Association Between CYP450 Polymorphisms and Treatment Outcomes of Escitalopram and Venlafaxine in Patients with Post Stroke Depression Authors: Yanyan Wang, Wenzhe Sun, Chensheng Pan, Zhou Zhu, Guo Li, Xin Zhao, Suiqiang Zhu Published: 2026 ABSTRACT AbstractBackgroundGenetic polymorphisms in drug metabolism play an important role in the wide inter-individual variability in antidepressant efficacy. This study aimed to determine whether cytochrome P450 2C19 (CYP2C19) and cytochrome P450 2D6 (CYP2D6) genotypes predict treatment response to escitalopram (ESC) and venlafaxine (VEN), respectively, in patients with post-stroke depression (PSD).MethodsIn this single-center prospective observational study, 313 acute stroke patients with PSD were consecutively enrolled and received ESC or VEN for 8 weeks. Efficacy was assessed using the 17-item Hamilton Depression Scale (HAMD-17). The primary outcome was the change in HAMD-17 score from baseline to week 8; secondary outcomes were response (≥50% reduction) and remission (score ≤7). Patients were stratified by CYP2C19 and CYP2D6 genotypes into fast- and slow-metabolizer subgroups. Subgroup comparisons utilized χ2-tests and Mann–Whitney U-tests.ResultsAt week 8, among 148 ESC-treated patients, CYP2C19 genotypes were distributed as poor/intermediate metabolizers (PM/IM) 82 (55.4%) and normal/rapid metabolizers (NM/RM) 64 (43.2%); among 113 VEN-treated patients, CYP2D6 genotypes showed marked imbalance (only 2 IM, 111 NM). No significant difference was found between ESC PM/IM and NM/RM subgroups in HAMD‑17 reduction (median: 10 vs 11, P=0.434), response (61.0% vs 69.7%, P=0.269), or remission (48.8% vs 51.5%, P=0.741). For patients treated with VEN, there was a significant imbalance in the distribution of CYP2D6 genotypes, which prevented us from further analyzing the association between CYP2D6 gene polymorphisms and the efficacy of VEN.ConclusionNo significant effect of CYP2C19 genetic polymorphism on ESC efficacy was observed in this study. These findings underscore the need for further exploration into the clinical utility of pharmacogenetics. Future antidepressant treatment should prioritize cost-effective, pragmatic approaches over costly genetic testing alone.Keywords: pharmacogenomics, cytochrome P450 enzymes, antidepressant, post-stroke depression Introduction Post-stroke depression (PSD) is the most prevalent neuropsychiatric complication following stroke.1–4 Given incidence rates peaking within the first six months and a cumulative prevalence of up to 52% within five years,5–8 PSD poses a major barrier to long-term functional recovery and prognosis in stroke survivors. While current guidelines recommend antidepressants and psychotherapy for prevention and early intervention,9,10 a major impediment to effective management is the inter-individual variability in antidepressant efficacy, the mechanisms of which remain inadequately understood.11 In recent years, genetic polymorphisms have emerged as key determinants of individual drug response, primarily by modulating pharmacokinetic profiles.12 Among these, cytochrome P450 (CYP450) enzymes, particularly CYP2C19 and CYP2D6, play pivotal roles in the metabolism of widely used antidepressants. Substantial evidence indicates that CYP2C19 variants significantly alter escitalopram (ESC) exposure; for instance, carriers of the CYP2C19 *17/*17 genotype typically achieve less than two-thirds of the steady-state concentration observed in CYP2C19 *1/*1 homozygotes.13,14 Similarly, there were studies showing that venlafaxine (VEN) concentrations are related to polymorphisms in CYP2D6.15,16 These findings underscore the potential of pharmacogenetics to explain differential drug response at the biochemical level. However, a critical translational gap remains. Although the influence of CYP polymorphisms on plasma concentrations is well-documented, it is far less clear whether these genotypes reliably predict clinical treatment outcomes in real-world psychiatric settings.17 Moreover, compared to the extensive literature on major depressive disorder, data regarding the impact of CYP2C19 and