Exceptional Circumstances Requiring Extended Database

| Home | | Pharmacovigilance |

Chapter: Pharmacovigilance: Withdrawal of Terodiline: A Tale of Two Toxicities

The ICH E1A guideline recognizes that a larger database and/or a longer period of exposure than usual may also be required in some circumstances.


The ICH E1A guideline recognizes that a larger database and/or a longer period of exposure than usual may also be required in some circumstances. To this end, it provides for exceptional circumstances when the harmonized general standards for clini-cal safety evaluation may not be applicable and an expanded database may be required. These excep-tions cover a diverse range of circumstances, and can best be discussed using drug-induced QT interval prolongation/torsade de pointes as an example. The approach is equally applicable to other rare but seri-ous adverse effects, such as clinical hepatotoxicity, gastro-intestinal haemorrhage, neutropenia and so on. Although there are a number of exceptional circum-stances specified in the guideline, six are particularly relevant to most NCEs.


Without doubt, any drug that shares a structural simi-larity with prenylamine is a candidate for an expanded clinical safety dataset, in order to better assess its potential to prolong the QT interval. Not surprisingly, terodiline, terfenadine, cisapride and pimozide all bear an obvious structural similarity to prenylamine, and would have called for an expanded clinical dataset to characterize their potential for QT interval prolon-gation and torsade de pointes. With regard to QT interval prolongation, many chemical classes have been implicated (Shah 2002; Aptula and Cronin, 2004; Aronov, 2005; Recanatini et al., 2005), and therefore a wide range of NCEs would require an expanded clinical dataset.


When an investigational drug is found in preclin-ical studies to block IKr or hERG channel and/or prolong the action potential, ICH E14 recommends that the clinical safety dataset focussing on ECG effects needs to be expanded, regardless of a negative ‘thorough QT/QTc study’ if the preclinical/clinical discrepancy cannot be explained. References have already been made to pharmacodynamic and pharma-cokinetic similarities between terodiline and preny-lamine. In retrospective preclinical studies conducted post-approval, prenylamine, terodiline, terfenadine, astemizole, pimozide, halofantrine, cisapride and levacetylmethadol have all been found to possess QT-prolonging properties, and would have called for an expanded clinical dataset had these studies been conducted prior to their approval. Focussed clinical studies with terodiline, albeit following its removal from the market, and other drugs confirmed that they had the potential to prolong the QT interval in man.


In compliance of the ICH E14 guideline, the clin-ical safety dataset needs to be expanded if ICH S7B-compliant in vivo studies are strongly positive, regardless of the status of the ‘thorough QT/QTc study’. The requirements for preclinical investiga-tions at the time of developing prenylamine were rudimentary. Information on findings from animal studies with prenylamine is now difficult to obtain. Although original preclinical studies with terodiline showed no effect on the QT interval in conscious dog or rat, ECG effects (including prolongation of the QT interval) were reported in anaesthetized cats. This finding in itself would have warranted further preclin-ical studies and an extended clinical safety database. Webster et al. (2001) have recently shown that terodi-line does induce QT prolongation in dogs and empha-sized that for compounds known to be clinical torsado-gens (terfenadine, terodiline, cisapride), there is little differentiation between the QT-prolonging and the clinically effective free plasma concentrations in man (<10-fold). This is reflective of their limited safety margins.


A range of ICH guidelines (ICH E1A, ICH E2E, ICH S7B and ICH E14) emphasize the need to take into account the pharmacological activities associated with other members of the same chemical or phar-macological class as the NCE under investigation. Therefore, this particular scenario requires that the safety database be expanded to exclude any class-related risks. Apart from prenylamine and lidoflazine, a number of other antianginal drugs such as bepridil, tedisamil, fendiline and aprindine have all been shown to prolong the QT interval and induce proarrhyth-mias. Therefore, during their clinical development, terodiline as well as any other antianginal drug would call for an expanded clinical safety database, for routinely evaluating their potential to prolong the QT interval. This is analogous to all non-steroidal anti-inflammatory drugs (NSAIDs) being evaluated for their gastro-intestinal toxicity. With regard to QT interval prolongation, many pharmacological classes have been implicated (Shah, 2002; Aptula and Cronin, 2004; Anson et al., 2005; Aronov, 2005; Recanatini et al., 2005), and therefore, again as stated above, a wide range of NCEs would require an expanded clin-ical dataset.

