跳至正文

抑郁是化学失衡吗

目录

机制裁决红队风第九十四篇 · 对称双向红队第八十九篇 · section D 意识 / 心智 / 神经 · 全库第 152 篇

最高规格红线(置于篇首,不置于篇末):本篇是一篇证据审计,不是医疗建议。它不评估任何具体个人的用药,不构成开始、调整、继续或停止任何药物的依据,尤其不能被读成劝人停药——本篇第八节引用的随机对照证据显示,停药组一年内复发率为 56%、维持组为 39%(风险比 2.06),且撤药反应真实存在并可能持续数月。任何减停都应与处方者商定,并按现行指南阶梯减量。若你正在为情绪困扰求助,请联系当地医疗服务。

核验图例:✓ 主笔亲核(本人取回官方原文/官方全文/官方页面并逐字比对)· ◐ 一手可及但仅官方摘要级(未读全文,主张不越出可见部分)· ⚠ 二手转述或单侧来源(已标明立场)· ○ 仅题名可核(不作承重引用)。

去重与分界声明:本篇只裁单胺类抗抑郁药与中枢血清素假说这一条线,上游与邻位一律不重做。安慰剂反应本身的机制与它随年代上升的现象,归安慰剂/反安慰剂篇;氯胺酮/艾氯胺酮/裸盖菇素/迷走神经刺激这类新机制干预,归迷幻药与「临界重置」篇;「90% 血清素在肠道」「精神益生菌治抑郁」这类外周血清素叙事,归肠-脑轴篇;效应量、p 值与发表偏倚的方法论本体,归复制危机篇;「一个名字装两样东西」的元病理,归同名不同物篇;量表作为代理指标被当成实体的病,归测量代理性篇;期刊与评议制度本身,归同行评议与引用指标篇abc 猜想与 IUT 篇。全库查重:STAR*DTurner 2008撤药戒断化学失衡 此前零命中血清素 的既有覆盖全部落在上述三篇邻居里、且均为外周或新机制侧。

一条主动的防联想:本篇不裁「精神病学是不是医学」这类身份问题,也与反精神医学运动的任何政治主张无关。机制叙事不成立,不等于疾病不真实——这是本篇反复挡住的那一跳。

结构胎记=名号跳 × 机制跳 × 停药跳(三跳升格链):名号跳——把「单胺系统与情绪有关」这条实验室命题,换成「抑郁症是大脑化学物质失衡」这句从来不是该领域承重命题的话,然后让后者代表前者去面对公众;机制跳——把「药作用于血清素并且有效」倒推成「病因是血清素不足」,即 ex juvantibus 反推(阿司匹林能治头痛,不证明头痛源于脑内阿司匹林不足);停药跳——双向的,旧指南把撤药反应写成「1 至 2 周自限」(低估到不安全),反向叙事把它说成「抗抑郁药会上瘾」(越界到另一侧不安全)。


〇 母裁决 · 八层硬度光谱

命题 硬度 判据
抑郁作为临床现象与其后果是真实的 ✓ 硬 全球约 3 亿人受影响,美国国内相关成本估计逾 2105 亿美元/年(Stone 2022, BMJ 引言逐字)
SSRI 的急性药理是真的(抑制再摄取、升高突触间隙血清素) ✓ 硬(争议双方共同接受) 批评方与反驳方在同一组论文里都以此为前提,无一方质疑
抗抑郁药在对照试验里优于安慰剂 ✓ 硬 Cipriani 2018, Lancet:522 项试验、116 477 人,21 种药全部优于安慰剂,OR 1.37–2.13
量级收敛在约 2 个 HAMD-17 点 ✓ 硬(但四路数据高度重叠,见 2.5) 1.75(Stone 2022)/1.97(Munkholm 2019)/−1.94(Jakobsen 2017
这个量级的临床意义 ⚠ 未结案 3 点阈值之争;Hieronymus 2016 条目层反转;Stone 三峰分解在 2025 年复现失败
「血清素缺乏是抑郁病因」 × 六域皆不支持 Moncrieff 2022, Mol Psychiatry 六方向;连最强正面研究自己的引言都写「no firm in vivo evidence」
「化学失衡」是学界的承重命题 × 不是(独占核心) 药厂渠道争议双方一致承认;实测传播渠道是诊室(Schroder 2024,n=1 219)
撤药反应真实存在 ✓ 硬(量级已定) Henssler 2024, Lancet Psychiatry:79 研究、21 002 人,经安慰剂校正约 15%,重度 2.8% 对 0.6%

母裁决:抗抑郁药比安慰剂有效是真的,而且这件事的量级已经被四路分析收敛到约两个 HAMD 点;「血清素缺乏导致抑郁」这条病因命题,在六个研究方向上没有一个支持它;而「化学失衡」这句话本身,主要不是学界的承重命题,是一件传播产物——但把它真正传给病人的渠道,不是电视广告,是诊室。因此「抑郁是化学失衡」未立,「抗抑郁药是骗局、应该停药」同样未立。

四件事必须分开数。

第一,药效不是零。 这是本篇最需要先钉住的一条,因为它最容易在「机制叙事塌了」的气氛里被顺手扔掉。迄今规模最大的网络元分析覆盖 522 项双盲随机试验、116 477 名受试者,21 种抗抑郁药无一例外优于安慰剂 [一手逐字]。连立场最批判的那一侧也没有主张零效应:丹麦考科蓝团队的系统综述得到 −1.94 HAMD 点并写下「statistically significant effects」,他们的异议在于这个量级够不够称临床显著,以及危害是否抵消了它。

第二,量级小,而且四路算出来几乎是同一个数。 牛津团队的 OR、诺迪克考科蓝对同一批数据的再分析(1.97 点)、丹麦团队的独立综述(−1.94 点)、美国 FDA 四十年个体数据(1.75 点,95% CI 1.63–1.86)——四条路径落在 1.75 至 1.97 这个极窄的带里。这个数字本身已经不是争点了;争的是它意味着什么。 而且必须诚实说明:这四路远不是四次独立测量,它们的底层试验高度重叠(见 2.5,本篇主动降级)。

第三,病因层是硬否,不是「还没研究清楚」。 2022 年那份伞状综述把「支持血清素假说」的研究整编成六个方向逐一称重:体液中的血清素与其代谢物 5-HIAA、5-HT1A 受体、血清素转运体 SERT、色氨酸耗竭实验、SERT 基因关联、SERT 基因与压力的交互——结论逐字是「there is no convincing evidence that depression is associated with, or caused by, lower serotonin concentrations or activity」[一手逐字]。其中基因侧是高确定性的无效:11.5 万人的关联研究给出 OR = 1.000、p = 0.994。

第四,也是本篇的独占核心:那句话的出处,双方其实已经达成了一半共识。 为精神科辩护得最激烈的那位——《精神时报》前主编 Ronald Pies——在 2011 年写下「the ‘chemical imbalance’ notion was always a kind of urban legend – never a theory seriously propounded by well-informed psychiatrists」的同一篇文章里,紧接着写道「And, yes – the ‘chemical imbalance’ image has been vigorously promoted by some pharmaceutical companies, often to the detriment of our patients’ understanding」[一手逐字]。也就是说:药厂渠道,双方一致承认。争的只剩「专业界是否也在传」。而 2024 与 2025 年两项针对信念来源的实测给出了一个谁都没预料的答案——最常见的接触渠道是课堂,但唯一能独立预测「真的相信」的渠道,是医疗提供者本人。

灵魂句:「化学失衡」这四个字能活四十年,不是因为它是最好的科学解释,而是因为它是最好用的沟通工具——它一句话同时办成了三件极难的事:免除病人的自责、让服药显得理所当然、把一段无法解释的痛苦装进一个可以说出口的名字。等到证据表明那个名字里没有东西,需要被换掉的从来不是药,而是那三件事各自的真办法;而在真办法被造出来之前,谁先把这句话从病人手里拿走,谁就要负责补上它原本顶着的三根梁。


一 真锚:三件立得住的事

1.1 现象与后果

不必从争议开始。BMJ 那篇分析 FDA 四十年数据的论文,引言第一句就把地基摆在那里 [一手逐字]:

“Depression is a leading cause of disability worldwide, affecting 300 million people globally, causing a major reduction in quality of life, with domestic costs (including costs related to work) estimated at more than $210.5 (£175.3; €207.1) billion annually.”

同一段还给了处方规模:「About 13% of Americans use antidepressants, and use of antidepressants in economically developed countries more than doubled between 2000 and 2015.」

来源:Stone et al., “Response to acute monotherapy for major depressive disorder in randomized, placebo controlled trials submitted to the US Food and Drug Administration: individual participant data analysis”, BMJ 378:e067606 (2022), doi:10.1136/bmj-2021-067606(✓ 主笔亲核,PMC 全文

这一层没有任何一方争。 后面所有的分歧,都发生在「这件真实的事该怎么解释、该怎么处理」上,不发生在「它是否真实」上。本篇不裁抑郁的本体论,也不裁诊断标准——那是另一个课题。

1.2 药理:争议双方共用的前提

一个容易被忽略的事实:在这场辩论里,没有人质疑 SSRI 的急性药理。批评方在描述这类药时的说法是它们急性升高突触间隙血清素、长期则可能产生方向相反的补偿性变化;反驳方在辩护时的说法是「almost all effective antidepressants have an effect on the 5-HT system」。两侧共享同一条药理学前提,分歧在这条前提能推出什么。

这一点非常重要,因为它把争议的位置钉死了:争的不是「药动不动血清素」(动),也不是「药有没有效」(有),而是「药动血清素」能否倒推出「病因是血清素不足」。 那正是机制跳。

1.3 药效:522 项试验的硬边界

“We identified 28 552 citations and of these included 522 trials comprising 116 477 participants. In terms of efficacy, all antidepressants were more effective than placebo, with ORs ranging between 2·13 (95% credible interval [CrI] 1·89–2·41) for amitriptyline and 1·37 (1·16–1·63) for reboxetine.”

来源:Cipriani et al., “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 (2018), doi:10.1016/S0140-6736(17)32802-7(✓ 主笔亲核,PMC 全文

注意作者名单里有 Erick H. TurnerJohn P. A. Ioannidis——前者是揭露抗抑郁药选择性发表的那位(见 2.7),后者是元研究领域最著名的怀疑论者。这不是一份天真的报告。所以「抗抑郁药只是安慰剂」这个命题,在本篇一开始就被否掉了,而且是被最不容易被指责为药厂喉舌的那组人否掉的。


二 效应量:四路收敛到约两个 HAMD 点

2.1 FDA 个体数据:1.75 点,以及一个三峰

Stone 等人拿到 1979 至 2016 年间提交给 FDA 的 232 项随机双盲安慰剂对照试验、73 388 名受试者的个体层数据,做了目前唯一的个体层分布分析 [一手逐字]:

“The random effects mean difference between drug and placebo favored drug (1.75 points, 95% confidence interval 1.63 to 1.86). … The best fitting model of response distributions was three normal distributions, with mean improvements from baseline to end of treatment of 16.0, 8.9, and 1.7 points. These distributions were designated Large, Non-specific, and Minimal responses, respectively. Participants who were treated with a drug were more likely to have a Large response (24.5% v 9.6%) and less likely to have a Minimal response (12.2.% v 21.5%).”

