Editors' Pick 2010

Risk/safety assessment

One of the recurring themes in this year's BJCP is the assessment of risk. The adage that ‘every drug is a poison’ should be foremost in the minds of every prescriber, or clinical research physician, but at what point does one call a halt to proceedings and declare that the harms outweigh any potential benefit? Sibille and colleagues at the French Club Phase I point out the lack of any consistent stopping rules in early phase studies, and the difficulties in applying toxicity grading systems from phase I/II/III cancer chemotherapy trials to studies in healthy volunteers [1]. Phase I participants clearly start at different baseline health levels, and have much less to gain (usually only a modest financial reward), as compared with cancer patients seeking a life-saving cure. Sibille et al. have started the ball rolling with their ‘points to consider’ in the development of safety stopping rules for phase I trials in healthy volunteers. These early proposals are by no means definitive, but will hopefully stimulate debate perhaps leading to some agreement on deciding at what point the ‘fat lady begins to sing’.

The subjective nature of assessing risk is of course a challenge to practitioners of evidence-based medicine, not least for those of us who have spent many hours debating real or imagined abnormalities in the QT intervals of patients taking antipsychotic and antidepressant drugs. The huge intra- and inter-individual variability, and the influence of heart rate and electrolyte concentrations make validation of any case definition difficult, and there is an urgent need to clarify the risk thresholds associated with particular drug-related QT/QT_c abnormalities. Waring and colleagues have retrospectively tested a new nomogram on hundreds of electrocardiograms from patients presenting with antidepressant overdose, and succeeded in detecting variation in QT_c effects amongst different antidepressants [2]. Again, like the ‘points to consider’ from Club Phase I, this work has not yielded ‘the’ definitive answer, but indicates it is time to shed the gloves, and knuckle down to the serious work of a prospective clinical validation study that will genuinely help clinicians quantify risk of arrhythmia according to prolongation of QT_c interval.

We recognize, of course, that the new broom does not necessarily sweep clean (or cleaner), and some proposed new tools will eventually fall by the wayside if they add little to our existing knowledge. Clinical pharmacology and the evaluation of new medicines cannot exist without adequate methodology. Some years ago we appealed for review papers about methods to study drug effects quantitatively. This appeal has not yet yielded a flood of papers but there has been a gratifying increase during 2010 [1, 3–9] and we hope that we will be able to refer to several more next year in our Editors' pick for 2011. While clinical pharmacologists have been closely involved for decades in the evaluation and monitoring of adverse drug reactions, a far more recent buzzword has been ‘patient safety’. So what does the new reporting scheme from the National Patient Safety Agency (NPSA) add to the established arena of pharmacovigilance? Not very much (other than perhaps an extra layer of paperwork), according to Davies and colleagues who used this scheme on over 700 adverse drug reactions, and found that only 2% would have met the threshold for detailed investigations according to NPSA guidance [10]. Indeed well over 95% of the adverse events would have been classified as low to moderate importance to patients, and of limited significance to the healthcare organization. This is reassuring, but probably means that deploying yet another method of reporting adverse reactions in parallel with established pharmacovigilance systems will be unproductive and could lead to confusion amongst physicians.

Topical debates: lifestyle drugs and generic substitution

‘Lifestyle’ drugs have been much in the news in 2010, with the advent in pubs and clubs of novel centrally active non-therapeutic drugs, some of them not covered by existing legislation. Even so, the use of these upstarts is eclipsed by more ‘traditional’ recreational drugs such as cocaine. In a themed issue of the Journal, three articles explored the molecular basis, clinical epidemiology and therapeutic approach to cocaine-induced dysrhythmias, one of the most important complications of cocaine intoxication [11–13]. Despite the importance of cocaine-induced dysrhythmias in cocaine-associated morbidity and mortality, Wood & Dargan make the point that statistics on the true prevalence of such arrhythmias are seriously deficient, and that, faced with a patient presenting with an arrhythmia, doctors are much less likely to ask about cocaine use than about classical cardiovascular risk factors such as smoking, diabetes and hypercholesterolaemia [12]. This goes some way to explain the lack of published data in this area. O'Leary & Hancox described the effects of cocaine on ion channels and how these, in conjunction with its indirect sympathomimetic effects, provides a substrate for arrhythmogenesis [11]. Hoffman reviewed the principles of treatment of cocaine-induced dysrhythmias [13], explaining the need to avoid class Ia and Ic anti-arrhythmics as well as β-adrenoceptor blockers. He also outlined the justification for novel (and in some cases unconventional) therapeutic approaches including intravenous administration of hypertonic sodium bicarbonate, magnesium chloride, lidocaine, rapidly eliminated α-adrenoceptor antagonists (e.g. phentolamine), and/or calcium channel blockers (e.g. nicardipine).

