The Effect of ABCB1 Polymorphisms on the Efficacy of Antidepressants

ABSTRACT

This scoping review investigates the association between ABCB1 polymorphisms and antidepressant efficacy in humans. A systematic search identified 630 records, of which 58 met the inclusion criteria, resulting in 42 unique studies (five randomised controlled trials (RCTs), two randomised studies (non‐RCTs), 30 prospective cohort studies, three case–control studies, one cross‐sectional clinical study and one phase I clinical trial). These studies examined single nucleotide polymorphisms (SNPs) in or near the ABCB1 gene and their association with antidepressant treatment response. Of the 42 studies, 30 focused on rs1045642, the most extensively studied SNP. Among these, only 20% reported statistically significant associations. Beyond rs1045642, rs2032582 and rs1128503 were also frequently studied, but statistically significant associations were reported in only a minority of cases (28% and 13%, respectively), often with conflicting directions. Haplotype analyses involving all three SNPs (the TTT haplotype) showed mixed results. Results were variable across antidepressants, likely due to overlapping pharmacokinetic pathways. Methodological differences, including study design, sample sizes and definitions of remission, likely contribute to these inconsistencies. This review highlights the complexity of linking ABCB1 polymorphisms to antidepressant treatment response and suggests the need for standardised methodologies and larger, diverse populations in future studies. Haplotype analyses could provide deeper insights and enhance personalised treatment strategies.

Summary

ABCB1 is a protein that helps move drugs across cell membranes, including at the blood–brain barrier, where it may affect how people respond to antidepressants. We reviewed 42 studies looking at whether small genetic differences in the ABCB1 gene are linked to treatment success. The most commonly studied genetic variants showed mixed and often conflicting results. Overall, no clear pattern emerged. More research, using consistent study methods, is needed to understand if ABCB1 genetics can help predict who will benefit from antidepressants.

1. Introduction

In recent years, depression has emerged as a growing public health issue, with a steadily increasing incidence. This trend is paralleled by a rise in prescriptions of antidepressant medications [1, 2]. The efficacy of antidepressants varies widely, and approximately 30–40% of patients with depression fail to respond to treatment. This variability, combined with the increasing prevalence of depression, constitutes a significant healthcare challenge—both for individuals and society. The reasons behind the differential effectiveness of antidepressants remain unclear. One hypothesis suggests that polymorphisms in specific drug transporters may play a role. The term ‘polymorphisms’ refers to the occurrence of two or more variants of a DNA sequence within a gene in a population. This study investigates the significance of polymorphisms in the ABCB1 transporter.

ABCB1, also known as P‐glycoprotein (P‐gp) or multidrug resistance protein 1 (MDR1), is an ATP‐binding cassette (ABC) transporter located in the cell membranes of various tissues, including the intestinal epithelium, bile canaliculi, renal tubules and endothelial cells of the blood–brain barrier [3]. ABCB1 functions as an efflux transporter, pumping xenobiotics (e.g., antidepressants) out of the cell. Since many antidepressants are substrates of ABCB1, it is plausible that single nucleotide polymorphisms (SNPs) or copy number variations (CNVs) in the ABCB1 gene could influence the efficacy of these medications. CNVs refer to variations in the number of copies of a specific DNA segment among individuals, while SNPs represent changes in a single nucleotide within the genome. While ABCB's role at the blood–brain barrier is often highlighted in relation to antidepressant efficacy, the transporter is expressed in many tissues, and this review does not assume a specific site of action.

SNPs can be synonymous, causing no change in the amino acid sequence, or nonsynonymous, resulting in an altered amino acid sequence and, potentially, a modified protein. SNPs can also occur in non‐coding regions of the genome, which may influence gene transcription regulation without altering the protein's amino acid sequence [4]. SNPs in the ABCB1 gene are hypothesised to affect the expression, activity and substrate specificity of the ABCB1 protein [5]. For instance, an SNP that increases the activity of the ABCB1 protein could reduce the concentration of antidepressants in the central nervous system, potentially rendering the treatment insufficient [6].

In 2008, Uhr et al. demonstrated that specific SNPs in the ABCB1 gene influenced the efficacy of antidepressants classified as ABCB1 substrates (e.g., escitalopram), but not of those considered non‐substrates (e.g., mirtazapine) [7]. However, a definitive answer to whether ABCB1 polymorphisms impact antidepressant efficacy remains elusive.

We aim to map and synthesise the existing evidence on the association between ABCB1 polymorphisms and antidepressant efficacy in humans by conducting a scoping review. Our goal is to provide a comprehensive overview of the existing literature on whether ABCB1 polymorphisms influence antidepressant efficacy.

2. Materials and Methods

2.1. Protocol and Registration

The protocol for our scoping review was developed following the Joanna Briggs Institute (JBI) guidelines for conducting scoping reviews and registered on the Open Science Framework (https://osf.io/) [8].

2.2. Eligibility Criteria

We included all clinical studies that investigate the association between ABCB1 polymorphisms and antidepressant treatment outcomes, especially remission.

We excluded:

• Non‐human studies

• Reviews

• Case reports

2.3. Search

The studies used in this project were identified via Medline and Embase. The initial systematic search strategy for Medline, Ovid was the following: ((ATP Binding Cassette Transporter, Subfamily B/ OR ABCB1 OR P‐gp.mp OR P‐glycoprotein.mp OR MDR1.mp OR Multidrug Resistance Protein 1.mp) AND (Depression/OR Antidepressive Agents/OR Depressi*.mp OR Antidepress*.mp) AND (Polymorphism, Genetic/OR Polymorphism.mp OR Single Nucleotide Polymorphism.mp OR SNP.mp OR SNPs.mp OR Mutation/OR Pharmacogenetic*.mp OR DNA Copy Number Variations/OR DNA Copy Number.mp OR CNVs.mp)).

The search strategy was developed for Medline and translated for Embase. We then searched the reference lists of all included studies for additional relevant studies.

This systematic search yielded 627 records, and an additional three records were identified from citation searching. A total of 572 records were excluded, leaving 58 reports of included studies. Of these, 42 unique studies were ultimately included in our scoping review. These studies included five RCTs, two randomised studies (non‐RCTs), 30 prospective cohort studies, three case–control studies, one cross‐sectional clinical study and one phase I clinical trial. The study selection process is illustrated in the flowchart shown in Figure 1.

FIGURE 1.

PRISMA flowchart illustrating the study selection process in accordance with PRISMA guidelines. The diagram presents the number of studies identified, screened and included in the scoping review, along with the main reasons for exclusion at each stage of the process.

3. Results

The included studies explored the association between ABCB1 polymorphisms and antidepressant efficacy, with most studies focusing on remission as the primary outcome. A total of 42 studies evaluated various antidepressants, including escitalopram, sertraline, duloxetine, paroxetine, venlafaxine, fluoxetine, citalopram, fluvoxamine, desvenlafaxine, tricyclic antidepressants (TCAs) and mirtazapine, in relation to SNPs within or near the ABCB1 gene. The characteristics of the data from the included studies are summarised in Table 1, and detailed information on each study can be found in Table 2.

TABLE 1.

General characteristics of the included studies.

	N = 42 studies
Mean age range (years)	14.7–54.8
Sample size range	15–683
Efficacy outcome measure, n (%)
HAMD	34 (81.0%)
MADRS	4 (9.5%)
QIDS‐SR	2 (4.8%)
Tested antidepressant, n (%)
Amitriptyline	7 (16.7%)
Citalopram	10 (23.8%)
Clomipramine	3 (7.1%)
Desipramine	1 (2.4%)
Desvenlafaxine	1 (2.4%)
Duloxetine	4 (9.5%)
Escitalopram	10 (23.8%)
Fluoxetine	11 (26.2%)
Fluvoxamine	2 (4.8%)
Mirtazapine	9 (21.4%)
Nortriptyline	3 (7.1%)
Paroxetine	15 (35.7%)
Sertraline	7 (16.7%)
Venlafaxine	18 (42.9%)
Study type, n (sample size interval)
RCT	5 (73–683)
Randomised trial (non‐RCT)	3 (74–246)
Prospective cohort	29 (15–519)
Case–control	3 (145–253)
Cross‐sectional	1 (79) a
Phase 1 clinical trial	1 (83) a
AD plasma concentrations, n (%)	19 (45.2%)
Haplotype analyses, n (%)	9 (21.4%)

Note: Mean age range, sample sizes, efficacy outcome measures, investigated antidepressants, study types and the number of studies reporting antidepressant plasma concentrations or performing haplotype analyses is presented. Values in parentheses indicate either percentage of studies (reporting efficacy measure or antidepressant) or sample size ranges (for study types).

Abbreviations: AD, antidepressant; HAMD, Hamilton Depression Rating Scale; MadRS, Montgomery–Åsberg Depression Rating Scale; QIDS‐SR, Quick Inventory of Depressive Symptomatology – Self Report.

^aOnly one study in this category.

TABLE 2.

Overview of included studies on antidepressant efficacy (response and/or remission).

