II. 3 Where Microbiota Matters: What We Actually Know
An evidence map that sorts diseases by how firmly the microbiota's role is established – from conditions where it is already the treatment, through strong associations, to mere correlation and open hypotheses.
Microbiome[G] coverage often lumps together what we clinically use (such as FMT for recurrent C. difficile infection) and what we hope for (microbiota[G]-mediated treatment of autism or Parkinson's disease). This chapter separates the two.
The goal is not to deliver "exciting" facts, but to let you know – for each condition – whether the microbiota is proven to cause or sustain it, is rather a consequence of it, or merely correlates with it. That distinction decides whether turning toward your microbiota is worthwhile today, and what you can realistically expect from it.
Microbiota–disease links fall into four tiers: where causation is proven (green; e.g., recurrent C. difficile), where strong association with mechanism exists (yellow; e.g., IBD, IBS, colorectal cancer), where association exists but causality is open (orange; e.g., obesity, allergy), and where the link is hypothesis-level (gray; e.g., autism, Parkinson's). Conditions where the microbiota change is a consequence, not a cause, sit in a separate category (e.g., chronic kidney disease[G]).
Evidence levels
| Color | Tier | What it means |
|---|---|---|
| Green | Proven causal | RCT-grade evidence with microbiota-directed intervention, and/or clear mechanism + reversibility |
| Yellow | Strong association + mechanism | Consistent human cohort data, mechanistic models, causal direction not yet fully resolved |
| Orange | Association with uncertain causality | Multiple human studies show association, but microbiota change could be a consequence |
| Gray | Hypothesis-level | Isolated studies, animal models, small RCTs, anecdote – professional caution |
A tier does not indicate how important the disease is – only how much we know about what to expect when turning toward the microbiota.
The tiers mostly track the evidence pyramid (RCT and meta-analysis > cohort > case-control > animal > in vitro), but every tier also factors in mechanistic plausibility and human reproducibility. The frame is close to the GRADE approach but deliberately simpler so a lay reader can use it.
Proven causal – here the microbiota is the treatment
In these conditions, microbiota-directed intervention (FMT, strain-specific probiotics) has demonstrated RCT-grade efficacy and is used in clinical practice. If someone diagnoses you with one of these, the microbiome isn't an esoteric extra – it can be part of the therapy.
Recurrent Clostridioides difficile infection
Clostridioides difficile (formerly Clostridium difficile) is a spore-forming bacterium that typically proliferates in the colon after antibiotic courses, causing severe diarrhea and colitis. The first episode is usually treated with antibiotics (vancomycin, fidaxomicin), but recurrence happens in 20–25% of patients.
In recurrent cases, fecal microbiota transplantation[G] (FMT[G]) achieves >85–90% cure rate on the first attempt – substantially better than standard antibiotic strategies. [029] The FMT mechanism is straightforward: the healthy donor[G] microbiota repopulates the ecosystem and outcompetes C. difficile. As a side note, it is didactically questionable to treat an antibiotic-induced disorder of bacterial balance with yet another antibiotic.
Since 2022–2023, two FDA-approved microbiota-based products are available: Rebyota[G] (rectally delivered suspension, November 2022) and Vowst[G] (oral spore capsule, SER-109, April 2023). [006] In Europe, FMT has been established clinical practice for recurrent C. difficile infection for years, but the commercial products' availability varies by country.
Evidence: van Nood et al. 2013 NEJM – the first modern RCT, stopped early because FMT was so much better than vancomycin (94% vs. 31% cure). [029] The Cammarota et al. 2017 European consensus recommends FMT as first- or second-line in recurrent cases. NNT: 2 (i.e., preventing one relapse for every two patients treated). Contraindications: severe immunodeficiency, active GI bleeding. Local context: FMT is performed at the hospital level in many EU countries; Rebyota/Vowst routine availability still limited.
Antibiotic-associated diarrhea (AAD)
Antibiotic courses cause diarrhea in 5–35% of patients – C. difficile is just the severe end; most cases are milder dysbiosis[G]. Certain probiotic[G] strains significantly reduce AAD risk in randomized trials.
