Place · Level 3 · Condition
Depression & Anxiety · Nutritional Psychiatry
5-HT 单胺假说不够 · 肠脑轴 + 炎症 + 营养缺失 · SMILES 2017 RCT · 不替代心理、药物治疗
Last updated
Story path
Chapter 1
More than 'feeling sad'
More than 'feeling sad'
Major depressive disorder (MDD) + anxiety disorders (GAD / panic / social) are the most common psychiatric illnesses of our time and one of the most under-discussed health topics on this atlas.
Global epidemiology (WHO 2024 + the China Mental Health Survey — Huang 2019 *Lancet Psychiatry*, not the Chen 2017 the site used to name):
Lifetime prevalence of any depressive disorder: China 6.8% (of which MDD 3.4%) / global 5-15%Any anxiety disorder: China 7.6% / global 14-20%Comorbidity: ~ 50% overlap between MDD + anxietyUndiagnosed / untreated: 80%+ in ChinaPost-pandemic: global depression + anxiety up 25-30% in 2020-2022
Why this atlas must cover it:
1 in 10 people will experience a serious mood disorderNutrition / exercise / sleep all influence symptoms, but nutritional psychiatry remains seriously neglectedImportant warning: this island is not a substitute for treatment — moderate / severe patients need medication + psychotherapy
Diagnosis (DSM-5):
MDD: ≥ 5 symptoms, ≥ 2 weeks, including at least one core symptom:
Depressed mood (most of the time)Anhedonia (loss of interest in previously enjoyable activities)Plus: appetite change / sleep change / psychomotor retardation or agitation / fatigue / worthlessness or guilt / poor concentration / thoughts of death or suicide
GAD: excessive anxiety + worry, difficult to control, ≥ 6 months + ≥ 3 symptoms (restlessness / easy fatigue / poor concentration / irritability / muscle tension / insomnia)
Screening tools (used both clinically and self-administered):
PHQ-9 (Patient Health Questionnaire-9): 9 items, 0-27; ≥ 10 = depressionGAD-7: 7 items, 0-21; ≥ 10 = anxietyChinese versions are validated; can be self-completed in 15 minutes
Red flags (immediate care):
Suicidal ideation / planSelf-harm behaviourSevere inability to self-carePsychotic symptoms (hallucinations / delusions)Emergency: call the national mental-health hotline 400-161-9995 or local emergency services
This atlas island covers:
Why the monoamine hypothesis (5-HT/DA/NE) is insufficientThe gut-brain axis + inflammation + nutritional-deficiency modelEvidence from SMILES RCT and other nutritional-psychiatry trialsThe practical intervention spectrum (exercise / sleep / diet / sunlight / medication / therapy)
Global epidemiology (WHO 2024 + the China Mental Health Survey — Huang 2019 *Lancet Psychiatry*, not the Chen 2017 the site used to name):
Lifetime prevalence of any depressive disorder: China 6.8% (of which MDD 3.4%) / global 5-15%Any anxiety disorder: China 7.6% / global 14-20%Comorbidity: ~ 50% overlap between MDD + anxietyUndiagnosed / untreated: 80%+ in ChinaPost-pandemic: global depression + anxiety up 25-30% in 2020-2022
Why this atlas must cover it:
1 in 10 people will experience a serious mood disorderNutrition / exercise / sleep all influence symptoms, but nutritional psychiatry remains seriously neglectedImportant warning: this island is not a substitute for treatment — moderate / severe patients need medication + psychotherapy
Diagnosis (DSM-5):
MDD: ≥ 5 symptoms, ≥ 2 weeks, including at least one core symptom:
Depressed mood (most of the time)Anhedonia (loss of interest in previously enjoyable activities)Plus: appetite change / sleep change / psychomotor retardation or agitation / fatigue / worthlessness or guilt / poor concentration / thoughts of death or suicide
GAD: excessive anxiety + worry, difficult to control, ≥ 6 months + ≥ 3 symptoms (restlessness / easy fatigue / poor concentration / irritability / muscle tension / insomnia)
Screening tools (used both clinically and self-administered):
PHQ-9 (Patient Health Questionnaire-9): 9 items, 0-27; ≥ 10 = depressionGAD-7: 7 items, 0-21; ≥ 10 = anxietyChinese versions are validated; can be self-completed in 15 minutes
Red flags (immediate care):
Suicidal ideation / planSelf-harm behaviourSevere inability to self-carePsychotic symptoms (hallucinations / delusions)Emergency: call the national mental-health hotline 400-161-9995 or local emergency services
This atlas island covers:
Why the monoamine hypothesis (5-HT/DA/NE) is insufficientThe gut-brain axis + inflammation + nutritional-deficiency modelEvidence from SMILES RCT and other nutritional-psychiatry trialsThe practical intervention spectrum (exercise / sleep / diet / sunlight / medication / therapy)
数字 · 有多常见, 又有多少人没被治
它太常见, 又太容易被当成意志问题: 大约每 10 个人就有 1 个会经历一次严重的情绪障碍, 而国内绝大多数患者从没被诊断、也没被治疗。把它当性格缺陷, 只会让人更晚求助、更深自责。重度抑郁 (MDD) 和焦虑障碍 (GAD、惊恐、社交焦虑) 是当代最常见的精神疾病。从流行病学看 (WHO 2024 + 中国心理健康调查 —— Huang 2019 Lancet Psychiatry, 不是站内此前写的 Chen 2017): 任一抑郁障碍的终身患病率在中国约 6.8% (其中重性抑郁 MDD 是 3.4%), 全球 5-15%; 任一焦虑障碍中国约 7.6%、全球 14-20%; MDD 和焦虑约一半的患者两者重叠; 国内未诊断未治疗的比例超过 80%。疫情后 2020-2022 全球抑郁加焦虑又上升了 25-30%。