Author, year (reference number)	Study design	N	Mean age	Methods	Tested antidepressant	Efficacy endpoint	Tested ABCB1 SNP(s)	Results on efficacy	Antidepressant blood/plasma levels measured
Breitenstein et al., 2016 (15)	RCT	73	Standard dose: 46.3 (SD = 16.9) High dose: 48.5 (SD = 13.5)	Patients were randomly assigned to a pharmacogenetically guided treatment group or standard antidepressant treatment. Treatment response was evaluated after 28 days of using HAMD‐17. 128 controls received treatment as usual.	Paroxetine, sertraline, citalopram, escitalopram, venlafaxine, amitriptyline, amitriptylinoxide, nortriptyline, trimipramine	HAMD‐17 percent change at D28 served as primary endpoint for the assessment of treatment efficacy.	rs2032583 rs2235015	ABCB1 genotype alone was not significantly associated with treatment outcome (rs2032583: p = 0.442; rs2235015: p = 0.513).	Yes
Dong et al., 2009 (32)	RCT	142	38 (SD = 10)	The first week consisted of a single‐blind placebo lead‐in phase to minimise placebo response. Subjects meeting inclusion criteria were then randomised to receive either fluoxetine or desipramine in a double‐blind manner for 8 weeks.	Fluoxetine, desipramine	Remission: final (week 8) HAMD‐21 score < 8. The relative response change was also computed as the difference in HAMD‐21 score between pre‐ and post‐treatment divided by the pretreatment HAMD‐21 score.	81 SNPs, including: rs1128503 rs2032583 rs2235015 rs2235040 rs3842 rs17064 rs10276036 rs2235020	Two ABCB1 haplotypes were significantly associated with remission: ACA (desipramine) and GCGCACACGAGAC (fluoxetine) (p < 0.05).	No
Perlis et al., 2010 (20)	RCT	250	44.2 (SD = 12.6)	Patients were randomised to receive once‐daily duloxetine (30 or 60 mg) at the start of a 7‐week double‐blind treatment phase. Active treatment commenced after a 1‐week placebo lead‐in phase, followed by fixed‐dose duloxetine 60 mg/day for 5 weeks.	Duloxetine	Changes in HAMD‐17 from baseline	rs10280101 rs7787082 rs2032583 rs2235040	No associations between the four polymorphisms and antidepressant treatment response were found for rs10280101, rs7787082, rs2032583 and rs2235040 (p = 0.7316, p = 0.7197, p = 0.6911 and p = 0.6993, respectively)	No
Roberts et al., 2002 (31)	RCT	160	Fluoxetine: 33.3 (SD = 11.6) Nortriptyline: 30.3 (SD = 10.9)	After baseline assessment, patients were randomised to receive fluoxetine or nortriptyline for 6 weeks.	Fluoxetine, nortriptyline	HDRS and MADRS after 3 weeks and 6 weeks	rs1045642	There was no relationship between response to treatment and genotype.	Yes
Schatzberg et al., 2015 (26)	RCT	683	38.6 (SD = 12.8)	Participants were randomly assigned to escitalopram, sertraline, or extended‐release venlafaxine. QIDS‐SR assessments were conducted at baseline, during treatment‐phase telephone monitoring (weeks 1, 2, 4 and 6), and at the 8‐week post‐treatment follow‐up. The study included 336 healthy controls.	Escitalopram, sertraline, venlafaxine	Remission: score ≤ 5 on QIDS‐SR after 8 weeks	rs10245483 rs3213619 rs2214102 rs2235015 rs10276036 rs2032588 rs2235033 rs28381916 rs2032583 rs7793196	Only rs10245483 significantly predicted remission (p < 0.001). Common G allele homozygotes responded better to escitalopram (p = 0.032) and sertraline (p = 0.020), whereas minor T allele homozygotes responded better to venlafaxine (p = 0.018).	No
Chang et al., 2015 (28)	Randomised prospective study	112	39.7 (SD = 12.4)	Patients were randomly assigned to fluoxetine or venlafaxine extended‐release. Symptom severity was assessed at weeks 0, 2, 4 and 6 using HDRS‐21.	Fluoxetine, venlafaxine	Percentage change in HDRS‐21 after 6 weeks.	rs1045642 rs2032582 rs1128503	Patients with the G/G genotype of rs2032582 had a worse antidepressant treatment response.	No
Perroud et al., 2011 (42)	Partly randomised study	74	Subjects without increasing suicidal ideation: 36.36 (SD = 10.42) Subjects with increasing suicidal ideation: 37.14 (SD = 11.22)	Patients progressed through up to seven sequential treatment steps based on MADRS scores at each visit.	Paroxetine, clomipramine, venlafaxine, lithium, triiodothyronine, nefazodone	Remission: MADRS score of 8 or less	61A>G rs2032582 rs1045642	None of the polymorphisms predicted antidepressant response.	Yes
Sarginson et al., 2010 (35)	Double‐blind, randomised trial	246	Range 70–72	15 mg mirtazapine or 20 mg paroxetine daily. Dose increased at days 14, 28 and 42 (max. 45 mg mirtazapine or 40 mg paroxetine) based on CGI scores. Assessments with HDRS‐21 at weeks 1–4, 6 and 8.	Paroxetine, mirtazapine	Remission: HDRS‐21 score less than 10	rs10245483 rs3213619 rs2214102 rs9282564 rs2235015 rs10276036 rs2229109 rs2032588 rs2235033 rs28381916 rs2235063 rs2235040 rs2032582 rs2032583 rs1045642	No association between genotype and time to remission was found after correction.	No
Blázquez et al., 2016 (27)	Naturalistic 1‐year follow‐up study	46	14.96 (SD = 1.74)	Patients were treated with fluoxetine for the first time. Symptom severity was assessed at baseline using K‐SADS‐PL and after 12 months using CDI, CGI‐S and GAF.	Fluoxetine	Remission: patients with a relatively asymptomatic period of at least 14 days. Recovery: no MDD diagnosis at 12 months.	rs2032582 rs1045642	The polymorphisms were not associated with remission or recovery.	No
Bousman et al., 2017 (18)	Prospective, open‐label cohort study	119	49 (SD = 13)	Patients receiving desvenlafaxine were assessed for symptom severity using the HDRS‐17 at baseline and at multiple time points post‐baseline (weeks 2, 4, 6, 8 and 10).	Desvenlafaxine	Comparison of the predicted dose (using the CNSDose support tool) and actual dose required for symptom remission (defined as an HDRS score of 7 or less by week 10 of the study).	rs1045642	Analysis of individual CNSDose variants demonstrated moderate concordance between actual and predicted antidepressant doses (p = 0.0001).	No
Breitenstein et al., 2014 (14)	Naturalistic study	58	48.53 (SD = 15.19)	The efficacy of different treatment strategies was assessed in patients with unfavourable ABCB1 genotypes. These included dose adjustments, switching to a non‐substrate antidepressant (e.g., mirtazapine), or augmentation. Depression severity was monitored weekly using HAMD‐21.	Amitriptyline oxide, amitriptyline, nortriptyline, trimipramine, doxepin, citalopram, paroxetine, sertraline, escitalopram, venlafaxine, mirtazapine, bupropion, fluoxetine, monoamine oxidase (MAO), inhibitors, reboxetine, imipramine, trazodone, lithium, lamotrigine, quetiapine, olanzapine	Differences in HAM‐D scores and remission rates (defined as a HAM‐D score < 10) at discharge among patients with and without ABCB1 testing during their hospital stay were compared.	rs2032583 rs2235015	An increase in the dose of an antidepressant with P‐gp substrate properties in patients with the TT/GG genotype was associated with a shorter stay at the psychiatric hospital. Patients whose ABCB1 test result was received during hospital stay were more likely to be remitted at discharge compared with patients whose test results were unknown at the time of treatment (p = 0.005) and had lower HAM‐D scores at discharge (p = 0.0195).	Yes
Gassó et al., 2014 (29)	Prospective observational study	83	14.7 (SD = 1.7)	All subjects began fluoxetine treatment in the initial phase of the study. Clinical improvement was assessed at weeks 8 and 12.	Fluoxetine	Response: Scores obtained on the CGI‐I scale at weeks 8 and 12 weeks. Clinical improvement: Differences in CDI, OCI‐CV, SCARED, CGI‐S and GAF/CGAS scales between baseline scores and scores obtained after 8 and 12 weeks.	rs2032582 rs1045642	rs2032582 was significantly associated with clinical improvement. After 8 weeks, T‐allele carriers showed significant improvement on CDI (p = 0.03), OCI‐CV (p = 0.001) and CGI‐I (p < 0.001).	Yes
Geers et al., 2022 (10)	Naturalistic prospective cohort study	152	Male: 49.8 (SD = 9.7) Female: 49.5 (SD = 11.2)	Depression severity was assessed using HAMD‐17 before treatment, at weeks 2 and 4. HAMD‐17 scoring was conducted within 2 days of study inclusion, and treatment commenced immediately thereafter.	Sertraline, paroxetine, escitalopram, fluoxetine, fluvoxamine, citalopram, trazopodone, trazopodone + paroxetine, clomipramine, pipofezine, amitriptyline, venlafaxine, duloxetine, mirtazapine, mianserin, agomelatine	HAMD‐17 score between baseline and week 2, weeks 2 and 4, and baseline and week 4.	rs2235040 rs4148739 rs2235015 rs1045642 rs9282564 rs2032582	No significant difference in the improvement in the ΔHAMD‐17 score in the total period of 4 weeks was found for any of the SNPs.	No
Gex‐Fabry et al., 2008 (40)	Prospective observational study	63	37 (range 19–62)	Patients received 20–30 mg paroxetine. Depression severity was assessed using MadRS at baseline, week 2, week 4 and every 2 weeks thereafter until remission, study discontinuation, or treatment change. Median study duration was 6 weeks (range 2–18 weeks).	Paroxetine	Clinical response: 50% improvement from baseline MADRS score.	rs2032582 rs1045642 rs9282564 (61A>G)	ABCB1 genotype contributed to improving the model for persistent response significantly for 61A.G (p = 0.043), but not rs2032582 (p = 0.068) and rs1045642 (p = 0.11).	Yes
Huang et al., 2013 (16)	Prospective, open‐label study	290	36.0 (SD = 13.2)	Patients were randomly assigned to one of four SSRI treatments. Depression severity was assessed using HAMD at baseline and at weeks 1, 2, 4 and 6.	Paroxetine, fluoxetine, citalopram, sertraline	Response: HAMD score > 50% reduction from the baseline after 6 weeks.	rs6946119 rs28401781 rs4148739 rs3747802	No associations with SSRI treatment response in week 6 after correction.	No
Islam et al., 2024 (22)	Prospective cohort study	177	35.43 (SD = 12.70)	Patients received escitalopram for 8 weeks, followed by either continued escitalopram monotherapy or augmentation with aripiprazole for an additional 8 weeks.	Escitalopram, aripripazole	Response: reduction of ≥ 50% in MADRS from baseline. Remission: MADRS total score of 10 or less.	rs1045642 rs1128503 rs2032582 rs2032583 rs2235015 rs2235040	No significant associations were observed between DNA methylation levels at individual CpG sites in ABCB1 gene in whole blood and treatment response or remission after correction for multiple testing.	Yes
Jelén et al., 2015 (43)	Prospective observational study	90	42.6 (SD = 11.1)	The 17‐item HDRS score was used to evaluate the severity of depressive symptoms before and after treatment. The control group consisted of 96 healthy blood donors recruited from the same geographical area as the patients.	SSRIs, SSRIs + another antidepressant, venlafaxine, venlafaxine + another antidepressant, agomelatine, a combination of other antidepressants	Severity of symptoms before (HDRS I) and after treatment (HDRS II) and change in the HDRS‐17 score (HDRS change = HDRS I—HDRS II).	rs1045642	Patients with the CC genotype had significantly higher change in HDRS scores (p = 0.0301). Likewise, patients with the T allele had decreased therapy effectiveness (p = 0.0142) compared to C allele carriers.	No
Jelén et al., 2019 (45)	Prospective observational study	92	49 (SD = 10)	The HDRS‐17 was used to assess depressive symptom severity before and after treatment. The control group included 96 healthy blood donors from the same geographical area as the patients.	SSRIs, SSRIs + another antidepressant, venlafaxine, venlafaxine + another antidepressant, agomelatine, a combination of other antidepressants	Effectiveness = initial HDRS–HDRS after treatment.	rs1128503 rs2032582	Mean reductions in HDRS scores were greater in wild‐type homozygotes (1236 CC and 2677 GG) compared to heterozygous 1236 CT carriers (p = 0.0307) and 2677 AG or AT (p = 0.0179), respectively.	No
Jelén et al., 2023 (44)	Prospective observational study	102	Median age 51 (range 17–63)	The HDRS‐17 was assessed at hospital admission during the symptomatic phase before treatment initiation and again after 8 weeks of antidepressant treatment. The control group comprised 94 Caucasian blood donors from the same geographic area.	SSRIs, SSRIs + another antidepressant, venlafaxine, venlafaxine + another antidepressant, agomelatine, a combination of other antidepressants	Clinical and genetic factors associated with a high HDRS score (> 22) at baseline and with a high HDRS change (> 15) after treatment.	T‐129C rs1128503 rs2032582 rs1045642	The presence of at least one variant allele (1236T, 2677T/A or 3435T) was significantly associated with declined treatment response. When particular SNPs were analysed in pairs, an association was also clearly seen. At least one 1236T in combination with 3435T significantly decreases the chances for reduced depression symptoms after treatment (p = 0.0030).	No
Kato et al., 2008 (39)	Prospective observational study	68	45.1 (SD = 15)	Following a 5‐day washout period, paroxetine was administered at a fixed dose of 20 mg/day. Depressive symptoms were assessed using HAMD‐21 at baseline and after 6 weeks of treatment.	Paroxetine	Percentage change in HAMD‐21 scores over time.	rs1045642 rs2032582 rs1128503	Analysis of covariance at week 6, controlling for baseline scores, showed a significant association between rs2032582 and response to paroxetine (p = 0.011, after correction p = 0.033). No significant association was found between HAMD score change and rs1045642 or rs1128503. The haplotype 3435C–2677G–1236T associated with poor response (p = 0.006).	Yes
Laika et al., 2006 (46)	Prospective observational study	50	Not reported	Patients received amitriptyline at a fixed dose of 75 mg twice daily for 3 weeks. HAMD and CGI scales were assessed weekly.	Amitriptyline	HAMD + CGI scale scores on day 21.	rs2032582	The SNP was not associated with therapeutic response.	Yes
