Antidepressant switching as a proxy phenotype for drug non-response: investigating clinical, demographic and genetic characteristics
====================================================================================================================================

* Chris Wai Hang Lo
* Alexandra C. Gillett
* Matthew H. Iveson
* Michelle Kamp
* Chiara Fabbri
* Win Lee Edwin Wong
* Dale Handley
* Oliver Pain
* Evangelos Vassos
* Naomi R. Wray
* Heather C. Whalley
* Danyang Li
* Allan H. Young
* Andrew M. Mcintosh
* AMBER Research Team
* Cathryn M. Lewis

## Abstract

**Background** Selective serotonin reuptake inhibitors (SSRIs) are a first-line pharmacological therapy in major depressive disorder (MDD), but treatment response rates are low. Clinical trials lack the power to study the genetic contribution to SSRI response. Real-world evidence from electronic health records provides larger sample sizes, but novel response definitions are needed to accurately define SSRI non-responders.

**Methods** In UK Biobank (UKB) and Generation Scotland, SSRI switching was defined using a ≤ 90-day gap between prescriptions for an SSRI and another antidepressant in primary care. Non-switchers were participants with ≥ 3 consecutive prescriptions for an SSRI. In UKB, clinical, demographic and polygenic score (PGS) associations with switching were determined, and the common-variant heritability was estimated.

**Results** In UKB, 5,133 (13.2 %) SSRI switchers and 33,680 non-switchers were defined. The mean time to switch was 28 days. Switching patterns were consistent across UKB and Generation Scotland (n = 498 switchers). Higher annual income and educational levels (OR [95% CI] for university degree compared to no qualifications: 0.727 [0.666-0.794]) were associated with lower levels of switching. PGS for non-remission, based on clinical studies, were associated with increased risk of switching (OR: 1.07 [1.02-1.12], p = 0.007). MDD PGS and family history of depression were not significantly associated with switching. The heritability (*h2*) of SSRI switching was approximately 4% on the observed scale.

**Conclusion** This study identified SSRI switching as a proxy of drug non-response, scalable across biobanks, capturing demographic and genetics of treatment non-response, and independent of the genetics of MDD.

Keywords
*   Antidepressant response
*   electronic health records
*   selective serotonin reuptake inhibitors (SSRIs)
*   GWAS
*   UK Biobank
*   Generation Scotland

## Introduction

Worldwide, approximately 300 million people suffer an episode of major depressive disorder (MDD) during their lifetime (1). Selective serotonin reuptake inhibitors (SSRIs) are a first-line pharmacological treatment in MDD based on safety profiles (2,3). However, considerable variability exists in antidepressant response (4), with only about one-third of antidepressant users achieving clinical remission with their first prescribed antidepressant (5). Another third of patients go on to develop treatment-resistant depression (TRD), defined as the lack of response to two antidepressants with adequate duration and dosage (6). Identification of factors that predict response and non-response to antidepressants would enable personalised prescribing and improve treatment for MDD.

Multiple factors have been associated with response and non-response to antidepressant treatment. For example, childhood trauma is associated with poorer response to antidepressants (7) and, in clinical trials, higher body mass index (BMI) and neuroticism scores were significantly associated with antidepressant response (8,9). Patients with TRD often have higher depression symptom severity (10,11), and observational studies suggest that TRD is correlated with sociodemographic characteristics, such as unemployment (11). However, baseline MDD severity was not associated with symptom-level response in a meta-analysis of 91 clinical trials (12). Biomarkers for antidepressant response, including brain-derived neurotrophic factor (13), cortisol (7) and inflammatory markers (14), show inconsistent results.

Genetic factors have been associated with antidepressant response. Cytochrome P450 variants play a minor role in adverse events and response by affecting metabolism of antidepressants (15–18), while the largest genome-wide association study (GWAS) of clinical studies to date estimated the single-nucleotide polymorphism (SNP)-based heritability for remission to be as high as 40% (19). However, genetic studies performed with clinical trials are under-powered to discover SNPs at genome-wide significance. Stringent inclusion criteria may also limit the generalisability of genetic findings to a population-wide level. Other study designs are therefore required to increase the power to identify the genetic component of antidepressant response. Real-world data from electronic health records (EHR) has the potential to fill this gap, since large sample sizes are available in biobanks, with pre-existing genetic data often available.

In EHR, defining treatment response phenotypes is challenging, as response (or resistance) to antidepressant treatment is not directly coded in most records. Using clinical records, proxy phenotypes can be captured from unstructured text using natural language processing algorithms (20). Alternatively, phenotypes can be defined from structured prescription records, which are more readily available and scalable in population-wide biobanks. One feasible strategy is to capture switching events between antidepressants as an indication of non-response (21). This approach reflects clinical guidelines, where patients who fail to respond to an antidepressant are recommended to switch to a different drug. Antidepressant switching in EHR has been used to define TRD, where two switches occur within a single episode of depression (22). Antidepressant switching is also used in clinical trial design as an alternative therapeutic strategy following inadequate response to the first antidepressant treatment (often an SSRI) (23,24).

In this study, we use primary care prescribing records in UK Biobank and Generation Scotland (25) to define a phenotype of switching from an SSRI to another antidepressant (of any class) within an episode of depression. We characterize the prescription patterns, and investigate the clinical, demographic, and polygenic predictors of SSRI switching. We further perform GWAS of SSRI switching, showing that switching is heritable, and we propose switching as a proxy measure of non-response to SSRIs.

## Methods and Materials

### Primary Sample – UK Biobank (UKB)

UKB is a prospective health study that recruited over 500,000 volunteers aged 40 to 69 years in the United Kingdom from 2007 to 2010 (26,27). All participants completed a baseline assessment on demographics, lifestyle, and mental health upon enrolment, and biological samples were collected (26,27). Genome-wide genotyping, available for all UKB participants, underwent standard quality control (QC) and imputation. Further description of UKB samples, as well as details on genotyping QC and imputation, is available in **Supplementary Methods,** and an analysis flowchart is shown in **Figure 1**.

![Figure 1.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/11/10/2024.11.09.24316987/F1.medium.gif)

[Figure 1.](http://medrxiv.org/content/early/2024/11/10/2024.11.09.24316987/F1)

Figure 1. Study design overview for primary sample (UK Biobank)
**Legends**

Icons used created by Team Iconify, DailyPm Studio and Freepik from Flaticon.