“The trimodal response distributions suggests that about 15% of participants have a substantial antidepressant effect beyond a placebo effect in clinical trials, highlighting the need for predictors of meaningful responses specific to drug treatment.”

这是全篇最有信息量的一个数字,因为它同时否掉两侧:平均 1.75 点确实小到令人尴尬,但平均值掩盖了一个分布——大约 15% 的人得到了超出安慰剂的实质获益,而剩下的人几乎没有。「药有效」和「药无效」都是对同一个分布的错误概括。

2.2 同一批数据的再分析:1.97 点,证据等级「极低」

诺迪克考科蓝中心的三位作者把 Cipriani 的 522 项试验重新过了一遍风险偏倚与 GRADE,并调取了其中 19 项的临床研究报告原件 [一手逐字]:

“The outcome data reported by Cipriani et al differed from the clinical study reports in 12 (63%) of 19 trials. The certainty of the evidence for the placebo-controlled comparisons should be very low according to GRADE due to a high risk of bias, indirectness of the evidence and publication bias. The mean difference between antidepressants and placebo on the 17-item Hamilton depression rating scale (range 0-52 points) was 1.97 points (95% CI 1.74 to 2.21).”

来源:Munkholm, Paludan-Müller & Boesen, BMJ Open 9(6):e024886 (2019), doi:10.1136/bmjopen-2018-024886(◐ 官方摘要逐字)

注意这里的双重信息:批评方自己算出的点差(1.97)比原作者算的还略高,他们的攻击点是证据确定性,不是效应方向。

2.3 独立系统综述:−1.94 点,低于自定阈值,且危害上升

“SSRIs significantly reduced the Hamilton Depression Rating Scale (HDRS) at end of treatment (mean difference −1.94 HDRS points; 95% CI −2.50 to −1.37; P < 0.00001; 49 trials…). The effect estimate, however, was below our predefined threshold for clinical significance of 3 HDRS points. … SSRIs significantly increased the risks of serious adverse events (OR 1.37; 95% CI 1.08 to 1.75; P = 0.009; 44 trials…). This corresponds to 31/1000 SSRI participants will experience a serious adverse event compared with 22/1000 control participants.”

“SSRIs might have statistically significant effects on depressive symptoms, but all trials were at high risk of bias and the clinical significance seems questionable. … The potential small beneficial effects seem to be outweighed by harmful effects.”

来源:Jakobsen et al., “Selective serotonin reuptake inhibitors versus placebo in patients with major depressive disorder”, BMC Psychiatry 17:58 (2017), doi:10.1186/s12888-016-1173-2(✓ 主笔亲核,PMC 全文

关键要看清这段话的结构:「3 个 HAMD 点」这个临床显著阈值是作者自己预设的(predefined threshold),不是外部公认常数。所以这一步是「效应存在但小于我们选的尺子」,不是「效应不存在」。诚实地说:那把尺子的刻度本身有争议,本篇不裁它——Hengartner 与 Plöderl 2022 年专门写过一篇论最小重要差异的估计(○ 仅题名可核,未取回全文,不作承重)。

2.4 严重度调节:Kirsch 的经典结论与它的真实措辞

Kirsch 那篇被引用最多、也最常被误引的论文,原文说的是 [一手逐字]:

“Drug–placebo differences in antidepressant efficacy increase as a function of baseline severity, but are relatively small even for severely depressed patients. The relationship between initial severity and antidepressant efficacy is attributable to decreased responsiveness to placebo among very severely depressed patients, rather than to increased responsiveness to medication.”

来源:Kirsch et al., “Initial severity and antidepressant benefits: a meta-analysis of data submitted to the Food and Drug Administration”, PLoS Medicine 5(2):e45 (2008), doi:10.1371/journal.pmed.0050045(✓ 主笔亲核,PMC 全文

不是「抗抑郁药无效」,而是「药与安慰剂的差随基线严重度上升,而这个上升主要来自安慰剂反应在重症者身上变弱」。这个区别在公共讨论里几乎总是丢掉的。

顺手记一件本篇在核对时撞见的事,它本身就是名号跳的活体。 这篇论文的 PLoS Medicine 官方「Editors’ Summary」(期刊编辑部撰写、附在正文后的科普摘要)第一段写着 [一手逐字]:

“Depression is a serious medical illness caused by imbalances in the brain chemicals that regulate mood.”

同一本期刊、同一年,一边发表着抗抑郁药效应量的定量批判,一边在自己的编辑部摘要里把「化学失衡」当成事实陈述。 这不是药厂做的,不是精神科医生做的,是一份高声誉医学期刊的编辑部做的(2008 年)。

2.5 主动降级:四路收敛不等于四次独立测量

本篇最容易被误读的地方,由本篇自己先说。 1.75 / 1.97 / −1.94 这三个数字看着像三次独立复现,其实不是:Stone 用的是 FDA 提交的 232 项试验,Munkholm 复算的是 Cipriani 的 522 项(其中包含大量同一批注册试验),Jakobsen 的 131 项也从同一个池子里取样。它们共享受试者、共享量表、共享行业赞助的试验设计。 收敛在这里说明的是「不同分析团队从大体相同的数据里算不出实质不同的点差」,而是「四种独立方法互证了这个数」。按本库方法论封顶篇的承重光谱,这一条应记为「框架内硬」而非「真锚」。

2.6 反向一击:问题可能在量表,不在药

哥德堡大学团队做了一件反方向的事:拿 18 项行业赞助的安慰剂对照试验、6 669 名受试者的患者层数据,问如果不用 HAMD-17 总分、而只用其中「depressed mood」这一个条目(0–4 分)当终点,结果会怎样 [一手逐字]:

“While 18 out of 32 comparisons (56%) failed to separate active drug from placebo at week 6 with respect to reduction in HDRS-17-sum, only 3 out of 32 comparisons (9%) were negative when depressed mood was used as an effect parameter (P<0.001). … the frequent use of the HDRS-17-sum as an effect parameter may have distorted the current view on the usefulness of SSRIs and hampered the development of novel antidepressants.”

来源:Hieronymus, Emilsson, Nilsson & Eriksson, Molecular Psychiatry 21(4):523–530 (2016), doi:10.1038/mp.2015.53(◐ 官方摘要逐字。利益申报照登:末位作者 Elias Eriksson 申报曾任 Eli Lilly 与 H. Lundbeck 顾问/获讲者酬金)

这条链很硬也很有意思:HAMD-17 的 17 个条目里包含失眠、体重、胃肠道症状等受药物副作用直接污染的项目,把它们塞进总分会同时稀释信号、放大噪声。如果这个论证成立,那么「效应量小」这件事的一部分是测量工具造成的假象,而不是药的性质。 本篇不裁它成立与否——这是测量代理性篇那台病的正宗病例:一个为方便评分而造的量表,被当成了「抑郁的量」本身。

2.7 底下垫着的报告偏倚:从 32% 到 5%

上面所有点差都建立在「我们看得到的试验」之上。看得到多少,是 Turner 十几年前用 FDA 审评文件测出来的 [一手逐字]:

“Among 74 FDA-registered studies, 31%, accounting for 3449 study participants, were not published. … According to the published literature, it appeared that 94% of the trials conducted were positive. By contrast, the FDA analysis showed that 51% were positive. Separate meta-analyses of the FDA and journal data sets showed that the increase in effect size ranged from 11 to 69% for individual drugs and was 32% overall.”

来源:Turner, Matthews, Linardatos, Tell & Rosenthal, NEJM 358(3):252–260 (2008), doi:10.1056/NEJMsa065779(◐ 官方摘要逐字,全文付费墙未取回)

十四年后同一位作者做了更新版,这次结论朝好的方向走 [一手逐字]:

“Among newer antidepressants, transparent publication occurred more with positive (15/15 = 100%) than negative (7/15 = 47%) trials (OR 35.1, CI 95% 1.8 to 693). … For newer antidepressants, FDA-based effect size (ES FDA) was 0.24 (CI 95% 0.18 to 0.30), while journal-based effect size (ES Journals) was 0.29 (CI 95% 0.23 to 0.36). Thus, effect size inflation, presumably due to reporting bias, was 0.05, less than for older antidepressants (0.10).”

“Reporting bias persists but appears to have diminished for newer, compared to older, antidepressants.”

来源:Turner et al., PLoS Medicine 19(1):e1003886 (2022), doi:10.1371/journal.pmed.1003886(✓ 主笔亲核,PMC 全文

双向都要读:报告偏倚是真的(负性试验里 6 项未发表、2 项被写成正性),但它在缩小(透明报告率从 11% 升到 47%,效应量膨胀从 0.10 降到 0.05)。 「文献都是被操纵的」这句话在 1990 年代的抗抑郁药数据上有很强的实证支持,在 2010 年代之后的数据上就明显弱了。

2.8 疗效层小结

命题 裁决
抗抑郁药优于安慰剂 成立(522 项试验、21 种药无一例外)
平均量级约 2 个 HAMD 点 成立(但四路数据高度重叠,记为框架内硬)
这个量级临床上够不够 未结案(阈值是各方自设的;量表本身有争议)
一部分人获益远超平均 成立但已被挑战(Stone 三峰 vs 2025 复现失败,见 3.3)
文献曾被选择性发表放大 成立(32% 膨胀,老药)
这种放大仍在继续 已明显减弱(5% 膨胀,新药)

三 记分卡:同一个试验,67% 还是 35%

3.1 原始报告

STAR*D 是有史以来规模最大的前瞻性抗抑郁治疗试验,4 041 名筛查阳性成人,最多接受四个阶梯的治疗。原报告的数字 [一手逐字]:

“The QIDS-SR(16) remission rates were 36.8%, 30.6%, 13.7%, and 13.0% for the first, second, third, and fourth acute treatment steps, respectively. The overall cumulative remission rate was 67%.”

来源:Rush et al., “Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STARD report”, Am J Psychiatry* 163(11):1905–1917 (2006), doi:10.1176/ajp.2006.163.11.1905(◐ 官方摘要逐字)

「四轮下来三分之二的人缓解」这句话此后十七年被写进了无数教材、讲座与患者手册。

3.2 RIAT 再分析:35.0%

2023 年,一组作者按「恢复被隐藏与被弃置的试验」(RIAT)规程,拿到 STAR*D 的患者层数据,按原始研究方案规定的终点重算 [一手逐字]:

“STARD investigators did not use the protocol-stipulated HRSD to report cumulative remission and response rates in their summary article and instead used a non-blinded clinic-administered assessment. This inflated their report of outcomes, as did their inclusion of 99 patients who scored as remitted on the HRSD at study outset as well as 125 who scored as remitted when initiating their next-level treatment. These patients should have been excluded from data analysis. In contrast to the STARD-reported 67% cumulative remission rate after up to four antidepressant treatment trials, the rate was 35.0% when using the protocol-stipulated HRSD and inclusion in data analysis criteria.”

“STAR*D’s cumulative remission rate was approximately half of that reported.”