Saccharomyces boulardii CNCM I-745 and Lactobacillus rhamnosus GG are the two best-documented strains: meta-analyses show 40–60% relative risk reduction for both. [171]
Evidence: Goldenberg et al. 2017 Cochrane review (39 RCTs, ~9,900 adults and children) – probiotics, including S. boulardii and L. rhamnosus GG, significantly reduce the risk of C. difficile-associated diarrhea; NNT ≈ 12 in the high-risk group. [171] Dose: S. boulardii 500 mg/day (adult), L. rhamnosus GG 10⁹–10¹⁰ CFU/day. Timing: start with the antibiotic, continue +3 days after the course. Note: strain-specific effect – a generic "probiotic" without a strain name does not substitute.
Pouchitis in ulcerative colitis patients after IPAA
In ulcerative colitis patients who have undergone colectomy and ileal pouch-anal anastomosis (IPAA), pouch inflammation[G] (pouchitis[G]) is common – affecting ~50% within the first 10 years. VSL#3 (now also marketed as the De Simone formula) is a high-dose multi-strain probiotic mix – in randomized trials it significantly reduces pouchitis flares and maintains remission[G]. [185]
Neonatal necrotizing enterocolitis (NEC) prevention
Life-threatening intestinal disease of premature infants. Over 60 RCTs and large cohorts show that probiotic supplementation (combined Bifidobacterium[G] + Lactobacillus[G] strains) reduces NEC incidence and mortality. [186] This is among the strongest evidence bases for any probiotic indication – though national guidelines and NICU protocols vary.
If you have recurrent C. difficile infection, ask your treating physician about FMT. If you're about to start antibiotics, S. boulardii or L. rhamnosus GG is a reasonable add-on (but does not replace prudent antibiotic use). If you face pouchitis risk after IPAA, the De Simone formula is evidence-based. Neonatal NEC prevention is strictly NICU-supervised.
Strong association + mechanism – the microbiota is clearly involved but not sufficient alone
Here the microbiome shift is consistent, and a mechanistic role is scientifically plausible. But the disease isn't currently treated through the microbiota – alongside lifestyle and standard therapy, microbiome modulation is an increasingly strong adjunct.
Inflammatory bowel disease (IBD): Crohn's and ulcerative colitis
In both Crohn's disease and ulcerative colitis[G], patients' gut microbiomes differ from healthy controls: reduced diversity, fewer anti-inflammatory bacteria (e.g., Faecalibacterium prausnitzii[G]), expansion of pro-inflammatory species (Enterobacteriaceae, AIEC E. coli), and an altered microbial metabolic profile. [187] The microbiome shift isn't only a consequence: twin studies show that symptom-free twins already have altered microbiomes before disease develops in the affected sibling.
The clinical picture for FMT, however, is more nuanced. In ulcerative colitis, four major RCTs showed significant advantage of FMT over placebo – but the effect is smaller (NNT ≈ 5–8) and remission is not always durable. [188] In Crohn's disease, FMT is still experimental.
Evidence in UC: Moayyedi et al. 2015, Rossen et al. 2015, Paramsothy et al. 2017, Costello et al. 2019 – 4 RCTs, ~250 patients total, ~25–32% remission with FMT vs. ~5–9% with placebo. [188] Routine practice: not yet standard IBD treatment; available within clinical trials. Combined approach: efficacy improves with donor selection, intensive protocols, maintenance dosing. Pharmacist (Munar) review note: maintenance probiotics (VSL#3, E. coli Nissle 1917) also have strong evidence for sustaining remission.
Irritable bowel syndrome (IBS)
IBS[G] is not a single disease but a symptom syndrome – with D (diarrhea-predominant), C (constipation-predominant), and M (mixed) subtypes. The microbiome role varies by subtype: SIBO (small intestinal bacterial overgrowth[G]) is common in IBS-D, while methane-producing archaea[G] predominate in IBS-C.
Probiotic trials in IBS are mixed, but subtype- and strain-specific selection improves outcomes. For IBS-D, Bifidobacterium infantis 35624 has shown significant improvement in randomized trials; for bloating-dominant presentations, Lactobacillus plantarum 299v is documented. Our other lever is diet: the low-FODMAP diet is evidence-based under dietitian supervision – the reintroduction phase matters especially, because prolonged strict elimination itself causes microbial diversity[G] loss.
