The effectiveness of a pharmacogenetic tool in increasing the success of treatment of major depression.

Meta-analysis of three clinical trials

Recently, several pharmacogenetic tests have appeared on the market, which help in the selection of drugs for patients with psychiatric diseases. The purpose of this meta-analysis is to determine the benefits of a pharmacogenetic tool in psychiatry (Neuropharmagen®) in the treatment of patients with depression.

A meta-analysis of random effects in clinical trials was conducted, in which the effect of this tool on improvement in patients with depression was studied. These effects were expressed in the form of standardized differences between different groups of patients. 450 patients were studied according to three clinical studies.

The random effects model evaluated a statistically significant effect for recommendations derived from pharmacogenetic technology (d = 0.34, 95% CI = 0.11–0.56, probability value = 0.004), which corresponded to an increase of about 1.8 times the probability of a particular clinical response taking into account the recommendations of the pharmacogenetic test compared with the absence of appropriate recommendations.

After excluding patients with mild depression from the analysis, the effect increased to 0.42 (95% CI = 0.19–0.65, probability value = 0.004, n = 287), with OR = 2.14 (95% CI = 1.40-3.27).

Such results prove the clinical benefit of this pharmacogenetic-based tool for improving the health of patients with depression, especially in patients with moderate and major depression. In addition, additional RCTs reinforced the results obtained in other patient samples.

• Introduction

Severe depression (Major depressive disorder – MDD) is a very common disease that has become one of the most pressing health problems. It was recently estimated that more than 300 million people in the world suffer from this disease, which is 4.4% of the world’s population [1]. Worldwide, depression is one of the main causes of disability and high health care costs [2-4]. Despite the growing number of psychotropic drugs, the field of psychiatry needs to improve the existing model of “study-errors” in the appointment of treatment. The levels of responses to individual antidepressants are low, especially in patients with mild depression [5, 6]. Treatment Resistant Depression (TRD), generally defined as the unsuccessful use of two pharmacological drugs in a row with sufficient dosage and duration of use, accounts for at least 20-30% of cases of severe depression. In addition, the effectiveness is limited by the low level of adherence to treatment [8]. High rates of unsuccessful treatment and adverse effects caused by drugs carry an increased economic and social burden [7, 9, 10].

It was found that genetic variability affects the reaction, metabolism and safety of a drug. In an effort to standardize and simplify the introduction of pharmacogenetics into clinical practice, international expert consortia, such as the Consortium for the Introduction of Clinical Pharmacogenetics (CPIC), have undertaken the mission of publishing and updating proven clinical instructions based on the confirmation of the relationship of drugs and genes, indicating the degree of reliability, standardized terminology and clinical recommendations [11-17]. Medical agencies are also aware of the influence of genetics on the response to drugs, and that the number of names of approved pharmaceuticals that contain information about biomarkers is constantly increasing [18, 19]. This has led to an increase in the number of pharmacogenetics-based tools that serve to support decisions made by clinics, as well as tests available on the market. However, the links between certain genetic changes and clinical outcome (clinical validation) are necessary, but insufficient to maintain the use of such a specific pharmacogenetic technology (PGx), exactly as indicated in the ACCE model (Analytical Validation, Clinical Validation, Clinical Benefit, as well as issues of Ethics and legal norms) [20]. For proper clinical implementation of pharmacogenetic tests, the clinical benefit must be confirmed. In other words, randomized controlled trials (RCTs) should be performed to confirm that the pharmacogenetic PGx technology, determined by a combination of genetic changes and an algorithm that outputs us a list of medical recommendations, will be better in terms of health indicators compared to traditional clinical practice. In addition, it is recommended that these studies be conducted on a representative sample of patients from real practice, that is, the studies should be pragmatic [21]. This is supported by the observation of significant differences in the immediate effectiveness of antidepressants themselves in different studies conducted under naturalistic conditions and on carefully selected groups of patients, according to stage III [22].

Neuropharmagen® (“AB-Biotics, CA”, Barcelona, Spain) is a pharmacogenetics-based technology that uses a proprietary approach to make comprehensive recommendations based on the following: (i) existing published pharmacogenetic instructions (for example, the instructions of the Consortium for the Introduction of Clinical Pharmacogenetics (CPIC) for dosage adjustments depending on combinations of genotypes), pharmacogenetic information provided in the instructions for use of drugs approved by the U.S. Food and Drug Administration (FDA) or other selected clinical studies; and (ii) safety concepts for classifying genetic changes that introduce a significant risk in terms of adverse effects, among other changes. Two prospective clinical studies conducted in parallel and one retrospective study evaluated the clinical benefits of this PGx pharmacogenetic test technology in the pharmacological treatment of patients with severe depression. These were multicenter, randomized, double-blind studies lasting twelve weeks (AB-GEN study, n = 280) [23], a randomized clinical trial, simple blind, for a duration of eight weeks (Korean study, n = 100) [24], as well as a follow-up study for twelve weeks, retrospective, naturalistic and multicenter (GENEPSI study, n = 70) [25]. The main variable used to evaluate effectiveness differed among the data from the three studies. Evaluation of the effectiveness of treatment was carried out on a scale of general impressions of the patient depending on severity (CGI-S, [26]), to compare patients whose attending physician used pharmacogenetic technology to obtain recommendations on the choice of drugs (recommendations based on PGx technology) and patients who were treated in accordance with traditional clinical practice. Two prospective randomized clinical trials also evaluated the improvement of symptoms in patients with depression using the Hamilton Classification Scale of 17 positions (HDRS-17, [27]).

Below is a meta-analysis of three clinical studies that evaluated the clinical benefits of pharmacogenetic information provided using this technology based on pharmacogenetics in order to select pharmacological treatment of patients with depression in comparison with traditional clinical practice.

• Materials and methods

The recommendations specifically applied to the PRISMA meta-analysis (Preferred Reporting Items for Systematic Reviews and Meta-analyses) were followed.

1. Collected data

All the subjects in the three studies were over the age of 18, with the main diagnosis of “severe depression”, a CGI-S score of ≥ 3 according to the assessment of the attending physician, and were included as those patients who needed treatment, de novo so are those who needed a replacement drug. The average values of the CGI-S and HDRS-17 scales at the beginning of the studies did not differ between the groups with and without PGx technology recommendations in any of the three studies. Patients with concomitant diagnoses were not excluded, except for a study conducted in South Korea. The collected data included the following elements:

• Design and duration of the study.
• Characteristics of patients for each treatment group: sample size, average age, gender distribution, ethnicity, concomitant psychiatric disorders, CGI-S indicators at the first and last visits, as well as incoming and outgoing HDRS-17 indicators, if available.