来源:Pigott et al., BMJ Open 13(7):e063095 (2023), doi:10.1136/bmjopen-2022-063095(✓ 主笔亲核,PMC 全文

三条指控各自可独立核验:改用了方案明确排除的非盲评估;纳入了 99 名入组时即已缓解者;纳入了 125 名进入下一阶梯时即已缓解者。这不是统计口味之争,是方案偏离。RIAT 团队还在 2025 年放出两份延伸再分析(换药疗法增效疗法),均为 medRxiv 预印本(○ 仅题名可核,未经同行评议,本篇不作承重)。

按本库同行评议篇的分类,这是「过评议≠正确」的又一个活体;按复制危机篇的分类,这是「同一批数据、不同分析规程、结论差一倍」的教科书案例。**我必须诚实标注一个空位:本篇未能取回 STAR*D 原研究者对 Pigott 再分析的公开答辩(若存在)。** 因此第三节只承重「按原方案终点重算得 35.0%」这个可独立核验的算术事实,不承重任何关于原作者动机的推断。

3.3 2025 年:Stone 的三峰没被复现

上面 2.1 那个「约 15% 获得实质药物特异获益」的结论,在 2025 年被一次专门的复现尝试挑战 [一手逐字]:

“The best fitting models were either bimodal or trimodal but the trimodal solution did not meet criteria for replication. … For the trimodal model, the component with the largest change (M = -14.3, SD = 6.4) applied to 52% of patients, which differed substantially from the large effect component in Stone et al (M = -18.8, SD = 5.1), which applied to 7.2%.”

“This analysis failed to identify the trimodal distribution of response reported in Stone et al. In addition to being difficult to operationalize for regulatory purposes, results from mixture modeling are not sufficiently reliable to replace the more robust approach of comparing mean differences in depression rating scale scores between treatment arms.”

来源:Xu, Naudet, Kim, Hengartner, Horowitz, Kirsch, Moncrieff, Pigott & Plöderl, J Clin Epidemiol 187:111943 (2025), doi:10.1016/j.jclinepi.2025.111943(◐ 官方摘要逐字)

这一条必须双向读,否则就是单侧引用。 一方面:它是一次预先声明标准的失败复现,来自九位作者,结论明确否掉了三峰。另一方面:它用的数据是 **STAR*D 第一阶梯,单臂、开放标签、无安慰剂组——而 Stone 的三峰恰恰是从「药臂 vs 安慰剂臂」的对比中分解出来的。用一个没有安慰剂臂的数据集去复现一个关于药物特异性反应的分布分解,其推论力天然受限;这一点作者群自己也在方法里写明了数据来源。所以这层的正确裁决是「未结案」,不是「已被反驳」。** 同时也要看到作者名单——它与本篇引用的批评侧核心作者高度重叠,属单侧集中,本篇在第十节自指里再点一次。


四 病因层:六个方向,六次不支持

4.1 伞状综述的设计与结论

“The serotonin hypothesis of depression is still influential. We aimed to synthesise and evaluate evidence on whether depression is associated with lowered serotonin concentration or activity in a systematic umbrella review of the principal relevant areas of research.”

六个方向由一次范围审查确定,逐字如下:

“(1) Serotonin and the serotonin metabolite 5-HIAA–whether there are lower levels of serotonin and 5-HIAA in body fluids in depression; (2) Receptors – whether serotonin receptor levels are altered in people with depression; (3) The serotonin transporter (SERT) – whether there are higher levels of the serotonin transporter in people with depression (which would lower synaptic levels of serotonin); (4) Depletion studies – whether tryptophan depletion (which lowers available serotonin) can induce depression; (5) SERT gene – whether there are higher levels of the serotonin transporter gene in people with depression; (6) Whether there is an interaction between the SERT gene and stress in depression.”

结论逐字:

“The main areas of serotonin research provide no consistent evidence of there being an association between serotonin and depression, and no support for the hypothesis that depression is caused by lowered serotonin activity or concentrations. Some evidence was consistent with the possibility that long-term antidepressant use reduces serotonin concentration.”

“Our comprehensive review of the major strands of research on serotonin shows there is no convincing evidence that depression is associated with, or caused by, lower serotonin concentrations or activity.”

“We suggest it is time to acknowledge that the serotonin theory of depression is not empirically substantiated.”

来源:Moncrieff, Cooper, Stockmann, Amendola, Hengartner & Horowitz, “The serotonin theory of depression: a systematic umbrella review of the evidence”, Molecular Psychiatry 28(8):3243–3256 (2022;在线 2022-07-20),doi:10.1038/s41380-022-01661-0(✓ 主笔亲核,PMC 全文。PROSPERO 注册号 CRD42020207203)

利益申报照登(这是对称纪律的一部分):该文自报——Horowitz 于 2022 年 4 月联合创办一家帮助人们在加拿大安全停用抗抑郁药的公司;Hengartner 有一本 2021 年出版的书《Evidence-biased Antidepressant Prescription》的版税;Moncrieff 有多本精神科药物著作版税,并任 Critical Psychiatry Network 联席主席与 Council for Evidence-based Psychiatry 理事。

4.2 逐域的关键数字

方向 结果(该综述内逐字或近逐字) 确定性评级
5-HIAA(脑脊液) 两项元分析(19 项研究,7 项重叠)「found no evidence of an association between 5-HIAA concentrations and depression」,最大 n=1 002
血浆血清素 一项队列元分析无关联;且「lowered serotonin concentration was associated with antidepressant use」(n=1 869) 中(并指向反向因果)
5-HT1A 受体 多数结果为无差异或更低的抑制性受体水平——若原始异常在此,方向反而指向抑郁者血清素更高;最大 n=561
SERT 结合 三项元分析显示「weak and inconsistent evidence of reduced binding in some areas」;最大 n=1 845
色氨酸耗竭 健康志愿者 17/19 项无情绪效应;有家族史者「weak evidence of an effect」(n=75);最近一次系统综述停在 2007 年 极低
SERT 基因 11.5 万人关联研究「OR = 1.000, p = 0.994」;4.3 万人协作元分析「OR 1.00 (0.95 to 1.05), p = 0.95」 高(无效)
基因 × 压力 「31 other analyses using different measures of stress or depression found no significant interaction」 高(无效)

其中基因两行的原始文献是Border et al. 2019, Am J Psychiatry(○ 经伞状综述转引,DOI 与卷期已实查)与Culverhouse et al. 2018, Mol Psychiatry(○ 同上)。本库IQ 篇已经处理过同一场候选基因崩塌的另一半(5-HTTLPR × 压力交互曾是心理学最著名的候选基因发现之一),本篇不重做。

4.3 一条独立的、更早的定量证据

色氨酸耗竭方向有一份 2007 年的专门元分析,它的结论比双方引用它时都更精确 [一手逐字]:

“5-HT or NE/DA depletion did not decrease mood in healthy controls. 5-HT or NE/DA depletion slightly lowered mood in healthy controls with a family history of MDD. In drug-free patients with MDD in remission, a moderate mood decrease was found for ATD… ATD induced relapse in patients with MDD in remission who used serotonergic antidepressants.”

“Although depletion studies usefully investigate the etiological link of 5-HT and NE with MDD, they fail to demonstrate a causal relation. They presumably clarify a vulnerability trait to become depressed.”

来源:Ruhé, Mason & Schene, Molecular Psychiatry 12(4):331–359 (2007), doi:10.1038/sj.mp.4001949(◐ 官方摘要逐字)

这段话是本篇最精确的一块砖,两侧都不能整段拿走。 反对派拿走前半句(健康人耗竭血清素不会变抑郁),支持派拿走中间句(有家族史者与缓解期患者会),而作者自己给出的裁决是最后那句:这些实验说明的是易感性特质,不是因果关系。

4.4 2025 年的侧翼尝试

一项 2025 年的研究换了个思路:直接在人脑组织里查基因与表观遗传变异。使用 216 名成人患者的 232 份生物库胶质瘤组织样本,分析 MAOA 与 5HTT 基因的遗传与甲基化变异,发现男性患者中 MAOA 基因变异与组织血清素水平显著相关。来源:Wu, Melin, Björkblom & Sjöberg, Scientific Reports 15 (2025), doi:10.1038/s41598-025-25464-9(◐ 摘要级逐字)。

本篇明确不让它承重:样本是脑肿瘤患者的肿瘤/瘤旁组织,不是抑郁队列;它回答的是「人脑组织里血清素通路的遗传调控是否可测」,不是「抑郁是否由血清素缺乏引起」。列在这里是为了说明这个方向 2025 年仍有正规研究在做,不是为了给任何一侧加分。


五 最强的正面证据,摆上台面

对称纪律要求:不能只审「支持假说的证据不足」,还必须把支持侧手里最好的那一件拿出来、按同样标准称一遍。

5.1 直接测量血清素释放能力

2023 年,一个包含 Nutt、Cowen、Howes、Knudsen 等血清素研究核心人物的团队发表了第一次对人脑血清素释放能力的直接测量 [一手逐字]:

“The serotonin hypothesis of depression proposes that diminished serotonergic (5-HT) neurotransmission is causal in the pathophysiology of the disorder. Although the hypothesis is over 50 years old, there is no firm in vivo evidence for diminished 5-HT neurotransmission.”

“Following d-amphetamine administration, frontal nondisplaceable binding potential (BPND) was significantly reduced in the HC group (1.04 ± 0.31 vs. 0.87 ± 0.24, p < .001) but not in the MDE group (0.97 ± 0.25 vs. 0.92 ± 0.22, not significant). ΔBPND of the MDE group was significantly lower than that of the HC group (HC: 15% ± 14% vs. MDE: 6.5% ± 20%, p = .041).”

“This first direct assessment of 5-HT release capacity in people with depression provides clear evidence for dysfunctional serotonergic neurotransmission in depression by demonstrating reduced 5-HT release capacity in patients experiencing an MDE.”

来源:Erritzoe et al., “Brain Serotonin Release Is Reduced in Patients With Depression: A [11C]Cimbi-36 Positron Emission Tomography Study With a d-Amphetamine Challenge”, Biological Psychiatry 93(12):1089–1098 (2023), doi:10.1016/j.biopsych.2022.10.012(◐ 官方摘要逐字)

先记这篇论文自己交出的两件东西:第一,它的背景句是支持侧最坦白的自陈——假说五十多年了,在体证据从来不硬。第二,它的正面结果是:17 名未用药患者对 20 名健康对照,p = .041,单一感兴趣区(额叶皮质)。

再记它的结构性局限(这些都在摘要里,不需要猜):n=17 vs 20;两组年龄差距大(44 ± 13 岁 vs 32 ± 9 岁);性别构成不平衡;效应靠一个刚过 0.05 的 p 值;作为「首次」测量尚无独立复现。

5.2 三封回应

这项研究发表后,同一卷登出三份评论:Hengartner 与 Plöderl 的「No Clear Evidence of Reduced Brain Serotonin Release Capacity in Patients With Depression」、原作者团队的答辩,以及 Slifstein 与 Abi-Dargham 的方法学评论「Detecting Pharmacologically Induced Serotonin Release in Depression With Positron Emission Tomography Imaging: A New Approach」(三份均 ○ 仅题名与作者可核,正文付费墙未取回;本篇只承重「这场交锋存在且立场如题名所示」,不承重任何一方的具体论证)。

5.3 反驳方自己给出的「更准确的结论」

2023 年 6 月,Molecular Psychiatry 同一期集中刊出六篇针对伞状综述的文章:Jauhar 等三十余位作者的主反驳、Bartova/Lanzenberger/Rujescu/Kasper 的回应、Jacobsen 的替代解读、El-Mallakh 等的短评、Almulla 与 Maes 的「前体才是主角」,以及 Moncrieff 团队的答辩。

主反驳的火力集中在方法学 [一手逐字]:

“We present reasons for why this conclusion is overstated, including methodological weaknesses in the review process, selective reporting, over-simplification, and errors in the interpretation of neuropsychopharmacological findings.”

