Whole-Food Diet vs Formula for Crohn’s: What the Trial Found

Real food was far easier to finish than eight weeks of formula

For children and young adults with mild-to-moderate Crohn's disease, exclusive enteral nutrition (EEN) has long meant replacing every meal with liquid formula for six to eight weeks. It can induce remission and has stronger evidence for mucosal healing than corticosteroids. But a treatment only works if people can complete it.

A 2025 clinician-blinded randomized trial found that 88% of participants assigned to a whole-food exclusion diet completed the eight-week protocol, compared with 52% assigned to EEN. The adjusted odds ratio was 7.7 (95% CI, 2.4–25; P<0.001). That is a large, statistically significant difference in tolerability—and the cleanest result in the study.

The trial does not prove that food-based treatment controls Crohn's as well as EEN. It found no statistically significant between-group differences in inflammatory markers, but it enrolled only 83 people and collected no endoscopic data. In Crohn's disease, feeling better and lowering fecal calprotectin are useful outcomes; neither confirms that intestinal ulcers have healed.

Still, the result changes the practical conversation. A diet excluding processed foods, gluten, red meat, and dairy may offer a more livable induction option for selected patients, particularly when formula-only treatment is unacceptable. The evidence supports discussion with an IBD team, not a solo attempt to replace prescribed care with a grocery list.

What the TASTI-MM trial actually tested

In the TASTI-MM randomized controlled trial, 83 biologic-naive participants aged 6 to 25 years with mild-to-moderate Crohn's disease received either eight weeks of Tasty&Healthy (T&H) or EEN. The average age was 14.5 years. T&H allowed whole foods while excluding processed food, gluten, red meat, and dairy; it used no formula supplementation and required no specific foods. EEN replaced solid food entirely with formula.

The completion gap was striking: 88% for T&H versus 52% for EEN. That difference matters beyond convenience. If a child stops EEN early because every meal comes from a carton, the theoretical efficacy of EEN cannot help that child during that course of treatment.

Symptomatic remission was less decisive. In the intention-to-treat analysis, remission occurred in 56% of the T&H group and 38% of the EEN group, with an adjusted odds ratio of 2.5 (95% CI, 0.98–6.3; P=0.1). The confidence interval includes no difference, and the trial did not find a statistically significant advantage for T&H.

The per-protocol analysis pointed numerically the other way: 67% remission with T&H and 76% with EEN (P=0.47). That comparison has a serious interpretive problem. EEN had much higher dropout, so the people who remained in its per-protocol group may have been unusually able to tolerate or respond to formula. The reversal does not prove EEN is superior, but it prevents any confident claim that T&H is superior for symptom remission.

Similar biomarkers are not proof of equivalent disease control

Both groups improved over eight weeks in fecal calprotectin, C-reactive protein, and erythrocyte sedimentation rate. The study found no statistically significant differences between diets on these inflammatory outcomes. For example, 34% of T&H participants and 33% of EEN participants reached fecal calprotectin below 250 μg/g (adjusted odds ratio 0.97; 95% CI, 0.37–2.6; P=0.84). A MINI score below 8 occurred in 44% and 31%, respectively (adjusted odds ratio 1.8; 95% CI, 0.7–4.5; P=0.33).

Those results are compatible with similar anti-inflammatory effects. They are also compatible with clinically meaningful differences that an 83-person trial could not detect. TASTI-MM was not an equivalence or noninferiority trial, so a non-significant result cannot establish that the treatments work equally well.

The missing outcome is endoscopic healing. A 2023 systematic review and meta-analysis of pediatric Crohn's studies found that EEN outperformed corticosteroids for mucosal healing. That is the benchmark a food-based alternative must meet before clinicians can treat it as interchangeable with EEN. TASTI-MM measured no endoscopic endpoints, so it cannot answer that question.

The microbiome result is intriguing, not a clinical endpoint

The trial also collected stool metagenomic data at baseline, four weeks, and eight weeks. Alpha diversity improved in the T&H group and declined in the EEN group, with higher species richness in T&H at both follow-up visits. Ruminococcus gnavus decreased with T&H and increased with EEN (q<0.001).

That direction fits broader IBD biology. A 2025 systematic review of microbiome studies reported that most included observational studies found lower microbial diversity in IBD than in healthy controls. A 2025 metagenomic study of newly diagnosed, treatment-naive IBD also found lower species-level diversity in Crohn's disease than in ulcerative colitis and healthy controls, with lower diversity associated with greater endoscopic severity in Crohn's disease.

But an eight-week change in diversity is not evidence that one diet prevents relapse, complications, or surgery. The microbiome is biologically relevant, but it is not a report card that can substitute for long-term clinical outcomes. Whether the T&H pattern persists during maintenance—or predicts durable remission—remains unknown.

Other CDED studies support a signal, with familiar dietary-trial problems

T&H is not identical to the Crohn's Disease Exclusion Diet (CDED), but both use whole-food exclusion strategies aimed at limiting components of a Western-style diet. A 2024 systematic review of CDED included nine studies and 352 participants, including three randomized trials. It found clinical-remission results in pediatric and adult Crohn's populations and better tolerability than EEN. Its evidence base, however, mixed randomized, uncontrolled, retrospective, and prospective studies. Consistency helps; it does not erase the limitations of the underlying designs.

A separate 2024 systematic review of solid-food diets, covering 14 studies and 11 randomized trials, also found a consistent remission signal for CDED. The authors judged nearly all included dietary randomized trials to have moderate-to-high risk of bias. Most did not provide the endoscopic-remission evidence needed to settle the EEN comparison.

Dietary restriction itself is not the treatment. In that review, low-FODMAP eating did not beat a sham diet for functional symptoms in quiescent Crohn's disease: Harvey-Bradshaw Index P=0.814, IBS Symptom Severity Score P=0.515, with no change in fecal calprotectin or CRP. That null result matters because it argues against a generic "eliminate foods and inflammation improves" story. The candidate benefit appears specific to the dietary pattern and clinical setting studied.

Evidence in older adults has started to appear. A 2026 real-world study of CDED in older adults with Crohn's disease and comorbidities reported that the diet was feasible and clinically effective. As an observational study, it extends feasibility beyond pediatric trial populations but cannot establish comparative efficacy against EEN.

Who should take this seriously—and what remains uncertain

The strongest current case for a whole-food exclusion approach applies to biologic-naive people aged 6 to 25 with mild-to-moderate Crohn's disease, the TASTI-MM population. It has a major adherence advantage over formula-only EEN and produced reductions in standard inflammatory markers that were not statistically distinguishable from EEN in that small trial.

It is not established for moderate-to-severe disease, biologic-exposed patients, or long-term maintenance. The T&H protocol also excluded broad food categories without standardized food lists or calorie targets, which makes consistent delivery across clinics harder than the label makes it sound.

A whole-food exclusion diet has earned its place as the best-supported food-based alternative to EEN for selected Crohn's patients. It has not yet earned the claim that it matches EEN for mucosal healing.

The next useful trial is straightforward in concept and demanding in execution: a larger randomized study with endoscopic outcomes, adequate follow-up, and a protocol that different centers can reproduce. Until then, EEN retains the stronger mucosal-healing evidence, while whole-food approaches offer a serious and far more tolerable option for patients who cannot complete formula therapy.