Although all areas of pharmacotherapy are equally important, managing some of them can be slightly more complicated when they involve administration of drugs with broad inter-individual pharmacokinetic variability, whose metabolic pathways are at times controlled by highly polymorphic enzymes such as CYP2D6 or CYP2C19. One of such areas where the management of drug therapy is often more complex is neuropsychiatry. The fact that neuropsychiatric drugs are characterized by a broad pharmacodynamic variability has hampered implementation of individualized management of such medicines.

It is not unusual for terms such as precision medicine and personalized dosing to be confused. On the one hand, precision medicine, defined in the standards of national health systems as the kind of medicine that “uses information on the genes, proteins and other characteristics of the condition affecting a person to establish the diagnosis or the treatment of the said disease”, allows selecting medication in such a way that a higher likelihood of obtaining a therapeutic response is theoretically possible. On the other hand, personalized medicine also ensures that the right medication is used at the right dose in any given patient. These two concepts are confused all too frequently.

Since the 1960's and 1970's, pharmacokinetic monitoring came to be gradually —if somewhat slowly— incorporated to clinical practice. Although drugs such as digoxin, theophylline and some antibiotics have been subjected to routine monitoring in some hospitals, most centers have failed to implement any surveillance measures. Recently, with the introduction of immunopharmacotherapy to specialties such as gastroenterology, dermatology, oncology and rheumatology, there seems to have been an increase in the clinical use of pharmacokinetics, although practitioners tend to confuse monitoring with quantification and application of a widely used algorithm.

Although the AGNP (Arbeitsgemeinschaft für Neuropsycopharmacologye und PharmaKopsychiatrie) group, made up of chemists, biochemists, pharmacists and psychiatrists, published its first consensus with evidence-based recommendations for pharmacokinetic monitoring in the realm of psychopharmacology back in 2004, application of such guidelines has been very limited1. Some hospitals do monitor clozapine, probably due to the restrictions on its use, but very few centers have extended their monitoring protocols to all antipsychotics, antidepressants, hypnotics and mood stabilizers.

At the same time, different platforms have been developed informing psychiatrists about the genotype of different enzyme isoforms involved in metabolic, transport and pharmacodynamic processes. Although such platforms are presented as precision medicine tools, the information they provide usually includes dosing recommendations, which gives rise to considerable ambiguity and even confusion as they are not in themselves useful tools to determine a patient's dosing schedule. For example, although a heterozygous genotype is associated with wide phenotypic variability, with drug clearance coefficients of variation as high as 65%, it has been abundantly demonstrated that one same dose cannot lead to the same outcome in all patients2. These platforms allow selection of the initial dosing regimen, albeit with a certain error margin, but they can under no circumstances be used as a basis for dosing individualization. Their dosing recommendations are usually inspired in the CPIC (Clinical Pharmacogenetics Implementation Consortium) guidelines3–5. Apart from genetic factors, there are other elements, i.e, anthropometry, comorbidity conditions, age, etc., that play a role in the determination of medicines’ serum or plasma concentrations. As shown by several studies, their potential benefit in terms of preventing toxicity and limiting adverse events has been reported in several studies on poor metabolizers. Use of these platforms should be made in due consideration of phenoconversion, defined as a mismatch between a given genotype and its functional expression (phenotype) as a result of an interaction with a drug, a food item or a natural product. In their study on venlafaxine, Preskorn et al.6 dwell extensively on this topic. It is therefore essential to carefully review the patient's entire medication to prevent phenoconversion phenomena that could alter dosing recommendations, i.e., inclusion of bupropion in a patient on venlafaxine7.

The debate that has arisen on the role of pharmacokinetics and pharmacodynamics in the management of narrow therapeutic index drugs prompts us to ask the key question in this systematic review: What evidence is there on the application of pharmacokinetic or pharmacodynamic criteria to dosing in the realm of neuropsychopharmacology?

Application of the inclusion and exclusion criteria yielded a total of 41 articles with a proper pharmacokinetics/pharmacogenetics balance. The reviewers carried out the selection in pairs, there being four reviewers the process was divided between two groups of two reviewers each.

