight loss. In 1995, the FDA asked American Home Products to remove the products from the market. It was recognized that fen-phen caused valvulopathies in a subgroup of patients. It was recommended that some patients using the drug have an echocardiogram to see if there was damage to the heart.

Not wanting to see another problem such as occurred with fen-phen, Arena Pharmaceuticals was asked by the FDA to conduct echocardiograms in the clinical development program of APD356. The Behavioral modification in Overweight and Obesity Management (BLOOM) trial, Arena conducted echocardiograms in approximately 3,200 patients.

Another use for echocardiography is to estimate the LVEF. The LVEF can be compromised in certain trial with cardiotoxic compounds. These include some antineoplastic agents that have inotropic effects on the myocardium.

Some investigative sites do not have access to or expertise in echocardiography. Therefore, the LVEF can be calculated by using a MUGA scan. Since these are mainly safety endpoints and usually not an efficacy endpoint, the combination of echocardiography and MUGA for LVEF measurements is common, but not interchangable within the same patient.

Diabetes

Cardiac safety is also a concern when developing treatments for diabetes, as evidenced by the issues seen with pioglitizone. In the warning label of this drug, there is a mention of causing or worsening heart failure.

Long-term trials in diabetes patient populations have demonstrated beneficial effects of glycemic control and lowering of HbA1c on both "macrovascular" outcomes and "microvascular" complications. Therefore, HbA1c remains an acceptable primary efficacy endpoint for approval of diabetes compounds.

Diabetes mellitus being associated with an elevated risk of cardiovascular disease, the FDA Endocrinologic and Metabolic Drugs Advisory Committee has recommended in 2008 that the cardiovascular risk should be now thoroughly addressed during drug development. In particular, "To establish the safety of a new antidiabetic therapy to treat type 2 diabetes, sponsors should demonstrate that the therapy will not result in an unacceptable increase in cardiovascular risk" (Guidance for Industry, "Diabetes Mellitus," December 2008).

The new responsibilities regarding cardiovascular safety for drug developers include:

• Safety committees to prospectively adjudicate cardiovascular events during all Phase II and Phase III trials. The mandatory cardiovascular events in the guidance are: cardiovascular mortality, myocardial infarction, stroke, hospitalization for acute coronary syndrome, urgent revascularization procedures.

• Ensure that clinical trials (both Phase II and Phase III) designs will allow conduct of meta-analysis and provide a detailed description of the statistical methodology for meta-analysis.

• Enroll patients at high risk of cardiovascular events (advanced disease, elderly patients, and patients with renal impairment.)

Conclusions

While the ECG is still the most important and easiest test used to assess cardiac safety in clinical trials, changes in regulations and technology have changed the way the ECG is both acquired and measured in clinical trials. Other cardiac safety tests are also important and have become part of an overall assessment of cardiac safety.

Timothy Callahan, PhD,* is Chief Scientific Officer, e-mail: tcallahan@biomedsys.com, and Pierre Maison-Blanche is Chief Medical Officer, both of Biomedical Systems, 77 Progress Pkwy, Maryland Heights, MO.

*To whom all correspondence should be addressed.