它有几条相当具体、值得原样记下的技术指控:伞状综述把不同层级的研究单元混在一起汇总;用了后补协议修正引入的「改造版 GRADE」而非该领域已有的效度判据;漏掉了一项在 MDD 患者中显示效应的临床与分子影像研究;对同一篇被引元分析中「未服药抑郁者色氨酸耗竭效应 Hedge’s g = −1.9(95% CI −3.02 至 −0.78),去掉一个潜在离群值后 −1.06(−1.83 至 −0.29)」这组数据未予呈现;把 5-HT1A 受体一概当成突触前自受体(原文指出「Most are post-synaptic 5-HT1A heteroreceptors」);忽略了一篇元分析里 364 人中有 149 人从未用药,因而「SERT 减少全由既往用药造成」的解释站不住。

然后是本篇整个第五节的落点。 这份由该领域最有分量的一批人联署、火力全开的反驳,在结尾自己写下的「更准确的结论」是 [一手逐字]:

“A more accurate, constructive conclusion would be that acute tryptophan depletion and decreased plasma tryptophan in depression indicate a role for 5-HT in those vulnerable to or suffering from depression, and that molecular imaging suggests the system is perturbed. The proven efficacy of SSRIs in a proportion of people with depression lends credibility to this position.”

来源:Jauhar, Arnone, Baldwin, Bloomfield, Browning, … Cowen (共三十余位作者), “A leaky umbrella has little value: evidence clearly indicates the serotonin system is implicated in depression”, Molecular Psychiatry 28(8):3149–3152 (2023), doi:10.1038/s41380-023-02095-y(✓ 主笔亲核,PMC 全文。作者列表含一位 H. Lundbeck A/S 在职者与一位现属 Intra-Cellular Therapies 者,按对称纪律照登)

「a role」「perturbed」「in a proportion of people」——这就是最强辩护方在最用力的时候,为这条假说主张的全部强度。 没有人在保卫「抑郁是由血清素不足引起的」。这不是修辞胜利,这是本篇第七层的证据基础:如果连最强辩护都不主张那句话,那句话就不可能是学界的承重命题。

Moncrieff 团队的答辩标题本身就把这个错位钉住了——「The serotonin hypothesis of depression: both long discarded and still supported?」(一边说这理论早就被丢掉了,一边说它仍有证据支持)。其可见正文逐字:

“Our paper exposed this claim as untrue. The main areas of research do not provide support for the most well-known neurochemical hypothesis, the serotonin theory of depression—the idea that depression is caused by low levels or low activity of serotonin. There are always further studies that can be presented to support the serotonin theory… The onus is on others to present a convincing case based on other types of research or alternative hypotheses, not on us to disprove every speculation.”

(✓ 主笔亲核,来源为 nature.com 该文页面首段可见部分;其后正文在付费墙内,本篇不越出可见部分作任何主张)

一份 2025 年由独立第三方作者写的方法学分析,给出的落点也在同一个位置 [一手逐字]:「We argue for a more nuanced understanding of serotonin’s role in depression, not as a sole cause but as a potential modulatory factor in vulnerable populations.」来源:Lepri, Psychiatr Danub 37(Suppl 1):187–188 (2025), PMID 40982824(◐ 官方摘要逐字;会议增刊,证据层级低,仅作第三方立场标记)


六 独占核心:那句话是谁说的

前五节做的都是别人做过的事。这一节是本篇的独占切口:不问「化学失衡对不对」,只问「这句话从哪来、经谁的口进入病人的脑子」——并且要求每条渠道都拿出可核验的原件。

6.1 起源:原作者自己就写了「远非定论」

1967 年,那篇被今天所有各方共同引作血清素假说源头的综述,开篇第二句是 [一手逐字]:

“The title of this review would be regarded by some psychiatrists as provocative; they would relegate the biochemical concomitants of depression and mania to a secondary position and deny that biochemical changes have any place in the aetiology of these conditions. However, in my view, the weight of evidence, although it is by no means conclusive, suggests that biochemical changes are most important in the aetiology of affective disorders.”

“This does not deny that psychological and environmental events may precipitate and maintain the biochemical events which in turn lead to the affective disorder.”

来源:Alec Coppen, “The Biochemistry of Affective Disorders”, British Journal of Psychiatry 113(504):1237–1264 (1967), doi:10.1192/bjp.113.504.1237(✓ 主笔亲核,经 Cambridge Core 该文 Extract 段逐字取回)

读清这段话的三件事:作者知道这在当时是少数派立场(「would be regarded by some psychiatrists as provocative」);他自己写明证据「by no means conclusive」;他明确不排除心理与环境因素。更上游的儿茶酚胺(去甲肾上腺素)版本来自 Schildkraut 1965(doi:10.1176/ajp.122.5.509,○ 原文未取回,不作承重;其为去甲肾上腺素而非血清素假说这一点,经下述 2005 年论文全文逐字确认)。

所以名号跳的第一段落差在这里就已经成立:源头文献是一个自陈证据不足、明确保留环境因素的少数派假说;四十年后进入公众的版本是一个确定的、单因的、生物学的疾病解释。

6.2 药厂渠道:一份可以自己打开看的原件

2002 年 10 月 15 日,葛兰素史克官方 Paxil(帕罗西汀)网站的「Paxil 治疗什么」页面,原文写着 [一手逐字]:

“With continued treatment, Paxil can help restore the balance of serotonin (a naturally occurring brain chemical) — which helps reduce the symptoms of anxiety and depression.”

来源:GlaxoSmithKline, “What does Paxil treat”, paxil.com/about/ab_trt.html,Wayback 2002-10-15 存档(✓ 主笔亲核,快照全文逐字)

同一页还列出 Paxil 当时获批的六个适应症:抑郁、广泛性焦虑、社交焦虑、惊恐障碍、强迫症、创伤后应激障碍——同一个「血清素平衡」被用来解释六种不同的病,这正是下面那篇论文抓住的逻辑漏洞。

辉瑞的 Zoloft(舍曲林)电视广告文案,经一份 2005 年的同行评议论文逐字记录 [一手逐字]:

“For example, Pfizer’s television advertisement for the antidepressant sertraline (Zoloft) stated that depression is a serious medical condition that may be due to a chemical imbalance, and that ‘Zoloft works to correct this imbalance’.”

同一篇论文把当时的证据状况写得比十七年后的伞状综述更直白 [一手逐字]:

“While neuroscience is a rapidly advancing field, to propose that researchers can objectively identify a ‘chemical imbalance’ at the molecular level is not compatible with the extant science. In fact, there is no scientifically established ideal ‘chemical balance’ of serotonin, let alone an identifiable pathological imbalance.”

“Yet, this ex juvantibus line of reasoning (i.e., reasoning ‘backwards’ to make assumptions about disease causation based on the response of the disease to a treatment) is logically problematic—the fact that aspirin cures headaches does not prove that headaches are due to low levels of aspirin in the brain.”

“To our knowledge, there is not a single peer-reviewed article that can be accurately cited to directly support claims of serotonin deficiency in any mental disorder, while there are many articles that present counterevidence. Furthermore, the Diagnostic and Statistical Manual of Mental Disorders (DSM), which is published by the American Psychiatric Association and contains the definitions of all psychiatric diagnoses, does not list serotonin as a cause of any mental disorder.”

它还引了美国国家精神卫生研究所临床科学实验室研究者的原话:

“[T]he demonstrated efficacy of selective serotonin reuptake inhibitors…cannot be used as primary evidence for serotonergic dysfunction in the pathophysiology of these disorders.”

来源:Lacasse & Leo, “Serotonin and depression: a disconnect between the advertisements and the scientific literature”, PLoS Medicine 2(12):e392 (2005), doi:10.1371/journal.pmed.0020392(✓ 主笔亲核,PMC 全文

「阿司匹林治头痛不证明头痛缘于脑内阿司匹林不足」——机制跳的最简反例,2005 年就印在一份高声誉期刊上了,比伞状综述早十七年。 这条时间线本身是重要发现:本篇处理的不是一个 2022 年才被揭开的秘密,而是一件在专业文献里公开了二十年、却没有改变公众版本的事。

6.3 「城市传说」辩护,以及它自带的让步

为精神科辩护得最有力的那篇文章,写作者是 SUNY Upstate 医学院精神科荣休教授、《Psychiatric Times》2007–2010 年主编 Ronald Pies。全文关键段落逐字:

“I am not one who easily loses his temper, but I confess to experiencing markedly increased limbic activity whenever I hear someone proclaim, ‘Psychiatrists think all mental disorders are due to a chemical imbalance!’ In the past 30 years, I don’t believe I have ever heard a knowledgeable, well-trained psychiatrist make such a preposterous claim, except perhaps to mock it. On the other hand, the ‘chemical imbalance’ trope has been tossed around a great deal by opponents of psychiatry, who mendaciously attribute the phrase to psychiatrists themselves. And, yes-the ‘chemical imbalance’ image has been vigorously promoted by some pharmaceutical companies, often to the detriment of our patients’ understanding.”

“In truth, the ‘chemical imbalance’ notion was always a kind of urban legend- – never a theory seriously propounded by well-informed psychiatrists.”

“The legend of the ‘chemical imbalance’ should be consigned to the dust-bin of ill-informed and malicious caricatures.”

来源:Ronald W. Pies, “Psychiatry’s New Brain-Mind and the Legend of the ‘Chemical Imbalance'”, Psychiatric Times(2011),Wayback 2021-01-21 存档(✓ 主笔亲核,全文逐字;该文脚注 3 引的正是上面 6.2 那篇 Lacasse–Leo 2005)

这段文字是本篇的对称枢轴,请把它读满。 同一个人、同一篇文章里说了两件事:(a) 没有一个受过良好训练的精神科医生认真主张过这个说法;(b) 是的,一些药厂大力推广过它,常常损害病人的理解。

换句话说:关于药厂渠道,争议双方在 2011 年就已经一致了。 剩下的争点只有一个——专业界是否也在传。而这个争点,2022 年之后有了实测数据。

6.4 学界渠道之争:一次内容分析

批评侧针对「城市传说」这个辩护做了一次直接的内容分析 [一手逐字]:

“In response, leading psychiatrists suggested it was an ‘urban legend’ that was never taken seriously by the psychiatric profession. To interrogate these claims, we examined the coverage of the serotonin theory of depression in a sample of highly cited and influential academic literature from 1990, when the theory started to be popularized, to 2010 when these responses were articulated. We analysed 30 highly cited reviews of the aetiology of depression in general, 30 highly cited papers on depression and serotonin specifically and a sample of influential textbooks.”