临床 · 医生怎么算数, 你怎么自评
临床上怎么算抑郁 (DSM-5)? MDD 需要 ≥ 5 个症状、持续 ≥ 2 周, 且至少包含一个核心症状——情绪低落 (大部分时间) 或快感缺失 (对过去喜欢的事不再感兴趣); 其余症状包括食欲改变、睡眠改变、精神运动迟缓或激越、疲倦、自责无用感、注意力难集中、以及关于死亡或自杀的想法。GAD 则是过度焦虑加担忧、难以控制、持续 ≥ 6 个月, 外加至少 3 个症状 (烦躁、易疲倦、注意力差、易怒、肌肉紧张、失眠)。自查和临床都常用两个量表: PHQ-9 (Patient Health Questionnaire-9) 共 9 题、0-27 分, ≥ 10 分提示抑郁; GAD-7 共 7 题、0-21 分, ≥ 10 分提示焦虑。两者中文版都验证可靠, 15 分钟内能自评完。
这一岛接下来会讲: 单胺假说 (5-HT/DA/NE) 为什么不够、肠脑轴加炎症加营养缺失的新模型、SMILES RCT 等营养精神病学证据, 以及完整的干预谱 (运动、睡眠、饮食、阳光、药物、心理)。
Chapter 2
4 neuro models
4 neuro models
Depression neuroscience has moved beyond the simplistic "low 5-HT" model; the modern view is four mutually complementary theories:
① Monoamine hypothesis (traditional):
Low synaptic serotonin (5-HT) / norepinephrine (NE) / dopamine (DA) → depressionThe main mechanism of SSRIs / SNRIs — block reuptake, ↑ synaptic monoamine availabilityLimitation: drug efficacy takes 2-4 weeks to appear, while monoamine concentrations rise within hours — direct monoamine replacement is not the fundamental mechanism
② Neuroplasticity + BDNF hypothesis:
BDNF (brain-derived neurotrophic factor) is significantly reduced in depression, especially in the hippocampusChronic stress → cortisol → ↓ hippocampal neurogenesis + dendritic atrophy → depressionSSRI's true action: indirectly ↑ BDNF + new neurons — explaining the 2-4 week onsetExercise + learning + sunlight + nutrient-dense diet all directly raise BDNF
③ Gut-brain axis:
Gut microbiota ↔ vagus + short-chain fatty acids + endotoxin + tryptophan metabolism → brainDepressive patients have significantly altered gut composition (Lactobacillus / Bifidobacterium ↓)"Psychobiotics" — the most concrete piece of evidence is one strain in one population: *Bifidobacterium longum* NCC3001 lowered depression scores over 6 weeks in 44 adults who had IBS plus mild-to-moderate anxiety/depression (Pinto-Sanchez 2017 Gastroenterology). That is a comorbid-IBS population, not a depression clinic, and the trial did not test L. helveticusDepression has an inflammatory basis: ↑ tumor necrosis factor alpha: A strong pro-inflammatory signal molecule that runs high in chronic inflammation. / interleukin-6: A pro-inflammatory signal molecule (cytokine) released by immune cells during inflammation. / C-reactive protein: A liver protein that rises with inflammation — a common blood marker for 'is the body inflamed'. → BBB → central inflammation → ↓ 5-HT synthesis + neuronal injuryThe mechanistic basis of nutritional intervention lives here
④ hypothalamic–pituitary–adrenal axis: The body's stress-response chain (hypothalamus → pituitary → adrenal) that releases cortisol.-axis dysregulation + chronic stress:
Hypothalamic-pituitary-adrenal axis: stress → CRH → ACTH → cortisolChronic stress → sustained high cortisol → hippocampal damage + feedback failure"Stress that can't be turned off": covered in atlas ashwagandha + endocrine/HPA L4
How the four models integrate:
```
stress → HPA activation → ↑ cortisol + inflammation
↓
gut dysbiosis + endotoxin
↓
blood–brain barrier: The 'security gate' on brain vessels that blocks most substances in blood from entering the brain. permeability + central inflammation
↓
↓ BDNF + ↓ neurogenesis
↓
↓ monoamine synthesis (5-HT requires tryptophan, but IDO diverts tryptophan to the kynurenine pathway)
↓
depression
```
This is why single interventions have limited effect but combined interventions stack:
SSRI repairs monoamines + indirectly BDNFExercise ↑ BDNF + ↓ inflammationNutrition repairs microbiota + ↓ inflammation + provides neurotransmitter precursorsPsychotherapy repairs cognition + behaviourSleep repairs BDNF + HPA
The "tryptophan → 5-HT" marketing trap:
"Eat more tryptophan (turkey / milk) and depression will improve"Truth: tryptophan must compete with other large neutral amino acids (LNAA) at the BBB; large protein meals actually reduce brain 5-HT synthesisThe old "high-carb meals → 5-HT" idea is over-simplified5-HTP supplements (intermediate of tryptophan): crosses BBB but bypasses feedback regulation → cardiovascular + serotonin-syndrome risk; depression evidence is weakTruly effective nutritional interventions are whole dietary patterns (next scene SMILES), not single amino acids
① Monoamine hypothesis (traditional):
Low synaptic serotonin (5-HT) / norepinephrine (NE) / dopamine (DA) → depressionThe main mechanism of SSRIs / SNRIs — block reuptake, ↑ synaptic monoamine availabilityLimitation: drug efficacy takes 2-4 weeks to appear, while monoamine concentrations rise within hours — direct monoamine replacement is not the fundamental mechanism
② Neuroplasticity + BDNF hypothesis:
BDNF (brain-derived neurotrophic factor) is significantly reduced in depression, especially in the hippocampusChronic stress → cortisol → ↓ hippocampal neurogenesis + dendritic atrophy → depressionSSRI's true action: indirectly ↑ BDNF + new neurons — explaining the 2-4 week onsetExercise + learning + sunlight + nutrient-dense diet all directly raise BDNF