**Abbreviations**

GCTA = Genome-wide Complex Trait Analysis; GCTB = Genome-wide Complex Trait Bayesian; GWAS = genome-wide association study; QC = quality control; SNP = single nucleotide polymorphism; SSRI = selective serotonin reuptake inhibitors; UKB = UK Biobank.

Linkage with primary care data is available for ∼230,000 UKB participants, containing clinical events (coded by READ v2 or CTV-3), and prescription records (coded by READ v2, READ v3, BNF or dm+d) (28). Prescriptions are mostly available from the 1990s to 2018, with the start date dependent on the databases linked to specific regions of practice. Prescription codes for SSRIs were extracted from all available records. Patients with depression were identified using diagnosis records for depressive disorders from primary care practitioners, using a previously validated algorithm (22). Participants with any primary care diagnosis for bipolar disorder, psychosis or substance abuse were excluded. Not all patients prescribed SSRIs had a depression diagnostic code assigned. To examine the impact of diagnosis, we analysed three datasets of participants prescribed an SSRI: (1) all patients; (2) patients with ≥ 1 depression diagnostic record; and (3) participants with ≥ 2 depression diagnostic records. The READ v3 diagnosis codes used are listed in **Supplementary table 1**, with antidepressants and mapped drug classes listed in **Supplementary table 2**.

Sociodemographic and clinical variables extracted from UKB were self-reported sex (field ID 31-0.0), educational qualifications (field ID 6138-0.0, with “none of the above” as reference level), annual income (field ID 738-0.0, with “less than £18,000” as reference level), Townsend Deprivation Index (field ID 189-0.0), BMI (field 21001-0.0) and family history of depression. A positive family history of depression was defined where at least one parent (field ID 20107-0.0 and 20110-0.0) or sibling (field ID 20111-0.0) was reported as having severe depression. Polygenic scores (PGSs) for MDD (UKB participants removed) (29), schizophrenia (30) and antidepressant non-remission (19) were computed using PRS-CS (31) with the GenoPred analysis pipeline (version 1) (32). PGSs were reference-standardized against 1000 Genomes Project European Populations (33), with full details of polygenic scoring methods and GWAS summary statistics in **Supplementary Methods** and **Supplementary table 3.**

### Replication sample – Generation Scotland

Replication analyses were performed in Generation Scotland to validate switching patterns across healthcare practices. Generation Scotland is a family-based longitudinal study which recruited over 24,000 volunteers from 2006 to 2011, with information on demographics, physical and mental health measurements (25). Linkage to prescription records in primary care was available for over 90% of participants (25). Details on Generation Scotland replication sample are summarised in **Supplementary Methods**.

### Phenotype definition – SSRI switching

SSRI switchers were defined as UKB participants being prescribed an SSRI who then received a prescription for another antidepressant within a 90-day window, from 5-95 days of the initial prescription (a *switching* event). The following additional criteria were applied: 

1.  a minimum of 5-day window between prescriptions to avoid capturing overlapping prescriptions of two antidepressants (augmentation) as switches;

2.  the pre-switch SSRI was prescribed ≤ 3 times in total to ensure transient exposure;

3.  the pre-switch SSRI was prescribed ≤ 2 times before the switch date, to capture early switchers specifically; and

4.  the pre-switch SSRI was prescribed ≤ 2 times after the switch date, to ensure augmentation is not captured, while giving a brief allowance period for cross-tapering.

SSRI non-switchers were defined as patients who did not switch from any SSRIs, and received ≥ 3 prescriptions for an SSRI.

The SSRI index date was defined as the first prescription date for the SSRI in switchers and non-switchers. A schematic figure of definitions of switchers and non-switchers is shown in **Supplementary figure 1**.

### Analysis of switching patterns, clinical and demographic variables, polygenic scores

Descriptive analyses were performed on SSRI switchers and non-switchers, by index SSRI drug, drug class post-switch, time to switch and age at index date. Distributions of BMI, and demographic variables across assessment centres in UKB were also assessed. Differences in the distribution of the variables between SSRI switchers and non-switchers were assessed by nonparametric statistical tests, including Pearson’s chi-square test (for binary variables), Kruskal-Wallis rank-sum test (for categorical variables of more than two levels), and Wilcoxon rank-sum test (for continuous variables).

Associations between switching and sociodemographic variables at baseline assessment and PGS were tested by logistic regression, with all models adjusted for self-reported sex, index date of SSRI, and assessment centre. Associations with PGS were further adjusted for 10 principal components for population stratification. For related individuals, one participant was removed based on third-degree relatedness (kinship coefficient < 0.044) by greedy matching, with cases being preferentially retained. Statistical significance was assessed by likelihood ratio test and corrected for multiple testing by Bonferroni correction within each sample (p≤ 0.0071, correcting for 7 variables).

### Genome-wide association study

Genome-wide association analysis was performed on SSRI switching using REGENIE (34), a two-step software for genome-wide analysis. Genotyped SNPs underwent linkage disequilibrium pruning (LD), using a *R2* threshold of 0.9 within a window size of 1000 markers and a block size of 100 markers. Given the low ratio of switchers / non-switchers, SNPs with low allele frequencies might go into quasi-complete separation when applying standard logistic regression models in GWAS (34). Therefore, SNP effect sizes underwent Firth correction to control for false positives as recommended in REGENIE documentation (34).

Two analyses were performed, for participants with at least one or two MDD primary care diagnostic records. For genetic analyses, we tested for differences in the distributions of assessment centre and genotyping batch between switchers and non-switchers by Kruskal-Wallis rank sum test, to assess which covariates to be included (see **Supplementary table 9**). Covariates included in the genome-wide association analyses were SSRI index date, genetic sex, assessment centre and six principal components for population stratification.

### SNP-based heritability (*h2*) estimation

SNP-based heritability was estimated using two genomic relatedness–based restricted maximum likelihood (GREML)-based methods, Genome-wide Complex Trait Analysis (GCTA) (35) and Genome-wide Complex Trait Bayesian (GCTB) (36). SNP-based heritability (*h2*) was reported on the observed scale since the sample was unselected for SSRI treatment. In GCTB, the *h2* estimates were constrained to be between 0 and 1 in each iteration, and the distributions of *h2* estimates can be skewed when the true *h2* is close to 0. To strengthen the robustness of our findings, the posterior mode and 95% highest posterior density (HPD) credible intervals were also reported for GCTB. The degree of polygenicity (*Pi*) and negative selection (*S*) were also reported. Full details of *h2*estimation are available in **Supplementary methods**.