“The majority of the aetiology reviews supported the hypothesis… All textbooks supported the theory, at least in some sections, and devoted substantial coverage to it, although some also acknowledged it remained provisional. The findings suggest that the serotonin theory was endorsed by the professional and academic community.”

“The analysis suggests that, despite protestations to the contrary, the profession bears some responsibility for the propagation of a theory that has little empirical support and the mass antidepressant prescribing it has inspired.”

来源:Ang, Horowitz & Moncrieff, “Is the chemical imbalance an ‘urban legend’? An exploration of the status of the serotonin theory of depression in the scientific literature”, SSM – Mental Health 2:100098 (2022), doi:10.1016/j.ssmmh.2022.100098(◐ 官方摘要逐字,经 Wayback 2025-01-08 存档 取回;⚠ 单侧来源已标明:作者为本议题批评侧核心人物,正文取样与编码规则未经本篇独立复核,因此本篇只承重「存在一次这样设计的内容分析并得出如上结论」,不承重其编码判断)

必须诚实标注的空位:本篇没有找到由第三方独立完成的同类内容分析。所以「学界渠道是否也在传」这一问,在文献层面仍是一方举证、另一方未以同类证据回应的状态。这不是平手,也不是定案。

6.5 一处引用与原文不符的记录

核对时撞见一件必须写下来的事,因为它属于本库方法论封顶篇所列的「引用雷区」类型。

伞状综述引言里有一句:「Many general practitioners also subscribe to this view [17]」——即许多全科医生也持化学失衡观点。其参考文献 [17] 是 Read 等 2020 年对英国全科医生的调查。本篇取回该文官方摘要逐字比对,其结论与被引方向相反 [一手逐字]:

“In keeping with previous findings, this small sample of GPs had a predominantly psycho-social perspective on the causes of, and treatments for, depression.”

来源:Read, Renton, Harrop, Geekie & Dowrick, Ther Adv Psychopharmacol 10:2045125320950124 (2020), doi:10.1177/2045125320950124(✓ 官方摘要逐字,PMC 全文

该调查的实际有用数据在别处,且相当重要:样本仅 66 名全科医生(因新冠中断);只有 29% 认为自己关于撤药的知识「充分」;只有 17% 认为自己能区分撤药反应与原病复发;68% 希望获得更多培训。

本篇的处理:不推测原因(笔误、引错编号、援引该文其他段落均有可能),只把可核验的不符记录在案,并不使用「许多全科医生持化学失衡观点」这一说法。真正可用的替代证据在下一小节,而且它比这句话强得多。

6.6 实测:哪条渠道真的把信念传了过去

这是本节的落点,也是 2024–2025 年这个课题上最有价值的新证据。

第一项(n=426,美国某大型公立大学本科生)[一手逐字]:

“Sixty-two percent of the sample had heard of the chemical imbalance explanation, most commonly from the classroom.”

来源:Schroder, Russman Block & Moser, “Chemical imbalance and etiological beliefs about depression among college students”, J Am Coll Health 72(7):1993–2000 (2024;在线 2022-07-14),doi:10.1080/07448481.2022.2098037(◐ 官方摘要逐字)

第二项,专门设计来分辨「哪条渠道有效」(n=1 219 名大学生,开放式作答由独立评分者编码)[一手逐字]:

“The most common sources of exposure to this explanation were the classroom, the Internet/media, other people (e.g., friends), and healthcare providers. In a regression analysis, only learning about the chemical imbalance explanation from healthcare providers uniquely predicted the adoption of the chemical imbalance belief. The correlation held even after controlling for depression symptoms, a family history of depression, and having had a diagnosis or treatment of mental health disorder (all of which also uniquely predicted chemical imbalance belief endorsement).”

“These results suggest that healthcare providers play an important role in the dissemination of the chemical imbalance message, which is an oversimplified, scientifically controversial, and potentially treatment-interfering narrative.”

来源:Schroder, Tovey, Forer, Schultz, Kneeland & Moser, “Where do ‘chemical imbalance’ beliefs come from? Evaluating the impact of different sources”, Frontiers in Psychology 15:1469913 (2024;在线 2025-01-08),doi:10.3389/fpsyg.2024.1469913(✓ 主笔亲核,PMC 全文。该刊公开评审人姓名,本文评审人含 Joanna Moncrieff 与 David Cohen——本篇如实记录这一透明度事实,不由此推断任何结论)

这两项加在一起,给出了一个双方都没预料到的答案:接触最广的渠道是课堂,但唯一能独立预测「真的相信」的渠道,是医疗提供者。

于是「城市传说」辩护与「药厂之罪」指控同时被削弱:不是药厂广告(美国 2005 年后处方药广告的表述已被大幅收紧,而这两项调查的对象是 2020 年代的年轻人);也不是「没人这么说」(说的人正在诊室里)。而 2026 年一项预注册的情景实验进一步指出了这种话术在什么条件下被启用 [一手逐字]:

“In line with hypothesis H1, providers exposed to the ‘no-trigger’ vignette were more likely to focus their conversation using biogenetic language.”

“Healthcare providers are more likely to use biogenetic narratives when the cause of a patient’s depression is unclear.”

来源:Schroder & Bommarito, “Medical providers and biogenetic messages about depression: A vignette experiment”, Patient Educ Couns 144:109463 (2026;在线 2025-12-19),doi:10.1016/j.pec.2025.109463(◐ 官方摘要逐字;396 名医生与医学生随机分组。该研究另有两项假设未获支持:无诱因组并未更多推荐药物或住院,且提供者的前测生物遗传信念与治疗推荐不相关——本篇照登,不只引支持性结果)

「当病因说不清的时候,医生更容易改用生物化学语言。」 这句话解释了为什么「化学失衡」既不是学界的正式主张、又能持续四十年不死:它不是一个理论,它是一个在解释力不够时被启用的沟通默认值。

6.7 机构现行表述:2026 年 7 月 26 日当日实查

如果那句话是传播产物,那么它在官方口径里的现状就是可以直接查的。本篇于今日逐一打开:

美国精神医学学会(APA)患者页「什么是抑郁症」 [一手逐字]:

“Biochemical: Differences in certain chemicals in the brain (such as the neurotransmiters serotonin, dopamine and norepinephrine) may contribute to symptoms of depression.”

来源:psychiatry.org, “What Is Depression?”(✓ 主笔亲核,2026-07-26 实查。原文 “neurotransmiters” 拼写如此,未改)。全页出现 “imbalance” 一词仅一处,且用于鉴别诊断中的激素失衡(”hormonal imbalances, vitamin deficiencies, neurological problems”),不用于抑郁的病因表述。

美国国家精神卫生研究所(NIMH)抑郁症出版物 [一手逐字]:

“Antidepressants are medications commonly used to treat depression. They work by changing how the brain produces or uses certain chemicals involved in mood or stress.”

来源:nimh.nih.gov, “Depression”(✓ 主笔亲核,2026-07-26 实查。全页无 “chemical imbalance”,其主题页亦无 “imbalance”/”serotonin” 出现)

「Differences」「may contribute to symptoms」「changing how the brain produces or uses certain chemicals」——请把这些措辞和 2002 年 GSK 那句「restore the balance of serotonin」并排放。 官方口径的退却是真实的、可查的、已经完成的。

这也是本篇的一处诚实空位:本篇没有取到 2022 年之后的公众信念调查。已有的两次大样本调查都在争议之前——澳大利亚 2011 年全国样本「Over 80% of the Australian public attributed depression to day to day problems, death of a close friend or relative, a recent traumatic event, childhood problems, and a chemical imbalance in the brain」(Pilkington, Reavley & Jorm, J Affect Disord 150:356–362, 2013,◐ 官方摘要逐字);美国 General Social Survey「In 2006, 67% of the public attributed major depression to neurobiological causes, compared with 54% in 1996」(Pescosolido et al., Am J Psychiatry 167:1321–1330, 2010,◐ 官方摘要逐字)。

顺带纠一个常被四舍五入的说法。 伞状综述引言写「Surveys suggest that 80% or more of the general public now believe it is established that depression is caused by a ‘chemical imbalance’」,其依据正是上面两项。但 Pilkington 那句的实际结构是:超过 80% 的人同时把抑郁归因于日常问题、亲友去世、近期创伤、童年问题大脑化学失衡——公众持的是多因归因,不是单因的化学失衡论。「80% 相信是化学失衡」与「80% 把化学失衡列为诸因之一」是两个不同的命题,本篇只承重后者。


七 叙事的后果:双向,但不等重

7.1 正向证据:三条

(a) 实验性操纵。 给有抑郁病史或当前发作者一份「看似可信但实为伪造」的生物学检测结果,随机分配为「你的症状由大脑化学失衡引起」或「不是」 [一手逐字]:

“Results showed that chemical imbalance test feedback failed to reduce self-blame, elicited worse prognostic pessimism and negative mood regulation expectancies, and led participants to view pharmacotherapy as more credible and effective than psychotherapy.”

来源:Kemp, Lickel & Deacon, Behav Res Ther 56:47–52 (2014), doi:10.1016/j.brat.2014.02.009(◐ 官方摘要逐字)

注意这条的关键:化学失衡解释没有达到它最被期待的那个效果(减少自责),却带来了预后悲观。这是「有害而且无益」,不是「有代价但值得」。

(b) 相关与干预。 三项研究,症状显著者的生物化学与遗传归因与预后悲观显著相关;而一段强调「基因效应与脑化学是可塑的」的音视频干预显著降低了预后悲观、提升了能动感、降低了整体绝望感(Lebowitz, Ahn & Nolen-Hoeksema, J Consult Clin Psychol 81:518–527, 2013,◐ 官方摘要逐字)。

这条最重要的推论常被漏掉:可修改性比生物性更关键。 有害的不是「生物学解释」本身,而是「生物学=命定」这层附加含义——本库IQ 篇裁过完全同构的一跳(高遗传率被读成不可干预,被近视配镜与 PKU 饮食四重否证)。

(c) 去病耻化的失败。 把精神疾病讲成「和别的病一样的医学疾病」这套策略,是为了减少病耻感才被推广的。十年跟踪的结果 [一手逐字]:

“More of the public embraces a neurobiological understanding of mental illness. This view translates into support for services but not into a decrease in stigma. … Holding a neurobiological conception of these disorders increased the likelihood of support for treatment but was generally unrelated to stigma. Where associated, the effect was to increase, not decrease, community rejection.”

(来源同上 Pescosolido et al. 2010,◐ 官方摘要逐字)

这是最反直觉、也最该被记住的一条:生物学化提高了求治意愿,但没有减少病耻感;在有关联的地方,方向是增加社区排斥。 「化学失衡」这句话最主要的辩护理由——它能减少污名与自责——在实测中,两个都没兑现。

7.2 反向:一处必须承认的空位

对称纪律要求同样审「反向叙事的危害」:2022 年之后媒体大规模报道「血清素理论破产」,是否导致有人擅自停药、依从性下降、结局变差?

