Clinical Utility and Implementation of Pharmacogenomics for the Personalisation of Antipsychotic Treatments

Decades of pharmacogenetic research have revealed genetic biomarkers of clinical response to antipsychotics. Genetic variants in antipsychotic targets, dopamine and serotonin receptors in particular, and in metabolic enzymes have been associated with the efficacy and toxicity of antipsychotic treatments. However, genetic prediction of antipsychotic response based on these biomarkers is far from accurate. Despite the clinical validity of these findings, the clinical utility remains unclear. Nevertheless, genetic information on CYP metabolic enzymes responsible for the biotransformation of most commercially available antipsychotics has proven to be effective for the personalisation of clinical dosing, resulting in a reduction of induced side effects and in an increase in efficacy. However, pharmacogenetic information is rarely used in psychiatric settings as a prescription aid. Lack of studies on cost-effectiveness, absence of clinical guidelines based on pharmacogenetic biomarkers for several commonly used antipsychotics, the cost of genetic testing and the delay in results delivery hamper the implementation of pharmacogenetic interventions in clinical settings. This narrative review will comment on the existing pharmacogenetic information, the clinical utility of pharmacogenetic findings, and their current and future implementations.


Introduction
Antipsychotic drugs are the mainstay treatment for psychotic symptoms of schizophrenia and other severe mental disorders.However, about 30-50% of people receiving antipsychotics do not respond adequately and/or develop severe and long-lasting side effects.The reasons behind treatment failure and adverse reactions are difficult to discern as the precise mechanism of action of antipsychotic drugs is still unclear.Clinical and environmental factors [1] as well as genetic factors [2] have been proposed as contributors to variability in response to antipsychotic treatment.
Pharmacogenetic and pharmacogenomic research investigate the genetic contribution to response variability at individual loci and genome-wise level, respectively.The main aims of pharmacogenetic and pharmacogenomic investigations are the unravelling of the mechanism of action of antipsychotics and the identification of response predictors that could help in the improvement and personalisation of treatment.Decades of pharmacogenetic research have revealed individual genetic biomarkers of clinical response to antipsychotics [3].However, prediction of antipsychotic response based on genetic biomarkers is far from accurate.Despite the clinical validity of several pharmacogenetic findings, the clinical utility remains unclear.Nevertheless, genetic information on metabolic enzymes involved in the biotransformation of most commercially available antipsychotics has proven to be useful for the reduction of treatment induced side effects [4,5].At present, the implementation of pharmacogenetic interventions in psychiatric settings is very limited, and further proof of their clinical utility is required.Pharmacogenomic research conducted in the last decade has advanced our knowledge on the mechanism of action of antipsychotic drugs by confirming the therapeutic value of known targets [2,6].Further research into genomic and epigenomic factors contributing to response may help to improve the accuracy and clinical applicability of genetic information in psychiatry.
The aim of this narrative review is to summarise the knowledge on the pharmacogenetics and pharmacogenomics of antipsychotic drugs and to give an overview of their current clinical applications.

