体质多基因风险评分:从先天禀赋到可量化的因果链

Written by

in

Language

Evidence level: A

中医所说的”先天禀赋”,在现代遗传流行病学中可以被多基因风险评分(PRS)精确量化——而且从基因→体质→疾病的因果链,正被纵向队列与孟德尔随机化方法逐步打通。

正文

中医体质学说有个核心主张:人的体质差异一部分”禀受于先天”,出生时就已经写进身体里,决定你容易得什么病、对什么环境敏感。过去这句话只能停留在经验层面——”先天禀赋”看不见摸不着,更像玄学。但过去十年遗传流行病学的发展,恰恰给这个古老概念配上了量化的尺子:多基因风险评分(PRS)。

PRS 的思路很直白:复杂疾病不是单个基因决定的,而是成百上千个微小效应的基因变异累加的结果。把一个人携带的所有风险等位基因按效应大小加权求和,得到一个分数,就是这个人的”基因彩票”数字。芬兰 FinnGen 队列 13.5 万人的研究给出了这个数字的威力:与平均 PRS 相比,高 PRS 人群的终身疾病风险高出 21%–38%,冠心病、糖尿病、房颤等五种常见病的发病年龄平均提前 4–9 年(PMID 32273609)。换句话说,”先天禀赋”在多大程度上影响你什么时候得病,现在可以直接算出数字。

这个逻辑放到中医体质上,会产生一个更深的发现。台湾生物库 11,030 人的一项研究,把抑郁症的多基因风险评分(MDD PRS)与中医体质分类(BCQ 量表)放在同一个模型里检验(PMID 39257026)。结果有三个层次:第一,MDD PRS 每升高一个标准差,阳虚体质(OR 1.07)、阴虚体质(OR 1.07)、血瘀体质(OR 1.06)的几率都显著上升——基因风险高的个体,确实更倾向于呈现”偏颇体质”;第二,阳虚体质(OR 2.07)和血瘀体质(OR 1.65)本身显著增加抑郁症风险;第三,也是最关键的一步——中介分析显示,MDD 多基因风险对抑郁症发病的效应,有 10.21% 经由阳虚体质传导、8.41% 经由血瘀体质传导。这是”基因→体质→疾病”因果链的直接统计证据:体质不只是疾病的相关标记,它可能是基因风险向疾病转化路径上的一个真实中间节点。

那么”体质是不是病因”这个问题,还能再进一步吗?这里需要引入孟德尔随机化(MR)。MR 利用一个天然实验:基因变异在受精时随机分配,不受后天环境、生活习惯、社会经济地位等混杂因素影响。把基因变异当作”工具变量”,就可以在观察数据里模拟随机对照试验,推断暴露与结局的因果关系,而不只是相关关系。中医体质能否作为 MR 中的暴露或中介变量,取决于一个前提——体质必须能被可靠地、量化地测量。2026 年一项 1,472 人的研究展示了这条路的可行性:研究者用验证性因子分析和项目反应理论把中医”证候负荷”构造成连续潜变量,再套用 target trial emulation(目标试验模拟)+ 双重机器学习框架做因果推断,发现证候负荷每升高一个标准差,甘油三酯 z 分数上升 0.18、LDL-C 上升 0.11、C反应蛋白上升 0.16,负对照结果为零,E值 1.54–1.92 表明结果对未测混杂具有中等稳健性(PMID 42519806)。这是中医证候量化第一次完整走进现代因果推断的方法学框架。

把三层证据拼起来,一条清晰的链路浮现:体质是可遗传的中间表型(GWAS 已定位相关位点)→ PRS 可以把”先天禀赋”变成可计算的连续分数 → 纵向队列和中介分析显示体质确实是基因风险传导的通路 → 孟德尔随机化与目标试验模拟提供了把”体质→疾病”从相关升级为因果的严格工具。对”先天禀赋”这个说法而言,这意味着一场身份转变:它从一句只能信的经验判断,变成一个可以检验、可以分层、甚至可以干预的临床工具。无锡队列 5,024 人 14 年随访给出了干预的可行性证据:高 PRS 人群如果把生活方式从糟糕(0–2 项健康因素)改进到理想(4–6 项),2 型糖尿病 10 年发病率从 6.77% 降到 3.28%(PMID 37432247)——基因风险高不是判决,而是提醒。