Another very topical debate is the place of generic and therapeutic substitution of drugs, and this has received some attention in the pages of the Journal. Duerden & Hughes [14] argued that generic substitution (switching from a branded to a generic product) is generally to be recommended, since the generic equivalents are usually cheaper (often considerably so) whilst being therapeutically equivalent to the branded product (although there are some notable exceptions, including some insulin preparations and some slow release preparations of antiepileptic drugs and of diltiazem). On the other hand, therapeutic substitution (switching to a cheaper, pharmacologically distinct but apparently equivalent, product, usually within the same drug class) is more contentious, and arguments for and against such substitution revolve around patient choice and evidence base on the one hand vs. potential cost saving on the other. Johnston and colleagues make the point [15] that medications are an easily identifiable target when healthcare costs are under review. In this context drug switching with the aim of reducing healthcare costs is common in the setting of hypertension treatment. Despite this widespread practice, the economic consequences may not always be as clear as envisaged by those advocating such switching (for example, drug switching may entail additional clinic visits and/or laboratory tests, and there exists also the possibility of reduced efficacy or of causing adverse effects which may necessitate additional clinic/surgery visits or even hospitalization). Indeed, the clinical and economic evidence for the usefulness and place of drug switching is weak, and properly controlled large outcome studies are needed.

Trifunctional antibodies: the leading edge in targeted cancer immunotherapy?

The Journal has enthusiastically embraced the explosion in scientific knowledge and clinical development (with uses ranging from multiple sclerosis to cancer) of targeted therapy with monoclonal antibodies [16]. A PubMed search performed during November 2010 using the term ‘monoclonal antibodies’ yielded 190 281 citations, whereas a search using the term ‘trifunctional antibodies’ yielded only 99 citations. Trifunctional antibodies are a new class of bispecific monoclonal antibodies with two different antigen binding sites, one in each of their hypervariable Fab regions and a Fc gamma receptor binding site in the functional constant (Fc) region. They have the capability to redirect T cells and accessory cells (e.g. macrophages, dendritic cells and natural killer cells) to a focus of disease such as a tumour deposit. In preclinical studies, trifunctional antibodies have demonstrated a capacity to activate these different immune effector cells and trigger a complex anti-tumour immune response [16].

During 2010 the Journal published a report by Ruf et al. of catumaxomab pharmacokinetics in cancer patients [17]. Catumaxomab is a trifunctional rat/murine hybrid antibody that has bi-specificity for the epithelial cell adhesion molecule, EpCAM, and the T-cell antigen, CD3. It also has a Fc region that preferentially binds to Fc-gamma receptors (FcgR) I/al and III. Ruf and coworkers report an open label, intraperitoneal dose escalation study in cancer patients. Thirteen cancer patients with symptomatic malignant ascites from EpCAM positive tumours, were treated with ascending doses of 10, 20, 50, and 150 µg catumaxomab infused intraperitoneally (i.p.) on days 0, 3, 6 or 7 and 10. Anti-catumaxomab antibody development was monitored. Ten out of 13 patients were evaluable for pharmacokinetic analysis. Catumaxomab became increasingly concentrated in ascites during treatment and effective anti-tumour concentrations (100–1000 pg ml^–1 range) persisted for 3 days even after the lowest (10 µg) dose administered. The systemic catumaxomab exposure was low (<1%), with a median maximal plasma concentration (C_max) of 403 pg ml^–1. The median plasma elimination half-life was 2.1 days. Catumaxomab remained immunologically active even after several days in the circulation. All patients developed anti-drug antibodies, but not until the last intraperitoneal infusion. High observed inter-individual variability and low systemic exposure may be explained by the inverse correlation between tumour burden, effector cell numbers and systemic antibody bioavailability. Based on the high and effective local catumaxomab concentrations and low systemic exposure, the authors confirmed the potential utility of this agent and dosing scheme in patients with malignant ascites.

Further studies of catumaxomab in cancer patients with malignant pleural effusions were reported with promising efficacy [18]. A phase II/III study in patients with epithelial cancers (EpCAM positive) and malignant ascites has also been published [19]. In this randomized study, 268 epithelial cancer patients with malignant ascites received either paracentesis alone or paracentesis plus catumaxomab given i.p. in the dose regimen reported by Ruf et al. [17]. The catuxomab-treated group had fewer symptoms and signs of ascites, a longer paracentesis-free survival time 46 vs. 11 days (P = 0.0001), and, in patients with gastric cancer, a longer median overall survival of 71 days vs. 43 days (P = 0.031). Adverse events were frequent, but manageable and generally reversible and mainly related to catumaxomab-mediated cytokine release. These data clearly contributed to the EMEA's approval of catumaxomab for malignant ascites. Other novel trifunctional antibodies are in early clinical development, e.g. ertumaxomab (Anti-HER2 x Anti-CD3) for metastatic HER2 positive breast cancer [20], and potential applications for this novel antibody class continue to expand.

