Name: Elimination Protocols
Creator: Realised
Published: 2026-06-01

Summary

An elimination diet is a time-limited diagnostic experiment, not a way of eating — its entire value is in the reintroduction phase that follows, and the most common failure mode is not running it wrong but never finishing it: people get relief, get scared, and stay restricted, trading a tractable gut symptom for a shrinking diet, a thinner microbiome, and a slide toward disordered eating.

Why Strong

Tier 1 for time-limited, structured, condition-specific elimination + reintroduction because: low-FODMAP for IBS has multiple RCTs and meta-analyses (including a 2024 blinded reintroduction RCT and an unfunded network meta-analysis) showing moderate-to-large symptom benefit; empiric elimination for biopsy-confirmed EoE is guideline first-line drug-free therapy. The principal limitation (dietary trials can't be blinded for the restriction phase) is inherent to the field, not a defect of these specific protocols, and the best reintroduction work was blinded.

Tier 3 for sensitivity-panel-guided or open-ended undifferentiated elimination because: IgG/IgA testing is rejected by major allergy bodies; applying elimination to non-GI, undifferentiated complaints has no supporting trial evidence and a documented harm profile.

NOT Tier 0.5 because: this is a targeted diagnostic intervention for specific indications, not a universal foundational practice — most people should never run one.

NOT a blanket Tier 1 because: the harm side (over-restriction, ARFID/orthorexia, nutrient and microbiome loss) is well-documented and the most common real-world use (skip reintroduction, trust a sensitivity test) is the low-tier version. The tier honestly splits by how it's run.

Practical takeaway

First: is elimination even warranted? Run it for a defined target, not for "feeling off."
• Warranted: IBS with a clear symptom pattern (low-FODMAP, ideally dietitian-guided); biopsy-confirmed EoE (step-up empiric elimination, specialist-managed); a specific suspected food with a plausible, reproducible reaction; confirmed enzymatic intolerance (e.g. lactose) where you're calibrating dose.
• Not warranted as an elimination protocol: undifferentiated fatigue/brain-fog/"inflammation" with no GI pattern; an IgG/IgA "sensitivity" panel result; a vague sense that "food is the problem." These need a different work-up, not a restriction.
• **Red-flag symptoms that need a doctor first, not a diet:** blood in stool, unintended weight loss, anaemia, nocturnal symptoms, onset over age 50, dysphagia, family history of IBD/coeliac/GI cancer. Don't self-eliminate your way past a diagnosable disease. (And get coeliac testing done before you remove gluten — the test needs you to be eating it.)

The three phases (non-negotiable structure — based on low-FODMAP, generalises):

1. Restriction (≤4-6 weeks, hard ceiling). Remove the target category as completely and as narrowly as the protocol requires — eliminate the fewest foods that test the hypothesis (the EoE field's step-up logic: start with two, not six). Substitute deliberately so total calories, protein, and fibre stay adequate. If you get no meaningful relief in 4-6 weeks, the hypothesis failed — stop and reintroduce everything, don't restrict harder.
2. Reintroduction (the phase that actually has the answers). Reintroduce one food/subgroup at a time, a small-then-larger dose over ~3 days, with a return-to-baseline washout between challenges. Log symptoms against dose. Expect to pass most challenges — that's the normal, healthy result and the point of doing it. For FODMAPs specifically, test each subgroup (fructans, GOS, lactose, fructose, polyols) separately; they're independent.
3. Personalisation. Keep restricted only the specific foods that reproducibly and dose-dependently triggered symptoms — and re-test even those every few months, because thresholds drift and tolerance often recovers.

What "working" looks like:
• You end with a shorter restriction list than you started the reintroduction with — usually far shorter.
• You can name your triggers and their doses ("a whole onion, yes; a little garlic oil, fine").
• Your diet at the end is more varied than at the peak of restriction, not less.
• Eating out and eating socially get easier over time, not harder.

What to track: the eliminated list length over time (it should peak early and shrink), symptom severity (a validated tracker like IBS-SSS if IBS), body weight and energy (guard against accidental under-eating), fibre and calcium intake, and — honestly — your anxiety around food. If the list is growing and the anxiety is rising, the protocol has failed even if individual symptoms improved.

Always pair with a clinician/dietitian when: the indication is a diagnosed disease (EoE, coeliac, IBD), the restriction will exceed 6 weeks, more than one food group is involved, the user is a child, pregnant, underweight, or has any eating-disorder history.

Evidence detail

Why This Entry Exists

A user has gut symptoms — bloating, pain, erratic bowels, reflux — reads that "FODMAPs" or "gluten" or "nightshades" or "the foods on my sensitivity test" are the culprit, and cuts them out. They feel better. So they cut more. Six months later they are eating fourteen foods, anxious at every restaurant, convinced each new symptom means another food to remove, and no closer to knowing what actually triggers them — because they never reintroduced anything. The relief that felt like a diagnosis was actually the start of a diagnostic process they abandoned at step one.