Lee et al., 2010 (17)	Prospective observational study	64	52.52 (SD = 15.79)	A 2‐week washout period was implemented for patients previously on medication. During the 8‐week treatment phase, all patients received citalopram (10–60 mg/day) without concurrent antidepressants. Clinical symptoms were assessed at baseline and at weeks 1, 2, 4 and 8. The control group comprised 100 healthy controls.	Citalopram	Remission: HDRS score lower than 8	rs12929977 rs4148330 rs35587 rs4148356 rs2074087 rs2239330 rs212087 rs4148379 rs212090	The synonymous SNP rs2239330 (c.4002G>A) was significantly associated with remission (p = 0.005) and forms a linkage disequilibrium block with three other SNPs, including rs212090 (c.5462T>A) in the 3′UTR. The haplotype showed a significant association with the remission state (p = 0.014). rs212090 also showed a potential association (p = 0.035).	No
Lin et al., 2011 (24)	Open‐label study design	100	42	Patients received escitalopram at a fixed dose of 10 mg for the first 4 weeks, followed by dose adjustments (10–30 mg/day) based on clinical response over an 8‐week treatment period. Assessments using HAMD‐21, HAM‐A and the GCI scale were conducted at weeks 2, 4, 6 and 8.	Escitalopram	Remission: HAMD score of less than 10 after 8 weeks.	rs1882478 rs2235048 rs2235047 rs1045642 rs6949448 rs10234411 rs3789246 rs1922242 rs2235046 rs1128503 rs2235018 rs2235016 rs10256836 rs1989831 rs1202184 rs3789243 rs2188524 rs28381796 rs4148732 rs1978095	Rs1882478 (p = 0.037), rs1045642 (p = 0.045) and rs10256836 (p = 0.021) were significantly associated with remission. The rs10256836 SNP had more nonremitters in the minor C allele than remitters. In contrast, rs1882478 and rs1045642, had more remitters for the minor allele (C and T, respectively) than nonremitters. The haplotype block, rs1882478‐rs2235048‐rs2235047‐rs1045642‐rs6949448 of ABCB1 was found strongly associated with remission rate (t = −4.92, p = 0.003) in which haplotype T‐T‐T‐C‐C was associated with a slower remission rate on S‐CIT treatment (p = 0.001). Another haplotype block also showed significant association with MDD remission in the T‐G‐C‐A‐C combination (t = −3.3, p = 0.047)	Yes
Ma et al., 2020 (21)	Prospective observational study	190	Not reported	Patients received a single antidepressant: venlafaxine (75–225 mg/day), mirtazapine (7.5–45 mg/day), or duloxetine (40–120 mg/day). Symptom severity was assessed before treatment and at weeks 1, 2, 4 and 6 of continuous treatment.	Venlafaxine, mirtazapine, duloxetine	Response: 50% or more reduction in the HAMD score at 6 weeks compared to baseline.	rs6946119 rs28401781 rs4148739 rs3747802 Haplotype rs28401781‐rs4148739	No SNP or haplotype showed a significant association with antidepressant treatment response after correction.	No
Menu et al., 2010 (11)	Prospective open naturalistic 4‐week study	117	45.4 (SD = 14.5)	Antidepressant treatment was selected by the treating psychiatrist under naturalistic conditions and prescribed within the therapeutic range. Patients were assessed at baseline and after 28 days of continuous treatment.	Fluoxetine, paroxetine, citalopram, escitalopram, other SSRIs, venlafaxine, TCAs (clomipramine, amitriptyline, dosulepine, imipramine), other antidepressants (mainly mirtazapine)	Response: > 50% decrease of HDRS‐17 score between baseline and week 4	rs1045642	No significant association of the SNP with treatment response	No
Mihaljevic Peles et al., 2008 (41)	Prospective, open‐labelled, study	127	52 (SD = 11)	Following a 5‐day washout period, paroxetine was administered at a fixed dose of 20 mg/day. Depressive symptoms were assessed using HAMD‐17 at baseline and after 6 weeks of treatment.	Paroxetine	Response: decrease in HAMD‐17 by ≥ 50% from baseline to week 6	rs2032582 rs1045642	No significant associations were observed between antidepressant treatment response and SNPs exon 21 G2677T (p = 0.25), exon 26 C3435 T (p = 0.37), or the associated haplotypes H1 (G–C), H2 (G–T), H3 (T–C) and H4 (T–T), based on rs2032582 (G2677T) and rs1045642 (C3435T) (p = 0.384).	No
Nikisch et al., 2008 (13)	Prospective observational study	15	36.2 (SD = 8.3)	All patients received escitalopram, starting at 20 mg once daily, increasing to 40 mg daily from day 5 onwards.	Citalopram	Response: ≥50% decrease in the HAMD‐21 after 4‐week treatment with R,S‐CIT.	rs1045642 rs2032582	The rs2032582 GG/GT genotype was associated with a better treatment response (p = 0.001) compared to the TT genotype.	Yes
Ozbey et al., 2014 (12)	Prospective, open‐label study	54	39.4 (SD = 13.4)	Citalopram was initiated at 20 mg/day, with dose adjustments up to 30 mg/day based on therapeutic response. The control group comprised 70 unrelated volunteers (34 females, 36 males) with no history of MDD.	Citalopram	Response: 50% reduction from the baseline HAMD‐17 score at the sixth week.	rs1045642	There was no significant difference for the distribution of 3435C/T between the responders and nonresponders (p = 1.000). The HAMD scores did not show any statistically significant differences according to genotype (p = 0.279)	Yes
Ozbey et al., 2017 (47)	Prospective, open‐label study	52	38.1 (SD = 10.8)	Venlafaxine was initiated at 37.5 mg/day, with dose adjustments based on therapeutic response. Treatment response was assessed using HDRS‐17 at baseline and at weeks 1, 2, 4 and 6.	Venlafaxine	HDRS‐17 after 6 weeks	rs1045642 rs2032582	The HDRS‐17 total scores did not show any statistically significant difference for the rs1045642 and rs2032582 genotypes (p = 0.850 and p = 0.577, respectively).	Yes
Santos et al., 2024	Prospective cohort study	80	40.48 (SD = 11.06)	The TMAP for MDD treatment algorithm guided the selection of optimal medication strategies.	Not reported	Remission: BDI < 10 after 6 weeks of adequate antidepressant treatment and no longer meeting SCID‐I criteria for MDD. Relapse: Any depressive episode after achieving remission during follow‐up. Treatment‐resistant depression: BDI ≥ 10 and continued MDD diagnosis (SCID‐I) despite two adequate antidepressant treatments.	rs1128503 rs1045642 rs2032582	The three SNPs showed no significant associations with relapse (all p > 0.05) or treatment‐resistant depression (all p > 0.05). For rs1045642, TT‐genotype carriers achieved remission significantly faster than CC or CT carriers (p = 0.028).	Not reported
Scherf‐clavel et al., 2022 (37)	Naturalistic cohort study	519	46.62 (SD = 14.12)	Patients were treated according to the doctor's choice.	Amitriptyline, venlafaxine, mirtazapine, quetiapine	Response: ≥50% reduction in HAMD‐21 from baseline. Remission: HAMD ≤ 7. Outcome assessments: week 7 (Wuerzburg sample) and week 6 (Munich sample).	rs1045642 rs1128503 rs2032582	No association was found between SNPs and response to drug therapy.	Yes
Simoons et al., 2020 (38)	Prospective cohort study	81	Responders: 44.8 (SD = 1.8) Nonresponders: 43.0 (SD = 1.3)	Patients received paroxetine 20 mg/day for 6 weeks. HDRS‐17 was administered at baseline and after 6 weeks of treatment.	Paroxetine	Response: ≥50% decrease in HDRS‐17 score	rs1045642 rs1128503 rs2032582 rs2235040	None of the four studied SNPs nor the rs1045642C‐rs2032582G‐rs1128503T‐haplotype were significantly associated with clinical response (p = 0.13, p = 0.28, p = 0.83, p = 0.20 and p = 0.72, respectively).	Yes
Singh et al., 2012 (25)	Prospective cohort study	98	Range 34–43	Patients initially received escitalopram (10 mg/day) or venlafaxine (75 mg/day), with dose adjustments at weeks 1, 4 and 8 based on clinical response and side effects.	Escitalopram, venlafaxine	Response: change by >50% on the HDRS‐17 after 8 weeks from baseline. Remission: ≤7 or less onHDRS‐17 after 8 weeks.	rs1045642 rs2032582 rs1128503	Patients carrying the TT genotype of rs1045642 required significantly lower escitalopram doses to achieve remission compared to TC or CC carriers (p = 0.0001). Among venlafaxine‐treated patients, remission rates were significantly higher in subjects carrying the TT genotype at rs1045642 compared to those with the CC genotype (p = 0.006). No significant associations were found for rs2032582 or rs1128503.	No
Uhr et al., 2008 (7)	Prospective cohort study	Not reported	For rs2032583: range 48–48.6 For rs2235015: range 46.9–49	Patients were grouped by antidepressant and treated for at least 4 weeks. 362 controls.	Citalopram, paroxetine, amitriptyline, venlafaxine, mirtazapine	Remission: HAMD‐21 score of 10 or less after 4, 5 and 6 weeks	rs1045642 rs2032582 rs2032583 rs4148739 rs2235040 rs2235015 rs1128503 rs2235067 rs4148740 rs10280101 rs7787082 rs11983225 rs1024842 rs12720067 67 other SNPs	Significant associations with antidepressant response were reported for rs2032583, rs2235067, rs4148740, rs10280101, rs7787082, rs2032583, rs4148739, rs11983225, rs10248420, rs2235040, rs12720067 and rs2235015. Specifically, carriers of the C allele (rs2032583) and T allele (rs2235015) showed better treatment outcomes. There was no significant effect for the non‐substrates.	Yes
Vancova et al., 2018 (5)	Prospective, open‐label study	61	40.85 (SD = 12.83)	Patients were assessed with HAMD‐21 at weeks 0, 2, 4 and 6. Following washout, paroxetine was initiated at 10–20 mg/day and increased to 40 mg/day from days 12–15.	Paroxetine	Response: decrease in HAMD‐21 of ≥50% at week 6. Remission: score ≤ 7 points on HAMD‐21 at week 6	rs2032582	Patients carrying at least one T allele showed significantly greater therapeutic efficacy with paroxetine at week 4 (p = 0.049) and week 6 (p = 0.001). However, improvement measured by HAMD‐21 score reduction was not significantly greater.	No
Zastrozhin et al., 2021 (36)	Prospective, open‐label study	108	35.2 (SD = 15.1)	The treatment regimen included mirtazapine at an average daily dose of 45 mg.	Mirtazapine	HADS and HAMD to evaluate efficacy on weeks 1, 4 and 8	rs1045642	No association was demonstrated	Yes
Zastrozhin et al., 2021 (34)	Prospective, open‐label study	105	37.5 (SD = 13.2)	The treatment regimen included fluvoxamine at an average daily dose of 117.6 ± 44.3 mg.	Fluvoxamine	HADS and HAMD to evaluate efficacy on weeks 1, 4 and 8	rs1045642	No association was demonstrated	Yes
Crisafulli et al., 2013 (33)	Case control association study	145	41.37 (SD = 14.07)	Hospitalised patients were assessed for illness severity at baseline and discharge using HAMD for major depression. A control group of 170 psychiatrically healthy Korean subjects underwent the same assessment.	Paroxetine, venlafaxine, fluoxetine, mirtazapine	Response: a priori ≥50% reduction from baseline to discharge in the HAMD. Remission: a HAMD score ≤ 7 at discharge for MD patients.	rs2235047 rs2229107 rs6961419 rs1922241 rs1202167 rs3789243	The study found no significant association between the investigated SNPs and related haplotypes, a diagnosis of MD and clinical response to treatment in the present study.	No
Peters et al., 2008 (9)	Two‐stage case–control study	Not reported	Not reported	Patients received citalopram for a planned 12‐week treatment period with flexible dosing (20–60 mg/day).	Citalopram	Response: subjects who had at least 42 days of treatment and whose QIDS‐SR score on their final clinical visit shows ≥50% reduction in score compared to baseline. Remission: a final QIDS‐SR score ≤ 5.	rs1128503 rs2032582 rs1045642 rs2235040 rs10280101 rs12720067	None of the polymorphism was significantly associated with the response phenotype.	No
Shan et al., 2019 (19)	Case–control study	253	30.89 (SD = 10.92)	Eligible patients received one of five antidepressants for 6 weeks. Assessments were conducted at screening, baseline, and at weeks 1, 2, 4 and 6. The study included 208 healthy controls.	Escitalopram, paroxetine, sertraline, duloxetine, venlafaxine	Response: changes in the HAMD‐17 total score of ≥50%	rs1045642 rs2032583 rs2032582 rs2235040 rs1128503 rs2235015	In the SSRI group, no significant correlation between SNPs and therapeutic response was found. In the SNRI group, only rs2032583 was significantly associated with therapeutic response. For rs2032583, patients with the TT genotype showed greater improvement in HAMD‐17 scores compared to CT carriers.	No
Magãlhaes et al., 2020 (30)	Cross‐sectional, observational, clinical study	79	54.8 (SD = 12.1)	This study aimed to describe and characterise the GnG‐PK/PD‐AD study and the population of depressed patients treated with fluoxetine, focusing on antidepressant pharmacokinetics, clinical outcomes and relevant genetic and non‐genetic factors.	Fluoxetine	Remission: HAMD score ≤ 7	rs1128503 rs2032582 rs1045642	Carriers of the TTT‐haplotype (comprising the T alleles of 1236C>T‐2677G>T/A‐3435C>T) showed a higher likelihood to be remitters compared to the non‐TTT and TTT–TTT haplotypes (p = 0.003 and p = 0.025, respectively).	Yes
Ray et al., 2015 (23)	Phase 1 clinical trial	83	47.34 (SD = 10.77)	A 12‐week trial with flexible dosing adjustments as needed. Starting and maximum doses were as follows: sertraline (50–200 mg/day), venlafaxine XR (37.5–375 mg/day) and escitalopram (10–20 mg/day). Symptom severity was assessed using HDRS‐24 at baseline and at weeks 2, 4, 6, 8, 10 and 12 in phase 1.	Sertraline, venlafaxine, escitalopram	Remission: HDRS‐24 score of ≤ 10 on the at any time point during treatment.	rs2235040 rs2032583 rs2235015 rs1045642 rs2032582 rs9282564	Minor allele carriers of rs2235040 and rs9282564 are significantly associated with remission and time to remission after controlling for the other ABCB1 SNPs.	No