## Results

### Descriptive analyses

In UKB, a total of 5,133 SSRI switchers and 33,680 non-switchers were identified from prescription records (full sample). Baseline characteristics of SSRI switchers and non-switchers are given in **Table 1 and Supplementary figure 2**. Patients in the full sample had a median of 18 SSRI prescriptions (IQR: 7-47; **Figure 2, Supplementary table 4**) across all prescribing history, primarily spanning the 1990s to 2018 (**Supplementary figure 3**). Of these participants, 3,088 (60%) switchers and 19,618 (58%) non-switchers had at least one diagnostic record for MDD, with 2,114 (41%) switchers and 12,667 (38%) non-switchers having at least two MDD diagnostic records (**Supplementary figure 2)**. In the full sample, 67% were female, and 96% were of white ethnicity (**Table 1**). Approximately half of participants had at least one prescription for a tricyclic antidepressant (TCA) (full sample: *N* = 18,125 [46.7%]; one depression record: *N* = 11,320 [49.9%]; two depression records: *N* = 7,749 [52.4%]; **Supplementary table 5**), and 13-18% received ≥ 1 prescription for a serotonin-norepinephrine reuptake inhibitor (SNRI) (**Supplementary table 5**). In Generation Scotland, a total of 498 SSRI switchers and 1,279 non-switchers with at least one diagnostic record for depression were identified (**Supplementary table 6**).

![Figure 2.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/11/10/2024.11.09.24316987/F2.medium.gif)

[Figure 2.](http://medrxiv.org/content/early/2024/11/10/2024.11.09.24316987/F2)

Figure 2. Number of antidepressant prescriptions in primary sample (UK Biobank)
**Caption**

(A) Number (Median [IQR]) of prescriptions of primary sample (by drug classes with number of patients > 500); (B) Proportion of patients receiving at least 1 prescription for a particular drug class. Only drug classes consisting of > 1% of sample sizes were labelled. Details on the statistics for both figures were available in supplementary materials.

**Abbreviations**

MAOI = monoamine oxidase inhibitors; NDRI = norepinephrine-dopamine reuptake inhibitors; NRI = norepinephrine reuptake inhibitors; SARI = serotonin antagonist and reuptake inhibitors; SNRI = serotonin–norepinephrine reuptake inhibitors; SSRI = selective serotonin reuptake inhibitors; TCA = tricyclic antidepressants.

View this table:
[Table 1.](http://medrxiv.org/content/early/2024/11/10/2024.11.09.24316987/T1)

Table 1. Primary sample (UK Biobank) summary

### Patterns of SSRI switching

In both UKB and Generation Scotland, the most prescribed SSRI antidepressants were fluoxetine and citalopram. Around one-third received fluoxetine (UKB [**Table 1**]: N = 1,687 [33%] switchers and 11,278 [33%] non-switchers; Generation Scotland [**Supplementary table 6**]: N = 165 [33%] switchers and 433 [34%] non-switchers). Another one-third received citalopram (UKB: N = 1,586 [31%] switchers and 12,448 [37%] non-switchers; Generation Scotland: N = 190 [38%] switchers and 604 [47%] non-switchers). Paroxetine and sertraline were also common prescribed, to over 10% of participants each.

Most SSRI switches occurred within six weeks of the index prescription in both UKB (median time to switch in days [IQR]: 28 [17-49], **Table 1**) and Generation Scotland (31 [31-61], **Supplementary table 6**). Distributions of time to switch were similar across index SSRIs in UKB (**Supplementary figure 4**). The proportion of switchers and time to switch were generally comparable across gender (**Supplementary figure 6**). In UKB, approximately half of switching events were to another SSRI (N = 2,380; 46%), or to a TCA (N = 1,597; 31%) (**Supplementary table 7, Figure 3**). Similar patterns were observed in Generation Scotland, with half switching to another SSRI (N = 237; 47.6%), with fewer patients changing to TCA after switching (N = 73; 14.7%) (**Supplementary figure 7, Supplementary table 8**).

![Figure 3.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/11/10/2024.11.09.24316987/F3.medium.gif)

[Figure 3.](http://medrxiv.org/content/early/2024/11/10/2024.11.09.24316987/F3)

Figure 3. Switching patterns and association testing with clinical, sociodemographic variables and polygenic scores in primary sample (UK Biobank)
**Caption**

(A) SSRI switching patterns in UK Biobank, stratified by index SSRI and drug class after switch; (B) Association testing between demographic variables and SSRI switcher statuses; (C) Association testing between polygenic scores and SSRI switcher status. Only SSRIs (pre-switch) and drug classes (post-switch) with at least a sample size of 70 are labelled.

**Abbreviations**

MDD = major depressive disorder; PGS = polygenic scores; SARI = serotonin antagonist and reuptake inhibitors; SNRI = serotonin– norepinephrine reuptake inhibitors; SSRI = selective serotonin reuptake inhibitors; TCA = tricyclic antidepressants.

Across the 22 assessment centres in UKB, the proportion of switchers differed significantly, varying between 9-19% (p = 2.5e-08; **Supplementary table 9, Supplementary figure 8).** The rate of switching decreased with more recent index dates (Pearson correlation coefficient between switching rate and time [*r*] = 0.43, p = 1.0e-05), and the period between index date and switching was shorter (*r* = -0.42, p = 1.5e-05) (**Supplementary figure 9**).

### Associations with sociodemographic variables and polygenic scores

In UKB, higher educational levels were associated with lower odds of SSRI switching (odds ratio for university degree [OR, 95% CI]: 0.727 [0.666-0.794], p = 1.53e-10, compared to the reference group of no qualifications) (**Supplementary table 10**). This effect was consistent when samples were limited to at least two MDD diagnoses in primary care (0.789 [0.686-0.908], p = 0.013) and one MDD diagnosis (0.778 [0.694-0.873], p = 0.001). Similar findings were observed in annual income, with higher income associated with lower risks of SSRI switching for annual income > £100,000 compared to the reference group of <£18,000 in the full sample (odds ratio [OR] 0.665 95% CI [0.529-0.827], p = 6.79e-15). Effect sizes were similar when constrained to at least two MDD diagnosis records (p = 1.43e-04) and one MDD diagnosis record (p = 4.92e-07). SSRI switching was not associated with Townsend Deprivation Index, a measurement of material deprivation.