③ Gut-brain axis:
Gut microbiota ↔ vagus + short-chain fatty acids + endotoxin + tryptophan metabolism → brainDepressive patients have significantly altered gut composition (Lactobacillus / Bifidobacterium ↓)"Psychobiotics" — the most concrete piece of evidence is one strain in one population: *Bifidobacterium longum* NCC3001 lowered depression scores over 6 weeks in 44 adults who had IBS plus mild-to-moderate anxiety/depression (Pinto-Sanchez 2017 Gastroenterology). That is a comorbid-IBS population, not a depression clinic, and the trial did not test L. helveticusDepression has an inflammatory basis: ↑ tumor necrosis factor alpha: A strong pro-inflammatory signal molecule that runs high in chronic inflammation. / interleukin-6: A pro-inflammatory signal molecule (cytokine) released by immune cells during inflammation. / C-reactive protein: A liver protein that rises with inflammation — a common blood marker for 'is the body inflamed'. → BBB → central inflammation → ↓ 5-HT synthesis + neuronal injuryThe mechanistic basis of nutritional intervention lives here
④ hypothalamic–pituitary–adrenal axis: The body's stress-response chain (hypothalamus → pituitary → adrenal) that releases cortisol.-axis dysregulation + chronic stress:
Hypothalamic-pituitary-adrenal axis: stress → CRH → ACTH → cortisolChronic stress → sustained high cortisol → hippocampal damage + feedback failure"Stress that can't be turned off": covered in atlas ashwagandha + endocrine/HPA L4
How the four models integrate:
```
stress → HPA activation → ↑ cortisol + inflammation
↓
gut dysbiosis + endotoxin
↓
blood–brain barrier: The 'security gate' on brain vessels that blocks most substances in blood from entering the brain. permeability + central inflammation
↓
↓ BDNF + ↓ neurogenesis
↓
↓ monoamine synthesis (5-HT requires tryptophan, but IDO diverts tryptophan to the kynurenine pathway)
↓
depression
```
This is why single interventions have limited effect but combined interventions stack:
SSRI repairs monoamines + indirectly BDNFExercise ↑ BDNF + ↓ inflammationNutrition repairs microbiota + ↓ inflammation + provides neurotransmitter precursorsPsychotherapy repairs cognition + behaviourSleep repairs BDNF + HPA
The "tryptophan → 5-HT" marketing trap:
"Eat more tryptophan (turkey / milk) and depression will improve"Truth: tryptophan must compete with other large neutral amino acids (LNAA) at the BBB; large protein meals actually reduce brain 5-HT synthesisThe old "high-carb meals → 5-HT" idea is over-simplified5-HTP supplements (intermediate of tryptophan): crosses BBB but bypasses feedback regulation → cardiovascular + serotonin-syndrome risk; depression evidence is weakTruly effective nutritional interventions are whole dietary patterns (next scene SMILES), not single amino acids
机制 · 模型 ①② · 递质与枝条
抑郁的神经科学早已不止5-HT 不够这一句, 现在的主流是 4 个互相补充的理论:① 单胺假说 (传统):
5-羟色胺 (5-HT) / 去甲肾上腺素 (NE) / 多巴胺 (DA) 突触水平低 → 抑郁SSRI / SNRI 主要机制 — 阻断再摄取, 突触可用单胺 ↑局限: 药效要 2-4 周才显现, 但单胺浓度上升几小时内就完成 — 说明直接补单胺并不是根本机制
② 神经可塑性 + BDNF 假说:
脑源性神经营养因子 (BDNF) 在抑郁时显著下降, 尤其在海马慢性应激 → 皮质醇 → 海马神经发生 ↓ + 树突萎缩 → 抑郁SSRI 的真正作用: 间接抬高 BDNF + 新生神经元, 这解释了为什么药需要 2-4 周起效运动、学习、阳光与富营养饮食都能直接抬高 BDNF
机制 · 模型 ③ · 肠子通向脑子的那条线
③ 肠脑轴 (Gut-Brain Axis):肠道菌 ↔ 迷走神经 + 短链脂肪酸 + 内毒素 + 色氨酸代谢 → 大脑抑郁患者菌群组成显著不同 (Lactobacillus / Bifidobacterium ↓)心理生物菌 (psychobiotics) — 证据最具体的一条是一个菌株、一个人群: Bifidobacterium longum NCC3001 在 44 名合并轻中度焦虑抑郁的肠易激 (IBS) 患者里, 6 周后抑郁分下降 (Pinto-Sanchez 2017 Gastroenterology)。请注意那是 IBS 共病人群, 不是抑郁门诊人群, 而且这项试验没有测 L. helveticus抑郁还带着炎症的底子: 炎症因子 (tumor necrosis factor alpha: A strong pro-inflammatory signal molecule that runs high in chronic inflammation. / interleukin-6: A pro-inflammatory signal molecule (cytokine) released by immune cells during inflammation. / C-reactive protein: A liver protein that rises with inflammation — a common blood marker for 'is the body inflamed'.) 一旦升高, 会穿过血脑屏障 (blood–brain barrier: The 'security gate' on brain vessels that blocks most substances in blood from entering the brain.) 引起中枢炎症, 使 5-HT 合成下降、神经受损营养干预的机制基础就落在这一条
机制 · 模型 ④ · 关不掉的压力轴
④ hypothalamic–pituitary–adrenal axis: The body's stress-response chain (hypothalamus → pituitary → adrenal) that releases cortisol. 轴失调 + 慢性应激:下丘脑-垂体-肾上腺轴 (HPA): 应激时下丘脑先放出促肾上腺皮质激素释放激素 (CRH), 它促使垂体分泌促肾上腺皮质激素 (ACTH), 再让肾上腺分泌皮质醇慢性应激下皮质醇长期偏高, 会损伤海马, 也让负反馈失灵关不掉 的应激: atlas ashwagandha + endocrine/HPA L4 详讲
机制 · 四条线怎么拧成一团
这 4 个模型如何整合: 把这四个模型串成一条链看, 就清楚多了。慢性应激先激活 hypothalamic–pituitary–adrenal axis: The body's stress-response chain (hypothalamus → pituitary → adrenal) that releases cortisol. 轴, 抬高皮质醇和炎症; 炎症再叠加肠道菌群失调, 让内毒素漏进血液、血脑屏障变得通透, 引起中枢炎症; 中枢炎症一边压低 BDNF 和新生神经元, 一边抢走造 5-HT 的原料——它会启动一个叫吲哚胺2,3-双加氧酶 (IDO) 的酶, 把色氨酸推去走犬尿氨酸通路, 而不去合成 5-HT; 几条线一起使劲, 最后就表现为抑郁。这就是为什么单一干预效果有限, 综合干预效果叠加:
SSRI 修单胺 + 间接 BDNF运动 ↑ BDNF + ↓ 炎症营养修肠菌 + ↓ 炎症 + 补充神经递质前体心理治疗修认知 + 行为睡眠修 BDNF + HPA
误区 · 多吃色氨酸就能补血清素?
