克服成瘾：物质、行为，以及从中牟利的产业
摘要
斯坦福大学精神病学教授 Keith Humphreys 博士就物质与行为成瘾提供了一个全面的分析框架——涵盖酒精、大麻、赌博到阿片类药物。他阐释了驱动成瘾的遗传、神经学及商业因素，并概述了最有效的循证戒瘾策略。
核心要点
- 成瘾的定义在于有害行为的持续性，而非频率——即便明显已毁掉自己的生活仍不停止，这才是成瘾的标志。
- 遗传风险真实存在且影响显著（遗传率估计为 0.3–0.5），而预测酒精使用障碍最有力的单一指标，是父母是否患有同样问题。
- 饮酒对心脏的任何潜在益处都被癌症风险所抵消——就整体健康而言，滴酒不沾优于任何饮酒量，尽管每周两杯的绝对风险极小。
- 现代大麻的效力已大幅提升——平均 THC 含量从约 4% 升至约 20%，与 1980 年代的典型使用模式相比，有效暴露量提高了约 65 倍。
- 大麻存在真实的精神病风险，尤其对年轻人以及有精神分裂症、双相障碍或相关疾病一级亲属的人群而言。
- 赌博成瘾是被精心设计出来的，机器围绕”伪装成赢利的亏损”和新奇刺激而设计——用户实际上并非在玩钱赢钱。
- 以成瘾牟利的产业刻意将成瘾潜力最大化；美国约 10% 的人口消费了大约一半的酒精——重度用户才是其核心商业模式。
- 社群与问责机制是克服任何成瘾或行为改变最有力的工具之一。
- 早期物质使用（14 岁之前）会大幅提升各类物质的终身成瘾风险。
- **GLP-1 激动剂与裸盖菇素（psilocybin）**被列为成瘾治疗领域最具前景的两大新兴药理方向。
详细笔记
成瘾的定义
- 成瘾并非单纯指频繁或强迫性地做某件事
- 其决定性特征是：在任何理性旁观者都会停止的情况下，仍继续做到自我毁灭的程度
- 经典案例：James Olds 实验中的大鼠，在食物就在旁边的情况下，仍不断自我刺激大脑直至饿死
- Addiction（成瘾）涉及奖励体验的逐步收窄——自然奖励（人际关系、工作、住所）逐渐消失，只剩物质成为唯一的快感来源
- 这解释了为何人们即便在客观上遭受伤害，仍紧抓物质不放：它可能是唯一仍能带来奖励的东西
成瘾术语
- **酒精使用障碍（AUD）**是一个更宽泛的谱系——轻度、中度或重度
- 只有重度 AUD 才类似于大多数人所认知的”成瘾”
- 轻度 AUD 的设计初衷是为了更早介入临床治疗，类似于治疗高血压前期
- “酗酒”（Alcoholism）在科学上仍具意义，且在康复社群（如 AA）中更为通用
遗传风险因素
- 成瘾的遗传率相当可观——跨物质的估计大约在 0.3 至 0.5 之间
- 酗酒父母的被收养子女，即便由不饮酒者抚养长大，酒精问题发生率也显著更高
- 父传子是研究中发现的最强遗传关联
- 某些遗传风险具有物质特异性（例如，缺乏代谢酒精所需酶的人，对酒精的愉悦感降低，风险随之降低）
- 其他遗传风险则跨物质共享——冲动性和寻求刺激等特质会提升多种物质的成瘾风险
- 遗传负荷若无接触暴露则无从体现——风险只有通过行为才会显现
- 长期康复中的人有时会出现新的成瘾行为（如性冲动障碍、暴食），这是潜在倾向寻找新出口的表现
主观反应与易感性
- 一小部分人（约 8–10%）体验到酒精具有高度的刺激性和欣快感，同时躯体反馈减弱（身体摇晃感减少、宿醉反应更少）——这一发现来自 Mark Schuckit 对酗酒父亲之子的研究
- 这些人缺乏正常的”停止信号”，在感受到负面后果之前可以喝更多
- 这种惩罚减少、奖励增加的模式在多种物质中都有体现，且在很大程度上由遗传决定
- 不同个体对不同药物的反应差异极大（例如，有人觉得阿片类药物令人极度不适，有人则在第一次使用时便感到一种深刻的完整感）
酒精：健康风险与营销现实
- J 形曲线（暗示不饮酒者死亡率高于少量饮酒者）存在混淆因素——包括戒酒的前重度饮酒者，以及已患病后戒酒的人群
- 红酒的健康益处从未有坚实的科学依据——葡萄酒中白藜芦醇含量极微，不足以产生效果；这一说法源于 1990 年代一档《60 Minutes》节目，后被酒精行业大力放大
- 酒精对心脏的任何潜在益处都小于其癌症风险——饮酒带来净死亡率获益并不存在
- 每周两标准杯的风险极小但不为零