Family history of severe depression was only nominally associated with SSRI switching status in the full sample (OR [95% CI]: 1.077 [1.001-1.158], p = 0.048), and was not associated with switching in patients with one (p = 0.084) or two (p = 0.226) MDD diagnosis records. The PGS for MDD was not associated with SSRI switching (p = 0.138, for full sample).

Higher PGS for antidepressant non-remission was associated with an increased risk of SSRI switching, with modest effect sizes. The associations were observed in all subgroups and were strongest in patients with two MDD diagnostic records (OR [95% CI]: 1.067 [1.018-1.119], p = 0.007). The associations were nominally significant in the full sample (1.034 [1.003-1.066], p = 0.029), and in participants with a single MDD diagnostic record (1.043 [1.004-1.05], p = 0.031), with similar directions of effect (**Figure 3, Supplementary table 10).**

### Genetic analyses

A GWAS was performed using REGENIE, on 2,868 SSRI switchers and 18,360 non-switchers for one MDD diagnosis record, reducing to 1,967 switchers and 11,853 non-switchers for two MDD diagnosis records. No variants were identified in either analysis at genome-wide significance (p < 5e-8), with Manhattan plots shown in **Supplementary figures 15-16)**. At a suggestive significance threshold (p<1e-5), 27 and 30 independent SNPs were identified with one or two MDD diagnoses respectively.

SSRI switching had a SNP-based heritability that was significantly different to zero on the observed scale in participants with one MDD diagnosis record (GCTA *h2* [SE]: 0.0451 [0.0269], p = 0.0409; GCTB (0.0399 [0.014], p = 0.0022) (**Figure 4, Supplementary table 11)**. The posterior mode for *h2* in GCTB was also different from zero (0.035; 95% HPD credible intervals [0.014-0.065]). Significant non-zero *h2* was also observed in GCTB for patients with at least two MDD diagnoses records (0.0383 [0.021], p = 0.034) but not in GCTA (p=0.27).

![Figure 4.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/11/10/2024.11.09.24316987/F4.medium.gif)

[Figure 4.](http://medrxiv.org/content/early/2024/11/10/2024.11.09.24316987/F4)

Figure 4. SNP-based heritability estimates for SSRI switching by GCTA and GCTB
**Caption**

SNP-based heritability (*SNP-h2*) on observed scale stratified by number of depression diagnoses in primary sample, expressed in **(A)** posterior mean (standard error) and **(B)** posterior mode (95% HPD credible intervals) for GCTB.

* p-value < 0.05.

** p-value < 0.01.

**Abbreviations**

GCTA = Genome-wide Complex Trait Analysis; GCTB = Genome-wide Complex Trait Bayesian; HPD = highest posterior probability; SE = standard error; SNP = single nucleotide polymorphism; SSRI = selective serotonin reuptake inhibitors.

## Discussion

EHRs offer promising opportunities to define antidepressant response outcomes from prescribing patterns, but the characteristics of these outcomes have not been well validated and characterised to date. In this study, we used SSRI switching as a proxy phenotype for non-response to antidepressants. The measure reflects the current clinical practice of moving patients from first-line SSRI treatment to a different antidepressant in cases of no-or poor response (2,37), the signs of which could be evident from two to four weeks at the earliest (38). We identify associations of switching with demographic and genetic profiles, and show a modest heritability for the switching phenotype.

Our phenotypic definition of switching in UKB aimed to capture antidepressant switchers, following previous work on over 260,000 participants prescribed antidepressants in the UK Clinical Practice Research Datalink (CPRD) database of primary care records (39). In the CPRD study, most SSRI users switched to another SSRI as second-line therapy (54.1%) (39), which is consistent with our findings. In CPRD, 9.3% of antidepressant users switched, with a median time to switch of 45 days (39). In comparison, SSRI users in UKB had a higher proportion of switching (13.2%), and a shorter time to switch (median: 28 days). The study period for the CPRD study was from 2005 (39), compared to the 1990s in UKB, the period when SSRIs became the first-line therapy in clinical guidelines (40).

In our definition of SSRI switching, we used a 90-day window between prescription dates of two different antidepressants to capture switching events. This contrasts with the shorter window applied in CPRD (39), where switching events were identified from a 30-day or less gap between the expected end dates of the first treatment and the start date of the second treatment. Using longer windows allows us to capture more switchers for genetic analyses, but is less specific to the exact cause of switching in the samples, such as lack of efficacy and side effects (37). Of note, using different window lengths did not substantially alter effect sizes for associations between switching and CYP2C19 metaboliser status in a previous UKB analysis (21). Our definition of switching events primarily relies on prescription dates of different antidepressants. This avoids making inferences of treatment duration where it cannot be accurately estimated from dosage instructions and quantity of prescriptions, as in UKB EHR data.

SSRI switching captures demographic and clinical variables associated with non-response to antidepressants. We showed that switching rates were lower in participants with higher incomes and higher educational levels, which aligns with evidence that higher PGS for education attainment were associated with remission in clinical trials (19). Higher levels of antidepressant response have been associated with higher socioeconomic status (SES) in a systematic review of clinical trials (41), and a Nordic registry study (42). SES was suggested previously to directly contribute to poor prognosis if it was causal to the development of depression itself (43). SES is also correlated with access to treatment and mental health services, which might in turn affect adherence to antidepressant and treatment outcomes (41). Our results confirm that SSRI switching in EHRs shows similar sociodemographic profiles as seen in antidepressant non-response in trials and retrospective clinical studies.

Results from the genetic analyses also support SSRI switching as a proxy phenotype for antidepressant non-response. SSRI switching in UKB was associated with PGS for non-remission, but not with MDD PGS or with family history of depression. These results indicate that the genetic basis of SSRI switching overlaps with the genetics of antidepressant response in clinical trials, but is independent of the genetics of MDD (19). MDD PGS captures the genetics of susceptibility and also of symptom severity (44,45) which is a strong predictor for antidepressant non-response. However, these genetic factors were not correlated with response outcomes in our analysis. Mixed evidence has been found in previous studies, with positive correlations between MDD PGS and poorer response in smaller clinical trials (15,46,47), but none survived multiple testing. Our genome-wide analysis of SSRI switching was severely underpowered to detect specific risk variants, but the validation of this EHR-based phenotype for antidepressant response/non-response opens opportunities for expanding the sample size in other real-world data sources. We sought replication of the UKB SSRI switching phenotype in Generation Scotland, which has 20,000 participants. However, with only 1,777 study members classified as SSRI switchers or non-switchers, limited analysis could be performed. We reported the results here for completeness.