CYP2D6 polymorphisms on antidepressant efficacy in the specific context of PSD are notably scarce. This paucity of clinical-outcome data limits the application of personalized medicine for stroke survivors. To bridge this gap, we conducted a prospective cohort study investigating CYP2C19/CYP2D6 effects on ESC/VEN efficacy in PSD. We selected these specific agents as probes because their metabolism is selectively governed by CYP2C19 and CYP2D6, respectively,18,19 allowing for a clear evaluation of genotype-treatment response linkages. Materials and Methods Study Design and Subjects A single-center prospective cohort study (ChiCTR2100045169) was conducted in the Department of Neurology, Tongji Hospital, Wuhan, Hubei Province, which was approved by the Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology (Approved No. of ethic committee: TJ-IRB20210142). From January to July 2021, 313 consecutive patients meeting DSM-5 criteria for PSD with HAMD-17 scores ≥14 were enrolled, with subsequent allocation to ESC or VEN based on the treating physician’s evaluation of clinical indications and contraindications. The dose of ESC was 5mg for the first two days and 10mg for the rest. The dose of venlafaxine was 37.5mg for the first two days, 75mg for the third to fourth days, and a fixed dose of 150mg from the fifth day. Finally, 261 patients were eligible for the research (Figure 1). All patients involved in this study or their family members signed written informed consents in accordance with the Declaration of Helsinki. The enrollment flow chart of this study. Inclusion and Exclusion Criteria The inclusion criteria of this study were as follows: (1) ages 18 to 75; (2) admitted to hospital within 14 days after stroke onset, including ischemic as well as hemorrhagic stroke; (3) stroke was confirmed by computed tomography (CT) or magnetic resonance imaging (MRI) scans; (4) patients who met DSM-V diagnostic criteria for PSD and had a HAMD-17 score of 14 or greater; (5) no antidepressant treatment was received prior to this hospitalization. Exclusion criteria were: (1) brain dysfunction caused by other non-vascular causes (such as primary brain tumor, brain metastasis, subdural hematoma, post-seizure paralysis, brain trauma, etc); (2) the history of dementia, depression and other mental illness; (3) aphasia, blindness, deafness, or cognitive dysfunction of a degree that precludes the ability to complete basic orientation, memory tasks, and follow simple instructions; (4) transient ischemic attack and subarachnoid hemorrhage; (5) unable to complete the follow-up; (6) severe PSD at risk of suicide; (7) PSD accompanied with psychotic symptoms; (8) Patients with bipolar disorder were identified during follow-up. Data Collection Demographic information and medical history were collected using standardized questionnaires. The HAMD-17,20 Hamilton Anxiety Scale (HAMA),21 9-item Patient Health Questionnaire (PHQ-9),22,23 and 15-item Patient Health Questionnaire (PHQ-15)24—all validated in Chinese populations—were administered to assess psychiatric symptoms. Neurological deficits and functional outcomes were evaluated using the National Institutes of Health Stroke Scale (NIHSS), Barthel Index (BI), and Modified Rankin Scale (mRS). Baseline assessment was conducted face-to-face during the patient’s hospital stay. The subsequent follow-up assessments (at the 2nd, 4th, and 8th weeks) were mainly conducted remotely via telephone interviews or WeChat video calls. For patients who had been discharged, this approach was adopted to ensure the continuity of longitudinal data collection while minimizing the burden on the patients. Genotyping Strategy Venous blood samples were collected from all patients at baseline and stored as whole blood in the Biobank Center of Tongji Hospital, Wuhan. After all patients had completed treatment assignment and clinical evaluations, genotyping to identify CYP2C19 and CYP2D6 alleles was outsourced to and performed by Beijing Bo Miao Biotechnology Co., Ltd. Three CYP2C19 major alleles, including *2(rs4244285, 681G>A), *3(rs4986893, 636G>A) and *17(rs12248560, −806C>T), were