证据与局限

已验证: 体质(阳虚/阴虚/血瘀)与 MDD 多基因风险显著相关,且 MDD PRS→MDD 的效应部分由阳虚(10.21%)、血瘀(8.41%)体质中介(PMID 39257026);高 PRS 使常见病发病提前 4–9 年(PMID 32273609);中医证候负荷可量化并进入 target trial emulation 因果框架,与血脂/CRP 显著关联(PMID 42519806)。

部分验证: “体质→疾病”的孟德尔随机化研究仍稀缺——多数 MR 用于中医药疗效与肠道菌群,直接以体质为暴露的 MR 设计尚待开展;台湾研究为横断面中介,因果时序需纵向数据确认。

待验证: ①体质量表能否像 PRS 一样构建”体质多基因评分”并在独立队列复现;②以体质为暴露的孟德尔随机化设计(用 GWAS 定位的体质相关位点作工具变量);③体质 PRS×生活方式干预的随机对照试验,检验”先天禀赋可被后天修正”的幅度边界。

相关阅读

证据等级: A
参考文献: PMID 39257026, PMID 37432247, PMID 32273609, PMID 42519806


免责说明:本文属传统知识与科学文献的整理、对照与研究设计讨论,仅供理解与参考,不构成医疗建议,不能替代专业诊断与治疗。文中涉及的效应量、频率区间与暴露结论多为研究报告层面的推断,个体情况差异很大;如有身体不适或健康疑虑,请就医并由专业医师判断。

Polygenic Risk Scores: From Innate Endowment to Quantifiable Causal Chains

Evidence level: A

In Traditional Chinese Medicine (TCM), “inherent endowment” refers to the constitution determined at birth. In modern genetic epidemiology, this can be precisely quantified via Polygenic Risk Scores (PRS)—and longitudinal cohorts combined with Mendelian Randomization are progressively establishing the causal chain from genes → constitution → disease.

Main Text

A core tenet of TCM constitution theory is that individual differences in constitution are partly “inherently endowed,” encoded at birth and determining susceptibility to specific diseases and sensitivity to environmental factors. Historically, this concept remained empirical—”inherent endowment” was intangible and often perceived as akin to metaphysics. However, the development of genetic epidemiology over the past decade has provided a quantitative metric for this ancient concept: the Polygenic Risk Score (PRS).

The logic of PRS is straightforward: complex diseases are not determined by a single gene but result from the cumulative effect of hundreds or thousands of minor genetic variants. By summing weighted risk alleles carried by an individual based on their effect sizes, one obtains a score representing that person’s “genetic lottery.” A study of 135,000 individuals in the Finnish FinnGen cohort demonstrated the power of this metric: compared to average PRS, high-PRS populations exhibited a 21%–38% higher lifetime disease risk and an average onset age advancement of 4–9 years for five common conditions, including coronary heart disease, diabetes, and atrial fibrillation (PMID 32273609). In other words, the extent to which “inherent endowment” influences disease onset timing can now be calculated directly.

Applying this logic to TCM constitution yields a deeper insight. A study involving 11,030 individuals from the Taiwan BioBank integrated MDD PRS with TCM constitution classification (BCQ scale) within a unified model (PMID 39257026). The results revealed three levels of evidence: First, for every standard deviation increase in MDD PRS, the likelihood of Yang Deficiency constitution (OR 1.07), Yin Deficiency constitution (OR 1.07), and Blood Stasis constitution (OR 1.06) increased significantly—individuals with higher genetic risk indeed tended to manifest “imbalanced constitutions.” Second, Yang Deficiency (OR 2.07) and Blood Stasis (OR 1.65) constitutions themselves significantly increased the risk of depression. Third, and most critically, mediation analysis indicated that 10.21% of the effect of MDD PRS on depression onset was mediated through Yang Deficiency constitution and 8.41% through Blood Stasis constitution. This provides direct statistical evidence for the “genes → constitution → disease” causal chain: constitution is not merely a correlate of disease but may represent a genuine intermediate node in the pathway from genetic risk to clinical manifestation.