Genetic analyses in UKB revealed a modest heritable signal for SSRI switching (SNP-based *h2* = ∼0.04). Antidepressant response was also found to be significantly heritable in a GWAS meta-analysis of clinical trials (19). Although genetic correlations cannot be calculated given the limited sample sizes, the positive association between non-remission PGS and SSRI switching confirms a common genetic component for these phenotypes. Our genetic heritability estimates differed between the GREML-based methods of GCTA and GCTB, and stratifying by one or two depression diagnosis records. The GCTA *h2* estimate with two MDD diagnosis records was not significantly different from zero, meaning we were unable to consistently confirm non-zero heritability across methods, likely due to heterogeneity in samples and insufficient statistical power. However, our results highlight the potential of using switching as a proxy phenotype to capture non-response from antidepressants, which is scalable across EHR data resources (25,48) and will allow future meta-analyses to obtain more robust estimates.

This study had several limitations from data availability in UKB primary care records and the sample size for genetic analyses. Genetic analyses were underpowered for GWAS, and at the margins of requirements for robust SNP-based heritability estimates. A further challenge is the moderate sample size for genetic studies in antidepressant clinical trials, from which non-remission PGSs were calculated. Antidepressants are prescribed for a wide range of conditions such as anxiety and insomnia (49), therefore diagnostic codes were necessary to ensure that the antidepressant was prescribed for depression. However, increasing specificity comes at the cost of reducing power, a widespread challenge in genetic studies. We chose to stratify SSRI switchers by those with at least one or two MDD diagnostic records in their primary care records, which reduced sample sizes to 58.5% and 38.1% of the full sample. For GWAS of mental disorders, broadening phenotypic definition increases power to detect associated loci, but reduces specificity (50,51). The impact on related phenotypes such as treatment response has yet to be assessed. Larger sample sizes for EHR-based SSRI switching are needed to replicate the current findings, and further refine appropriate definitions for proxy phenotypes of antidepressant response. While UKB has rich EHR data on prescribing, it lacks the response measures available in clinical trials, including depression symptom scores at baseline and during treatment. We are also unable to determine the exact causes of switching, which likely include side effects and non-compliance to antidepressant treatment, in addition to non-response

## Conclusion

Using primary care records in UKB, we successfully characterised SSRI switching as a data-driven proxy for SSRI non-response, validated using clinical, demographic and genetic variables. The SSRI switching phenotype was consistent with current clinical practices. SSRI switching appeared to capture the genetics of antidepressant response, and was distinct from the genetics of susceptibility to MDD. We also identified a modest, but significant, heritability for SSRI switching, although further replication is necessary. In summary, this study identified and validated SSRI switching as a valuable phenotype defined from electronic health records to capture the genetics of SSRI response. SSRI switching is highly scalable, and can be more broadly applied in genetic analyses, contributing to pharmacogenetic research aimed at personalised prescribing.

## Disclosures

CL sits on the Scientific Advisory Board for Myriad Neuroscience and has received consultancy fees from UCB. OP provides consultancy services for UCB. AHY is the editor of Journal of Psychopharmacology and the Deputy Editor of BJPsych Open. AHY received paid lectures and advisory boards for the following companies with drugs used in affective and related disorders: Flow Neuroscience, Novartis, Roche, Janssen, Takeda, Noema pharma, Compass, Astrazenaca, Boehringer Ingelheim, Eli Lilly, LivaNova, Lundbeck, Sunovion, Servier, Livanova, Janssen, Allegan, Bionomics, Sumitomo Dainippon Pharma, Sage, Neurocentrx. He is the Principal Investigator in the Restore-Life VNS registry study funded by LivaNova, the Principal Investigator on ESKETINTRD3004: “An Open-label, Long-term, Safety and Efficacy Study of Intranasal Esketamine in Treatment-resistant Depression.”, Principal Investigator on “The Effects of Psilocybin on Cognitive Function in Healthy Participants” and “The Safety and Efficacy of Psilocybin in Participants with Treatment-Resistant Depression (P-TRD)”studies, including:‘’A Double-Blind, Randomized, Parallel-Group Study with Quetiapine Extended Release as Comparator to Evaluate the Efficacy and Safety of Seltorexant 20 mg as Adjunctive Therapy to Antidepressants in Adult and Elderly Patients with Major Depressive Disorder with Insomnia Symptoms Who Have Responded Inadequately to Antidepressant Therapy.’’, ‘’ An Open-label, Long-term, Safety and Efficacy Study of Aticaprant as Adjunctive Therapy in Adult and Elderly Participants with Major Depressive Disorder (MDD).’’, ‘’A Randomized, Double-blind, Multicentre, Parallel-group, Placebo-controlled Study to Evaluate the Efficacy, Safety, and Tolerability of Aticaprant 10 mg as Adjunctive Therapy in Adult Participants with Major Depressive Disorder (MDD) with Moderate-to-severe Anhedonia and Inadequate Response to Current Antidepressant Therapy.’’, and ‘’ A Study of Disease Characteristics and Real-life Standard of Care Effectiveness in Patients with Major Depressive Disorder (MDD) With Anhedonia and Inadequate Response to Current Antidepressant Therapy Including an SSRI or SNR.’’ AHY is also the UK Chief Investigator for Compass COMP006 & COMP007 studies, and Novartis MDD study MIJ821A12201. He has received grant fundings (past and present) from the following organisations: NIMH (USA); CIHR (Canada); NARSAD (USA); Stanley Medical Research Institute (USA); MRC (UK); Wellcome Trust (UK); Royal College of Physicians (Edin); BMA (UK); UBC-VGH Foundation (Canada); WEDC (Canada); CCS Depression Research Fund (Canada); MSFHR (Canada); NIHR (UK). Janssen (UK) EU Horizon 2020. AHY has no shareholdings in pharmaceutical companies. All other authors have no competing interests to disclose.