Literature Review
The recommendations of the Preferred Reporting Items for Reviews and Meta-Analyses statement (PRISMA, 2021) were followed for the literature search.We searched the ISI Web of Science using the Boolean method with the following terms: (1) "antipsychotic" OR "chlorpromazine" OR "droperidol" OR "fluphenazine" OR "haloperidol" OR "loxapine" OR "perphenazine" OR "pimozide" OR "prochlorperazine" OR "thioridazine" OR "thiothixene" OR "trifluoperazine" OR "aripiprazole" OR "asenapine" OR "brespiprazole" OR "cariprazine" OR "clozapine" OR "iloperidone" OR "lurasidone" OR "olanzapine" OR "paliperidone" OR "quetiapine" OR "risperidone" OR "ziprasidone"; (2) "pharmacogenetic" OR "pharmacogenetics OR "pharmacogenomic OR pharmacogenomics".We restricted the search from January 2010 to November 2023 (see Figure 1 for description of process).The references and discussions of all pooled articles were carefully scanned for additional publications.Before screening, we excluded studies (1) not performed in humans; (2) not published in English; and (3) not published in scientific journals.A total of 1406 studies were identified.Articles were excluded from this revision for the following reasons: (1) not drug of interest (not antipsychotics); (2) other areas of research (Alzheimer disease, depression, implementation of informatic tools, etc.); (3) not original new data (reviews of publications); (4) case-report studies; (5) studies with exclusive pharmacokinetic design and outcomes; (6) studies with not reported response phenotype.A total of 269 relevant papers were identified that reported original pharmacogenetic or pharmacogenomic results on a variety of antipsychotic response phenotypes (influence of genetic variants on treatment response, adverse reactions and/or drug clearance).Several limitations to this review need to be considered.Firstly, it is likely that the most recent publications that are not yet listed in search engines, as well as other relevant papers that did not fit our search terms, have been missed.Secondly, many important findings regarding the pharmacogenetics of antipsychotics were published before the 2010 cut-off, although most clinically relevant findings have been replicated in more recent years.Furthermore, there is publication bias in favour of studies with positive findings, and negative or contradictory findings may be underrepresented.Despite these limitations, the next sections will summarise the most significant findings in recent years and will discuss the current applicability of those findings for the personalisation of antipsychotic treatment.

Pharmacogenetic Studies
Pharmacogenetic or candidate-gene studies are mainly focussed on a limited number of genes selected from the kinetic or dynamic profile of antipsychotic drugs, although genetic variants determining response may also be related to disease pathology.Findings of candidate gene studies are rarely universally replicated.Limited samples sizes, differences in the duration and type of treatment, patient ethnicity, etc., may hinder the replication of results [2].Furthermore, treatment response and clinical outcome are difficult to measure and may complicate the interpretation of results and explain discrepancies between studies.Replication of candidate-gene findings in independent studies and meta-analyses of published results are necessary to confirm and clarify the strength of the associations.A summary of candidate-gene studies reporting significant findings published in the last decade can be found in Table 1.Older papers with significant results reported or included in later publications have been excluded.While not fully comprehensive, this section will discuss the most relevant and clinically valuable findings to date.

Pharmacogenetic Associations with Treatment Response
Given the multitarget profile of most antipsychotics, it is not surprising that many genes related to neurotransmitter receptors and transporters have been associated with variability in treatment response.Dopamine and serotonin neurotransmitter pathways have been thoroughly investigated for possible pharmacogenetic markers [2].Although the mechanism of action of antipsychotics is not fully understood, it is known that dopamine blockade is directly associated with antipsychotic efficacy [2].Numerous studies have described associations between genetic variants in dopamine receptors and treatment response [7][8][9][10], with the DRD2 rs1799732 (−141C Ins/Del) deletion significantly associated with poorer antipsychotic drug response (OR = 0.65 for the Ins allele in meta-analyses) [2].Although initially a non-significant trend of association of the DRD3 Ser allele with poor clozapine response was observed, a later meta-analysis [11] confirmed the association of the 9Gly allele with good response in Caucasians, but not in other ethnic groups (OR = 0.72; p = 0.002).Genetic variants in serotonin receptors have also been associated with treatment response, particularly to second-generation antipsychotics.Recent studies confirm the association, with HTR2A rs6313 T allele carriers showing generally better improvement than non-carriers [2,12].HTR3 variants have been associated with clozapine response in Indian patients [13], although the strength of the association varied according to the response phenotype investigated, illustrating the difficulties in assessing clinical outcome.The associations of HTR2A and HTR3 variants with clozapine response have also been confirmed via meta-analyses (OR = 0.68 and 0.45, respectively) [14].Histaminic receptors, a target of several second-generation antipsychotics (SGA) of unclear therapeutic validity, may contain genetic variants influencing response [15][16][17].Other genes not directly targeted by antipsychotics but involved in monoamine pathways and/or mental disorders have been associated with response.The COMT rs4680 variant (Val158Met), associated with reduced cortical dopamine [18], has been associated with response to second-generation antipsychotics [19,20].A recent meta-analysis confirmed the strength of this association (Z = 6.709, p = 9.8 × 10 −12 ) [21].Variants of the BDNF gene were associated with risperidone and clozapine response [22] and with treatment resistance [23].