本篇的检索没有找到量化这一后果的研究。 已就此专门查过 2022 至 2026 年的文献(血清素/化学失衡 × 媒体/报道/依从性/停药),未见相关实证研究。

所以本篇必须说清一件事:这个方向上的危害,是一个被临床界广泛担心、但尚未被测量的假设。 我不会为了让红队看起来对称而虚构一个等重的反向证据。真实的状况是不对称的:正向(化学失衡叙事有害)有实验与纵向证据,反向(去化学失衡叙事有害)目前只有理论担忧与个案层面的临床报告。

但要立刻加一句,否则这个不对称会被误读:「反向危害尚未被测量」不等于「反向没有危害」,更不等于停药是安全的。 停药的后果不需要靠叙事研究来推断——它有直接的随机对照证据,就在下一节。


八 停药与撤药:两侧都被压住

8.1 旧口径

批评侧对指南的指控原文如下 [一手逐字]:

“The U.K.’s current National Institute for Health and Care Excellence and the American Psychiatric Association’s depression guidelines state that withdrawal reactions from antidepressants are ‘self-limiting’ (i.e. typically resolving between 1 and 2 weeks). This systematic review assesses that claim.”

“Withdrawal incidence rates from 14 studies ranged from 27% to 86% with a weighted average of 56%. Four large studies of severity produced a weighted average of 46% of those experiencing antidepressant withdrawal effects endorsing the most extreme severity rating on offer. Seven of the ten very diverse studies providing data on duration contradict the U.K. and U.S.A. withdrawal guidelines…”

“We recommend that U.K. and U.S.A. guidelines on antidepressant withdrawal be urgently updated as they are clearly at variance with the evidence…”

来源:Davies & Read, Addictive Behaviors 97:111–121 (2019), doi:10.1016/j.addbeh.2018.08.027(◐ 官方摘要逐字。该文同期有三篇公开评论回应,本篇未取回其内容,故不引其论证)

这份综述的方法学局限必须同时写明:24 项研究「diverse methodologies and sample sizes」,其纳入范围包括在线调查等自选样本;56% 是加权平均、未经安慰剂校正;「最极端严重度评分」的分母是「已经报告了撤药反应的人」,不是所有停药者。

8.2 经安慰剂校正的量级

2024 年一份更严格的元分析给出了目前最可用的数字 [一手逐字]:

“From 6095 articles screened, 79 studies (44 RCTs and 35 observational studies) covering 21 002 patients were selected… 16 532 patients discontinued from an antidepressant, and 4470 patients discontinued from placebo. Incidence of at least one antidepressant discontinuation symptom was 0·31 (95% CI 0·27-0·35) in 62 study groups after discontinuation of antidepressants, and 0·17 (0·14-0·21) in 22 study groups after discontinuation of placebo. Between antidepressant and placebo groups of included RCTs, the summary difference in incidence was 0·08 [0·04-0·12]. The incidence of severe antidepressant discontinuation symptoms after discontinuation of an antidepressant was 0·028 (0·014-0·057) compared with 0·006 (0·002-0·013) after discontinuation of placebo.”

“Considering non-specific effects, as evidenced in placebo groups, the incidence of antidepressant discontinuation symptoms is approximately 15%, affecting one in six to seven patients who discontinue their medication. … our findings can inform clinicians and patients about the probable extent of antidepressant discontinuation symptoms without causing undue alarm.”

来源:Henssler, Schmidt, Schmidt, Schwarzer, Bschor & Baethge, Lancet Psychiatry 11(7):526–535 (2024), doi:10.1016/S2215-0366(24)00133-0(◐ 官方摘要逐字。该文有勘误,并引发五封公开通信,其中包括 Read 与 Davies 的异议原作者答辩——○ 仅题名可核,本篇只承重「交锋存在」)

两侧同时被压住:撤药反应的存在是硬事实(31% vs 17%,差值 8 个百分点,重度 2.8% vs 0.6% 即约五倍风险)——「1 至 2 周自限」的旧口径站不住;但「超过一半人会有撤药反应」这个未校正数字也站不住,经安慰剂校正后是约 15%、约六到七人中一人。并且撤药反应绝大多数不是重度。

8.3 停药的后果:随机对照证据

这是全篇唯一一处需要读者特别注意其临床含义的证据 [一手逐字]:

“By 52 weeks, relapse occurred in 92 of 238 patients (39%) in the maintenance group and in 135 of 240 (56%) in the discontinuation group (hazard ratio, 2.06; 95% confidence interval, 1.56 to 2.70; P<0.001). Secondary outcomes were generally in the same direction as the primary outcome. Patients in the discontinuation group had more symptoms of depression, anxiety, and withdrawal than those in the maintenance group.”

“Among patients in primary care practices who felt well enough to discontinue antidepressant therapy, those who were assigned to stop their medication had a higher risk of relapse of depression by 52 weeks than those who were assigned to maintain their current therapy.”

来源:Lewis et al.(ANTLER 试验),NEJM 385(14):1257–1267 (2021), doi:10.1056/NEJMoa2106356(◐ 官方摘要逐字。150 家英国全科诊所,478 人随机双盲,ISRCTN15969819)

读法要精确:入组者是「有至少两次抑郁发作史或已服药两年以上、且自我感觉好到可以考虑停药」的人;在这群人里,随机分到停药组的一年复发率显著更高。这意味着任何人都不该停药——试验里毕竟有 44% 的停药者一年内没有复发;它意味着停药是一个有实质风险的临床决定,必须与处方者一起做

同时注意依从性数据:维持组 70%、停药组 52%——近一半停药组受试者没能按方案完成,这本身就说明停药在真实条件下有多难。

8.4 指南实际改了什么:现行条文逐字

指南层面的变化不是传闻,可以直接读。英国 NICE 现行抑郁症指南 NG222(2022)在「停用抗抑郁药」一节写着 [一手逐字]:

“Explain that it is usually necessary to reduce the dose in stages over time (called ‘tapering’) but that most people stop antidepressants successfully.”

“withdrawal symptoms can be mild, may appear within a few days of reducing or stopping antidepressant medication, and usually go away within 1 to 2 weeks; withdrawal can sometimes be more difficult, with symptoms lasting longer (in some cases several weeks, and occasionally several months); withdrawal symptoms can sometimes be severe, particularly if the antidepressant medication is stopped suddenly.”

“slowly reduce the dose to zero in a step-wise fashion, at each step prescribing a proportion of the previous dose (for example, 50% of previous dose); consider using smaller reductions (for example, 25%) as the dose becomes lower; if, once very small doses have been reached, slow tapering cannot be achieved using tablets or capsules, consider using liquid preparations if available; ensure the speed and duration of withdrawal is led by and agreed with the person taking the prescribed medication… recognise that withdrawal may take weeks or months to complete successfully.”

来源:NICE, Depression in adults: treatment and management, NG222, 条 1.4.12–1.4.16,nice.org.uk/guidance/ng222(✓ 主笔亲核,2026-07-26 实查)

这段条文本身就是一次完整的双向裁决,而且是官方做的。 它采纳了批评侧的核心技术主张——按比例阶梯减量(这正是 Horowitz 与 Taylor 2019 年在 Lancet Psychiatry 提出的双曲减量方案,○ 仅题名可核,未取回全文)、必要时用液体制剂、承认撤药可持续数周至数月且可能严重;同时保留了对反向叙事的限制——「大多数人能成功停药」、撤药「不影响每个人」。

所以停药层的裁决是:批评侧赢了指南,但没有赢到他们主张的强度。 旧的「1 至 2 周自限」被官方撤下了;「抗抑郁药会上瘾」也没有被官方写进去。


九 双向裁决 · 对称三向红线 · 对称金句

9.1 「抑郁是化学失衡」为什么未立

  • 病因层六个方向无一支持,基因侧是高确定性无效(OR = 1.000, p = 0.994)。
  • 现有的正向影像证据,其自己的引言写着五十年来「no firm in vivo evidence」,且首次直接测量是 n=17 对 20、p = .041、单区域、无独立复现。
  • 最强的辩护方,在最用力的反驳里主张的强度只到「a role」「perturbed」「in a proportion of people」。
  • 源头文献(1967)自陈证据「by no means conclusive」,且明确保留环境因素。
  • ***ex juvantibus* 倒推在逻辑上不成立**,而且这一点 2005 年就印在同行评议期刊上(阿司匹林反例)。
  • 官方口径(APA/NIMH/NICE)在 2026 年都已不用这个说法

9.2 「抗抑郁药是骗局、该停药」为什么同样未立

  • 522 项试验、116 477 人,21 种药全部优于安慰剂;作者名单里有揭露报告偏倚的人和最著名的元研究怀疑论者。
  • 连最批判的独立综述也报「statistically significant effects」,其异议是量级与危害权衡,不是零效应。
  • FDA 个体数据显示一个可辨识的高应答亚群(该分解 2025 年被挑战,但挑战数据无安慰剂臂,状态是未结案)。
  • 有证据表明量表本身在稀释信号(单条目分析把阴性比较从 56% 降到 9%)。
  • 报告偏倚是真的,但在缩小(效应量膨胀 0.10 → 0.05;透明报告 11% → 47%)。
  • 停药有直接随机对照证据的风险:一年复发率 56% 对 39%,HR 2.06。
  • 撤药反应存在但经安慰剂校正约 15%、重度约 2.8%,且官方指南明确写「大多数人能成功停药」。

9.3 独占裁决:那句话的真实身份

「化学失衡」不是一个被推翻的科学理论,它从来没当上过科学理论。 它是一件沟通器物:在源头文献里是一个自陈不确定的少数派假说;在 1990 至 2000 年代的药厂材料里被写成产品机制(GSK 官网原句、Pfizer 电视广告原句);在专业界的教科书与高引综述里以「有保留的支持」形态长期在场(一方举证,另一方未以同类证据回应);在今天的实测里,它主要经由诊室传给病人,而且在病因说不清的时候最容易被启用

因此对这句话的正确处置不是「揭穿它」,而是「替换它顶着的功能」。 它同时顶着三根梁:免除自责、让服药正当、给痛苦一个可说出口的名字。实测数据显示,它连第一根都没顶住(化学失衡反馈没有减少自责),却带来了预后悲观。所以替换它不需要牺牲什么——需要做的是把「可修改性」补进解释里,而这一条是有实验证据支持的(可塑性干预显著降低预后悲观、提升能动感)。

9.4 对称三向红线(最高规格)

一、不升格。 本篇不主张「抑郁与神经化学无关」——这是一个比「化学失衡成立」同等程度的过度主张。证据支持的是:血清素缺乏作为病因没有立住;血清素系统与情绪有关联、并在易感与缓解期人群中可被扰动,这是双方都承认的。「没有证据支持 X」不等于「有证据反对一切生物学解释」。

二、不虚无化。 本篇不主张抗抑郁药无效、不主张精神科是伪科学、不主张任何人应停药。药效是真的,撤药风险是真的,停药复发风险是真的,指南的修订是真的进步。证据审计的目的是让说法与证据对齐,不是让治疗停摆。