genotyped. Three variant CYP2D6 alleles (*2, *4, *10) and the wild-type 1 reference were genotyped. This panel, targeting the predominant Han Chinese alleles (*1, *2, *10)25 plus the loss-of-function *4, balanced clinical rigor with economic feasibility. Although *5, *14, and *41 can alter phenotypes in rare compound states, their omission was a pragmatic trade-off to avoid costly custom assays. Phenotypic Criteria Patients treated with ESC were categorized by CYP2C19 metabolizer status using the common consensus star allele nomenclature.26 CYP2C19 poor metabolizers (PMs) were characterized by the presence of two no function alleles (*2/*2, *2/*3 and*3/*3). CYP2C19 intermediate metabolizers (IMs) were characterized by the presence of one normal function allele and one no function allele (*1/*2, *1/*3), or one no function allele and one increased function allele (*2/*17, *3/*17). CYP2C19 normal metabolizers (NMs) were characterized by the presence of two normal function alleles (*1/*1), which were called extensive metabolizers (EMs) in the past.27 CYP2C19 *1/*17 was classified as rapid metabolizers (RMs) and CYP2C19 *17/*17 was classified as ultra-metabolizers (UMs).28 Patients treated with VEN were categorized by CYP2D6 metabolizer status. Patients with CYP2D6 *4/*4 were classified as PMs, those with CYP2D6 *10/*10 were classified as intermediate metabolizers IMs, those with CYP2D6 *1/*1 or *1/*2 or *2/*2 or *1/*10 or *2/*10 were classified as NMs and individuals carrying CYP2D6 *1/*1×N or *1/*2×N or *2/*2×N were classified as UMs. N is categorical and indicates the number of copy variants (eg, *1×2, *1×3).29 Statistical Analysis The primary outcome was the decrease of HAMD-17 score from baseline to week 8. Secondary outcomes included response and remission at week 8 according to HAMD-17. Response was defined as ≥ 50% decrease of HAMD-17 score from baseline.18 Remission was defined as HAMD-17 score ≤ 7.12 Group differences were assessed by χ2 tests for categorical variables and nonparametric Mann–Whitney U-test for continuous variables. Categorical variables were expressed as frequencies (percentages) while continuous variables were presented as medians [interquartile range (IQR)]. In addition, multiple linear regression analysis was used for finding factors of the decrease of HAMD-17 score and multivariate logistic regression analysis was conducted to explore independent predictor of response and remission at week 8. Variables with P<0.05 in the univariate logistic analysis were regarded as potential risk factors and included in the multivariate logistic regression model, which was processed using the backward stepwise method. Overall, a 2-sided value of P<0.05 was considered statistically significant. All data analysis were performed using the Statistical Program for Social Sciences (SPSS) statistical software (version 25, Chicago, IL, USA). Results A total of 313 stroke patients were enrolled and 261 of them were finally eligible for the research. 188 (72.0%) were male and 73 (28.0%) were female with a median age of 59 years. Of 148 patients treated with ESC, 18 were PMs (*2/*2, n=15; *2/*3, n=3), 64 were IMs (*1/*2, n=56; *1/*3, n=8), 63 were NMs (*1/*1, n=63) and only 3 were RMs (*1/*17, n=3). Of 113 patients treated with VEN, 2 were IMs (*10/*10, n=2) and 111 were NMs (*1/*1, n=15; *1/*2, n=15; *1/*10, n=61; *2/*10, n=20). As shown above, there was a significant imbalance in the distribution of CYP2D6 genotypes, which prevented us from further analyzing the association between CYP2D6 gene polymorphisms and the efficacy of VEN. As for patients treated with ESC, we combined PMs and IMs into one group, and NMs and RMs into one group for the purpose of keeping the number of people in each subgroup roughly equal. The grouping method was referred to previous studies.18,27 As shown in Table 1, there was no significant difference between CYP2C19 PMs/IMs and CYP2C19 NMs/RMs in the 8week decrease of HAMD-17 score (median: 10 vs 11, P=0.434), with a Hodges–Lehmann estimated difference of −1.0 points (95% CI: −3.0 to 1.0). Similarly, no significant between‑group