Pharmacogenetic Associations with Antipsychotic-Induced Adverse Reactions
Antipsychotic treatment induces severe side effects such as extra-pyramidal side effects (EPS) and weight gain (WG).There is strong evidence on the contribution of genetic factors to the development of adverse reactions during antipsychotic treatment.The influence of functional polymorphisms in metabolic enzymes on the clearance of antipsychotic drugs has been proven in numerous studies [2,4,24].Cytochrome P450 (CYP) functional variants associated with slower (poor metabolisers) or faster (rapid or ultrarapid metabolisers) metabolic rates are known to significantly contribute to antipsychotic efficacy and toxicity [2,25,26].More robust data have been obtained for functional variants in CYP2D6 and CYP2C19 enzymes.CYP1A2 and CYP2D6 functional variants have been associated with tardive dyskinesia (TD) [27,28].Patients with poor or intermediate CYP2D6 metaboliser phenotypes are at greater risk for risperidone and haloperidol adverse reactions [29,30].Currently, genetic variants of CYP enzymes are the most used biomarkers to personalise psychiatric treatment, as will be discussed later.Several studies have found a relationship between oxidative stress and antipsychotic-induced TD based on genes involved in the antioxidant defence mechanism, dopamine turnover and metabolism [31].DRD2 polymorphisms have been related to akathisia and other movement disorders [32].Expression of the HSPG2 gene is altered by antipsychotic treatment and has been associated with movement disorders in animal studies.Numerous studies have investigated its association with EPS, and a meta-analysis found a significant association between the HSPG2 rs2445142 variant and TD [33].Regarding weight gain, variants in HTR2C, LEP, and MC4R genes have been confirmed as genetic markers with moderate genetic effects (ORs = 1.47-1.96)[34,35].HTR2C is involved in the regulation of appetite, and its contribution to weight gain and body mass index during antipsychotic treatment has been thoroughly investigated.
The HTR2C rs3813929 (-759-C/T) variant is the most frequently studied single-nucleotide polymorphism and one of the most strongly associated with antipsychotic-induced WG, with the rs3813929-T allele predicting less WG and body mass index (BMI) [36][37][38][39].Gene variants in LEP and LEPR, two genes involved in the regulation of lipid metabolism, have been related to antipsychotic-induced WG and short-term dyslipidaemia [40,41].A recent meta-analysis including 12 studies found no association between the LEP rs7799039 polymorphism and WG, although a significant association was found within a subgroup of first-episode schizophrenia patients (OR = 2.32, p = 0.0009) [42].One of the most robust findings is the association of genetic variants in the MC4R gene with antipsychotic-induced WG.The MC4R was previously associated with obesity in the general population, adding plausibility to the finding, which has been replicated in later studies [43,44].Finally, an HLA-DQB1 variant (6672G/C, rs113332494) was robustly associated with clozapine-induced agranulocytosis [45].The association of HLA variants with antipsychotic-induced agranulocytosis was later corroborated in genome-wide association studies [46,47].The strength of the association increases when patients' ancestry is considered (OR = 16.67-53.70)[45].
Other findings reported in single studies (see Table 1) require confirmation in independent samples to determine the validity and strength of the associations.In general, the strength of the reported genetic associations is moderate (odds ratios rarely >3).This is not surprising given the multitarget profile of most currently used antipsychotics and the polygenic characteristics of treatment response.However, this suggests that the clinical utility of these findings is limited, as discussed later.