Three clinical trials were approved by the Institutional Supervisory Boards (IRB) with the participation of the relevant centers, and were conducted in accordance with the Helsinki Declaration of Human Rights. The GENEPSI study protocol was approved by the IRB Boards of the San Carlos Clinical Hospital in Madrid, Spain (protocol code: ALM-PSI-2013-01, approval date: 20.12.2013); the AB-GEN-2013 study protocol was approved by the IRB Boards of the Clinical Hospital of Barcelona, Spain, acting as the basic central IRB, and as the IRB of the participating clinic (protocol code: AB-GEN-2013, approval date: 15.11.2013); the protocol of the South Korean study was approved by the IRB of St. Mary’s Hospital of the Catholic University of Bucheon, South Korea (approval number: HC16EIMI0015, approval date: 02.03.2016). Written informed consent was obtained from all participants of the AB-GEN study and the Korean study before registering for participation. In accordance with the policy of AB-Biotics, all patients participating in the GENEPSI study provided their written consent to take genetic samples.

In accordance with informed consent, all patients submit samples for DNA extraction and genotyping of selected genetic variations. The details of the laboratory analysis are also described in [23].

Pharmacogenetic data are extracted from the analysis of genetic polymorphisms in 30 genes associated with drug efficacy, metabolism and specific adverse effects (Table S1). The conclusion of Neuropharmagen® is available on a computer via an Internet site, and it provides information about the differences in the effect of antidepressants, antipsychotics, mood stabilizers and other drugs on the central nervous system (Figure S1). The PGx conclusion consists of the following sections: (1) a summary table based on the “safety first” approach, which presents priority warnings using color codes similar to traffic lights: red indicates an increased risk of adverse reactions to the drug, yellow indicates the need to control the dosage of the drug and/or a lower probability of positive reactions, green color is associated with an increase in the probability of a positive reaction and/or a low risk of adverse reactions, and white color means “use according to the instructions” – relevant genetic variations were not detected, and (2) PGx results, with a breakdown and explanations for certain medications, as well as recommendations contained in the instructions for the use of medicines approved by the US Food and Drug Administration (FDA) [18] and pharmacogenetic recommendations, as well as information on selected clinical trials [29-34].

2. Results

The measure of the effectiveness of the intervention (the magnitude of the effect) was calculated in the form of standardized differences, using an improved assessment developed by Hedge de la de Cohen [35], based on the indicator on the CGI-S scale, according to the conclusion of the attending physician from the first to the final visit for each study. The CGI-S score was the only common variable for all studies. Remission of symptoms was also assessed in two RCTs on the HDRS-17 scale. Thus, the secondary evaluation criterion in this meta-analysis was the effect value calculated based on the change in the HDRS-17 score from the first to the final visit in each study.

3. Statistics

Continuous variables were expressed as mean ± standard deviation (SD). Discrete variables were expressed as numbers and percentages. The basic characteristics of individual studies were compared using version 20 of the statistical data processing software package for social sciences SPSS (IMB Corp., Chicago, Illinois, USA).

The meta-analysis of random effects was calculated using the Meta [36] and Metafor [37] packages, as well as the R studio application version 1.2.1335 [38] version 3.6.0 of the statistical software package R [39]. The mean standard differences (SSR) and their corresponding 95% confidence intervals (CI) were calculated based on data on the clinical response (changes in CGI-S and HDRS-17 indicators) for each patient, as well as in the form of their combinations. The positive effect with 95% CI showed that the treatment in the group with PGx technology recommendations was more effective than in the control group (the probability coefficient

Statistical heterogeneity in the studies was determined by the statistics of Inconsistencies (H2). H2 is the total percentage of change attributed to differences between studies, and has traditionally been interpreted on the basis of indicators of 25, 50 and 75% to determine low, medium or high heterogeneity, respectively [40]. If the heterogeneity was statistically significant, the combined effect values for each recent clinical assessment were calculated using random effect models, and their significance was determined using Z-tests.

Finally, the odds ratios (OR) were estimated based on standardized differences in the analyzed continuous variables (the change in the score on the CGI-S and HDRS-17 scales) by the method of logical approximation developed by Hasselblad and Hedges [41,42].

4. Error estimation

The overall risk of RCT error was defined as CV (control value) and JE (joint estimation) based on six directions recommended by the Cochrane error risk assessment tool [43]: sequence generation, distribution of hidden data, blinding, data on incomplete results, sample statements of results and other sources of bias. A tool for assessing the risks of errors in non-randomized trials, with interventions (ROBINS-I) [44] was used to assess the risk of error in GENEPSI study.

• Results

Table 1 provides a summary of the main characteristics of three clinical trials with an assessment of the clinical results of prescriptions issued with or without Neurofarmagen® recommendations for patients with severe depression. The studies covered 450 patients with severe depression as the first diagnosis, with an average age of 50 years (range 28.1 – 70.4) and a female-male ratio of 2.6:1. In the South Korean study and GENEPSI, there should have been unsuccessful treatment courses in at least one preliminary averaging due to lack of efficacy and/or tolerability, whereas untreated patients were not excluded from the AB-GEN study. The average rate in studies with antidepressants that previously had no effect for the current episode was 2.6 ± 2.2 in the AB-GEN study and 2.3 ± 1.9 in the South Korean study. Two RCTs identified patients who met the criteria for inclusion both for PGx recommendations and for groups with traditional treatment (TAU), through a random list generated by a computer that blocked and unblocked clinics’ access to PGx results, respectively. In the South Korean study, patients were monitored for eight weeks after randomization, and in the AB-GEN study, such control was carried out for ten weeks. The variables with the clinical assessment of patients (indicated in Table 1) were recorded at the time of randomization, in the middle of the study period and at the last visits. A retrospective GENEPSI study [25] defined an initial visit as one in which a saliva sample was taken, after which patients were monitored for 12 weeks after the initial visit. In this study, patients were retrospectively divided into two groups: (1) patients to whom psychiatrists prescribed pharmacological treatment according to recommendations issued using pharmacogenetic PGx technology, or (2) patients whose treatment was carried out without taking into account such recommendations – without adding or removing drugs with green warnings, and/or adding preparations with yellow and/or red warnings – (see Figure S1 as an example of a pharmacogenetic PGx statement).

Table 1. Characteristics of the covered patients and studies.