differences were observed in the response (61.0% vs 69.7%, P=0.269) or remission rates (48.8% vs 51.5%, P=0.741). Figure 2 visually illustrates the longitudinal response patterns of ESC, showing generally consistent HAMD-17 trajectories (Figure 2A) and response/remission curves (Figure 2B) between groups, aligning with the statistical results presented in Table 1. Table 1Baseline and Follow-up Characteristics of Patients Treated with Escitalopram Stratified by Metabolizer PhenotypeCovariateEscitalopram Stratified by Metabolizer Phenotype of CYP2C19PMs/IMs (n=82)NMs/RMs (n=66)PAge, y59(49–66)58(49–65)0.808BMI24.2(21.9–26.1)23.1(21.4–25.4)0.075Female, N (%)26(31.7)15(22.7)0.225Sleep time<5 h18(22.0)21(31.8)0.176Physical activity, N (%)9(11.0)7(10.6)0.943Smoking, N (%)39(47.6)35(53.0)0.508Drinking, N (%)41(50.0)29(43.9)0.463Diabetes, N (%)24(29.3)26(39.4)0.195Hypertension, N (%)59(72.0)45(68.2)0.618Hyperlipidemia, N (%)14(17.1)9(13.6)0.566CHD, N (%)7(8.5)7(10.6)0.669Ischemic Stroke, N (%)63(76.8)55(83.3)0.328Baseline HAMD-1718(15–21)17(15–22)0.678Baseline HAMA11.5(10–15)11(10–14)0.890Baseline PHQ-911(10–14)12(9–15)0.489Baseline PHQ-159(7–11)9(7–11)0.559Baseline NIHSS3(1–6)3(1–8)0.428Baseline BI82.5(45–100)87.5(34–100)0.560Baseline mRS3(1–4)2(1–4)0.9842week HAMD-1714.5(11–17)14(11–16)0.4002week Decrease of HAMD-17 score4(1–7)3.5(0–8)0.9942week Response, N (%)9(11.0)11(16.7)0.3142week Remission, N (%)6(7.3)5(7.6)1.0004week HAMD-1710(7–14)11.5(8–14)0.4654week Decrease of HAMD-17 score7.5(4–11)7.5(5–10)0.6364week Response, N (%)34(41.5)22(33.3)0.3114week Remission, N (%)22(26.8)15(22.7)0.5678week HAMD-178(5–12)7(5–10)0.6258week Decrease of HAMD-17 score10(6–14)11(7–14)0.4348week Response, N (%)50(61.0)46(69.7)0.2698week Remission, N (%)40(48.8)34(51.5)0.741Abbreviations: PMs, Poor Metabolizers; IMs, Intermediate Metabolizers; NMs, Normal Metabolizers; RMs, Rapid Metabolizers; BMI, Body mass index; CHD, Coronary heart disease; HAMD-17, 17-item Hamilton Depression Scale; HAMA, Hamilton Anxiety Scale; PHQ-9, 9-item Patient Health Questionnaire; PHQ-15, 15-item Patient Health Questionnaire; NIHSS, The National Institutes of Health Stroke Scale; BI, Barthel Index; mRS, Modified Rankin Scale. Baseline and Follow-up Characteristics of Patients Treated with Escitalopram Stratified by Metabolizer Phenotype Abbreviations: PMs, Poor Metabolizers; IMs, Intermediate Metabolizers; NMs, Normal Metabolizers; RMs, Rapid Metabolizers; BMI, Body mass index; CHD, Coronary heart disease; HAMD-17, 17-item Hamilton Depression Scale; HAMA, Hamilton Anxiety Scale; PHQ-9, 9-item Patient Health Questionnaire; PHQ-15, 15-item Patient Health Questionnaire; NIHSS, The National Institutes of Health Stroke Scale; BI, Barthel Index; mRS, Modified Rankin Scale. Response patterns of escitalopram in CYP2C19 phenotypic subgroups. (A) HAMD-17 trajectories; Shaded areas represent IQR. (B) Response and remission rates. Table 2 presents the multivariable linear regression analysis of the 8-week HAMD-17 score reduction. The model adjusted for CYP2C19 genotype, demographic characteristics, lifestyle factors, comorbidities, stroke type, and baseline psychiatric and neurological scales. The multiple linear regression model explained 47.5% of the variance in 8week decrease of HAMD-17 score (Adjusted R2 =0.475, F=7.653, P<0.001). Baseline HAMD-17 (P<0.001) and baseline mRS (P=0.008) were proved to be significantly associated with 8week decrease of HAMD-17 score. However, no relationship between CYP2C19 polymorphisms and 8week decrease of HAMD-17 score was found (P=0.439). Among VEN-treated patients, multivariable linear regression adjusting for the aforementioned covariates (excluding genetic polymorphisms) demonstrated that the model accounted for 66.8% of the variance in the 8-week HAMD-17 score reduction (adjusted R2= 0.668, F=12.880, P<0.001). Similarly, baseline HAMD-17 (P<0.001) was found to be significantly associated with 8week decrease of HAMD-17 score. Table 2Multiple Linear Regression Analysis for 8 Week Decrease of HAMD-17 Score of Escitalopram and VenlafaxineCovariateEscitalopram (n=148)Venlafaxine (n=113)BSEBetatPFAdjusted R2BSEBetatPFAdjusted R2Constant−6.346.89−0.920.3597.6530.475−8.335.523−1.5080.13512.8800.668NMs/RMs0.550.710.050.780.439Age0.550.040.060.800.4230.010.030.020.360.718BMI0.030.120.040.570.5690.190.110.121.790.077Female0.071.07−0.06−0.660.510−0.580.73−0.06−0.790.429Sleep time<5h−0.710.86−0.01−0.060.951−1.190.59−0.13−2.010.047*Physical activity−0.051.150.101.550.1230.121.160.010.100.921Smoking1.780.980.050.560.5800.790.760.091.030.304Drinking0.550.96−0.09−1.080.284−0.990.68−0.11−1.450.152Diabetes−1.030.76−0.03−0.380.706−0.430.68−0.04−0.640.524Hypertension−0.290.81−0.05−0.730.466−0.540.59−0.06−0.930.357Hyperlipidemia−0.600.99−0.11−1.630.1050.540.730.050.740.459CHD−1.611.200.101.540.127−0.980.78−0.07−1.250.214Ischemic Stroke1.840.95−0.04−0.590.5570.040.910.0030.050.962Baseline HAMD-17−0.560.160.927.91<0.001*1.130.121.149.49<0.001*Baseline HAMA1.240.14−0.17−1.510.133−0.040.09−0.04−0.480.631Baseline PHQ-9−0.210.15−0.01−0.060.951−0.250.12−0.25−2.060.043*Baseline PHQ-15−0.010.150.030.310.754−0.180.10−0.14−1.730.086Baseline NIHSS0.050.12−0.04−0.420.679−0.120.13−0.10−0.930.356Baseline BI−0.050.03−0.13−0.770.4450.040.030.231.530.130Baseline mRS−0.020.62−0.45−2.690.008*−0.030.60−0.01−0.050.962Note: *Statistically significant at P<0.05 level, 2-sided.Abbreviations: NMs, Normal Metabolizers; RMs, Rapid Metabolizers; BMI, Body mass index; CHD, Coronary heart disease; HAMD-17, 17-item Hamilton Depression Scale; HAMA, Hamilton Anxiety Scale; PHQ-9, 9-item Patient Health Questionnaire; PHQ-15, 15-item Patient Health Questionnaire; NIHSS, The National Institutes of Health Stroke Scale; BI, Barthel Index; mRS, Modified Rankin Scale. Multiple Linear Regression Analysis for 8 Week Decrease of HAMD-17 Score of Escitalopram and Venlafaxine Note: *Statistically significant at P<0.05 level, 2-sided. Abbreviations: NMs, Normal Metabolizers; RMs, Rapid Metabolizers; BMI, Body mass index; CHD, Coronary heart disease; HAMD-17, 17-item Hamilton Depression Scale; HAMA, Hamilton Anxiety Scale; PHQ-9, 9-item Patient Health Questionnaire; PHQ-15, 15-item Patient Health Questionnaire; NIHSS, The National Institutes of Health Stroke Scale; BI, Barthel Index; mRS, Modified Rankin Scale. As shown in Table 3, CYP2C19 genotype did not independently predict ESC response or remission. Notably, baseline disease severity emerged as a key prognostic factor: greater depressive symptom burden (higher HAMD-17) correlated with a higher likelihood of response (adjusted OR=1.29, 95% CI: 1.14–1.45, P<0.001), whereas greater functional disability (higher mRS) significantly diminished the likelihood of both response (adjusted OR=0.49, 95% CI: 0.36–0.67, P<0.001) and remission (adjusted OR=0.54, 95% CI: 0.42–0.69, P<0.001). As shown in Table 4, baseline BI score was independently associated with an increased likelihood of 8-week treatment response (adjusted OR = 1.03, 95% CI: 1.01–1.05, P<0.001) and remission (adjusted OR = 1.04, 95% CI: 1.02–1.06, P=0.001) among VEN-treated patients. Conversely, a higher baseline PHQ-9 score was associated with reduced odds of remission (adjusted OR = 0.81, 95% CI: 0.72–0.92, P<0.001). Table 3Univariate and Multivariate Logistic Regression Analysis for 8 Week Response (Left) and 8week Remission(Right) of EscitalopramCovariateUnivariate Logistic AnalysisMultivariate Logistic AnalysisUnivariate Logistic AnalysisMultivariate Logistic AnalysisPCrude OR (95% CI)PAdjusted OR (95% CI)PCrude OR (95% CI)PAdjusted OR (95% CI)NMs/RMs0.2701.47(0.74–2.93)0.7411.12(0.58–2.13)Age0.7251.01(0.98–1.04)0.3970.99(0.96–1.02)BMI0.5650.97(0.86–1.09)0.9821.00(0.90–1.11)Female0.8761.06(0.50–2.27)0.8541.07(0.52–2.20)Sleep time<5h0.0702.17(0.94–5.02)0.1931.64(0.78–3.43)Physical activity0.3731.71(0.52–5.61)0.5971.33(0.47–3.77)Smoking0.7310.89(0.45–1.75)0.7420.90(0.47–1.71)Drinking0.0640.52(0.26–1.04)0.3240.72 (0.38–1.38)Diabetes0.6020.83(0.41–1.68)0.4870.79(0.40–1.55)Hypertension0.1841.64(0.79–3.39)1.0001.00(0.49–2.02)Hyperlipidemia0.6630.82(0.33–2.04)0.4970.73(0.30–1.80)CHD0.2270.51(0.17–1.53)0.2670.52(0.17–1.64)Ischemic Stroke0.5100.75(0.31–1.78)0.6831.18(0.53–2.64)Baseline