Pharmacogenomic Studies
As we have seen in the previous section, candidate-gene studies have succeeded in the identification of treatment response markers, some of which are currently in use for the personalisation of psychiatric medication.However, pharmacogenetic studies are based on existing knowledge on the selection of candidate genes, and their contribution to the understanding of the mechanism of action of antipsychotics is limited.Advances in technology have allowed thorough genomic investigations of disorders and treatment outcomes.No previous knowledge is required as the whole genome is investigated.However, genomic studies on antipsychotics have been marred by limited sample sizes that increase the probability of false-negative and -positive findings.Most pharmacogenomic studies on antipsychotics have been conducted with moderate sample sizes (n < 1000).Several studies included replication in independent samples to confirm their findings.Table 2 includes a summary of genome-wide and whole exome sequencing studies on antipsychotic response and their side effects published during the last decade.
A BeadChip study of 6789 SNPs conducted by Drago et al. [48] found association between response to SGA and first-generation antipsychotics (FGA) and two ARID5B SNPs, a gene involved in autoimmune disorders that are often found in schizophrenia patients.A study investigating variants in 1996 genes on patients treated with olanzapine and risperidone also found associations of SNPs in chromosome 6, a region where the HLA gene is located, with treatment response, reinforcing the evidence on the influence of genetic variant in immunologic pathways in clinical outcomes [49].A genome-wide association study (GWAS) on Indian patients also found a variant (rs4795893) in a gene involved in pro-inflammatory response (CCL2) as well as a variant in a gene involved in mental disorders (NRG1 rs13250975), amongst others, associated with response to SGA and FGA [50].Two genes not previously associated with treatment response or mental disorders, MYOTB and MTRR, were found to contain variants associated with antipsychotic response using whole exome sequencing (WES) [51].This strategy also identified polymorphisms in the genes NMDA and AMPA, involved in glutamatergic pathways, associated with response [52].Alterations in glutamatergic pathways have long been associated with the aetiology of schizophrenia, although no clear association with treatment response has been described to date.Interestingly, a recent GWAS also found association between genetic variants in the glutamatergic GRM7 gene and treatment response [53].The same study identified novel associations of GPR12 and MAP2K3, two genes involved in glutamatergic pathways, with response using WES techniques, although no finding was replicated by both strategies.Finally, a sequencing study of 143 genes revealed statistically significant association between polymorphisms in DRD3 (rs342026) and in a neighbouring gene (PLK5 rs1261027) with olanzapine response, confirming the findings of candidate-gene studies [54].
Regarding adverse reactions, Tanaka et al. described an association between a variant (rs6977820) in the gene DPP6 and treatment-induced TD, although the biological relation of this gene with the adverse reaction is not clear [55].Alkelai et al. [56] performed WES in patients treated chronically with antipsychotics and found a significant enrichment of RIMS2 variants associated with TD.The authors report that the expression of this gene was observed to be increased in the brains of schizophrenic patients in a previous study.One of the most robust pharmacogenetic findings to date is the association of genetic variants in the melanocortin 4 receptor (MC4R) with antipsychotic-induced WG, an association that has been confirmed in GWAS and candidate-gene studies (see Table 1).Several novel associations with antipsychotic-induced WG have been reported in subsequent genomic studies but have not yet been replicated and require further study [57][58][59].Two GWAS studies have confirmed the previously reported association between genetic variants in the HLA complex and clozapine-induced agranulocytosis [46,47].Furthermore, a GWAS study described a mutation in the AKCR1 gene, rs2814778, previously associated with lower neutrophil counts, associated with clozapine-induced neutropenia [60].Finally, a relatively large GWAS study identified variants in two genes, ATAD3B and SKIL, related to antipsychotic-induced QTc interval change.The ATAD3B gene plays a role in inflammatory pathology, whereas alterations in the expression of the SKIL gene may be related to the aetiology of schizophrenia [61].Despite their moderate power, these genomic studies have confirmed the importance of genes involved in inflammatory response and in glutamatergic pathways on treatment variability in the development of antipsychotic-induced side effects.However, some of the findings are difficult to interpret given the modest sample sizes and the difficulty in determining the interaction with other environmental and clinical factors involved in response.Independent candidate-gene studies may be required to confirm the validity of findings of underpowered GWAS and WES studies.
Aside from alterations in the DNA sequence, other genomic factors such as transcription regulation and epigenomic events may play a role in antipsychotic response variability.The study of gene expression and epigenetic changes may provide information on gene × environment interactions and yield more accurate predictions of clinical outcomes.Animal studies have shown that risperidone treatment induces gene expression changes associated with susceptibility to EPS, with the involvement of the mTOR pathway [62].A study suggested that the basal expression of four genes can predict antipsychotic response.However, the estimation was based on a limited number of patients (n = 30) and needs replication in independent studies [63].Several studies have provided evidence suggesting that antipsychotic medications modulate DNA methylation and that these alterations may influence treatment response [64,65].Animal studies have shown that methylation levels of dopaminergic genes are altered by olanzapine treatment [64].DNA methylation changes have been correlated with clinical improvement in MDD patients treated with antidepressants, and similar effects may occur in antipsychotic-treated patients [66].Human studies have shown that global methylation levels are decreased in schizophrenia patients in comparison with controls, and antipsychotic treatment may partially explain the decrease [65].Methylation within a recognition sequence for HES transcriptional repressors was found to correlate with clinical improvement after antipsychotic treatment [67].Research on the interaction of genomic events and environmental factors on antipsychotic response is ongoing despite the difficulty in obtaining large, detailed cohorts for study.