Study	Study design	Country	N	Demographics	Patient Characteristics
Korean study [24]	Eight weeks, multicenter, prospective, double-blind RCT. Two groups: with PGx technology recommendations and without recommendations (TAU). Variables for clinical evaluation: CGI-S, HDRS-17, FIBSER, PHQ-9/15, GAD-7, SDI, defined by clinics	South Korea	100 (with PGx recommendations n = 52, TAU n= 48)	With recommendations for PGx versus TAU. Gender (% of women): 76.9 compared to 72.9 (ns). Age (years): average (SD): 44.2 (16.1) compared to 43.9 (13.8) (ns) Ethnic composition (%): Koreans 100 compared to 100	Age ≥ 20 years The main diagnosis is “Severe depression” on the DSM-VCGI-S scale ≥3 despite the current treatment of AD with sufficient dosage and duration (at least 6 weeks), or intolerance. The main moderate to severe depression on the HDRS-17 scale. Past courses of treatment with antidepressants were unsuccessful, the average indicator (SD): 2.3 ± 1.9 Patients with drug abuse or hospitalized in the period of eight weeks before the study were excluded
AB-GEN study [23]	Twelve weeks, multicenter double-blind prospective RCT. Two groups: with PGx recommendations and without recommendations (TAU). Variables for clinical evaluation:  CGI-S, HDRS-17, FIBSER, PHQ-9/15, GAD-7, SDI, defined by clinics	Spain	280 (With recommendations PGx n = 136, TAU n = 144)	With PGx recommendations compared to the traditional TAU course: Gender (% of women): 63.9 compared to 63.4 (ns) Age (years), Average (SD): 51.74 (12.02) compared to 50.74 (13.12) (ns) Ethnicity (%): Caucasians 93.5 versus 91.3, Latinos 4.5 versus 6.2, others 2.0 versus 2.5 (ns)	Age ≥ 18 years. The main diagnosis is “Major depression” according to DSM-IVCGI-S ≥ 4 Medication requirement new or replacement of AD. The previous course of antidepressants was unsuccessful, average (SD): 2.6 ± 2.2 18% of patients suffered from borderline depression at the beginning of the study (indicator on the HDRS scale <14) 13% of patients abused medications at the beginning of the study. None of the patients were hospitalized at the beginning of the study
GENEPSI study [25]	Twelve weeks, multicenter, retrospective, observational and naturalistic study. Two groups: treatment with and without PGx recommendations. Variables for clinical evaluation: CGI-S scores determined by clinics, adverse effects data were taken from medical records	Spain	70 (with PGx recommendations n = 38, without recommendations n = 32)	Comparison of cases where PGx recommendations were taken into account and cases where they were not: Gender (% of women): 76.3 relative to 81.25 (ns) Age (years), Average (SD): 54.3 (14.5) relative to 55.2 (15.2) (ns) Ethnicity: Mostly Caucasians	Age ≥18 years. Psychiatric diagnosis according to ICD-10. Previous unsuccessful treatment (lack of effect and/or tolerability) of CGI-S ≥3. 30% of patients were hospitalized at the beginning of the study. 10% of patients intentionally concealed drug abuse. HDRS-17 scale – without evaluation

The main variable differed between the three clinical studies: the CGI-S, PGI-I (Patient Global Impression of Improvement) and HDRS-17 scales were used in the GENEPSI, AB-GEN and South Korean studies, respectively, but the CGI-S scale was taken for evaluation in all studies to determine the improvement of patients’ condition. Two RCTs covering only patients with the main diagnosis of severe depression also took the HDRS-17 scale as a variable to assess the response. On the contrary, the retrospective study was evaluated only on the CGI-S scale, since the study covered groups of patients with different major psychiatric diagnoses – MDD (severe depression), bipolar disorder, schizophrenia and general anxiety. For the purposes of this meta-analysis, only the data of patients with the main diagnosis of severe depression were taken into account. The average value of the initial indicators of CGI-S and HDRS-17 did not differ between the groups taking into account and not taking into account the PGx recommendations in any of the three studies (Table 2). The primary severity, assessed on the CGI-S scale, was the same in all three studies. However, when depression symptoms were assessed on the HDRS-17 scale, patients from the AB-GEN study showed an average of 19.2 ± 5.8, while participants in the South Korean study had an average value on the HDRS-17 scale of 23.8 ± 4.8 (probability value

Table 2. The main indicators on the CGI-S and HDRS-17 scale in groups that take into account pharmacogenetic recommendations and in control groups for three clinical trials.

Study	Variable	Taking into account PGx recommendations	Control	p–value
Korean study [24]	CGI-S, average ± SD	4.90 ± 0.80	4.60 ± 0.70	0.063
HDRS-17, average ± SD	24.50 ± 4.60	23.10 ± 5.00	0.159
AB-GEN study [23]	CGI-S, average ± SD	4.50 ± 0.62	4.40 ± 0.57	0.166
HDRS-17, average ± SD	19.47 ± 5.96	19.01 ± 5.71	0.482
GENEPSI study [25]	CGI-S, average ± SD	4.29 ± 0.57	4.26 ± 0.72	0.836
HDRS-17, average ± SD	na	na	na

The results on the risk of error of the AB-GEN RCT and the South Korean study, as well as the randomized GENEPSI study, are presented in Tables 3 and 4, respectively. GENEPSI RCTs and the South Korean study took the same risk of error. The errors of selection, screening and confirmation are considered low. However, since the clinics of both RCTs knew exactly what kind of treatment was being carried out (a non-blind method), there was a possibility of errors in the results and definition. The South Korean study was not sponsored by companies, but AB-GEN included funding from one of the companies to develop tests using PGx pharmacogenetic technology. Taking into account the quality of the GENEPSI study, its retrospective design allowed to reduce the risk of error inherent in traditional open research. Patients were averaged by major severity (CGI-S), age, gender, drug abuse and concomitant diseases. The treatment, as well as the main visits and follow-up were clearly defined. All the previously obtained results were reported, and the patients who disappeared from observation were distributed uniformly.

Table 3. Assessment of the risk of error in the RCT of the Korean study and AB-GEN.

Error	Korean study [24]	AB-GEN study [23]
Sequencing (selection error)	Low: “Randomization was performed by the research center with a 1:1 ratio for the PGx and TAU groups, using a computer-generated random list”	Low: “Randomization was performed by the research center with a 1:1 ratio for the intervention group and the control group, using a computer-generated random list”
Hiding the location (selection error)	Low:	Low:
Random list compiled by an independent center	Random list compiled by an independent center
Hiding participants and researchers. (error of conducting)	High:	High:
Hidden patients	Hidden patients
Undisclosed clinics	Undisclosed clinics
Hiding the evaluation of the result (error of determination)	High:	High:
Assessments on the CGI-S and HDRS-17 scales made by non-hidden clinics.	Assessments on the CGI-S and HDRS-17 scales made by non-hidden clinics.
Data with incomplete results (error due to dropping out)	Low:	Low:
The patients who dropped out during the follow-up were distributed uniformly	The patients who dropped out during the follow-up were distributed uniformly
Selective statements (statement error)	Low:	Low:
Ready results were brought	Ready results were brought
Other sources of error	High:	High:
Patients selected by clinics	Patients selected by clinics
Without industry sponsorship	With financing from the industry