HAMD-170.012*1.13(1.03–1.24)<0.001*1.29(1.14–1.45)0.2410.95(0.88–1.03)Baseline HAMA0.6441.02(0.94–1.10)0.029*0.91(0.84–0.99)Baseline PHQ-90.2421.06(0.96–1.15)0.1400.94(0.86–1.02)Baseline PHQ-150.4901.04(0.93–1.15)0.3050.95(0.86–1.05)Baseline NIHSS0.009*0.90(0.83–0.97)0.001*0.86(0.79–0.94)Baseline BI0.002*1.02(1.01–1.03)<0.001*1.03(1.01–1.04)Baseline mRS0.001*0.68(0.53–0.86)<0.0010.49(0.36–0.67)<0.001*0.54(0.42–0.69)<0.001*0.54(0.42–0.69)Note: *Statistically significant at P<0.05 level, 2-sided.Abbreviations: NMs, Normal Metabolizers; RMs, Rapid Metabolizers; BMI, Body mass index; CHD, Coronary heart disease; HAMD-17, 17-item Hamilton Depression Scale; HAMA, Hamilton Anxiety Scale; PHQ-9, 9-item Patient Health Questionnaire; PHQ-15, 15-item Patient Health Questionnaire; NIHSS, The National Institutes of Health Stroke Scale; BI, Barthel Index; mRS, Modified Rankin Scale. Univariate and Multivariate Logistic Regression Analysis for 8 Week Response (Left) and 8week Remission(Right) of Escitalopram Note: *Statistically significant at P<0.05 level, 2-sided. Abbreviations: NMs, Normal Metabolizers; RMs, Rapid Metabolizers; BMI, Body mass index; CHD, Coronary heart disease; HAMD-17, 17-item Hamilton Depression Scale; HAMA, Hamilton Anxiety Scale; PHQ-9, 9-item Patient Health Questionnaire; PHQ-15, 15-item Patient Health Questionnaire; NIHSS, The National Institutes of Health Stroke Scale; BI, Barthel Index; mRS, Modified Rankin Scale. Table 4Univariate and Multivariate Logistic Regression Analysis for 8 Week Response (Left) and 8week Remission(Right) of VenlafaxineCovariateUnivariate Logistic AnalysisMultivariate Logistic AnalysisUnivariate Logistic AnalysisMultivariate Logistic AnalysisPCrude OR (95% CI)PAdjusted OR (95% CI)PCrude OR (95% CI)PAdjusted OR (95% CI)Age0.6590.99(0.94–1.04)0.3520.98(0.94–1.02)BMI0.2031.13(0.94–1.36)0.7150.98(0.85–1.11)Female0.9031.07(0.38–3.03)0.2450.61(0.26–1.41)Sleep time<5h0.044*2.66(1.03–6.90)0.016*2.66(1.20–5.88)Physical activity0.8591.22(0.14–11.01)0.4490.53(0.10–2.75)Smoking0.8990.94(0.37–2.40)0.2861.54(0.70–3.38)Drinking0.1740.52(0.20–1.33)0.7991.11(0.51–2.43)Diabetes0.3311.67(0.60–4.66)0.8420.92(0.41–2.05)Hypertension0.1180.30(0.06–1.36)0.9431.03(0.41–2.61)Hyperlipidemia0.1862.82(0.61–13.09)0.6961.22(0.45–3.29)CHD0.1540.42(0.13–1.39)0.039*0.31(0.10–0.94)Ischemic Stroke0.2092.31(0.63–8.49)0.2701.97(0.59–6.57)Baseline HAMD-170.9171.00(0.90–1.10)0.002*0.86(0.78–0.94)Baseline HAMA0.3700.96(0.88–1.05)0.011*0.85(0.75–0.96)Baseline PHQ-90.1420.93(0.84–1.03)<0.001*0.80(0.72–0.89)<0.001*0.81(0.72–0.92)Baseline PHQ-150.3430.94(0.83–1.07)0.004*0.83(0.73–0.94)Baseline NIHSS0.002*0.82(0.73–0.93)<0.001*0.80(0.70–0.90)Baseline BI0.001*1.03(1.01–1.05)<0.001*1.03(1.01–1.05)<0.001*1.04(1.02–1.06)0.0011.04(1.02–1.06)Baseline mRS0.002*0.55(0.38–0.80)<0.001*0.43(0.29–0.64)Note: *Statistically significant at P<0.05 level, 2-sided.Abbreviations: BMI, Body mass index; CHD, Coronary heart disease; HAMD-17, 17-item Hamilton Depression Scale; HAMA, Hamilton Anxiety Scale; PHQ-9, 9-item Patient Health Questionnaire; PHQ-15, 15-item Patient Health Questionnaire; NIHSS, The National Institutes of Health Stroke Scale; BI, Barthel Index; mRS, Modified Rankin Scale. Univariate and Multivariate Logistic Regression Analysis for 8 Week Response (Left) and 8week Remission(Right) of Venlafaxine Note: *Statistically significant at P<0.05 level, 2-sided. Abbreviations: BMI, Body mass index; CHD, Coronary heart disease; HAMD-17, 17-item Hamilton Depression Scale; HAMA, Hamilton Anxiety Scale; PHQ-9, 9-item Patient Health Questionnaire; PHQ-15, 15-item Patient Health Questionnaire; NIHSS, The National Institutes of Health Stroke Scale; BI, Barthel Index; mRS, Modified Rankin Scale. Discussions This prospective cohort study yielded several principal findings. First, CYP2C19 genotype-based subgrouping did not lead to significant differences in ESC efficacy among patients with PSD, indicating that CYP2C19 polymorphism exerts minimal direct impact on treatment outcome. Second, baseline disease severity as shown by higher HAMD‑17 scores and by higher mRS scores in ESC patients emerged as a strong