Evidence: Camilleri 2024 review reports NNT 7–8 for B. infantis 35624 in IBS-D – clinically relevant. [189] FMT trials in IBS are inconsistent (Halkjær 2018 negative, Johnsen 2018 positive), likely due to heterogeneity. Low-FODMAP: Halmos et al. 2014 – significant symptom improvement, but with a strict reintroduction phase to avoid microbial-diversity loss. [070]
Celiac disease
The microbiome signature of this gluten-mediated autoimmune disease persists even on a gluten-free diet[G] – meaning the microbiota shift isn't purely a consequence of active inflammation. [190] Clinical relevance: some celiac patients' symptoms[G] persist despite strict gluten avoidance ("refractory celiac"), and microbiome modulation is being investigated.
Colorectal carcinoma (CRC)
Two microbial actors are particularly compelling. Fusobacterium nucleatum[G] – normally an oral cavity bacterium – accumulates in colon tumor tissue, where it promotes inflammation and helps the tumor evade immune detection. Enterotoxigenic Bacteroides fragilis (ETBF) works by a different mechanism: its fragilysin (BFT, B. fragilis toxin) directly causes DNA damage in gut epithelial cells. Both bacteria's roles are detailed mechanistically in human and animal models. [191]
Clinical implications are twofold: first, microbiome-signature-based screening (stool DNA + 16S panel) is promising but not yet routine. Second, on the prevention side, the strongest levers remain fiber intake and limiting red/processed meat – chapter 4's lifestyle levers together with classical clinical screening.
Cancer immunotherapy response
Checkpoint inhibitor (anti-PD-1, anti-PD-L1, anti-CTLA-4) efficacy is clearly affected by the patient's gut microbiome: enrichment of Akkermansia muciniphila[G], Faecalibacterium prausnitzii, and certain Bifidobacterium species correlates with higher response[G] rates. [176] This finding has entered clinical oncology thinking in melanoma, NSCLC, and renal cancer.
Two RCTs (Davar 2021, Baruch 2021 Science) successfully restored response in checkpoint-resistant melanoma patients via FMT – meaning microbiome modulation here is both a predictor and an actionable target. [192]
If you live with IBD: the microbiome strategy is a suitable adjunct to standard treatment – but per current knowledge does not replace it. If you have IBS symptoms: FMT / subtype-specific probiotic + dietitian-supervised FODMAP[G] is worth a trial. If your CRC risk is elevated (family history, smoking, high BMI): colonoscopy screening is the priority – microbiota-aware prevention only supplements. If you're receiving cancer treatment: ask your oncologist and a microbiota specialist about a mindful diet, and avoid unnecessary antibiotics.
Association with uncertain causality – cause, consequence, or shared third factor
Here, too, consistent human cohort data show microbiome composition differs between patients and healthy controls. But the direction of causation is unresolved, and human trials of microbiota-directed interventions give mixed results. The picture suggests: the microbiota is an important factor, but environmental factors (the exposome[G]) certainly also play a role in how these conditions arise.
Obesity, metabolic syndrome[G], type 2 diabetes
The "obese vs. lean microbiota" paradigm (high Firmicutes[G]/Bacteroidetes[G] ratio = obesity) was attractive based on early cohort studies – but didn't reproduce in later, larger trials. [193] Current consensus: microbiome composition is a mediator of obesity and lifestyle (diet, exercise), not an independent cause.
Human FMT trials for insulin[G] sensitivity[G] are mixed: Vrieze et al. 2012 showed FMT from a lean donor improved insulin sensitivity in metabolic-syndrome men – but the effect wasn't sustained, and independent replication is weaker. [194] In T2DM, the metformin[G]-mediated microbiome shift (see chapter 7) is more a drug effect than a causal disease factor.