Table 4. Error risk assessment in a GENEPSI retrospective study

Error	GENEPSI study [25]
Distribution error	Low:
“The patient data were grouped, respectively, according to three psychiatric clinics in Madrid (see the author’s affiliation), in which Neuropharmagen was used”
Uniform distribution according to the main assessment of severity (CGI-S), age, gender, drug abuse and concomitant diseases
“A genetic test can show an increase in the effect of a placebo. To avoid these effects, we intend to conduct this retrospective study only on patients who underwent genotyping, instead of comparing patients to whom pharmacogenetic tests were applied and patients with traditional treatment courses.”
Selection of patients (initial error)	Low:
It was established that patients’ saliva should be sampled at the first visit”
The observation period is set for 12 weeks after the first visit
Erroneous classification of treatment courses (error due to erroneous classification)	Moderate:
The status of the intervention is well defined, albeit retrospectively
Deviations from the provided courses (error in implementation)	Low:
Retrospective distribution of treatment courses. No difference in the treatment methods of the groups.
Data loss (error due to dropping out)	Low:
The patients who dropped out during the follow-up were distributed
Measurement of results (measurement error)	Moderate:
The CGI-S score was recorded in the medical record by attending physicians before the retrospective protocol was determined.
Selective extracts (the error of the results extract)	Low:
Ready results were brought

Based on the results of three studies (n=450), random effects were evaluated at the level of statistically significant effects – calculated based on changes in values on the CGI-S scale determined by the attending physician – at the level of d=0.34 (95% CI = 0.11–0.56, probability value = 0.004) for the course prescribed by the recommendations of the pharmacogenetic PGx test (Figure 1), with a heterogeneity of H2 = 23.13%. Based on the results of two RCTs (South Korean study and AB-GEN, n=380), the average standard difference of random effects (SSR) calculated from changes in the HDRS-17 index also significantly contributed to the choice of a drug based on PGx technology (d = 0.33, 95% CI = 0.03–0.63, probability value = 0.030) (Figure 2). Moderate heterogeneity was observed in this analysis (H2 = 44.8%). To determine whether the estimated magnitude of the effect changed relative to the severity of the patients, an auxiliary analysis was performed covering a group of Korean patients together with AB-GEN study participants with the main values on the HDRS-17 scale greater than 17 (patients with moderate to severe depression, n=287). When patients with mild depression were excluded from the combined analysis based on changes in HDRS-17 indicators, the estimated effect value increased to 0.42 (95% CI = 0.19–0.65, probability value = 0.004).

Figure 1. Effect values separately and together with the clinical response based on changes in CGI-S indicators.

Figure 2. Effect values separately and together with the clinical response based on changes in HDRS-17 indicators.

The calculated values of the clinical response for patients who were prescribed PGx technology recommendations in comparison with patients without recommendations – obtained on the basis of SSR calculated on the basis of changes in CGI-S indicators – were 1.84 (95% CI = 1.27 – 2.81). Similarly, OR 1.1 (95% CI = 1.05 – 3.12) was estimated on the basis of SSR calculated based on changes in HDRS-17 indicators, and this value increased to 2.14 (95% CI = 1.40 – 3.27) when only patients with moderate and severe depression were taken into account (with HDRS-17 over 17).

• Discussion

To date, three studies have been published on the benefits of prescribing recommendations for the pharmacogenetic PGx technology using the Neuropharmagen® test on adult patients suffering from severe depression [23-25]. A combined analysis of these three studies showed that treatment according to PGx recommendations is associated with much greater effectiveness in terms of improving symptoms in patients, according to estimates, as CGI-S and HDRS-17 indicators change, compared with traditional clinical practice. The cumulative effect values correspond to an increase of approximately 1.8 times for clinical response indicators when selecting drugs taking into account PGx recommendations compared to their absence. A similar effect indicator was calculated for antidepressants (d=0.30), namely in the framework of phase III studies [45, 46]. It should be noted that during the phase III studies, antidepressants were controlled by placebo, the studies included in our combined analysis were controlled by groups undergoing active treatment with antidepressants and other psychoactive drugs without restriction by type, quantity and dosage.

RCTs involving patients with severe depression often exclude patients with common characteristics for many patients in conventional clinics. Several studies have shown that 65–90% of patients undergoing treatment for depression can be excluded from RCTs [47–49]. Concomitant psychiatric conditions are the most significant reasons for excluding patients. Paradoxically, more than 60–70% with depression have at least one more concomitant mental illness, while 30–40% have two or more, among which anxiety and medication use were most often observed [50–54]. In addition, although it is generally accepted that an indicator equal to or less than 7 on the HDRS-17 scale is a sign of remission of depression [55], a value exceeding 17–19 is usually required for inclusion in the RCT, since it has been observed that patients with a high degree of severity of the disease at the initial point of the study respond better to the course of treatment [6, 56]. First of all, this practice of reaching carefully selected patients has shown that there is a significant overestimation of the benefits of treating depression [5]. Considering that confirmations from the practice of treating depression have been demonstrated in patients with moderate and severe depression, RCTs also established the HDRS-17 index higher or equal to 18 for the coverage of patients with a study or an analysis of efficacy [57, 58].

Among the three studies included in the meta-analysis, the AB-GEN study, covering a large number of patients (n=280), was conducted in a real clinic (Table 1). During randomization, the average HDRS-17 score was 19.2 (±5.8), and up to 17.8% of patients suffered from mild depression with an HDRS-17 score of less than 14, who were excluded according to the criteria for exceptions in the STAR*D study [59]. The average value of previous failures in the studies was 2.6 ± 2.2, in the range of 0–15. In addition, concomitant psychiatric diseases were not excluded (35.8% of patients suffered from anxiety, and 12.6% had problems with drug abuse) [23]. A recent reanalysis of AB-GEN data showed that the clinical benefit of PGx technology was influenced by the characteristics of patients, including the main severity [60]. To compare the results of this meta-analysis with the results of phase III clinical studies on the effectiveness of antidepressants, as well as phase III studies on the effectiveness of a PGx-based course, which usually applied stricter criteria for coverage and exclusion, an auxiliary analysis was performed covering a group of South Korean patients together with patients from the AB-GEN study with an HDRS score of ≥18, which showed an increase in SSR from 0.33 for the entire group of studied patients (including patients with mild depression) to 0.42 in a subgroup of patients with moderate and severe depression. The corresponding OR for treatment according to the course prescribed by the pharmacogenetic test increased from 1.81 to 2.14 (95% CI =1.40–3.27).

Other tests and a combined meta-analysis of the results of clinical trials using different combinatorial tests based on the pharmacogenetic PGx technology in psychiatry have shown an increase in effectiveness compared to traditional treatment of adult patients diagnosed with severe depression [61, 62]. However, different PGx tools are based on certain proprietary algorithms and differ from each other in the sets of genes and selected genetic variants for the interpretation of the patient’s genetic profile, in the way they are combined and translated into clinical recommendations, as well as in the form in which the results are demonstrated, not counting other factors. Therefore, any pharmacogenetic instrument should be supported by specialized RCTs [63, 64]. The results of our meta-analysis are consistent with the results of previous analyses, but primarily show the clinical benefits of pharmacogenetic tests using specific tools compared to traditional treatment, especially for patients with mild to moderate depression. It is particularly noted that the benefits of this pharmacogenetic test were demonstrated in a sample with a predominant proportion of patients suffering from treatment resistant depression. All participants in the South Korean study have suffered at least two failures in the treatment using antidepressants in the past regarding the current case of severe depression, up to 65% of participants in the AB-GEN study were resistant initially, and for the GENEPSI study it was required that there was at least one unsuccessful treatment in patients as a criterion for coverage.