This is the central trap of elimination diets, and it is extremely common. The clinical protocols (low-FODMAP, empiric food elimination for EoE) are genuinely effective — but they are three-phase protocols where restriction is only phase one, and the published guidance is emphatic that the restriction phase should last no more than 4-6 weeks before structured reintroduction begins. The version that escapes into the wild keeps only phase one. That open-ended version is where the harm is.

What bad advice does this protect against?
• "Cut out [food group] and see how you feel" as a permanent lifestyle, with no plan to reintroduce.
• The belief that feeling better on a restricted diet proves the eliminated food is harmful (it usually doesn't — see Mechanism).
• IgG/IgA "food sensitivity" tests sold direct-to-consumer as a shortcut to knowing what to eliminate. These are not validated and routinely generate long restriction lists from foods you eat often and tolerate fine.
• The framing of elimination as low-risk because "it's just food." Prolonged restriction has real costs: nutrient gaps, microbiome depletion, social and psychological narrowing, and a well-documented on-ramp to orthorexia and ARFID — particularly in the GI-symptom population, who are already at elevated risk.
• The opposite error: dismissing elimination entirely. For a confirmed indication, run correctly and time-boxed, it is first-line and evidence-backed.

Evidence

Strongest first.

1. Low-FODMAP diet for IBS — efficacy of the protocol is well established; the reintroduction phase is now its own evidence base.
A 2024 network meta-analysis (23 RCTs, 1,689 IBS patients, explicitly unfunded — authors declared no external funding) found the low-FODMAP diet significantly superior to a standard diet on symptom severity (IBS-SSS mean difference −46.3, p<0.01) and quality of life. Notably the Mediterranean diet ranked highest on most outcomes (less restrictive, but only one trial — thin), and a gluten-free diet showed no significant advantage over either. The review flagged the field's two structural weaknesses: dietary trials cannot be blinded (high performance-bias risk), and there is "a lack of data concerning adherence to restrictive diets," i.e. nobody is tracking what happens long-term to people kept restricted. (independent / unfunded)

A landmark 2024 Gastroenterology study did what almost no elimination research does: it blinded the reintroduction. Six-week low-FODMAP responders entered a 9-week blinded randomised reintroduction with individual FODMAP powders (fructans, fructose, GOS, lactose, mannitol, sorbitol) vs glucose control. Result: symptom recurrence was personalised — fructans and mannitol were the most prevalent triggers, but the pattern differed per person, and many participants tolerated FODMAPs they had eliminated. The blinded design is the cleanest available demonstration that the eliminated list is far longer than the actually-triggering list — which is the entire argument for reintroduction and against staying broadly restricted. (academic/clinical)

Real-world effectiveness: with a trained dietitian, the low-FODMAP diet improves symptoms in roughly 50-80% of IBS patients (per the StatPearls clinical synthesis and NICE positioning). NICE recommends it selectively, under professional supervision.

**2. Empiric elimination for eosinophilic oesophagitis (EoE) — the case where elimination is genuinely first-line — and the field has moved toward less restriction.
EoE is an immune/allergic oesophageal disease confirmed by biopsy. Empiric elimination is "the most effective drug-free treatment." Crucially, the field abandoned the old, aggressive six-food elimination diet in favour of a step-up (2-4-6) approach**: start by removing just milk and gluten-containing cereals (a two-food elimination), re-scope at 8-12 weeks, and only escalate if needed. A two-food elimination achieves remission in ~43% of children and adults, identifies most single-trigger responders early, and "avoid[s] unnecessary dietary restrictions." This is the clinical world independently arriving at Realised's position: eliminate as little as possible for as short as possible. (academic/clinical — the 2-4-6 study, J Allergy Clin Immunol 2017; CGH systematic review 2023)

3. Microbiome cost of the restriction phase is real and measurable.
A systematic review of nine studies found a **consistent reduction in faecal *Bifidobacterium*** during the FODMAP-restriction phase — appearing within 3-4 weeks. FODMAPs are prebiotic fibres; starving them starves the bacteria that ferment them. The reduction is reversible on reintroduction (and one RCT showed a probiotic restored Bifidobacterium without losing symptom benefit). This is a concrete biological reason the restriction phase is time-boxed: it is not a neutral holding pattern. (academic/clinical)

4. Nutritional adequacy degrades with prolonged restriction.
Prolonged FODMAP restriction is associated with reduced intake of calcium, B vitamins, iron, fibre, and prebiotics. The longer the restriction and the broader the list, the larger the gap — and this is the FODMAP diet, which is among the better-structured protocols. Ad-hoc elimination of multiple food groups with no nutritional substitution is worse. (clinical synthesis — StatPearls, Cambridge Proc Nutr Soc long-term review)