色氨酸 → 5-HT 营销陷阱:常见说法是多吃色氨酸 (火鸡、牛奶) 抑郁就好真实情况: 色氨酸通过 blood–brain barrier: The 'security gate' on brain vessels that blocks most substances in blood from entering the brain. 时要与其它大中性氨基酸 (LNAA) 抢同一转运体, 大量蛋白餐反而会降低脑内 5-HT 合成高碳水餐促 5-HT 这条旧规则太简化了5-HTP 补剂 (色氨酸的中间产物): 越过 BBB 但绕过反馈调节 → 心血管 + 5-HT 综合征风险, 抑郁的证据也弱真正有效的营养干预是整体饮食模式 (下一幕 SMILES), 而不是单一氨基酸
Chapter 3
SMILES 2017 trial
SMILES 2017 trial
The SMILES Trial (Jacka et al 2017 BMC Medicine) is the first causal RCT in nutritional psychiatry:
Design:
N = 67 moderate-to-severe MDD patients (DSM-IV-TR diagnosis + MADRS ≥ 18, with a dietary screen confirming poor diet quality — SMILES never used the PHQ-9)Modified Mediterranean diet (ModiMedDiet) vs social support × 12 weeksKey: most participants were already on SSRI / psychotherapy — this is "diet as adjunct," not "diet as replacement"
ModiMedDiet content:
Whole grains / vegetables / fruit / legumes / low-fat dairy / nuts / fish / eggs / olive oil / moderate red meatRestrict: sweets / refined grains / fried foods / processed meat / soft drinks / alcohol= modified Mediterranean = low-UPF diet
Results:
MADRS (Montgomery-Åsberg Depression Rating Scale):Diet arm ↓ 11 pointsControl ↓ 4 points7-point difference, clinically significantRemission rate (MADRS < 10):Diet arm 32%Control 8%NNT = 4 (1 in 4 treated reaches remission — lower than many drugs)Anxiety scores also improvedSide effects: near zero
Replications:
HELFIMED 2019 (Parletta; 152 recruited, 95 completing at 3 months): Mediterranean diet plus fish oil vs social groups — depression fell further (p = 0.03) and held at 6 months. ⚠️ The fish oil is part of the intervention, so it is not a diet-only replicationAMMEND 2022 (Bayes, 72 completers): Australian men aged 18-25, 12 weeks of Mediterranean diet vs befriending, BDI-II between-group difference 14.4 points. ⚠️ It is an open-label single-site RCT, not a meta-analysis — and the year and author this page used to print were both wrong
Conclusions (nutritional-psychiatry 2024 consensus):
Mediterranean / DASH / low-UPF diets as adjunctive therapy: Level B-A evidenceCannot replace medication / psychotherapy — but can substantially synergisePrevention (healthy people): long-term low UPF + nutrient-dense diet shows an inverse dose-response with depression risk (atlas UPF story Lane 2023 BMJ)
Key nutrients (individual evidence):
① Omega-3 (EPA-dominant):
The higher the EPA dose, the larger the effect, and the benefit is clearer in people already taking an antidepressant (Mocking 2016 meta: 13 RCTs / 1233 people, overall SMD 0.398, p = 0.006). ⚠️ The paper tested the EPA:DHA ratio as a moderator and it was not significant — so a 60% EPA product rule does not come from this metaSolo effect small; adjunct to SSRI synergy is more pronounced
② Vitamin D:
Deficiency (< 30 ng/mL) associated with depression (Anglin 2013 meta)But RCT evidence for supplementation improving depression is inconsistent — possibly association without causation, or efficacy only in deficiency
③ Folate + B12:
Deficient individuals (especially older / vegan / MTHFR polymorphism) have higher depression riskL-methylfolate (Deplin) is an SSRI adjunct (Papakostas 2012 Am J Psychiatry)
④ Mg:
Small trials show Mg supplementation improves depression / anxiety (Tarleton 2017)Mechanism likely NMDA-receptor modulation + GABA + hypothalamic–pituitary–adrenal axis: The body's stress-response chain (hypothalamus → pituitary → adrenal) that releases cortisol.
⑤ Zn:
Deficiency associated with depression + supplementation improves it (Wang 2018 meta)
Warnings:
5-HTP / SAMe / St John's wort: with SSRI → serotonin syndrome (dangerous)L-tryptophan: blood–brain barrier: The 'security gate' on brain vessels that blocks most substances in blood from entering the brain. transport competition limits effect + ordinary dietary amounts are fine; high-dose supplement evidence is weakAny nutritional intervention + prescription medication → discuss with the psychiatrist, do not self-titrate
Design:
N = 67 moderate-to-severe MDD patients (DSM-IV-TR diagnosis + MADRS ≥ 18, with a dietary screen confirming poor diet quality — SMILES never used the PHQ-9)Modified Mediterranean diet (ModiMedDiet) vs social support × 12 weeksKey: most participants were already on SSRI / psychotherapy — this is "diet as adjunct," not "diet as replacement"
ModiMedDiet content:
Whole grains / vegetables / fruit / legumes / low-fat dairy / nuts / fish / eggs / olive oil / moderate red meatRestrict: sweets / refined grains / fried foods / processed meat / soft drinks / alcohol= modified Mediterranean = low-UPF diet
Results:
MADRS (Montgomery-Åsberg Depression Rating Scale):Diet arm ↓ 11 pointsControl ↓ 4 points7-point difference, clinically significantRemission rate (MADRS < 10):Diet arm 32%Control 8%NNT = 4 (1 in 4 treated reaches remission — lower than many drugs)Anxiety scores also improvedSide effects: near zero
Replications:
HELFIMED 2019 (Parletta; 152 recruited, 95 completing at 3 months): Mediterranean diet plus fish oil vs social groups — depression fell further (p = 0.03) and held at 6 months. ⚠️ The fish oil is part of the intervention, so it is not a diet-only replicationAMMEND 2022 (Bayes, 72 completers): Australian men aged 18-25, 12 weeks of Mediterranean diet vs befriending, BDI-II between-group difference 14.4 points. ⚠️ It is an open-label single-site RCT, not a meta-analysis — and the year and author this page used to print were both wrong
Conclusions (nutritional-psychiatry 2024 consensus):
Mediterranean / DASH / low-UPF diets as adjunctive therapy: Level B-A evidenceCannot replace medication / psychotherapy — but can substantially synergisePrevention (healthy people): long-term low UPF + nutrient-dense diet shows an inverse dose-response with depression risk (atlas UPF story Lane 2023 BMJ)
Key nutrients (individual evidence):
① Omega-3 (EPA-dominant):
The higher the EPA dose, the larger the effect, and the benefit is clearer in people already taking an antidepressant (Mocking 2016 meta: 13 RCTs / 1233 people, overall SMD 0.398, p = 0.006). ⚠️ The paper tested the EPA:DHA ratio as a moderator and it was not significant — so a 60% EPA product rule does not come from this metaSolo effect small; adjunct to SSRI synergy is more pronounced
② Vitamin D:
Deficiency (< 30 ng/mL) associated with depression (Anglin 2013 meta)But RCT evidence for supplementation improving depression is inconsistent — possibly association without causation, or efficacy only in deficiency
③ Folate + B12:
Deficient individuals (especially older / vegan / MTHFR polymorphism) have higher depression riskL-methylfolate (Deplin) is an SSRI adjunct (Papakostas 2012 Am J Psychiatry)
④ Mg:
Small trials show Mg supplementation improves depression / anxiety (Tarleton 2017)Mechanism likely NMDA-receptor modulation + GABA + hypothalamic–pituitary–adrenal axis: The body's stress-response chain (hypothalamus → pituitary → adrenal) that releases cortisol.