Some limitations have been identified that could affect the interpretations made in this meta-analysis. Firstly, a small number of studies are covered as an open retrospective study, which could have an impact on the calculated values of the effect. To eliminate this limitation, an auxiliary analysis was carried out, covering only data from two RCTs and showing similar results. Secondly, the clinical studies taken contained potential sources of error. Regarding the quality of RCTs, the selection error was considered low, since randomized lists were used in both studies, in which patients were distributed by the research center in a ratio of 1:1 for the group taking into account the recommendations of PGx and the control group, and the randomized list was generated on a computer. Patients who dropped out during the follow-up were distributed between two groups of two RCTs, presenting a slight risk of error. Both RCTs showed all the previously indicated results. On the other hand, the blind method of research could be considered as a certain limitation in these studies. Although it was not known for all the participants covered by both RCTs what kind of treatment they received, the attending physicians who chose one or another treatment in both studies could disclose some information, introducing the risk of a different approach during the course. The AB-GEN study minimized this risk by conducting telephone surveys with the help of independent evaluators who collected the main variables (PGI-I). However, the assessment of high degrees, such as on the HDRS-17 scale, was not considered timely when conducting telephone surveys. Personal surveys by independent evaluators could be allowed to blindly evaluate all secondary variables, but logical difficulties could also occur, probably increasing the number of respondents who dropped out. Other authors have resorted to alternatives such as the use of false genotyping data in control groups [65]. However, the risk of using false genetic profiles for the selection of drugs makes this method unacceptable from an ethical point of view. Regarding the quality of the open study, the overall risk of error was found to be low. When participants knew that they were being treated according to pharmacogenetic recommendations, there was an increase in the placebo effect due to the expectations of receiving personalized treatment. To eliminate this limitation, all participants covered by the GENEPSI study underwent retrospective genotyping. A uniform distribution was obtained between both study groups according to the main severity (CGI-S), age, gender, drug abuse and concomitant diseases. The retrospective design also made it possible to minimize other risks of error. The main variable (CGI-S) was registered in the medical record by a competent clinic, but before the study protocol was determined. The intervention status, although determined retrospectively, was determined correctly, and both groups were balanced in terms of patient characteristics (the main severity in terms of CGI-S, age, gender, drug abuse and concomitant diseases) and patient departures during follow-up. The initial visits and follow-up were also clearly defined, and all the previously indicated results were recorded. In terms of funding, two of the studies under consideration (AB-GEN and GENEPSI) were partially funded by AB-Biotics (the company that developed and marketed the pharmacogenetic technology that is the subject of the study). In addition, the staff of the AB-Biotics company assisted in the analysis of the results and in the preparation of relevant certificates. However, as far as we know, the South Korean study can be considered the first study of the use of pharmacogenetic technologies in psychiatry that was not funded by companies. Regarding the ethnic composition, most of the GENEPSI and AB-GEN participants were of European origin. However, 100 patients covered by the study in South Korea (approximately 20% of the analyzed samples) had a characteristic appearance of residents of the Asian region, as confirmation that the results apply not only to people of Caucasian appearance. The last limitation that was determined was the possibility of randomly assigning patients to the PGx recommendations group, but the course of their treatment would not coincide with prescription of pharmacological drugs. However, the form of extract from the AB-GEN study made it possible to control this restriction, under which 6.1% of patients from the study groups fell.

• Conclusions

This meta-analysis has shown the benefits of certain pharmacogenetic technologies for the selection of drugs for adult patients suffering from severe depression. However, due to the low number of studies taken, the generalization of these results may be limited. To regularly reinforce the benefits of PGx technology for patients with severe depression, further pragmatic RCTs may be required to assess the actual impact of PGx tests on certain groups of patients (for example, patients with a weak clinical picture of depression, on different ethnic and/or age groups). Given the high percentage of patients who do not meet the criteria for RCT coverage, although undergoing treatment for depression, the criteria for attribution/exclusion in future studies should be revised to close the existing gaps between the study groups and the actual patients of the clinics.

Additional materials: available online at www.mdpi.com , Table S1: list of analyzed genes and polymorphisms, Figure S1: example of an extract with interpretation using the Neuropharmagen test for one of the patients covered by the study, showing (a) color differentiation of drugs (b) detailed information on one of the drugs.

We express our gratitude to: for the development of the concept: S.V., M.T., E.V., E.A. and J.E.; for data storage: S.V., M.T. and J.E.; for in-depth analysis: S.V., M.T. and J.E.; for methodology: S.V., M.T. and H.E.; for project management: H.E.; for control: E.V., E.A. and J.E.; for editing: the original version – S.V.; for editing: checking and editing – S.V., M.T., E.V., E.A. and J.E.

Funding: This study did not receive external funding.

Conflicts of interest: S.V. and M.T. are employees of the company “AB-Biotics”. J.E. is an employee and a minority shareholder of the company “AB-Biotics”. MONTE and H.E. participated in the design of the study, analysis, interpretation of data and revision of documents with the main results of GENEPSI and AB-GEN studies. AB-Biotics partially funded the GENEPSI and AB-GEN studies, and conducted genotyping and provided the results from the PGx extract for three covered studies. E.V. received grants and acted as a consultant, expert or data source on the minimum effective concentration for the following companies: AB-Biotics, Abbott, Allergan, Angelini, Dainippon Sumitomo Pharma, Galenica, Janssen, Lundbeck, Novartis, Otsuka, Sage, Sanofi -Aventis and Takeda. E.A. received remuneration for consulting and training for Eli Lilly, Lundbeck, Otsuka, Pfizer and Sanofi-Aventis. E.A. participated as the scientific director of clinical trials funded by AB-Biotics, Bristol-Myers Squibb, Eli Lilly and Sanofi-Aventis, and acted as the national coordinator of clinical trials funded by Servier and Lundbeck.