三、不作诊疗建议、不评价任何在世研究者。 本篇不评估任何个人的用药,不构成任何临床决定的依据;对所有被引作者只裁其公开文本与数据,不推断动机与人格。两侧的利益申报一律照登:批评侧核心作者有停药服务公司、相关著作版税与倡导组织职务;反驳侧联署名单含药企在职人员;条目层分析作者申报药企顾问与讲者酬金。照登利益申报是为了让读者自己加权,不是为了暗示任何人不诚实。

一条主动的防联想(补在三向之外):本篇与「精神病学是否是医学」这一身份争论无关,也与任何反精神医学的政治主张无关。把「一句话的证据不足」升格成「一个学科的正当性破产」,正是本篇通篇在挡的那种跳跃。

9.5 对称金句

防升格侧(取自为精神科辩护得最有力的那位,同一篇文章里的两句):

“In truth, the ‘chemical imbalance’ notion was always a kind of urban legend – never a theory seriously propounded by well-informed psychiatrists.” “And, yes – the ‘chemical imbalance’ image has been vigorously promoted by some pharmaceutical companies, often to the detriment of our patients’ understanding.” —— Ronald W. Pies(《Psychiatric Times》2007–2010 年主编),2011

防虚无侧(取自这场辩论中最强正面研究自己的背景句,与撤药量级元分析自己的结语):

“Although the hypothesis is over 50 years old, there is no firm in vivo evidence for diminished 5-HT neurotransmission.” —— Erritzoe et al., Biological Psychiatry, 2023(该研究本身是为支持假说而做的)

“…our findings can inform clinicians and patients about the probable extent of antidepressant discontinuation symptoms without causing undue alarm.” —— Henssler et al., Lancet Psychiatry, 2024

两侧合起来正是本篇的母裁决:为精神科辩护的人说这句话从来不是理论;为血清素假说找证据的人说这个假说五十年来没有硬的在体证据;量化撤药风险的人说数字要让人知情、但不要引起不必要的恐慌。三句话里没有一句支持「抑郁是化学失衡」,也没有一句支持「抗抑郁药是骗局」。


十 自指:这一篇自己的软处

一、本库自己的机制叙事有同一种病。 PHD(预测性内稳态调度)、「距临界距离」、「调度带宽」这些构念在本库里承担了大量解释工作,而它们的经验支持强度未必高于「血清素系统在抑郁中被扰动」。本篇裁别人「机制故事跑在证据前面」时,落刀必须同样落在自己身上——睡眠架构篇刚刚就 PHD 执行层给出「只兼容、不确证」的裁决,那是同一把刀。

二、本篇的核心数字全部以一个代理指标为单位。 1.75 点、1.97 点、−1.94 点、3 点阈值——全是 HAMD-17 的刻度。本库测量代理性篇警告过的正是这件事:当一个为方便评分而造的量表变成「疗效」本身的单位,关于它的争论就无法只靠更多数据解决。 第 2.6 节那条反向证据(换单条目就把阴性比较从 56% 降到 9%)说明这不是理论顾虑。本篇没有能力裁定哪把尺子对,只能把尺子的存在说清楚。

三、四路收敛不是四次独立测量。 已在 2.5 主动降级,此处再记一次,因为这是本篇最容易被引用者放大的地方。

四、批评侧作者高度重叠。 Moncrieff、Horowitz、Hengartner、Plöderl、Kirsch、Pigott 这六个名字反复出现在本篇的伞状综述、影像批评通信、STAR*D 再分析、三峰复现失败等多个关键位置。这是一个紧密的作者群,不是多个独立来源。 本篇的做法是:凡这一侧的结论,都尽量配一份非该群体的原件(Cipriani/Stone/Turner/Henssler/NICE/Schroder 等),并在引用时标注单侧集中。反驳侧同样是一个紧密网络(Jauhar 等三十余人联署且多来自英国少数几家机构),处理方式相同。

五、本篇的检索工具本身有偏。 本会话的内置网页搜索工具持续报错不可用(见排障记录),发现层改用 Europe PMC/PubMed/Crossref/OpenAlex/Unpaywall 的官方接口与 Wayback。这套工具对被索引的期刊文献覆盖很好,对非索引材料(专业媒体评论、机构声明、灰色文献)覆盖差。 本篇因此可能系统性低估了「机构与媒体层的表态」这一类证据——例如 6.4 那处「无第三方独立内容分析」的空位,有一部分可能是工具造成的,而不是文献真的不存在。这一条我无法自己修正,只能标明。

六、本篇没有做、且明确不做的事。 不裁抑郁的诊断标准;不裁「该不该用药」这个临床问题;不裁精神科的建制正当性;不给任何个人任何建议。本篇能承担的全部责任是:让每一句关于证据的话,都能被读者用链接自己走回原件。


十一 后续问题 · 不确定点 · 可信度 · 关联笔记

11.1 后续问题

  1. 2022 年之后有没有新的公众信念调查? 这是本篇最想要而没有的数据。官方口径已经退却,公众版本有没有跟着动,目前无人测量。
  2. 诊室渠道能不能被干预? 2024 与 2026 那三项研究都指向医疗提供者,而 2026 那项还指出触发条件是「病因说不清」。针对提供者的沟通训练是否能同时保住「不自责」和「不悲观」这两个目标,是可做的实验。
  3. Stone 的高应答亚群到底在不在? 需要在有安慰剂臂的个体层数据上重做混合模型复现,而不是在单臂开放数据上。
  4. HAMD-17 该不该被换掉? 若条目层分析成立,过去三十年所有效应量估计的分母都需要重算。这件事的影响远超本课题。
  5. 长期用药是否反向压低血清素? 伞状综述提出这一可能(血浆血清素降低与抗抑郁药使用相关,另有 5-HIAA 长期治疗后降低的证据),并援引「对立性耐受」模型(Fava 2020,○ 仅题名可核)。这是一个可被前瞻性设计直接检验的问题。

11.2 不确定点(本篇明确不承重的部分)

  • **STAR*D 原研究者是否有公开答辩**:未取回,故第三节只承重算术,不推断动机。
  • 学界渠道是否也在传播那句话:一方举证,无第三方同类分析,状态是「未结案」。
  • 反向叙事的危害:无量化研究,本篇拒绝为求对称而虚构。
  • Erritzøe 交锋的具体论证:三份通信均付费墙,只承重「交锋存在」。
  • Schildkraut 1965 原文:未取回(○);其为去甲肾上腺素版本这一点经 2005 年论文全文确认。
  • 2025 年 medRxiv 两份 RIAT 延伸再分析:预印本,未经同行评议,不作承重。
  • 3 点临床显著阈值:各方自设,本篇不裁其对错。

11.3 可信度

  • Overall:high(承重层几乎全为官方原文逐字或官方摘要逐字,四十余条来源全部挂链接可复核)。
  • 最硬的部分:药效存在与其量级(③④);病因层六域不支持(⑥);撤药量级(⑧);「化学失衡」的药厂渠道(争议双方一致承认);机构现行表述(今日实查)。
  • 框架内硬:四路点差收敛(数据重叠,已降级)。
  • 未结案:量级的临床意义;高应答亚群是否存在;学界渠道之争。
  • 诚实空位:2022 后公众信念;反向叙事危害;STAR*D 原作者答辩。

11.4 关联笔记


来源清单

第一层 · 主笔亲核全文或全页逐字(✓)

  1. Moncrieff, Cooper, Stockmann, Amendola, Hengartner & Horowitz. The serotonin theory of depression: a systematic umbrella review of the evidence. Molecular Psychiatry 28(8):3243–3256, 2022. https://doi.org/10.1038/s41380-022-01661-0https://pmc.ncbi.nlm.nih.gov/articles/PMC10618090/
  2. Jauhar, Arnone, Baldwin, Bloomfield, Browning, … Cowen. A leaky umbrella has little value: evidence clearly indicates the serotonin system is implicated in depression. Molecular Psychiatry 28(8):3149–3152, 2023. https://doi.org/10.1038/s41380-023-02095-yhttps://pmc.ncbi.nlm.nih.gov/articles/PMC10618084/
  3. Lacasse & Leo. Serotonin and depression: a disconnect between the advertisements and the scientific literature. PLoS Medicine 2(12):e392, 2005. https://doi.org/10.1371/journal.pmed.0020392https://pmc.ncbi.nlm.nih.gov/articles/PMC1277931/
  4. Pies. Psychiatry’s New Brain-Mind and the Legend of the “Chemical Imbalance”. Psychiatric Times, 2011. 存档:http://web.archive.org/web/20210121061206/https://www.psychiatrictimes.com/view/psychiatrys-new-brain-mind-and-legend-chemical-imbalance
  5. GlaxoSmithKline. What does Paxil treat(官方产品站页面). 存档:http://web.archive.org/web/20021015100904/http://www.paxil.com/about/ab_trt.html
  6. Cipriani et al. Comparative efficacy and acceptability of 21 antidepressant drugs… The Lancet 391(10128):1357–1366, 2018. https://doi.org/10.1016/S0140-6736%2817%2932802-7https://pmc.ncbi.nlm.nih.gov/articles/PMC5889788/
  7. Stone et al. Response to acute monotherapy for major depressive disorder in randomized, placebo controlled trials submitted to the US FDA. BMJ 378:e067606, 2022. https://doi.org/10.1136/bmj-2021-067606https://pmc.ncbi.nlm.nih.gov/articles/PMC9344377/
  8. Turner et al. Selective publication of antidepressant trials and its influence on apparent efficacy: Updated comparisons and meta-analyses. PLoS Medicine 19(1):e1003886, 2022. https://doi.org/10.1371/journal.pmed.1003886https://pmc.ncbi.nlm.nih.gov/articles/PMC8769343/
  9. Jakobsen et al. Selective serotonin reuptake inhibitors versus placebo in patients with major depressive disorder. BMC Psychiatry 17:58, 2017. https://doi.org/10.1186/s12888-016-1173-2https://pmc.ncbi.nlm.nih.gov/articles/PMC5299662/
  10. Kirsch et al. Initial severity and antidepressant benefits: a meta-analysis of data submitted to the FDA. PLoS Medicine 5(2):e45, 2008. https://doi.org/10.1371/journal.pmed.0050045https://pmc.ncbi.nlm.nih.gov/articles/PMC2253608/(含期刊编辑部「Editors’ Summary」中的化学失衡表述)
  11. Pigott et al. What are the treatment remission, response and extent of improvement rates after up to four trials of antidepressant therapies in real-world depressed patients? A reanalysis of the STARD study. BMJ Open* 13(7):e063095, 2023. https://doi.org/10.1136/bmjopen-2022-063095https://pmc.ncbi.nlm.nih.gov/articles/PMC10373710/
  12. Schroder, Tovey, Forer, Schultz, Kneeland & Moser. Where do “chemical imbalance” beliefs come from? Evaluating the impact of different sources. Frontiers in Psychology 15:1469913, 2024. https://doi.org/10.3389/fpsyg.2024.1469913https://pmc.ncbi.nlm.nih.gov/articles/PMC11752450/
  13. Coppen. The Biochemistry of Affective Disorders. British Journal of Psychiatry 113(504):1237–1264, 1967. https://doi.org/10.1192/bjp.113.504.1237 (Cambridge Core Extract 段逐字)
  14. NICE. Depression in adults: treatment and management, NG222, 条 1.4.10–1.4.17. https://www.nice.org.uk/guidance/ng222/chapter/Recommendations (2026-07-26 实查)
  15. American Psychiatric Association. What Is Depression? https://www.psychiatry.org/patients-families/depression/what-is-depression (2026-07-26 实查)
  16. National Institute of Mental Health. Depression. https://www.nimh.nih.gov/health/publications/depressionhttps://www.nimh.nih.gov/health/topics/depression (2026-07-26 实查)
  17. Read, Renton, Harrop, Geekie & Dowrick. A survey of UK general practitioners about depression, antidepressants and withdrawal. Ther Adv Psychopharmacol 10, 2020. https://doi.org/10.1177/2045125320950124https://pmc.ncbi.nlm.nih.gov/articles/PMC7457636/
  18. Moncrieff et al. The serotonin hypothesis of depression: both long discarded and still supported? Molecular Psychiatry 28(8):3160–3163, 2023. https://doi.org/10.1038/s41380-023-02094-z (首段可见部分逐字;其后在付费墙内)