Clinical Utility of Findings
The two previous sections have reviewed numerous studies that have identified genetic markers of variability in antipsychotic treatment.Pharmacogenetic and pharmacogenomic studies have succeeded in validating antipsychotic targets of therapeutic value as well as provided novel information highlighting new areas of putative therapeutic value.However, despite many of these findings having been confirmed and classified as true pharmacogenetic markers, their clinical utility for the personalisation of antipsychotic treatment is not clear.Even with the strong evidence confirming the association of dopaminergic and serotonergic variants and treatment response, their moderate to low genetic effects indicate that their contribution is not determinant on antipsychotic treatment outcome, and that taken individually, they could not predict response accurately.Similarly, HT2RC and MC4R associations with antipsychotic-induced WG have been confirmed in independent studies; however, they have no clear clinical utility to discriminate between different antipsychotics despite their varied pharmacological profiles.The strong association of genetic variants in the HLA complex with clozapine-induced agranulocytosis may help to predict if the patient has a high or low risk of developing severe neutropenia [68], but the relatively modest sensitivity of these markers (31-41%) [46,47] does not reach the 50% threshold for clinical applicability [69], although the addition of patient's ancestry in the prediction algorithm may increase the clinical value [45].Algorithm combinations of information in several genes and polygenic risk scores have been proposed to improve prediction of antipsychotic response and induced side effects.An algorithm combination of four SNPs in genes involved in mTOR regulation had a 66% accuracy to predict EPS in a sample of n = 356 antipsychotic-treated patients [70].A panel of genetic variants in 15 genes predicted efficacy in 60% of cases and medication tolerability in 71% of the n = 352 patients included in the study [71].Patients with higher polygenic risk scores for schizophrenia tended to have less improvement with antipsychotic drug treatment [72].Interestingly, polygenic risk scores for schizophrenia, diabetes and BMI were also associated with AIGW [73].Although these strategies may be useful as prognostic tools, their clinical utility for the selection of treatment remains unclear.Furthermore, most of the pharmacogenetic markers in drug targets may not help to discriminate between antipsychotic treatments given their multitarget profile and the low-to-moderate genetic effects observed in the reported associations.
While pharmacogenetic and pharmacogenomic research is still ongoing to improve response prediction and clinical utility, it is likely that the best genetic biomarkers with the greatest effect and clinical utility have already been discovered.Certainly, functional CYP polymorphisms may have both clinical utility and applicability.Functional polymorphisms in CYP enzymes have been clearly associated with the clinical outcome of antipsychotic treatments.The presence of CYP poor-metaboliser variants is strongly associated with increased adverse reactions, whereas ultrarapid-metaboliser variants may contribute to treatment failure and toxicity [2].Guiding treatment according to CYP phenotypes has been proven useful in children [74].Furthermore, a recent study showed that long-acting antipsychotics may still be affected by CYP genetic variants [75].These associations are of a higher magnitude than the ones observed with genetic variants in antipsychotic targets and may offer useful information on appropriate dosing according to the functional variants carried by the patient.In the case of antidepressants, clinical dosing recommendations based on the patient's genetic profile have already been issued by the Clinical Pharmacogenetics Implementation Consortium (CPIC) [76,77].Several studies have proven the clinical benefits of using CYP information for the adjustment of clinical doses or selection of alternative treatments when genetic contraindications are observed.Growing evidence suggests that information on CYP1A2, CYP2C19 and CYP2D6 may be useful for the adjustment of clinical doses of antipsychotics [78,79].In a previous study, we demonstrated that the dose adjustment of the commonly used antipsychotics haloperidol, risperidone, aripiprazole, clozapine and olanzapine according to the genetically determined phenotype of CYP2D6, CYP1A2 and CYP2C19 resulted in a decrease of side-effects [4].A similar pharmacogenetic intervention for children and adolescents with autism spectrum disorders resistant to treatment resulted in clear improvement in 80% of cases as well as a reduction in health-associated costs [80].CYP2D6 and CYP2C19 genotyping improved response in addition to clinician's and patient's satisfaction in patients given antipsychotics or antidepressants [81].As further evidence of the clinical and economic benefits emerges, the implementation of CYP information for the improvement of antipsychotic response will surely increase in clinical settings.
Further improvement of pharmacogenetic predictions may be achieved by investigating the interaction between genes, environment, clinical and demographic factors.A recent study suggests that the consideration of pharmacogenetic and phenoconversion data that considers concomitant treatment and other dose-altering environmental factors may increase the accuracy of antipsychotic dose selection [82].An additional area that merits further research is the possible interaction of pharmacogenetics x age.Older children are more likely to experience EPS than younger children, whereas antipsychotic adverse reactions such as sedation, weight increase and fatigue in particular are more prevalent in children than in adults [83,84].Despite this evidence, few pharmacogenetic studies have been conducted in children.In summary, antipsychotic selection based on pharmacogenetic findings may be difficult to achieve, although information on pharmacokinetic genes has a real value for the selection of appropriate clinical doses resulting in improved efficacy and less toxicity.