List of sources

1. WHO. Depression and Other Common Mental Disorders; WHO: Geneva, Switzerland, 2017.
2. Steel, Z.; Marnane, C.; Iranpour, C.; Chey, T.; Jackson, J.W.; Patel, V.; Silove, D. The global prevalence of common mental disorders: A systematic review and meta‐analysis 1980–2013. Int. J. Epidemiol. 2014, 43, 476–493.
3. König, H.; König, H.‐H.; Konnopka, A. The excess costs of depression: A systematic review and meta‐ analysis. Epidemiol. Psychiatr. Sci. 2019, 1–16, doi:10.1017/S2045796019000180
4. Vos, T.; Abajobir, A.A.; Abate, K.H.; Abbafati, C.; Abbas, K.M.; Abd‐Allah, F.; Abdulkader, R.S.; Abdulle, A.M.; Abebo, T.A.; Abera, S.F.; et al. Global, regional, and national incidence, prevalence, and years lived with disability for 328 diseases and injuries for 195 countries, 1990–2016: A systematic analysis for the Global Burden of Disease Study 2016. Lancet 2017, 390, 1211–1259.
5. Rush, A.J.; Trivedi, M.H.; Wisniewski, S.R.; Nierenberg, A.A.; Stewart, J.W.; Warden, D.; Niederehe, G.; Thase, M.E.; Lavori, P.W.; Lebowitz, B.D.; et al. Acute and Longer‐Term Outcomes in Depressed Outpatients Requiring One or Several Treatment Steps: A STAR*D Report. Am. J. Psychiatry 2006, 163, 1905– 1917.
6. Fournier, J.C.; DeRubeis, R.J.; Hollon, S.D.; Dimidjian, S.; Amsterdam, J.D.; Shelton, R.C.; Fawcett, J. Antidepressant drug effects and depression severity: A patient‐level meta‐analysis. JAMA 2010, 303, 47– 53.
7. Johnston, K.M.; Powell, L.C.; Anderson, I.M.; Szabo, S.; Cline, S. The burden of treatment‐resistant depression: A systematic review of the economic and quality of life literature. J. Affect. Disord. 2019, 242, 195–210.
8. Sheehan, D.V.; Keene, M.S.; Eaddy, M.; Krulewicz, S.; Kraus, J.E.; Carpenter, D.J. Differences in medication adherence and healthcare resource utilization patterns: Older versus newer antidepressant agents in patients with depression and/or anxiety disorders. CNS Drugs 2008, 22, 963–973.
9. Warden, D.; Rush, A.J.; Trivedi, M.H.; Fava, M.; Wisniewski, S.R. The STAR*D Project results: A comprehensive review of findings. Curr. Psychiatry Rep. 2007, 9, 449–459.
10. Kelly, K.; Posternak, M.; Alpert, J.E. Toward achieving optimal response: Understanding and managing antidepressant side effects. Dialogues Clin. Neurosci. 2008, 10, 409–418.
11. Swen, J.J.; Nijenhuis, M.; de Boer, A.; Grandia, L.; Maitland‐van der Zee, A.H.; Mulder, H.; Rongen, G.A.; van Schaik, R.H.; Schalekamp, T.; Touw, D.J.; et al. Pharmacogenetics: From bench to byte—An update of guidelines. Clin. Pharmacol. Ther. 2011, 89, 662–673.
12. Leckband, S.G.; Kelsoe, J.R.; Dunnenberger, H.M.; George, A.L., Jr.; Tran, E.; Berger, R.; Muller, D.J.; Whirl‐ Carrillo, M.; Caudle, K.E.; Pirmohamed, M. Clinical Pharmacogenetics Implementation Consortium Guidelines for HLA‐B Genotype and Carbamazepine Dosing. Clin. Pharmacol. Ther. 2013, 94, 324–328.
13. Caudle, K.E.; Rettie, A.E.; Whirl‐Carrillo, M.; Smith, L.H.; Mintzer, S.; Lee, M.T.; Klein, T.E.; Callaghan, J.T. Clinical Pharmacogenetics Implementation Consortium Guidelines for CYP2C9 and HLA‐B Genotypes and Phenytoin Dosing. Clin. Pharmacol. Ther. 2014, 96, 542–548.
14. Hicks, J.K.; Sangkuhl, K.; Swen, J.J.; Ellingrod, V.L.; Muller, D.J.; Shimoda, K.; Bishop, J.R.; Kharasch, E.D.; Skaar, T.C.; Gaedigk, A.; et al. Clinical pharmacogenetics implementation consortium guideline (CPIC) for CYP2D6 and CYP2C19 genotypes and dosing of tricyclic antidepressants: 2016 update. Clin. Pharmacol. Ther. 2017, 102, 37–44.
15. Hicks, J.K.; Bishop, J.R.; Sangkuhl, K.; Muller, D.J.; Ji, Y.; Leckband, S.G.; Leeder, J.S.; Graham, R.L.; Chiulli, D.L.; LLerena, A.; et al. Clinical Pharmacogenetics Implementation Consortium (CPIC) guideline for CYP2D6 and CYP2C19 genotypes and dosing of selective serotonin reuptake inhibitors. Clin. Pharmacol. Ther. 2015, 98, 127–134.
16. Caudle, K.E.; Keeling, N.J.; Klein, T.E.; Whirl‐Carrillo, M.; Pratt, V.M.; Hoffman, J.M. Standardization can accelerate the adoption of pharmacogenomics: Current status and the path forward. Pharmacogenomics 2018, 19, 847–860.
17. Bank, P.C.D.; Caudle, K.E.; Swen, J.J.; Gammal, R.S.; Whirl‐Carrillo, M.; Klein, T.E.; Relling, M.V.; Guchelaar, H.J. Comparison of the Guidelines of the Clinical Pharmacogenetics Implementation Consortium and the Dutch Pharmacogenetics Working Group. Clin. Pharmacol. Ther. 2018, doi:10.1002/cpt.762.
18. Drozda, K.; Muller, D.J.; Bishop, J.R. Pharmacogenomic testing for neuropsychiatric drugs: Current status of drug labeling, guidelines for using genetic information, and test options. Pharmacotherapy 2014, 34, 166– 184.
19. Ehmann, F.; Caneva, L.; Prasad, K.; Paulmichl, M.; Maliepaard, M.; Llerena, A.; Ingelman‐Sundberg, M.; Papaluca‐Amati, M. Pharmacogenomic information in drug labels: European Medicines Agency perspective. Pharmacogenom. J. 2015, 15, 201.
20. Haddow, J.E.; Palomaki, G.E. ACCE: A Model Process for Evaluating Data on Emerging Genetic Tests. In Human Genome Epidemiology: A Scientific Foundation for Using Genetic Information to Improve Health and Prevent Disease; Khoury, M.J., Little, J., Burke, W., Eds.; Oxford University Press: Oxford, UK, 2004; pp. 217– 233. ISBN 0195146743.
21. Bousman, C.A.; Hopwood, M. Commercial pharmacogenetic‐based decision‐support tools in psychiatry.
22. Wisniewski, S.R.; Rush, A.J.; Nierenberg, A.A.; Gaynes, B.N.; Warden, D.; Luther, J.F.; McGrath, P.J.; Lavori, P.W.; Thase, M.E.; Fava, M.; et al. Can phase III trial results of antidepressant medications be generalized to clinical practice? A STAR*D report. Am. J. Psychiatry 2009, 166, 599–607.
23. Perez, V.; Salavert, A.; Espadaler, J.; Tuson, M.; Saiz‐Ruiz, J.; Saez‐Navarro, C.; Bobes, J.; Baca‐Garcia, E.; Vieta, E.; Olivares, J.M.; et al. Efficacy of prospective pharmacogenetic testing in the treatment of major depressive disorder: Results of a randomized, double‐blind clinical trial. BMC Psychiatry 2017, 17, 250.
24. Han, C.; Wang, S.‐M.; Bahk, W.‐M.; Lee, S.‐J.; Patkar, A.A.; Masand, P.S.; Mandelli, L.; Pae, C.‐U.; Serretti,
25. A Pharmacogenomic‐based Antidepressant Treatment for Patients with Major Depressive Disorder: Results from an 8‐week, Randomized, Single‐blinded Clinical Trial. Clin. Psychopharmacol. Neurosci. 2018, 16, 469–480.
26. Espadaler, J.; Tuson, M.; Lopez‐Ibor, J.M.; Lopez‐Ibor, F.; Lopez‐Ibor, M.I. Pharmacogenetic testing for the guidance of psychiatric treatment: A multicenter retrospective analysis. CNS Spectr. 2017, 22, 315–324.
27. Guy, W. ECDEU Assessment Manual for Psychopharmacology: Revised; DHEW Publication No. ADM 76–338; US Department of Health, Education and Welfare: Rockville, MD, USA, 1976.
28. Hamilton, M. A rating scale for depression. J. Neurol. Neurosurg. Psychiatry 1960, 23, 56.
29. Liberati, A.; Altman, D.G.; Tetzlaff, J.; Mulrow, C.; Gotzsche, P.C.; Ioannidis, J.P.A.; Clarke, M.; Devereaux, P.J.; Kleijnen, J.; Moher, D. The PRISMA statement for reporting systematic reviews and meta‐analyses of studies that evaluate healthcare interventions: Explanation and elaboration. BMJ 2009, 339, b2700.
30. Porcelli, S.; Fabbri, C.; Serretti, A. Meta‐analysis of serotonin transporter gene promoter polymorphism (5‐ HTTLPR) association with antidepressant efficacy. Eur. Neuropsychopharmacol. 2012, 22, 239–258.
31. Niitsu, T.; Fabbri, C.; Bentini, F.; Serretti, A. Pharmacogenetics in major depression: A comprehensive meta‐ analysis. Prog. Neuro‐Psychopharmacol. Biol. Psychiatry 2013, 45C, 183–194.