5. Disordered-eating overlap — strongest harm signal, and population-specific.
The GI-symptom population is already at elevated eating-disorder risk, and restriction protocols can tip it. In coeliac disease — a condition that mandates lifelong gluten elimination — 14-57% of patients meet criteria for ARFID (avoidant/restrictive food intake disorder), driven by ongoing symptoms, fear of exposure, conditioned food aversions, and the social burden of restriction. ARFID is restriction not motivated by body image (distinguishing it from anorexia) but it produces the same malnutrition, weight loss, and psychosocial impairment. Orthorexia — pathological fixation on "clean"/"safe" eating — has the same shape: cut, then cut more, until the safe-food list is tiny. Elimination dieting is a recognised on-ramp for both. (clinical — DSM-5 / StatPearls; coeliac-ARFID prevalence data)

6. IgG/IgA "food sensitivity" tests — the elimination shortcut that is not validated.
The AAAAI, EAACI, and CSACI (US, European, and Canadian allergy bodies) all advise against IgG/IgA food testing for diagnosing sensitivities. IgG to a food reflects normal exposure and tolerance — often you produce more IgG to foods you eat more of. Using these panels to build an elimination list reliably over-diagnoses, generates unnecessary restriction, and risks nutritional harm with no outcome benefit. (independent specialty-society consensus)

Mechanism

Why elimination produces relief — and why relief is not a diagnosis.

Three separate things can make you feel better when you cut a food, and only one of them means the food is a genuine trigger:

1. True reaction. A real IgE allergy, a confirmed enzymatic intolerance (lactase deficiency), or a dose-dependent osmotic/fermentative effect (FODMAPs draw water into the gut and are fermented to gas). Removing it removes the mechanism. This is the real signal.
2. Confounded removal. Cutting "gluten" usually means cutting bread, pasta, beer, and most processed food simultaneously — you've changed a dozen variables, eaten less overall, and slowed down. Any of those could be the actual driver. The single named food gets the credit.
3. Expectation and attention. Elimination diets cannot be blinded in real life. You expect to feel better, you're attending closely to your gut, and symptom-reporting is highly suggestible. This is exactly why the 2024 blinded reintroduction study matters: when participants didn't know which FODMAP they were getting, the trigger list shrank and personalised.

Because relief is so easily produced by routes (2) and (3), the eliminated list always over-states the triggering list. That is the core mechanism the whole protocol is built around. Restriction generates a hypothesis ("something I cut matters"). Only reintroduction — ideally one variable at a time, with a washout, watching for dose-dependent symptom return — tests it. Skip reintroduction and you never learn which of the three mechanisms was operating, so you stay restricted on foods you tolerate fine.

Why time-boxing the restriction is biological, not arbitrary. FODMAPs and other fermentable fibres are food for your gut bacteria. Remove them and Bifidobacterium (and other fermenters) decline within weeks. The gut you are trying to calm is partly fed by the things you are removing — so an indefinite restriction phase progressively depletes the ecosystem you depend on, and the longer you stay, the more the diet's own side-effects can masquerade as the original problem. The 4-6 week ceiling exists to extract the diagnostic signal before the depletion sets in.

Risks And Contraindications

• Disordered eating / ARFID / orthorexia — the headline risk. Elimination dieting is a recognised on-ramp. The GI population is already elevated (14-57% ARFID in coeliac). Hard contraindication to unsupervised elimination: any current or past eating disorder, or a current trajectory of a shrinking food list with rising food anxiety. If the elimination is expanding rather than resolving, that is the disorder, not the diet, and it needs eating_disorder_body_image_diagnostic-level support — stop eliminating.
• Nutritional deficiency. Calcium, iron, B vitamins, fibre most commonly. Risk scales with breadth and duration of restriction and with lack of deliberate substitution. See micronutrient_deficiency_screening.
• Microbiome depletion. Bifidobacterium and other fermenters decline within weeks of FODMAP/fibre restriction. Reversible on reintroduction — another reason not to linger.
• Masking serious disease. Self-eliminating can suppress symptoms of IBD, coeliac, or malignancy and delay diagnosis. Red-flag symptoms (above) override any diet plan. Test for coeliac before removing gluten.
• Social and psychological narrowing. Restriction makes shared meals, travel, and spontaneity harder; the cost is real and compounds quietly.
• Children: elimination diets in children risk growth and development and should never be run unsupervised.
• The "felt better so I'll cut more" reflex is itself the risk pattern — flag it explicitly to the user. Feeling better is the cue to reintroduce, not to escalate.

Run correctly — single defined target, ≤6 weeks, structured reintroduction, supervision where indicated — the risks of a time-limited trial are modest. Nearly all the harm in this entry comes from the open-ended version.