Implementation of Pharmacogenetics for Antipsychotics
Despite the growing evidence of the clinical benefits, pharmacogenetic information is rarely used in psychiatric settings as a prescription aid.There are several reasons for this: limited evidence on the clinical benefits and cost-effectiveness of pharmacogenetic interventions, few clinical guidelines based on pharmacogenetic biomarkers, and delay in delivery of results, amongst others [2,85].
The clinical utility of pharmacogenetic-guided antidepressant treatment has been proved in several studies Genotyping for key CYP functional polymorphisms and variants in the serotonin transporter gene (SLC6A4), the main target of SSRIs, can be used to select antidepressant type and dose with significant improvement in efficacy [79].A recent study showed that children presenting pharmacogenetic contraindications for the antidepressant treatments they were receiving showed poorer response than those without contraindications [86].However, evidence on the clinical benefits of pharmacogenetic interventions for antipsychotics is sparse.A large clinical trial on the impact of pharmacogenetic interventions for 26 drugs of different medical areas did not find clear benefits, although a reduction in side effects on the subgroup of patients treated with antipsychotic and antidepressant medications was observed [87].A strategy including gathering of clinical and environmental information, study of pharmacological interactions, pharmacogenetic characterisation of the patients and adjusting the pharmacological treatment according to the information obtained has given good results in patients resistant to treatment with neu-roleptics [88].Pre-emptive strategies can also be clinically beneficial.It has been estimated that pre-emptive pharmacogenetic testing could be useful to predict 95% of vulnerable patients' exposure to inadequate drugs [89].Routinary pharmacogenetic testing of patients that require treatment with antipsychotics may significantly decrease the prevalence of side effects, as shown in our previous study [4].Pharmacogenetic interventions for the selection of antipsychotic treatment results in higher improvement in comparison with treatment as usual [90].Pharmacogenetic testing increases treatment adherence, leading to greater efficacy [91].However, further evidence in different clinical settings is required to confirm the clinical utility of genetic testing for antipsychotics.
Emerging evidence illustrates the cost-effectiveness of the implementation of pharmacogenetics in psychiatry.Pharmacogenetic-guided antidepressant treatment resulted in significant savings in medication and clinical care [92,93].CYP2D6 genotyping is cost-effective in patients treated with neuroleptics and antidepressants [91, [94][95][96].The combination of clinical, environmental and pharmacogenetic information for the personalisation of treatment in resistant patients achieved a cost-benefit ratio of 3.31-3.59with a reduction of direct and total costs in most patients [95].The cost-effectiveness of pharmacogenetic testing to prevent clozapine-induced agranulocytosis has been proven but only in patients with HLA susceptibility alleles [97].Rapid and low-cost pharmacogenetic interventions in treatment-resistant autistic patients resulted in a reduction in hospital stays and clinical visits [80].Further evidence of the cost-effectiveness of pharmacogenetic interventions will facilitate their implementation in national health services.
A significant barrier is the existence of pharmacogenetic guidelines for the adjustment of only a limited number of antipsychotics.The Dutch Pharmacogenetics Working Group (DPWG) has recently issued guidelines for the adjustment of aripiprazole, brexpiprazole, haloperidol, pirmozide, risperidone, zuclopenthixol and quetiapine according to CYP2D6 and CYP3A4 polymorphisms [5].However, they conclude that no useful pharmacogenetic information is available for the commonly used antipsychotics clozapine and olanzapine.In a previous study, we used local clinical guidelines for the adjustment of doses of several FGA and SGA antipsychotics [4].Dose adjustment according to CYP2D6-predicted metabolising rates resulted in a significant reduction of the toxicity induced by the antipsychotics aripiprazole, haloperidol and risperidone, confirming the utility of the existing guidelines.Furthermore, the adjustment of clozapine and olanzapine according to CYP1A2 and CYP2C19 polymorphisms also resulted in a reduction of side effects and improved efficacy in treatment-resistant patients [80].Further research to provide evidence and specific guidelines for clozapine and risperidone is required.
Long waits for test results can diminish the clinical utility of pharmacogenetics.Commercial pharmacogenetic tests may take a minimum of 2-3 weeks to provide results.Polygenic risk scores may predict treatment response with a higher level of accuracy [72].However, the increase in genotyping expense and in the time until results delivery significantly diminishes clinical utility.Biochemical measures (determination of metabolite plasma levels) obtained after a few weeks from the start of the treatment can provide more accurate information on drug clearance as the effect of concomitant treatments, diet and other environmental factors are reflected.Nevertheless, although genetic prediction of antipsychotic response may not achieve 100% accuracy, pharmacogenetic interventions for dose adjustment can be highly beneficial when applied at the start of the treatment, thus allowing for the selection of the right dose or alternative treatments.Rapid pharmacogenetic interventions that provide clinical recommendations within 24-48 h of request are of significant clinical utility [2,80].The genotyping of key polymorphisms in CYP enzymes can be easily performed in most clinical laboratories and produce information at the start of the treatment, facilitating personalisation and increasing clinical value.