Abbreviations: BDI, Beck Depression Inventory; CGAS, Children's Global Assessment Scale; CGI‐I, Clinical Global Impression–Improvement; CGI‐S, Clinical Global Impression Severity; CNSDose, commercial pharmacogenetic‐based decision support tool; ESC, escitalopram; FLX, fluoxetine; GAF, Global Assessment of Functioning; HAMD‐17/HAMD‐21, Hamilton Depression Rating Scale (17‐ or 21‐item version); HadS, Hospital Anxiety and Depression Scale; HAM‐A, Hamilton Anxiety Rating Scale; HDRS‐17/HDRS‐21, Hamilton Depression Rating Scale (17‐ or 21‐item version); MADRS, Montgomery–Åsberg Depression Rating Scale; MDD, major depressive disorder; OCD‐CV, Obsessive‐Compulsive Inventory–Child Version; QIDS‐SR, Quick Inventory of Depressive Symptomatology – Self Report; SCARED, Screen for Child Anxiety Related Disorders; SCID‐I, Structured Clinical Interview for DSM‐IV Axis I Disorders; SD, standard deviation; SNRI, serotonin–norepinephrine reuptake inhibitor; SSRI, selective serotonin reuptake inhibitor; TMAP, Texas Medication Algorithm Project; VEN, venlafaxine; XR, extended‐release; YMRS, Young Mania Rating Scale.