## Supporting information

Supplementary Materials [[supplements/316987_file02.docx]](pending:yes)

## Data Availability

The UK Biobank data was accessed via project 82087. Please visit [https://www.ukbiobank.ac.uk/enable-your-research](https://www.ukbiobank.ac.uk/enable-your-research) for access. Generation Scotland data are available on reasonable request. Researchers may request access to Generation Scotland data through [https://www.ed.ac.uk/generation-scotland/for-researchers](https://www.ed.ac.uk/generation-scotland/for-researchers).

## Funding

This research was funded by Wellcome Mental Health Award (226770/Z/22/Z) and part-funded by the National Institute for Health and Care Research (NIHR) Maudsley Biomedical Research Centre (BRC). GS:SFHS was funded by a grant from the Chief Scientist Office of the Scottish Government Health Directorates (CZD/16/6) and the Scottish Funding Council (HR03006). Genotyping of the GS:SFHS samples was carried out by the Genetics Core Laboratory at the Edinburgh Clinical Research Facility, University of Edinburgh, Scotland, and was funded by the Medical Research Council UK and the Wellcome Trust (Wellcome Trust Strategic Award ‘STratifying Resilience and Depression Longitudinally’ Reference 104036/Z/14/Z). CF was partly supported by #NEXTGENERATIONEU (NGEU), funded by the Ministry of University and Research (MUR), National Recovery and Resilience Plan (NRRP), project MNESYS (PE0000006)—a multiscale integrated approach to the study of the nervous system in health and disease (DN. 1553 11.10.2022). OP is supported by a Sir Henry Wellcome Postdoctoral Fellowship [222811/Z/21/Z]. MHI is supported by the HDR UK DATAMIND hub, which is funded by the UK Research and Innovation grant MR/W014386/1, by the Wellcome Trust (220857/Z/20/Z; 226770/Z/22/Z, 104036/Z/14/Z; 216767/Z/19/Z) and by a Research Data Scotland Accelerator Award (RAS-24-2). AHY is funded by the National Institute for Health and Care Research (NIHR) Maudsley Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King’s College London. The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.

## Data sharing statement

The UK Biobank data was accessed via project 82087 — for access, go to [https://www.ukbiobank.ac.uk/enable-your-research](https://www.ukbiobank.ac.uk/enable-your-research). Generation Scotland data are available on reasonable request. Researchers may request access to Generation Scotland data through [https://www.ed.ac.uk/generation-scotland/for-researchers](https://www.ed.ac.uk/generation-scotland/for-researchers).

## Acknowledgements

For the purposes of open access, the author has applied a Creative Commons Attribution (CC BY) licence to any Accepted Author Manuscript version arising from this submission. We are grateful to all the families who took part in Generation Scotland, the general practitioners and the Scottish School of Primary Care for their help in recruiting them, and the whole Generation Scotland team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists, healthcare assistants and nurses.

*   Received November 9, 2024.
*   Revision received November 9, 2024.
*   Accepted November 10, 2024.


*   © 2024, Posted by Cold Spring Harbor Laboratory

This pre-print is available under a Creative Commons License (Attribution-NonCommercial-NoDerivs 4.0 International), CC BY-NC-ND 4.0, as described at [http://creativecommons.org/licenses/by-nc-nd/4.0/](http://creativecommons.org/licenses/by-nc-nd/4.0/)

## References

1.  1.GBD 2019 Mental Disorders Collaborators (2022): Global, regional, and national burden of 12 mental disorders in 204 countries and territories, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet Psychiatry 9: 137–150.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/s2215-0366(21)00395-3&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=35026139&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

2.  2.Depression in Adults: Treatment and Management (2022): London: National Institute for Health and Care Excellence (NICE). Retrieved May 23, 2024, from [http://www.ncbi.nlm.nih.gov/books/NBK583074/](http://www.ncbi.nlm.nih.gov/books/NBK583074/)
    
    

3.  3.Guideline Development Panel for the Treatment of Depressive Disorders (2022): Summary of the clinical practice guideline for the treatment of depression across three age cohorts. Am Psychol 77: 770–780.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34843274&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

4.  4.Cipriani A, Furukawa TA, Salanti G, Chaimani A, Atkinson LZ, Ogawa Y, et al. (2018): 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 391: 1357–1366.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/S0140-6736(17)32802-7&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=29477251&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

5.  5.Trivedi MH, Rush AJ, Wisniewski SR, Nierenberg AA, Warden D, Ritz L, et al. (2006): Evaluation of outcomes with citalopram for depression using measurement-based care in STAR*D: implications for clinical practice. Am J Psychiatry 163: 28–40.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1176/appi.ajp.163.1.28&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=16390886&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000234367100010&link_type=ISI) 

6.  6.Fabbri C, Corponi F, Souery D, Kasper S, Montgomery S, Zohar J, et al. (2019): The Genetics of Treatment-Resistant Depression: A Critical Review and Future Perspectives. Int J Neuropsychopharmacol 22: 93–104.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=29688548&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

7.  7.Nikkheslat N, McLaughlin AP, Hastings C, Zajkowska Z, Nettis MA, Mariani N, et al. (2020): Childhood trauma, HPA axis activity and antidepressant response in patients with depression. Brain Behav Immun 87: 229–237.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=31794798&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

8.  8.Dreimüller N, Lieb K, Tadić A, Engelmann J, Wollschläger D, Wagner S (2019): Body mass index (BMI) in major depressive disorder and its effects on depressive symptomatology and antidepressant response. J Affect Disord 256: 524–531.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=31280076&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

9.  9.Quilty LC, Meusel L-AC, Bagby RM (2008): Neuroticism as a mediator of treatment response to SSRIs in major depressive disorder. J Affect Disord 111: 67–73.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.jad.2008.02.006&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=18384882&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000260663800009&link_type=ISI) 

10. 10.De Carlo V, Calati R, Serretti A (2016): Socio-demographic and clinical predictors of non-response/non-remission in treatment resistant depressed patients: A systematic review. Psychiatry Res 240: 421–430.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.psychres.2016.04.034&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=27155594&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

11. 11.Gronemann FH, Jorgensen MB, Nordentoft M, Andersen PK, Osler M (2020): Socio-demographic and clinical risk factors of treatment-resistant depression: A Danish population-based cohort study. J Affect Disord 261: 221–229.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=31655377&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