32. Breitenstein, B.; Scheuer, S.; Bruckl, T.M.; Meyer, J.; Ising, M.; Uhr, M.; Holsboer, F. Association of ABCB1 gene variants, plasma antidepressant concentration, and treatment response: Results from a randomized clinical study. J. Psychiatr. Res. 2016, 73, 86–95.
33. Uhr, M.; Tontsch, A.; Namendorf, C.; Ripke, S.; Lucae, S.; Ising, M.; Dose, T.; Ebinger, M.; Rosenhagen, M.; Kohli, M.; et al. Polymorphisms in the drug transporter gene ABCB1 predict antidepressant treatment response in depression. Neuron 2008, 57, 203–209.
34. van der Weide, K.; van der Weide, J. The influence of the CYP3A4*22 polymorphism on serum concentration of quetiapine in psychiatric patients. J. Clin. Psychopharmacol. 2014, 34, 256–260.
35. Mas, S.; Gasso, P.; Ritter, M.A.; Malagelada, C.; Bernardo, M.; Lafuente, A. Pharmacogenetic predictor of extrapyramidal symptoms induced by antipsychotics: Multilocus interaction in the mTOR pathway. Eur. Neuropsychopharmacol. 2015, 25, 51–59.
36. Hedges, L. V Distribution Theory for Glass’s Estimator of Effect size and Related Estimators. J. Educ. Behav. Stat. 1981, 6, 107–128.
37. Schwarzer, G.; Carpenter, J.R.; Rücker, G. Meta‐Analysis with R; Springer: New York, NY, USA, 2015. ISBN 9783319214153.
38. Pishva, E.; Drukker, M.; Viechtbauer, W.; Decoster, J.; Collip, D.; van Winkel, R.; Wichers, M.; Jacobs, N.; Thiery, E.; Derom, C.; et al. Epigenetic genes and emotional reactivity to daily life events: A multi‐step gene‐environment interaction study. PLoS ONE 2014, 9, e100935.
39. RStudio Team. RStudio: Integrated Development for R; RStudio, Inc.: Boston, MA, USA, 2015.
40. R Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing, Vienna, Austria, 2015.
41. Higgins, J.P.T.; Thompson, S.G.; Deeks, J.J.; Altman, D.G. Measuring inconsistency in meta‐analyses. BMJ 2003, 327, 557–560.
42. Hasselblad, V.; Hedges, L.V. Meta‐Analysis of Screening and Diagnostic Tests. Psychol. Bull. 1995, doi:10.1037/0033‐2909.117.1.167.
43. Chinn, S. A simple method for converting an odds ratio to effect size for use in meta‐analysis. Stat. Med. 2000, 19, 3127–3131.
44. Higgins, J.P.T.; Altman, D.G.; Gøtzsche, P.C.; Jüni, P.; Moher, D.; Oxman, A.D.; Savovic, J.; Schulz, K.F.; Weeks, L.; Sterne, J.A.C.; et al. The Cochrane Collaboration’s tool for assessing risk of bias in randomised trials. BMJ 2011, 343, d5928.
45. Sterne, J.A.; Hernán, M.A.; Reeves, B.C.; Savović, J.; Berkman, N.D.; Viswanathan, M.; Henry, D.; Altman, D.G.; Ansari, M.T.; Boutron, I.; et al. ROBINS‐I: A tool for assessing risk of bias in non‐randomised studies of interventions. BMJ 2016, 355, i4919.
46. Gibertini, M.; Nations, K.R.; Whitaker, J.A. Obtained effect size as a function of sample size in approved antidepressants. Int. Clin. Psychopharmacol. 2012, 27, 100–106.
47. Cipriani, A.; Furukawa, T.A.; Salanti, G.; Chaimani, A.; Atkinson, L.Z.; Ogawa, Y.; Leucht, S.; Ruhe, H.G.; Turner, E.H.; Higgins, J.P.T.; et al. Comparative efficacy and acceptability of 21 antidepressant drugs for the acute treatment of adults with major depressive disorder: A systematic review and network meta‐ analysis. Lancet 2018, 391, 1357–1366.
48. Zimmerman, M.; Mattia, J.I.; Posternak, M.A. Are Subjects in Pharmacological Treatment Trials of Depression Representative of Patients in Routine Clinical Practice? Am. J. Psychiatry 2002, 159, 469–473.
49. Zetin, M.; Hoepner, C.T. Relevance of Exclusion Criteria in Antidepressant Clinical Trials. J. Clin. Psychopharmacol. 2007, 27, 295–301.
50. Blanco, C.; Olfson, M.; Goodwin, R.D.; Ogburn, E.; Liebowitz, M.R.; Nunes, E.V.; Hasin, D.S. Generalizability of clinical trial results for major depression to community samples: Results from the National Epidemiologic Survey on Alcohol and Related Conditions. J. Clin. Psychiatry 2008, doi:10.4088/JCP.v69n0810.
51. Kessler, R.C.; Berglund, P.; Demler, O.; Jin, R.; Koretz, D.; Merikangas, K.R.; Rush, A.J.; Walters, E.E.; Wang,P.S. National Comorbidity Survey Replication the Epidemiology of Major Depressive Disorder. JAMA 2003, 289, 3095–3105.
52. Hasin, D.S.; Goodwin, R.D.; Stinson, F.S.; Grant, B.F. Epidemiology of Major Depressive Disorder. Arch. Gen. Psychiatry 2005, 62, 1097–1106.
53. Zimmerman, M.; Chelminski, I.; McDermut, W. Major depressive disorder and axis I diagnostic comorbidity. J. Clin. Psychiatry 2002, 63, 187–193.
54. Melartin, T.K.; Rytsälä, H.J.; Leskelä, U.S.; Lestelä‐Mielonen, P.S.; Sokero, T.P.; Isometsä, E.T. Current comorbidity of psychiatric disorders among DSM‐IV major depressive disorder patients in psychiatric care in the Vantaa Depression Study. J. Clin. Psychiatry 2002, 63, 126–134.
55. Rush, A.J. STAR*D: What Have We Learned? Am. J. Psychiatry 2007, 164, 201–204.
56. Romera, I.; Pérez, V.; Menchón, J.M.; Polavieja, P.; Gilaberte, I. Optimal cutoff point of the Hamilton Rating Scale for Depression according to normal levels of social and occupational functioning. Psychiatry Res. 2011, 186, 133–137.
57. Cameron, I.M.; Reid, I.C.; MacGillivray, S.A. Efficacy and tolerability of antidepressants for sub‐threshold depression and for mild major depressive disorder. J. Affect. Disord. 2014, 166, 48–58.
58. Singh, A.B. Improved Antidepressant Remission in Major Depression via a Pharmacokinetic Pathway Polygene Pharmacogenetic Report. Clin. Psychopharmacol. Neurosci. 2015, 13, 150–156.
59. Bradley, P.; Shiekh, M.; Mehra, V.; Vrbicky, K.; Layle, S.; Olson, M.C.; Maciel, A.; Cullors, A.; Garces, J.A.; Lukowiak, A.A. Improved efficacy with targeted pharmacogenetic‐guided treatment of patients with depression and anxiety: A randomized clinical trial demonstrating clinical utility. J. Psychiatr. Res. 2018, 96, 100–107.
60. Rush, A.J.; Fava, M.; Wisniewski, S.R.; Lavori, P.W.; Trivedi, M.H.; Sackeim, H.A.; Thase, M.E.; Nierenberg, A.A.; Quitkin, F.M.; Kashner, T.M.; et al. Sequenced treatment alternatives to relieve depression (STAR*D): Rationale and design. Control. Clin. Trials 2004, 25, 119–142.
61. Menchon, J.M.; Espadaler, J.; Tuson, M.; Saiz‐Ruiz, J.; Bobes, J.; Vieta, E.; Alvarez, E.; Perez, V. Patient characteristics driving clinical utility in psychiatric pharmacogenetics: A reanalysis from the AB‐GEN multicentric trial. J. Neural Transm. 2019, doi:10.1007/s00702‐018‐1879‐z.
62. Bousman, C.A.; Arandjelovic, K.; Mancuso, S.G.; Eyre, H.A.; Dunlop, B.W. Pharmacogenetic tests and depressive symptom remission: A meta‐analysis of randomized controlled trials. Pharmacogenomics 2019, 20, 37–47.
63. Rosenblat, J.D.; Lee, Y.; McIntyre, R.S. The effect of pharmacogenomic testing on response and remission rates in the acute treatment of major depressive disorder: A meta‐analysis. J. Affect. Disord. 2018, 241, 484– 491.
64. Bousman, C.A.; Zierhut, H.; Müller, D.J. Navigating the Labyrinth of Pharmacogenetic Testing: A Guide to Test Selection. Clin. Pharmacol. Ther. 2019, doi:10.1002/cpt.1432.
65. Bousman, C.A.; Dunlop, B.W. Genotype, phenotype, and medication recommendation agreement among commercial pharmacogenetic‐based decision support tools. Pharmacogenom. J. 2018, 18, 613–622.
66. Zeier, Z.; Carpenter, L.L.; Kalin, N.H.; Rodriguez, C.I.; McDonald, W.M.; Widge, A.S.; Nemeroff, C.B. Clinical Implementation of Pharmacogenetic Decision Support Tools for Antidepressant Drug Prescribing. Am. J. Psychiatry 2018, 175, 873–886.