3.1. General Findings

Most studies used the Hamilton Depression Rating Scale (HAMD, HDRS) to assess treatment response, typically defining remission as a 50% reduction in score from baseline. While specific SNPs were associated with remission in some studies, other studies investigating the same SNPs reported no statistically significant associations or even conflicting results. This variability was observed across nearly all antidepressants. None of the studies examined CNVs in relation to antidepressant response. In the following section, we report the effects of various antidepressants in relation to the investigated SNPs across the 42 included studies. An ‘rs’ followed by a number indicates an rsID, a reference number used to identify specific SNPs. For each assessed SNP and drug, we describe if the study found a positive or negative association. Details on the identified associations can be found in Table 2. Table 3 illustrates the associations between SNPs from the included studies and antidepressant efficacy, measured as response and/or remission. Figure 2 illustrates association results for the three most investigated SNPs.

TABLE 3.

Overview of ABCB1 single nucleotide polymorphisms (SNPs) investigated in relation to antidepressant efficacy (response and/or remission) across the included studies.

Author, year (reference number)	rs1045642	rs2032582	rs2032583	rs2235015	rs1128503	rs2235040	rs4148739	Other SNPs	Haplotypes	Antidepressant
Blázquez et al., 2016 (27)										Fluoxetine
Bousman et al., 2017 (18)										Desvenlafaxine
Breitenstein et al., 2014 (14)			TT	GG						Amitriptyline, nortriptyline, imipramine, trimipramine, doxepin, citalopram, escitalopram, sertraline, paroxetine, venlafaxine, fluoxetine, trazodone, reboxetine and olanzapine
										Amitriptyline oxide, mirtazapine, bupropion, monoamine oxidase inhibitors, lithium, lamotrigine and quetiapine.
Breitenstein et al., 2016a (15)										Paroxetine, sertraline, citalopram, escitalopram, venlafaxine, amitriptyline, amitriptylinoxide, nortriptyline, trimipramine
Chang et al., 2015 (28)		GG								Fluoxetine, venlafaxine
Crisafulli et al., 2013 (33)								rs2235047, rs2229107, rs6961419, rs1922241, rs1202167, rs3789243		Paroxetine, venlafaxine, fluoxetine, mirtazapine
Dong et al., 2009a (32)								77 other SNPs, including: rs3842,	GCGCACACGAGAC in block 2	Fluoxetine
								rs17064, rs10276036, rs2235020	ACA in block 1	Desipramine
Gassó et al., 2014 (29)		T allele								Fluoxetine
Geers et al., 2022 (10)								rs9282564		Sertraline, paroxetine, escitalopram, fluoxetine, fluvoxamine, citalopram, trazopodone, trazopodone + paroxetine, clomipramine, pipofezine, amitriptyline, venlafaxine, duloxetine, mirtazapine, mianserin, agomelatine
Gex‐Fabry et al., 2008 (40)								61A>G (rs9282564)		Paroxetine
Huang et al., 2013 (16)								rs6946119, rs28401781, rs3747802		Paroxetine, fluoxetine, citalopram, sertraline
Islam et al., 2024 (22)										Escitalopram, aripripazole
Jelén et al., 2015 (43)	CC									SSRIs, SSRIs + another antidepressant, velafaxine, venlafaxine + another antidepressant, agomelatine, a combination of other antidepressants
Jelén et al., 2019 (45)		GG (compared to AG or AT)			CC (compared to CT)					SSRIs, SSRIs + another antidepressant, velafaxine, venlafaxine + another antidepressant, agomelatine, a combination of other antidepressants
Jelén et al., 2023 (44)	3435 T	2677 T/A			1236 T			T‐129C		SSRIs, SSRIs + another antidepressant, velafaxine, venlafaxine + another antidepressant, agomelatine, a combination of other antidepressants
Kato et al., 2008 (39)									3435C–2677G–1236T	Paroxetine
Laika et al., 2006 (46)										Amitriptyline
Lee et al., 2010 (17)								rs2239330 (c.4002G>A), rs212090 (c.5462T>A)	rs2239330 ‐ rs212090 ‐ rs212087 ‐ rs4148379	Citalopram
								rs12929977, rs4148330, rs35587, rs4148356, rs2074087, rs212087, rs4148379
								rs1882478 (minor allele, C)	rs1882478‐rs2235048‐rs2235047‐rs1045642‐rs6949448 (T‐T‐T‐C‐C)
Lin et al., 2011 (24)	Minor allele (T)							rs10256836 (minor allele, C)	rs1922242‐rs2235046‐rs1128503‐rs2235018‐rs10256836 (T‐G‐C‐A‐C)
								rs2235048, rs2235047, rs6949448, rs10234411, rs3789246, rs1922242, rs2235046, rs2235018, rs2235016, rs1989831, rs1202184, rs3789243, rs2188524, rs28381796, rs4148732, rs1978095		Escitalopram
Ma et al., 2020 (21)								rs6946119, rs28401781, rs3747802	rs28401781‐rs4148739	Venlafaxine, mirtazapine, duloxetine
Magalhães et al., 2020 (30)									TTT‐haplotye (1236C>T‐2677G>T/A‐3435C>T)	Fluoxetine
Menu et al., 2010 (11)										Fluoxetine, paroxetine, citalopram, escitalopram, other SSRIs, venlafaxine, TCAs (clomipramine, amitriptyline, dosulepine, imipramine), other antidepressants (mainly mirtazapine)
Mihaljevic Peles et al., 2008 (41)									H1 (G–C), H2 (G–T), H3 (T–C) and H4 (T–T), based on rs2032582 (G2677T) and rs1045642 (C3435T)	Paroxetine
Nikisch et al., 2008 (13)		GG/GT (compared to TT)								Citalopram
Ozbey et al., 2014 (12)										Citalopram
Ozbey et al., 2017 (47)										Venlafaxine
Perlis et al., 2010a (20)								rs10280101, rs7787082		Duloxetine
Perroud et al., 2011 (42)								61A>G (rs9282564)		Paroxetine, clomipramine, venlafaxine, lithium, triiodothyronine, nefazodone
Peters et al., 2008 (9)								rs10280101, rs12720067		Citalopram
Ray et al., 2015 (23)						Minor allele		rs9282564 (minor allele)		Sertraline, venlafaxine, escitalopram
Roberts et al., 2002a (31)										Fluoxetine, nortriptyline
Santos et al., 2024	TT (compared to CC or CT)									Not reported
Sarginson et al., 2010 (35)								rs10245483, rs3213619, rs2214102, rs9282564, rs10276036, rs2229109,		Paroxetine, mirtazapine
								rs2032588, rs2235033, rs28381916, rs2235063
Schatzberg et al., 2015a (26)								rs10245483 (common G allele homozygotes for escitalopram and sertraline; minor T allele homozygotes for venlafaxine)		Escitalopram, sertraline, venlafaxine
								rs3213619, rs2214102, rs10276036, rs2032588, rs2235033, rs28381916, rs7793196
Scherf‐clavel et al., 2022 (37)										Amitriptyline, venlafaxine, mirtazapine, quetiapine
										Escitalopram, paroxetine, sertraline
Shan et al., 2019 (19)			TT (compared to CT)							Duloxetine, venlafaxine
Simoons et al., 2020 (38)									rs1045642C‐rs2032582G‐rs1128503T	Paroxetine
Singh et al., 2012 (25)	TT (compared to TC or CC)									Escitalopram
	TT (compared to CC)									Venlafaxine
			C allele	T allel				rs2235067, rs4148740, rs10280101, rs7787082, rs11983225, rs10248420, rs12720067
Uhr et al., 2008 (7)								67 other SNPs		Citalopram, paroxetine, amitriptyline, venlafaxine
								rs2235067, rs4148740, rs10280101, rs7787082, rs11983225, rs10248420, rs12720067, 67 other SNPs		Mirtazapine
Vancova et al., 2018 (5)		T allele								Paroxetine
Zastrozhin et al., 2021 (36)										Mirtazapine
Zastrozhin et al., 2021 (34)										Fluvoxamine