Article selection

Figure 1 shows the process followed to review and select the articles included in this study.

Figure 1.

Process followed to review and select the articles included in this study.

(0,08MB).

Of the 899 articles initially selected, 39 were discarded after reading their title as they were duplicates of studies already included. The title and abstract screen resulted in 137 articles being exhaustively reviewed by the two pairs of reviewers. The majority of them were eventually excluded because they were association studies and did not provide any information on the efficacy of the dosing individualization techniques used. The screening process yielded a total of 31 studies, which were supplemented by another 10 articles, all of them of a pharmacogenetic nature, drawn from the literature analysis included in some of the papers selected. The final sample comprised 11 pharmacogenetic and 16 pharmacokinetic studies.

The present systematic review seeks to answer an apparently simple question: What are the clinical benefits of pharmacokinetics and pharmacogenetics for individualizing the dosing of psychotropic drugs? Contrary to what may be expected, the number of clinical studies on the subject is relatively scarce, which means that any conclusions drawn should be taken with caution. Over the last decade there has been a mushrooming of genetic studies aimed mainly at correlating genetic variants with exposure to psychoactive drugs. An example of this is an exhaustive review published in Molecular Psychiatry in 200437 on the genotype-phenotype relations that exist between the different antidepressants and antipsychotics and on different dosing modification proposals intended to compensate for differences in plasma concentrations. A metanalysis published in JAMA-Psychiatry in 202138 found a strong association between exposure to different psychotropic drugs and different CYP2C19 and CYP2D6 genotypes. These studies have led to the publication of several clinical guidelines such as the CPIC (Clinical Pharmacogenetics Implementation Consortium) guidelines, the DPWG (Dutch Pharmacogenetics Working Group) guidelines, the CPNDS (Canadian Pharmacogenomics Network for Drug Safety) guidelines, and the RNPGx (French National Network of Pharmacogenetics) guidelines39, among others. They have also resulted in the inclusion of the relevant information in the summary of product characteristics (SmPC) of drugs like aripiprazole. These clinical guidelines made recommendations both related to indications and dose modifications based on the CYP2C19 and CYP2D6 genotypes. Nonetheless, the information selected to make such recommendations comes from studies correlating genotypes with drug exposure, which is indicative that evidence based on clinical results is too scarce to make recommendations. The information contained in these guidelines makes reference to “potential risks”, suggesting that the metabolization of these drugs is subjected to the action of other genes such as CYP1A2, CYP2C9 and CYP3A4 as well as that of other environmental, epigenetic or dietary factors, comorbidities or concomitant medication40.

This systematic review responds to the need to analyze the clinical evidence for improving the efficacy and toxicity profile of psychotropic drugs when pharmacokinetics and pharmacogenetics are used for treatment individualization. Although clinical trials that use pharmacogenetics as a tool to individualize treatment with antipsychotics do not show significant differences regarding the effectiveness of treatment or the reduction of adverse events27,28, certain patients with a rapid or slow metabolizer CYP2D6 genotype treated with drugs metabolized by this pathway could benefit from a reduction of the dose at the beginning of their treatment27.

Data is less clear in the realm of clinical trials on antidepressants. It should be underscored that the nine clinical trials that use pharmacogenetics as a tool to select the right medication and individualize the dose to be administered to the patient tend to be based on different designs. None of the studies focuses on a specific antidepressant but rather they select the antidepressant to be used based on different tests and selection algorithms for each patient, which makes it difficult to draw any conclusions for any specific drug. They also use different genotyping, dose calculation and therapeutic indication algorithms such as Genecept®41, GeneSight®36, CNSDose®34, Neuropharmagen®33, and Neuro IDgenetix®32, in addition to the above mentioned CPIC guidelines. These platforms have been approved to support physicians in their decisions on both the indication and the dosing of psychotropic drugs, establishing usage alerts for the different drugs as a function of genotype. It must be said, however, that there is wide variability across such platforms, both in terms of the number of genes analyzed and therapeutic recommendations, which makes it difficult to compare the various studies analyzed in this review. Around 40 pharmacogenetic platforms are currently being used to provide information and recommendations on the use of psychotropic drugs, which is indicative of the wide variability that exists. The review performed in this paper only found clinical trials that used five of these platforms, with negative results having been obtained with Genecept®41 and GeneSight®36. Other public platforms such as Sequecnce2Script42 (sequence2script.com) make an attempt to provide more than just genotyping information, with their dose-calculation and indication algorithms including other aspects like the potential occurrence of phenoconversion as a result of the concomitant medication used.