Discussion
As described in previous sections, the information on functional variants in CYP enzymes for the adjustment of antipsychotic clinical doses has proven to be of significant clinical utility and should be incorporated as a prescription tool.Furthermore, the genotyping of genetic variants in CYP1A2, CYP2C19, CYP2D6 and CYP3A4 for the adjustment of the dose of commonly used antipsychotics can be cost-effective.However, given the variability of genotype and clinical recommendations provided by pharmacogenetic tests [98], a standardisation of the CYP markers investigated and the pharmacogenetic-guided recommendations should be performed.Currently available genetic tests interrogate a variety of CYP genes and variants, providing different information and clinical recommendations that should be standardised.In cases where sequencing technologies to detect novel or known functional variants are not available, a list of CYP pharmacogenetic markers useful to determine the metabolic status of patients of different ethnicities should be compiled.Additionally, the clinical recommendations based on pharmacogenetic information should also be standardised.Detailed information on both key CYP genetic markers and relevant dose changes for antidepressants and antipsychotics have been issued by the CPIC and DWPG consortiums [5,76,77].Both organisations compile clinical, pharmacokinetic and pharmacogenetic information to provide precise information on required dose changes, and their recommendations should be followed to elaborate reports on pharmacogenetic-guided clinical recommendations.Alternative recommendations with insufficient confirmation should be avoided.Standardised pharmacogenetic tests and recommendations will increase their reliability and facilitate their understanding and interpretation.
Finally, it is necessary to increase the education and training of clinical staff on the use and utilisation of pharmacogenetic information as a prescription tool for antipsychotic medications, either as a help in treatment-resistant patients or as a routinary pre-emptive prescription tool to increase the efficacy and safety of treatments.Although the teaching of pharmacogenetic strategies is being introduced in faculties in several countries, further education should be provided at different stages of training and in clinical settings to facilitate the use of pharmacogenetics for the personalisation of antipsychotic treatment.