- 美国 10% 的饮酒者消费了约一半的酒精销量——该行业的盈利依赖于这一重度用户群体，由此形成扩大并维持成瘾的结构性动机
酒精与女性
- 从 1990 年代末开始，酒精行业刻意针对女性开展营销（例如”妈妈喝酒文化”），甚至人为制造看似自发形成的网络社群
- 女性饮酒量因此大幅增加
- 女性每饮一杯所受的身体伤害大于男性，原因在于体型差异及可能的激素因素
- 乳腺癌、卵巢癌及其他激素相关癌症的风险因饮酒而特别升高——这一信息已被证明对于激励女性减少或停止饮酒尤为有效
饮酒的社会动态
- 文化上存在一种不对等：人们鲜少被问及为何喝酒，却常被问及为何不喝——解释的负担落在了戒酒者身上
- 健康信息（尤其是癌症风险）在戒酒通常需要理由的环境中，充当了一种社会许可，让人得以公开拒绝
- 酒精具有抗焦虑作用，这在很大程度上解释了其社交价值——部分人在喝了一杯后，因焦虑减轻而真的更善于社交
大麻：效力、风险与政策
- 平均 THC 含量已从 1980–90 年代的 3–5% 上升至今日的约 20%
- 约 42% 的当前大麻使用者每天或几乎每天使用（而历史上的典型频率是每周一两次）
- 综合效应：现代典型使用量对大脑的暴露约为 1980 年代典型使用量的 65 倍——效力比相当于可卡因叶与纯可卡因之间的差距
- Cannabis use disorder（大麻使用障碍）的风险和危害远超上一代人的经验所提示的程度
- 去罪化（取消对使用者的处罚）对使用率的影响极小；合法化（商业生产与营销）则会显著提高消费量
- 大麻是一种损害表现的药物——长期使用会损害短期记忆、专注力、对细节的关注以及动机
- “无法起飞”（即无力维持独立的成年生活）是青少年重度使用者中有据可查的一种模式
- Cannabis-induced psychosis（大麻诱发精神病）的风险是真实存在的，高效能大麻的最新数据进一步强化了这一结论——尤其对青少年和年轻成人而言
- 若一级亲属中有精神分裂症、双相障碍或分裂型人格障碍患者，请勿使用大麻
- CBD（非致幻成分）具有合法的医疗应用，包括一款获 FDA 批准的儿科癫痫发作治疗药物
大麻食品与吸食大麻
- 大麻食品的起效更慢、更难把握——用户常在感受到效果之前摄入过多
- 合法市场中的大麻产品往往生产不均匀，导致剂量难以预测
- 关于吸食者能有效自我调节 THC 摄入量的说法，缺乏实验室证据支持
大麻作为入门毒品
- 所有物质都可能充当入门毒品——大麻具有独特性的说法是一种歪曲
- 入门机制包括：享受某种物质后进一步探索其他物质；社交圈向使用毒品的同伴转移；以及可能存在的神经敏化效应，使后续物质的奖励感增强
- 就大多数年轻人而言，酒精在统计上是更大的入门风险，但在公众话语中并不被视为毒品——这一框架对酒精行业有利
赌博成瘾
- 现代老虎机围绕新奇刺激而设计，而非赢钱——用户被新奇的符号组合所刺激，而非被真正的财务奖励所驱动
- “伪装成赢利的亏损”（LDWS）：即便净结果是亏损，机器也会将部分返还庆祝为”赢了”，从而维持
English Original 英文原文
Overcoming Addiction: Substances, Behaviors, and the Industries That Profit From Them
Summary
Dr. Keith Humphreys, professor of psychiatry at Stanford, provides a comprehensive framework for understanding addiction across substances and behaviors — from alcohol and cannabis to gambling and opioids. He explains the genetic, neurological, and commercial forces that drive addiction, and outlines the most effective evidence-based strategies for overcoming it.
