Precision Dosing: The Clinical Pharmacology of Goldilocks

Abstract

If you had to summarise clinical pharmacology in one word, that word could be dose. So much of what clinical pharmacologists do is about understanding how best to dose drugs to maximise their benefit and/or minimise the harm they can do. For most drugs today, there is a single dose level recommended for most patients with the exception of well characterised situations such as different ages or ethnicities, the presence of organ failure, genetic differences, or the presence of interacting concomitant medication. In these cases, the drug’s pharmacokinetics/pharmacodynamics, usually its concentrations in plasma and site of action, are altered, and a dose adjustment is recommended to compensate for this change. This is a first step along the pathway to fully individualised precision dosing, that is already how some drugs, for example insulin, must be dosed as there is no one dose that fits all patients. The theme of this issue of CPT is precision dosing. The article topics range from making it easier to identify and manage subpopulations that require a different dose to when and how to support truly individualised dosing. Considering first improving the management of dosing in subpopulations for which an imprecise drug dose may be recommended. Imprecise drug dosing originates from gaps in knowledge, which leads to poor clinical outcomes due to sub- or supratherapeutic exposure and increased clinical care costs. With regard to the effect of interacting drugs, prescribing information may advise dose adjustments in the presence of strong enzyme inhibitors or inducers with lesser changes for less potent inhibitors or inducers to ensure proper dosing. There is a need here for better standards for which drugs are strong, moderate, or weak inducers or inhibitors to ensure consistent clinical interpretation, appropriate categorization of each inhibitor and inducer, and proper dosing. In this issue of CPT, two papers provide important contributions to meet this need.1, 2 The work done by Yang et al.2 emphasizes the need for collaboration between researchers in academia, industry, and regulatory agencies to maintain consistency in the interpretation and clinical management of drug-drug interaction and ensure proper dosing to patients. Recent advances in genomic medicine and applications of liquid biopsy represent another way of improving dosing recommendation in subpopulations, or some individuals, by increasing the ability to characterise the level of enzyme activity in individual patients and offering the potential to guide dose selection based on each patient’s drug metabolising profile.3 The induction of transporters is also an essential consideration for the pharmacokinetics and dosing of some drugs.4 Identification and implementation of dosing changes based on patient genotype can be challenging to implement; however, it is hoped the decade of experience at Vanderbilt University Medical Center will be helpful in overcoming challenges that hinder the smooth implementation of genetics in guiding pharmacotherapeutic dosing recommendation.5 Likewise, consortia’s combined expertise, such as PharmGKB,6 serves as a freely available comprehensive resource that facilitates the implementation of pharmacogenetics in clinical practice, including through CPIC guidelines,7 which can be hugely influential in guiding dosing recommendations across different patients. Whilst the impacts of co-variates are often studied in isolation, patients often have multiple co-variates together, for example organ failure and an interacting concomitant medication. One route to help these patients is to use modelling to understand the interaction between co-variates and guide dosing for these more complex situations.8 Indeed, real-world patients are often different from the patients in the clinical trials on which the drug dosing recommendation is based. For real-world patients, model-informed precision dosing can be a useful tool to guide the selection of the right dose in many such cases.9 The variability intrinsic to patients in the real world with multiple combinations of relevant co-variates plus the often poorly studied influence of variation in disease severity or disease pathogenesis or truly unknown factors presents significant challenges to successful precision dosing for real-world patients. Such variation in the relationship of exposure to effect may be even more important than the impact of variation in drug concentrations after a given dose. As with other complex multi-variate challenges, machine learning approaches are being developed to tackle this problem, and they bring the possibility of determining the optimal doses from real-world data after drug approval.10 What about those drugs where individualised dosing is required? This is where the Goldilocks principle should be introduced (Figure 1). How should Goldilocks treat the panic attack that she must have suffered when the bears came home and found her asleep in their house? Did she use the dose for the father or mother bear, or did she find that baby bear’s dose suited her best? As an experimentalist, let’s hope she learned from her earlier experience and took baby bear’s dose first, thereby avoiding the sedative effects of too high a dose. It would also give her the option to titrate up to the desired effect if needed. This is a recognised precision dosing strategy for some drugs and one that regulatory authorities support and indeed encourage when appropriate.11 It is encouraging that such strategies are used but also noteworthy that there are many suitable cases where they are not. Given the urgency of Goldilocks’ situation though, there was probably limited time for her to explore dose titration so it is to be hoped she adopted a model-informed approach to identify the right dose for her, based on her previous experience of how the differences in the three bears preferences affected her. Why are precision dosing strategies not used more widely? For a detailed answer, the reader is encouraged to read the article by Darwich et al.12 In brief, there is added complexity from precision dosing because it usually needs additional tests, such as drug concentrations or a drug effect biomarker, and these tests require interpretation and additional action. This all increases the demands on patient and physician time and extra costs for the payers. However, technology is being developed to help.13 Plasma exposure-based dose adjustments are widely used to maximise the chance of vancomycin efficacy and avoid its serious adverse effects. Automated tools to select the most suitable pharmacokinetic model improve the clinical benefit from exposure guided vancomycin dosing.14 Setting up an advisory service to guide prescribers15 or the use of modelling tools integrated seamlessly into the electronic medical record and electronic prescribing system16 removes much of the burden from the treating clinician to make implementation much easier. Further advances in, and the broader use of, such clinical decision support tools will help enable wider uptake for those drugs that can benefit from precision dosing. Additional demonstrations of drug-disease combinations, where precision dosing is feasible and shows clear clinical benefit, will also encourage wider uptake and investigation of the benefits for other drugs and diseases. This issue includes examples of model-informed precision dosing to improve the clinical benefit for sickle cell anaemia patients treated with hydroxyurea17 and optimize pharmacological care in neonates with opioid withdrawal syndrome,18 a significant clinical need during the current opioid abuse crisis in the United States. Current drug labels are often inadequate to support precision dosing for many drugs, and some relatively simple changes could help increase the acceptance and use of precision dosing.19 Regulatory authorities recognise the value that precision dosing can bring for some drugs11 and a recent FDA workshop brought together many interested parties from academia, industry, patients, and the health authorities, delivering some critical viewpoints and ideas to increase uptake.20 Ideally, model-informed precision dosing will be built into drug development for those drugs where it is most likely to be useful.21 This will take several years to have a clinical effect but, for drugs already on the market, there is no need to wait. There is already an opportunity for model-informed precision dosing leveraging real-world data9 perhaps combined with novel machine learning techniques10 and a comprehensive framework for enabling precision dosing for approved drugs and those newly in development is presented in this issue.22 The current COVID pandemic represents an opportunity for precision dosing,23 and it is to be hoped the urgency of the situation will bring the key stakeholders together to find ways to make precision dosing a broader reality that will subsequently also benefit many other diseases and drugs. In conclusion, precision dosing has great potential to improve patients’ health care by maximizing therapeutic benefits while minimizing risks associated with drug use. Although precision dosing has a potential impact that is substantial for some drugs, it may not be necessary, feasible, or cost-effective to implement for every drug or therapeutic class. Akin to the concept of precision dosing, Goldilocks used a model-informed approach that incorporates prior knowledge related to how the differences in the three bears’ preferences affected her. Using this model-informed precision dosing approach helped Goldilocks to find the drug strength that was “just right” for her, and she lived happily ever after!

Topics & Concepts