Evidence level: B
一个在痰湿质里”耗竭即需补充”的产丁酸菌,在阳虚质里却可能是”升高即稳态”——同一菌属在不同体质基线上的方向相反,应答标志物直接套用会反向干预。
正文
当我们用现代医学思路设计中医体质相关的菌群干预时,第一个、也是最隐形的假设是:“在体质 A 里能预测应答的菌,在体质 B 里还是预测应答的菌”。但这条假设的成立性,最近两年被真实数据动摇——一个叫 Faecalibacterium(主要是 F. prausnitzii)的属,在不同中医体质研究里方向并不一致。判断”一菌标志”能否跨体质套用,本身就是一个需要被裁决的工程问题,而不是可以默认成立的前提。
方向分裂:同一种菌,在不同体质里”该补”还是”该削”
Faecalibacterium 在西方研究里长期被视为”健康肠道指示菌”——它产丁酸、刺激 IL-10、抑制 NF-κB、巩固肠屏障,Sokol 2008 的奠基性工作至今仍是机制锚点(PMID 25275569 用 IL-17 抑制与结肠炎缓解再次确认)。这似乎暗示:哪一类人体内这个菌少,哪一类人就该补它。
但把视线转向中医体质研究,方向出现了明显分裂:
- 痰湿质(PDC)肥胖人群:Shin 2022(PMID 35719350)30 例 PDC vs 30 例非 PDC 肥胖 vs 30 例平和质对照,PDC 组该菌显著耗竭、Prevotella 富集、F/B 比值改变,菌群与代谢紊乱指标的相关性强于与 BMI 的相关性——”耗竭即需补充”成立。
- 阳虚质(YADC)健康人群:Zhang 2022(PMID 35294128)4 例健康 YADC vs 5 例平和质对比发现反向信号——”有害菌”(Prevotellaceae,LDA>4,P=0.0141)和”有益菌”(Ruminococcaceae LDA>4、P=0.0025、Faecalibacterium LDA>4、P=0.0484)双向升高。作者解读为”健康阳虚者的稳态维持机制”——该菌在 YADC 里不是缺陷而是调动起来的力量。
- 气阴两虚(QYLX)代谢综合征:Shang 2022(PMID 36440222)60 例 MS 患者+30 例对照,TSZE(痰湿)组关联 Faecalibacterium 耗竭,QYLX 关联不同菌属与代谢路径富集。两组与健康对照的距离、菌属-临床相关结构不重叠。
- 湿热质(DHC)多中心 249 例:2025 年 Pan 等(PMC 11764936)发现 DHC 组特征是 Barnesiella、Coprobacter、Butyricimonas 等产 SCFA 属显著下降,Faecalibacterium 不是该研究的最强信号。
方向为什么会反
要把”观察巧合”升格为”可裁决的工程问题”,需要三条机制候选:(1) 炎症基线差异——痰湿质常伴低度慢性炎症,该菌与炎症互为因果,补充等于直接降火;阳虚质健康者炎症基线低,升高反映”养能量”需求而非缺陷。(2) 肠屏障状态差异——湿热、痰湿常伴肠通透性增加(”湿热内蕴”),菌减少既是结果也是推动者;阳虚质若屏障完好,再补不带来额外收益。(3) 菌属内部异质性——Lopez-Siles 等(PMID 24713205)指出 F. prausnitzii 内部 phylogroup I/II 在疾病中表现不同,真正可用的标志物可能在种、株甚至基因型层面。
一个最易踩的坑:把 IBS 应答标志物直接搬到体质 RCT
《NC-IBS 应答者基线富集 Coriobacteriaceae/C. aerofaciens》(Gargari 2024,PMID 38178601,本系列 #696 母文)这类 IBS 标志物,搬到痰湿/阳虚/湿热质上方向不是”必然一致”——因为 IBS-D 患者的该菌水平本就多样(Duboc 2012、Rigsbee 2012 多份研究未在 IBS-D 看到一致耗竭),把它定为”应答菌”本身就有数据噪声;把这种”带有体质混合背景的菌属信号”直接用于单一中医体质的预测模型,方向翻转几乎是默认会发生的事。
方向一致性裁决的三步法