⑤ Zn:
Deficiency associated with depression + supplementation improves it (Wang 2018 meta)
Warnings:
5-HTP / SAMe / St John's wort: with SSRI → serotonin syndrome (dangerous)L-tryptophan: blood–brain barrier: The 'security gate' on brain vessels that blocks most substances in blood from entering the brain. transport competition limits effect + ordinary dietary amounts are fine; high-dose supplement evidence is weakAny nutritional intervention + prescription medication → discuss with the psychiatrist, do not self-titrate
试验 · SMILES 是怎么设计的
SMILES Trial (Jacka et al 2017 BMC Medicine) 是营养精神病学第一个 RCT 证因果:设计:
N = 67 中-重度 MDD 患者 (DSM-IV-TR 诊断 + MADRS ≥ 18, 且饮食筛查确认饮食质量差 —— SMILES 从未用过 PHQ-9)改良地中海饮食 (ModiMedDiet) vs 社交支持 × 12 周关键: 大部分受试者已经在用 SSRI / 心理治疗 — 这是饮食辅助 不是饮食替代
ModiMedDiet 内容:
全谷、蔬菜、水果、豆类、低脂奶、坚果、鱼 / 鸡蛋、橄榄油、适量红肉限制: 甜食、精制谷物、油炸、加工肉、软饮、酒精= 改良地中海饮食 = 低 UPF 饮食
数字 · 饮食组到底好了多少
结果:MADRS (Montgomery-Åsberg 抑郁量表):饮食组 ↓ 11 分对照组 ↓ 4 分差异 7 分, 临床显著缓解率 (MADRS < 10):饮食组 32%对照组 8%NNT = 4 (每 4 个干预 1 个缓解, 比许多药物 NNT 还低)焦虑分也改善副作用: 几乎零
证据 · 别人复制出来没有
复制研究:HELFIMED 2019 (Parletta, 招募 152 人, 3 个月 95 人完成): 地中海饮食加鱼油 vs 社交组, 抑郁分下降更多 (p = 0.03), 6 个月仍在。⚠️ 多了鱼油这一味, 所以不是纯饮食的复现AMMEND 2022 (Bayes, 72 人完成): 澳洲 18-25 岁男性, 12 周地中海饮食 vs 陪伴对照, BDI-II 组间差 14.4 分。⚠️ 它是一项开放标签的单中心 RCT, 不是 meta 分析 —— 站内此前写的年份与作者也都是错的
结论 (营养精神病学 2024 共识):
地中海 / DASH / 低 UPF 饮食 作为辅助治疗 B-A 级证据不能替代药物、心理治疗 — 但能显著协同预防 (健康人群): 长期低 UPF + 富营养饮食与抑郁风险呈反向剂量响应 (atlas UPF 故事 Lane 2023 BMJ)
营养素 · 鱼油、维生素 D、叶酸
关键营养素 (个别证据):① Omega-3 (EPA-dominant):
EPA 剂量越高效果越好, 而且已经在吃抗抑郁药的人获益更明显 (Mocking 2016 meta: 13 项 RCT / 1233 人, 总体 SMD 0.398, p = 0.006)。⚠️ 这篇把 EPA:DHA 比例当调节因素检验过, 结果不显著 —— 所以EPA 要占 60% 以上这条产品规则并不出自它单独使用效应小, 辅助 SSRI 时协同更明显
② 维生素 D:
缺乏 (< 30 ng/mL) 与抑郁关联 (Anglin 2013 meta)但补充改善抑郁的 RCT 证据不一致 — 可能只是关联而非因果, 或者仅对缺乏者有效
③ 叶酸 + B12:
缺乏者 (尤其老年、vegan 或 MTHFR 多态) 抑郁风险 ↑L-methylfolate (Deplin) 可作为 SSRI 辅助药 (Papakostas 2012 Am J Psychiatry)
营养素 · 镁与锌
④ Mg:部分小试验显示 Mg 补充改善抑郁、焦虑 (Tarleton 2017)机制可能是 NMDA 受体调节 + GABA + hypothalamic–pituitary–adrenal axis: The body's stress-response chain (hypothalamus → pituitary → adrenal) that releases cortisol.
⑤ Zn:
缺乏与抑郁相关, 补充可改善 (Wang 2018 meta)
安全 · 和处方药一起吃的红线
警告:5-HTP / SAMe / 圣约翰草与 SSRI 同用 → 5-HT 综合征 (危险), 建议避免L-tryptophan: blood–brain barrier: The 'security gate' on brain vessels that blocks most substances in blood from entering the brain. 转运竞争限制 + 不超过常规饮食量没问题, 大剂量补剂证据弱任何营养干预与处方药合用, 都建议先和精神科医生讨论, 不要自行加药
Chapter 4
Complete tx spectrum
Complete tx spectrum
Depression + anxiety treatment — the full spectrum (by evidence level + indication):
First-line: psychotherapy (preferred for mild-to-moderate):
CBT (cognitive behavioural therapy): Level A, 8-20 sessions, long-term effect not inferior to medicationIPT (interpersonal therapy): when relationships are centralMBCT (mindfulness-based cognitive therapy): Level A for relapse preventionMindfulness + ACT (acceptance and commitment): Level B-A for anxietyDigital CBT (iCBT): online platforms; not a complete substitute for in-person, but markedly better than no treatment
First-line: medication:
SSRI (fluoxetine / paroxetine / sertraline / citalopram / escitalopram): onset 2-4 weeks, Level ASNRI (venlafaxine / duloxetine): SSRI + dual NE channelMirtazapine: commonly used in older adults + poor appetite + insomniaBupropion: no sexual side effects + aids smoking cessationtricarboxylic acid (Krebs) cycle: The mitochondrial hub cycle that fully oxidizes fuel and harvests electrons for energy. (tricyclics): older drugs, effective but more side effects, rarely first-line nowMAOI: rarely used now
New drugs (2020s):
Esketamine (Spravato) nasal spray: treatment-resistant MDD, rapid onset (hours-days), FDA-approved 2019Brexanolone / Zuranolone: postpartum depression, GABA-A modulationPsychedelic research (psilocybin, MDMA): positive Phase III trials, regulatory review in some countries — important new direction but far from clinically widespread
That is the clinical menu of medication + psychotherapy. Turn the page for the three common SSRI misconceptions + four non-drug levers (exercise / light / sleep / social) + when to seek professional help.