Conclusions
Pharmacogenetic and pharmacogenomic research have helped to unravel the mechanism of action of antipsychotics and to identify biomarkers of response.However, most biomarkers have modest clinical utility, as their capacity to discriminate between different antipsychotics is reduced.Functional biomarkers in CYP metabolic enzymes are of greater utility as they can be used to adjust clinical doses of antipsychotics, resulting in greater efficacy and safety.Nevertheless, the implementation of pharmacogenetic interventions in psychiatry is minimal.Further evidence of the clinical benefits and cost-effectiveness, rapid and low-cost pharmacogenetic tests and specific pharmacogenetic guidelines for antipsychotics are required to increase their implementation.It is envisaged that artificial intelligence combining information from different areas (genetic, epigenetic, clinical, pharmacological and demographic) will produce a much more accurate prediction of response and will further the personalisation of antipsychotic treatment.However, given the complexity of such an approach combining information on multiple factors that need to be measured before the start of the treatment and the high cost that it will imply, it may be years until it is practicable in clinical settings.GWAS and WES 189 + 222 Chinese Risperidone GWAS revealed a significant association between GRM7 SNPs (rs141134664, rs57521140 and rs73809055) and treatment response [53] Abbreviations: AIWG: antipsychotic-induced weight -gain; FGA: first-generation antipsychotics; GWAS: genome-wide association study; SGA: second-generation antipsychotics; SNP: single nucleotide polymorphism; TD: tardive dyskinesia; WES: whole exome sequencing; WG: weight gain.

Summative Paragraph
• Genetic variants in genes coding for drug targets -dopamine and serotonin receptors in particular-may influence the efficacy and safety of antipsychotic medications.

•
Functional variants in CYPs are associated with antipsychotic availability.

•
Dose adjustment according to CYP functional variants present may help to improve adherence, efficacy and safety of antipsychotics.

•
Clinical implementation of pharmacogenetic interventions for personalisation of antipsychotic treatment is limited.

•
Improved clinical guidelines based on pharmacogenetic data, education and training in pharmacogenetics, reduced costs and shorter delivery times may increase implementation.

•
Further research on the combined effect of pharmacogenetics, phenoconversion, and clinical and environmental factors is required.

Figure 1 .
Figure 1.PRISMA description of literature search.

Table 1 .
Summary of candidate-gene studies reporting significant associations.

Table 2 .
Summary of genomic studies on antipsychotic medications.