Nevertheless, pharmacogenetic recommendations for individualizing the dosing of psychoactive drugs can only be useful if they are implemented before initiation of treatment as pharmacogenetic information ceases to be clinically useful if the treatment is already underway. Pharmacogenetic information therefore provides a one-time snapshot, with limited information to follow up patients and make dosing adjustments according to their changing clinical situation. This means that the evolution of a psychoactive drug's plasma levels during treatment does not depend on the genotype but rather on the environment that surrounds the patient and on their clinical situation. In other words, although the genotype does not change, plasma concentrations may do so.

As regards studies on the usefulness of clinical pharmacokinetics (Tables 3a and 3b), a total of 16 of these studies were selected after ensuring that they met al.l the inclusion criteria. Half of these studies (n = 8) reported on the effects of antipsychotics in patients diagnosed with schizophrenia and the other half on the effects of antidepressants typically used in subjects with major depression. One study22 analyzed subjects with obsessive-compulsive disorder and another24 evaluated the role of lamotrigine as a mood stabilizer in patients with bipolar disorder.

Many of the excluded studies were merely descriptive, with a significant amount of these describing the concentrations and metabolic quotients of a certain drug as a function of the polymorphisms of an enzyme isoform, generally the one most closely involved in the drug's metabolism. However, this did not result in the use of distinctive doses in those genetic subgroups and therefore did not lead to distinctive clinical results.

Unlike the pharmacogenetic studies analyzed, most of which were randomized clinical trials, all the pharmacogenetic studies in the sample were prospective, observational and usually naturalistic. Another difference between both groups was the sample size. Pharmacokinetic studies included fewer patients, albeit enough to draw meaningful conclusions. This is to a certain extent indicative of the lack of economic support for pharmacokinetic studies, generally based on clinical practice. Although pharmacokinetics was introduced into hospitals in the 1960's, its implementation has faced many difficulties. It has often been used as a way of controlling toxicity rather than increasing efficacy, for example in the case of digoxin. Its use has been cast aside because of alleged increases in healthcare costs.

In the field of neuropsychiatry, although scientific evidence has supported the implementation of pharmacokinetics techniques more firmly than in the case of, say, aminoglycosides1, few hospitals use such procedures routinely, least of all those that extend their practice beyond clozapine and tricyclic antidepressants. The results obtained in this systematic review suggest the existence of significant differences in drug exposure-related pharmacokinetic parameters (concentrations, concentration-dose indices) between responding and non-responding patients treated for depression. This tool might be instrumental in optimizing drug therapy, which is a priority in this population of patients. Needless to say, more randomized clinical trials are needed to clearly establish the benefit of pharmacokinetic techniques. Pharmacokinetic studies carried out in this area by the pharmaceutical industry, which is the only actor capable of defraying the cost of a clinical trial, are limited to the items required for them to be able to register their products. The same can be said about other areas such as therapeutic monitoring of antibiotics, antifungals or cytostatics, which has been based on findings made by isolated researchers, groups or scientific societies and not on the information provided in their SmPCs.

At any rate, the most logical thing to do would be to implement optimization strategies based on the study of potential drug-drug interactions in order to avoid genetic interpretation errors. Treatment should be initiated based on the different genetic variants known to be associated with neuropsychiatric disorders and pharmacokinetic techniques should be applied to optimize dosing and identify potential unreported drug-drug interactions.

No funding.

No conflict of interests.