First-line: psychotherapy (preferred for mild-to-moderate):
CBT (cognitive behavioural therapy): Level A, 8-20 sessions, long-term effect not inferior to medicationIPT (interpersonal therapy): when relationships are centralMBCT (mindfulness-based cognitive therapy): Level A for relapse preventionMindfulness + ACT (acceptance and commitment): Level B-A for anxietyDigital CBT (iCBT): online platforms; not a complete substitute for in-person, but markedly better than no treatment
First-line: medication:
SSRI (fluoxetine / paroxetine / sertraline / citalopram / escitalopram): onset 2-4 weeks, Level ASNRI (venlafaxine / duloxetine): SSRI + dual NE channelMirtazapine: commonly used in older adults + poor appetite + insomniaBupropion: no sexual side effects + aids smoking cessationtricarboxylic acid (Krebs) cycle: The mitochondrial hub cycle that fully oxidizes fuel and harvests electrons for energy. (tricyclics): older drugs, effective but more side effects, rarely first-line nowMAOI: rarely used now
New drugs (2020s):
Esketamine (Spravato) nasal spray: treatment-resistant MDD, rapid onset (hours-days), FDA-approved 2019Brexanolone / Zuranolone: postpartum depression, GABA-A modulationPsychedelic research (psilocybin, MDMA): positive Phase III trials, regulatory review in some countries — important new direction but far from clinically widespread
That is the clinical menu of medication + psychotherapy. Turn the page for the three common SSRI misconceptions + four non-drug levers (exercise / light / sleep / social) + when to seek professional help.
临床 · 心理治疗清单
抑郁 + 焦虑治疗 — 完整谱 (按证据等级 + 适应症):一线: 心理治疗 (轻-中度首选):
CBT (认知行为治疗): A 级证据, 8-20 次, 长期效果不输药物IPT (人际治疗): 关系问题为主MBCT (正念认知治疗): 复发预防 A 级正念 + 接纳承诺治疗 (ACT): 焦虑 B-A 级数字 CBT (iCBT): 网络平台, 不能完全替代真人但比无治疗显著
临床 · 药物清单 (含新药)
一线: 药物:SSRI (氟西汀、帕罗西汀、舍曲林、西酞普兰、艾司西酞普兰): 起效 2-4 周, A 级SNRI (文拉法辛、度洛西汀): 同 SSRI + NE 双通道米氮平 (Mirtazapine): 老年 + 食欲差 + 失眠常用安非他酮 (Bupropion): 不影响性功能 + 助戒烟三环类 (tricarboxylic acid (Krebs) cycle: The mitochondrial hub cycle that fully oxidizes fuel and harvests electrons for energy.): 老药, 有效但副作用多, 现在很少一线MAOI: 现在罕用
新药 (2020s):
Esketamine (Spravato) 鼻喷: 难治性 MDD, 起效快 (小时-天), FDA 2019 批Brexanolone / Zuranolone: 产后抑郁, GABA-A 调节致幻剂研究 (Psilocybin, MDMA): III 期试验阳性, 部分国家审批中 —— 重要新方向但远未临床普及
SSRI myths + non-drug treatments + when to seek help
Common SSRI misconceptions:"SSRIs change personality / are addictive": wrong — not addictive substances, but tapering must be gradual (weight / dizziness / "brain zaps" discontinuation syndrome)"SSRI sexual side effects": real, 30-50 % in some patients; alternatives exist (bupropion / mirtazapine)"SSRI is slow = useless": standard 2-4 week window, full evaluation at 6-8 weeks
Exercise — antidepressant evidence Level A:
150 min/week moderate aerobic → significant depression-score dropCooney 2013 Cochrane meta: exercise vs control SMD −0.62 (moderate); but restricted to the six methodologically strongest trials it fell to −0.18 and lost significance — that caveat comes from the review authors themselves. Head-to-head against medication there was no significant differenceHeissel 2023 *BJSM* (41 RCTs, 2264 people) found a larger effect (SMD −0.95), still −0.67 in the low-bias subsetMechanism: BDNF + endorphins + self-efficacy + sleep + socialThree elements: intensity + frequency + persistence; a 30-minute walk is more sustainable than occasional HIIT
Light therapy: first-line for seasonal affective disorder (SAD) — morning 10,000 lux light box for 30 minutes; non-seasonal MDD as adjunct (Level B).
Sleep: depression + insomnia have bidirectional causality (atlas `insomnia` + `shift-work-circadian`); CBT-I doesn't conflict with SSRI — combination effect is stronger.
Social support + relationships: isolation is one of the strongest modifiable risk factors for depression. Not "see friends more" — meaningful interpersonal connection.
When to see a professional:
PHQ-9 ≥ 10 or GAD-7 ≥ 10: see psychiatry / psychological counselling≥ 2 weeks of persistence + functional impairment: sameSuicidal ideation → immediate careIf self-adjustment isn't working long-term: don't blame yourself — seek professional help
Key cautions: depression + anxiety are not "lack of willpower" or "melodrama" — they are diseases with a neurobiological basis. "Just think positively" / "you have nothing to worry about" is misaligned comfort that worsens guilt. Stigma is one of the main reasons for low Chinese clinic-attendance rates — this island is written to reduce that stigma.