要把”一菌标志跨体质是否成立”做成可裁决的工程问题,建议走这三步:(1) 基线分布标定——在每种偏颇质+平和质分层队列里,先报 Faecalibacterium 属、种(F. prausnitzii phylogroup I/II)、关键功能基因(丁酰-CoA:乙酸 CoA-转移酶)的基线均值与 95% CI,明确”少”或”多”的真实水平。(2) 方向一致性检验——用三体质+平和质四组队列跑同种菌群干预 RCT,逐亚组估计”基线菌量-应答幅度”斜率与置信区间:同号且 CI 重叠 → 方向一致;异号或不重叠 → 方向翻转;同号但量级差 ≥2 倍 → 方向一致但量级不可外推。Bootstrap 检验斜率差异,配 FDR 校正。(3) 功能 vs 菌属双轨对比——同步测粪便丁酸浓度与宏基因组丁酸通路基因丰度,分别建模”菌属预测 vs 功能预测”在四亚组的 ΔAUC、方向翻转率。如果功能标志物比菌属更稳定(参考本系列 #701),应升级到功能层;如果功能层也翻转,就不要把这套菌标志物跨体质套用。
工程结论
跨体质方向一致性问题不是哲学问题,是可被三步法裁决的工程问题。在裁决完成之前,最安全的做法是每种偏颇质独立训练应答预测模型,把”一菌标志跨体质套用”当成待证假设而不是默认前提;本系列 #696(C. aerofaciens 跨体系迁移)和 #701(菌属 vs 功能粒度)会进一步给出具体裁决设计。
证据与局限
已验证:
– 痰湿质该菌耗竭 + Prevotella 富集(PMID 35719350,n=90 三组对照)
– 阳虚质该菌双向升高作为稳态(PMID 35294128,n=9 小样本 + 120 大群体验证)
– 痰湿 vs 气阴两虚 MS 菌属-临床相关结构不重叠(PMID 36440222,n=90)
– 湿热质多中心产 SCFA 属减少(PMC 11764936,n=249)
– F. prausnitzii 经丁酸-IL-10-NF-κB 通路抑制 IL-17 缓解结肠炎(PMID 25275569)
– 该菌与 E. coli 联合指数区分 IBS/IBD 表型(PMID 24713205,n=111)
部分验证:
– Faecalibacterium 属水平信号在 IBS-D 上不一致(多份研究阴性/阳性并存),属内异质性是大概率原因
– 阳虚质 Faecalibacterium 升高的样本仅 4+5 例(PMID 35294128),需大样本复制
待验证:
– 三体质同种菌群干预 RCT 中 Faecalibacterium 斜率方向是否翻转(无直接头对头研究)
– 属水平 vs 菌株水平(F. prausnitzii phylogroup I/II)方向稳定性差异的量化
– 功能层(粪便丁酸、丁酰-CoA 转移酶基因)方向是否比菌属层更稳定
相关阅读
证据等级: B
参考文献: PMID 35719350, PMID 35294128, PMID 36440222, PMID 25275569, PMID 24713205, PMID 38178601, PMID 11764936, PMID 41853345
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。
Can the response direction of Faecalibacterium remain consistent across imbalanced constitutions in Traditional Chinese Medicine constitution × microbiome response?
Evidence level: B
A butyrate-producing Faecalibacterium that is “depleted and requires supplementation” in the Phlegm-Dampness constitution may instead represent a “rise toward homeostasis” in the Yang Deficiency constitution—opposing directions across different constitutional baselines mean that directly applying response markers can lead to counter-intervention.
