Past, present and future of drug safety: Editorial

Abstract

For most of recorded history, drug therapy was primarily conducted based on traditional beliefs and anecdotal reports with agents that were often ineffective, frequently not expensive and frequently fairly innocuous. This changed dramatically with the Therapeutic Revolution when the discovery in 1935 that bacterial infection could be cured with specific therapy produced paradigm changing effects on health care and medical practice.1 The excitement over the use of sulfanilamide therapy was however soon muted with the Elixir of Sulfanilamide Tragedy.2 Over the following decades, as the benefits of drug therapy have increasingly evolved, so has our appreciation of the potential for drug harms (Figure 1). Given that new opportunities and challenges in drug safety field are constantly arising, some current insights and future directions deserve to be outlined according to the evidence-based medicine approach. Data arising from meta-analyses of observational studies reported that 5%–10% of hospital admissions were caused by adverse drug reactions and that approximately half of those hospitalizations could be prevented. Thus, in this issue of the British Journal of Clinical Pharmacology, we are considering the theme of drug safety with a number of manuscripts exploring drug harms and their detection. An already well-established but still growing area in drug safety is pharmacogenetics (PGx), which has the potential as a component of medicine optimization to enhance drug safety, improve drug efficacy and reduce the risk of toxicity. In this issue, Medwid and Kim share the Canadian experience on broad implementation of pre-emptive dihydropyrimidine dehydrogenase (DPYD) genotyping in patients taking fluoropyrimidines, providing an overview on current models of PGx, what are clinically relevant PGx tests and their associated drugs and challenges moving forward, for example, lack of diversity in PGx studies. Future directions in PGx are also discussed with a focus on new clinical targets, methodologies and analysis approaches.3 Bearing this in mind, Nyangwara et al provide insights on the prevalence of doxorubicin-induced cardiotoxicity in black Zimbabwean breast cancer cohort and attempt to evaluate the predictive value of genetic biomarkers for cardiotoxicity.4 Mechanistic insight into drug toxicity can be obtained also by using the data from metabolomic analysis to identify a drug candidate's off-targets; thus, metabolomics may also help to lay another pathway toward drug safety. Lim et al have identified glycine, hippuric acid sulfate, 3-hydroxydecanedioc acid and suberate as predictive biomarkers of predisposition to cisplatin-induced acute kidney injury (AKI) and speculated that metabolites indicative of mitochondrial dysfunction may serve as early markers of subclinical AKI.5 A ‘big data’ and modern approach to pharmacovigilance was well presented by Jacob et al. They have provided a bioinformatics-guided disproportionality analysis of sevoflurane-induced nephrogenic diabetes insipidus using the FDA Adverse Event Reporting System database. With a ROR of 76.012 (95% CI: 44.67–129.35) and PRR of 75.72 (χ2: 934.688), positive signals for sevoflurane were observed; of note, of the 14 cases, 50% required hospitalization, and 14% were fatal.6 Now something about the roots of this themed issue … with the number of substances on the market, it is conceivable that mistakes (e.g. the wrong dose) can be made when practitioners prescribe or dispense drugs, emphasizing the importance of taking steps to minimize those preventable medication errors and improve drug safety. As an example of this, Zieting et al reported what can happen and what should be done if a patient with Hodgkin lymphoma receives a 10 times higher dose of bleomycin (150 USP) than that usually used as part of the doxorubicin, bleomycin, vindesine and dacarbazine (ABVD) protocol. The authors review potential toxicities, clinical management and the selection of further chemotherapy after bleomycin overdose. Of import, besides management measures, they also discuss how to prevent medication errors and how to apply this in other hospital settings.7 Clinical trials assess risk–benefit for drugs in well-defined and controlled setting, given for a relatively brief period of time and to a small and homogeneous population with robust monitoring. Thus, to obtain a clear and wider picture (more extensive and for specific subpopulations) on drug safety profile, post-marketing (phase IV) studies and real-world evidence are of critical importance. As well, alternate data sources that could be used to support observational pharmacovigilance studies of treatment during pregnancy and breastfeeding are multi-year administrative claims datasets, which provide cross-sectional and longitudinal data and consequently can inform maternal and fetal/newborn safety. In the present themed issue, we have covered both domains. Rong et al explored avapritinib-related AEs in real-world practice and speculated that clinicians should be particularly careful when prescribing avapritinib to elderly male patients, especially within 30 days.8 In terms of drug safety concerns during pregnancy, Lavu et al have conducted a meta-analysis that demonstrated that pregnant patients exposed to antiseizure drug (especially in case of polytherapy) have a significantly increased risk of adverse fetal growth outcomes including small for gestational age and low birth weight compared to unexposed pregnant patients.9 Keeping in mind that the skin is the largest organ in the body, a ‘reflection of our inner health’, and the organ most frequently affected by drug reaction (up to 10% of hospitalized patients and 1%–3% of multimedicated patients), there are a number of manuscripts on this topic on drugs and the skin.10 Important insights from the literature are outlined in the review of cutaneous adverse drug reactions by Del Pozzo-Magaña and Liy-Wong and the review on pathophysiology of drug hypersensitivity by Elzagallaai and Rieder.10, 11 New considerations are provided by Martin et al who used the French pharmacovigilance database to identify drug-induced Sweet's syndrome cases. Among the 994 789 reports recorded in the database, they identified 136 cases of Sweet's syndrome with a total of 224 drugs mentioned as suspects (21.0% were antibacterials, 19.2% were antineoplastics, and 12.1% were immunosuppressants); the highest RORs were observed with bortezomib, azacitidine, pegfilgrastim, azathioprine and bendamustine.12 An additional case report by Ouni et al describes a case of toxic epidermal necrolysis due to pristinamycin, an oral streptogramin antibiotic with bactericidal activity against Gram-positive bacteria.13 Bearing this in mind, it is evident that adverse drug reactions and drug-related adverse events may be associated with medical malpractice, poor prescription practice, variable adherence, PKPD alteration or intentional overdosing.14, 15 Gaps in current data collection are also among the list of indirect factors; there is a responsibility of healthcare professionals to be accurate, timely and comprehensive in the description of adverse events and ‘bodyguard’ their patients. Given all this, it is clear that clinical pharmacologists need to continue to study drug safety to ensure the safe and effective use of drugs while clinicians should conduct punctual monitoring and always conduct risk–benefit analysis as part of rationale prescribing.16 An integrated, mechanistic evaluation of safety signals is essential, especially since the present/future of clinical safety should focus on prediction rather than remediation. In summary, for clinical pharmacologists these are good but challenging times to be a responsible ‘bodyguard of drug safety’. None. The authors declare no conflict of interest. Not applicable.

Topics & Concepts