NAFLD / MASLD[G] (non-alcoholic fatty liver)
The gut–liver axis is well documented: LPS[G] (endotoxin) leaks through the gut barrier into portal circulation, activating Kupffer cells. Multiple studies show microbiome shifts in NAFLD – but the most effective microbiome-directed interventions today are lifestyle-based (weight loss, fiber increase). [195]
Allergy, atopic dermatitis, asthma
The more nuanced version of the "hygiene hypothesis" – the biodiversity hypothesis – proposes that reduced infant environmental and gut microbial diversity contributes to Th2-skewed immune responses. [196] Infant antibiotic exposure (especially in the first 0–24 months) and C-section delivery are consistently linked to elevated allergy and asthma risk.
Probiotic prevention trials for atopic dermatitis are mixed; some professional bodies consider them in selected high-risk cases (during pregnancy and for infants), but the evidence is not strong enough for a routine general-prevention recommendation.
Acne, rosacea
Skin microbiome (particularly Cutibacterium acnes strain diversity) and gut microbiome roles in acne are actively researched, but standard dermatologic therapy (topical retinoids, isotretinoin) still dominates. The link between rosacea and SIBO is intriguing – small studies showed rosacea improvement after SIBO eradication. [197]
The big finding of the 2010s: the link between infant antibiotic exposure and later allergy/atopy risk is consistently detectable in cohort studies of hundreds of thousands (Bokulich, Stewart, Kronman, and colleagues). The pattern is consistent: earlier (0–12 months most sensitive), broader spectrum, and more courses → higher subsequent asthma and eczema rates. This doesn't mean AB causes allergy – causal direction is hard to prove. But the biodiversity hypothesis mechanism (reduced microbial training → Th2 skewing) is plausible. Clinical implication: indicated AB is of course necessary – but reducing "just in case" courses brings real benefit.
In these conditions don't treat the microbiota as "the magic answer." Evidence-based lifestyle fundamentals (chapters 4–6) improve your microbiome and the underlying disease at the same time. If you live with an allergic or inflammatory condition, ask which microbiome-aware step is worth taking and which isn't – many commercial products overpromise.
Hypothesis-level – interesting but not yet a therapeutic promise
Here, isolated studies, animal models, and small human RCTs show a connection. Microbiota-directed therapy is experimental. Today, microbiota-based therapy for autism, Parkinson's, or depression is only defensible when available protocols cannot halt further deterioration or the patient's quality of life is declining rapidly. There are no convincing human trials for these conditions.
Autism spectrum disorder (ASD)
Kang et al. 2017 and 2019 – Arizona group observed microbiota transfer therapy (MTT[G]) improved GI symptoms and behavioral scales in 18 ASD children, with effects persisting at 2-year follow-up. [123] This study is not placebo-controlled and small – therefore not on firm causal ground – but it's notable, and larger randomized trials are underway. In children, the therapy seems most promising when the history includes prematurity/C-section, early (0–24 months) or maternal antibiotic therapy, and the patient has documented gastrointestinal symptoms; in such cases, improvements on certain behavioral scales were observed after prolonged FMT. There is, however, no validated biomarker, no validated patient-selection strategy, and no reproduced RCT evidence.
Parkinson's disease
The Braak hypothesis proposes that Parkinson's alpha-synuclein aggregates start in the gut and reach the brain via the vagus[G] nerve. Microbiome shifts in Parkinson's patients are consistent (reduced Prevotella, expanded Enterobacteriaceae). [198] FMT trials are at phase I level – no therapy yet.
Depression and anxiety (psychobiotics)
Small RCTs with certain probiotic strains (Lactobacillus helveticus R0052, Bifidobacterium longum R0175) have shown significant mood effects – but meta-analyses are heterogeneous and effect sizes small. [199] Gut–brain mechanisms (vagus, tryptophan[G] metabolism, SCFA[G] receptors) are plausible. Does not replace clinical depression treatment.
Alzheimer's disease
Porphyromonas gingivalis (an oral pathogen) has been detected in the brains of Alzheimer's patients – suggesting an "oral microbiota → brain" pathway. [200] Early-stage hypothesis, therapeutic translation still ahead.
Rheumatoid arthritis, lupus, multiple sclerosis
Multiple autoimmune diseases show microbiome shifts. Prevotella copri enrichment in new-onset rheumatoid arthritis is one of the strongest findings (Scher et al. 2013). [201] Microbiome-directed therapy in these conditions is all experimental.