Note: Each cell contains the specific genotype associated with the reported outcome. Green indicates a statistically significant positive association; blue indicates a statistically significant negative association; red indicates no statistically significant association; and purple indicates an association without specified direction.

^aIndicates that the study is a randomised controlled trial (RCT).

FIGURE 2.

Illustration of association findings for the three most studied SNPs. The three most studied SNPs were (A) rs1045642 (n = 30), (B) rs2032582 (n = 25) and (C) rs1128503 (n = 15). Green indicates studies reporting a statistically significant association with antidepressant response or remission; red indicates no statistically significant association. For each SNP, the study type distribution and the number of found associations for each study type are described.

3.2. Findings by Antidepressants

3.2.1. Citalopram

rs1045642: No statistically significant association with treatment response was reported in six studies [7, 9, 10, 11, 12, 13].

rs2032582: Three studies found no statistically significant association with treatment response [7, 9, 10]. However, one study reported a positive association with treatment response [13].

rs2032583: One study demonstrated that the implementation of ABCB1 genotyping as a diagnostic tool influenced clinical decisions and improved treatment outcomes when higher doses of P‐gp substrate antidepressants were administered [14], while another found no association [15]. A third study found a positive association with treatment response [7].

rs2235015: One study demonstrated that the implementation of ABCB1 genotyping as a diagnostic tool influenced clinical decisions and improved treatment outcomes when higher doses of P‐gp substrate antidepressants were administered [14], while two other studies found no association [10, 15]. Another study found a positive association with treatment response [7].

rs1128503: No statistically significant association with treatment response was reported in two studies [7, 9].

rs2235040: No statistically significant association with treatment response was reported in two studies [9, 10]. One study found an association with treatment response [7].

rs4148739: Two studies found no association with treatment response [10, 16], while one found an association [7].

rs2239330 (c.4002G>A) and rs212090 (c.5462T>A): One study found that c.4002G>A was strongly associated with remission status. Additionally, c.5462T>A demonstrated a possible association with remission status [17].

Other SNPs: No statistically significant association with treatment response was reported for rs9282564 [10], rs6946119, rs28401781 and rs3747802 [16], rs12929977, rs4148330, rs35587, rs4148356, rs2074087, rs212087 and rs4148379 [17]. A positive association was reported for the haplotype rs2239330‐rs212090‐rs212087‐rs4148379 [17]. Another study examined 67 additional SNPs, none of which demonstrated a significant association with the outcome [7].

3.2.2. Desvenlafaxine

rs1045642: One study highlighted moderate concordance between actual and predicted doses based on ABCB1 genotyping [18].

3.2.3. Duloxetine

rs1045642: Two studies reported no statistically significant association with response in duloxetine‐treated patients [10, 19].

rs2032582: No statistically significant association with duloxetine response was observed in two studies [10, 19].

rs2032583: Conflicting results were reported. While one study identified a positive association with response in duloxetine‐treated patients [19], another found no statistically significant association [20].

rs2235015: No statistically significant association with remission was observed in two studies [10, 19].

rs1128503: One study found no statistically significant association with response [19].

rs2235040: Three studies reported no statistically significant association with response [10, 19, 20].

Other polymorphisms: No statistically significant association with duloxetine response was found for rs9282564 [10], rs10280101, rs7787082 [20], rs4148739, rs6946119, rs28401781 and rs3747802 [21]. Additionally, rs28401781‐rs4148739 was not statistically significantly associated with treatment response [21].

3.2.4. Escitalopram

rs1045642: Five studies reported no statistically significant association with remission [10, 11, 19, 22, 23], while two studies identified a positive correlation [24, 25].

rs2032582: Five studies found no statistically significant association [10, 19, 22, 23, 25].

rs2032583: One study demonstrated that the implementation of ABCB1 genotyping as a diagnostic tool influenced clinical decisions and improved treatment outcomes when higher doses of P‐gp substrate antidepressants were administered [14], whereas five studies found no evidence of a relationship [15, 19, 22, 23, 26].

rs2235015: One study demonstrated that the implementation of ABCB1 genotyping as a diagnostic tool influenced clinical decisions and improved treatment outcomes when higher doses of P‐gp substrate antidepressants were administered [14], whereas six studies found no evidence of a relationship [10, 15, 19, 22, 23, 26].

rs1128503: No statistically significant association with response was observed in four studies [19, 22, 24, 25].

rs2235040: One study reported a statistically significant positive association with response [23]; whereas three studies found no evidence of an association [10, 19, 22].

rs9282564: A positive association with response was identified in one study [23], while one study found no significant correlation [10].

Other polymorphisms: Several other SNPs showed no statistically significant association with response, including rs4148739 [10], rs3213619, rs2214102, rs10276036, rs2032588, rs2235033, rs28381916, rs7793196 [26], rs2235048, rs2235047, rs6949448, rs10234411, rs3789246, rs1922242, rs2235046, rs2235018, rs2235016, rs1989831, rs1202184, rs3789243, rs2188524, rs28381796, rs4148732 and rs1978095 [24]. rs1882478 was associated with a positive response and rs10256836 with a negative response [24]. One study found that rs10245483 was associated with a positive response [26]. Haplotypes rs1882478‐rs2235048‐rs2235047‐rs1045642‐rs6949448 (T‐T‐T‐C‐C) and rs1922242‐rs2235046‐rs1128503‐rs2235018‐rs10256836 (T‐G‐C‐A‐C) were negatively associated with response [24].

3.2.5. Fluoxetine

rs1045642: No statistically significant association with antidepressant treatment response was reported in seven studies [10, 11, 27, 28, 29, 30, 31].

rs2032582: Three studies found no statistically significant association with treatment response [10, 27, 30]. One reported a negative association with antidepressant treatment response [28], while one study identified a significant association with clinical improvement following fluoxetine treatment [29].

rs2032583: One study demonstrated that the implementation of ABCB1 genotyping as a diagnostic tool influenced clinical decisions and improved treatment outcomes when higher doses of P‐gp substrate antidepressants were administered [14], while one found no association [32].

rs2235015: One study demonstrated that the implementation of ABCB1 genotyping as a diagnostic tool influenced clinical decisions and improved treatment outcomes when higher doses of P‐gp substrate antidepressants were administered [14]. Two studies found no statistically significant association with antidepressant treatment response [10, 32].

rs1128503: No statistically significant association with treatment response was found in three studies [28, 30, 32].

rs2235040: No statistically significant association with antidepressant treatment response was reported in two studies [10, 32].

rs4148739: No statistically significant association with antidepressant treatment response was reported in two studies [10, 16].

Other SNPs: No association was reported for rs2235047, rs2229107, rs6961419, rs1922241, rs1202167, rs3789243 [33], rs3832, rs17064, rs10276036, rs2235020 [32], rs9282564 [10], rs6946119, rs28401781 and rs3747802 [16].