Can the question “Is constitution a cause?” be further explored? Here, Mendelian Randomization (MR) must be introduced. MR utilizes a natural experiment: genetic variants are randomly allocated at fertilization, independent of postnatal environmental factors, lifestyle habits, or socioeconomic status. By treating genetic variants as “instrumental variables,” one can simulate randomized controlled trials within observational data to infer causal relationships between exposure and outcome rather than mere correlation. Whether TCM constitution can serve as an exposure or mediator variable in MR depends on a prerequisite: constitution must be reliably and quantitatively measured. A 2026 study involving 1,472 individuals demonstrated the feasibility of this approach: researchers constructed a continuous latent variable for TCM “syndrome burden” using confirmatory factor analysis and item response theory, then applied a target trial emulation plus double machine learning framework for causal inference. They found that for every standard deviation increase in syndrome burden, z-scores for triglycerides rose by 0.18, LDL-C by 0.11, and C-reactive protein by 0.16; null results for negative controls indicated robustness against unmeasured confounding (E-value 1.54–1.92) (PMID 42519806). This marks the first time TCM syndrome quantification has fully entered modern causal inference methodological frameworks.

Synthesizing these three layers of evidence reveals a clear pathway: constitution is a heritable intermediate phenotype (GWAS has identified relevant loci) → PRS transforms “inherent endowment” into a calculable continuous score → longitudinal cohorts and mediation analyses confirm constitution acts as a conduit for genetic risk → Mendelian Randomization and target trial emulation provide rigorous tools to upgrade the “constitution → disease” relationship from correlation to causation. For the concept of “inherent endowment,” this signifies a shift in identity: it transitions from an empirical judgment based on faith into a testable, stratifiable, and potentially intervention-driven clinical tool. Evidence for intervention feasibility comes from the Wuxi cohort of 5,024 individuals followed over 14 years: high-PRS populations improved their lifestyle from poor (0–2 healthy factors) to ideal (4–6 healthy factors), reducing the 10-year incidence of type 2 diabetes from 6.77% to 3.28% (PMID 37432247)—high genetic risk is not a verdict, but a warning.

Evidence and Limitations

Verified: Significant association between constitution (Yang/Yin/Blood Stasis) and MDD PRS, with partial mediation of the MDD PRS→MDD effect via Yang Deficiency (10.21%) and Blood Stasis (8.41%) constitutions (PMID 39257026); high PRS associated with earlier onset of common diseases by 4–9 years (PMID 32273609); TCM syndrome burden quantifiable and integrated into target trial emulation causal frameworks with significant associations to lipid profiles and CRP (PMID 42519806).

Partially Verified: Mendelian Randomization studies specifically examining “constitution → disease” remain scarce—most MR studies focus on TCM efficacy and gut microbiota; direct MR designs using constitution as exposure are still pending; the Taiwan study was cross-sectional regarding mediation, requiring longitudinal data to confirm causal temporality.

To Be Verified: ① Can constitution scales be used to construct a “constitution PRS” replicable in independent cohorts? ② MR designs using constitution as exposure (utilizing GWAS-identified loci as instrumental variables)? ③ Randomized controlled trials examining the interaction of constitution PRS × lifestyle interventions to determine the magnitude boundary of whether “inherent endowment” can be modified postnatally.

Related Reading

Evidence Level: A
References: PMID 39257026, PMID 37432247, PMID 32273609, PMID 42519806


Disclaimer: This article compiles traditional knowledge and scientific literature for comparison and research-design discussion only. It is not medical advice and cannot replace professional diagnosis or treatment. Effect sizes, frequency ranges and exposure conclusions cited here are largely inferences at the research-report level, and individual circumstances vary widely. If you have any discomfort or health concerns, please consult a physician.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *