Effectiveness of herbal medicine as an add-on to antipsychotics in patients with schizophrenia spectrum disorders accompanied by depression: A systematic review and meta-analysis

Abstract

Background

Depression is a common comorbidity of schizophrenia spectrum disorder (SSDs) that affects functional outcomes and quality of life. This systematic review and meta-analysis evaluated the effectiveness of herbal medicine as an adjunct therapy to antipsychotics in patients with SSDs and comorbid depression.

Methods

Eight databases were searched from inception to January 2025 for randomized controlled trials (RCTs) evaluating herbal medicine combined with antipsychotics vs antipsychotics alone in patients with SSDs and comorbid depression. The primary outcome measure was a depression-specific assessment instrument. Risk of bias was assessed using the Cochrane RoB 2 tool, and evidence certainty was assessed using GRADE.

Results

Overall, of 12 RCTs, 884 participants were included. Compared to antipsychotics, combination therapy significantly improved depressive symptoms (standardized mean difference [SMD] –0.91, 95 % confidence interval (CI) –1.05 to -0.76, p<0.00001, N=10, n=768, moderate certainty), enhanced efficacy for schizophrenia symptoms (SMD –0.60, 95 % CI –0.89 to -0.31, p=0.0014, N=9, n=666, moderate certainty), and showed higher response rates (relative risk [RR] 1.26, 95 % CI 1.04 to 1.52, p=0.0277, N=5, n=372, low certainty). Adverse event rates showed no significant difference between groups (RR 1.18, 95 % CI 0.69 to 2.01, p=0.1609, N=2, n=136, low certainty).

Conclusions

This review provides evidence that herbal medicine, as an adjunct to antipsychotics, may improve both depressive and psychotic symptoms in patients with SSDs and comorbid depression. Although promising, methodological limitations and the exclusive Chinese origin of the studies indicate the need for more rigorous and diverse trials to establish definitive clinical recommendations.

Protocol registration

PROSPERO, CRD42025643148.

1. Introduction

Schizophrenia spectrum disorders (SSDs) encompass a group of psychiatric conditions characterized by psychotic symptoms, including schizophrenia, delusional disorder, brief psychotic disorder, schizophreniform disorder, and schizoaffective disorder.¹ These disorders affects approximately 1 % of the global population,² causing significant disability and imposing substantial burdens on patients, families, and healthcare systems.³ Depression is a common comorbid condition in individuals with SSDs, with prevalence rates ranging from 25 % to 60 %.^4,5 This comorbidity is associated with poorer overall outcomes, including impaired quality of life, increased risk of suicide, reduced treatment adherence, and higher rates of relapse.5, 6, 7

Post-schizophrenic depression, specifically defined in the International Classification of Diseases (ICD-10) as a depressive episode arising in the aftermath of a schizophrenic illness where some schizophrenic symptoms may still be present, presents particular challenges in clinical management.⁸ The complex interrelationship between the negative symptoms of schizophrenia and depressive symptoms further complicates diagnosis and treatment.⁹

Current treatment approaches typically involve augmentation strategies, including the addition of antidepressants (e.g., selective serotonin reuptake inhibitors) to antipsychotics.⁴ Meta-analyses have showed only modest benefits of antidepressants in treating depression in schizophrenia, with response rates often remaining suboptimal and significant tolerability concerns reported.¹⁰ Additionally, polypharmacy strategies increase the risk of drug interactions, metabolic complications, and medication non-adherence.^11,12 These limitations of conventional pharmacotherapy highlight the urgent need for alternative or complementary therapeutic approaches.

Herbal medicine, particularly those from East Asian traditional medicines, has been used for centuries to treat various psychiatric conditions.¹³ Several herbal formulations have showed potential antidepressant and antipsychotic properties through mechanisms including modulation of monoaminergic neurotransmission, anti-inflammatory effects, hypothalamic-pituitary-adrenal (HPA) axis regulation, and neuroprotection.14, 15, 16 Preliminary research suggests that certain herbal medicines might effectively complement antipsychotics by addressing depressive symptoms while potentially mitigating some adverse effects of conventional treatments.^15,17

Despite the growing interest in integrative treatment approaches, evidence regarding herbal medicine as an adjunctive therapy for depression in SSDs has not been systematically evaluated. Previous systematic reviews have examined herbal medicines for schizophrenia or depression separately,^16,18 but none have specifically addressed their roles in treating the comorbidity of these conditions.

Therefore, this systematic review and meta-analysis aimed to evaluate the effectiveness of herbal medicine as an add-on to antipsychotics in improving depressive and overall psychiatric symptoms in patients with SSDs accompanied by depression. This synthesis of evidence will inform the clinical practice and future research on integrative treatment approaches for these challenging comorbidities.

2. Methods

The protocol for this systematic review was registered in PROSPERO (registration number: CRD42025643148). This systematic review complied with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 Statement (Supplement 1).¹⁹

2.1. Criteria for inclusion and exclusion

2.1.1. Study types

Only randomized controlled trials (RCTs) were included in this review. Quasi-RCTs wherein the allocation sequences were either non-random or predictable were excluded. No restrictions were applied regarding the publication status or language.

2.1.2. Participant types

We included studies involving patients of any age diagnosed with SSDs who had comorbid depression. SSDs include schizophrenia, delusional disorder, brief psychotic disorder, schizophreniform disorder, and schizoaffective disorder, as defined by established diagnostic criteria, such as the Diagnostic and Statistical Manual of Mental Disorders (DSM), ICD, or Chinese Classification of Mental Disorders (CCMD). For the comorbid depression component, we included studies that explicitly described participants as having comorbid depression with SSDs, regardless of whether standardized depression rating scales were used. Studies that focused on post-schizophrenic depression were also included. Studies were excluded if they involved patients with substance/medication-induced psychotic disorders or were primarily focused on treating the side effects of antipsychotic medication rather than comorbid depression.

2.1.3. Intervention types and controls

The intervention of interest was oral herbal medicine based on East Asian traditional medicine principles used in combination with antipsychotic medication. Herbal medicines include decoctions, granules, tablets, capsules, and oral formulations.

The eligible control groups were administered either antipsychotic medication alone or antipsychotic medication combined with a placebo. Concurrent treatments, such as general nursing care and rehabilitation training, were permitted only when administered identically in both the intervention and control groups. Studies were excluded if they incorporated additional treatments such as acupuncture, moxibustion, or cupping in either the intervention or control groups, as these would confound the assessment of herbal medicine effectiveness.

2.1.4. Outcomes measures

The primary outcome measures were validated instruments specifically designed to assess depression, including the Calgary Depression Scale for Schizophrenia (CDSS)²⁰ and Hamilton Depression Rating Scale (HDRS).²¹ Secondary outcomes included positive and negative symptom scores, social function scores, quality of life scores, relapse rates, adherence to antipsychotic medication, good response rates, and adverse event rates. Good response for depression was defined as ≥ 50 % reduction in CDSS or HDRS²² scores, and good response for schizophrenia was defined as ≥ 50 % reduction in PANSS²³ or BPRS²⁴ scores. The overall adverse event rate was calculated as the number of participants who experienced adverse events divided by the total number of participants. Excluding relapse rate, for all other outcomes, data were extracted at treatment completion. The relapse rate was assessed at the end of the observation period as specified in each study.

2.2. Literature searches

We conducted a comprehensive search across eight electronic databases: MEDLINE via PubMed, EMBASE, Cochrane Central Register of Controlled Trials, China National Knowledge Infrastructure, Citation Information by NII, Korean Studies Information Service System, ScienceON, and Oriental Medicine Advanced Searching Integrated System. We also searched the World Health Organization International Clinical Trials Registry Platform to identify relevant ongoing or completed trials. The search was conducted on January 15, 2025, with no restrictions on language or publication date. The complete search strategy for this systematic review is provided in Supplement 2. We also manually searched the reference lists of the included studies and relevant review articles to identify additional eligible studies. Gray literature was explored by searching conference proceedings, dissertations, and unpublished reports.

2.3. Data selection

Two reviewers independently screened the titles and abstracts of all retrieved citations against the predefined eligibility criteria. The full texts of potentially relevant studies were then obtained and independently assessed by the same two reviewers. Any disagreements at either stage were resolved through discussion and a third reviewer was consulted when necessary.

2.4. Data extraction

For each included study, two reviewers independently extracted data using a standardized pre-piloted form. The extracted information encompassed: (1) Study characteristics: first author, publication year, country, study design, sample size, and trial duration; (2) Participant characteristics: diagnostic criteria, inclusion criteria based on symptom scores and illness duration, pattern identification (if reported), demographic data (sex, age), duration of illness, and treatment setting; (3) Intervention details: herbal medicine name, dosage form, main and additional herb components, manufacturer, frequency and treatment duration; (4) Psychotropic medication details: name, dose, frequency, and treatment duration; (5) Outcome measures: primary and secondary outcomes; (6) Results: mean values, standard deviations, response rates, and other relevant statistics; and (7) Information on funding and potential conflicts of interest. Data inconsistencies or disagreements between the reviewers were resolved through discussion, and a third reviewer was consulted when necessary. When the reporting was unclear or incomplete, we attempted to contact the corresponding authors for clarification.

2.5. Risk of bias assessment of included studies

The risk of bias was assessed using the Cochrane Risk of Bias 2 (RoB 2) tool,²⁵ which evaluates five domains: randomization process, deviations from intended interventions, missing outcome data, outcome measurement, and selection of reported results. Based on these domains, each study was classified as having a low risk, some concerns, or high risk of bias. Two reviewers independently assessed the risk of bias using Excel for RoB version 2. Disagreements were resolved through discussion and a third reviewer was consulted when necessary.

2.6. Data analysis

Considering the anticipated clinical heterogeneity among the included studies on herbal medicines, we used a random effects meta-analysis model. For continuous outcomes, the mean differences were calculated when studies used the same measurement scale or standardized mean differences (SMD) with 95 % confidence intervals (CI) across different scales. For dichotomous outcomes, we computed the relative risk (RR) with a 95 % CI.

Heterogeneity among studies was assessed using the I² statistic, with values of 25 %, 50 %, and 75 % considered as low, moderate, and high heterogeneity, respectively.²⁶ Publication bias was assessed by visual inspection of funnel plots if ≥10 studies were included for a particular outcome. All analyses were performed using RevMan version 8.20 (The Cochrane Collaboration, London, UK).

2.7. Grading of recommendations, assessment, development, and evaluations (GRADE) assessment

The certainty of the evidence for each outcome was evaluated using the GRADE approach.²⁷ We assessed five domains that may reduce the evidence certainty: risk of bias, inconsistency, indirectness, imprecision, and publication bias. For each outcome, the evidence was categorized as “high,” “moderate,” “low,” or “very low certainty.” Evidence from the RCTs started with high certainty and was downgraded by one level for serious concerns and two levels for very serious concerns in any domain. The final GRADE assessments and justifications for downgrading are summarized in a table.

3. Results

3.1. Study selection

Overall, 959 records were identified after removing duplicates. Of these, 906 were excluded during the initial screening process, and full-text articles were sought for the remaining 53. Among them, two documents could not be retrieved in full text, leaving 51 articles for detailed evaluation. After full-text assessment, four studies were excluded for not being RCTs, seven studies were excluded because of the absence or unclear description of diagnostic criteria, and 28 studies were excluded because they did not evaluate the add-on effects of herbal medicine on antipsychotics (Supplement 3). Consequently, 12 RCTs 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39 were included in this systematic review (Fig.1).

Fig. 1.

PRISMA Flow Diagram.

3.2. Study characteristics

All included studies were conducted in China. The diagnostic criteria used included CCMD-3 in eight studies28, 29, 30^,32,34,36, 37, 38 and ICD-10 in four studies^31,33,35,39. The sample sizes ranged from 56 to 92 participants per study, with 884 participants across all studies (443 in treatment groups and 441 in control groups). Most studies included both male and female participants, with one study³⁶ exclusively focusing on female patients. The mean age of the participants ranged from 31.7 to 45.7 years, showing that most studies targeted adult populations.

The control group received antipsychotic medications, including risperidone, olanzapine, haloperidol, ziprasidone, or amisulpride, either as monotherapy or in combination. The treatment duration ranged from 6 to 8 weeks. The most commonly used outcome measures were the Hamilton Depression Rating Scale (HDRS; n = 7)^28,29,32, 33, 34^,37,38 and CDSS (n = 4)^30,31,34,35 for assessing depressive symptoms and the Positive and Negative Syndrome Scale (PANSS; n = 5)^29,33, 34, 35, 36 and Brief Psychiatric Rating Scale (BPRS; n = 5)^{28,32,34,37,38} to evaluate schizophrenia symptoms. Seven studies29, 30, 31, 32, 33^,36,39 reported adverse events, with the treatment groups generally showing fewer or milder side effects than the control group. Commonly reported treatment-emergent adverse events include headaches, insomnia, anxiety, extrapyramidal symptoms, sedation, and gastrointestinal disturbances (Table 1, Supplement 4).

Table 1.

Characteristics of included studies on herbal medicine for schizophrenia with depression.

Study	Diagnostic Criteria & Inclusion Criteria	Sample Size (M/F) & Mean Age	Interventions	Comparisons	Duration of intervention (follow-up)	Outcomes	Effect estimate of each outcome, MD or RR [95 % CI]
Sun, 201328	Schizophrenia (CCMD-3), acute, with depression; HDRS ≥ 12	TG: 18/22, 38.6 yrs CG: 16/24, 36.9 yrs	Risperidone (2-6 mg/d) + Anshen Buxin Wan	Risperidone (2-6 mg/d)	8 wks (Not done)	1) Depression (HDRS) 2) Schizophrenia (BPRS) 3) Schizophrenia response (BPRS ≥50 % reduction)	1) -2.40 [-3.46; -1.34] 2) -4.60 [-8.88; -0.32] 3) 1.36 [0.98; 1.90]
Li et al., 201429	Post-schizophrenic depression (CCMD-3); HDRS ≥ 18	TG: 22/16, Range: 22-51 yrs CG: 25/15, Range: 25-55 yrs	Continued antipsychotics + Jieyu Anshen Granule	Continued antipsychotics	6 wks (Not done)	1) Depression (HDRS) 2) Depression response (HDRS ≥50 % reduction) 3) Schizophrenia (PANSS) 4) Negative symptoms (PANSS-N) 5) Positive symptoms (PANSS-P)	1) -1.39 [-2.60; -0.18] 2) 1.38 [0.86; 2.22] 3) -1.10 [ -2.13; -0.07] 4) -1.50 [-2.68; -0.32] 5) -0.80 [-2.00; 0.40]
Zhang, 201430	Schizophrenia (CCMD-3) with depression; PANSS ≥ 60; CDSS ≥ 6	TG: 25/16, 32.1 yrs CG: 26/15, 32.7 yrs	Risperidone + Shugan Jieyu Capsule	Risperidone	8 wks (Not done)	1) Depression (CDSS) 2) Depression response (CDSS ≥50 % reduction) 3) Schizophrenia response (PANSS ≥50 % reduction)	1) -4.55 [-6.38; -2.72] 2) 1.40 [1.06; 1.84] 3) 1.00 [0.78; 1.28]
Yang, 201531	Schizophrenia (ICD-10), with depression; CDSS ≥ 6	TG: 26/14, 37.4 4.7 yrs CG: 25/15, 36.2 yrs	Risperidone (4-8 mg/d) + Xuefu Zhuyu Decoction	Risperidone (4-8 mg/d)	8 wks (Not done)	1) Depression (CDSS) 2) Depression response (HDRS ≥50 % reduction) 3) Adverse event rate	1) -3.70 [-5.11; -2.29] 2) 1.28 [1.03; 1.57] 3) 1.25 [0.36; 4.32]
Liu and Ma 201633	Post-schizophrenic depression (ICD-10)	TG: 20/16, 43.2 yrs CG: 21/15, 42.6 yrs	Amisulpride (200-600 mg/d) + Shugan Jieyu Capsule	Amisulpride (200-600 mg/d)	6 wks (Not done)	1) Depression (HDRS) 2) Depression response (HDRS ≥50 % reduction) 3) Schizophrenia (PANSS)	1) -4.40 [-6.61; -2.19] 2) 1.04 [0.80; 1.34] 3) 0.30 [-3.21; 3.81]
Xu and Jin 201632	Schizophrenia (CCMD-3), acute, with depression; HDRS ≥ 12	TG: 13/17, 39.1 yrs CG: 12/18, 38.4 yrs	Risperidone (2-6 mg/d), Olanzapine (5-20 mg/d), Haloperidol (4-20 mg/d) + Yangxue Qingnao Granule; CG:	Risperidone (2-6 mg/d), Olanzapine (5-20 mg/d), Haloperidol (4-20 mg/d)	8 wks (Not done)	1) Depression (HDRS) 2) Schizophrenia (BPRS) 3) Schizophrenia response (BPRS ≥50 % reduction)	1) -2.20 [-3.85; -0.55] 2) -4.90 [-9.84; 0.04] 3) 1.3750 [0.9233; 2.0477]
Sun, 201734	Schizophrenia (CCMD), acute, with depression; PANSS ≥ 60; CDSS ≥ 6; HDRS ≥ 20; BPRS ≥ 35	Total: 32/42, 40.4 yrs	Risperidone (2-6 mg/d), Olanzapine (5-20 mg/d), Haloperidol (4-20 mg/d) + Shugan Jieyu Decoction	Risperidone (2-6 mg/d), Olanzapine (5-20 mg/d), Haloperidol (4-20 mg/d)	8 wks (Not done)	1) Depression (CDSS, HDRS) 2) Schizophrenia (PANSS, BPRS)	1) -1.60 [-2.24; -0.96] (CDSS), -3.70 [-5.08; -2.32] (HDRS) 2) -7.40 [-13.64; -1.16] (PANSS) -8.20 [-12.37; -4.03] (BPRS)
Xu and Han 201735	Schizophrenia (ICD-10), with depression; PANSS ≥ 60	TG: 22/10, 32.5 yrs CG: 19/9, 33.5 yrs	Ziprasidone 160 mg/d + Shugan Jieyu Capsule	Ziprasidone 160 mg/d	6 wks (Not done)	1) Depression (CDSS) 2) Schizophrenia (PANSS) 3) Negative symptoms (PANSS-N) 4) Positive symptoms (PANSS-P)	1) -3.85 [-5.80; -1.90] 2) -11.03 [-17.82; -4.24] 3) -3.01 [-4.93; -1.09] 4) -0.72 [-2.34; 0.90]
Lin and Chen 201836	Post-schizophrenic depression (CCMD-3), female; HDRS ≥ 18, PANSS ≥ 60	TG: 0/30, 37.5 yrs CG: 0/30, 36.5 yrs	Olanzapine 10 mg/d + Shugan Jieyu Capsule	Olanzapine 10 mg/d	6 wks (Not done)	1) Depression response (HDRS ≥50 % reduction) 2) Schizophrenia (PANSS) 3) Negative symptoms (PANSS-N) 4) Positive symptoms (PANSS-P)	1) 1.86 [0.86; 4.00] 2) -15.81 [-21.22; -10.40] 3) -6.68 [-9.08; -4.28] 4) -1.32 [-3.30; 0.66]
Qi et al., 201837	Schizophrenia (CCMD-3), acute, with depression; Duration ≥ 1 year	TG: 24/21, 44.2 yrs CG: 27/18, 45.7 yrs	Risperidone (2 mg/d), Olanzapine (5 mg/d), Haloperidol (4 mg/d)+ Yangxue Qingnao Granule; CG:	Risperidone (2 mg/d), Olanzapine (5 mg/d), Haloperidol (4 mg/d)	8 wks (Not done)	1) Depression (HDRS) 2) Schizophrenia (BPRS)	1) -1.80 [-2.67; -0.93] 2) -5.40 [-8.66; -2.14]
Li, 201938	Schizophrenia (CCMD-3), acute, with depression; Duration ≥ 1 year	TG: 36/10, 44.2 yrs CG: 35/11, 44.2 yrs	Risperidone (2 mg/d), Olanzapine (5 mg/d), Haloperidol (4 mg/d)+ Yangxue Qingnao Granule	Risperidone (2 mg/d), Olanzapine (5 mg/d), Haloperidol (4 mg/d)	8 wks (Not done)	1) Depression (HDRS) 2) Schizophrenia (BPRS) 3) Schizophrenia response (BPRS ≥50 % reduction)	1) -1.83 [-2.68; -0.98] 2) -5.37 [-8.57; -2.17] 3) 1.27 [0.93; 1.73]
Chen et al, 202139	Schizophrenia (ICD-10), with depression; PANSS ≥ 60, CDSS ≥ 6	TG: 19/9, 31.7 yrs; CG: 18/10, 32.1 yrs	Risperidone + Shugan Jieyu Capsule	Risperidone	8 wks (Not done)	1) Adverse event rate 2) Schizophrenia response (PANSS ≥50 % reduction) 3) Quality of life (GQOLI-74)	1) 1.14 [0.48; 2.72] 2) 1.44 [1.00; 2.07] 3) 36.89 [27.02; 46.76]

BPRS, Brief Psychiatric Rating Scale; CCMD, Chinese Classification of Mental Disorders; CDSS, Calgary Depression Scale for Schizophrenia; CG, Control Group; GQOLI-74, Generic Quality of Life Inventory-74; HDRS, Hamilton Depression Rating Scale; ICD, International Classification of Diseases; MD, Mean Difference; NR, Not Reported; PANSS, Positive and Negative Syndrome Scale; RR, Risk Ratio; TG, Treatment Group.

3.3. Details of herbal medicine used

Regarding intervention characteristics, five distinct herbal medicines were utilized as adjunctive therapy to conventional antipsychotic treatment: Shugan Jieyu Capsule (n = 5),^{30,33,35,36,39} Yangxue Qingnao Granule (n = 3),^32,37,38 Shugan Jieyu Decoction (n = 1),³⁴ Anshen Buxin Wan (n = 1),²⁸ Jieyu Anshen Granule (n = 1),²⁹ and Xuefu Zhuyu Decoction (n = 1).³¹ The composition of these herbal preparations varied, with some studies providing detailed information while others did not report the specific composition. Among the reported formulations, several herbs appeared frequently across different preparations. Bupleuri Radix is a common component found in Jieyu Anshen Granules, Xuefu Zhuyu Decoction, Shugan Jieyu Decoction, and Angelicae Gigantis Radix. Other commonly used herbs include Paeoniae Radix Alba, Ligustici Rhizoma, Poria, and Glycyrrhizae Radix. The Yangxue Qingnao Granule preparations shared similar compositions across studies, containing Paeoniae Radix Alba, Angelicae Gigantis Radix, Rehmanniae Radix Preparat, Ligustici Rhizoma, Spatholobi Caulis, Margaritifera Usta Concha, Uncariae Ramulus et Uncus, Cassiae Semen, Prunellae Spica, Asari Herba Cum Radix, and Corydalis (Tuber) Rhizoma. None of the included studies reported any conditional modifications to the herbal formulations based on individual patient symptoms (Table 2).

Table 2.

Characteristics of Chinese herbal medicine interventions.

Author, Year	Preparation Name	Dosage Form	Main Composition (daily dose)	Administration
Sun, 201328	Anshen Buxin Wan	Pills	Not reported (Manufacturer: Not reported)	15 pills/day
Li et al., 201429	Jieyu Anshen Granule	Granules	Bupleuri Radix (Bupleurum falcatum L.), Curcumae Radix (Curcuma longa L.), Gardeniae Fructus (Gardenia jasminoides J.Ellis) (parched), Arisaema Praeparatus cum Bile (Arisaema erubescens (Wall.) Schott), Poria (Poria cocos (Schw.) Wolf), Acori Graminei Rhizoma (Acorus gramineus Aiton), Polygalae Radix (Polygala tenuifolia Willd.) (processed), Lilii Bulbus (Lilium lancifolium Thunb.), Zizyphi Spinosae Semen (Ziziphus jujuba Mill. var. spinosa (Bunge) Hu ex H.F.Chow) (parched), Glycyrrhizae Radix (Glycyrrhiza uralensis Fisch. ex DC.) etc.	5g, bid
Zhang, 201430	Shugan Jieyu Capsule	Capsule	Not reported (Manufacturer: Not reported)	2 capsules (0.72g), bid
Yang, 201531	Xuefu Zhuyu Decoction	Decoction	Persicae Semen (Prunus persica (L.) Batsch) 20g, Carthami Flos (Carthamus tinctorius L.) 15g, Bupleuri Radix (Bupleurum falcatum L.) 15g, Cyathulae Radix (Cyathula officinalis K.C.Kuan) 15g, Angelicae Gigantis Radix (Angelica gigas Nakai) 12g, Polygalae Radix (Polygala tenuifolia Willd.) 12g, Ligustici Rhizoma (Ligusticum striatum DC.) 12g, Paeoniae Radix Rubra (Paeonia lactiflora Pall.) 10g, Platycodi Radix (Platycodon grandiflorus (Jacq.) A.DC.) 10g, Hoelen cum Pini Radix (Poria cocos (Schw.) Wolf) 10g, Glycyrrhizae Radix (Glycyrrhiza uralensis Fisch. ex DC.) (roasted) 6g	bid
Liu and Ma 201633	Shugan Jieyu Capsule	Capsule	Not reported (Manufacturer: Not reported)	1.44g/day
Xu and Jin 201632	Yangxue Qingnao Granule	Granules	Paeoniae Radix Alba (Paeonia lactiflora Pall.), Angelicae Gigantis Radix (Angelica gigas Nakai), Rehmanniae Radix Preparat (Rehmannia glutinosa (Gaertn.) DC.), Ligustici Rhizoma (Ligusticum striatum DC.), Spatholobi Caulis (Spatholobus suberectus Dunn), Margaritifera Usta Concha (Pteria martensii (Dunker) or Hyriopsis cumingii (Lea)), Uncariae Ramulus et Uncus (Uncaria rhynchophylla (Miq.) Miq. ex Havil.), Cassiae Semen (Cassia obtusifolia L.), Prunellae Spica (Prunella vulgaris L.), Asari Herba Cum Radix (Asarum sieboldii Miq.), Corydalis (Tuber) Rhizoma (Corydalis turtschaninovii Besser) (Manufacturer: Tianjin Tasly Pharmaceutical Co., Ltd.)	1 packet (4g), tid
Sun, 201734	Shugan Jieyu Decoction	Decoction	Albizziae Cortex (Albizia julibrissin Durazz.) 60g, Gardeniae Fructus (Gardenia jasminoides J.Ellis) 30g, Bupleuri Radix (Bupleurum falcatum L.) 20g, Curcumae Radix (Curcuma longa L.) 20g, Salviae Miltiorrhizae Radix (Salvia miltiorrhiza Bunge) 20g, Poria (Poria cocos (Schw.) Wolf) 20g, Paeoniae Radix Alba (Paeonia lactiflora Pall.) 15g, Atractylodis Rhizoma Alba (Atractylodes macrocephala Koidz.) 15g, Angelicae Gigantis Radix (Angelica gigas Nakai) 15g, Aucklandiae Radix (Aucklandia costus Falc.) 15g, Zingiberis Rhizoma Recens (Zingiber officinale Roscoe) (slow-baked) 12g, Ligustici Rhizoma (Ligusticum striatum DC.) 10g, Menthae Herba (Mentha arvensis L.) 10g, Glycyrrhizae Radix (Glycyrrhiza uralensis Fisch. ex DC.) 10g	tid
Xu and Han 201735	Shugan Jieyu Capsule	Capsule	Not reported (Manufacturer: Chengdu Kanghong Pharmaceutical)	2 capsules (0.72g), bid
Lin and Chen 201836	Shugan Jieyu Capsule	Capsule	Not reported (Manufacturer: Not reported)	0.72g, bid
Qi et al., 201837	Yangxue Qingnao Granule	Granules	Paeoniae Radix Alba (Paeonia lactiflora Pall.), Angelicae Gigantis Radix (Angelica gigas Nakai), Rehmanniae Radix Preparat (Rehmannia glutinosa (Gaertn.) DC.), Ligustici Rhizoma (Ligusticum striatum DC.), Spatholobi Caulis (Spatholobus suberectus Dunn), Margaritifera Usta Concha (Pteria martensii (Dunker)), Uncariae Ramulus et Uncus (Uncaria rhynchophylla (Miq.) Miq. ex Havil.), Cassiae Semen (Cassia obtusifolia L.), Prunellae Spica (Prunella vulgaris L.), Asari Herba Cum Radix (Asarum sieboldii Miq.), Corydalis (Tuber) Rhizoma (Corydalis turtschaninovii Besser) (Manufacturer: Not reported)	4g, tid
Li, 201938	Yangxue Qingnao Granule	Granules	Paeoniae Radix Alba (Paeonia lactiflora Pall.), Angelicae Gigantis Radix (Angelica gigas Nakai), Spatholobi Caulis (Spatholobus suberectus Dunn), Ligustici Rhizoma (Ligusticum striatum DC.), Margaritifera Usta Concha (Pteria martensii (Dunker)), Uncariae Ramulus et Uncus (Uncaria rhynchophylla (Miq.) Miq. ex Havil.), Cassiae Semen (Cassia obtusifolia L.), Rehmanniae Radix Preparat (Rehmannia glutinosa (Gaertn.) DC.), Prunellae Spica (Prunella vulgaris L.), Corydalis (Tuber) Rhizoma (Corydalis turtschaninovii Besser), Asari Herba Cum Radix (Asarum sieboldii Miq.) (Manufacturer: Not reported)	1 packet, tid
Chen et al., 202139	Shugan Jieyu Capsule	Capsule	Not reported (Manufacturer: Not reported)	2 capsules (0.72g), bid

All studies reported no conditional modifications to the herbal formulations.

Abbreviations: bid, twice daily; tid, three times daily.

3.4. Risk of bias of included studies

Assessment using the RoB 2 tool revealed a consistent pattern of risk of bias across the included studies. Only the study by Xu et al. (2017) ³⁵ showed a low risk of bias in the randomization process and outcome measurement domains, whereas the remaining studies exhibited concerns regarding the randomization process. All studies consistently showed a low risk of bias in the domains related to deviations from intended interventions and missing outcome data. However, bias in outcome measurement was predominantly rated as “some concerns” across all studies, except for that by Xu et al. (2017) ³⁵. Bias in selection of reported results was rated as “some concerns” in all studies as no pre-specified analysis plan was available. Consequently, the overall bias assessment was predominantly classified as “some concerns” for most studies, as one or more domains were rated as some concerns, with no domains rated as high-risk (Fig.2, Supplement 4).

Fig. 2.

Risk of bias assessment for trials examining depressive symptoms (CDSS, HDRS).

3.5. Intervention effects

3.5.1. Primary outcome

The primary outcome measured by CDSS and HDRS showed significant improvement in depressive symptoms in patients receiving combination therapy (herbal medicine plus antipsychotics) compared to that of controls, contributing to the overall depression symptom improvement with an SMD of –0.91, 95 % CI –1.05 to -0.76, p<0.00001, N=10, n=768, moderate certainty (Fig.3A).

Fig. 3.

Forest plots comparing herbal medicine plus antipsychotics versus antipsychotics alone. (A) Depressive symptoms (CDSS, HDRS), (B) Depression response rate (≥50 % reduction in CDSS, HDRS), (C) Schizophrenia symptoms (PANSS, BPRS), and (D) Adverse events. Abbreviations: BPRS, Brief Psychiatric Rating Scale; CDSS, Calgary Depression Scale for Schizophrenia; HDRS, Hamilton Depression Rating Scale. PANSS, Positive and Negative Syndrome Scale.

3.5.2. Secondary outcomes

For secondary outcomes, the response rate, defined as ≥50 % improvement from baseline on CDSS or HDRS scores, was significantly higher in the combination therapy group (RR 1.28, 95 % CI 1.09 to 1.49, p=0.0277, N=5, n=372, low certainty) (Fig.3B). Schizophrenia symptoms, measured by PANSS total or BPRS total scores, also showed greater improvement in the combination therapy group (SMD –0.60, 95 % CI –0.89 to -0.31, p=0.0014, N=9, n=666, moderate certainty) (Fig.3C). Notably, the overall adverse event incidence showed no significant difference between groups (RR 1.18, 95 % CI 0.69 to 2.01, p=0.1609, N=2, n=136, low certainty), suggesting that the addition of herbal medicine did not increase side effects (Fig.3D). Meta-analysis results for other outcome variables are presented in the supplementary figures (Supplement 4).

3.6. Subgroup analyses

Although subgroup analyses were initially planned, they could not be conducted owing to the limited number of relevant studies.

3.7. Publication bias

The meta-analysis of depression symptom scores assessed using the CDSS and HDRS included 10 RCTs, allowing for publication bias evaluation through visual inspection of funnel plots. No asymmetry was observed in the funnel plots, suggesting a low risk of publication bias (Supplement 4). This finding was supported by Egger's test, which showed no significant evidence of publication bias (intercept = -2.093, 95 % CI: -9.88 to 5.69, t = -0.527, p = 0.613).

3.8. GRADE assessment

Evidence certainty of outcomes was assessed using the GRADE approach. Among the multiple outcomes assessed in this review, four key outcomes were selected for GRADE evidence certainty assessment based on their clinical importance. For depressive symptoms, our critical outcome, the evidence was rated as moderate-certainty, primarily because most studies had concerns regarding the overall risk of bias. Similarly, evidence for schizophrenia symptoms was assessed with moderate-certainty, because most studies had “some concerns” about the overall risk of bias. Heterogeneity was observed between the studies, but most showed the same direction of effect and evidence was not downgraded owing to inconsistency. For the response rate (defined as ≥50 % improvement in CDSS or HDRS), evidence certainty was rated as low, due to risk of bias concerns and imprecision (as the total number of events for this binary outcome was <300). The adverse event rate also received a low-certainty evidence rating, which was downgraded for risk of bias concerns because the CI for the effect estimate included 1, indicating imprecision (Table 3).

Table 3.

Summary of findings.

Outcome	Importance	Total patients (Studies)	Certainty of the evidence (GRADE)	Relative effect (95 % CI)	Anticipated absolute effects (95 % CI)
					Control group	Intervention group
Depressive symptoms (CDSS or HDRS)	Critical	768 (10 RCTs)	$\oplus$$\oplus$$\oplus$$◯$Moderatea	-	-	SMD 0.91 lower [-1.05, -0.76]
Response rate (number of participants with a 50 % or greater improvement in CDSS or HDRS)	Important	372 (5 RCTs)	$\oplus$$\oplus$$◯$$◯$ Lowa,c	RR 1.26 [1.04, 1.52]	556 per 1,000	701 per 1,000 [578, 845]
Schizophrenia symptoms (PANSS or BPRS)	Important	666 (9 RCTs)	$\oplus$$\oplus$$\oplus$$◯$ Moderated	-	-	SMD 0.60 lower [-0.89, -0.31]
Adverse event rate	Important	136 (2 RCTs)	$\oplus$$\oplus$$◯$$◯$ Lowa,b	RR 1.18 [0.69, 2.01]	162 per 1,000	191 per 1,000 [111, 325]

a, Most studies had 'some concerns' regarding overall risk of bias; b, confidence interval for binary outcome effect estimate includes 1; c, total number of events for binary outcome is less than 300; d, heterogeneity exists between studies, but most studies show the same direction of effect. Abbreviations: BPRS, Brief Psychiatric Rating Scale; CI, Confidence Interval; CDSS, Calgary Depression Scale for Schizophrenia; HDRS, Hamilton Depression Rating Scale; PANSS, Positive and Negative Syndrome Scale; RCT, Randomized Controlled Trial; RR, Risk Ratio; SMD, Standardized Mean Difference.

4. Discussion

4.1. Summary of main findings

This meta-analysis of 12 RCTs involving 884 participants showed that herbal medicine as an adjunctive therapy to antipsychotics significantly improved both depressive and psychotic symptoms in patients with SSDs and comorbid depression. The combination therapy showed substantial effect sizes for depressive symptoms (SMD -0.91, moderate certainty), enhanced effectiveness for schizophrenia symptoms (SMD -0.60, moderate certainty), and higher response rates (RR 1.26, low certainty) without significantly increasing adverse events (RR 1.18, low certainty). The GRADE assessment revealed moderate-certainty evidence for primary outcomes, primarily limited by methodological concerns rather than result inconsistency. Future well-designed trials may alter the magnitude of the estimated effects, though the direction is likely to remain consistent given the uniform favorable results across all included studies and absence of publication bias.

4.2. Comparison with previous studies

To the best of our knowledge, this is the first systematic review to specifically examine the role of herbal medicines in treating depression comorbid with SSDs. Previous systematic reviews have separately investigated herbal medicines for either schizophrenia or depression.^15,18 Deng and Xu (2017) found that Wendan decoction combined with antipsychotics may have beneficial effects in schizophrenia,¹⁵ while Wang et al. (2019) reported that herbal medicine has promise as a depression treatment.¹⁸ Our findings advance this knowledge by addressing the specific therapeutic challenges of comorbid conditions. The relatively large effect sizes observed are consistent with previous studies on integrative approaches to psychiatric disorders, suggesting that multi-target herbal interventions may complement the receptor-targeting mechanisms of conventional antipsychotics.

4.3. Clinical and scientific implications

The mechanisms underlying the therapeutic effects of herbal medicines in SSDs with comorbid depression likely involves multi-target actions,⁴⁰ which complemented the receptor-targeting mechanisms of conventional antipsychotics. This multitarget approach may be particularly advantageous for complex comorbid conditions involving multiple neurobiological pathways. Among the commonly used herbs, preclinical studies have shown that Bupleuri Radix, Angelicae Gigantis Radix, and Paeoniae Radix Alba affect monoaminergic neurotransmission, neuroinflammation, and HPA axis regulation,41, 42, 43, 44 which may contribute to their antidepressant properties. Specifically, saikosaponins from Bupleuri Radix modulate stress responses through regulation of corticotropin-releasing factor and HPA axis activity,^41,42 whereas nodakenin, decursin, and demethylsuberosin from Angelicae Gigantis Radix exhibit neuroprotective effects by inhibiting NF-κB and MAPK activity and inducing Nrf-2 activity.⁴³ Additionally, paeoniflorin from Paeoniae Radix Alba influences serotonin and norepinephrine metabolism, contributing to antidepressant effects.⁴⁴

Among the identified herbal interventions, the Shugan Jieyu Capsule/Decoction was the most frequently studied (n = 6), followed by the Yangxue Qingnao granules (n = 3). From a traditional East Asian medicine perspective, many of these formulations follow the therapeutic principle of “soothing the liver and relieving depression,” which aligns with the traditional understanding that stagnation of liver qi contributes to depressive symptoms.⁴⁵ Herbs such as Bupleuri Radix and Angelicae Gigantis Radix are commonly employed to address this pattern.^41,44,46 For Yangxue Qingnao Granule, the constituent herbs appear to focus on improving cerebral blood flow, reducing oxidative stress, and regulating neurotransmitter balance.⁴⁷ The alkaloids found in components such as Uncariae Ramulus et Uncus may influence both dopamine and serotonin signaling pathways,^48,49 potentially addressing both schizophrenia and depressive symptoms. This dual action may be particularly valuable in addressing the complex neurobiological underpinnings of comorbidities.

From a clinical perspective, these findings suggest that herbal medicines could serve as reasonable adjunctive options for patients with SSDs and comorbid depression, particularly when conventional antidepressant augmentation proves insufficient or causes intolerable side effects. However, clinicians should remain aware of the current evidence limitations when making treatment decisions.

4.4. Strengths and limitations

The strengths of this review include its comprehensive search strategy across multiple databases, strict inclusion criteria focusing specifically on the comorbid condition, and robust methodology using established assessment tools (RoB 2, GRADE).Although our findings are promising, several methodological considerations must be made. First, all included studies were conducted in China, potentially limiting data generalizability to other populations and healthcare contexts. Cultural factors influencing treatment expectations as well as genetic differences in drug metabolism may contribute to regional variations in treatment effects. Second, the risk-of-bias assessment revealed consistent concerns across most studies, particularly regarding the randomization process, outcome measurement, and selective reporting. Although no study was rated as having a high risk of bias in any domain, the predominance of “some concern” ratings reduced confidence in the precision of the reported effect sizes. Third, the relatively short treatment durations (6–8 weeks) in the included studies precluded conclusions regarding long-term effectiveness and safety. Depression in schizophrenia often requires extended treatment,⁵⁰ and the sustainability of the benefits of herbal adjunctive therapy remains uncertain. Fourth, the lack of standardized depression assessment tools for participant inclusion criteria in some studies^37,38 may have contributed to heterogeneity in our results, as the identification and severity of comorbid depression at baseline may have varied across studies. Finally, the lack of detailed reports on herbal medicine standardization in several studies raises questions regarding preparation consistency and active component concentrations. This limitation is common in herbal medicine research, and highlights the need for improved reporting standards in future trials.

4.5. Conclusions

This systematic review provide evidence that herbal medicines as an adjunct to antipsychotics may effectively improve both depressive and psychotic symptoms in patients with SSDs and comorbid depression. While evidence certainty ranges from low to moderate, primarily because of the methodological limitations in the included studies, the consistent direction of effects across studies suggests considerable promise for this therapeutic approach. Future research requires well-designed multicenter trials with larger sample sizes, rigorous methodologies, longer follow-up periods, mechanistic studies investigating neurobiological effects, comparative effectiveness studies examining different formulations, and incorporation of patient-centered outcomes to establish more definitive clinical recommendations for this challenging clinical presentation.

CRediT authorship contribution statement

Chan-Young Kwon: Conceptualization, Writing – original draft. Kyoung-Eun Lee: Investigation, Writing – review & editing. Min-Jae Kim: Investigation, Writing – review & editing. Ji-Won Kim: Investigation, Writing – review & editing. Ji-Won Oh: Investigation, Writing – review & editing. Hye-Li Jeon: Investigation, Writing – review & editing. Boram Lee: Conceptualization, Writing – review & editing. Pyung-Wha Kim: Conceptualization, Writing – review & editing. Yujin Choi: Conceptualization, Investigation, Writing – review & editing.