Chapter 5
Decision tree
Decision tree
"I might have depression / anxiety" — self-check + action:
Step 1 · Self-check (5 minutes):
Take PHQ-9 + GAD-7 online (national health platforms + some apps)Scoring:0-4: none / minimal — just maintenance5-9: mild — start with lifestyle + counselling10-14: moderate — strongly recommend professional help15+: severe — see a doctor immediately
Step 2 · Severity → corresponding pathway:
Mild (PHQ < 10, minor functional impact):
The 8 lifestyle levers (below)Digital CBT (e.g., "Simple Psychology" / "KnowYourself" / "HaoXinQing" Chinese platforms)Re-score PHQ-9 monthlyNo improvement → escalate
Moderate (PHQ 10-14):
Professional counselling / therapy (CBT preferred)+ consider medication (discuss with physician, not mandatory)+ lifestyle interventionRe-assess every 2-4 weeks
Severe (PHQ ≥ 15):
Must see psychiatry / psychosomatic medicineCombined medication (SSRI/SNRI) + psychotherapyAssess suicide risk (Columbia scale)Some require brief hospitalisation
Suicide risk:
Immediately call: National mental-health hotline 400-161-9995Beijing Suicide Research and Prevention Center: 010-82951332Shanghai hotline: 021-12320-5Local emergency (any time)
Step 1 · Self-check (5 minutes):
Take PHQ-9 + GAD-7 online (national health platforms + some apps)Scoring:0-4: none / minimal — just maintenance5-9: mild — start with lifestyle + counselling10-14: moderate — strongly recommend professional help15+: severe — see a doctor immediately
Step 2 · Severity → corresponding pathway:
Mild (PHQ < 10, minor functional impact):
The 8 lifestyle levers (below)Digital CBT (e.g., "Simple Psychology" / "KnowYourself" / "HaoXinQing" Chinese platforms)Re-score PHQ-9 monthlyNo improvement → escalate
Moderate (PHQ 10-14):
Professional counselling / therapy (CBT preferred)+ consider medication (discuss with physician, not mandatory)+ lifestyle interventionRe-assess every 2-4 weeks
Severe (PHQ ≥ 15):
Must see psychiatry / psychosomatic medicineCombined medication (SSRI/SNRI) + psychotherapyAssess suicide risk (Columbia scale)Some require brief hospitalisation
Suicide risk:
Immediately call: National mental-health hotline 400-161-9995Beijing Suicide Research and Prevention Center: 010-82951332Shanghai hotline: 021-12320-5Local emergency (any time)
路径 · 每一档具体怎么走
我可能有抑郁、焦虑 · 自查 + 行动:Step 1 · 自查 (5 分钟):
在线做这两张量表 (国内卫生信息平台 + 部分 APP 有)分数区间见前一页
Step 2 · 严重程度对应路径:
轻度 (PHQ < 10, 影响功能小):
生活方式 8 大杠杆 (下一页)数字 CBT (例如简单心理 / KnowYourself / 好心情 等中国平台)每月复评 PHQ-9不改善 → 升级
中度 (PHQ 10-14):
专业心理咨询、治疗 (CBT 优先)可以考虑加药物 (与医生讨论, 不强制)加生活方式干预每 2-4 周复评
重度 (PHQ ≥ 15):
应尽快看精神科、心身医学科通常会联合药物 (SSRI/SNRI) 与心理治疗评估自杀风险 (Columbia 量表)部分需要短期住院
Step 3-4 · lifestyle + nutrition
Step 3 · 8 lifestyle levers (ROI order):Exercise: 150 min/week moderate aerobic → significant depression-score drop (Level A, comparable to medication)
Sleep: 7-9 h + regular + CBT-I for insomnia
Nutrition: Mediterranean / DASH / low-UPF diet (SMILES Level B-A)
④ Sunlight + outdoors: morning light + natural environment + vitamin D
⑤ Social: meaningful interpersonal connection (isolation is the strongest risk)
⑥ Cut / reduce alcohol: alcohol relieves briefly + worsens depression long-term
⑦ Limit screens + social media: chronic SoMe → anxiety + depression signal
⑧ Mindfulness / meditation: apps (Headspace / Calm / 一念) + in-person classes
Step 4 · Nutritional support (adjunctive, not replacement):
Omega-3 (EPA-dominant): 1-2 g/day, choose EPA > 60% productsVitamin D: ≥ 800-2000 IU/day if deficientB12 + folate: monitor in long-term vegetarian / older / MTHFR-polymorphism individualsMg: 300-400 mg/day (glycinate or citrate)Zn: 8-15 mg/day in deficient (vegetarian + low red meat)Avoid: 5-HTP / St John's wort / SAMe + SSRI combination (serotonin syndrome)
Step 5 · maintenance + warnings + stance
Step 5 · Long-term maintenance:Continue applying CBT skills (even after symptom remission)Relapse prevention: MBCT + regular exercise + life rhythmSlow taper of medication (with the doctor, minimum 8-12 months)Recognise relapse signals early: sleep changes / appetite changes / loss of interest → intervene immediately
Important warnings:
"Influencer psychology advice": use cautiously, most lack professional credentials"Emotional management courses": some commercial courses overpromise, do not replace real treatment"Antidepressant supplements": most have weak evidence + commercial markup"TCM for depression": individual signals exist (Xiaoyaosan etc.) but insufficient evidence for first-line use
Atlas position (reiteration):
Depression + anxiety are diseases — not weak character / weak will / dramaStigma is the main reason for low Chinese clinic-attendance rates — this island is written so you feel less aloneLifestyle + nutrition are adjuncts, not replacements for medication / psychotherapyEarly intervention + complete treatment + long-term maintenance = most patients improve substantiallyFinding the right doctor / therapist may take 2-3 tries — do not give up after the first mismatchRecovery / remission ≠ perfection — it is "returning to a state in which you can live the life you want"
Stigma + finding therapy in China
"I have depression but I'm afraid to see a doctor" — China's mental-health clinic-attendance rate is only 9.5% vs 40-50% in the West. Stigma + not knowing where to go + concerns about records = the three main barriers.Debunking stigma:
"Seeing a psychiatrist = I'm crazy" → wrong. Depression is a neurochemical brain disorder, like diabetes"My employer will find out" → in China, public-hospital records are in principle not actively disclosed to employers"It will affect dating / promotion" → some careers (civil service / pilots) have restrictions on psychiatric history, most ordinary jobs have no explicit restrictions"Others tough it out, so I can too" → wrong. You can't see what others are quietly treating
Practical paths to finding counselling / therapy:
A. Public hospitals (insurance + prescribing + appropriate for severe cases):
General-hospital psychiatric / psychosomatic departments (most Tier-3 hospitals)Specialty centers: Anding (Beijing) / Shanghai Mental Health Center / Nanjing Brain HospitalPros: insurance + prescriptions; cons: short visit (10-15 min) + limited psychotherapy resources + waits
B. Private psychological counselling (CBT / talk therapy):
Platforms like Simple Psychology / 壹心理 / KnowYourselfNational Level-2 / Level-3 counsellor credentialsFees: ¥300-1500 per 50 minutesFrequency: weekly × 12-20 weeksPros: ample time + genuine psychotherapy + privacy; cons: out-of-pocket + variable quality + no prescriptions
C. Online video counselling + digital CBT (apps, starter):
Same platforms above with video + textApps: HaoXinQing / XinQing / Woebot, Headspace, CalmAdvantages: cheap + anytime + anonymous; limits: moderate-or-above needs in-person
Criteria for choosing a counsellor / physician:
1. Credentials: national Level-2 / Level-3 counsellor, or licensed psychiatrist
2. Modality: CBT (most evidence) / psychodynamic / EMDR / mentalisation
3. Feeling in the first session: can you relax enough to speak in 10 minutes? If not, switch
4. Does not push intensive packages / does not sell courses / does not insist on private WeChat contact
5. No felt improvement after 2-3 sessions → switch (it's not you, it's the fit)
When to escalate to psychiatry + medication:
PHQ-9 ≥ 15 (severe)Suicidal ideation / self-harm behaviourSevere impact on work / studies / lifeNo significant improvement with counselling after 3 monthsBipolar (history of mania / hypomania) — must see psychiatry + mood stabilisers
What family + friends can do:
Listen ≠ solve — don't rush to give advice"You need to pull yourself together" / "Stop overthinking" = worsens stigmaRecommend seeking care + accompany to appointments + reduce questioningLearn the basics of depression and anxietyTake care of yourself too (caregiver burnout is real)
Atlas connections: insomnia (insomnia is a sentinel + comorbidity of depression) + perimenopause (E2 withdrawal + depression risk) + menstrual-cycle (PMDD) + sleep-apnea + adaptogens (Rhodiola).