Main Text
When designing microbial interventions related to TCM constitutions using modern medical logic, the first and most invisible assumption is: “The bacteria that predict response in Constitution A also predict response in Constitution B.” However, the validity of this assumption has been shaken by real-world data over the last two years. A genus called Faecalibacterium (primarily F. prausnitzii) shows inconsistent directions across different TCM constitution studies. Determining whether a “single-bacteria marker” can be applied across constitutions is an engineering problem requiring adjudication, not a premise that can be assumed true.
Directional Split: Should We Supplement or Reduce the Same Bacteria in Different Constitutions?
In Western research, Faecalibacterium has long been viewed as an “indicator of a healthy gut”—it produces butyrate, stimulates IL-10, inhibits NF-κB, and consolidates the intestinal barrier. Sokol 2008’s foundational work remains a mechanism anchor (PMID 25275569 confirmed via IL-17 inhibition and colitis resolution). This suggests: “In which group is this bacteria low, should we supplement it in that group.”
However, shifting focus to TCM constitution research reveals a clear split:
- Phlegm-Dampness Constitution (PDC) Obese Populations: Shin 2022 (PMID 35719350) compared 30 PDC vs. 30 non-PDC obese vs. 30 normal constitution controls. The PDC group showed significant depletion of Faecalibacterium, enrichment of Prevotella, and altered F/B ratios. Correlations between microbiota and metabolic disorder indicators were stronger than with BMI—”depletion implies supplementation” holds true.
- Yang Deficiency Constitution (YADC) Healthy Populations: Zhang 2022 (PMID 35294128) found a reverse signal comparing 4 healthy YADC vs. 5 normal constitution cases—”harmful bacteria” (Prevotellaceae, LDA>4, P=0.0141) and “beneficial bacteria” (Ruminococcaceae LDA>4, P=0.0025; Faecalibacterium LDA>4, P=0.0484) showed bidirectional elevation. Authors interpret this as a “homeostasis maintenance mechanism in healthy Yang Deficiency individuals”—this bacteria in YADC is not a defect but mobilized power.
- Qi-Yin Deficiency (QYLX) Metabolic Syndrome: Shang 2022 (PMID 36440222) compared 60 MS patients + 30 controls. The TSZE (Phlegm-Dampness) group associated with Faecalibacterium depletion, while QYLX associated with different genera and metabolic pathway enrichment. The distance from healthy controls and the microbiota-clinical correlation structures did not overlap between groups.
- Damp-Heat Constitution (DHC) Multi-center 249 Cases: Pan et al. (PMC 11764936) in 2025 found that the DHC group was characterized by significant decreases in SCFA-producing genera such as Barnesiella, Coprobacter, and Butyricimonas; Faecalibacterium was not the strongest signal in this study.
Why Do Directions Reverse?
To elevate “observational coincidence” to a “judgable engineering problem,” three mechanism candidates are needed: (1) Inflammatory Baseline Differences—Phlegm-Dampness often accompanies low-grade chronic inflammation, where the bacteria and inflammation are mutually causal; supplementation equals direct fire reduction. Healthy Yang Deficiency individuals have low inflammatory baselines, so elevation reflects “energy nurturing” needs rather than defects. (2) Intestinal Barrier State Differences—Damp-Heat and Phlegm-Dampness often accompany increased intestinal permeability (“damp-heat accumulation”); bacterial reduction is both a result and a driver; if the Yang Deficiency barrier is intact, further supplementation yields no additional benefit. (3) Intra-genus Heterogeneity—Lopez-Siles et al. (PMID 24713205) noted that F. prausnitzii phylogroups I/II show different performances in disease; the truly usable marker may lie at the species, strain, or genotypic level.
A Common Pitfall: Directly Transplanting IBS Response Markers to Constitution RCTs
Markers like those in “NC-IBS Responders Baseline Enriched with Coriobacteriaceae/C. aerofaciens” (Gargari 2024, PMID 38178601, source of this series #696) do not necessarily align directionally when applied to Phlegm-Dampness/Yang Deficiency/Damp-Heat constitutions. This is because Faecalibacterium levels in IBS-D patients are inherently diverse (Duboc 2012, Rigsbee 2012 found no consistent depletion in IBS-D across multiple studies); defining it as a “response bacterium” introduces data noise. Applying such a “microbiota signal with mixed constitutional backgrounds” directly to a prediction model for a single TCM constitution almost inevitably leads to direction reversal.
Three-Step Adjudication for Directional Consistency
To make the question of “whether a single-bacteria marker applies across constitutions” a judgable engineering problem, we recommend three steps: (1) Baseline Distribution Calibration—In stratified cohorts for each disparate constitution + normal constitution, first report baseline means and 95% CIs for the Faecalibacterium genus, species (F. prausnitzii phylogroup I/II), and key functional genes (butyryl-CoA:acetate CoA-transferase), clarifying the true levels of “low” or “high.” (2) Directional Consistency Testing—Run the same microbial intervention RCT across four groups (three constitutions + normal) and estimate subgroup slopes and confidence intervals for “baseline bacterial load vs. response magnitude”: same sign with overlapping CIs → consistent direction; opposite signs or non-overlapping CIs → direction reversal; same sign but magnitude difference ≥2x → consistent direction but magnitude not extrapolatable. Perform Bootstrap tests for slope differences with FDR correction. (3) Dual-Track Comparison: Function vs. Genus—Simultaneously measure fecal butyrate concentration and macro-genomic butyrate pathway gene abundance. Model “genus prediction vs. functional prediction” for ΔAUC and direction reversal rates across four subgroups. If functional markers are more stable than genus markers (refer to this series #701), upgrade to the functional layer; if the functional layer also reverses, do not apply this bacterial marker across constitutions.
Engineering Conclusion
The question of cross-constitution directional consistency is not a philosophical issue but an engineering problem adjudicable by the three-step method. Until adjudication is complete, the safest approach is to independently train response prediction models for each disparate constitution, treating “cross-constitution application of a single-bacteria marker” as a hypothesis to be tested rather than a default premise; series #696 (C. aerofaciens cross-system migration) and #701 (genus vs. functional granularity) will further provide specific adjudication designs.
Evidence and Limitations
Verified:
– Faecalibacterium depletion in Phlegm-Dampness + Prevotella enrichment (PMID 35719350, n=90 three-group comparison)
– Bidirectional elevation of Faecalibacterium in Yang Deficiency as homeostasis (PMID 35294128, n=9 small sample + n=120 large cohort validation)
– Non-overlapping microbiota-clinical correlation structures between Phlegm-Dampness and Qi-Yin Deficiency MS (PMID 36440222, n=90)
– Multi-center reduction of SCFA-producing genera in Damp-Heat constitution (PMC 11764936, n=249)
– F. prausnitzii inhibits IL-17 via butyrate-IL-10-NF-κB pathway to resolve colitis (PMID 25275569)
– Combined index with E. coli distinguishes IBS/IBD phenotypes (PMID 24713205, n=111)
Partially Verified:
– Genus-level signals for Faecalibacterium are inconsistent in IBS-D (mixed negative/positive results across multiple studies); intra-genus heterogeneity is likely the cause
– Only 4+5 samples showed elevated Faecalibacterium in Yang Deficiency (PMID 35294128); requires large-sample replication
To Be Verified:
– Whether Faecalibacterium slope direction reverses in same-genus microbial intervention RCTs across three constitutions (no direct head-to-head study exists)
– Quantification of directional stability differences between genus-level vs. strain-level (F. prausnitzii phylogroup I/II)
– Whether the functional layer (fecal butyrate, butyryl-CoA transferase genes) is more stable than the genus layer
Related Reading
- Directional Stability of Same Marker Across Different Constitutions
- Cross-system Migration of C. aerofaciens Response Marker
- Cross-constitution Transfer of TCM Constitution Response Prediction Models
- Multi-center Microbiome Study of Damp-Heat Constitution
Evidence Level: B
References: PMID 35719350, PMID 35294128, PMID 36440222, PMID 25275569, PMID 24713205, PMID 38178601, PMID 11764936, PMID 41853345
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.
Leave a Reply