independent predictor of response and remission whereas genetic variation had no predictive effect. Third, the VEN group exhibited a markedly imbalanced CYP2D6 genotype distribution, precluding a robust pharmacogenetic analysis of VEN efficacy. The 8week response rate and remission rate in this study were significantly higher than those in a previous randomized controlled trial which enrolled outpatients diagnosed with major depressive disorder and with inadequate response to at least one antidepressant.12 However, the subjects in our study were inpatients who met DSM-V criteria for PSD along with a HAMD-17 score ≥14 and they were treated with antidepressant for the first time. Differences in study design and study population may be the key to the different results. Both CYP2C19 PMs/IMs and NMs/RMs achieved clinically meaningful HAMD‑17 reductions (HL estimate −1.0, 95% CI: −3.0 to 1.0). These findings are compatible with, at most, a minimal or absent effect of CYP2C19 polymorphism on ESC efficacy, aligning with Hodgson et al, who observed a clear genotype–concentration relationship but found that CYP2C19 status did not predict clinical response.17 Reis et al held the view that the difference in blood drug concentrations between individuals was resulted by not only the genetic variability in CYP450 enzymes but also environmental factors such as comedication, diet and comorbidities.30 Kirchheiner el al were ever committed to establish genotype specific dose adjustments for a range of antidepressants by exploring the relationship between CYP2D6 as well as CYP2C19 and serum concentrations of drug, although they did not examine the effect on treatment outcome.31,32 However, other studies on the role of CYP450 genetic variation in antidepressant treatment response reported discrepant findings.33–35 Hall-flavin et al demonstrated that outcomes of patients with depression can be improved when treatment is guided by pharmacogenomic algorithms containing cytochrome P450 genotype information.36,37 A large randomized controlled trial also showed that pharmacogenomic testing did not significantly improve mean symptoms but did significantly improve response and remission rates for difficult-to-treat depression patients over standard of care.12 Furthermore, NEJM Journal Watch took note of this trial and reviewed it critically. Because it was funded by the manufacturer that developed the gene test and several statistical tests were not corrected.38 In addition, our study unexpectedly found that baseline HAMD-17 score was an independent predictor of 8-week response rate of ESC. Moreover, the higher the baseline HAMD-17 score, the higher the response rate, which is highly consistent with previous findings.39,40 A meta-analyses shown that initial depression severity predicted symptom improvement in patients who took antidepressant medication.41 Irving Kirsch et al also reported that the mean differences between groups treated with antidepressant medication and placebo become larger as baseline depression severity increases.40 Given that this finding was not the main purpose of the study, we did not elaborate on it at length. Robust association analyses between CYP2D6 polymorphisms and VEN efficacy were precluded by several factors. Primarily, constrained by the initial research budget and the configuration of the commercial kit, our genotyping failed to comprehensively interrogate all clinically relevant CYP2D6 variant sites. We examined only the *1, *2, *4, and *10 alleles, which capture approximately 80% of CYP2D6 alleles in the Han Chinese population.25 Failure to detect low-frequency variations such as *5, *14 and *41 to some extent will affect the precise inference of metabolic phenotypes. However, Sun et al detected no PMs or UMs among 2467 Han Chinese patients, despite broader genotyping coverage including *5 and *14. In that large cohort, NMs accounted for 72.1%, and IMs (*10/*10) comprised 27.7%, robust evidence confirming the intrinsic rarity of PMs and UMs in the Han Chinese population.42 Consequently, the distribution observed in our VEN cohort (n=113)—111 NMs and 2 IMs (*10/*10), with zero PMs/UMs—aligns perfectly with established population genetics. This marked phenotypic imbalance served as another critical factor underlying the inability to perform association analyses between CYP2D6 polymorphisms and VEN efficacy. Dictated by the intrinsic genetic architecture of the Han Chinese population rather than experimental error, this constraint must be carefully considered when interpreting the VEN-related findings. In summary, whether gene polymorphism plays an important role in antidepressant therapy remains controversial. The existing and growing array of pharmacogenomic tests will pose clinical implementation challenges. Although genetic testing is not harmful to the human body, its high price greatly increases the financial burden of depressed patients. In addition, genetic testing may have other drawbacks, such as insurance coverage restrictions and genetic discrimination.43 In a nutshell, the choice of an antidepressant medication is unlikely to be significantly aided by determination of cytochrome genotypes but must await the development of readily accessible biomarkers of clinical response. The main limitations include the single-center design and relatively small sample size. For ESC-treated patients, the limited numbers of CYP2C19 PMs (n=18) and RMs (n=3) required pooling of metabolic subgroups, restricting inferences regarding the independent effect of the PM phenotype. For VEN-treated patients, the sample size, coupled with the intrinsic rarity of CYP2D6 PMs/UMs in Han Chinese, led to extreme phenotypic imbalance (zero PMs/UMs), precluding robust association analyses. Therefore, our findings warrant confirmation in larger, multi-center cohorts. Additionally, by enrolling only hospitalized patients during the acute/subacute phases of stroke with HAMD-17 scores ≥14, our cohort inherently represents moderate-to-severe cases, thereby restricting the extrapolation of our findings to milder presentations or the chronic stable phase managed in community settings. Conclusions The extent to which genetic variation can influence drug efficacy needs to be further explored. Pharmacogenetic testing may not have reached the threshold of clinical utility in antidepressant treatment; a more cost-effective and pragmatic approach should be adopted in future practice. Funding Statement This work was financially supported by the National Natural Science Fund of China (grant number 82171465, 82101605, 82001218), the National Key R&D Program of China (grant number 2017YFC1310000), and Hubei Technological Innovation Special Fund (grant number 2019ACA132). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Abbreviations CYP2C19, cytochrome P450 2C19; CYP2D6, cytochrome P450 2D6; PSD, post-stroke depression; VEN, venlafaxine; ESC, escitalopram; BMI, Body mass index; CHD, Coronary heart disease; HAMD-17, 17-item Hamilton Depression Scale; HAMA, Hamilton Anxiety Scale; PHQ-9, 9-item Patient Health Questionnaire; PHQ-15, 15-item Patient Health Questionnaire; NIHSS, The National Institutes of Health Stroke Scale; BI, Barthel Index; mRS, Modified Rankin Scale; PMs, Poor Metabolizers; IMs, Intermediate Metabolizers; NMs, Normal Metabolizers; EMs, Extensive Metabolizers; RMs, Rapid Metabolizers; UMs, Ultra Metabolizers; CT, computed tomography; MRI, magnetic resonance imaging. Data Sharing Statement Data are available from the corresponding author (Xin Zhao, 2018tj5452@hust.edu.cn) upon reasonable request. Ethics Approval and Consent to Participate The approval of the study for experiments was obtained from the Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology (Approved No. of ethic committee: TJ-IRB20210142). All patients involved in this study or their family members gave written informed consents according to the Declaration of Helsinki. Author Contributions All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work. Disclosure The authors report no conflicts of interest in this work. References Associated Data Data Availability Statement Data are available from the corresponding author (Xin Zhao, 2018tj5452@hust.edu.cn) upon reasonable request.