One study found that the 1236T‐2677T‐3435T (TTT) haplotype (rs1128503 (1236C>T), rs2032582 (2677G>T/A), rs1045642 (3435C>T)) was associated with better clinical outcomes with fluoxetine, specifically a higher likelihood of remission [30]. Another haplotype was found to be significantly associated with remission status: GCGCACACGAGAC in block 2 for fluoxetine‐treated patients [32].

3.2.6. Fluvoxamine

The effects of fluvoxamine on various SNPs were assessed in two studies. Both studies found no statistically significant association with treatment response for rs1045642 [10, 34]. One of the studies assessed an additional five SNPs (rs2032582, rs2235015, rs2235040, rs4148739 and rs9282564), but no association with treatment response was found [10].

3.2.7. Mirtazapine

Nine studies consistently found that ABCB1 polymorphisms did not influence response to mirtazapine, a non‐substrate antidepressant [7, 10, 11, 14, 21, 33, 35, 36, 37].

3.2.8. Paroxetine

rs1045642: No statistically significant association with remission was reported in 10 studies [7, 10, 11, 19, 35, 38, 39, 40, 41, 42].

rs2032582: No statistically significant association with paroxetine response was observed in eight studies [7, 10, 19, 35, 38, 40, 41, 42]. Two studies reported a positive association with response [5, 39].

rs2032583: One study demonstrated that the implementation of ABCB1 genotyping as a diagnostic tool influenced clinical decisions and improved treatment outcomes when higher doses of P‐gp substrate antidepressants were administered [14]. One found a positive association [7], whereas three studies reported no significant link between this SNP and treatment outcomes [15, 19, 35].

rs2235015: One study demonstrated that the implementation of ABCB1 genotyping as a diagnostic tool influenced clinical decisions and improved treatment outcomes when higher doses of P‐gp substrate antidepressants were administered [14]. One found a positive association [7], whereas four studies reported no significant link between this SNP and treatment outcomes [10, 15, 19, 35].

rs1128503: No statistically significant association with therapeutic response was observed in four studies [7, 19, 38, 39].

rs2235040: One study found a statistically significant association with treatment response [7], while three others did not identify any association [10, 19, 35, 38].

rs4148739: Two studies did not identify any statistically significant association [10, 16]. One study found an association with treatment response [7].

rs9282564: One study found a positive association [40], whereas three did not find an association [10, 35, 42].

Other SNPs: No association was found for rs2235047, rs2229107, rs6961419, rs1922241, rs1202167, rs3789243 [33], rs6946119, rs28401781, rs3747802 [16], rs10245483, rs3213619, rs2214102, rs9282564, rs10276036, rs2229109, rs2032588, rs2235033, rs28381916 and rs2235063 [35]. An association with treatment response was found for rs2235067, rs4148740, rs10280101, rs7787082, rs11983225, rs10248420 and rs12720067 [7].

One study reported that the wild variant haplotype 3435C–2677G–1236T based on rs1045642 (C3435T), rs2032582 (G2677T) and rs1128503T (1236T) was associated with poor response [39], while another study found no association [38].

No association was found for the following haplotypes: H1 (G–C), H2 (G–T), H3 (T–C) and H4 (T–T), based on rs2032582 (G2677T) and rs1045642 (C3435T) [41].

3.2.9. Sertraline

rs1045642: No statistically significant association was found in three studies [10, 19, 23].

rs2032582: Three studies reported no significant association with treatment response [10, 19, 23].

rs2032583: One study demonstrated that the implementation of ABCB1 genotyping as a diagnostic tool influenced clinical decisions and improved treatment outcomes when higher doses of P‐gp substrate antidepressants were administered [14]. However, four studies reported no association [15, 19, 23, 26].

rs2235015: One study reported that the implementation of ABCB1 genotyping as a diagnostic tool influenced clinical decisions and improved treatment outcomes when higher doses of P‐gp substrate antidepressants were administered [14]. However, five studies showed no association with treatment response [10, 15, 19, 23, 26].

rs1128503: One study did not find an association [19].

rs2235040: Two studies found no statistically significant association with treatment response [10, 19], while one study reported a significant positive association [23].

rs4148739: Two studies did not identify any significant association [10, 16].

rs9282564: No statistically significant association with treatment response was reported in one study [10], whereas another identified a significant positive association [23].

Other polymorphisms: No statistically significant association with treatment response was reported for rs6946119, rs28401781, rs3747802 [16], rs3213619, rs2214102, rs10276036, rs2032588, rs2235033, rs28381916 and rs7793196 [26]. A positive association was found for rs10245483 [26].

3.2.10. Selective Serotonin Reuptake Inhibitors (SSRIs)

Three studies only stated that they investigated ‘SSRIs’; these are therefore described in the section below.

rs1045642: One study focusing on SSRIs observed a positive correlation with treatment response [43], while another focusing on SSRIs observed a negative correlation [44].

rs2032582: One study focusing on SSRIs observed a positive correlation with treatment response [45], while another focusing on SSRIs observed a negative correlation [44].

rs1128503: One study focusing on SSRIs observed a positive correlation with treatment response [45], while another focusing on SSRIs observed a negative correlation [44].

Other SNPs: T‐129C was not associated with response [45].

3.2.11. Tricyclic Antidepressants

rs1045642: No statistically significant association with treatment response was reported for clomipramine, amitriptyline, dosulepin or imipramine in one study [11] as well as no association for nortriptyline in another study [31] and no association for clomipramine in a third study [42]. No association with treatment response was found for amitriptyline and clomipramine in one study [10] and none was found for amitriptyline in two other studies [7, 37].

rs2032582: No association with treatment response was found for amitriptyline and clomipramine in one study [10]. No significant association with amitriptyline treatment response was reported in three studies [7, 37, 46], and no association with clomipramine was reported in another study [42].

rs2032583: One study reported for amitriptyline and nortriptyline that the implementation of ABCB1 genotyping as a diagnostic tool influenced clinical decisions and improved treatment outcomes when higher doses were administered [14], but another did not find an association [15]. A positive association for amitriptyline was found in a third study [7]. No association was found for desipramine [32].

rs2235015: One study reported for amitriptyline and nortriptyline that the implementation of ABCB1 genotyping as a diagnostic tool influenced clinical decisions and improved treatment outcomes when higher doses were administered [14], but another did not find an association [15]. An association with treatment response was not found for amitriptyline and clomipramine in one study [10] while a positive association for amitriptyline was found in another study [7]. No association was found for desipramine [32].

rs1128503: No association was found for amitriptyline in two studies [7, 37]. No association was found for desipramine [32].

rs2235040: An association with treatment response was not found for amitriptyline and clomipramine in one study [10]. One study found an association with treatment response for amitriptyline [7]. No association was found for desipramine [32].

rs4148739: An association with treatment response was not found for amitriptyline and clomipramine in one study [10]. One study found an association with treatment response for amitriptyline [7].

rs9282564: No statistically significant association with treatment response for amitriptyline and clomipramine was observed in one study [10], and no association for clomipramine was found in another study [42].

Other polymorphisms: rs3842, rs17064, rs10276036, rs2235020 and 73 other SNPs were not associated with response for desipramine, but the haplotype ACA was associated with a positive response [32].

3.2.12. Venlafaxine

rs1045642: Two studies reported a statistically significant positive association with treatment response [25, 43], while one reported a negative association [44]. Nine studies found no statistically significant association [7, 10, 11, 19, 23, 28, 37, 42, 47].

rs2032582: Two studies found a negative association with treatment response [28, 44]; whereas another study reported an association in the opposite direction [45]. No statistically significant association was reported in eight studies [7, 10, 19, 23, 25, 37, 42, 47].

rs2032583: One study demonstrated that the implementation of ABCB1 genotyping as a diagnostic tool influenced clinical decisions and improved treatment outcomes when higher doses were administered [14]. Two studies reported a positive association [7, 19]. However, three studies found no statistically significant association [15, 23, 26].

rs2235015: One study demonstrated that the implementation of ABCB1 genotyping as a diagnostic tool influenced clinical decisions and improved treatment outcomes when higher doses were administered [14]. One study reported a positive association [7]. However, five studies found no statistically significant association [10, 15, 19, 23, 26].

rs1128503: No statistically significant association with treatment response was observed in five studies [7, 19, 25, 28, 37], whereas one study identified a positive statistically significant association [45] and one identified a negative association [44].

rs2235040: An association was found in one study [7], and a positive association was observed in another study [23]. Two other studies found no association [10, 19].

rs4148739: Two studies found no association with treatment response [10, 21], while one study found an association [7].

rs9282564: One study reported a statistically significant positive association with treatment response [23]; while two studies found no association [10, 42].

Other polymorphisms: No statistically significant association with treatment response was found for rs2235047, rs2229107, rs6961419, rs1922241, rs1202167, rs3789243 [33], T‐129C [44], rs6946119, rs28401781, rs3747802 [21], rs3213619, rs2214102, rs10276036, rs2032588, rs2235033, rs28381916, rs7793196 [26].

rs10245483 [26] was associated with a positive treatment response, and rs2235067, rs4148740, rs10280101, rs7787082, rs11983225, rs10248420 and rs12720067 [7] were associated with treatment response.