第二层 · 官方摘要或官方元数据逐字(◐)

  1. Turner, Matthews, Linardatos, Tell & Rosenthal. Selective publication of antidepressant trials and its influence on apparent efficacy. NEJM 358(3):252–260, 2008. https://doi.org/10.1056/NEJMsa065779
  2. Munkholm, Paludan-Müller & Boesen. Considering the methodological limitations in the evidence base of antidepressants for depression: a reanalysis of a network meta-analysis. BMJ Open 9(6):e024886, 2019. https://doi.org/10.1136/bmjopen-2018-024886
  3. Hieronymus, Emilsson, Nilsson & Eriksson. Consistent superiority of SSRIs over placebo in reducing depressed mood in patients with major depression. Molecular Psychiatry 21(4):523–530, 2016. https://doi.org/10.1038/mp.2015.53
  4. Ruhé, Mason & Schene. Mood is indirectly related to serotonin, norepinephrine and dopamine levels in humans: a meta-analysis of monoamine depletion studies. Molecular Psychiatry 12(4):331–359, 2007. https://doi.org/10.1038/sj.mp.4001949
  5. Erritzoe et al. Brain Serotonin Release Is Reduced in Patients With Depression: A [11C]Cimbi-36 PET Study With a d-Amphetamine Challenge. Biological Psychiatry 93(12):1089–1098, 2023. https://doi.org/10.1016/j.biopsych.2022.10.012
  6. Rush et al. Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STARD report. Am J Psychiatry* 163(11):1905–1917, 2006. https://doi.org/10.1176/ajp.2006.163.11.1905
  7. Xu, Naudet, Kim, Hengartner, Horowitz, Kirsch, Moncrieff, Pigott & Plöderl. Large responses to antidepressants or methodological artifacts? A secondary analysis of STARD. J Clin Epidemiol* 187:111943, 2025. https://doi.org/10.1016/j.jclinepi.2025.111943
  8. Henssler, Schmidt, Schmidt, Schwarzer, Bschor & Baethge. Incidence of antidepressant discontinuation symptoms: a systematic review and meta-analysis. Lancet Psychiatry 11(7):526–535, 2024. https://doi.org/10.1016/S2215-0366%2824%2900133-0 (勘误 https://doi.org/10.1016/S2215-0366%2824%2900253-0
  9. Davies & Read. A systematic review into the incidence, severity and duration of antidepressant withdrawal effects: Are guidelines evidence-based? Addictive Behaviors 97:111–121, 2019. https://doi.org/10.1016/j.addbeh.2018.08.027
  10. Lewis et al. Maintenance or Discontinuation of Antidepressants in Primary Care(ANTLER). NEJM 385(14):1257–1267, 2021. https://doi.org/10.1056/NEJMoa2106356
  11. Pilkington, Reavley & Jorm. The Australian public’s beliefs about the causes of depression: associated factors and changes over 16 years. J Affect Disord 150:356–362, 2013. https://doi.org/10.1016/j.jad.2013.04.019
  12. Pescosolido et al. “A disease like any other”? A decade of change in public reactions to schizophrenia, depression, and alcohol dependence. Am J Psychiatry 167:1321–1330, 2010. https://doi.org/10.1176/appi.ajp.2010.09121743
  13. Kemp, Lickel & Deacon. Effects of a chemical imbalance causal explanation on individuals’ perceptions of their depressive symptoms. Behav Res Ther 56:47–52, 2014. https://doi.org/10.1016/j.brat.2014.02.009
  14. Lebowitz, Ahn & Nolen-Hoeksema. Fixable or fate? Perceptions of the biology of depression. J Consult Clin Psychol 81:518–527, 2013. https://doi.org/10.1037/a0031730
  15. Schroder, Russman Block & Moser. Chemical imbalance and etiological beliefs about depression among college students. J Am Coll Health 72(7):1993–2000, 2024. https://doi.org/10.1080/07448481.2022.2098037
  16. Schroder & Bommarito. Medical providers and biogenetic messages about depression: A vignette experiment. Patient Educ Couns 144:109463, 2026. https://doi.org/10.1016/j.pec.2025.109463
  17. Ang, Horowitz & Moncrieff. Is the chemical imbalance an “urban legend”? SSM – Mental Health 2:100098, 2022. https://doi.org/10.1016/j.ssmmh.2022.100098 (摘要经 http://web.archive.org/web/20250108134417/https://www.sciencedirect.com/science/article/pii/S266656032200038X 取回)
  18. Lepri. The myth of serotonin theory of depression: an analysis of umbrella review methodologies and clinical implications. Psychiatr Danub 37(Suppl 1):187–188, 2025. https://pubmed.ncbi.nlm.nih.gov/40982824/
  19. Wu, Melin, Björkblom & Sjöberg. Addressing the serotonin hypothesis of depression through analyses of genetics, methylation and metabolite variations in glioma patients. Scientific Reports 15, 2025. https://doi.org/10.1038/s41598-025-25464-9

第三层 · 仅题名与元数据可核,不作承重(○)

  1. Schildkraut. The catecholamine hypothesis of affective disorders: a review of supporting evidence. Am J Psychiatry 122(5):509–522, 1965. https://doi.org/10.1176/ajp.122.5.509
  2. Border et al. No Support for Historical Candidate Gene or Candidate Gene-by-Interaction Hypotheses for Major Depression Across Multiple Large Samples. Am J Psychiatry 176:376–387, 2019. https://doi.org/10.1176/appi.ajp.2018.18070881
  3. Culverhouse et al. Collaborative meta-analysis finds no evidence of a strong interaction between stress and 5-HTTLPR genotype… Molecular Psychiatry 23:133–142, 2018. https://doi.org/10.1038/mp.2017.44
  4. Hengartner & Plöderl. No Clear Evidence of Reduced Brain Serotonin Release Capacity in Patients With Depression. Biological Psychiatry 93(12):e61, 2023. https://doi.org/10.1016/j.biopsych.2022.11.020
  5. Rabiner, Agnorelli, Howes, Nutt, Cowen & Erritzoe. Reply to: No Clear Evidence of Reduced Brain Serotonin Release Capacity… Biological Psychiatry 93(12):e63–e64, 2023. https://doi.org/10.1016/j.biopsych.2022.11.021
  6. Slifstein & Abi-Dargham. Detecting Pharmacologically Induced Serotonin Release in Depression With PET Imaging: A New Approach. Biological Psychiatry 93(12):1056–1058, 2023. https://doi.org/10.1016/j.biopsych.2023.04.008
  7. Horowitz & Taylor. Tapering of SSRI treatment to mitigate withdrawal symptoms. Lancet Psychiatry 6(6):538–546, 2019. https://doi.org/10.1016/S2215-0366%2819%2930032-X
  8. Read & Davies. Incidence of antidepressant withdrawal symptoms(通信). Lancet Psychiatry 11(10):788–789, 2024. https://doi.org/10.1016/S2215-0366%2824%2900273-6 | 原作者答辩 https://doi.org/10.1016/S2215-0366%2824%2900287-6
  9. Hengartner & Plöderl. Estimates of the minimal important difference to evaluate the clinical significance of antidepressants… BMJ Evid Based Med 27:69–73, 2022. https://doi.org/10.1136/bmjebm-2020-111600
  10. Fava. May antidepressant drugs worsen the conditions they are supposed to treat? The clinical foundations of the oppositional model of tolerance. Ther Adv Psychopharmacol 10, 2020. https://doi.org/10.1177/2045125320970325
  11. Bartova, Lanzenberger, Rujescu & Kasper. Reply to: “The serotonin theory of depression…” Molecular Psychiatry 28(8):3153–3154, 2023. https://doi.org/10.1038/s41380-023-02093-0
  12. Jacobsen. Serotonin and depression – an alternative interpretation of the data in Moncrieff et al. Molecular Psychiatry 28(8):3158–3159, 2023. https://doi.org/10.1038/s41380-023-02090-3
  13. El-Mallakh, Doroodgar, Elsayed & Kidambi. The serotonin theory of depression. Molecular Psychiatry 28(8):3157, 2023. https://doi.org/10.1038/s41380-023-02091-2
  14. Almulla & Maes. Although serotonin is not a major player in depression, its precursor is. Molecular Psychiatry 28(8):3155–3156, 2023. https://doi.org/10.1038/s41380-023-02092-1
  15. Priest, Vize, Roberts, Roberts & Tylee. Lay people’s attitudes to treatment of depression: results of opinion poll for Defeat Depression Campaign just before its launch. BMJ 313:858–859, 1996. https://doi.org/10.1136/bmj.313.7061.858
  16. Pigott et al. Restoring STAR*D(两份 RIAT 延伸再分析,medRxiv 预印本). https://doi.org/10.1101/2025.02.10.25321991https://doi.org/10.1101/2025.10.27.25338365

第四层 · 检索与元数据渠道

  1. Europe PMC REST API(检索与全文 XML);PubMed E-utilities(官方摘要);Crossref REST API(DOI 元数据与卷期页码);OpenAlex 与 Unpaywall(开放获取定位);Internet Archive Wayback CDX 与快照(历史页面)。本篇所有 DOI 的卷期页码均经 Crossref 或 Europe PMC 实查,未凭记忆填写;两处带括号的 DOI 已按本库排障记录百分号编码并逐一测试返回 200。

纪律声明

本篇不凭记忆写。四十余条来源中,十八条为主笔取回官方全文或官方页面逐字比对,十九条为官方摘要逐字,其余仅题名与元数据可核并明确不作承重。全部承重引用挂原始链接。核对过程中发现的三处问题已随文注明:伞状综述引言一处引用与被引原文结论相反(6.5);「80% 公众相信化学失衡」的实际调查结构是多因归因而非单因(6.7);四路效应量收敛的数据高度重叠、不构成独立复现(2.5,本篇主动降级)。本篇不裁抑郁的诊断与本体,不裁临床决策,不作任何诊疗建议,尤其不构成停药建议;对所有被引作者只裁公开文本与数据,两侧利益申报一律照登。

〔机制裁决第 94 篇 · 对称双向第 89 篇 · section D 意识 / 心智 / 神经 · 全库第 152 篇 · 作者 Claude Opus 5〕