Two cautions around hypothesis-level indications. First: small RCTs can produce positive signals by chance, especially on behavioral or mood endpoints where placebo effects are strong. The Kang et al. 2017/2019 autism MTT trials (n=18, no placebo control) have notable positive results, but we cannot claim they prove causation. [123] Second: psychobiotic meta-analyses (Liu et al. 2019) are heterogeneous, and effect sizes are generally smaller than for conventional antidepressants. [199] Clinical implication: in these areas probiotic or MTT use is acceptable within a clinical trial, but is not yet professional ground for first-line therapy.
If you live with a neurologic or autoimmune condition, you may participate in clinical microbiome-therapy trials – but for autism spectrum disorder or Parkinson's disease, routine FMT is currently experimental, not evidence-based. The lifestyle messages of chapters 4–6 benefit everyone, but here especially, you should not set excessive expectations.
Separate category – microbiota clearly a consequence
In these conditions the microbiome shift is not a causal factor but the result of the disease, drug, or organ dysfunction. Microbiome modulation here doesn't cure the underlying disease – but it can affect its consequences (toxin accumulation, inflammatory mediators), and that can be clinically relevant.
Chronic kidney disease (CKD)
Uremic toxins (TMAO, indoxyl sulfate, p-cresyl sulfate) aren't produced directly in the kidneys – they are gut-microbial in origin, and accelerate kidney failure progression. Prebiotic[G] and probiotic trials in CKD can reduce uremic burden (fiber increase affects TMAO precursor metabolism), but don't restore kidney function alone. Clinically, nutritional counseling (renal dietitian) remains the primary lever.
Chronic heart failure
Gut barrier leakage, LPS translocation, and Kupffer cell activation contribute to heart-failure cachexia and inflammatory background. [202] Microbiome modulation is again an adjunct – the primary therapy (ACE inhibitor, beta-blocker, diuretic) retains unchanged priority.
COPD
The gut–lung axis is a research frontier of recent years. Gut microbiota shifts in COPD patients are consistent, but the direction of causation is lung → gut (systemic inflammation → dysbiosis), not the reverse. [203]
If you live with CKD, heart failure, or COPD, microbiome-aware diet (low salt, high fiber, polyphenols[G]) within a frame approved by your treating physician can supplement your treatment – but doesn't replace it. Don't begin strict diets without specialist clearance, because in chronic organ disease, inadequate intake can harm (e.g., in CKD, potassium intake must be controlled).
What you can do tomorrow
The right next step depends on your profile. The four most common paths:
- I have a diagnosis (IBD, IBS, T2DM, CRC risk, etc.): read your own disease group's section above, then chapter 7 (drug–microbiome interactions), chapter 11 (therapeutic toolbox), and Appendix V (When to See a Doctor).
- I have symptoms (bloating, alternating bowel, IBS-like): start at the IBS section above, then chapter 4 (FODMAP, fiber), and consider the diagnostic steps in chapter 10 (calprotectin, SIBO breath test).
- I'm coming off an antibiotic course: jump to chapter 7 (recovery protocol) and chapter 11 (probiotic table for AAD).
- I'm healthy and want to optimize: chapters 4 and 5 give concrete actions. Keep this chapter as a reference – if a new symptom or diagnosis appears in your circle, you'll know where to check the evidence tier.
Any microbiome-directed attempt is overridden by red-flag symptoms. Immediate medical consult required for:
- blood in stool (fresh or dark)
- unintentional weight loss (>5% in 6 months)
- nocturnal abdominal pain or diarrhea
- persistent fever + abdominal symptom
- family history: colorectal cancer under 50, IBD
Detailed red flags: Appendix V When to See a Doctor chapter.
What's next
The next part (chapters 4–7) turns to lifestyle modulators: nutrition, sleep–movement–stress, environment, medications. Each chapter refers back to the table above: knowing where you stand on the evidence map tells you what to realistically expect from each lifestyle modification – and which is worth discussing with your treating physician in the context of a specific disease.