When discussing the ‘pharmacological class’ of a drug, the notion of its ‘therapeutic class’ deserves a comment. Following structural modifications of a lead compound or following the approval of a drug, it is often discovered to have more potent activity at a pharmacological target other than that intended orig-inally. Therefore, drugs are often intended for devel-opment in one specific therapeutic area but are later developed or used clinically in an entirely different therapeutic area. Thus, drugs frequently cross ‘ther-apeutic boundaries’ (Shah, 2002). Therefore, lack of a safety concern in drugs of a therapeutic class is not altogether wholly reassuring when developing another drug in the same therapeutic class – what really matters is the chemical or the pharmacologi-cal class. Terodiline itself was re-developed for use in a completely different therapeutic area (urinary incontinence) that was not associated with any proar-rhythmic risk. Terfenadine is another typical example. It was discovered through a central nervous system programme aimed at synthesizing new antipsychotic agents, but because of its more potent secondary phar-macological effects at the H1-antihistamine receptor, its development was diverted to market it as the first non-sedating H1-antihistamine. However, like other antipsychotic agents, it was sooner or later bound to attract regulatory attention because of the potential of antipsychotics-related chemical structures to have an effect on the QT interval. As an antihistamine, terfe-nadine remained a highly successful and popular drug until withdrawn, due to reports of torsade de pointes resulting from drug interactions. Sildenafil, originally intended for development as an antianginal drug, was developed instead for male erectile dysfunction, and it is not surprising that at high concentrations, it too has been shown to prolong cardiac repolarization by blocking the rapid component of the delayed rectifier potassium current (Geelen et al., 2000). At clinical doses, however, a significant effect on QT inter-val is most unlikely (Morganroth et al., 2004), espe-cially since the drug is used intermittently. However, its further development for use in pulmonary hyper-tension may present interesting dilemmas (Shah, 2005b).


Depending on whether a drug is a class III antiar-rhythmic drug or not, the frequency of QT inter-val prolongation and/or torsade de pointes can vary widely. For a number of antianginal or non-cardiac drugs, these are low frequency events associated with their use. It is therefore self-evident that an expanded clinical safety database would be required for a new antianginal drug. The size of the database would be determined by the preclinical data and the anticipated frequency of the event to be detected, as well as the confidence with which the risk is to be excluded. Since the risk of torsade de pointes is often as low as 1 in 10 000 or even lower, requirements for very large databases can be counter-productive to the extent that they delay the introduction of otherwise beneficial medicines to the market.


A dataset that is larger and/or of longer exposure may also be appropriate when a specific serious adverse event that represents an alert is observed unexpect-edly in early clinical trials. When the potency of an NCE to delay ventricular repolarization is high, signals are often detected during early clinical trials, frequently pharmacology studies in healthy volunteers or early dose-ranging studies in patients. Pimozide, for example, was found to prolong the QT interval in about 10% of the patients in one study in 1989. Similarly, halofantrine was also found to produce an effect on the QT interval during early clinical trials. This is especially important when an event is a ‘moving target’ depending on the presence of other risk factors, such as drug interactions or other inter-current events.

As it was, the clinical trials database on terodi-line was comparable with those for other contem-porary drugs intended for urinary incontinence. In retrospect, however, it was not large enough for a drug with its chemical and pharmacological pedi-gree. It had included 8 controlled = 229 and 6 uncontrolled =147 studies with a total popula-tion of 376 patients exposed to terodiline. Of these, 241 had received the drug for up to 1 month, and a further 39 for 2–3 months. Seventy-five patients had been treated for 4–12 months. In the aftermath of its withdrawal, a number of studies investigated the ECG effects of terodiline. Apart from the study by Thomas et al. (1995) referred to earlier, other stud-ies have shown that adequate ECG monitoring of the patients during clinical trials ought to have identified the proarrhythmic risk. In the study by Yoshihara et al. (1992) in 109 Japanese patients receiving 24 mg daily of terodiline for 4 weeks, side effects such as ortho-static hypotension and arrhythmia were observed, and these symptoms disappeared following discontinua-tion of the treatment. Of note is the prospective study by Stewart et al. (1992) in 8 elderly in-patients treated with terodiline for urinary incontinence. They found that after 7 days of treatment with 12.5 mg twice daily, terodiline significantly increased the QT interval by a mean of 29 ms and the QTc interval by 15 ms and decreased the resting heart rate by a mean of 6.7 beats per minute.

As a result of experiences with some of the estab-lished as well as newly introduced drugs, clinical trials programmes now usually include ECG monitor-ing in at least one or two large studies, particularly those investigating high doses or studying the effect of inhibition of drug elimination (e.g. drug interaction studies). Depending on the ECG findings from these ‘exploratory’ studies, the database may require expan-sion to address the proarrhythmic risk more fully.

Contact Us, Privacy Policy, Terms and Compliant, DMCA Policy and Compliant

TH 2019 - 2022 pharmacy180.com; Developed by Therithal info.