Controversy

Position A — "Find your trigger foods and avoid them; restriction is the cure."
Common in functional-medicine and wellness spaces, often paired with IgG/sensitivity panels. Treats the eliminated list as a permanent answer. The relief is real but the interpretation is wrong: it conflates relief (easily produced by confounding and expectation) with proof of harm, and has no exit from restriction.

Position B — "Elimination diets are pseudoscience / harmful fad dieting."
A reaction to Position A's excesses and to the commercial testing industry. Over-corrects: it sweeps the genuinely first-line, evidence-backed protocols (low-FODMAP for IBS, empiric elimination for EoE) in with the fad version and can leave real sufferers without an effective tool.

The funding/bias dimension. Two opposite pulls. (1) Commercial sensitivity-testing and supplement/"gut-healing" industries profit from broad, indefinite restriction and from the anxiety that drives re-testing — their incentive is more elimination, longer. (2) The cleanest efficacy evidence here (the 2024 IBS network meta-analysis) was unfunded, and the strongest reintroduction RCT was academic — there is no commercial sponsor for the "reintroduce and stop restricting" message, because nobody sells reintroduction. The least-biased finding in the field is also the one with the least money behind it: most of what you cut, you can put back.

Realised Position.
An elimination diet is a time-limited diagnostic experiment with a mandatory reintroduction phase, not a way of eating. For a confirmed indication, run narrowly and time-boxed (≤4-6 weeks restriction), it is genuinely first-line. Its value is almost entirely in the reintroduction — that is where you learn your real, usually-short trigger list. Staying broadly restricted is the failure mode, not the success state. We will not recommend elimination off the back of an IgG panel, and we treat a growing restriction list and rising food anxiety as a clinical signal to stop and seek support, not as the diet working.

Cross-Pillar Connections

• Diet (ibs_diagnostic_lifestyle, gerd_diagnostic_lifestyle, histamine_intolerance_and_mast_cell_activation, small_intestinal_bacterial_overgrowth): these are the conditions where a targeted elimination trial is most often warranted; this entry is the how to run it safely companion to their what it is.
• Diet (diet_gut_microbiome, dietary_fiber_diversity_and_microbiome_health): the microbiome cost of restriction — fermentable fibre is bacterial food — is why restriction is time-boxed; fibre diversity is what reintroduction restores.
• Diet (micronutrient_deficiency_screening, individual_metabolism_variation_and_personalised_nutrition, diet_foundations_for_baseline): nutrient-gap monitoring during restriction; the personalisation phase is individual-variation in practice; foundations are what you return to.
• Mental (eating_disorder_body_image_diagnostic): the disordered-eating on-ramp is the headline risk; a shrinking food list with rising anxiety routes here.

What would change our mind

Falsifiability: explicit upgrade/downgrade criteria from source

We would upgrade the personalised/long-term-restriction position if: blinded, controlled trials showed that maintaining broad elimination long-term (rather than reintroducing to a minimal trigger set) produced durably better symptom and quality-of-life outcomes without nutritional or microbiome cost — i.e. if the reintroduction phase turned out to be discardable.

We would upgrade commercial sensitivity testing if: adequately powered, blinded RCTs showed IgG/IgA-guided elimination outperformed sham-guided elimination on objective outcomes. (Current consensus is the opposite.)

We would downgrade even the time-limited protocols if: long-term follow-up showed that a course of structured elimination/reintroduction measurably raised eating-disorder incidence net of benefit, or that the microbiome changes during restriction were not reliably reversible.

Industry bias note

Structural incentives the evidence base may reflect

Bias risk here is HIGH but unusual in direction — it pushes toward more restriction, not less.
• Who profits from broad, indefinite elimination: direct-to-consumer IgG/"food sensitivity" testing companies (each panel sells a restriction list and seeds re-testing), supplement and "gut-repair"/"leaky-gut" protocol sellers (restriction creates the deficiency the supplement then addresses), and specialty "free-from" product lines. Their incentive is a longer list followed longer.
• Who funds the honest message: essentially no one. "Reintroduce most of what you cut and stop restricting" has no product attached. Tellingly, the strongest efficacy evidence (2024 network meta-analysis) was explicitly unfunded and the cleanest reintroduction evidence was academic. This is the inverse of the usual pharma-bias pattern: here the under-monetised finding (reintroduction works; restriction should be brief) is the well-supported one, and the well-monetised claim (stay restricted, buy the test) is the weak one.
• Counter-direction check: the "elimination is all pseudoscience" backlash also has a constituency (general anti-wellness positioning) and over-states by erasing the legitimate first-line protocols. Both poles are wrong; the honest middle is "right tool, narrow indication, time-boxed, reintroduce."

Elimination Protocols