12. 12.Maslej MM, Furukawa TA, Cipriani A, Andrews PW, Sanches M, Tomlinson A, et al. (2021): Individual Differences in Response to Antidepressants: A Meta-analysis of Placebo-Controlled Randomized Clinical Trials. JAMA Psychiatry 78: 490–497.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=33595620&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

13. 13.Castrén E, Monteggia LM (2021): Brain-Derived Neurotrophic Factor Signaling in Depression and Antidepressant Action. Biol Psychiatry 90: 128–136.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.biopsych.2021.05.008&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34053675&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

14. 14.Kofod J, Elfving B, Nielsen EH, Mors O, Köhler-Forsberg O (2022): Depression and inflammation: Correlation between changes in inflammatory markers with antidepressant response and long-term prognosis. Eur Neuropsychopharmacol 54: 116– 125.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34598835&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

15. 15.Wigmore EM, Hafferty JD, Hall LS, Howard DM, Clarke T-K, Fabbri C, et al. (2020): Genome-wide association study of antidepressant treatment resistance in a population-based cohort using health service prescription data and meta-analysis with GENDEP. Pharmacogenomics J 20: 329–341.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41397-019-0067-3&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=30700811&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

16. 16. GENDEP Investigators,  MARS Investigators, STAR*D Investigators (2013): Common genetic variation and antidepressant efficacy in major depressive disorder: a meta-analysis of three genome-wide pharmacogenetic studies. Am J Psychiatry 170: 207– 217.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1176/appi.ajp.2012.12020237&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=23377640&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000314201500013&link_type=ISI) 

17. 17.Tansey KE, Guipponi M, Perroud N, Bondolfi G, Domenici E, Evans D, et al. (2012): Genetic predictors of response to serotonergic and noradrenergic antidepressants in major depressive disorder: a genome-wide analysis of individual-level data and a meta-analysis. PLoS Med 9: e1001326.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1371/journal.pmed.1001326&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=23091423&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

18. 18.Li D, Pain O, Chiara F, Wong WLE, Lo CWH, Ripke S, et al. (2024): Metabolic activity of CYP2C19 and CYP2D6 on antidepressant response from 13 clinical studies using genotype imputation: a meta-analysis. Transl Psychiatry 14: 296.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=39025838&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

19. 19.Pain O, Hodgson K, Trubetskoy V, Ripke S, Marshe VS, Adams MJ, et al. (2022): Identifying the Common Genetic Basis of Antidepressant Response. Biol Psychiatry Glob Open Sci 2: 115–126.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=35712048&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

20. 20.Davidson EM, Casey A, Grover C, Alex B, Wu H, Campbell A, et al. (2023, April 4): The epidemiological characteristics of stroke phenotypes defined with ICD-10 and free-text: a cohort study linked to electronic health records. doi:10.1101/2023.04.03.23288096
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NzoibWVkcnhpdiI7czo1OiJyZXNpZCI7czoyMToiMjAyMy4wNC4wMy4yMzI4ODA5NnYxIjtzOjQ6ImF0b20iO3M6NTA6Ii9tZWRyeGl2L2Vhcmx5LzIwMjQvMTEvMTAvMjAyNC4xMS4wOS4yNDMxNjk4Ny5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

21. 21.Wong WLE, Fabbri C, Laplace B, Li D, van Westrhenen R, Lewis CM, et al. (2023): The Effects of CYP2C19 Genotype on Proxies of SSRI Antidepressant Response in the UK Biobank. Pharmaceuticals (Basel*)* 16: 1277.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=37765085&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

22. 22.Fabbri C, Hagenaars SP, John C, Williams AT, Shrine N, Moles L, et al. (2021): Genetic and clinical characteristics of treatment-resistant depression using primary care records in two UK cohorts. Mol Psychiatry 26: 3363–3373.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41380-021-01062-9&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=33753889&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

23. 23.Brent D, Emslie G, Clarke G, Wagner KD, Asarnow JR, Keller M, et al. (2008): Switching to another SSRI or to venlafaxine with or without cognitive behavioral therapy for adolescents with SSRI-resistant depression: the TORDIA randomized controlled trial. JAMA 299: 901–913.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1001/jama.299.8.901&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=18314433&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000253413200019&link_type=ISI) 

24. 24.Rush AJ, South C, Jha MK, Jain SB, Trivedi MH (2020): What to Expect When Switching to a Second Antidepressant Medication Following an Ineffective Initial SSRI: A Report From the Randomized Clinical STAR*D Study. J Clin Psychiatry 81: 19m12949.
    
    

25. 25.Milbourn H, McCartney D, Richmond A, Campbell A, Flaig R, Robertson S, et al. (2024): Generation Scotland: an update on Scotland’s longitudinal family health study. BMJ Open 14: e084719.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=38908846&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

26. 26.Bycroft C, Freeman C, Petkova D, Band G, Elliott LT, Sharp K, et al. (2018): The UK Biobank resource with deep phenotyping and genomic data. Nature 562: 203–209.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41586-018-0579-z&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=30305743&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

27. 27.Elliott P, Peakman TC, on behalf of UK Biobank (2008): The UK Biobank sample handling and storage protocol for the collection, processing and archiving of human blood and urine. International Journal of Epidemiology 37: 234–244.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/ije/dym276&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=18381398&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000254713400004&link_type=ISI) 

28. 28. The UK Biobank (2022, September 26): Health-related outcomes data. Retrieved June 26, 2023, from [https://www.ukbiobank.ac.uk/enable-your-research/about-our-data/health-related-outcomes-data](https://www.ukbiobank.ac.uk/enable-your-research/about-our-data/health-related-outcomes-data)
    
    

29. 29.Wray NR, Ripke S, Mattheisen M, Trzaskowski M, Byrne EM, Abdellaoui A, et al. (2018): Genome-wide association analyses identify 44 risk variants and refine the genetic architecture of major depression. Nat Genet 50: 668–681.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41588-018-0090-3&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=29700475&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

30. 30.Trubetskoy V, Pardiñas AF, Qi T, Panagiotaropoulou G, Awasthi S, Bigdeli TB, et al. (2022): Mapping genomic loci implicates genes and synaptic biology in schizophrenia. Nature 604: 502–508.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/S41586-022-04434-5&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=35396580&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

31. 31.Ge T, Chen C-Y, Ni Y, Feng Y-CA, Smoller JW (2019): Polygenic prediction via Bayesian regression and continuous shrinkage priors. Nat Commun 10: 1776.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=30992449&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

32. 32.Pain O, Al-Chalabi A, Lewis CM (2024): The GenoPred Pipeline: A Comprehensive and Scalable Pipeline for Polygenic Scoring. Bioinformatics btae551.
    