Of rat poison and WARF!

Little did the Wisconsin Alumni Research Foundation appreciate that more than 50 years after their coumarin-based anticoagulant rodenticide warfarin was licensed for human use, it would remain the subject of much clinical pharmacological research, even though newer anticoagulants are on the prowl. Some clinical pharmacologists specialize in old drugs! Four publications in the Journal in 2010 highlight new aspects of the clinical pharmacology of warfarin under the themes of drug–drug interactions with non prescription and herbal medicines and pharmacogenetics and ethnicity.

Abdul and colleagues investigated pharmacokinetic and pharmacodynamic interactions of Echinacea and policosanol with warfarin in healthy subjects [21]. Both the more commonly used Echinaceas, purpurea and augustifolia inhibit or induce the function of many CYP enzymes including inhibiting CYP2C9, the major CYP involved in the metabolism of the more active S(–) isomer of warfarin. Abdul and colleagues conducted a well-designed study in healthy subjects and showed that E. purpurea and augustifolia increased the exposure of S-(–)-warfarin by about 10% (90% CI 3, 18%) with no effect on the less active R-(+)-warfarin (which is mainly metabolized by CYP3A4), but had no effect on international normalized ratio (INR) measurements, which are the standard pharmacodynamic measure of the effect of warfarin. This small increase in exposure is very unlikely to have major clinical consequences in likely recipients of warfarin, including elderly patients. Thus, this herbal medicine could, if appropriate, be safely used with warfarin. Importantly the outcome of the study supports previous in vitro metabolic studies.

Fang and colleagues studied the time-dependent inhibition of CYP3A4 and CYP2C9 by noscapine and showed that this can explain the interaction of noscapine with warfarin [22]. Several case reports have documented adverse effects of warfarin (bleeding, elevated INR) in patients who take the over-the-counter antitussive noscapine. Using human liver microsomes, Fang and colleagues found that noscapine is a noncompetitive inhibitor of CYP2C9 and a competitive inhibitor of CYP3A4 with IC₅₀ values of about 10 µm. Using a validated in vitro-in vivo equation to predict drug–drug interactions via inhibition of metabolism, they predicted that noscapine would increase the exposure to warfarin by less than 2%. However, the inhibition of CYP2C9 by noscapine was time-dependent reaching a maximum of 50% at 30 min. When this factor (inactivation mechanism) was incorporated into their model, they predicted that S-(–)-warfarin exposure would increase by over 40% in patients taking noscapine. Thus a time-dependent inhibition mechanism could explain case reports of warfarin toxicity in patients also taking noscapine. The study highlights the increasingly recognized importance of over-the-counter medicines in adverse interactions with prescription medicines.

Lee and colleagues showed that warfarin sensitivity in mechanical heart valve replacement patients is associated with a haplotype of CYP2C9. CYP2C9*2 and *3 are the most common loss-of-function alleles in Caucasians and, together with VKORC1 variant allele(s) are associated with low warfarin dosage requirements [23]. In Asians, warfarin dosage requirements are lower than those for Caucasians and this is attributed almost solely to the VKORC1 genotype and not CYP2C9 as variant *2 and *3 alleles of the latter are believed to occur infrequently in Asian subjects. Lee and colleagues resequenced the CYP2C9 gene in 50 Korean subjects and found 39 variant alleles of which seven were new but rare. One specific haplotype containing promoter and intrinsic single nucleotide polymorphisms (SNPs) that had been previously identified in a Japanese cohort, CYP2C91e, was found in 15% of a small sample of patients who required low warfarin doses. They concluded that the development of a warfarin dosage algorithm in the Asian population may require determination of this type of CYP2C9 haplotype rather than single CYP2C9*2 and *3 evaluation.

Cen and colleagues studied Han Chinese patients with mechanical heart valve replacements and found that CYP4F2 rs2108622 is a minor but significant genetic determinant of warfarin dose requirement [24]. Apart from CYP2C9 and VKORC1 genotypes, interest a few years ago centred on a CYP4F2 variant allele (frequency ∼25%), as it could contribute substantially to warfarin dose requirements as this CYP is a vitamin K1 oxidase; that is the ‘missing’ 40%. Unfortunately, this was not the case for Caucasians. In what appears to be the first study in Han Chinese, the variant had an almost identical frequency to that in Caucasians and its contribution to warfarin dose requirements was only 4% in a population of 222 Han Chinese patients with mechanical heart valve replacement. This study, although having a negative finding, is important as it was conducted using a sound mechanistic hypothesis that Han Chinese would have a lower frequency of the variant allele, thus contributing to the low dose requirements of warfarin in this ethnic group.