The haplotype rs28401781‐rs4148739 was not associated with response [21].

4. Discussion

The studies present highly mixed results regarding the impact of ABCB1 polymorphisms on the efficacy of antidepressants. For many of the investigated SNPs, some studies identify an association; while others do not.

4.1. Study Design

For many SNPs, contradictory results were observed across multiple studies, with no consistent correlation to the specific drug used. In three studies, there may be issues with the methods used to assess depression and remission [19, 23, 26]. For instance, Schatzberg et al. [26] employ a different scale (QIDS‐SR) compared to most other studies, which use HDRS [26]. This discrepancy could lead to differences in diagnosis and in assessing treatment effects/remission. The use of varying scales for such critical parameters may contribute to the divergence in study conclusions. Depression is generally more challenging to measure compared to many other diseases, as its diagnosis largely relies on patient self‐reporting rather than objective diagnostic criteria, making studies prone to bias. Comparing results from different methods for measuring primary endpoints introduces additional uncertainty regarding the influence of polymorphisms on antidepressant efficacy.

The findings across studies do not indicate any clear association between ABCB1 polymorphisms and specific types of antidepressants. For example, SSRIs did not seem to produce more consistent results compared to serotonin–norepinephrine reuptake inhibitors (SNRIs). This lack of differentiation may stem from the fact that both classes often rely on overlapping pharmacokinetic pathways and similar mechanisms of action, potentially limiting the influence of transporter‐related genetic variations.

Positive findings may result from small sample sizes, a limitation that Shan et al. acknowledge [19]. Although this study has a sample size of 292 patients, larger than most of the other studies, many have fewer than 100 patients. Out of the 42 studies included, only five were RCTs. However, all but one of these RCTs failed to identify associations between ABCB1 polymorphisms and antidepressant efficacy. This raises important questions about whether observed associations in non‐RCT studies might be influenced by methodological biases, such as smaller sample sizes, shorter observation periods or the lack of a control group.

Another limitation is the relatively short observation period of the studies. Geers et al. observed patients for only 4 weeks, while most other studies limited their observations to 6 weeks [10]. While the effects of antidepressants can typically be seen within 1–2 weeks, their full impact is most pronounced after 6–8 weeks or even later [48]. Some studies assessed remission after 8 weeks, but many applied the shorter periods mentioned above. These durations seem insufficient for evaluating the association between SNPs in the ABCB1 gene and clinical response. Studies with shorter observation periods may have missed associations that could have emerged over longer periods.

As shown in Table 2, another important study design aspect in evaluating SNPs as pharmacogenetic biomarkers is whether antidepressant plasma concentrations were measured, since the concentration–phenotype relationship is central to understanding their clinical relevance. Notably, only 19 of the 42 studies assessed both measures concurrently, despite this often representing a gold standard [49]. It is, however, important to note that plasma concentrations primarily reflect systemic drug exposure, while the ABCB1 transporter at the blood–brain barrier may affect antidepressant levels in the central nervous system independently of plasma values. Therefore, studies without plasma concentration data do not necessarily limit the interpretation of clinical outcomes, but the inclusion of both measures can provide a better understanding of how genetic variation might affect treatment response. This review, however, focused specifically on clinical treatment outcomes and therefore did not analyse the data on antidepressant plasma concentrations in detail.

4.2. Lack of Association for Specific Polymorphisms

The observation that many polymorphisms were not associated with antidepressant remission raises questions about their biological relevance. The lack of effect could indicate that these SNPs do not affect ABCB1 function. This aligns with the broader hypothesis that individual SNPs may have limited utility in predicting antidepressant efficacy unless they are part of larger haplotype structures that interact with other genomic regions. Future research should prioritise exploring these haplotype interactions across diverse populations to uncover potential genetic predictors of treatment response.

4.3. Non‐Substrate Antidepressants and ABCB1 Polymorphisms

Findings consistently showed that ABCB1 polymorphisms did not influence the response to non‐substrate antidepressants, such as mirtazapine. This reinforces the theoretical expectation that non‐substrate drugs are not impacted by transporter‐related genetic variation. These results are clinically relevant, as they suggest that mirtazapine could be a viable alternative for patients with genetic profiles unfavourable for ABCB1‐substrate antidepressants.

4.4. Haplotype‐Based Predictions

Another possible explanation for the inconsistent results is that individual SNPs may not directly influence remission. Instead, SNPs may be linked to other gene regions that affect ABCB1 function, thus exerting an indirect impact. Different demographic groups may carry the same SNP but experience different effects on remission, depending on the associated haplotype. For instance, a specific SNP might correlate with higher remission rates in one population due to its linkage with a haplotype that enhances antidepressant efficacy, whereas the same SNP in another group might be linked to a different haplotype with lower efficacy.

For rs1045642, the most extensively studied SNP in the ABCB1 gene, a statistically significant association with antidepressant treatment response was reported in 6 out of 30 included studies. The C allele was most frequently associated with an improved treatment response. Furthermore, 4 of the 30 studies examined rs1045642 as part of a haplotype analysis, with varying results. The most investigated haplotype, involving rs1045642‐rs2032582‐rs1128503, was analysed in three studies. In one study, the CGT haplotype showed no association with treatment response, while another study reported a negative association. Conversely, the TTT haplotype was positively associated with treatment response in the same study. The role of haplotypes in antidepressant efficacy is further supported by population‐based differences. If individual SNPs were solely responsible for antidepressant efficacy, results would be consistent across studies. Instead, genetic linkages likely contribute, as seen in Huang et al. [16], where rs4148739 was linked to remission in a Chinese Han population but not in Geers et al. [10], which studied a Russian population, despite both focusing on SSRIs. These findings highlight the importance of considering haplotype structures rather than individual SNPs to better understand the complex genetic mechanisms underlying antidepressant efficacy. Thus, it remains questionable whether individual SNPs can reliably predict antidepressant efficacy. Future studies should focus on haplotypes rather than isolated SNPs. These interpretations, however, require further investigation, as haplotypes were assessed in only a limited number of studies.

4.5. Synonymous and Non‐Synonymous SNPs

Numerous studies have explored synonymous and non‐synonymous SNPs to understand their potential role in antidepressant response. Non‐synonymous SNPs, like rs2032582 (G2677T), theoretically alter protein structure and function by causing amino acid changes. While rs2032582 was often found not associated with remission, this absence of association suggests that the amino acid change does not significantly impact ABCB1 function in this context. Similarly, rs1045642, a synonymous SNP, has yielded mixed results, likely due to its lack of direct influence on protein function. However, rs1045642 is frequently part of larger haplotypes linked to remission, highlighting the importance of considering SNP interactions within genomic regions. These findings underscore the complexity of interpreting the impact of individual SNPs and the potential need to shift focus toward haplotype‐level analyses.

4.6. Implications for Clinical Practice and Methodological Considerations

The findings highlight the complexity of using genetic markers for guiding antidepressant treatment. While some SNPs show promise as predictors of treatment response, the inconsistencies across studies underscore the limitations of relying solely on genetic data. Integration of genetic markers with clinical, demographic, and environmental factors may enhance the precision of personalised treatment strategies. Furthermore, the consistent lack of influence of ABCB1 polymorphisms on non‐substrates like mirtazapine emphasises the importance of considering pharmacological properties when tailoring treatments. Aligning study parameters such as depression scales, sample sizes and observation periods is critical for achieving reliable findings. The diverse populations examined highlight the need for more inclusive research that considers population‐specific genetic backgrounds and environmental influences. Such approaches may facilitate more consistent and actionable outcomes in pharmacogenomics.

5. Conclusion

This scoping review highlights the uncertainty regarding the impact of ABCB1 polymorphisms on antidepressant efficacy. Of the 42 studies included in this review, 30 investigated rs1045642, making it the most extensively studied SNP in the ABCB1 gene. Among these, only 20% reported a statistically significant association with antidepressant treatment response. SNPs such as rs2032582 and rs1128503 were frequently studied, but statistically significant associations were reported in only 28% and 13% of the studies, respectively, often with conflicting directions. Haplotype data also showed mixed results. Studies varied substantially in terms of design, methodology, population and observation period—and biases in non‐RCTs may have been substantial.

Taken together, this scoping review highlights the need for standardised research methodologies to clarify the role of ABCB1 polymorphisms in antidepressant efficacy. Future research should prioritise larger, standardised studies with consistent methodologies and diverse populations to better clarify the role of ABCB1 polymorphisms in antidepressant efficacy. Incorporating haplotype analyses alongside individual SNP investigations could provide deeper insights into the genetic mechanisms underlying treatment response, ultimately improving personalised pharmacogenomics.

Conflicts of Interest

The authors declare no conflicts of interest.

Thorsen S., Bülow C., Dalhoff K., and Sonne D., “The Effect of ABCB1 Polymorphisms on the Efficacy of Antidepressants,” Basic & Clinical Pharmacology & Toxicology 137, no. 4 (2025): e70097, 10.1111/bcpt.70097.

Data Availability Statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.