References · 12
- World Health Organization. (2024). Mental health: strengthening our response. WHO Fact sheet. www.who.int/news-room/fact-sheets/detail/mental-health-strengthening-our-response
- Huang, Y., Wang, Y., Wang, H., Liu, Z., Yu, X., Yan, J., et al. (2019). Prevalence of mental disorders in China: a cross-sectional epidemiological study. The Lancet Psychiatry, 6(3), 211-224. 10.1016/S2215-0366(18)30511-X
- Malhi, G. S., & Mann, J. J. (2018). Depression. The Lancet, 392(10161), 2299-2312. 10.1016/S0140-6736(18)31948-2
- Cryan, J. F., O'Riordan, K. J., Cowan, C. S. M., Sandhu, K. V., Bastiaanssen, T. F. S., Boehme, M., et al. (2019). The microbiota-gut-brain axis. Physiological Reviews, 99(4), 1877-2013. 10.1152/physrev.00018.2018
- Pinto-Sanchez, M. I., Hall, G. B., Ghajar, K., Nardelli, A., Bolino, C., Lau, J. T., et al. (2017). Probiotic Bifidobacterium longum NCC3001 reduces depression scores and alters brain activity: a pilot study in patients with irritable bowel syndrome. Gastroenterology, 153(2), 448-459.e8. ⚠️ SCOPE: one strain, one population — 44 adults with IBS (diarrhoea or mixed) AND mild-to-moderate anxiety and/or depression, 6 weeks, double-blind. It is not evidence in a depression clinic population, and it did not test L. helveticus. 10.1053/j.gastro.2017.05.003
- Jacka, F. N., O'Neil, A., Opie, R., Itsiopoulos, C., Cotton, S., Mohebbi, M., et al. (2017). A randomised controlled trial of dietary improvement for adults with major depression (the SMILES trial). BMC Medicine, 15, 23. 10.1186/s12916-017-0791-y
- Mocking, R. J. T., Harmsen, I., Assies, J., Koeter, M. W. J., Ruhé, H. G., & Schene, A. H. (2016). Meta-analysis and meta-regression of omega-3 PUFA supplementation for major depressive disorder. Translational Psychiatry, 6(3), e756. 10.1038/tp.2016.29
- Bayes, J., Schloss, J., Sibbritt, D. (2022). The effect of a Mediterranean diet on the symptoms of depression in young males (the “AMMEND: A Mediterranean Diet in MEN with Depression” study): a randomized controlled trial. The American Journal of Clinical Nutrition, 116(2), 572-580. AMMEND: a 12-week open-label single-site RCT in 72 Australian men aged 18-25 with moderate-severe depression, befriending control. It is a single trial, not a pooled review of other trials; its year is 2022 and its first author is Bayes. The story used to list it as a 2024 review by a different author. 10.1093/ajcn/nqac106
- Parletta, N., Zarnowiecki, D., Cho, J., Wilson, A., Bogomolova, S., Villani, A., et al. (2017). A Mediterranean-style dietary intervention supplemented with fish oil improves diet quality and mental health in people with depression: A randomized controlled trial (HELFIMED). Nutritional Neuroscience, 22(7), 474-487. HELFIMED: Mediterranean diet PLUS fish oil versus social groups — the fish oil is part of the intervention, so it is not a clean replication of a diet-only trial. 152 recruited, 95 completed at 3 months; depression fell more in the diet arm (p = 0.03) and held at 6 months. 10.1080/1028415X.2017.1411320
- Cooney, G. M., Dwan, K., Greig, C. A., Lawlor, D. A., Rimer, J., Waugh, F. R., et al. (2013). Exercise for depression. Cochrane Database of Systematic Reviews, (9), CD004366. 10.1002/14651858.CD004366.pub6
- Heissel, A., Heinen, D., Brokmeier, L. L., Skarabis, N., Kangas, M., Vancampfort, D., et al. (2023). Exercise as medicine for depressive symptoms? A systematic review and meta-analysis. British Journal of Sports Medicine, 57(16), 1049-1057. 10.1136/bjsports-2022-106282
- Cipriani, A., Furukawa, T. A., Salanti, G., Chaimani, A., Atkinson, L. Z., 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. The Lancet, 391(10128), 1357-1366. ⚠️ ITS EFFECT METRIC IS THE ODDS RATIO FOR RESPONSE — it reports neither Cohen's d nor an NNT. The site used to compare a single trial's d, and a single trial's NNT, against numbers attributed to this paper that are not in it. Do not put it on the same ruler as a small single-site trial. 10.1016/S0140-6736(17)32802-7