Silvia Vilchez1, *, Mikel Tucson 1, Eduard Vieta 2,3,4, Enrique Alvarez 5,6 and Jordi Espadaler 1, *

1. “AB-Biotics, S.A.” (AB‐Biotics, S.A.), Av. de la Torre Blanca 57, 08172 Sant Cugat del Valles, Barcelona, Catalonia, Spain (Av. de la Torre Blanca 57, 08172 Sant Cugat del Valles, Barcelona, Catalunya, España)
2. Department of Psychiatry and Psychology, Institute of the Brain, 170 Villaroel Street, 08036 Barcelona, Catalonia, Spain (Departamento de Psiquiatría y Psicología, Instituto de Neurocencias, C/ Villaroel, 170, 08036 Barcelona, Catalunya, España)
3. August Pi I Sunyer Institute for Biomedical Research (IDIBAPS), University of Barcelona, 149 Rossello Street, 08036, Barcelona, Catalonia, Spain (Institut d’investigacions Biomèdiques August Pi I Sunyer (IDIBAPS), University of Barcelona, C/ Rosselló, 149, 08036 Barcelona, Catalunya, España)
4. Center for Biomedical Research in the Mental Health System (CIBERSAM), Av. Monforte de Lemos, 3-5, 28029 Madrid, Spain (Centro de Investigación Biomédica en Red de Salud Mental (CIBERSAM), Av. Monforte de Lemos, 3-5, 28029 Madrid, España)
5. Psychiatric Department, Santa Creu and Sant Pau Hospital, St. Anthony Maria Claret, 167, 08025 Barcelona, Catalonia, Spain (Servei de Psiquiatria, Hospital de la Santa Creu i Sant Pau, C/ Sant Antoni Maria Claret, 167, 08025 Barcelona, Catalunya, España)
6. Sant Pau Institute of Biomedical Research, Autonomous University of Barcelona, 08025, Barcelona, Catalonia, Spain (Institut d’investigacions Biomèdica Sant Pau, Universitat Autònoma de Barcelona, 08025 Barcelona, Cataluña, España)

* Contacts: vilches@ab-biotics.com (S.V.); espadaler@ab-biotics.com (J.E.)

LINK TO THE ORIGINAL STUDY (in English)