Key Takeaways
- Addiction is defined by persistence of harmful behavior, not just frequency — continuing despite obvious destruction of one’s life is the hallmark.
- Genetic risk is real and large (heritability estimated at 0.3–0.5), and the single best predictor of alcohol use disorder is whether a parent had it.
- Any cardiac benefit from alcohol is outweighed by cancer risk — zero alcohol is better than any for overall health, though two drinks per week carries very small absolute risk.
- Modern cannabis is dramatically stronger — average THC content has risen from ~4% to ~20%, making it roughly 65x more potent in effective exposure compared to typical 1980s use patterns.
- Cannabis carries a real psychosis risk, especially for young people and those with first-degree relatives with schizophrenia, bipolar disorder, or related conditions.
- Gambling addiction is engineered, with machines designed around “losses disguised as wins” and novelty stimulation — users are not actually playing to win money.
- Addiction-for-profit industries deliberately maximize addictive potential; ~10% of the U.S. population drinks roughly half of all alcohol consumed — heavy users are the core business model.
- Community and accountability are among the most powerful tools for overcoming any addiction or behavior change.
- Early substance use (before age 14) dramatically increases lifetime addiction risk across all substances.
- GLP-1 agonists and psilocybin are cited as the two most promising emerging pharmacological frontiers for addiction treatment.
Detailed Notes
Defining Addiction
- Addiction is not simply doing something frequently or compulsively
- The defining feature is continuing behavior to the point of self-destruction when any rational observer would stop
- Classic example: rats in James Olds studies self-stimulating their brains while starving next to food
- Addiction involves a progressive narrowing of rewarding experiences — natural rewards (relationships, work, housing) fall away, leaving the substance as the only remaining source of pleasure
- This explains why people cling to substances even when objectively harmful: it may be the only thing still producing reward
Addiction Terminology
- Alcohol use disorder (AUD) is a broader spectrum — mild, moderate, or severe
- Only severe AUD resembles what most people recognize as addiction
- Mild AUD was designed for earlier clinical intervention, similar to treating pre-hypertension
- “Alcoholism” remains scientifically meaningful and is preferred in recovery communities (e.g., AA)
Genetic Risk Factors
- Genetic heritability of addiction is substantial — roughly 0.3 to 0.5 across substances
- Adopted children of alcoholic parents show significantly higher rates of alcohol problems even when raised by non-drinkers
- Father-to-son transmission is the strongest genetic link identified in studies
- Some genetic risk is substance-specific (e.g., lacking the enzyme to metabolize alcohol reduces alcohol’s pleasure and thus risk)
- Other genetic risk is cross-substance — traits like impulsivity and sensation-seeking raise risk across multiple substances
- Genetic loading without exposure is irrelevant — the risk only manifests through behavior
- People in long-term recovery sometimes develop new addictive behaviors (e.g., sexual compulsions, binge eating) as the underlying predisposition finds a new outlet
Subjective Response and Vulnerability
- A subset of people (~8–10%) experience alcohol as highly stimulating and euphoric with reduced physical feedback (less body sway, fewer hangovers) — research by Mark Schuckit on sons of alcoholic fathers
- These individuals lack normal “stop signals” and can drink far more before negative consequences register
- This pattern — less punishment, more reward — is seen across substances and is largely genetic
- Individual responses to different drugs vary enormously (e.g., someone who finds opioids deeply unpleasant vs. someone who feels a profound sense of wholeness from first use)
Alcohol: Health Risks and the Marketing Reality
- The J-shaped curve (suggesting non-drinkers have higher mortality than light drinkers) is confounded by former heavy drinkers and those already ill who quit
- Red wine benefits were never scientifically sound — resveratrol levels in wine are too trace to matter; the narrative originated from a 1990s 60 Minutes segment and was amplified by the alcohol industry
- Any potential cardiac benefit from alcohol is smaller than the cancer risk — net mortality benefit from drinking does not exist
- Two standard drinks per week carries very small but nonzero risk
- 10% of U.S. drinkers consume approximately half of all alcohol sold — the industry’s profitability depends on this heavy-use group, creating a structural incentive to expand and maintain addiction
Alcohol and Women
- Starting in the late 1990s, the alcohol industry deliberately targeted women with marketing (e.g., “mommy wine culture”), including engineering apparently organic online communities
- Women’s drinking increased substantially as a result
- Alcohol causes more damage per drink in women than men, due to body size and likely hormonal factors
- Breast cancer, ovarian cancer, and other hormone-related cancers are specifically elevated by alcohol — and this information has proven particularly effective in motivating women to reduce or stop drinking
Social Dynamics of Drinking
- There is a cultural asymmetry: people are rarely asked why they are drinking, but frequently asked why they are not — placing the burden of explanation on those who abstain
- Health information (especially cancer risk) functions as social permission to quit in environments where abstinence otherwise requires justification
- Alcohol is anxiolytic, which explains much of its social utility — some people are genuinely more socially engaged after one drink due to reduced anxiety
Cannabis: Potency, Risk, and Policy
- Average THC content has risen from 3–5% in the 1980s–90s to approximately 20% today
- ~42% of current cannabis users use it every day or nearly every day (vs. a modal once or twice per week historically)
- Combined effect: modern typical use represents approximately 65x the brain exposure of typical 1980s use — the same potency ratio as between coca leaf and cocaine
- Cannabis use disorder risk and harms are substantially higher than previous generations’ experience suggests
- Decriminalization (removing penalties for users) has minimal effect on use rates; legalization (commercial production and marketing) substantially increases consumption
- Cannabis is a “performance-degrading drug” — regular use undermines short-term memory, concentration, attention to detail, and motivation
- “Failure to launch” (inability to sustain independent adult life) is a documented pattern among heavy adolescent users
- Cannabis-induced psychosis risk is real, strengthened by newer high-potency data — particularly in adolescents and young adults
- Do not use cannabis if you have first-degree relatives with schizophrenia, bipolar disorder, or schizoid personality disorder
- CBD (non-intoxicating component) has legitimate medical applications, including an FDA-approved medication for pediatric seizure disorders
Edibles vs. Smoked Cannabis
- Edibles have slower, harder-to-gauge onset — users frequently consume too much before effects register
- Cannabis products in legal markets are often unevenly manufactured, creating unpredictable dosing
- Claims that smokers self-regulate THC intake effectively are not supported by lab evidence
Cannabis as a Gateway
- All substances can act as gateway drugs — the claim that cannabis is uniquely so was a distortion
- Gateway mechanisms include: enjoying the substance and exploring others; social network shifts toward drug-using peers; possible neurological sensitization that increases reward from subsequent substances
- Alcohol is statistically the greater gateway risk for most young people, but is not framed as a drug in public discourse — a framing that benefits the alcohol industry
Gambling Addiction
- Modern slot machines are engineered around novelty stimulation, not winning — users are stimulated by novel symbol combinations, not actual financial reward
- “Losses disguised as wins” (LDWS): machines celebrate partial returns as wins even when the net result is a loss, sust