    

33. 33.1000 Genomes Project Consortium, Auton A, Brooks LD, Durbin RM, Garrison EP, Kang HM, et al. (2015): A global reference for human genetic variation. Nature 526: 68–74.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/nature15393&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=26432245&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

34. 34.Mbatchou J, Barnard L, Backman J, Marcketta A, Kosmicki JA, Ziyatdinov A, et al. (2021): Computationally efficient whole-genome regression for quantitative and binary traits. Nat Genet 53: 1097–1103.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/S41588-021-00870-7&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34017140&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

35. 35.Yang J, Lee SH, Goddard ME, Visscher PM (2011): GCTA: a tool for genome-wide complex trait analysis. Am J Hum Genet 88: 76–82.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.ajhg.2010.11.011&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=21167468&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

36. 36.Zeng J, Xue A, Jiang L, Lloyd-Jones LR, Wu Y, Wang H, et al. (2021): Widespread signatures of natural selection across human complex traits and functional genomic categories. Nat Commun 12: 1164.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41467-021-21446-3&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=33608517&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

37. 37.Boyce P, Hopwood M, Morris G, Hamilton A, Bassett D, Baune BT, et al. (2020): Switching antidepressants in the treatment of major depression: When, how and what to switch to? J Affect Disord 261: 160–163.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=31630037&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

38. 38.Cleare A, Pariante C, Young A, Anderson I, Christmas D, Cowen P, et al. (2015): Evidence-based guidelines for treating depressive disorders with antidepressants: A revision of the 2008 British Association for Psychopharmacology guidelines. J Psychopharmacol 29: 459–525.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1177/0269881115581093&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=25969470&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

39. 39.Mars B, Heron J, Gunnell D, Martin RM, Thomas KH, Kessler D (2017): Prevalence and patterns of antidepressant switching amongst primary care patients in the UK. J Psychopharmacol 31: 553–560.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=28460603&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

40. 40.Anderson IM, Nutt DJ, Deakin JF (2000): Evidence-based guidelines for treating depressive disorders with antidepressants: a revision of the 1993 British Association for Psychopharmacology guidelines. British Association for Psychopharmacology. J Psychopharmacol 14: 3–20.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1177/026988110001400101&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=10757248&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000086020400001&link_type=ISI) 

41. 41.Elwadhi D, Cohen A (2020): Social inequalities in antidepressant treatment outcomes: a systematic review. Soc Psychiatry Psychiatr Epidemiol 55: 1241–1259.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=32666210&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

42. 42.Sundell KA, Waern M, Petzold M, Gissler M (2013): Socio-economic determinants of early discontinuation of anti-depressant treatment in young adults. Eur J Public Health 23: 433–440.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/eurpub/ckr137&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=21953063&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000319468600021&link_type=ISI) 

43. 43.Kelly KM, Mezuk B (2017): Predictors of remission from generalized anxiety disorder and major depressive disorder. J Affect Disord 208: 467–474.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=27863710&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

44. 44.Kwong ASF, Morris TT, Pearson RM, Timpson NJ, Rice F, Stergiakouli E, Tilling K (2021): Polygenic risk for depression, anxiety and neuroticism are associated with the severity and rate of change in depressive symptoms across adolescence. J Child Psychol Psychiatry 62: 1462–1474.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=33778956&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

45. 45.Halldorsdottir T, Piechaczek C,  Soares de Matos AP, Czamara D, Pehl V, Wagenbuechler P, et al. (2019): Polygenic Risk: Predicting Depression Outcomes in Clinical and Epidemiological Cohorts of Youths. Am J Psychiatry 176: 615–625.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=30947532&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

46. 46.García-González J, Tansey KE, Hauser J, Henigsberg N, Maier W, Mors O, et al. (2017): Pharmacogenetics of antidepressant response: A polygenic approach. Prog Neuropsychopharmacol Biol Psychiatry 75: 128–134.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.pnpbp.2017.01.011&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=28159590&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

47. 47.Fanelli G, Benedetti F, Kasper S, Zohar J, Souery D, Montgomery S, et al. (2021): Higher polygenic risk scores for schizophrenia may be suggestive of treatment non-response in major depressive disorder. Prog Neuropsychopharmacol Biol Psychiatry 108: 110170.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.pnpbp.2020.110170&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=33181205&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

48. 48.Davies MR, Kalsi G, Armour C, Jones IR, McIntosh AM, Smith DJ, et al. (2019): The Genetic Links to Anxiety and Depression (GLAD) Study: Online recruitment into the largest recontactable study of depression and anxiety. Behav Res Ther 123: 103503.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=31715324&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom) 

49. 49.Wong J, Motulsky A, Eguale T, Buckeridge DL, Abrahamowicz M, Tamblyn R (2016): Treatment Indications for Antidepressants Prescribed in Primary Care in Quebec, Canada, 2006-2015. *JAMA* 315: 2230.
    
    

50. 50.McIntosh AM, Lewis CM, Mark J Adams for the Psychiatric Genomics Consortium Major Depressive Disorder Working Group (2024, May 1): Genome-wide study of half a million individuals with major depression identifies 697 independent associations, infers causal neuronal subtypes and biological targets for novel pharmacotherapies. doi:10.1101/2024.04.29.24306535
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NzoibWVkcnhpdiI7czo1OiJyZXNpZCI7czoyMToiMjAyNC4wNC4yOS4yNDMwNjUzNXYyIjtzOjQ6ImF0b20iO3M6NTA6Ii9tZWRyeGl2L2Vhcmx5LzIwMjQvMTEvMTAvMjAyNC4xMS4wOS4yNDMxNjk4Ny5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

51. 51.Cai N, Revez JA, Adams MJ, Andlauer TFM, Breen G, Byrne EM, et al. (2020): Minimal phenotyping yields genome-wide association signals of low specificity for major depression. Nat Genet 52: 437–447.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41588-020-0594-5&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=32231276&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F10%2F2024.11.09.24316987.atom)